Intensive Care Medicine

, 34:93

21st ESICM Annual Congress

Open Access
Abstracts

DOI: 10.1007/s00134-008-1241-2

Cite this article as:
Intensive Care Med (2008) 34(Suppl 1): 93. doi:10.1007/s00134-008-1241-2

0352

Effect of CSF from Subarachnoid Hemorrhage-Patients with or Without Vasospasm on Endothelin-1 Sensitivity and Production on Isolated Rat Basilar Artery

B. Assenzio*1, E. L. Martin1, E. Stankevicius2, U. Simonsen2, I. Mastromauro1, S. Viberti1, M. M. Fontanella3, R. Boccaletti3, A. Ducati3, L. Mascia1

1Department of Anaesthesiology and Critical Care, University of Turin, Turin, Italy,

2Department of Pharmacology, Aarhus University, Aarhus, Denmark, 3Department of Neuroscience, University of Turin, Turin, Italy

Introduction

After subarachnoid hemorrhage (SAH) the presence of breakdown blood products is considered the initial trigger of vasospasm; however the subsequent pathogenic mechanisms remain unclear. The aims of this study were to: 1) compare the response to endothelin-1 (ET1) on isolated cerebral vessel incubated for 24 hours with CSF collected from patients with or without vasospasm, and 2) quantify ET1 production during incubation with CSF.

Methods

CSF was collected daily from patients with SAH and the occurrence of vasospasm was diagnosed by angiography. Rat basilar arteries were dissected and incubated with 5% of CSF from patients with or without vasospasm or artificial CSF. After 24 hours, the vessels were mounted on a wire myograph. The contractile response to ET1 was assessed and ET1 production was measured in the culture media of these incubated vessels.

Results

The vessels incubated for 24 hours with CSF from SAH patients with vasospasm showed an enhanced contractile response to ET1 compare to patients without vasospasm or artificial CSF. Incubation with CSF from both types of SAH patients induced a biphasic doseresponse curve, while artificial CSF resulted in a sigmoidal curve. The pEC50(1) and pEC50(2) of the dose response following incubation with CSF from vasospasm patients were significantly lower than non vasospasm indicating an increased sensitivity to ET1. Production of ET1 was significantly up-regulated in the arteries stimulated with CSF from vasospasm patients (1.87±0.36 pg/ml) compared to vessels stimulated with CSF from patients without vasospasm or artificial CSF (0.88 ± 0.04 and 0.83 ± 0.17 pg/ml respectively; p<0.05). Open image in new window

Conclusion

These results suggest that mediators specific to the CSF of patients with vasospasm alter the behaviour of normal cerebral vessels through modulation of the ET1 pathway.

Grant Acknowledgement

Regione Piemonte, Fondi ex 60%, University of Turin.

0353

Red Blood Cell Transfusion and Cerebral Oxygenation in Patients with Severe Traumatic Brain Injury

V. Padilla*1, Y. Corzia2, M. Jimenez2, V. Arellano2, C. Ferrandiz2, S. Leal-Noval2

1Intensive care, Hospital Universitario Virgen del Rocio, Sevilla, Spain, 2

Introduction

To investigate the long-term influence of erythrocyte transfusion on cerebral oxygenation (PtiO2) in patients with severe traumatic brain injury.

Methods

Prospective and observational study. Neurotrauma intensive care unit of trauma center level I. Sixty consecutive, hemodynamically stable patients with severe traumatic brain injury, pre-transfusion hemoglobin < 10 g/dL, non-bleeding and monitored through intracranial pressure and brain tissue partial oxygen pressure (PtiO2) catheters were included. All patients were transfused with 1–2 units of red blood cells.

Results

Ten sets of variables (pre-transfusion, end of transfusion, and 1, 2, 3, 4, 5, 6, 12 and 24 hours after transfusion) were recorded, including: PtiO2, cerebral perfusion pressure (CPP), end-tidal CO2, peripheral oxygen saturation, temperature, hemoglobin, lactate and PaO2/FiO2 ratio. Transfusion was associated with an increase in PtiO2 during a 6-hour period, with a peak at 3 hours (26.2%; P = 0.0001) in 78.3% of the patients. No relationship was observed between PtiO2, CPP and hemoglobin increments. The relative increment in PtiO2 at hour 3 was only correlated with baseline PtiO2; r2 0.166; P = 0.001. All of the patients with basal PtiO2<15 mmHg showed an increment in PtiO2 versus 74.5% of patients with basal PtiO2 ≥ 15 mmHg (P<0.01, hour 3).

Conclusion

Erythrocyte transfusion is associated with a variable and prolonged increment of cerebral tissue oxygenation in anemic patients with severe traumatic brain injury. Low baseline PtiO2 levels (<15 mmHg) could define those patients who benefit the most from erythrocyte transfusion.

Grant Acknowledgement

Supported by Spanish Goverment founds (FIS) PI 04296.

0354

Effect of Osmotherapy with Mannitol and Hypertonic Saline on Cerebral Oxygenation in Patients with Severe Traumatic Brain Injury and Refractory Intracranial Hypertension

M. Oddo*1, J. M. Levine1, S. Frangos1, E. Maloney-Wilensky1, E. MacMurtrie1, A. Kofke1, P. D. LeRoux2

1Departments of Neurosurgery and Neurocritical Care, 2Department of Neurosurgery, University of Pennsylvania Medical Center, Philadelphia, United States

Introduction

To analyze the effects on brain tissue oxygen pressure (PbtO2) of mannitol and hypertonic saline (HTS) in patients with severe traumatic brain injury (TBI) and refractory intracranial hypertension.

Methods

Twelve consecutive patients with severe TBI (GCS ≤8), admitted to our Neuro-ICU, monitored with continuous PbtO2, and treated with mannitol (25%, 0.75 g/kg) and HTS (7.5%, 250 ml) for recurrent episodes of elevated intracranial pressure (ICP>20 mm Hg), were retrospectively identified from a prospective observational database. PbtO2, ICP, mean arterial pressure (MAP), cerebral perfusion pressure (CPP), central venous pressure (CVP) and cardiac output, were recorded every 30 minutes for 120 minutes after osmotherapy.

Results

A total of 42 episodes of intracranial hypertension, treated with mannitol (n = 28 boluses) and HTS (n = 14 boluses), were analyzed. Baseline PbtO2, ICP, CVP, cardiac output, MAP, and CPP were comparable. Mean CPP and MAP did not differ significantly at each time-point analyzed. HTS treatment was associated with a significant improvement of cerebral oxygenation in patients with severe TBI (Table 1). HTS was also associated with better ICP control and improved systemic hemodynamics.

VARIABLE

MANNITOL (n=28)

HTS (n=14)

P VALUE

PbtO2 (mm Hg) baseline

28.6 ± 11.0

27.4 ± 13.7

0.77

30 min

27.7 ± 11.8

34.2 ± 17.6

0.18

60 min

26.0 ± 11.0

36.3 ± 17.1

0.04

120 min

26.9 ± 11.2

40.4 ± 17.4

0.007

ICP (mm Hg) baseline

28 ± 7

27 ± 8

0.73

30 min

20 ± 7

17 ± 7

0.20

60 min

21± 9

15 ± 6

0.03

120 min

22 ± 7

15 ± 6

0.002

CVP (cm H2O)

6 ± 3

7 ± 5

0.20

30 min

6 ± 3

9± 6

0.08

60 min

5 ± 3

8 ± 4

0.04

120 min

5 ± 2

8 ± 5

0.04

CO (L/min) baseline

6.4 ± 1.3

6.4 ± 1.8

0.95

30 min

6.0 ± 1.0

7.6 ± 1.4

0.002

60 min

6.2 ± 1.2

7.7 ± 1.7

0.007

120 min

6.1 ± 1.1

7.5 ± 1.4

0.004

Conclusion

In patients with severe TBI and refractory intracranial hypertension, osmotherapy with hypertonic saline solutions is associated with a significant improvement of cerebral oxygenation and systemic hemodynamics.

Grant Acknowledgement

SICPA Foundation, Lausanne, Switzerland.

0355

Cerebral Oxygenation in Severe Traumatic Brain Injury and Storage time of Transfused Blood

C. M. Ferrándiz*, V. Padilla, M. V. Arellano, Y. Corcia, M. Jiménez, S. R. Leal

Unidad de Cuidados Críticos y Urgencias, Hospital Universitario Virgen del Rocío, Sevilla, Spain

Introduction

Prolonged erythrocyte storage time might reduce the efficacy of transfusion. In this study, the effects of transfusion of erythrocytes with four different storage periods (<10 days, n = 18; 10–14 days, n = 15; 15–19 days, n = 17; and >19 days, n = 16 patients) on brain tissue oxygen tension (PtiO2) in stable male patients with severe traumatic brain injury, were investigated during a 24 hours follow-up period.

Methods

Prospective, observational study carried out at Neurotrauma critical care unit of a University Hospital. Sixty-six male, non-bleeding, hemodynamically stable anemic patients (hemoglobin < 9.5 g/dL) with Glasgow Coma Scale < 9 were included. PtiO2, cerebral perfusion pressure, mean arterial pressure, intracranial pressure, peripheral oxygen saturation, CO2 pressure at the end of expiration and intracerebral temperature were recorded in all patients at baseline, immediately after the completion of transfusion and 1, 2, 3, 4, 5, 6, 12 and 24 hours post-transfusion.

Results

All four groups were homogeneous with respect to multiple baseline variables, except for storage time of transfused erythrocytes (P<0.0001). There was a significant short-lasting (3–4 hours) increase in PtiO2 values after transfusion of erythrocytes stored for <10 days, 10–14 days, or 15–19 days, when compared to those at baseline. In contrast, no significant changes in PtiO2 were observed after transfusion of erythrocytes stored >19 days. Not inter group differences were observed.

Conclusion

Transfusion of erythrocytes increased cerebral oxygenation in patients with severe traumatic brain injury, except in those transfused with erythrocytes stored for more than 19 days.

Grant Acknowledgement

FIS 04/296.

0356

500ml of Blood Loss Does Not Decrease Non Invasive Tissue Oxygen Saturation (Sto2) Measured by Near Infrared Spectroscopy in Healthy Adults

V. Jeger*1, S. Fontana2, N. Baur2, M. Luginbuühl3, A. Ferretti1, H. Zimmermann1, A. Exadaktylos1

1Emergency Medicine, University Hospital Inselspital, 2Blood Donation Service, SRK,

3Anesthesia, University Hospital Inselspital, Bern, Switzerland

Introduction

In the past, non-invasive tissue oxygenation (StO2) measurement by near infrared spectroscopy (NIRS) and iatrogenic ischemic stress test (IIST) has been used to detect under-perfusion in shock patients, mainly in ICU. IIST has not been performed in a prospective controlled setting with a standardized blood loss. Therefore to investigate if blood loss of 500ml is leading in healthy blood donors to detectable changes by NIRS.

Methods

Prospective evaluation of 20 healthy female blood donors (500ml, ca.10% of circulating blood volume) with a body weight between 50 and 65kg. StO2 was continuously measured on the thenar eminence with InSpectra® (Model 650, Hutchinson Technology, Minn. USA). From the StO2 values and the tissue hemoglobin index (THI) HbO2 was calculated. Blood pressure, StO2 and HbO2 values at the beginning and immediately after blood donation(BD) were calculated. The ischemic stress test was performed by occluding the brachial artery by a sphygmomanometer cuff inflated 50mmHg over systolic blood pressure for 3 minutes. Pre- defined variables were calculated out of the recorded StO2 and HbO2 values. To detect a difference in variable before and after blood donation, the non-parametric Wilcoxon matched-pairs signed-ranks test was used.

Results

Median age was 30.5 yrs (range 19–62), median body weight 58 kg (range 50–65). There was a significant decrease in systolic blood pressure (BP) from 121.8 ± 9.9 (mean ± SD) to 112.8 ± 10.7 (p< 0,0001).

However variables measured by NIRS did not change significantly: StO2 79.3 ± 5.0% to 78.6 ± 5.3% (p = 0.35); HbO2 1067.5 ± 162.8 to 1056.8 ± 176.6; (p = 0,44). Variables measured after ISST did not change significantly, too: StO2 slope of recovery 3.0 ± 0.8%/sec to 3.1 ± 0.6%/sec (p = 0,26); HbO2 slope of recovery 59.2 ± 17.9/sec to 61.6 ± 13.6/sec (p = 0.6); delta StO2 13.7 ± 3.1% to 14.6 ± 4.1% (p = 0.4). There was a trend in delta HbO2 429.5 ± 153.9 to 468.5 ± 129.0 (p = 0.058). The study was underpowered to show a possible significant change.

Conclusion

Although systolic blood pressure decreased significantly after blood donation of 500ml, StO2 did not show any significant change in baseline. Furthermore, there were no significant changes after ischemic stress tests. Only HbO2 increased nearly significantly after blood donation. In order to understand the physiological background of these findings, more systematic research is needed.

Oral Presentations ICU admission: 0357–0362

0357

Are “Off Hours” Icu Admission Associated with Higher Mortality?

R. Antunes*1, E. Diz1, O. Ribeiro2, A. Costa-Pereira2, T. Cardoso1, A. Carneiro1

1UCIP, Hospital Geral Santo Antońio, 2Servic'o de Bioestatística e Inform'tica Médica, FMUP, Porto, Portugal

Introduction

Staffing levels in acute care hospitals tend to be lower during evening, night, weekend and holyday shifts (referred to as ‘off hours’) even in intensive care units (UCI's). Only sparse data are available on the relationship between time of admission and mortality, with the few published reports yielding contraction results. We sought to determine whether ICU mortality rates differ according to the time of admission: day or “off-hours”.

Methods

Retrospective cohort study on patients admitted to our mixed ICU from December 2001 to November 2007. For each admission, age, sex, category of admission, date and hour of admission, SAPS II, ICU and hospital lengths of stay and ICU and hospital mortalities were recorded. Patients were grouped according to their day and time of admission in weekday day shifts and off hours (from 8:30pm to 8:29am for night shifts and from Friday 8:30pm to Monday 8:29am for weekends) and compared using univariable and multivariable analyses. The main outcome measure was in-hospital mortality.

Results

Of the 1031 patients admitted during the study period, 775 (75%) were admitted within the “off hours” period (34% over night and 41% during weekend). No significant difference was found, between both groups, regarding age, gender and SAPS II score. Regarding category of admission a great proportion of trauma patients were admitted “off hours” (34% vs 24%, p=0.01). Comparing the group admitted during daytime and ‘off-hours’ no significant difference was found regarding ICU mortality (28% vs 29%, p=0.662) or hospital mortality (33% vs 34%, p=0.715). In a logistic regression model with ICU mortality as end point, adjusted to age, gender, SAPS II, category of admission and ICU admission time, for patients admitted during daytime we found an OR =0.886 (95% CI =0.620–1.265).

Conclusion

Admission in ICU is more common during “off hours”. The results of the few studies that examined the effect of the time of admission on outcome were contradictory probably reflecting the different organizations in the care of critical ill patients and type of ICU. In our unit, there is a consultant on site 24h/day, every day and the nurse-to-patient ratio is at least 1:2. As long as adequate staffing is maintained, and necessary diagnostic and therapeutic modalities are available, weekend and night admission should not be associated with differences in patients outcome.

0358

Off Hour Admission to the Intensive Care is not Associated with Increased Mortality

I. A. Meynaar*1, J. I. van der Spoel2, J. H. Rommes3, M. van Spreuwel-Verheijen1, R. J. Bosman2, P. E. Spronk3

1ICU, Reinier de Graaf Hospital, Delft, 2ICU, Onze Lieve Vrouwe Gasthuis, Amsterdam,

3ICU, Gelre Hospitals, Apeldoorn, Netherlands

Introduction

Off hour admission to the ICU has been associated with increased as well as with decreased mortality. We evaluated whether ICU admission during off hours affects hospital mortality.

Methods

All 3 ICUs are mixed surgical/medical intensivist led units in non-academic teaching hospitals. Intensivists are available 24/7 and intensivists on duty have no obligations outside the ICU. During daytime the intensivist is available at the bedside in the ICU and rounds are made at least twice daily. During off hours the intensivist is generally not present in the hospital, but is available within 15 minutes. Off hours were defined as the period between 10 PM and 8 AM during weekdays and between 6 PM and 9 AM in the weekend. All patients fulfilling APACHE II criteria (age>15, no burns, no cardiac surgery, LOS>8 hrs) admitted in 2004–2007 were included. A standardised dataset for calculation of APACHE II and SAPS II scores including standardised mortality ratios (SMR) was prospectively collected.

Results
A total of 6725 patients were admitted in the study period. Table 1 shows that off hours patients are sicker than daytime patients. Hospital mortality is higher in off hour patients, but SMR is not significantly different (Table 2). In logistic regression analysis off hour admission is not an independent risk factor for increased mortality (Table 2).
Table 1

Patients

 

Daytime

Off hours

p

Number of patients

4553

2172

 

Mean age (SD)

65.9 (15.3)

63.5 (18.0)

<0.001

Mean APACHE II (SD)

16.3 (8.6)

17.9 (9.0)

<0.001

Mean SAPS II

35.7 (18.3)

40.6 (18.2)

<0.001

Table 2

Hospital Mortality

 

Daytime

Off hours

Crude OR (95% CI)

Adjusted OR (95% CI)*

Hospital mortality

767 (16.8%)

469 (21.6%)

1.36 (1.20–1.55)

1.13 (0.97-1.31)

APACHE II SMR (95% CI)

0.65 (0.60–0.71)

0.68 (0.61–0.75)

  

SAPS II SMR (95% CI)

0.68 (0.63–0.74)

0.69 (0.62–0.76)

  

*including age and APACHE II expected mortality in the model

Conclusion

Off hour admission to the ICU is associated with increased mortality. This is seems to be explained though by higher illness severity in patients who were admitted off hours.

0359

The Effect on In-Hospital Mortality of Delayed Admission to Critical Care

L. Parker*1, D. Raw2, N. Bhuiyan2

1Critical Care, 2Department of Anaesthesia, University Hospital Aintree, Liverpool, United Kingdom

Introduction

In the United Kingdom there is evidence of insufficient critical care facilities (1) which leads to delays in intensive care unit admission (2) and inter-hospital transfers. Critically ill patients are at increased risk of morbidity and mortality during transport (3) and therefore, wherever possible, in our institution we use theatre recovery as a critical care overflow facility until an internal bed becomes available. An audit in our institution in 2007 showed an increased mortality for patients cared for in recovery compared with those admitted directly to critical care(4). However, this did not take into account illness severity.

Methods

Patients admitted indirectly to critical care via theatre recovery over a two year period from April 2006 – March 2008 were identified. Their APACHE II score, hospital outcome and predicted mortality were compared with those patients admitted directly to ICU.

Results

Year

Source of admission

Number

Mortality %

Mean APACHE II

Predicted Mortality %

P value Chi square

2006/07

Recovery

68

43

14.5

27.5

0.012

2006/07

Direct

757

30.1

16.7

32.8

0.12

2007/08

Recovery

55

30.1

14.6

26.4

0.44

2007/08

Direct

805

25

16.1

30.2

0.002

Conclusion

Patients admitted to critical care via the recovery overflow facility have a higher mortality than those admitted directly. They also have a higher mortality than those patients transferred between UK ICUs (hospital mortality 26%/mean APACHE II 16.3/predicted mortality 31.5%)(5). This cannot be explained by illness severity. More ICU beds should be made available and care levels increased in recovery. Early patient transfers should be considered in the absence of an immediately available critical care bed.

0360

Recognition of the Deteriorating Patient Reduces Unplanned Admissions to the Intensive Care Unit

I. A. Mitchell*1, C. Van Leuvan1, B. Avard1, N. Slater2, R. Berry3, P. Lamberth4, H. McKay3

1Intensive Care, 2Resuscitation, The Canberra Hospital, 3Patient Safety and Quality Unit, ACT Health, 4Intensive Care, Calvary Healthcare, Canberra, Australia

Introduction

Unplanned intensive care unit (ICU) admissions from the ward are often triggered by the failure to recognize and appropriately manage deteriorating patients. This is evidenced by delays in admission to the ICU (1), unexpected referrals to intensive care and unexpected deaths often being preceded by significant physiological disturbances (2). The ability to detect early deterioration in patients allows early appropriate intervention. Early intervention can reduce unplanned admissions to the ICU and unexpected deaths (3). The aim of the project was to reduce the number of unplanned ICU admissions by introducing an intervention onto four wards which included: — The implementation of a new education program “COMPASS”, the introduction of a new observation chart and the use of a track and trigger system.

Methods

A four month prospective, controlled, before and after intervention trial in four wards at a district and tertiary hospital was undertaken. The ethics committee waived the need for consent. Demographic data collected on all admissions included: age, gender, and admission diagnosis. Outcome data included: frequency of observations, unplanned ICU admissions, cardiac arrests, medical emergency team reviews and hospital outcome.

Results

There were a similar number of ward admissions for the two periods (1196 and 996 respectively). The frequency of vital sign measurement increased for respiratory rate (2.7 to 4.7 per day, p<0.0001) and arterial oxygen saturation (4.3 to 6.8 per day, p=0.02). There was a decrease in unplanned ICU admissions (21 to 5, p=0.005) and cardiac arrests (4 to 0, p=0.03). Medical emergency team reviews increased from 27 to 51 (p < 0.001) and the number of hospital deaths decreased from 35 to 16 (2.9% to 1.6% of all first admissions, p=0.05).

Conclusion

A three pronged approach to the recognition of the deteriorating patient increases their detection and appears to reduce the number of unplanned admissions to ICU. This simple intervention has the potential to improve patient outcome.

Grant Acknowledgement

Supportive grant from ACT Health.

0361

The Inclusion of Pulse Oximetry as a Component of an Early Warning Score Improves its Ability to Predict Mortality Following a Cardiac Arrest Call or Unplanned Intensive Care Unit (ICU) Admission

S. L. Toft*1, S. E. Clark2, P. D. G. Alexander3

1Resusitation Training Department, 2Critical Care Outreach, 3Consultant, Acute Intensive Care Unit, University Hospital of South Manchester, Manchester, United Kingdom

Introduction

It is well recognised that ward based early warning score systems can be used to predict outcome (1). In 2007 the UK's National Institute for Health and Clinical Excellence (NICE) recommended (2) that physiological scoring systems should include the following variables: heart rate (HR), systolic blood pressure (BP), respiratory rate (RR), temperature (T), assessment of consciousness level (Alert, Verbal response, Pain response, Unresponsive — AVPU) and oxygen saturation (SpO2). The aim was to assess several published scoring systems against our existing Modified Early Warning Score (MEWS) which includes HR, BP, T, RR, AVPU and urine output (UO).

Methods

For a period of 3 months from January to April 2008 physiological data were collected for all unplanned ICU (uICU) admissions and cardiac arrest (CA) calls in our institution. Up to 24 hours of physiological parameters prior to the event and the documented MEWS were taken from ward observation charts and collated on a database. The maximum pre-event value for each patient was then compared to corrected MEWS (recalculated from documented physiology), PMEWS (HR, BP, RR, T, AVPU, SpO2), NMEWS (HR, BP, RR, T, AVPU, SpO2, FiO2) and SEWS (HR, BP, RR, T, AVPU, SpO2, UO).

Results
Hospital discharge data were available on 86 patients (overall mortality 45.3%) with 44 CA calls (mortality 54.5%) and 42 uICU admissions (mortality 38.5%). Results are reported as median (IQR) and P values from Mann Whitney test. Receiver Operating Characteristic (ROC) curve areas were different among the scoring systems with the best prediction being those containing a SpO2 score.
 

Alive

Dead

P

ROC Area

95% CI

Documented MEWS

4 (2–6)

4.5 (2–8)

0.094

0.62

0.49 to 0.74

Corrected MEWS

3 (1–5)

4 (2–6)

0.175

0.60

0.46 to 0.73

PMEWS

3 (2–6)

6 (3-7.5)

0.003

0.71

0.59 to 0.83

NMEWS

5 (2–7)

6.5 (4–9)

0.022

0.66

0.53 to 0.79

SEWS

3 (1–4.5)

5 (2.5–7)

0.008

0.69

0.56 to 0.81

Conclusion

The addition of SpO2 to early warning score systems improve their ability to predict hospital discharge mortality following a cardiac arrest call or unplanned ICU admission. Further work is ongoing regarding defining appropriate weightings for each sub-group.

0362

The Characteristics of Patients Receiving a Repeat Medical Emergency Team Activation

P. Calzavacca*1, E. Licari2, A. Tee2, R. Bellomo2

1ICU, Ospedale Uboldo, Cernusco sul Naviglio, Italy, 2ICU, Austin hospital, Heidelberg, Australia

Introduction

Several studies have evaluated epidemiological characteristics of patients receiving a Medical Emergency Team (MET) call. However, no published data is available on patients receiving more than one MET call (2nd MET call patients) and how they differ from those with a single activation. This study aims to analyze the epidemiology of these 2nd MET call patients.

Methods

A retrospective observational study carried out in a single tertiary care Australian teaching hospital. We retrieved information about patient demographics, reasons for the MET call, procedures performed by the MET and the outcome of the MET patients during the hospital stay. For 2nd MET call patients, we referred to their characteristics presented at the 1st MET activation. The Ethics committee of the hospital waived the need for informed consent. For statistical analysis we used SPSS version 13.0 for Windows. Data are given as mean (and median) ± standard deviation (SD) for continuous variables or as percentage for categorical variables. Chi-square, Wilcoxon Sum Rank or Mann-Whitney U tests were used as indicated to compare different subgroups, p0.05 was considered statistically significant.

Results
We analyzed 2237 MET calls activated in 1667 patients during the period 16 August 2005 – 15 August 2007. Three hundred eighty-two patients (22.9%) received more than 1 MET call and up to 13 times during the hospital stay (mean 2.53 calls, median 2). Table 1 summarizes our major epidemiological findings for the entire population as well as the two subgroups of single and 2nd MET call patients. The mean age of the study group was 69.8 years (±16.8), mean length of stay (LOS) 23 days (±34.4), 55.4% of the patients were men and 46% were post-surgical. The MET instituted Not For Resuscitation (NFR) orders in 12.5% of the patients reviewed and 212 unplanned Intensive Care Unit (ICU) admissions occurred during the study period. Overall, in-hospital mortality was 34.3%. Single MET call patients tend to already have an existing NFR order compared to 2nd MET call patients. Second MET call patients were more likely to be surgical and dysrhythmias were a more common trigger. They also had a 50% longer LOS and a 30% increase in mortality.
Table 1

Epidemiological and Outcome Characteristics of Overall Population, Single and 2nd Met

 

Overall

Single MET call group

2nd MET call group

p value

Age*

69.8 (74) ± 16.8

69.5 (74) ± 17.1

70.5 (74) ± 15.5

0.287

Sex (male)

923/1667 (55.4%)

715/1287 (55.6%)

208/380 (54.7%)

0.778

Surgical

766/1667 (46%)

560/1287 (43.5%)

206/380 (54.2%)

< 0.001

NFR before

377/1667 (22.6%)

305/1287 (23.7%)

72/380 (18.9%)

0.052

NFR after

209/1667 (12.5%)

168/1287 (13.1%)

41/380 (10.8%)

0.242

Unplanned ICU admission

212/1667 (12.7%)

171/1287 (13.3%)

41/380 (10.8%)

0.199

LOS*

23 (15) ± 34.4

20.5 (13) ± 34.3

31.7 (22) ± 33.2

< 0.001

Hospital mortality

571/166 (34.3%)

409/1287 (31.8%)

162/380 (42.6%)

< 0.001

* For continuous variables mean (median) ± Standard Deviation are given

Conclusion

Patients requiring a 2nd MET call have a higher in-hospital mortality and longer LOS compared to those receiving only a single activation Delayed institution of NFR orders could be one of the possible reasons for this difference.

Oral Presentations Molecular mechanisms of acute lung injury: 0363–0368

0363

Acute Lung Injury: Apoptosis in Effector and Target Cells of the Upper and Lower Airway Compartment Upon Endotoxin and Hypoxia-Induced Injury

B. V. Roth Zgraggen*1, J. Tornic1, L. Reyes1, C. Booy1, D. R. Spahn2, B. Beck-Schimmer3

1Institute of Physiology and Center for Integrative Human Physiology, University of Zurich,

2Institute of Anesthesiology, University Hospital Zurich, 3Institute of Anesthesiology and Institute of Physiology and Center for Integrative Human Physiology, University of Zurich and University Hospital Zurich, Zurich, Switzerland

Introduction

Apoptosis has been considered as an underlying mechanism in acute lung injury (ALI) and acute respiratory distress syndrome (ARDS). Endotoxin-induced injury is an experimental in vitro and in vivo model closely resembling ALI and ARDS. Previous studies show that hypoxia can exert a pro-inflammatory effect on the lung, and might be an essential co-factor for ALI/ARDS. Therefore, apoptosis was determined in effector cells (alveolar macrophages, AM and neutrophils) and target cells (tracheobronchial and alveolar epithelial cells) of the respiratory compartment, measuring caspase-3 activity upon endotoxinand hypoxia-induced injury.

Methods

Rat alveolar epithelial cells (AEC), rat primary tracheobronchial epithelial cells (TBEC) (1), AM and neutrophils were placed in a hypoxic incubator with 5% oxygen for 4 hours, control cells were left at 21% oxygen. Some cells were stimulated with lipopolysaccharide (LPS, 20 ug/ml) for 4 hours. Camptothecin was used as a positive control for induction of apoptosis. To determine apoptosis rate, Caspase-3 activity was measured by the proteolytic cleavage of the fluorogenic caspase-3 substrate Ac-Asp-Glu-Val-Asp-AMC. ANOVA was performed to assess the statistical significance of differences. P values <0.05 were considered significant.

Results

Apoptosis rate of AM was increased by 75% (p<0.05) under stimulation with LPS, while hypoxia did not affect caspase-3 activity. Neutrophils, however, showed a decreased apoptosis rate of 39% upon hypoxia (p<0.05). However, LPS did not change caspase-3 activity. TBEC experienced an enhanced caspase-3 activity upon LPS stimulation (increase of 121%, p<0.05) with no change under hypoxia. Similar results were found in AEC with a LPS-induced increase of apoptosis rate of 45% (p<0.05) and an unaffected caspase-3 activity under hypoxia.

Conclusion

Our data demonstrate that the three cell types from the upper and lower respiratory compartment AEC and TBEC as well as AM show the same pattern of apoptosis upon exposure to hypoxia and endotoxin. The apoptotic answer of neutrophils, however, is different. Our results support findings with increased apoptosis rate in pulmonary cells (2). The functional implications of these inflammatory answers have further to be analyzed.

Grant Acknowledgement

Olga Mayenfisch Stiftung, Zurich, Switzerland

0364

Decreased Ventilator-Induced Lung Injury in Matrix Metalloproteinase-8 Deficient Mice

G. M. Albaiceta*1, D. Parra1, E. Garcá Prieto1, A. Astudillo2, A. Fueyo3, X. S. Puente4, F. Taboada5, C. Lopez-Otin4

1Intensive Care Unit, 2Anatomia Patoloǵica, Hospital Universitario Central de Asturias,

3Biologia Funcional, 4Bioquimica y Biologia Molecular, 5Medicina, Universidad de Oviedo, Oviedo, Spain

Introduction

Matrix metalloproteinases (MMP) modulate extra cellular matrix turnover and cytokine processing after different injuries. MMP-8 is released by neutrophils at the sites of inflammation. We hypothesized that MMP-8 could play a relevant role in ventilator-induced lung injury (VILI).

Methods

Mice lacking MMP-8 (Mmp8−/−) and their wild type counterparts were ventilated for 2 hours using low (peak 15 cmH2O, PEEP 2 cmH2O) or high (25 cmH2O, PEEP 0 cmH2O) pressures. Lung injury was assessed by gas exchange, wet/dry weight ratio, histology and protein content, cells and myeloperoxidase activity in bronchoalveolar lavage fluid (BALF). Levels of MMP-2, -8 and -9 were measured by gelatin zymography and western blotting. Collagen, cyto and chemokines (IL1eta, IFNgamma, IL-4, IL-10, MIP-2 and LIX) were measured in lung tissue and BALF. Data are given as mean±SD. Variables were compared using a two-way ANOVA, with a significance level of 0.05.

Results
MMP-8 increased in lung tissue and BALF in wild type mice after high pressure ventilation. Mm8−/− mice were more resistant to VILI than their counterparts (Table 1). There were no differences in lung collagen, MMP-2 or -9 between genotypes. Absence of MMP-8 was related to lower levels of IFNgamma and LIX in basal conditions. After injurious ventilation, there was a dampened increase in MIP-2 with significant increases in IL-4 and IL-10 in knockout mice, suggesting a shift to an anti-inflammatory response.
Table 1

Measurements of Lung Injury. MPO: Myeloperoxidase; PIP: Peak Inspiratory Pressure

 

Mmp8+/+ PIP 15

Mmp8−/− PIP 15

Mmp8+/+ PIP 25

Mmp8−/− PIP 25

PaO2/FiO2

408±59

386±46

290±49#

345 ±60*

Wet/Dry weight

4.10±0.75

4.48 ±0.24

6.00±0.69#

5.12±0.96#

Histological score

0.67±0.82

0.75±0.53

3.1±1.24#

1.44±0.73*

BALF proteins

6.3±2.1

5.1±0.9

36.3±12.1#

21.8±11.2*

BALF cells

0.6±0.2

0.6±0.2

2.5±1.3#

0.9±0.5*

BALF MPO

4.0±3.0 1.6±1.3

7.6±2.3#

2.3±1.0*

 

# p < 0.05 vs PIP 15, * p < 0.05 vs Mmp8+/+

Conclusion

MMP-8 promotes lung inflammation after high pressure ventilation. Absence of this enzyme results in an anti-inflammatory response that protects against VILI. Grant Acknowledgement. FIS PI-07/0597.

0365

Phosphoinositide-3 Kinase Gamma Required for Lps-Induced Transepithelial Neutrophil Trafficking in the Lung

J. Reutershan*1, M. S. Saprito2, D. Wu3, T. Ru:ckle4, K. Ley5

1Dept. of Anaesthesiology and ICM, University Hospital Tu:bingen, Tu:bingen, Germany,

2Robert M. Berne Cardiovascular Research Center, University of Virginia, Charlottesville,

3Department of Pharmacology, Yale University School of Medicine, New Haven, United States, 4Department of Operational Excellence Research, Merck Serono International S.A., Geneva, Switzerland, 5Division of Inflammation Biology, La Jolla Institute for Allergy & Immunology, La Jolla, United States

Introduction

In acute lung injury (ALI), excessive recruitment of polymorphonuclear leukocytes (PMNs) to the lung aggravates pulmonary inflammation and promotes organ failure. Phosphoinositide 3-kinase |gamma (PI3K|gamma) is a critical mediator of directional cell movement and has been implicated as an attractive target in experimental ALI. Here, we sought to characterize the role of PI3K|gamma in mediating the different steps of PMN trafficking in the lung. In addition, the contribution of endothelial and leukocytic PI3K|gamma was studied in bone marrow chimeras.

Methods

In wildtype and for PI3K|gamma gene-deficient mice (PI3K|gamma-/-), ALI was induced by inhalation of LPS. PMN accumulation in the different compartments of the lung (intravascular space, interstitium, alveolar space) was determined by flow cytometry. Chimeric mice were created by transfer of bone marrow between wildtype and PI3K|gamma-/- mice. We evaluated the small molecule PI3K|gamma inhibitor AS-605240 for its efficiency to block PMN transmigration in vivo and in vitro. LPS-induced microvascular permeability was determined by the extravasation of Evans blue.

Results

LPS-induced PMN adhesion to the pulmonary endothelium and transendothelial migration into the lung interstitium was enhanced in PI3K|gamma gene-deficient (PI3K|gamma-/-) mice. However, transepithelial migration into the alveolar space was reduced in these mice significantly. When irradiated PI3K|gamma-/- mice were reconstituted with bone marrow from wildtype mice, migratory activity into the alveolar space was restored only partially, suggesting a proinflammatory role for endothelial PI3K|gamma. A small molecule PI3K|gamma inhibitor reduced chemokine-induced PMN migration in vitro when PMNs but not when endothelial cells were treated. The inhibitor also reduced LPS-induced PMN migration in vivo but did not affect pulmonary microvascular permeability.

Conclusion

We conclude that PI3K|gamma is required for transepithelial but not for transendothelial migration in LPS-induced lung injury. Pharmacologic inhibition of PI3K|gamma activity preferably on hematopoietic cells may be effective at curbing inflammatory lung tissue damage.

Grant Acknowledgement

This study was supported by the German Research Foundation (grant RE 1683/3-1 to J. Reutershan), by NIH grant HL73361 to K. Ley.

0366

Recombinant Human Dnase Reduces Duration of Ventilation in Non-Surgical But Not in Surgical Mechanically Ventilated Intensive Care Patients

N. Deschner*1, R. Vonthein2, W. Brehm3, J. Riehtmueller3

1Department of Anaesthesiology and ICM, University Hospital Tuebingen, 2Department of Medical Biometry, 3Department of Pediatrics, University Hospital Tuebingen, Tuebingen, Germany

Introduction

RhDNase is an established part in the treatment of children with cystic fibrosis [1] and in children following cardiac surgery rhDNAse reduced duration of ventilation [2]. The aim of this study was to investigate whether rhDNase is able to reduce the duration of ventilation in adult mechanically ventilated intensive care patients.

Methods

A double-blind, placebo-controlled, randomised, multi-centre national trial was conducted after approval of local ethics committees. Patients were stratified in a surgical and a non-surgical group. The trial was started within 48h after start of mechanical ventilation and was terminated when weaning was successful or after 21 days. Patients in the treatment group received 2,5ml of rhDNase endotracheally twice a day. Patients in the placebo group received normal saline.

Results

In this study, 123 surgical and 162 non-surgical patients were included. Characteristics like gender, weight, APACHE score, chronic pre-existing diseases and prevalence of COPD were distributed equally in both groups. In non-surgical patients more smokers were randomized to the rhDNase group and patients in the NaCl group had a lower GCS score. Acute burn patients were randomized to the rhDNase group only. In the rhDNase group, 12 patients (2 surgical) died versus 16 (4 surgical) in the placebo group. In surviving surgical patients median duration of ventilation was 16.6 days (CI 11.5 to 21 days) in the rhDNase and 11.7 days (CI 8.4 to 15.6 days, p=0.39) in the placebo group. In surviving non-surgical patients median duration of ventilation was 7.8 days (CI 6 to 9.3 days) in the rhDNase and 12.6 days (CI 7.9 to 16.9 days; p=0.038) in the placebo group. Pulmonary function parameters like oxygenation index, peak pressure and compliance and SOFA and CPIS scores did not differ. Cost-effectiveness analysis showed, that 9 days of treatment with rhDNase migth save 2,9 days on the ventilator which might save EUR 3000 in the treatment of an average nonsurgical ICU patient.

Conclusion

In adult non-surgical intensive care patients, rhDNase significantly shortens the duration of ventilation. This effect is not seen in surgical patients. Particulary patients with pneumonia respond favourably to the treatment with rhDNase.

Grant Acknowledgement

This study was supported by a restricted grant from Roche Pharma AG, Grenzach-Wyhlen, Germany.

0367

Inhibition of Acute Lung Injury by Dexmedetomidine

D. E. Schwartz*1, G. Pirc1, M. Castellon2, R. Minshall2

1Anesthesiology, 2Pharmacology, University of Illinois Chicago, Chicago, United States

Introduction

Acute lung injury (ALI) and the more severe acute respiratory distress syndrome (ARDS) are life threatening conditions characterized by inflammation of lung parenchyma and vascular leakage leading to impaired gas exchange and hypoxemia. The triggering mechanisms for ARDS/ALI are not completely understood, and current treatments provide supportive care rather than target specific mechanisms that have a definitive effect upon outcome. A recent study revealed a protective effect of dexmedetomidine in a rat model of sepsis, demonstrating a decrease in mortality and inflammatory response (1). With this in mind, we investigated the effect of dexmedetomidine on lung edema formation and albumin permeability in an airway endotoxin model of ALI.

Methods

Mice were divided into four groups: no lung injury/no treatment (control), lung injury/no treatment (LPS/saline), no lung injury/ dexmedetomidine treatment (saline/Dex), and lung injury/dexmedetomidine treatment (LPS/Dex). Under general anesthesia, the subclavian vein was cannulated after which the mouse was allowed to recover for approximately 60 minutes. The mice were subsequently placed in a closed chamber and exposed to nebulized normal saline (no injury) or 10 mg of LPS in saline solution for 60 minutes to induce an inflammatory response. After an additional 60 minutes, mice were given an IV bolus injection of saline or Dex (10 ug/kg) followed by a constant infsuion (10 ug/kg/hr) for 2 hrs. Mice were also injected with 125I-labeled albumin 60 min prior to euthanasia under general anesthesia at which time blood and lung tissue were collected for determination of extravascular lung water (ELW) content, lung edema formation (wet/dry weight ratio), and albumin permeability.

Results

Exposure of mice to nebulized LPS increased ELW from 15.5 ± 4.0 ul to 40.5 ± 5.5 ul (P<0.05) following LPS exposure, whereas ELW in mice treated with Dex was significantly reduced (26.9 ± 4.5 ul) and not different from control or Dex treatment alone (17.8 ± 3.1 ul). In addition, LPS-induced increase in the lung W/D ratio (16% increase over control; P<0.05) was prevented by Dex treatment. Lastly, exposure of mice to nebulized LPS increased vascular albumin permeability from 5.2 ± 0.6 to 6.6 ± 0.8%, whereas treatment with Dex 2 hrs after the onset of LPS exposure tended to block the permeability increase (5.5 ± 0.8% in lungs exposed to Dex alone vs. 5.7 ± 0.7% in LPS/Dex group) though this effect did not reach statistical significance.

Conclusion

In conclusion, treatment with dexmedetomidine after induction of lung injury by LPS significantly decreased extravascular water accumulation in the lung, thereby attenuating edema formation. 125I-albumin permeability may also be reduced. These data suggest that dexmedetomidine can block LPS-induced acute lung injury in mice and thus may be the sedative of choice for patients with ALI or ARDS in the critical care setting.

0368

Mechanisms By Which Ethanol Ingestion Predisposes To Acute Lung Injury (Ali)

M. Lois*1, D. M. Guidot2

1Medicine, John Peter Smith, Fort Worth, 2Medicine, Emory University/ Atlanta VAMC, Atlanta, United States

Introduction

The chronic effects of ethanol abuse on the lung are not known. Recently chronic alcohol abuse was shown to increase the incidence and mortality from ALI in patients with sepsis. We postulated that ethanol ingestion via glutathion depletion and derangement of the interstitial matrix predisposes to ALI.

Methods

Endotoxin-primed (2mg/kg IP, 2 hrs prior to isolation), ex vivo perfused lungs isolated from ethanol-fed and control rats. Vascular and alveolar protein permeability in vivo: 131I and 125I-albumin flux into the vascularspace or alveolar epithelium. Alveolar type II cell barrier function in vitro: % of 14C-inulin leak.

Apical sodium channel function in vitro: % of membrane patches with cation channel activity.

The lung tissue and lung lavage fluid were prepared and assayed for MMP-9 and MMP-2 activity.

Results

Ethanol ingestion increased acute edema formation in lungs isolated from endotoxemic rats. there was decreased plasma and lung tissue levels of reduced glutathione (GSH) and surfactant secretion. Matrix metalloproteinases (MMP) activity was increased due to activation of latent enzymes along with elevations in the 7-S fragment of type IV collagen in the lung lavage fluid. In addition there was increased bi-directional protein permeability across the alveolar epithelium in vivo. As a compensatory response there was increased alveolar epithelial transcellular sodium transport in vitro. Open image in new window

Conclusion

Chronic alcohol abuse decreases tissue glutathione levels and predisposes to ALI.

Oral Presentations Metabolism endocrinology: 0369–0374

0369

Energy Supply Level Correlates With Icu Mortality: A Multicentre Study in a Cohort of 1209 Patients

C. Pichard*1, G. Kreymann2, A. Weimann3, H. Buchardi4, H. J. Herrmann5, H. Schneider6

1Nutrition, University Hospital Geneva, Geneva, Switzerland, 2Critical Care, Universitätsklinikum Eppendorf, Hamburg, 3General Surgery, Klinikum St. Georg, Leipzig,

4Anaesthesiology, Emergency and Critical Care, Universitätsklinikum Göttingen, Göttingen,

5Medical Affairs, Baxter GmbH, Munich, Germany, 6HealthEcon, HealthEcon AG, Basel, Switzerland

Introduction

Early vs. late start of nutrition is associated with improved clinical outcome in ICU patients. The level of energy supply on clinical outcome remains unclear.

Methods

1209 ICU patients with ICU LOS>4 days from the DIVI (Deutsche Interdisziplinare Vereinigung fur Intensiv- und Notfallmedizin) database containing 3548 patients recruited in 14 ICUs at 9 German university hospitals, were evaluated with respect to the effect of early vs. late enteral and/or parenteral energy supply (providing 1500 kcal or more in the first 3 days), on SOFA Score, mortality, sepsis and acute kidney failure identified by ICD 10 classification.

Results
583 early and 626 late energy supply receiving patients aged 60.0±17.2 vs. 61.2±17.7 yr (mean±SD, p=0.25), 367M/215F vs. 393/231 (p=0.98), SOFA 9.1±4.2 vs. 9.1±3.9 (p=0.96) were evaluated. With the exception of a reduced occurrence of acute kidney failure, there was no positive effect of early energy supply during the first 3 days on ICU and hospital mortality and sepsis incidence. Provision of >1500 kcal, apart from separate parenteral glucose, was associated with reduced sepsis incidence (p=0.56), acute kidney failure (p=0.04), ICU (13.4 vs. 18.6%, p=0.026) and hospital mortality (18.7 vs. 24.2%, p=0.035). Increasing energy from 800 to 1300 kcal (Table) was correlated (R=0.99, p=0.0001) with the Odds Ratio for ICU mortality, which suggests a positive effect of early energy supply corresponding to an increasing reduction in mortality from 2 to 6%, respectively.
Table 1

Level of Energy Supply and Clinical Outcome

kcal for 3 first days

Early: Mortality / Survival (%mortality)

Late: Mortality / Survival (%mortality)

Odds Ratio

800

97/510 (16.0%),N=607

108/494 (17.9%),N=602

1.149

900

86/485 (15.1%),N=571

119/519 (18.7%),N=638

1.293

1100

70/425 (14.1%),N=495

135/579 (18.9%),N=714

1.416*

1200

63/400 (13.6%),N=463

142/604 (19.0%),N=746

1.492**

1300

57/382 (13.0%),N=439

148/622 (19.2%),N=770

1.601***

*p=0.03, **p=0.014, ***p=0.005, Chi2-test; early vs. late

Conclusion

Early (3 first days) energy supply, intravenous glucose not counted, is associated with lower mortality and morbidity in ICU patients. The level of early energy supply (800–1300kcal), apart from separate parenteral glucose, is linearly correlated with ICU mortality risk.

Grant Acknowledgement

Baxter GmbH, Munich, Germany

0370

Negative Energy Balance is Associated with Infectious Complications after Subarachnoid Hemorrhage

N. Badjatia*1, L. Fernandez1, M. Schlossberg1, J. Schmidt1, J. Claassen1, K. Lee1, E. Connelly2, S. A. Mayer1, M. Rosenbaum3

1Neurocritical Care, 2Neurosurgery, 3Pediatrics, Columbia University, New York, United States

Introduction

Aneurysmal subarachnoid hemorrhage (SAH) patients are frequently hypermetabolic and at high risk for development of medical complications. This study examined the relationship between energy balance and complications following SAH.

Methods

Prospective observational study conducted in fifty eight consecutive SAH patients aged 58 (range: 26 – 86, 38 (66%) women) who were fed only via enteral nutrition (EN) during the first seven days after hemorrhage between October 2005 and October 2007. We recorded demographic data, time to initiation of EN, energy intake from EN, resting energy expenditure (REE)(measured by indirect calorimetry) and in — hospital complications over 14 days. Energy balance (EB) was calculated as the difference between REE and caloric intake. Fever was defined as core body temperatures >38.3 C and hyperglycemia as blood glucose levels >11 mmol/L. EB data is expressed as mean +/− SD and compared between groups by t test. Cumulative EB was compared to the frequency of infectious and non infectious complications with a Spearman's correlation coefficient. Linear regression was conducted to determine whether EB was correlated with the frequency of specific post-SAH complications.

Results

Median time to feeding was 1 day post bleed (range: 0 – 5). Daily target caloric intake was estimated at 27.5 +/− 3.1 kcal/kg/day and the amount delivered was 14.2 +/− 5.1 kcal/kg/day. EN accounted for 67% of the caloric intake, with propofol and dextrose infusions accounting for 23% and 10% respectively. Cumulative average EB for the first seven days post bleeding was −116.5 +/− 53.0 kcal / kg. A greater average negative EB was associated with urinary tract infections (−134.2 +/− 32.8 kcal/kg v. −106.2 +/− 52.0 kcal/kg v. P = 0.04), sepsis (−152.2 +/− 25.0 kcal/kg v. −110.2 +/− 48.2 kcal/kg, P = 0.03), pneumonia (−128.6 +/− 36.2 kcal/kg v. −97.7 +/− 56.0 kcal/kg, P = 0.02), hyperglycemia (−120.3 +/− 42.1 kcal/kg v. −74.0 +/− 40.2 kcal/kg, P = 0.04), and fever (−126.6 +/− 40.9 kcal/kg v. −94.2 +/− 55.4 kcal/ kg, P = 0.03). The average negative EB during the first seven days after SAH correlated with the total number of infectious complications (r = 0.53, P <0.001), but not non infectious complications (r = 0.21, P = 0.13). After adjusting for Hunt Hess score, fever, hyperglycemia, and anemia, the cumulative number of infectious complications was predicted by average negative EB in the first seven days post bleed (P = 0.01).

Conclusion

A greater negative energy balance is associated with an increased risk of infectious complications after SAH. Future studies need to better understand the impact negative EB has upon outcome after SAH.

Grant Acknowledgement

Neeraj Badjatia received a K 12 Career Development Award for this project by Grant Number KL2 RR024157 from the National Center for Research Resources (NCRR).

0371

Hypoglycemia Counterregulatory Response in Critically Ill Patients

E. A. Vlot1, H. J. Ruven2, H. S. Biemond-Moeniralam*1

1Intensive Care, 2Clinical Chemistry, St Antonius Hospital, Nieuwegein, Netherlands

Introduction

The application of Intensive Insulin Therapy (IIT) in critically ill patients resulted in an increase of the incidence of hypoglycemia in the ICU. With the similar IIT treatment regimen, the incidence of hypoglycemia was higher in septic patients compared to postsurgical non-septic patients. Hypoglycemia in septic patients can also occur without insulin treatment. Little is known about the counterregulatory hormonal response to hypoglycemia in critically ill patients. The aim of this study was to investigate if the counterregulatory response to hypoglycemia in septic patients is blunted compared to postsurgical non-septic patients. This could be an important factor in spontaneous hypoglycemia and the higher incidence of hypoglycemia in septic patients while treated with IIT.

Methods

Prospective, observational study in adult ICU patients with severe sepsis or septic shock and a first hypoglycemia versus postoperative non-septic ICU patients with a first hypoglycemia. Blood samples for measurement of the counterregulatory hormones (glucagon, epinephrine, cortisol and growth hormone) were taken directly after the first severe hypoglycemia (blood glucose level <3.0 mmol/l, a level below wich a significant increase in the counterregulatory hormone levels is expected). Directly thereafter the hypoglycemia was corrected. All patients were treated according to an adjusted IIT regimen aiming at blood glucose levels between 4.5–8.0 mmol/l. Student-t-test was performed for data analysis.

Results

The counterregulatory hormone response to the first hypoglycemic event was measured in 4 non-septic post-cardiosurgical and 8 severe septic patients. The plasma glucagon response to a severe hypoglycemic event was lower in the severe septic compared to the non-septic post-cardiosurgical patients (mean 79±17 ng/l versus 168±46 ng/l, p<0.05). Growth hormone response was 3.6±1.3 ng/l in the septic group compared to 12.3±6.0 ng/l in the non-septic group, p=0.08). Cortisol and the catecholamine responses were not significantly different between the two groups.

Conclusion

The counterregulatory response to hypoglycemia in septic patients is blunted compared to postsurgical non-septic patients. These findings support the need for more frequent blood glucose measurements during IIT in septic ICU patients, in order to reduce the incidence of hypoglycemia during IIT.

0372

Patients with Hyperglycaemia During Critical Illness are at Increased Risk of Developing Diabetes

I. Gornik*1, A. Vujaklija1, O. Gornik2, E. Lukić3, G. Madžarac3, V. Gašparović1

1Department of ICM, University Hospital Rebro, 2Department of biochemistry and molecular biology, University of Zagreb, Faculty of pharmacy and biochemistry, 3Medical school, University of Zagreb, Zagreb, Croatia

Introduction

Hyperglycaemia occurring in a non-diabetic patient is a known complication of critical illness, in diseases such as sepsis, acute coronary syndromes and other conditions in intensive care. Its origin is in the acute stress response, in most part mediated by stress hormones such as adrenalin and cortisol, which increase gluconeogenesis and glicogenolysis, probably by increased insulin resistance. Yet, some non-diabetic patients with critical illness do and some do not develop hyperglycaemia. We have hypothesised that patients who have hyperglycaemia have impaired glucose control mechanisms and are at increased risk of developing type 2 diabetes mellitus in the period after ICU discharge.

Methods

We included adult patients discharged alive from the hospital after being treated in the medical ICU during the time of 3 years (2000–2002). For better congruence, only patients with sepsis and acute coronary syndrome were selected. Patients whose venous blood glucose during the ICU stay never exceeded 7.7 mmol/l formed the normoglycaemia group, while the patients who had hyperglycaemia (venous blood glucose >7.7 mmol/l) on at least two occasions formed the hyperglycaemia group. Patients with only one hyperglycaemic episode were excluded from the study. Patients with terminal illness or those receiving corticosteroids were excluded. Absence of diabetes was confirmed before discharge. Follow-up time was at least five years.

Results

There were 331 patients with selected diagnoses discharged alive willing to participate and with no terminal illness, 168 formed the normoglycaemia group and 90 patients were left in the hyperglycaemia group after excluding newly diagnosed diabetes and patients on corticosteroids. In the normoglycaemia group 115 patients finished follow-up: 95 remained normoglycaemic, 16 developed impaired glucose tolerance (IGT) or increased fasting glucose (IFG), while 4 developed type 2 diabetes. In the hyperglycaemia group 51 patients finished follow-up: 29 remained normoglycaemic, 14 developed IGT or IFG, 8 developed type 2 diabetes. For patients with critical care hyperglycaemia, relative risk for development of IGT or IFG was 2.26 (95% CI 1.21–4.22) and for development of type 2 diabetes 5.35 (95% CI 1.71-16-72).

Conclusion

Our results show that patients with critical care hyperglycaemia are at increased risk of developing impaired glucose control or type 2 diabetes mellitus in the 5 years follow-up period. This supports the theory that this is a group of patients with latent endocrine disorder manifesting initially during the critical illness and developing to a manifest disorder later. Critical care hyperglycaemia should be considered a risk factor for diabetes.

0373

Reduced Arginine Production and Increased Arginase Activity are Independent of the Cause of Sepsis

K. A. P. Wijnands*1, M. Poeze1, Y. C. Luiking2, P. Breedveld3, C. H. C. Dejong3, G. Ramsay4, N. E. P. Deutz2

1Department of Surgery and ICM, University Hospital Maastricht, 2Department of Surgery, University Hospital Maastricht, currently: University of Arkansas for Medical Sciences, Center for T, 3Department of Surgery, University Hospital Maastricht, Maastricht, Netherlands,4., West Hertfordshire Hospitals, London, United Kingdom

Introduction

During sepsis plasma arginine (Arg) levels are hypothesized to be reduced, while whole body arginase activity is increased and de novo arginine production is reduced. It has been suggested, that the abnormalities in arginine metabolism during sepsis are related to the origin of the sepsis. Equally, mortality of sepsis is dependent on its cause. Aim: To study arginine metabolism and arginase activity during surgical and non-surgical sepsis.

Methods

25 ICU patients were studied: 12 with severe sepsis/septic shock (surgical sepsis) within 48h of diagnosis after surgery, 6 with severe sepsis due to pneumonia (nonsurgical sepsis), and 7 non-septic ICU controls; 16 age-matched healthy subjects served as controls. Arg metabolism was studied using primed-continuous infusion of stable isotopes of arginine, citrulline and urea, and subsequent measurement of arterial amino acid concentrations and tracer-tracee ratios. De novo Arg production is the Cit to Arg conversion and whole body arginase activity was measured as Arg to Urea conversion. None of the patients received blood transfusions. Statistical analysis by 2-way ANOVA with Bonferroni correction between groups; data are means ± SEM.

Results

Plasma arginine concentration was similarly reduced in septic patients with surgical cause compared to non-surgical. Also, whole-body arginine production was similar between surgical and non-surgical sepsis. Whole-body arginine de novo synthesis was reduced both in surgical and non-surgical sepsis, while arginase activity was increased in both groups compared to ICU controls and healthy subjects.

Conclusion

In septic patients, plasma arginine levels are reduced early during sepsis. This appears to be related to the increased plasma arginase activity and reduced de novo arginine production. However, these observations are not related to the origin of the sepsis.

0374

Growth Hormone Secretion Pattern in Acute and Prolonged Critical Illness After Multiple Trauma

F. Duska*, M. Fric, P. Waldauf, J. Pazout, J. Pachl

Dept. of Anaesthesia and Int. Care Med, Charles University, 3rd Faculty of Medicine, Prague, Czech Republic

Introduction

Acute and prolonged phase of critical illness represent different neuroendocrine milieus. According to animal models [1], anterior pituitary dysfunction, particularly flat growth hormone (GH) secretion pattern may contribute to muscle wasting syndrome during protracted critical illness. In this human study we ask, whether GH secretion pattern differs in presence/absence of traumatic brain injury (TBI) and how it changes in time.

Methods

Subjects:Multiple trauma patients expected to require at least 2 weeks of ICU care, with (n=12) and without (n=8) TBI defined as GCS <8 on the scene and intracranial pathology on CT. Male=17, Female=3, aged 40+−16 years, BMI 27+−4 kg.m−2, ISS=39+−14, APACHE II 24+−8. Weexcluded patients with diabetes insipidus or direct trauma of hypothalamus/pituitary. Design: Arterial blood was sampled every 30 min from 22:00 to 06:30 on day 4 and 17 of their ICU stay and GH was sampled using RIA. As a measure of GH secretion we use basal level, number and amplitude of pulses (derived manually from individual GH vs. time plots), and area under the GH curve. For these experiments, we used control groups from a previously publihed study [2]. Mann-Whitney or Wilcoxon tests are used for comparisons, as appropriate.

Results
During acute illness (day 4) patients without TBI tend to secrete more GH than those with TBI (median AUC 74 vs. 130 mIU/l, p=.46), because GH pulses tend to have higher amplitude (p=.13), but there is no difference in No of pulses (p=.52). During acute-to-protracted transition (i.e. from day 4 to 17), we did not observe any significant change in GH secretion pattern (see Table), but the overall amount of GH secreted tend to decrease in patients without TBI (p=.098).
Table 1

GH Secretion Pattern

Units: mIU/l

Day 4 Pt. with TBI (n=12)

Day 4 Pt. without TBI (n=8)

Day 17 Pt. with TBI (n=10)

Day 17 Pt. without TBI (n=8)

Basal level

2 (1–7)

4 (2–9)

2 (1–5)

2 (1–3)

No of pulses/night

2 (0–3)

2.5 (2–3)

3 (0–4)

2.5 (1.5–3)

Avrg. pulse amplitude

4 (3–8)

9 (6–16)

6 (3–8)

6 (5–9)

Summ of GH secreted

74 (28–197)

130 (79–258)

73 (31–197)

70 (47–105)

Data presented as median (interquartile range)

Conclusion

In contrast with a previous animal study [1], we did not demonstrate any significant differences in GH secretion pattern between acute and protracted critical illness in humans after multiple trauma. TBI seem to impair GH secretion capacity only during the acute phase.

Grant Acknowledgement

IGA NR 8230-4/2004.

Oral Presentations Septic shock: 0375–0380

0375

Bacterial Genoma Detection at Early Stage of Septic Shock: A Feasibility Study

A. Dugard*1, J. Amiel1,M. Ploy2,M. Clavel1,N. Pichon1,D. Chainier2, P. Vignon1,B. François1

1Medical-surgical Intensive Care Unit, 2Bacteriology and Virology, Dupuytren Teaching Hospital, Limoges, France

Introduction

Early microbiological documentation is a key point for adapted antibiotherapy in patients sustaining a septic shock. This is not consistently possible to achieve using conventional bacteriological methods which require a substantial delay to obtain final results and have a fairly high rate of false negative results. Molecular biology using nucleic acid based techniques may provide an earlier and better identification of the pathogen. We conducted a prospective study to compare the results of conventional microbiology with those obtained from molecular biology at the early stage of septic shock.

Methods

Patients admitted to our medical-surgical intensive care unit with a high clinical suspicion of septic shock during daytime were studied. Blood sample and oriented biological sample(s) (urine, tracheal aspirate, CSF, etc…) were obtained. Conventional microbiological assessment and molecular tests based on bacterial DNA identification (Staphylococcus aureus [S.a], methicillin-resistant Staphylococcus aureus [MRSA], Pseudomonas aeruginosa [P.a], Streptococcus pneumoniae [S.p], Enterobacteraceae [Ent.]) in normally biological sterile fluids and in serum were performed using real-time PCR (SmartCycler®, Cepheid). A universal bacterial DNA sequence identification was also performed followed by DNA sequencer analysis for species identification.

Results
Two patients with positive candidemia were excluded from this feasibility study and 28 patients were studied (age [mean±SD]: 57±18 years, APACHE II score: 23±7). Septic shock was of pulmonary (n=12), abdominal (n=5), urinary (n=8), cutaneous (n=2) or meningeal (n=1) origin. A bacteremia was documented in 7 patients. Of them, molecular serum analysis was in agreement with conventional bacteriology in 3 patients. At least one pathogen was identified using conventional microbiological assessment in other biological samples in 13 patients (Table). Of them, 11 had concordant molecular analysis. In addition, PCR analysis allowed the identification of at least one pathogen in the biological samples obtained from 2 patients with undocumented sepsis using the conventional microbiological approach.
 

]S.a

SAMR

P.a

S.p

Ent.

Conventional microbuology (n)*

2

1

3

2

7

PCR analysis (n)*

2

1

2

4

9

*: 1 biological sample may be positive for several pathogens

Conclusion

In this pilot study, we showed that bacterial DNA identification in biological samples other than blood is feasible at the early phase of septic shock. The molecular approach promises to provide early identification of the causative pathogen in either blood or biological normally sterile fluids. Further studies are needed to define the field of use of this promising technique on clinical grounds.

0376

Searching the Surviving Sepsis Campaign Database for Avoidable Deaths

J. Koeze*1, V. M. Silderhuis1, E. K. A. H. Metz2, T. Halaby3, G. J. Kootstra4, R. M. L. Brouwer5

1Internal Medicine, 2Cardiology, Medisch Spectrum Twente, 3Microbiology, laboratorium microbiologie twente achterhoek, 4Internal Medicine and Infectious diseases, 5Internal medicine and Critical Care, Medisch Spectrum Twente, Enschede, Netherlands

Introduction

Recently it has been estimated that in the Netherlands yearly about 1700 deaths in hospitalised patients could prevented by timely, accurate, diagnosis and adequate medical care. The surviving sepsis campaign (SSC) is an international campaign aiming at a reduction in mortality with 25% in 5 years time. In March 2007 the SSC guidelines were introduced in the ICU, ER and the departments of Internal Medicine and Surgery of our hospital.

Methods

Retrospective analysis of prospectively obtained data from the SSC database a large non-academic teaching hospital.

Results

Hundred and thirty two patients (age 65 ± 1,3 mean ± s.e.m., 65% male) were included. 72 patients (55%) were admitted to the ICU, 60 patients (45%) were treated in the wards. Thirty four (27%) patients did not survive, 12 of these patients were not admitted to the ICU because of limits on escalation of treatment (LOT). Twenty two patients were admitted to the ICU. Nineteen died in the ICU, cause of death multiorgan failure (n=15) and withdrawal of futile treatment (n=4). Three patients died after transfer from the ICU to the ward (all three with LOT).

Compliance to the resuscitation bundle in the deceased patients was >80%. In 31 patients blood cultures were taken. Blood cultures were positive in only five. In 15 patients a primary pulmonary focus was suspected however, in only one a suitable sputum specimen was obtained (Streptococcus Pneumoniae, n=1). Seven patients were thought to have a urinary tract infection, in six urinary cultures were taken (E.Coli, n=1; Klebsiella Pneumoniae, n=1). An abdominal focus was suspected in twelve patients, urine cultures were taken in six. In three patients the same microorganism was found in the blood and urinary culture (E.Coli (2), Klebsiella Pneumoniae). Microbiological data to narrow the antibiotic regimen were only available in a limited number of patients.

Conclusion

Compliance to the SSC bundles was excellent, but more microbiological samples of suspected sources of infection should be obtained. Many patients who died in our SSC cohort had limits on escalation of treatment and cannot be counted as preventable deaths. Detailed analysis of cause of death detects flaws in the diagnostic and therapeutic process and should be used to improve this process.

0377

Septic Shock Due to Icu-Acquired Candidemia; a Fatal Condition?

R. Tchokonte1, J. A. Guzman*2

1Medicine, Wayne State University, Detroit, 2Pulmonary and Critical Care Medicine, Cleveland Clinic Foundation, Cleveland, United States

Introduction

Bloodstream infection caused by Candida spp. is becoming increasingly prevalent in ICU patients and it carries a high morbidity and mortality. Although progress has been made in the management of sepsis, limited data are available addressing the outcome of fungal septic shock.

This study was performed to assess the outcomes of patients who developed septic shock due to ICU acquired candidemia.

Methods

The medical records of patients who had at least one episode of Candidemia developed after their ICU admission over a five year span were reviewed. Case patients were defined as individuals who had at least one positive blood culture for Candida species collected > 48 h after ICU admission who developed septic shock within 48hr of the positive blood culture. Patients with endocarditis were excluded. Results are expressed as means ± SD.

Results

There were 83 episodes of candidemia (1.55/1000 patient days). Eighteen patients met shock criteria and were included in the study. Patients were female (55%), age 61.2 ± 13.4 years, and had an APACHE II score of 26.8 ± 8.8. Main species causing candidemia were, C. glabrata (61%) C. albicans (33%) and C. lusitaniae (6%). Hospital and ICU length of stay were 27.3 ± 21.8 and 19.0 ± 17.0. days, respectively. Candidemia developed 8.4 ± 13.2 days after ICU admission. Most patients had non-neutropenic immunesuppression, eight (44%) were on corticosteroid treatment, 6 (33%) were colonized (4 combined urine and sputum, 2 sputum), 10 (55%) patients were on parenteral nutrition and all patients were on antibiotics at the time of candidemia. Clearance of fungemia (n=7) occurred 7.1 ± 7.4 days after the initiation of treatment. The majority of patients (67%) required at least 2 vasopressor agents to keep blood pressure at goal. Mortality was 89%.

Conclusion

Prevalence of ICU-acquired candidemia was high in our cohort. The limited size of our sample precludes assessing whether non-albicans spp are associated with different outcomes, although they were the more prevalent species in patients with shock. Development of fungal septic shock, regardless of the species, is associated with a nearly fatal outcome.

0378

Comparison of Procalcitonin and C-Reactive Protein for Differentiation Between Systemic Inflammatory Response Syndrome, Sepsis and Septic Shock

S. S. Zurnic*1, R. M. Popovic2

1Department of Intensive Care Therapy, 2Department of Microbiology, General Hospital, Valjevo, Serbia

Introduction

Numerous clinical studies have shown that early diagnosis of sepsis and adequate antibiotics therapy during the first six hours after development of sepsis are essential to the outcome of therapy and survival of the septic patients. Procalcitonin is the most sensitive early laboratory marker of sepsis.

Methods

The prospective study included thirty- three critically ill patients that were hospitalized in Intensive Care Unit with suspicion on infection. The American College of Chest Physicians/ Society of Critical Care Medicine Consensus Conference definition of sepsis was used to identify patients with systemic inflammatory response syndrome (SIRS), sepsis or septic shock. Sequential organ failure assessment (SOFA) score was used to describe the sequence of complications and the severity of organ dysfunction in critically ill patients. Patients were split into three groups on the basis of clinical, laboratory and bacteriological findings: SIRS group; sepsis group and septic shock group. The level of Pct and CRP in differential diagnosis of SIRS, sepsis and septic shock was analyzed. Data are presented as means±standard deviation (SD) and in percentage values in addition to absolute numbers. To compare two independent samples, an unpaired t-test was used, and the chi-square test to compare proportions. Among the Pct, CRP and SOFA score, linear (Pearson,s) correlation and the regression formula were calculated (y= a+ bx). Statistical significance was accepted at p < 0,05.

Results

In the group of 33 critically ill patients there were 10 with SIRS, 15 with sepsis and 8 with septic shock. Pct and CRP levels were higher in patients with sepsis and septic shock, although correlation with SOFA score was weak for CRP(0,47 in septic patients, p>0,05). CRP levels were near their maximum already during lower SOFA scores, whereas correlation of the Pct level and SOFA score was 0,98 in infected patients. In patients with SIRS, Pct (CRP) concentrations were 0,40ng/ml (116mg/l), in patients with sepsis 4,6 ng/ml (150mg/l) and 18,6ng/ml (182mg/l) in septic shock. The kinetics of both parameters were also different, and Pct concentrations reacted more quickly than CRP.

Conclusion

Results of the study have shown that Pct is a more accurate diagnostic parameter for differentiating SIRS and sepsis, and therefore daily determinations of Pct may be helpful in the follow up of critically ill patients.

0379

Predictive Factors of Renal Replacement Therapy in Patients with Severe Sepsis or Septic Shock

L. Zoric*1, L. Muller1, S. Alonso2, S. Jaber3, J. E. de La Coussaye1, J. Y. Lefrant1

1Division Anesthesie reanimation Douleur Urgence, 2departement d Information Medicale, CHU Nîmes, Nimes, 3Service d'anesthésie réanimation B, CHU Montpellier, Montpellier, France

Introduction

Acute renal failure (ARF) is associated with a poor outcome in patients with severe sepsis or septic shock (1). We looked for the predictive factors of acute renal failure requiring renal replacement therapy (RRT) in a data base of 445 patients with severe sepsis or septic shock.

Methods

All patients with chronic renal failure needing dialysis were excluded from the study from the database of 445 patients who had participated in an observational study on severe sepsis in 15 ICUs during 2006. We studied the morphological parameters, organ dysfunction at admission, the plasma creatinine level at admission, and the quantity of fluid resuscitation and the type of fluid during the first 24 hours. These parameters were entered into a univariate and multivariate analysis with logistic regression analysis. The results are presented in real value, percentage or median value (5th and 95th percentile), depending on the studied parameter. P<0.05 is significant.

Results
423 patients were screened for the study, 113 (27%) of them needed RRT. The comparison of the patients with and without RRT is shown in the Table. In the multivariate analysis including all significant variables, male sex (OR=2.045 [1.154–3.625]), oliguria (OR=2.665 [1.575–4.510]), red blood cell transfusion (OR=2.349 [1.39–3.969]) and a plasma creatinine level >113 μMol/l (OR=3.406 [1.988–5.936] were associated with a risk of RRT (validity of the model verified by a Hosmer and Lemeshow test, p = 0,6). However, the type of fluid infused did not influence the need for RRT(p = 0.83).
Table 1

Comparison Between Patients With RRT and without RRT

 

RRT (113)

Without RRT (310)

p

age (years)

68 [36–88]

66 [32–84]

0.09

Women

31 (28%)

113 (37%)

0.09

IGS II

70 (62%)

43 [23–75]

<0.01

Initial oliguria

70 (62%)

113 (36%)

< 0.01

Initial plasma creatinine (μMol/l)

168 [58–456]

102 [42–314]

< 0.01

Shock

86 (76%)

222 (72%)

0.36

Fluid resuscitation at 24 h (ml)

3000 [500–7000]

2500 [800–7500]

0.04

Crystalloids (ml)

2800 [500–7000]

2500 [500–7500]

0.14

Hydroxyethyl starch (ml)

1500 [500–2800]

1000 [500–3000]

0.09

Albumine (ml)

400 [200–1300]

400 [200–1500]

0.78

Transfusion

71 (63%)

132 (43%)

< 0.01

Conclusion

Initial oliguria, the baseline plasma creatinine level and the sex are associated with ARF requiring RRT in patients with severe sepsis. However, the type of fluid infused doesn't seem to influence the occurrence of ARF.

Grant Acknowledgement

The study was supported by CHU Nîmes.

0380

Procalcitonin and C-Reactive Protein as Markers for Infection and Mortality in Patients with Severe Sepsis and Septic Shock

J. Koeze*1, V. M. Silderhuis1, R. M. L. Brouwer2

1Internal Medicine, 2Internal Medicine and Critical Care, Medisch Spectrum Twente, Enschede, Netherlands

Introduction

Procalcitonin (PCT) is used as a marker to differentiate sepsis from other non-infectious causes of the systemic inflammatory response syndrome (SIRS). C-reactive protein (CRP) is also used as a marker for sepsis and severity of disease. The Surviving Sepsis Campaign (SSC) has been developed to improve the management, diagnosis, and treatment of sepsis. In March 2007 the SSC guidelines were introduced in the ICU and the departments of Internal Medicine and Surgery of our hospital. We studied the clinical application of PCT and CRP plasma concentrations in the detection of severe sepsis /septic shock and the assessment of severity of disease.

Methods

Prospective observation study in patients admitted to the departments of internal medicine, surgery and the ICU of a large non-academic teaching hospital who were enrolled in the SSC. PCT and CRP plasma levels were determined at inclusion (0) and 24 hour (24) after inclusion in de SSC registration. Patients were classified according to outcome (hospital discharge / mortality), presence of infection confirmed by microbiological culture and PCT / CPR levels. Data were analyzed using non-parametric statistical methods (Pearson Chi-Square and Kruskal-Wallis test).

Results

Hundred and thirty two patients (age 65 ± 1,3 mean ± s.e.m., 65% male) were included. 72 (55%) were admitted to the ICU, 60 patients (45%) were treated in the wards. Infection was confirmed by cultures in 50% of patients. There was no difference in PCT (0/24) between patients with and without microbiological culture proven infection. Fifty percent of patients were classified as severe sepsis and 50% as septic shock. Overall hospital mortality was 27% and there was no difference in PCT(0/24) and CRP (0/24) between survivors and non-survivors.

Conclusion

PCT does not differentiate patients with culture confirmed infection and severe sepsis/septic shock from patients with severe SIRS and suspected but no proven infection. PCT and CRP are no indicators of outcome in patients with severe sepsis and septic shock.

Oral Presentations Cultural variations at the end-of-life: 0381-0386

0381

Attitudes of Intensive Care Physicians Toward End-of-Life Decision — Czech National Survey Study

R. Parizkova*1, V. Cerny2, K. Cvachovec3, I. Novak4, V. Sramek5, D. Nalos6

1Anaesthesiology and ICM, University Hospital, Hradec Kralove, 2Anaesthesiology and ICM, University Hospital, Hradec Kralove, 3Anaesthesiology and ICM, University Hospital Motol, Prague, 4ICU Internal Medicine I, University Hospital, Plzen, 5Anaesthesiology and ICM, St Ann0s University Hospital, Brno, 6Anesthesiology and ICM, Masaryk0s Hospital, Usti nad Labem, Czech Republic

Introduction

The development of intensive care has brought the ability to increase patients survival and quality of life but the prolongation of treatment in terminally ill patients without chance to improve outcome has been considered as a futile (1). The aim of study was obtaining information concerning ICU physician's attitudes in the Czech Republic toward end-of-life decision (EOLD) and real implementation into the clinical practice.

Methods

A structured questionnaire was sent to each member of Czech Society of Anaesthesiology and Intensive Care. The opinion of physicians on ways of EOLD including euthanasia was explored. The answers were compared with respect to physicians religion, social and demographic characteristics, length of practice and type of hospital.

Results
A total 870 questionnaires were sent, response rate was 213 (26%). Ninghty percentage of all physicians considered limiting therapy in terminally ill patients as a acceptable, however the implementation to the clinical practice was much less frequent (Table 1). In incompetent patients 96% responders considered physicians as a key person for decision making. Only 57 (28,9%) resp. 87 (43%) physicians include family resp. nursing staff in decision-making process. Physicians from small hospitals and with shorter clinical practice include relatives and nurses to EOLD significantly fewer. The majority of responders agree with maintenance of analgosedation and infusion therapy, 76% keep ventilatory support, 43% nutrition, 49% oxygen and 14% antibiotics. Euthanasia is acceptable in 39 (19,1%) physicians, from those most of them were unbelievers. Lack of privacy and absence of family being present was considered as a key barriers for providing death dignity.
Table 1

Attitudes and Practice of Physicians

 

Limiting therapy — attitudes

Limiting therapy — real practice

Withholding treatment

205 (99)

158 (75.9)

Withdrawing treatment

191 (91)

100 (48.3)

Terminal weaning

107 (51.4)

19 (9.3)

Do not resuscitate

209 (99.5)

148 (70.5)

Descriptive statistics was used, data are presented as Numbers (%)

Conclusion

Withholding and withdrawing treatment is acceptable for most Czech ICUs physicians however implementation in practice is lower. Analgosedation and fluids are the most procedures kept during withholding therapy. Lack of privacy and impersonal environment were considered as the main obstacle for death dignity on ICUs.

0383

End of Life Decisions in an Indian Icu

A. K. Mandal*, S. Bal, Y. Javeri, D. Nama, H. Tewari, P. Pandey, T. Rawat, R. Kumar, R. Uttam, R. Mani

PulmonologyThoracic Surgery and MICU, Fortis Flt. Lt. Rajan Dhall Hospital, New Delhi, India

Introduction

Limitation of life support in Indian ICUs is less frequently practiced as compared to that in the West. This may be attributable to lack of physician awareness, appropriate legislation and social factors. There is a paucity of data on end of life care in Indian patients.

Methods

Retrospective chart review of all patients who expired during the period from May 2006 till December 2007. Data collected included age, gender, disease category, comorbidities, mechanical ventilation days, length of stay in ICU and hospital, APACHE IV, premorbid functional status and End of Life decisions. Interventions, antibiotic (carbapenems), vasopressor and diagnostic studies performed within 3 days of death were also collected. EOL decisions were taken on the lines of ISCCM recommendations. Setting: A 12 bed Medical — Surgical ICU in a tertiary care centre in India.

Results

There were 830 admissions of which 88 (10.6%) expired. The study population was divided into 2 groups — Full Support (FS) and Support Limitation (EOL). The population was further subdivided into: young (female ≤59 years and male ≤64 years) and elderly (female ≥60 years and male ≥65 years). FS group comprised of 33 (73.3%) males &12 (26.7%) females and 26 (60%), 17 (39.5%) respectively in the EOL group (NS). 19 (42.2%) & 26 (57.8%) were young and elderly in the FS group and 26 (60.4%) & 17 (39.5%) in the EOL group (NS). The primary organ involvement was: respiratory system 16 (35.6%), sepsis 7 (15.6%), neurological 5 (11.1%), renal 2 (4.4%) & cardiac 1 (2.2%) in the FS group and 26 60.4%), 10 (23.2%), 6 (13.9%), 0, 0 & 3 (6.97%) respectively in the EOL group (NS). Co morbidities were 37 (82.2%) in the FS group and 41 (95.3%) in the EOL group (p = 0.05). Mean APACHE IV was 80.8 in the FS group and &77.34 in the EOL group (NS).The mean LOS in the ICU was 7.3 in the FS group and 15.4 in the EOL group (p = 0.02).Functional status in the FS group were: 25 (55.5%) independent, 17 (37.7%) partially dependent & 3 (6.6%) fully dependent and in the EOL group: 12 (27.9%), 20 (46.5%) and 11 (25.6%) respectively (NS). Interventions were done in 44 (97.7%) in FS group and 29 (67.4%) (NS). Antibiotic change to carbapenems in 14 (31.1%) in the FS group and 5 (11.6%) in the EOL group (p = 0.02). Vasopressors were added in 9 (20%) in FS group and 1 (2.3%) in the EOL group (p = 0.01).Diagnostic studies were performed in 28 (62.2%) in the FS group and 7 (16.3%) in the EOL group (p = 0.00) The End of Life decisions taken were taken in 43 (48.8%). Death was preceded byDNR(do not resuscitate) 15 (34%),WH (with holding of therapy) 25 (58.1%) and WD (withdrawal of therapy) 3 (7%).

Conclusion

End of life decisions preceded 49% of deaths. These decisions led to a reduction in aggressive interventions immediately prior to death as compared to full support decisions. However, limitation of therapy did not reduce the ICU length of stay.

0384

Comparison of Deaths Occurred in the Wards to Those Occurred in the Intensive Care Unit (ICU) at an University Hospital

F. O. Machado*, R. D. Moritz, G. Beduschi, M. Heerdt, B. Rosso

Critical Care Medicine, Federal University — University Hospital, Florianopolis, Brazil

Introduction

The majority of deaths in Brazil happen in hospitals and more specifically in ICU. In this country there is few studies about the dying process and there is not clear legal definitions about withhold or withdraw support (WWS). It is important to know about the reality of the dying process to provide the best treatment for dying patients. The objectives of this study are: To analyze the deaths occurred in the HU/UFSC; To compare the profile of the patients who died in the ICU to those who died in the wards; To identify if the decisions of WWS, comfort care plans, patients’ identification as dying and/or do-not-resuscitate orders prior to the death were recorded.

Methods

Retrospective cohort study, approved by Ethical Committee. Deaths occurred in patients, admitted to the HU/UFSC from July 2004 to June 2007, were analyzed. Demographic characteristics, clinical features and the treatment performed for the patients who died were evaluated. It was considered if WWS preceded the death. Data was allocated in two groups: G1 to the patients who died in the ICU and G2 to the patients who died in the Wards. Data were analyzed using t Student and χ2 tests (p<value <0,05).

Results

The analyzed hospital has 165 beds (158 in wards and 7 in ICU). During the study period 14330 patients were admitted at hospital. The hospital mortality rates was 4,95%. This index in ICU was 23,77% and in Wards was 3,39%. The data of 710 patients had been analyzed. Two hundred and sixty one died in ICU (G1) and 449 in the wards (G2). The most of patients were men in both groups (56,9% in G1, 53% in G2) and the patients mean age was 57,4±18,1 years (G1) and 69,7±15,8 years (G2) (p<0.001). WWS preceded 37,5% of the deaths in G1 and 9,6% in G2 (p<0,001). Vasopressor drugs were the most frequently WWS in G1, while intensive care admission was the most WWS in G2. Do-not-resuscitate orders was documented in 0,76% of patient's records (G1) and in 2,44% (G2). Comfort care plans were present in 0,38% in cases (G1) and in 11,13% (G2). Patients were identificate as dying in 2,68% (G1) and 29,84% (G2). Cardiopulmonary resuscitation (CPR) prior to death was present in 27,20% of the cases (G1) and in 20,26% (G2).

Conclusion

In G1 the patients were younger and the death was more frequently preceded by WWS. In G2 the patients were older, more commonly considered “dying” and the death was more frequently preceded by comfort care plans. A CPR maneuver was more common in G1. Do not resuscitate orders were unusually documented.

Grant Acknowledgement

Federal University of Santa Catarina.

0385

Lawyers in ICU: 12 Month Audit of In-House Legal Assistance to ICU Patients

C. Eynon*1, S. Dench2, A. Dinsmore2

1Neurosciences ICU, Wessex Neurological Centre, Southampton, 2Stewarts solicitors, 63 Lincoln's Inn Fields, London, United Kingdom

Introduction

The Neurosciences ICU (NICU) sees over 650 adult patients per year. 2/3 are emergency admissions following traumatic brain injury, spinal injuries, subarachnoid haemorrhage or intracerebral haemorrhage. Legal issues regarding power of attorney, compensation claims, dealing with the police etc are not taught at medical school yet are often immediate concerns for both patients and their families.

Methods

A legal firm specialising in medicolegal issues, particularly brain and spinal injuries, was approached about providing pro-bono advice to patients on NICU. After approval by the Hospital Trust a 12 month pilot project was established with lawyers spending one day per week on the NICU. Formal consent/assent was sought from the patient or their relatives by the ICU consultant.

Results

31 cases were reviewed. Demographics were similar to the overall patient mix on NICU. 10 cases were road traffic accidents (passenger or pedestrian), 6 RTA (driver), 7 medical cases (SAH or ICH), 2 industrial accidents, 1 sport related injury, 3 alleged assaults and 2 falls. Breach of care and the possibility of compensation was identified in 22 cases. Non-compensation issues were identified for all cases and a total of 123 hours spent on noncompensation issues. The legal firm involved were instructed to pursue compensation claims in 7/22 cases. Feedback from families has been universally positive. The service has been extended to all patients in the Neurological Centre.

Conclusion

Many ICU patients have immediate legal concerns. The use of a legal team in ICU has significant benefits in terms of access to legal aid and the time taken for completion of legal issues. Overall wellbeing of patients and their family is improved by relief of anxieties regarding legal matters. Identification of compensation claims will potentially allow earlier access to funding for ongoing care.

0386

Rapid Response Systems and the Demand From Do-Not-Attempreanimation Patients

F. L. de La Vega*1, E. O. Ribas1, R. B. Albuquerque1, P. P. De Leon1, J. A. Victorino2

1Rapid Response System, 2Intensive Care Unit, Hospital Mãe de Deus, Porto Alegre, Brazil

Introduction

Rapid response systems (RRS) are developed as a tool to decrease the assistential gap between the acute critical patients in the wards and the advanced care avaliable just on critical care areas. Indeed, the RRS calls are usually made for all instable patients in the wards, independently of their resuscitation status. So, some calls are made for patients with do-not-attemp-reanimation (DNAR) orders, and others, have a DNAR order registred after the evaluation of the instability. We quantify here the utilization of resourses of RRS to attend DNAR patients in 6 months.

Methods

We describe the proportion of calls made for DNAR patients, the call criteria and the management of this group of patients.

Results

In 6 months, the RRS received 536 calls, 33 was for patients with a previous DNAR order and 31 for patients that received a DNAR order during the RRS attending. This sums a total of 12% of calls for DNAR patients. The median age was 80 ± 11 years. The more frequent calling criteria for DNAR patients was respiratory distress (30%, 19 calls), low oxigen saturation (28%, 18 calls), death constatation (27%, 17 calls) and hypotension (13%, 8 calls). The RRS prescribed medications for 34 (53%) DNAR patients and solicited adictional exams for 12 (19%). Eleven (17%) DNAR patients were submited to Non-Invasive Ventilation (NIV) and one (2%) was intubated. From the 64 DNAR patients, 1 (2%) was transferred to ICU and 3 (5%) to High Dependency Unit (HDU).

Conclusion

These data show that in our experience, a significative proportion of RRS work is dedicated to assist DNAR patients. This group of DNAR patients is old, demand medical attention for respiratory disfunction and has a low rate of investigatory exams performed, low ICU transferences, but high rates of HDU transferences and NIV utilization rates. We suggest that better evaluating the characteristics and needs of this group of patients may provide tools to better organization of care.

Oral Presentations Abstract awards winners — The best preselected abstracts submitted to the congress: 0387–0390

0387

Impact of Subglottic Secretions Drainage on Ventilator-Associated Pneumonia: A Randomized Multicentre Trial

J. Lacherade*1, P. Guezennec2, K. Debbat3, J. Hayon4, A. Monsel1, B. De Jonghe1, P. Fangio1, C. Appere de Vecchi1, H. Outin1, S. Bastuji-Garin5

1Medico-surgical ICU, Poissy Hospital, Poissy, 2Medico-surgical ICU, Andre Mignot Hospital, Le Chesnay, 3Medico-surgical ICU, Avignon Hospital, Avignon, 4Medico-surgical ICU, Saint-Germain Hospital, Saint-Germain en Laye, 5Public Health Department, Henri-Mondor Hospital, University Paris XII, Créteil, France

Introduction

Previous randomized studies have showen that subglottic secretions drainage (SSD) could lower the incidence of ventilator-associated pneumonia (VAP), especially early-onset VAPs. Nevertheless, based on the monocentric design of the available studies, the recent guidelines for prevention of VAP have only suggested, not recommended, the use of SSD.

Methods

Randomized controlled trial conducted in 4 ICUs. Adult patients who were expected to require mechanical ventilation for ≥ 48 hours and had been intubated with the specific Hi-Lo Evac tube® (Mallinckrodt Medical, Ireland) were eligible. Patients admitted after cardiac arrest, drug overdose or acute alcohol intoxication and those with a tracheostomy on ICU admission were excluded. Within the 12 hours after intubation, patients were randomly assigned to discontinuous (SDD+ group) or no (SDD- group) subglottic secretions drainage. Clinical suspicion of VAP was based on previously published criteria. Diagnosis of VAP was confirmed when the quantitative culture of a protected distal sample or bronchoalveolar lavage grew >1000 colony-forming units (CFU/ml) or >10000 CFU/ml, respectively, of at least one microorganism.

Results

333 patients were included, 169 in the SSD+ group and 164 in the SSD− group. Age, gender, SAPS II score (52.9 in the SSD+ group vs 53.8 in the SSD− group) and SOFA score (8.8 vs 8.6) at admission, and percentage of medical patients (84% vs 86.6%) were similar in the 2 groups. 67 episodes of VAP occurred, 25 in the SSD+ group (14.8%) and 42 in the SSD− group (25.6%) (p=0.02), yielding an incidence rate of 17 (SDD+ group) and 34 (SDD− group) episodes of VAP per 1000 ventilator days (p = 0.002). 12 episodes of laryngeal oedema after extubation were observed (8 in the SDD+ group, vs 4 in the SDD− group, p=0.3). Median duration of mechanical ventilation prior to the first episode of VAP (8.0 days in both groups), duration of mechanical ventilation (11.1 days in the SDD+ group vs. 10.9 days in the SDD− group) and ICU mortality (42.0% vs 40.0%) were not statistically different between the 2 groups.

Conclusion

In this multicentre randomized trial, subglottic secretion drainage significantly reduced the incidence of VAP without increasing the risk of laryngeal oedema.

0388

Influence of Intensive Care Unit Size on Outcome of Severe Sepsis

M. Reinikainen*1, S. Karlsson2, I. Parviainen3, T. Varpula4, E. Ruokonen3, M. Varpula4, T. Ala-Kokko5, V. Pettilä4

1Dept of Intensive Care, North Karelia Central Hospital, Joensuu, 2Dept of Intensive Care, University Hospital, Tampere, 3Dept of Intensive Care, University Hospital, Kuopio, 4Dept of Intensive Care, University Hospital, Helsinki, 5Dept of Intensive Care, University Hospital, Oulu, Finland

Introduction

In many disease categories, higher hospital volumes are associated with improved outcomes. We wanted to find out whether the risk of death from sepsis is influenced by the size of the intensive care unit (ICU).

Methods

In the Finnsepsis study, all ICU patients (n = 4500) admitted to 24 ICUs during a 4-month period (1 Nov 2004 – 28 Feb 2005) were screened. The criteria for severe sepsis were fulfilled in 470 patients. 18 patients were excluded because of treatment in more than one ICU. We divided the units into three groups: university hospital ICUs (230 patients), large non-university hospital ICUs (145 patients) and small non-university hospital ICUs (77 patients).

Results

There were no significant differences between the ICU groups in mean severity of illness (SAPS II scores). The overall hospital mortality rate was 29.2%. In post-operative patients, the hospital mortality rate was 22.9% for patients treated in large ICUs (including university and large non-university hospital ICUs) but 42.3% for patients treated in small ICUs, P = 0.045. The survival curves of post-operative patients are shown in the figure. In medical patients, there were no differences between ICU groups in patient outcomes. Open image in new window

Conclusion

Treatment of surgical patients with severe sepsis in small ICUs was associated with increased mortality.

Grant Acknowledgement

Finnish Society of Intensive Care.

0389

Are We Able to Predict the Quality of Life 1 Year After Intensive Care ?

P. Merlani*1, M. Verdon1, T. Perneger2, B. Ricou1

1Service of Intensive Care, 2Quality of Care Unit, Geneva University Hospitals and the University of Geneva, Geneva, Switzerland

Introduction

Research on the outcome of critically ills was mainly focused on mortality. However patients are more concerned about their future quality of life (QOL). Future QOL is often central in the decision making of limiting treatments. We tested the ability of patients (P), families (F), nurses (N) and physicians (Ph) to predict the QOL of the patient 1 year after discharge of ICU.

Methods

We included adults admitted to our surgical ICU, who stayed >36h in the unit and consented to participate to the study. We collected patient's characteristics and ICU data. At ICU discharge we asked the patient, the family, the attending nurse and the senior physician to predict the QOL of the patient at 1 year after ICU. After 1 year, we re-contacted the patient to obtain his/her actual QOL. QOL was assessed by Euro-QOL: 5 dimensions EQ-5D and the visual analogue scale (EQ-VAS).

Results

We included 762 among 3723 screened patients and obtained data at 1 year after ICU from 642 (84%) patients. 579 (76%) had finally survived and were analysed. The interclass correlations between the EQ-VAS at 1 year and the prediction by P (0.389), F (0.392) and N (0.330) were bad, and the worst by Ph (0.196). The agreement by Bland and Altmann between the prediction by P, F, N, Ph and the measured QOL was poor (+6±40,+6±39,+3±44,−2±44 with P, F, N optimistic and Ph pessimistic). Regarding EQ-5D: in Mobility, Self-care, Activity, the agreement (and Kappas) were satisfactory, but always better for P and F than N and Ph. Regarding Pain and Anxiety, the agreement and Kappas were worse, with no statistical correlation between the predictions by N and Ph and the measured QOL. Correlations and agreements were not significantly affected by types of diagnosis or degrees of QOL at 1 year. Especially in elderly patients ([65y), there was no significant correlation between EQ-VAS predicted by Ph or any of the EQ-5D predicted by Ph and the measured QOL after 1 year.

Conclusion

Patients and families were very imprecise in their prediction of future QOL at 1 year after ICU. Nurses, and particularly physicians were worse in their prediction of QOL of patients, especially regarding the elderlies, that was not correlated to the measured QOL. Caregivers should be very careful when trying to integrate the notion of future QOL in therapeutic decisions.

Grant Acknowledgement

This work was supported by the Swiss National Science Foundation 3200B0-100789, the Käthe-Zingg-Schwichtenberg Fonds (ASSM), the Société Académique de Gene`ve and the Fonds de péréquation Recherche et Developpement des HUG.

0390

Amplitude Integrated EEG (aEEG) Predicts Outcome in Hypothermia Treated Cardiac Arrest Patients

M. Rundgren*1, E. Westhall2, T. Cronberg3, I. Rosén4, H. Friberg1

1Anaesthesia and Intensive Care, Lund University Hospital, Lund, Sweden, 2Neurophysiology,, 3Neurology, 4Neurophysiology, Lund University Hospital, Lund, Sweden

Introduction

The amplitude integrated EEG (aEEG) pattern has shown good correlation to outcome in neonates exposed to asphyxia. Studies evaluating its value in adult cardiac arrest patients are scarce. We have studied the aEEG-patterns and evolution of aEEG-patterns in 101 adult hypothermia treated cardiac arrest patients and correlated the major patterns to outcome.

Methods

From feb 2004 to feb 2008, 101 consecutive hypothermia treated cardiac arrest patients were monitored using the Nervus®-monitor. The monitor was applied on arrival in the ICU, and data was linked to the department of neurophysiology, where assessment was made without knowledge of patient outcome. Patients were sedated using propofol and fentanyl during hypothermia treatment. The evolution of aEEG-patterns at normothermia was correlated to patient outcome (regaining consciousness). Six months evaluation is ongoing.

Results

101 patients were monitored, 6 died from circulatory failure before normothermia was resumed, and in one case registration was lost, leaving 94 registrations for evaluation. The patients had a mean age of 61 years, 80% were male, the majority of the cardiac arrests were witnessed (83%) and out-of-hospital (84%). The initial rhythm was ventricular fibrillation in 58% of cases. Normothermia was resumed at a mean of 36 h after cardiac arrest. The aEEG-pattern at start of registration did not correlate to patient outcome. At normothermia, 28 patients had a flat or suppression-burst aEEG, two of whom regained consciousness. Fifty-two of 56 patients with a continuos aEEG without electrographic status epilepticus regained consciousness. Twenty-six patients developed electrographic status epilepticus, 10 from a continuous background pattern, and 16 from a discontinuous pattern. Only one regained consciousness.

Conclusion

Electrographic status epilepticus is common among hypothermia treated cardiac arrest patients and the outcome is poor. aEEG-monitoring should be performed routinely and treatment initiated as early as possible. A continuous pattern at normothermia is strongly correlated to recovering consciousness.

Grant Acknowledgement

This study was supported by Skane county council's research and development foundation and by the governmental funding of clinical research within the National Health Service.

Oral Presentations Microcirculation in sepsis: 0391–0395

0391

Microcirculation and Vascular Reactivity During Development of Endotoxin Tolerance in Humans

Draisma*1, R. Bemelmans2, J. G. van der Hoeven1, P. Spronk2, P. Pickkers1

1ICM, Radboud University Nijmegen Medical Centre, Nijmegen, 2ICM, Gelre hospital, location Lukas, Apeldoorn, Netherlands

Introduction

Changes in microcirculation and mitochondrial dysfunction appear to be key mechanisms in sepsis, since they can lead to regional mismatch of oxygen supply and demand. Lipopolysaccharide (LPS, endotoxin) can be used to induce endotoxemia as a model of inflammation in humans, but the effects on microcirculatory perfusion have not been tested before. We were particularly interested in microcirculatory changes during repeated administrations of LPS when endotoxin tolerance developed. The aim of our study was to compare 3 methods to investigate microcirculation and vascular reactivity during endotoxemia and endotoxin tolerance.

Methods

Endotoxin tolerance was induced in 9 volunteers by intravenous injection of 2ng/kg/day lipopolysaccharide on 5 consecutive days. Microcirculation and vascular reactivity was monitored before (t=0) and after (t=2 and 4 hrs) LPS administrations on day 1 and 5. Near Infrared Spectroscopy (NIRS) measured thenar muscle tissue saturation before, during and after arterial occlusion (ischemia) was induced by inflating a cuff above the elbow to 50mmHg above systolic blood pressure for 90 seconds. Orthogonal Polarization Spectral imaging (OPS) measured microvascular perfusion sublingually using side stream darkfield imaging in small, medium and large sized microvessels. Forearm blood flow (FBF) was measured by strain-gauge plethysmography during local intra-arterially infusion of endothelial-dependent vasodilatory acetylcholine.

Results

Endotoxin tolerance developed during 5 consecutive days of LPS administration as demonstrated by the attenuated release of pro-inflammatory cytokines and absence of symptoms and fever on the fifth day. Both NIRS and OPS did not demonstrate changes in microcirculation during endotoxemia or endotoxin tolerance in vivo. FBF measurements showed an acetylcholine dose-dependent attenuation (from 5.7±2.6 to 2.6±2.0) after the first administration of LPS (p=0.01), but not after acetylcholine stimulation when tolerance was present (from 6.1±2.1 to 5.8±3.7) on day 5(p=0.25).

Conclusion

In this human endotoxin tolerance model, no significant effect of repeated LPS administrations on microcirculatory perfusion could be observed using NIRS (thenar muscle) and OPS (sublingual). The forearm blood flow however, was attenuated after the first administration of LPS, indicating an endothelial dysfunction in endotoxemia in vivo. This attenuated response to acetylcholine was not present after the fifth administration of LPS, which may indicate that tolerance is also present at the levels of the endothelium.

0392

Microvascular Density is Strongly Correlated to Oxygen Extraction and Cardiac Output in Septic Shock Patients

A. Bruhn*, C. Ruiz, G. Izquierdo, R. Lopez, M. Andresen, G. Hernandez

Departamento de Medicina Intensiva, Pontificia Universidad Catoólica de Chile, Santiago, Chile

Introduction

Several authors have reported that there is no relation between systemic hemodynamics and microcirculation1,2. These reports have focused on microvascular flow but not density. In contrast, a recent study by Trzeciak3 showed that microvascular density was inversely related to mixed venous oxygen saturation (SmvO2). The goal of this study was to determine if SmvO2 and cardiac output are related to sublingual microcirculation in SS patients.

Methods

By using side dark field videomicroscopy (Microscan ®, Microvision medical) we evaluated sublingual microcirculation in 10 SS patients who had already been fluid resuscitated, within the first 24 hours after admission to ICU. Each patient's microcirculation was evaluated looking at 3 to 6 different sublingual areas (10–20 seconds/image). Simultaneously, we assessed systemic hemodynamic parameters (MAP, mean arterial pressure; NA, Noradrenaline dose; CI, cardiac index and SmvO2). Images were analyzed by semiquantitative scores of flow (MFI, Mean flow index and PPV, proportion of perfused vessels) and density (TVD, total vascular density; PVD, perfused vascular density) of small vessels (<20µm). Correlations between parameters were determined by Pearson coefficient and considered significant if p < 0.05.

Results

We found that SmvO2 and CI are inversely correlated to microvascular density, but not to scores of microvascular flow. MAP and NA were not correlated to density and flow scores. Open image in new window

Conclusion

Sublingual microvascular density is strongly correlated to oxygen extraction and cardiac output and may be a critical determinant of systemic hemodynamics.

0393

Both Passive Leg Raising and Volume Expansion Improve Sublingual Microcirculation in Preload-Dependent Septic Patients

J. Pottecher*1, S. Deruddre1, J. F. Georger2, E. Vicaut3, D. Benhamou1, J. L. Teboul2, J. Duranteau1

1Département d'Anesthésie-Réanimation, 2Service de Réanimation Médicale, CHU Bicêtre, Le Kremlin-Bicêtre, 3Laboratoire Etude de la microcirculation, CHU Fernand Widal, Paris, France

Introduction

We conducted a prospective study to assess sequential sublingual microcirculatory changes associated with passive leg raising (PLR) and volume expansion (VE) in preload-dependent septic patients.

Methods

After IRB approval, 25 ventilated septic patients with suspected fluid-responsiveness (respiratory variation in pulse pressure (DeltaPP>13%)) underwent a 5 steps trial. At Baseline, systemic hemodynamics and sublingual microcirculatory parameters (MicroScan®, MicroVision Medical, The Netherlands) including the proportion of perfused vessels (PPV) were acquired. Measurements were repeated after PLR, when returning to baseline position (Post-PLR), when VE induced the same DeltaPP value than PLR (Idem PLR) and at the end of VE (Figure). Values were analyzed with Friedman, Wilcoxon and Spearman tests.

Results

Twenty septic shock and 5 severe sepsis patients (57±17 years, APACHE II : 23±7, SOFA : 11±4) were included within the first day following sepsis onset. Both PLR and VE improved sublingual microcirculation (increased PVP) but microcirculatory parameters did not correlate with systemic hemodynamics. Open image in new window

Conclusion

In preload-dependent septic patients, PLR and VE induced equal sublingual microcirculatory improvement. This result suggests that VE-induced microcirculatory changes were rather due to hemodynamic effects (systemic or local) than rheologic ones.

0394

Fluid Resuscitation Improves Sublingual Microcirculatory Flow in Septic Shock Patients

C. Luengo*1, M. R. Losser1, M. Legrand1, P. Goedhart2, C. Ince2, D. Payen1

1Anesthesiology and Critical Care, Lariboisiére Hospital, AP-HP, Paris Diderot University, Paris, France, 2Department of Physiology, Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands

Introduction

Impaired microcirculation in septic shock (SS) is frequent and has been related to outcome. Fluid challenge (FC) is a hallmark of resuscitation and shifts hypodynamic to hyperdynamic shock at the macrocirculation level. Since microcirculation may behave independently from macrocirculation, little is known about the relation between macro- and micro-circulation after fluid challenge. In this study, we used SDF (1, MicroScan, MicroVision Medical Amsterdam, NL) imaging to assess the sublingual microcirculation in patients with SS. We hypothesized that microcirculatory blood flow may change after fluid challenge relatively independently from macrohemodynamic response.

Methods

Fourteen patients (mean age 64±17 yrs) with SS of less than 48 hrs, mechanically ventilated and receiving norepinephrine, underwent a fluid challenge (500ml saline over 15 min). SAPS II and SOFA score were collected. SDF imaging of the sublingual microcirculation were performed in each patient, at 3 different spots, before and after FC. Heart rate (HR), invasive mean arterial pressure (MAP), central venous pressure (CVP), and cardiac output (CO) were registered at the same time.

Movie clips of the microcirculation were analysed by 2 independent researchers in a blind manner. Microcirculatory flow Index (MFI) was calculated on a semi-quantitative basis according to the Boerma score (2). Small vessels (diameter<20µm) changes were considered for statistical analysis using Student's t-test. p < 0.05 was considered significant. Results are shown as mean±SD.

Results

SAPS II score was 56 ± 18 and SOFA score was 11 ± 4. MFI was constantly impaired at baseline. HR, MAP, CO and CVP did not change significantly before and after the fluid challenge (100 vs 99 bpm, p= 0.68; 75 vs 79 mmHg, p=0.27; 6.9 vs 7.5 l/min, p=0.31 and 10 vs 14 mmHg, p=0.12, respectively). Small vessels MFI score improved significantly after fluid resuscitation (2.1 vs 2.5, p= 0.004). Open image in new window

Conclusion

In this cohort, fluid challenge improved microcirculatory flow despite no significant effect on systemic hemodynamic variables. This suggests that microcirculatory parameters are more sensitive in assessing responsiveness to fluid than macro-parameters. Further research should be performed to better clarify the kinetic and the threshold for micro- and macro-hemodynamic modifications in SS and to evaluate such a functional test for outcome prediction.

Grant Acknowledgement

Plan quadriennal EA322.

0395

Microcirculatory Changes after Activated Protein-C Infusion in Severe Sepsis

A. Donati*, M. Romanelli, L. Romagnoli, L. Botticelli, C. Anastasi, P. Pelaia

Anesthesia and Intensive Care Unit, Marche Polytechnical University, Ancona, Italy

Introduction

In sepsis the link between the systemic inflammatory response and the development of MOF is represented by Microcirculatory and Mitochondrial Distress Syndrome (MMDS) that causes an important cellular impairment of aerobic methabolism not corrigible exclusively with a restoration of a normal hemodynamics and oxygen delivery. We observed the changes caused by MMDS in the tissue oxygenation and in the microcirculation with NIRS and microcirculatory analysis in patient with severe sepsis or septic shock, before, during and after Activated Protein C (APC) infusion. We evaluated if APC influences tissue saturation (an index of O2ER) and if alterations of hemodynamics are linked to these changes.

Methods

Prospective observational study. We evaluated ten septic patients treated with APC from December 2005 to September 2007. Microcirculation images were registered and analyzed by the SDF technic and MAS software that calculate the MFI (Microvascular Flow Index), mean velocity and FCD (Functional Capillary Density) parameters. We carried out evaluation with NIRS of the StO2 with the spectrometer InSpectra (Hutchinson Technology Inc., Minn) putting a probe of 15mm in the brachioradialis muscle of the patients. The measurements were made in 5 steps: pre-APC, at 24h, 48h, 72h and 24h after the end of the infusion (post-APC). Each measurement (of the basic StO2 and of the slope during and after the ischemia) was registered and transformed from the software InSpectra Analysis. The parameters analyzed with the non-parametric test of Wilcoxon for repeated measurement (P<0.05).

Results

MICROCIRCULATION: There was a significant increase of MFI and of FCD started after 72 h from the beginning of infusion for small and medium vessels (P<0.001) and an increase of mean velocity at 72h for small vessels and in post-APC for small and medium vessels (P<0.05)

NIRS: The increase of the basal StO2 during and after APC treatment and its decrease during the arterial occlusion are statistically significant (P<0.05). The increase of the StO2 slope after arterial occlusion is statistically significant starting from the second day of infusion of APC (P<0.05).

Conclusion

Microcirculation analysis throw SDF technic let visualize the significant modifications which begin in septic syndrome (low capillary density, reduced flow velocity, heterogeneity of the microcirculation) and also the improvements of the microcirculatory state during the APC infusion.

There is an improvement of all the NIRS parameters after the APC infusion, which means an increase of tissutal O2ER. We have to verify if that increase is linked either with a reduced shunt effect in the microcirculation or with the end of metabolic down-regulation that involves the mitochondrial system. APC seams improve microcirculatory state in septic patients acting most of all on small vessels vases. NIRS and the microcirculation monitoring are useful tools to evaluate therapy and the outcome of severe sepsis and septic shock.

Oral Presentations Perioperative organ failure: 0396–0400

0396

Application of the Sequential Organ Failure Assessment Score in Prediction of Morbidity and Mortality after Cardiac Transplantation

N. Qedra*1, T. Gromann2, K. Yilmaz2, M. Kadry2, H. Kuppe2, H. Kuppe3, H. Lehmkuhl2, R. Hetzer2

1The Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, 2The Department of Cardiothoracic and Vascular Surgery, 3The Department of Anesthesiology, Deutsches herzzentrum berlin, Berlin, Germany

Introduction

BACKGROUND: The sequential organ failure assessment score SOFA has been shown to predict mortality and morbidity in heterogeneous cardiac surgical patients but not after heart transplantation (HTx) (1, 2, 3). As patients after HTx need early postoperative catecholamines we evaluated the application of SOFA in prediction of 30-day mortality and morbidity following HTx.

Methods

We retrospectively studied 126 consecutive heart transplant recipients (age: median 47, 12–70 years). The SOFA was calculated postoperatively and daily until intensive care unit (ICU) discharge or for a maximum of 7 days. C-reactive protein (CRP) values and white blood cell (WBC) count were reviewed. Lengths of ICU stay and 30-day mortality were assessed.

Results

From the 1st until the 7th postoperative day (POD) only SOFA values, not CRP or WBC counts, were significantly higher in non-survivors (12.5%) than in survivors (Mann-Whitney test: p<0.001). For SOFA area under the receiver operating characteristic curve (ROC-AUC) for risk of 30-day mortality at ICU admission was 0.90 (95% CI 0.83 to 098). The highest value (0.94, 95% CI 0.88 to 0.99) was reached on the 4th POD. A SOFA value of > 12 points as a predictor for 30-day mortality had a specificity of 79% and sensitivity of 88%. In survivors the maximum of SOFA, but not of CRP or WBC counts, correlated significantly with the length of ICU stay (p < 0.001).

Conclusion

Although patients after HTx need catecholamines in the early postoperative period, SOFA can be used to grade the severity of morbidity and to identify the risk of 30-day mortality without specific modifications. As an independent score, SOFA is therefore helpful in early therapeutic decision making and resource planning in heart transplant recipients.

0397

Respiratory Tract Infections after Cardiac Surgery

M. Riera*1, J. Herrero1, J. Ibáñez1, F. Enriquez2, J. Saez de Ibarra2, O. Bonnin2

1Intensive Care Unit, 2Cardiac Surgery, Son Dureta Hospital, Palma, Spain

Introduction

Nosocomial pneumonia (NP) and tracheobronchitis (TBX) after cardiac surgery are associated with worse outcomes. The aim of this study was to identify risk factors associated with NP and TBX after cardiac surgery and to determine their impact on mortality and morbidity.

Methods

Retrospective observational cohort study of 1600 adult patients operated under extracorporeal circulation and who stayed in the ICU for more than 24 hours. NP was diagnosed in accordance with the American Thoracic Society guidelines. All NP and TBX episodes were confirmed by a quantitative culture of endotracheal aspirate. Univariate and logistic regression analysis were done.

Results
The frequency of NP in our population was 1.2% (15.6 episodes per 1,000 days of mechanical ventilation) and that of TBX was 1.6% (21 episodes per 1,000 days of mechanical ventilation). The duration of mechanical ventilation with NP was longer (median 21.8 days, IQR 8.9 – 35) than without respiratory tract infections (median 5.5 hours, IQR 4.1 – 10). Significant independent risk factors for respiratory tract infections are summarized in Table 1. The median length of stay in the ICU and hospital were significantly longer in patients with NP (30 and 42 days respectively) and TBX (10 and 28 days) than in patients without any respiratory tract infection (3 and 11 days, p < 0.0001). Mortality in patients without respiratory tract infection was 0.9% (14 of 1555), in patients with NP was 42% (8 of 19) and 12% (3 of 25) in TBX, differences that were statistically significant (p < 0.0001).
Table 1

Risk Factors for Respiratory Tract Infections (NP and TBX, N = 44).

 

Relative risk (95% CI)

P Value

LV Ejection Fraction <30%

4.7 (1.9 – 11.4)

0.001

Urgent surgery

4.3 (2.2 – 8.6)

< 0.0001

Chronic Renal Failure

4.2 (2.0 – 8.8)

< 0.0001

Constant

0.003

< 0.0001

Conclusion

Patients undergoing cardiac surgery have a low frequency of NP and TBX. Both respiratory tract infections are associated with a poor prognosis.

0398

High Tidal Volume as an Independent Risk Factor for Acute Lung Injury after Cardiac Surgery

D. Pasero*1, A. Davi1, F. Guerriero1, N. Rana1, G. Merigo1, I. Mastromauro1, S. Viberti1, L. Mascia1, M. Rinaldi2, M. Ranieri1

1Department of Anesthesia and ICM, 2Department of Cardiovascular Surgery, Ospedale S.Giovanni Battista, University of Turin, Turin, Italy

Introduction

Despite a widespread use of fast-track protocols, patients undergoing cardiopulmonary by pass (CPB) may need prolonged mechanical ventilation, based with high tidal volume (10–15 ml/kg) and low positive end expiratory pressure (PEEP). Experimental and clinical data showed that this kind of ventilation might increase pulmonary and systemic inflammation in patients with acute lung injury (ALI), and this has been recently suggested in patients undergoing CPB.

Our hypothesis was that mechanical ventilation with large tidal volumes during the post operative period was a risk factor that may contribute to the development of acute lung injury.

Methods

A prospective observational study was conducted on patients undergoing CPB. Exclusion criteria were age < 18 years old, off pump surgery, heart and lung transplantations.

Results
Among 307 patients, admitted to the cardiac ICU from April to September 2007, 200 met inclusion criteria. 14 (7%) developed ALI on day 3.7±1.4. Multivariate logistic regression analysis adjusted for baseline patient characteristics (age, gender, Simplified Acute Physiology Score (SAPS II)) and underlying peri-operative ALI risk factors (CPB, blood transfusions, total time of surgery), identified high tidal volume, SAPS II, incremental PEEP level and total time of surgery as risk factors for the development of ALI (Table).

Variables

OR

CI Lower

CI Upper

P

SAPS II

1.221

1.065

1.400

.004

Tot Time Surg

1.009

1.002

1.018

.020

Mean TV/PBW

2.038

1.025

4.054

.042

PEEP

2.085

1.269

3.425

.004

Conclusion

Together with SAPS II and total time of surgery, high tidal volume and incremental level of PEEP represent independent risk factors of acute lung injury in patient undergoing CPB.

0399

Validation of Cardiac Output Measurements with the Lidco Pulse Contour System in Patients with Reduced Left Ventricular Function after Heart Surgery

B. Mora*, I. Ince, A. Moritz, B. Birkenberg, B. Steinlechner, M. Dworschak

Cardiothoracic and Vascular Anaesthesia and ICM, Medical University of Vienna, Vienna, Austria

Introduction

The LiDCO™ is a new continuous, minimally invasive method that employs the pulse contour technique to determine cardiac output (CO). In swine (1) and patients (2,3) with normal ventricular function there seems to be good agreement between LiDCO™ and the thermodilution CO method using a pulmonary artery catheter (TDCO). However, LiDCO™ has not yet been validated in patients with a compromised left ventricular ejection fraction (LVEF) after heart surgery.

Methods

After institutional approval and obtaining informed consent we studied 27 ASA IV patients with a LVEF < 40%. Postoperatively multiple duplicate CO measurements were carried out at various haemodynamic states with the TDCO and the LiDCO™ technique. The correlation coefficient was determined by a simple regression analysis and mean bias as well as the upper and lower limits of agreement were calculated. Open image in new window

Results

A total of 204 measurements were performed. TDCOs ranged from 2.3 to 11.2 and those determined by LiDCO™ from 2.8 and 10.6 L/min. The correlation coefficient r2 between these two methods was 0.59 (P < 0.05). Mean bias as well as lower and upper limits of agreement (i.e. mean bias ± 2 SD) were 0.32, −2.22 and 2.86, respectively. 96% of the differences between the two methods lied between the limits of agreement.

Conclusion

These preliminary results suggest that LiDCO™ seems to correlate fairly well with TDCO even in patients with a reduced pump function after cardiac surgery. In spite of the wide range between the limits of agreement LiDCO™ may thus be an alternative to TDCO in cases were insertion of a pulmonary artery catheter is either contra-indicated or not considered opportune.

0400

The Postoperative Vasopressin and Copeptin Response in Noncardiac Surgery Patients: A Prospective, Controlled Trial

S. Jochberger*1, G. Luckner1, V. Mayr1, C. Torgersen1, V. Wenzel1, H. Ulmer2, W. Hasibeder3, N. Morgenthaler4, M. Dünser1

1Anaesthesiology and Critical Care Medicine, 2Department of Medical Statistics, Informatics and Health Economics (MSIG), Innsbruck Medical University, Innsbruck, 3Anaesthesiology and Critical Care Medicine, Krankenhaus der Barmherzigen Schwestern, Ried i. Innkreis, Austria, 4Department of Research, B.R.A.H.M.S. Aktiengesellschaft, Henningsdorf, Germany

Introduction

The systemic inflammatory response syndrome (SIRS) refers to the general activation of the immune system by non-infectious stimuli, which may cause cardiovascular failure and subsequent multiple organ dysfunction (1). Comparable to the clinical presentation of septic shock, hemodynamic instability associated with SIRS is characterized by hypovolemia and peripheral vasodilatation with or without myocardial dysfunction (2). Further information on the endogenous arginine vasopressin (AVP) response in patients with the postoperative systemic inflammatory response syndrome (SIRS) and vasodilatory shock would give more insight into the pathophysiology of SIRS-associated cardiovascular failure and help to indicate AVP therapy.

Methods

Patients after uncomplicated abdominal surgery without SIRS (n=10), critically ill patients after non-cardiac surgery with SIRS (n=9) and patients with SIRS plus vasodilatory shock (n=22) were included into this prospective trial. Plasma AVP (radioimmunoassay) (3) and copeptin (immunoluminometric assay) (4) concentrations together with clinical parameters were documented daily during the first seven postoperative days.

Results

The AVP response significantly differed between the three groups. Patients without SIRS had lower AVP concentrations than SIRS patients with (p=0.001) or without shock (p=0.003). Patients with SIRS and shock had higher AVP levels than patients with SIRS alone (p < 0.001). Without a group difference, AVP decreased over time (p=0.007). Non-survivors had higher AVP levels than survivors at day 28 (p < 0.001). In SIRS patients without shock, serum osmo-larity was indirectly associated with AVP levels, while mean arterial blood pressure and se-rum osmolarity were associated with AVP in SIRS patients with shock. AVP and copeptin correlated significantly with each other (p < 0.001; r=0.76). In patients without hemofiltration, copeptin levels could predict 28 day mortality with a high sensitivity and specifity.

Conclusion

The postoperative AVP response in non-cardiac surgery patients seems well preserved. A contributory role of AVP to the failure to restore the vascular tone in patientswith vasodilatory shock cannot be excluded but seems less important than in septic or postcardiotomy shock.

Oral Presentations Paediatric intensive care: 0401–0405

0401

Measured and Predicted Energy Expenditure and Delivered Energy in Critically Ill Children after Extracorporeal Circulation

J. Depeyre*1, C. Jotterand1, M. Ranson2, M. Perez3, P. Stücki3, J. Cotting3

1Dietetics, University of Applied Sciences of Western Switzerland, Carouge, 2Clinical Nutrition Unit, 3Pediatric Intensive Care Unit, University Hospital, Lausanne, Switzerland

Introduction

In pediatric intensive care units, energy needs are complex to determine. Extracorporeal circulation (EC) that induces an inflammatory state could influences energy expenditure. The aim of this study was to compare energy expenditure measured (REEm) by indirect calorimetry with theoretical REE and energy intake (EI) in patients with and without EC.

Methods

Children with expected mechanical ventilation ≥72 hours and a FiO2 ≤60% were prospectively studied and separated into 2 groups according to EC or not. REEm was measured daily by indirect calorimetry until extubation; results were compared to predicted REE by the equation from Schofield WH (1985). EI were recorded daily and energy balance was calculated. Comparisons were tested by the t-test and the Bland-Altman method.

Results

62 critically ill children, 36 males and 26 females were included. In the 24 EC patients, mean age (±SD) was 19.8±20.1 months with a body weight and body size of 8.0±4.3 kg and 72.7±18.1 cm, respectively. In the 38 non EC patients, mean age was 22.5±1.2 months with a body weight and body size of 9.8±5.1 kg and 77.3±20.7 cm, respectively. These results did not differ between groups. On day 1, CRP was 110.5±65.1 mg/dl in EC and 78.8±87.7 mg/dl (ns) in non EC. 337 indirect calorimetry measurements (5.4±2.5 per patient) were performed; 138 in EC and 199 in non EC. REEm was higher in EC compared to non EC with 57.9±12.3 kcal/kg/d (median 57.5 range: 31.5 to 97.0 and IQ 47.6 to 65.8 kcal/kg/d) in EC and 53.7±10.8 kcal/kg/d (median 52.9, range: 27.8 to 94.1 and IQ 45.6 to 60.3 kcal/kg/d) in non EC (p<0.01). Respiratory quotient was 0.84±0.1 in both groups. Schofield estimated REE correctly for both groups. The mean bias was −1.4±14.4 kcal/kg/d in EC (ns; 95% CI −3.8 to 1.0) and 0.3±14.0 kcal/kg/d in non EC (ns; 95% CI −1.6 to 2.3). During the stay, EI was 47.0±27.0 kcal/kg/d and 52.2±24.3 kcal/kg/d in EC and non EC, respectively (p < 0.001). The balance was strongly negative during the first 6 days in EC and during the first 2 days in non EC. The balance became clearly positive since day 8 in EC and day 7 in non EC.

Conclusion

Extracorporeal circulation increases slightly REE in critically ill children, but this rise does not meet the criteria of hypermetabolism (REEm>10% predicted REE). The Schofield equation predicts very accurately REE in both groups. The energy balance remains negative longer in patients with EC compared to patients without EC.

0402

Coulter LH750: Clinical Usefulness of Neutrophil VCS Research Population Data in the Diagnosis of Sepsis in Neonates — Correlation with ITR

E. Grimaldi1, F. Scopacasa1, F. Raimondi2

1Dip. di Biochimica e Biotecnologie Mediche, Università degli Studi Federico II, 2Dip. di Pediatria — Sezione di Neonatologia, Università degli Studi Federico II, Napoli, Italy

Introduction

Neutrophils (NE) and bands counting are auxiliary test in detection of bacterial infections, even if the band count is a low-specificity test with high inaccuracy (1). In neonates absolute NE count is one of the Rochester criteria ad is also used to obtain Immature/mature NE ratio (ITR), a useful index in neonatal sepsis diagnosis. ITR has been considered to have an high variability in terms of sensibility (SE) and specificity (SP)(2). Coulter LH700 series hematology analyser provides CBC/diff data, including 24 research population data (RPD) referred to mean and standard deviation of Volume, Conductivity ans Scatter (VCS) measures of WBC subpopulations. Several experiences have been reported about the clinical usefulness of NE RPD in detecting sepsis in adults (3,4). We daily analyze samples of the neonatal intensive therapy providing CBC/Diff and ITR to those with sepsis suspect. The aim of this study was to evaluate the clinical usefulness of NE RPD in detecting neonates’ sepsis in order to find a time-saving test with high SE and SP.

Methods

We collected Coulter LH750 data of 62 neonatal samples with suspect of sepsis and 36 age-matched normal samples. ITR was calculated with 200-cell count and sepsis was assessed with blood culture.

Results
Results are reported in Table 1. NE RPD values of septic patients, except SDS, are statistically different from normal. On the basis of sepsis diagnosis we analysed the performances of ITR and NE RPD. The current cut-off of 0.2 for ITR has SE=53.8%, SP=77.6%, NEVM=147 has SE=92.3%, SP=65.3% and VSD=26.83 has SE=84.6%, SP=63.3%. Since our major goal is the detection of sepsis in non-neutrophilic patients, we studied the combination of NE#<5.5 and NE RPD. VM=148 has SE=100%, SP=72.5% while VSD=26.83 has SE=100%, SP=70.0%.
Table 1

NE RPD

 

Cut-off

SE%

SP%

AUC

NE#

5.5

46.2

81.6

0.68

ITR

0.2

53.8

77.6

0.79

VM

147

92.3

65.3

0.83

VSD

26.8

84.6

63.3

0.71

VM (NE#<5.5)

148

100

72.5

0.91

VSD (NE#<5.5)

26.8

100

70

0.82

Conclusion

A fast and correct diagnosis of sepsis is essential in the Intensive Care treatment of neonates and our data clearly show that NE RPD of Coulter LH700 series improve the process of detecting sepsis in neonates. RPD have better performance than the current tests in signalling the suspect of sepsis. Moreover they represent a time-saving and faster test than ITR determination.

0403

Audit of Different Ways of Estimating Blood Glucose Levels in Paediatric Intensive Care

H. C. Hill*1, L. Woodgate2, P. Barton3, P. B. Baines2

1Paediatric Intensive Care, Royal Liverpool Children's NHS Trust, 2Paediatric Intensive Care, 3Biochemistry, Royal Liverpool Children's NHS Trust, Liverpool, United Kingdom

Introduction

Aggressive ‘tight’ control of glucose improves outcome in randomised controlled trials in critically ill adults. However, there are concerns that the adverse consequences of hypoglycaemia may outweigh the benefits of tight glucose control in paediatrics. Prior to starting a randomised trial of aggressive glucose control in the Paediatric Intensive Care Unit (PICU) the results from 3 different methods for measuring glucose in critically ill children were compared.

Methods

The glucose concentrations obtained from automated laboratory equipment, ward based blood gas analyser(BGA) and bedside glucometer (MediSense)were compared from critically ill children having routine blood tests, in a 20 bed regional PICU in England.

Results

101 routine samples of blood were analysed to determine the glucose concentrations when measured by automated laboratory equipment, by ward based blood gas analyser (BGA) and by bedside glucometer (MediSence),a total of 303 analyses. In the absence of a gold standard, these measurements were compared one with another. Only 53.5% of the glucometer and blood gas analyser (BGA) estimations of glucose (n=101) and 56.4% of glucometer and laboratory measurements (n=101) were within 1.0 mmol/l of each other. 86% of the comparisons between the BGA and the laboratory measurements were within 1.0mmol/L (n=101).

Conclusion

There is considerable disagreement between the glucose concentrations measured using the different methods. This may produce difficulties with attempts to aggressively control glucose.

0404

Reducing Adverse Drug Events on Paediatric Intensive Care — are We Winning?

K. J. Parkins1, N. Reilly*2, S. Ellis1

1Paediatric Intensive Care, 2Pharmacy, Royal Liverpool Children's NHS Trust, Liverpool, United Kingdom

Introduction

Data from national audits has previously shown approximately 10% of patients admitted to hospital experience an adverse event. Adverse drug events have the potential to cause significant morbidity and mortality. The PICU at RLCH admits approximately 1000 patients per year from neonates to 18 years and as drugs are prescribed on a weight or surface area basis the risk of adverse drug events is high.

Methods

Using the last 3 years of PICU adverse event database and a Pharmacy intervention audit (June 2006) we identified high alert drugs. Assessed whether awareness of events and interventions are reducing incidents.

Results
Top 5 groups of drugs identified requiring intervention: Prophylactic Cardiac antibiotics, sedation and analgesia, Potassium chloride, dinoprostone and inotropes, and heparin. Interventions identified as effective: change in written guidelines for cardiac antibiotics, sedation and analgesia, dinoprostone and heparin. We have simplified guidelines (cardiac antibiotics, heparin and sedation/analgesia), and changed unit policy to use of standard concentrations of inotropes/prostin. Each change in guidelines has been backed up with notification and training by our education team. All events involving nursing staff are flagged to our nurse education team so that individual training needs can be met, and any general training issues identified. To increase awareness of adverse incidents we also use simulation training. This has helped improve recognition of events and awareness of methods to minimise effects. Adverse incidents are now discussed at medical, nursing and AHP new staff induction and at annual mandatory training.
Table 1

Reported Adverse Events Over Last 3 Years on PICU

 

Total adverse events

Numbers drug events

% Drug events

2005

223

84

37.7

2006

344

146

42.4

2007

375

116

30.9

Table 2

Picu Pharmacist Interventions Over 1 Month June 2006

Drug

Number of Interventions (total 109)

% of total interventions

Prophylactic cardiac antibiotics

28

25.6

Sedation/analgesia

27

24.8

Potassium chloride

3

2.8

Vasoactive drugs

4

3.7

Heparin

7

6.4

Conclusion

Simplifying drug guidelines and increasing their availability appear to have helped reduce adverse drug events on PICU. Increasing awareness through multidisciplinary adverse events team (responsible for investigation and feedback) and education and training are also significant factors.

0405

Prospective Incidence Study of Nosocomial Infections in the Picu

I. Kasmi, G. Kasmi, S. Sallabanda, E. Kola, I. Klironomi*, R. Lluka, F. Zavalani, K. Marku, G. Sallabanda, A. Kola

Pediatric, University Hospital Center Mother Theresa, Tirana, Albania

Introduction

Nosocomial infections are important causes of mortality morbidity and prolonged hospital stay in PICU.

Methods

A prospective surveillance study was performed in PICU at Tirana University Hospital during one year period February 2007– January 2008 to describe the epidemiologic profile of NI. CDC criteria were used as standard definitions for NI. Data including extrinsic risk factors invasive devices associated with NI were recorded.

Results

During the study period were admitted 484 patients — 405 no surgical and 79 surgical patients with mean age of 30.4 mo.

The incidence rate of NI during the study period was 10.5 per cent with a total number of NI of 52. 19 surgical patients and 33 no surgical patients had NI.

Urinary tract infections the respiratory tract infections blood stream infections were the most encountered with a rate of 4 per cent 3 per cent 1.4 per cent respectively.

The etiologic profile was as follow — gram negative 71.15 per cent gram positive 23.1 per cent and fungi 5.75 per cent.

Mortality was not influenced by the episodes of nosocomial infections.

The length of hospital stay in infected patients was longer than in no infected patients 19.5 days and 3.6 days respectively P< 0.01.

Conclusion

This study highlights the importance of NI surveillance and the necessity to evaluate preventive measures and design control guidelines which could reduce their consequences.

Oral Presentations Lung imaging and monitoring: 0406–0410

0406

Functional Residual Capacity Guided Alveolar Recruitment Strategy Improves Oxygenation

H. Heinze*, T. Meier, B. Sedemund-Adib, M. Heringlake, W. Eichler

Anaesthesiology, University of Lübeck, Lübeck, German

Introduction

Open endotracheal suctioning may lead to alveolar decrecruitment 1, which can be monitored by means of functional residual capacity (FRC) measurements 2. We hypothesized that a recruitment strategy based on FRC measurements would improve oxygenation after an open endotracheal suctioning manoeuvre.

Methods

We studied 65 postoperative mechanically ventilated cardiac surgery patients. FRC was assessed by oxygen washout using a sidestream O2-analyser (LUFU system (Dräger Medical AG, Luebeck, Germany)). Patients were ventilated with BIPAP, PEEP of 10 cmH2O, and a tidal volume of 6 – 8 ml / kg. All patients received an open suctioning procedure (20 sec, 14 F catheter, 200 mmHg negative pressure). Prospectively patients were stratified into two groups by the post suctioning FRC value (A: FRC >94% of baseline; B: FRC <94% of baseline). Both groups were randomized into either a recruitment manoeuvre (RM) group (PEEP 15 mbar, PIP 35–40 mbar for 30 sec, group RM) or a non RM group, in which venitilation was resumed without a RM (group NRM), resulting in four groups. FRC and paO2/FiO2 (PF-ratio) were recorded at baseline (T1), after suctioning (T2), after RM or NRM (T3), and 30 min after T3 (T4).

Results

B_RM and B_NRM showed lower FRC values at T2 compared to A_RM or A_NRM, which remained lower at T3 and T4 only in B_NRM. At T3 B_NRM showed lower FRC values compared to B_RM. A_RM and A_NRM showed no differences in FRC values (see Figure 1). B_NRM showed lower values of PF-ratio at T3 (A_RM: 392 ± 24 mmHg; A_NRM: 378 ± 21; B_RM: 381 ± 24; B_NRM: 301 ± 27; p = 0.036). This difference did not reach significance at T4 (p = 0.098). Open image in new window

Conclusion

A FRC guided alveolar recruitment strategy after open endotracheal suctioning restored reduced FRC and showed short term improvements of oxygenation in patients after cardiac surgery.

0407

Use of Cine CT to Assess Cyclic Recruitment — Derecruitment (R/D). Effects of Peep, Tidal Volume and Respiratory Rate on R/D

A. Bruhn*, D. Bugedo, F. Riquelme, J. Varas, C. Cabrera, G. Bugedo

Departamento de Medicina Intensiva, Pontificia Universidad Catolica de Chile, Santiago, Chile

Introduction

The use of high PEEP levels has been the main approach advocated to avoid R/D. Nevertheless, a recent clinical trial failed to show any protective effect derived from the use of high compared to moderate PEEP levels. R/D is difficult to assess in patients and the use of static CT may not adequately predict dynamic changes taking place during ongoing mechanical ventilation. The goal of this study was to evaluate the effect of PEEP and tidal volume on R/D using cine-CT during uninterrupted MV.

Methods

We studied ventilated patients fulfilling criteria for ARDS, who underwent a protocol including 4 ventilatory modes: 1)HTV: High tidal volume (TV) (12 ml/kg)/Low respiratory rate(RR) (15/min)/PEEP 9cmH2O, 2) LTV/HRR: Low TV (6ml/kg)/High RR (30/min)/PEEP9 cmH2O, 3)LTV/LRR Low TV (6ml/kg)/Low RR (15/min)/PEEP9 cmH2O, and 4)HP: Low TV (6ml/kg)/High RR(30/min)/High PEEP (15cmH2O). All patients were sedated and paralyzed, connected to a pneumotach (Hans Rudolph, Inc.) including esophageal pressure and proximal airway flow and pressure monitoring. After having a thorax CTscan, a transverse region 2 cm over the diaphragm was chosen and cine-CTs of 8 seconds at this fix transverse region were performed at each ventilator mode (LightSpeed VCT, GE; image reconstruction at 0,25sec-32 images, 0.4 sec / rotation, 16 × 1,25 mm slices, matrix of 512 × 512). Images were analyzed manually with a software (Maluna®, University of Gottingen, Germany), and R/D was determined as the change in non aerated tissue (NAT,-100 to +100HU), along the respiratory cycle, expressed as % of total volume. The effect of PEEP, TV and RR, on R/D and end-expiratory NAT, was evaluated by paired t-tests between the corresponding ventilator modes. A p<0.05 was considered statistically significant. Open image in new window

Results

We have already analyzed 5 patients (52±20y.o, 4 male) who were on their 2nd to 6th day of MV (PaFi 169±25, compliance 24±9 ml/cmH2O). R/D was maximal at HTV (7.8±5.8%) and decreased significantly at lower TVs (LTV/LRR:4.9±4.9%,p=0.02 vs HTV; LTV/HRR:5.0±4.4%, p=0.02 vs HTV). Increasing PEEP while using low TV did not further decrease R/D (HP:4.7±4.6%, p=0.5 vs LTV/HRR). Respiratory rate did not influence R/D at low TV. There was a non significant trend to decreased end-expiratory NAT with HP and HTV compared to both LTV/HRR and LTV/LRR.

Conclusion

Our preliminary data suggest that decreasing tidal volume is the most effective intervention to decrease R/D. At low TV, increasing PEEP adds no further protection against R/D despite a trend to decreased end-expiratory NAT. These data are consistent with results observed in large clinical trials on protective strategies. Cine-CT is a suitable and interesting method to assess R/D during uninterrupted MV.

Grant Acknowledgement

Fondecyt 11060350.

0408

Computed Tomography Correlates of “Best Positive End-Expiratory Pressure” in A Decremental Peep Trial

P. M. Spieth*1, A. R. Carvalho1, P. Pelosi2, A. Barrenengoa1, T. Koch1, M. Gama de Abreu1

University Hospital Dresden, Department of Anesthesiology and Intensive Care Therapy, Dresden, Germany, 2Department of Ambient, Health and Safety, University of Insubria, Varese, Italy

Introduction

Titration of the “best positive end-expiratory pressure” (PEEP) level in patients with acute lung injury (ALI) remain controversial. This work aimed at comparing four commonly used criteria for determination of the best PEEP in experimental ALI and their computed tomography (CT) correlates.

Methods

Ten pigs (30–35.6 kg) were anesthetized, intubated and mechanically ventilated with a tidal volume (VT) of 10 ml/kg, PEEP = 5 cmH2O, FIO2 of 1.0 and respiratory rate (RR) between 10–20/min. ALI was then induced by means of surfactant depletion until PaO2 ≤ 200 mmHg. The VT was then set at 6 ml/kg and RR increased to keep the PaCO2 lower than 55 mmHg. After stabilization, a recruitment maneuver was performed (50 cmH2O, 40 s). Thereafter, PEEP was set at 24 cmH2O and decreased to zero in steps of 4 cmH2O (10 min/step). At each step, arterial and mixed venous blood samples were obtained and cardiac output (CO) was measured. In six animals (CT-Group), whole lung helical CT-scans (1 mm thickness) were obtained at each step during end-expiratory and end-inspiratory pauses. Airway pressure and flow signals were recorded and elastance of the respiratory system (Ers), resistance (Rrs) and PEEP continuously computed by the least squares method. The best PEEP was then assessed considering four definitions: 1) PEEP of minimal Ers; 2) PEEP of minimal Ers and highest PaO2; 3) last PEEP with PaO2+PaCO2 ≥ 400 mmHg; 4) last PEEP before the PEEP of minimal Ers. Results are presented as median and inter-quartiles. Comparisons were performed with the Wilcoxon's test and adjusted with the Bonferroni-Holm procedure. Significance was accepted at p < 0.05.

Results

The stepwise reduction of PEEP resulted in an increase of cardiac output and arterial blood pressure, as well as a reduction of total lung volume (TLV) and volume of intrapulmonary gas. The intrapulmonary shunt increased and PaO2 decreased significantly at the PEEP of 8 cmH2O (p < 0.05). No significant differences were observed among the PEEP of minimal Ers (12.5 cmH2O [12.4–12.6]), PEEP of minimal Ers+highest PaO2 (12.7 cmH2O [12.4–15.7]) and last PEEP with PaO2+PaCO2≥400 mmHg (12.5 cmH2O [9.7–15.3]). The PEEP level 4 cmH2O higher than the PEEP of minimal Ers (16.5 cmH2O [12.4–16.9]) differed significantly from the PEEP levels according the other criteria, and resulted in a compromise among hyperaerated, normally aerated poorly aerated and non aerated lung areas.

Conclusion

In this model of ALI, a PEEP level set 4 cmH2O higher than the PEEP of minimal Ers in a decremental PEEP trial resulted in a compromise between oxygenation, lung mechanics and lung aeration. This criterion could be useful to achieve improved pulmonary function and aeration simultaneously.

0409

Effects of Prone and Supine Position on Regional Pulmonary Perfusion Measured by Small Animal Positron Emission Tomography

T. Richter*1, R. Bergmann2, M. Ragaller1, I. Közle2, J. Pietzsch2

1Anesthesia and Critical Care, Technische Universität Dresden, 2Institute of Radiopharmacy, Forschungszentrum Dresden-Rossendorf, Dresden, Germany

Introduction

There are only limited data about the influence of posture on the pulmonary perfusion (Qr) in small animals. Ga-68 radiolabeled human albumin microspheres (Ga-68-DOTA-HSAM) with a diameter of 20 µm are expected to lodge in the pulmonary capillaries. The purpose of this study was to quantify Qr in prone and supine position in terms of their imaging manifestations in healthy rats.

Methods

The animal research committee of the Regierungspräsidium Dresden approved the animal facilities and the experiments according to institutional guidelines and the German animal welfare regulations. Seven anesthetized, spontaneous breathing Wistar rats (297 ± 53 g), were positioned either in prone (n=3) or supine (n=4) position and Ga-68-DOTA-HSAM were infused. Qr was achieved by imaging the radioactivity distribution in the lungs. The animal PET 3D volume data were reconstructed with 3D OSEM MAP algorithm. After first measurement animals were arranged in opposite position and the PET measurement was repeated. The 3D data were at first manually coregistered and residual differences in the relative positioning of the lungs were corrected user independently (Rover, ABX GmbH, Germany). Masks for defining regions of interest (ROI) were set in the coregistered volume data for all animals. Mean normalized Qr values (Qmean) of the dorsal and ventral parts of the lungs were calculated on the basis of an automatic ROI-setting including threshold analysis.

Results
Vertical gradient of regional perfusion was significantly steeper in the supine, −0.131±0.01 %/cm, than in the prone animals −0.055±0.01 %/cm (P = 0.002), indicating that the vertical distribution of regional perfusion in dependent regions was more accentuated in the supine than in the prone infused animals. Changes in the vertical gradient after rotation in the opposite position resulted in a vertical gradient of −0.093 ± 0.031 %/cm in prone position. Vertical gradient in the supine position of prone infused animals was −0.093 ± 0.015 %/cm. Position changes did not produce significant changes in vertical gradient in supine (P = 0.125) and in prone infused animals (P = 0.25). Qmean was not affected by posture.

position of Ga-68-infusion

position of acquisition

Qmean ± SD prone

Qmean ± SD supine

ROI dorsal

prone

0.18 ±0.04

0.16 ± 0.03

ROI dorsal

supine

0.19 ± 0.05

0.16 ± 0.02

ROI ventral

prone

0.16 ± 0.04

0.14 ± 0.03

ROI ventral

supine

0.17 ± 0.04

0.14 ± 0.02

Conclusion

The influence of posture on regional perfusion was demonstrated, for our knowledge, for the first time in rodents. Qr was significantly decreased from dorsal to ventral regions in supine infused animals. Ga-68-DOTA-HSAM (20 µm) can be used for non-invasive pulmonary perfusion studies in rats with small-animal PET. Lung structure distribution changes were minor. Distribution pattern of regional perfusion in prone and supine position in normal rat lungs will be used as reference data set for further studies on injured rat lungs.

0410

Peep Optimizing Using Electric Impedance Tomography in Porcine Lung Injury

T. Muders*1, H. Wrigge1, J. Zinserling1, U. Günther1, H. Lüpschen2, C. Putensen1

1Anesthesiology and Critical Care Medicine, University Hospital, Bonn, 2Medical Information Technology, RWTH University, Aachen, Germany

Introduction

Potential for alveolar recruitment is individually different in patients with Acute Lung Injury (ALI) and can not be predicted by origin of ALI. Delay of regional ventilation (RVD) measured by Electric Impedance Tomography (EIT) during slow inflation has recently been shown to correlate with alveolar recruitment (1). We hypothesized that tidal recruitment can be reduced by individual PEEP optimizing using EIT.

Methods

ALI was induced (oleic acid + abdominal hypertension) in 8 anesthetized pigs that received PEEP of 0–25cmH2O in steps of 5 in randomized order. PEEP level derived from ARDSnet protocol table was compared to lowest PEEP resulting in minimal differences of local RVD measured by EIT (see Figure).

Results

PEEP from ARDSnet protocol (10(8–10), Median(Range)) was lower than PEEP that minimized tidal recruitment estimated by EIT (20(20–25), p < 0.05). PaO2/FiO2 increased with each step of PEEP up to 25 cmH2O and was lower with ARDSnet PEEP (179(139) vs 362(118), mean(SD), p<0.01). PaCO2 was not different (38(6) vs 37(5)). Open image in new window

Conclusion

In this model of ALI EIT was helpful to find lowest PEEP level that minimizes tidal recruitment. ARDSnet PEEP was lower and did not entirely avoid tidal recruitment.

Oral Presentations Ventilator-associated pneumonia: Can we prevent?: 0411–0414

0411

Prophylaxis of Ventilator Associated Pneumonia by Continuous Lateral Rotation Therapy: A Prospective Randomized Trial

T. Staudinger*1, A. Bojic1, U. Holzinger2, F. Mallner1, M. Rohwer1, G. Heinz3, P. Schellongowski1, K. Laczika1, G. J. Locker1

1Internal Medicine I, 2Internal Medicine III, 3Internal Medicine II, Medical University of Vienna, Vienna, Austria

Introduction

Continuous lateral rotation therapy (CLRT) has been shown to be effective for prophylaxis of ventilator associated pneumonia (VAP). However, neither hospital stay nor days on ventilator or mortality could be positively influenced by CLRT. Most studies, however, lack to report on standards of VAP prophylaxis as well as on duration and practicability of CLRT. We performed a prospective randomized trial of the efficacy of CLRT in mechanically ventilated patients with respect to VAP prophylaxis.

Methods

Patients were eligible if not suffering from pneumonia or acute lung injury but in the need for mechanical ventilation. Patients were randomized within 48 hours after intubation to undergo CLRT in a specially designed bed allowing continuous rotation of the upper part of the body through an arc of 90° or to be treated in supine position. Rotation had to be performed for at least 18 hours per day. All patients were in a semirecumbent position of at least 30°. Suctioning and change of ventilator circuits were standardized. Sedation was allowed only if clinically indicated. All patients received enteral nutrition if possible. Endpoints of the study were spontaneous breathing, death, lack of tolerance, or development of pneumonia defined as newly acquired infiltrate in two consecutive chest x-rays plus signs of inflammation plus purulent bronchial secretions.

Results

Of 150 consecutive patients enrolled, 75 underwent CLRT (mean age 59 ± 16, 50 male) while 75 (age 60±15, 48 male) were randomized into the control group. Patients suffered mostly from resuscitation after cardiac arrest, cardiogenic shock, neurological diseases, or COPD. Groups were comparable with respect to diagnoses and severity of disease. In the CLRT group, 7 patients developed VAP (9%) versus 20 patients (27%) in the control group (p <0.01). ICU and hospital survival were 70% and 63% in the CLRT group and 74% and 63% in the control group, respectively (p = n.s.). Comparing CLRT to control group, days on ventilator were 8 ± 5 and 14 ± 23 days (p=0.02), length of hospital stay was 24 ± 22 and 39 ± 45 days (p=0.02), and ventilator free days during the first 28 days were 15 ± 10 versus 11 ± 10 (p=0.04), respectively. Development of pressure sores, duration of sedation, days on enteral feeding as well as use antibiotics was comparable. Intolerance of CLRT during weaning was observed in 40%.

Conclusion

CLRT is an effective prophylaxis for VAP. It exerts additive effects to other prophylactic measures like semirecumbent position and standardized ventilation and suctioning protocols. Lower VAP rate seems to result in shorter duration of ventilation and length of hospital stay. CLRT is not well tolerated in non-sedated patients during the weaning process.

0412

The Influence of Tracheostomy Tube with Intermittent Subglottic Drainage Compared to Conventional Tracheostomy Tube on the Incidence of Ventilator-Associated Pneumonia

E. Nagy*, F. Veroniki, C. Toronidis, C. Tsokantaridis

Intensive Care Unit, Kavala General Hospital, Kavala, Greece

Introduction

Subglottic secretion drainage (SSD) via an endotracheal or a tracheostomy tube has recently been introduced in clinical practice for prevention of ventilator associated pneumonia (VAP), since it is cosidered to reduce micro-aspiration of secretions from the area around the cuff. The purpose of this study was to investigate the influence of a tracheostomy tube with SSD versus conventional tracheostomy tube on the incidence of VAP in ICU patients.

Methods

A randomized comparative clinical study, which included 126 patients, admitted for both medical and surgical reasons, with a length of stay >7 days in the ICU. Group A included 64 patients ventilated via conventional tracheostomy tube (TT-C) and Group B included 62 patients ventilated via a tracheostomy tube with subglottic secretion drainage (TT-SSD). SSD was performed hourly, applying a suction of -20cmH2O. Clinical pulmonary infection score plus quantitative culture of protected specimen brushing were used for the diagnosis of ventilator-associated pneumonia. The following data were recorded: age, APACHE II, predicted mortality rate (PDR), day on which tracheostomy was performed (T-Day), days of mechanical ventilation (MV-Days), length of stay (LOS), incidence of pneumonia, and outcome (death or discharge). Demographic data were compared by the nonparametric Mann-Whitney test. The incidence of pneumonia between the groups was compared by the x2 test and outcome was compared by the Fisher test. P<0,05 was considered statistically significant.

Results
The results of this comparative study are shown on the table below.
 

Group A (TT-C)

Group B (TT-SSD)

p

Age

50.6±22.7

56.3±19.6

0.268 (NS)

APACHE II

18.5±5.7

18.2±6.2

0.880 (NS)

PDR

33.13±17.8

32.1±18.0

0.880 (NS)

T-Day

4.3±3.4

3.8±2.3

0.830 (NS)

MV-Days

14.6±10.8

20.1±16.9

0.051(NS)

LOS

19.5±11.4

28.6±25.5

0.117 (NS)

VAP

28 (43.75%)

28 (45.16%)

0,170 (NS)

Death

16 (25%)

20 (32.25%)

0.585 (NS)

Conclusion

According to the results of this study, the use of a tracheostomy tube with intermittent subglottic drainage does not reduce the incidence of ventilator-associated pneumonia, compared to conventional tracheostomy tube, in patients staying in the ICU longer than 7 days.

0413

Oral Decontamination with Clorhexidine 2% as a Single Intervention to Decrease Ventilator Associated Pneumonia in a Medical-Surgical ICU

M. Granados*1, A. Cossio1, J. Mora1, M. Badiel2, M. Vargas1, M. Gomez1, J. Martinez1, C. Salas1, F. Rosso3

1Critical Care Unit, 2Clinical Research Institute, 3Infection Surveillance Committee, Fundacion Valle del Lili, Cali, Colombia

Introduction

Ventilador Asociated Pneumonia (VAP) is a significant cause of in-hospital morbidity and mortality, especially in the developing countries. Select oral care interventions, such as topical antiseptic agents, could reduce the occurrence of VAP. Some studies have suggested the benefit of the use of Chlorhexidine (CHX) for oral decontamination, nevertheless this benefit has not been completely proven. We hypothesized that oral decontamination with CHX, 2% would reduce the incidence of VAP in our ICU.

Methods

Design: Observational study. Bidirectional cohort.Patients admitted to a mixed medical-surgical ICU of a University affiliated Hospital who needed mechanical ventilation for 48 h or more were eligible for recruitment. Since July 1st, 2007, the use of use CHX 2% solution three times per day as oral decontamination was implemented. VAP adjusted Incidence rates of were compared with a cohort from October 2006 thru 2007(before CHX) and after the implementation from 2007 –2008. A bivariate and multivariate analysis were performed. P value<0.05 was considered significant statistical.

Results
Eight hundred thirty six patients needed mechanical ventilation for more than 48 hours during the study period (2006–2008). The overall VAP incidence rate was 11.2/1000 days-ventilator. The table shows demographic characteristics of patients. VAP incidence rate was decreased after protocol implementation (13.1 vs. 9.1 VAP/1000-days ventilator, p=0.05, unadjusted OR: 1.4 (IC95% 0.92–2.2). Adjusted OR by APACHE II difference was maintained (OR=1.3, IC95% 0.89–2.01).
Table 1

Demographic Characteristics of Patients

Variable

October 1st 2006 to

July 1st -2007 to Ma

p

Age, mean ± SD

54.5±18.7

53.7±19.6

0.57

Males, %

55.6

57.2

0.3

Apache II,, mean± SD

21.5±5.9

23.1±8.1

0.002

Medical Admision, %

65.1

63.0

0.08

VAP/1000-days-ventila

13.1

9.3

0.05

Days ventilator, mean

6.6(3.75–11)

6.4 (4–10.8)

0.9

ICU Length of stay, d

8

9

0.41

Overall Mortality, %

26.5

22

0.1

Source: Fundacion Valle del Lili ICU Database

Conclusion

The simple implementation of CHX 2% in mechanical ventilated patients was effective in reducing incidence rate VAP. This simple intervention could be beneficial in the devolping countries were VAP incidence is higher.

Grant Acknowledgement

Study supported by Hospital Infection Surveillance Committee and Clinical Research Institute of Fundación Valle del Lili.

0414

Lateral Position to Prevent Gastric Aspiration in Intubated Patients in a Surgical Intensive Care Unit

T. Mauri*1, K. Kumwilaisak2, L. Berra2, S. Pivi2, C. Crimi2, J. W. Ufberg3, F. Kueppers3, L. M. Bigatello2

1Perioperative Medicine and Intensive Care, San Gerardo Hospital, University of Milano-Bicocca, Monza, Italy, 2Anesthesia and Critical Care, Massachusetts General Hospital, Harvard Medical School, Boston, 3Emergency Medicine, Temple University School of Medicine, Philadelphia, United States

Introduction

Elevation of the head of the bed of intubated individuals to ≥ 30 degrees (semi-recumbent position) is recommended as a way to prevent aspiration of gastric content in the airways that may lead to nosocomial pneumonia. However, recent data suggest that horizontal orientation of the endotracheal tube (ETT) decreases bacterial colonization of the lower respiratory tract, in comparison to semi-recumbent position. We designed a study to test the hypothesis that lateral, head down position, to achieve horizontal position of ETT, decreases aspiration of gastric contents in comparison to semi-recumbent position.

Methods

We enrolled twenty adult patients intubated from less than 48-hour and without signs of pulmonary infection. The first 10 consecutive patients were ventilated in the semi-recumbent position (SR group) and the following 10 patients in the lateral, head down position (LHD group). Patients were studied for 64 hours in the SR group and for 12–24 hours in the LHD group. Tracheal secretions were collected every 8 hours (SR group) or every 4 hours (LHD group) and tested for pepsin presence. Oral and gastric suctions were collected and tested for pepsin at baseline and every 12 hours. Data were analyzed by Chi-square, Fisher's exact test, or T-test, as appropriate.

Results
The two groups were similar at baseline (Table 1). Ten patients from the SR group (100% of samples) and 8 (80% of samples) had pepsin-positive gastric samples. Four patients in each group had pepsin in oral secretions (23% of oral samples for both groups). Pepsin was detected in tracheal secretions of 7 patients (33% of tracheal samples) from the SR group and 5 patients (38% of tracheal samples) from the LHD group (p = 0.35).
Table 1

Baseline Characteristics of the 2 Study Groups

 

SR group N = 10

LHD group N = 10

P-value

Age, yrs.

63 ± 18

64 ± 14

0.9

Female, %

50

60

1

BMI, kg/m2

27.4 ± 9

28.5 ± 9.9

0.83

APACHE II

16.8 ± 3.4

14.7 ± 2

0.14

CPIS

4 ± 2

4 ± 2

0.78

RASS

(−)1 ± 2

(−)2 ± 1

0.06

Tube feeding, %

40

20

0.62

Opioids use, %

70

70

1

Conclusion

Intubated patients ventilated in the semi-recumbent position manifest a high incidence of a marker of aspiration of gastric contents. Lateral, head down position doesn't decrease significantly this phenomenon.

Grant Acknowledgement

Jenney Fund, Dept. of Anesthesia, MGH, Boston, USA.

Poster Sessions Sedation, analgesia, pharmacology: 0415–0428

0415

A Multimodal Short Sedation Regimen Avoiding Secondary Induced Hyperalgesia in the Intensive Care Unit : A New Standardized Approach for Critically Ill Patients

F. Meurant, L. Vandebrouck*

Intensive Care, Kirchberg Hospital, Luxemburg, Luxembourg,

Introduction

Using a combination of a N-Methyl-D-Aspartate antagonist (NMDAa) and remifentanil we attempted to use their synergistic effects in order to avoid hypnotic drug accumulation. We assessed the benefits of this Multimodal Short Sedation Regimen by comparing it to more conventional sedative protocols.

Methods

Medical (n=396), surgical (n=314) and trauma patients (n=88) who needed more than 48 hours of sedation were prospectively allocated into one of 4 groups :Cat.A (patients in shock)- B1(respiratory failure)- B2 (hemodynamic failure)- C (stable patients)) as defined by different hemodynamic (inotropic support) and respiratory (PaO2/FiO2 ratio) parameters. All patients were sedated using a narcotic (remifentanil(0.04–0.45mcg/kg/min)) and a hypnotic agent (propofol(0–7mg/kg/h)+/− midazolam(0–4.5mcg/kg/min)). Titration of the sedative agents was done hourly in order to achieve a Sedation Agitation Score of 4 (yawning; frown). Patients were randomized into two groups (G:n=399). In G1 we added an NMDAa namely Magnesium(0.08g/kg/d)+/− Ketamine(0–5 mcg/kg/min)+/− Clonidine 0–0.02mcg/kg/min) to the baseline sedative regimen and in G2 we added a placebo. The different groups were compared as to Midazolam and Propofol requirements (Mi.; Po.), awakening/weaning times(A/Wt: patient calm, comfortable, cooperative / spontaneously breathing; spontaneously moving)) as well as the global hospitalisation stay (GHs). For statistical analysis a Shapiro-Wilk test, Wilcox and a Student T-test were used.

Results

6160 patient/days were studied during a two year period. Demographic data were comparable in terms of mean Severity Acute Physiologic Score, age and gender. In G1 patients were ventilated invasively during 10+/−1 days and in G2 during 15+/−2days. The average A/Wtime necessary to shift through the Cat.A to Cat.B was 10+/−5 min in G1 and 28+/−6 hours in G2 (p <0.05); to Cat.B through Cat.C it takes 6+/−2 min(B1) or 5+/−3 min (B2) in G1 compare to 22+/−7 hours in G2(p < 0.05). The Mi. (mcg/kg/min) in Cat.A was 2–8.5 in G2 and 0–0.9 in G1 (p < 0.02); in Cat.B2 it was 1.5–4.5 in G2 and 0–0.8 in G1 (p < 0.05). The Pr. (mg/kg/h) in Cat B1 was 0.5–7 in G2 and 0–3.5 in G1; in Cat.C it was 0–10 in G2 and 0–3 in G1. The mean GHs was 2.6 days longer for patients in G2 compared to G1.

Conclusion

The described Short Sedation Regimen is usable for medical, surgical and trauma patients. The titration synergy concept used by combining NMDAa, remifentanil and hypnotic agents allows a dose reduction for each individual drug, thus avoiding the risk of hypnotic accumulation. As a result it decreases Intensive Care stay and subsequently has longterm financial implications.

0416

Risk Factors Leading to Midazolam Therapeutic Failure During Continuous Sedation in Critically Ill Ventilated Patients

Chamorro*, M. Romera, B. Balandín, M. Valdivia, S. Alcantara

Intensive Care, Hospital Puerta Hierro, Madrid, Spain

Introduction

Nearly 17% of critically ill ventilated patients treated with continuous intravenous midazolam experience therapeutic failure (taquifilaxia or tolerance). These phenomena greatly hinder patient management. The aim of this study is to analyse and describe the factors that could influence its onset

Methods

Prospective and descriptive study carried out in an intensive care unit (ICU) of a University hospital. We included 1015 ventilated patients that received midazolam to achieve the proposed sedation aims: 3 to 5 level on the Ramsay scale and ventilator synchrony. We defined therapeutic failure as the need to administer more than 0.23 mg/Kg/h to reach these aims. All patients were treated with intravenous opiates. We analyzed demographical, clinical, and outcome data. Univariate and multivariate analyses were performed to investigate significant risk factors and logistic regression was used to fit the model. In each variable we calculated the odds ratio (OR) with its corresponding confidence interval (CI), significance assumed for P<0.05.

Results

176 patients had therapeutic failure. On univariate and multivariate analysis the associated variables were: Male gender OR 1.4 (1.02–1.86, CI 95%), p=0.03; Age under 45 years, OR 2.1 (1.6–2.7, CI 95%), p < 0.001, compared with ages between 45 to 65 years; and OR 5.1 (3.4–5.7, CI 95%), p < 0.001, compared with ages over 65. Ages between 45 to 65 have an OR 2.4 (1.6–3.8, CI 95%), p < 0.001, compared with ages over 65; Ventilation due to respiratory insufficiency, OR 2.8 (2.1–3.9, CI 95%), p < 0.001; Need to administer midazolam during 5 or more days, OR 7.4 (5.1–10.8 CI 95%,), p < 0.000.

This logistic regression formula P=ez / 1+ ez, could estimate tolerance risk in up to 77% (70.8–80, IC 95%) of the cases.

Where z=constant +0.47(male gender) + 1.03 (age 45–65) + 1.06 (ventilation due to respiratory insufficiency) + 2.41 (midazolam administration ≥ 5 days; or z= constant +0.47(male gender) + 2.34 (age under 45) + 1.06 (ventilation due to respiratory insufficiency) + 2.41 (midazolam administration ≥ 5 days).

Conclusion

Male gender, being younger, being mechanically ventilated for respiratory insufficiency and receiving midazolam during 5 or more days are the main risk factors leading to midazolam therapeutic failure.

0417

Monitoring and Analgesic, Sedative and Neuromuscular Blockade Strategies Used in Spain. An In-Person Survey

C. Chamorro*1, J. Domínguez-Roldan2, F. Barturen3, J. Borrallo4

1Intensive Care, Hospital Puerta Hierro, Madrid, 2, Hospital Virgen del Rocío, Sevilla, 3, Policlinica Miramar, 4, Hospital Guadalajara, Palma de Mallorca, Spain

Introduction

There is great variability in the sedo-analgesic strategies used in critically ill ventilated patients. The aim of this study is to investigate the current sedation, analgesia and neuromuscular blockade practises in adult ICUs in Spain.

Methods

Descriptive study based on an in-person survey of sixty physicians, representative of 60 ICUs. All were critical care specialists or anaesthesia specialists with direct responsibility in critical care management. A questionnaire was developed for this study and consisted of 19 questions on different aspects of usual practises. Each specialist surveyed, anonymously, chose one of five options for each answer. Each answer represents the usual practises in their ICUs, not their personal thought.

Results

Midazolam (MID) is the most common sedative employed for tracheal intubation (TI) (42% of ICUs), followed by propofol (PRO) (36%).Thirty per cent of ICUs do not use NMBs for TI, 27% employ succinylcholine, 24% cisatracurium and 16% rocuronium. For short-term sedation, patients in shock, a midazolam-opioid combination is preferred (50%), followed by remifentanyl combined or not with low doses of other sedatives. For long-term (i.e. ARDS) a midazolam-opioid combination is employed by 71%; 13% employ a midazolam-propofol-opioid combination. 40% of the ICUs do not have a protocolised midazolam or propofol maximum dose limit. For sedation monitoring 51% of the ICUs use the Ramsay scale, 5% the SAS and 3% the RASS; 41% never use sedation scales. 21% use train-of-four monitoring to titrate continuous neuromuscular blockers. No ICU uses pain scales in non-communicative patients nor scales to detect delirium. 51% do not use strategies to avoid midazolam accumulation, 11% employ sequential sedation (replace midazolam with propofol or remifentanil), 11% employ a daily sedation break, 8% bispectral index monitoring and 19% use two or more of the previous strategies. Bispectral monitoring is used by 49% of ICUs, 6% to adjust deep sedation, 6% to titrate barbiturate treatment, 3% to monitor sedation during neuromuscular blockade and 34% in two or more of these situations. Morphine, fentanyl and remifentanil are the most used analgesics in critically ill patients, but ketorolac is still employed in patients with hemorrhagic risk and meperidina in patients with severe acute pancreatitis.

Conclusion

There is considerable variation in clinical practises. Some of the employed strategies could produce morbidity. It is recommended that a standardised approach to analgesic, sedation and neuromuscular blockade use and assessment be developed and tested.

0418

Risk Factors Leading to Propofol Therapeutic Failure During Continuous Sedation in Critically Ill Ventilated Patients

C. Chamorro*, M. Romera, B. Balandín, M. Valdivia, M. Pérez

Intensive Care, Hospital Puerta Hierro, Madrid, Spain

Introduction

There is little information about propofol therapeutic failure incidence during its employment for critically ill ventilated patients. These phenomena greatly hinder patient management and may lead to employing excessive and toxic doses. The aim of this study is to analyse and describe the factors that could influence its onset.

Methods

Prospective and descriptive study carried out in an intensive care unit (ICU) of a University hospital. We included 1460 ventilated patients that received propofol (1% or 2%) to achieve proposed sedation aims: 3 to 5 level on the Ramsay scale and ventilator synchrony. We defined therapeutic failure as the need to administer more than 350 mg/h to reach these aims. Propofol was administered, preferably, in patients with haemodynamic stability and foreseen sedative requirements lower than 3 days, and in patients with need of frequent neurological evaluations independently of foreseen sedative time requirements. Most of the patients were also treated with intravenous opiates. We analyzed demographical, clinical, and outcome data. Univariate and multivariate analyses were performed to investigate significant risk factors and logistic regression was used to fit the model. In each variable we calculated the odds ratio (OR) with its corresponding confidence interval (CI), significance assumed for P<0.05.

Results
54 patients had therapeutic failure, most of them during the first 48 hours of propofol administration. On univariate analysis the associated variables were: Male gender OR 1.4 (1.02–1.86 CI 95%), p=0.03; Age under 45, OR 5.2 (2.9–9.0, CI 95%), p < 0.001; Polytraumatism as admission reason, OR 2.9 (1.7–4.9, IC 95%), p < 0.01. On multivariate analysis polytraumatism as admission reason lots its statistically significance because a subordinated relationship with age under 45.
Table 1

Equation Variables

 

B

ET

Sig

EXp (B)

C.I. 95,0% Lower/Upper

polytraumatism

0.342

0.317

0.282

1.407

0.755/2.622

sex

1.160

0.397

0.003

3.190

1.466/6.941

age

1.597

0.329

0.000

4.940

2.590/9.421

constant

−5.034

0.418

0.000

0.07

 

a:Variable(s) inserted in step 1: polytraumatism, sex, age45

Conclusion

In our series being younger (age under 45) is the main factor leading to propofol therapeutic failure. Male have more risk to present it than women.

0419

Sedation and Analgesia Practice in Mechanically Ventilated Patients Admited to 9 ICUS of University Hospitals in Spain

A. Sandiumenge*1, C. Chamorro2, H. Torrado3, M. Jiménez4, T. Muñoz5, C. Pardo6, M. Alonso7, J. Alonso8

1ICM, University Hospital Joan XXIII, Tarragona, 2ICM, University Hospital Puerta de Hierro, Madrid, 3ICM, University Hospital Bellvitge, Barcelona, 4ICM, University Hospital Clinico de San Carlos, Madrid, 5ICM, Hospital Txagorritxu, Vitoria, 6ICM, University Hospital Fuenlabrada, 7ICM, University Hospital 12 de Octubre, Madrid, 8ICM, University Hospital Vall d'Hebron, Barcelona, Spain

Introduction

To evaluate the analgesia and sedation practice and their complications in mechanically ventilated (MV) patients in 9 ICUs of different university hospitals in Spain.

Methods

Demographic and sedo-analgesia (SA) procedural data was retrospectively recorded in all mechanically ventilated patients admitted during a 30 day period in 9 Spanish ICUs (1 traumatic, 1 surgical, 1 medical and 6 medical-surgical). Time under sedation and analgesia, agents used and administration patterns were recorded. Sedative failure (defined as an insufficient sedation level after 0.25 mg/kg/h for midazolam (MZ) or 4.5 mg/kg/h for propofol (PF)), and deprivation (defined as agitation during sedation withdrawal excluding organic causes) were also recorded in all patients under continuous sedation. Statistical analysis was performed using SPSS 13.0, significance level p < 0.05.

Results

A total of 244 patients were mechanically ventilated for a mean time of 114.7 ± 190.7 hours (range 1–1556) (35.7% MV >72h; n=87). Mean age was 58.9 ± 17.2 y/o, 73% were male and most patients were post-surgical (53.3%) followed by medical (33.6%) and traumatic (7.8%) with a mean APACHEII of 15.2 ± 7.9. Previous history of alcohol, illegal or prescribed psychotropic drugs intake was documented in 39 (16%), 17 (7%) and 42 (17.2%) patients. ICU mortality was 18.9 (n=46) and mean ICU length of stay (LOS) was 195.5±245.1 hours (range 1–1604).184 patients (75.5%) under MV received continuous infusion of sedatives(S) (mean time 92.7±143.8 hours) and/or analgesics (A) (mean time 136.4±195.1 hours) (75.5% SA; 23% only S; 5.9% only A). PF (mean time 60.7±97.2 hours) was the preferred sedative (64.0%). MZ (42.4%; mean 144.4±218.8 hours) and Morphine chloride (53.3%; mean 144.9±157.4 hours) were the preferred agents for long term sedation (p < 0,05) Remifentanil (29.3%) and Fentanyl (22.0%) were used during 62.4±72.1, and 149.7±279.5 hours respectively. The use of other sedatives like Clonidine (1.6%) and barbiturates (1.1%) was low. Fifty four (29.3%) patients received more than 2 sedatives simultaneously, mostly due to sequential sedation (substitution of long half-life sedatives for short half-life ones) (17.4%), sedative failure (7.6%) or deprivation (7.6%). History of alcohol abuse and illegal or prescribed psychotropic drugs intake did not influence LOS, MV or sedation duration. Alcohol intake was associated with sedation failure and deprivation (p=0.03).

Conclusion

One out of four patients with MV receives sedation without analgesia. Patients with previous history of alcohol abuse are at risk of presenting sedation failure. There is a wide variability in SA utilization patterns among the different centres.

0420

Cost-Consequence Simulation of Different Sedation Regimes in Germany

M. J. Al1, J. Martin2, J. Bakker3, R. Welte*4

1Institute for Medical Technology Assessment, Erasmus MC, Rotterdam, Netherlands, 2Klinik für Anästhesiologie, op. Intensivmedizin und Schmerztherapie, Klinik am Eichert, Göppingen, Germany, 3Department of Intensive Care, Erasmus MC, Rotterdam, Netherlands, 4Reimbursement and Health Outcomes, GlaxoSmithKline, Munich, Germany

Introduction

Economic considerations become increasingly important in the ICU due to scarce resources, medical progress and new reimbursement systems. Health economic models offer a promising tool for optimising resource allocation.

Methods

We performed a cost-consequence analysis of remifentanil-based sedation (RS) vs conventional sedation (CS) of critically ill patients in Germany using a probabilistic Markov model. The patient flow, the length of stay, the duration of mechanical ventilation (MV) and the direct medical costs of RS vs CS in the ICU were simulated. Input data for the model were obtained from UltiSAFE, a Dutch open-label trial with 205 critically ill patients with the exception of unit costs. In UltiSAFE patients received CS according to Dutch guidelines (predominantly fentanyl or morphine combined with midazolam or propofol) or RS (remifentanil, combined with propofol if required) for up to 10 days. Unit costs were measured in a separate microcosting study in a German mixed 12-bed adult ICU using the hospital perspective and prices of 2006. Material, staff and overhead costs were considered. Based on UltiSAFE we performed two analyses distinguished by the considered patient population: (1) Inclusion of all patients (2) Subgroup analysis according to the UltiSAFE target population, thereby including only patients who started weaning within 72 hours of start of treatment.

Results

Compared to CS, RS lowered the mean duration of MV from 5.9 to 4.9 days in the total population and from 3.2 to 2.2 days in the subgroup. Similarly, it reduced the length of ICU stay from 8.4 to 7.4 days or 5.7 to 4.9 days in the total population or in the subgroup, respectively. As a result, RS decreased the ICU costs per patient by € 1128 in the total and € 1003 in the subgroup population. The probability of RS being cost-saving was estimated at 85% for all patients and 91% for the subgroup.

Conclusion

From an economic perspective, RS seems to be the preferred regimen compared to CS: It leads to a substantially shorter length of MV and length of ICU stay. The associated cost-savings more than compensated the additional drug costs of RS resulting in net-savings to the hospital.

Grant Acknowledgement

This study was funded by GlaxoSmithKline.

0421

Morphine or Fentanyl for Mechanically Ventilated Patients with Hemodynamic Instability?

R. P. Oliveira1, A. C. Costa1, A. F. Meregalli1, D. M. Dallegrave1, E. C. Zignani1, F. D. Neto1, J. A. Höher1, G. Friedman*2

1ICU, Complexo Hospitalar Santa Casa, 2ICU, Universidade Federal Ciências da Saúde de Porto Alegre, Porto Alegre, Brazil

Introduction

One major goal of analgesia and sedation for critically ill patients is to provide control of pain and anxiolysis to facilitate mechanical ventilation. Fentanyl or morphine is commonly used to provide analgesia. However, both medications can promote adverse consequences, including hypotension, prolonged mechanical ventilation, increased ICU duration of stay and high costs. Comparative trials of both opioids have not been performed in critically ill patients and the selection of an agent depends on its pharmacology, adverse effects and costs. Many ICU physicians preferred fentanyl to morphine in hemodynamically unstable patients due to the risk of additional hypotension. However, both drugs accumulate and may cause similar adverse effects when continuously infused. The aim of this study was to compare the hemodynamic effects of morphine and fentanyl in hemodynamically unstable patients under mechanical ventilation.

Methods

Prospective, randomized, open-label trial including mechanically ventilated patients >16 years and hemodynamically unstable (cathecolamine requirement >1 hour). Exclusion criteria included patients after cardiac arrest or with Do Not Resuscitate (DNR) order. All patients received continuous I.V. midazolam in association with either morphine or fentanyl. The endpoint level of sedation was a Ramsay scale 2–4 and of analgesia was a Behavioral Pain Scale (BPS)<5. Mean arterial pressure (MAP), heart rate (HR), rate of opioid and of benzodiazepines, rate of catecholamine (noradrenaline), BPS score and adverse effects (bowel distention and vomit) were measured every 6 hours. The protocol was finalized with hemodynamic stabilization or sedative infusion interruption. Variables were expressed in frequencies and means. Student's t test was used to compare means and p<0.05 was significant.

Results

A total of 29 patients were enrolled in the study; 16 were randomly assigned to the morphine group and 13 to the fentanyl group. Although not significant, the fentanyl group was older (51±20 yrs vs. 63±12 yrs; p=0.069) but less sicker (APACHE II score, 27 vs.16; p=0.054) then the morphine group. There were no differences for heart rate (103 ±23 vs.101±23) and MAP (82.5±16.1 vs. 83.5±17.5). However, noradrenaline rate was higher in the morphine group (0.165±0.15mg/kg/min vs. 0.11±0.10mg/kg/min; p=0.009). The mean BPS was similar in the two groups (3.5±0.8 vs .3.6±0.9). Bowel distention occurred in 6 patients in the morphine group and in 3 patients in the fentanyl group (p=0.278). Five patients in the morphine group were discharged from ICU as compared with 3 patients in the fentanyl group (p=0.678).

Conclusion

These preliminary results indicate that both drugs were associated with similar adverse effects. Noradrenaline rate was higher with morphine then with fentanyl. However, patients with morphine were sicker.

0422

The Potential Economic Impact of Alterations to Sedation Protocols in a District General Hospital Intensive Care Unit

D. Dutta*, A. Krishnamurthy, P. Petson

Dept of Anaesthesia and Critical Care, Princess Alexandra Hospital NHS Trust, Harlow, United Kingdom

Introduction

The objective of this research was to assess the potential for improved resource utilization in a District General Hospital (DGH) Intensive Care Unit (ICU) through alterations to existing sedation protocols. We also examined whether other factors needed to be considered to ensure such changes brought long-term benefits. Between 2003–2006, the ICU at the Princess Alexandra Hospital (PAH) operated 4–5 Level 3 beds with a mean occupancy of 103%, admitting a mean of 326 patients with an mean LOS of 6.27 days, requiring a mean of 35 non-clinical transfers (NCT).There was a need to maximize capacity by investigating methods of reducing LOS and the need for NCT via efficient use of available resources. One likely method was through modifications to the ICU sedation protocol.

Methods

The original PAH sedation protocols utilized morphine and midazolam for long stay ICU patients, with propofol and alfentanil used for short stay patients. In 2006, protocol modifications introduced the use of remifentanil and propofol in Specifically selected patients. Published evidence suggests a reduction in LOS of a day is possible under similar protocols(1,2). A mobile simulation course on the use of remifentanil was arranged. A retrospective audit was carried out, using the controlled drug register, the ICU admission register and ICU database. Data was collected on the total number of patients admitted to the ICU, those who received remifentanil, the total duration of use of remifentanil and the total LOS of those patients receiving remifentanil, and compared to the average LOS in the PAH's ICNARC data(3).

Results

Our data shows 29%(n=110) of our patients received remifentanil with average duration of 2.2 days. Remifentanil was used in line with the protocol in 84.5% cases and LOS in this group is one day. This shows a reduction in LOS of 1.1 days compared with 2.1 days LOS by cumulative ICNARC CMPD data(3). We have also noticed a marked reduction in number of nonclinical transfers to 4(from 35), reduction in average occupancy to 97% (from 103%) and average LOS to 5.18 (from 6.27)against an increase ICU admission to 337 (from 326).

Conclusion

Our data demonstrated a reduction in LOS of 1.1 days, based on appropriate use of Remifentanil in 84.5% of cases. We have also noticed improvement in efficiency of the unit by indirect measurement i.e. increased number of admissions with decreased occupancy and number of nonclinical transfers. Proposed daily drug costs of sL70 resulted in a saving of approximately sL1500(4) through the reduction in LOS enabling more efficient use of available resources. These findings have shown that there is not only economic benefit to be gained with introduction of remifentanil but that we have improved overall performance and efficiency of our Intensive Care Unit. Further ongoing audit is required to assess whether further improvements in resource utilization might be possible on satisfying outstanding training requirements.

0423

Descritive Analysis of Psycho-Affective Disorders and Psychotropic Medication in Patients Admitted to 9 Spanish ICUS

Sandiumenge*1, H. Torrado2, T. Muñoz3, C. Pardo4, J. Alonso5, M. A. Alonso6, C. Chamorro 7, M. Jiménez8

1ICM, University Hospital Joan XXIII, Tarragona, 2ICM, University Hospital Bellvitge, Barcelona, 3ICM, Hospital Txagorritxu, Vitoria, 4ICM, University Hospital Fuenlabrada, Madrid, 5ICM, University Hospital Vall d'Hebron, Barcelona, 6ICM, University Hospital 12 de Octubre, 7ICM, University Hospital Puerta de Hierro, 8ICM, University Hospital Clinico de San Carlos, Madrid, Spain

Introduction

Psycho-affective disorders have a negative impact on long term outcome of patients surviving ICU(1). However, little information exists about their influence on ICU complications and on their management by intensive care practitioners.

Methods

All patients admitted to 9 Spanish ICUs (1 medical, 1 trauma, 1 surgical, 6 medical-surgical) during a 30 day period were followed until death or discharge. Demographic data and previous history of alcohol, illegal or prescribed psychotropic use was recorded at admission. Mechanical ventilation as well as analgesia and sedation procedures were recorded. Agitation and the presence of emotional disorders, as well as prescription of psychotropic medication during ICU stay were also recorded. Statistical analysis was performed using SPSS 13.0. Significance level p < 0.05.

Results

A total of 471 patients (66.9% males), with a mean age of 58.9±17.6 y/o, were admitted to the ICU after a medical (52.2%), post-surgical (38.4%) or traumatic (7.4%) event. Forty-eight patients (10.2%) had depression and 91 (19.3%) took psychotropic medication on regular basis at the time of admission being benzodiazepines (BZD) (n=71; 15.0%), antidepressants (ADP) (n=43; 9.1%) (72% serotonin reuptake inhibitors (SRI)), and neuroleptics (NRL) (n=9; 1.9%) the most common agents. Mean ICU length of stay (LOS) was 5.7±8.3 days. Half of the patients (51.8%) were mechanically ventilated and 48.2% were under continuous sedation and/or analgesia. Agitation and depression was diagnosed in 14.8% (n=70) and 1.6% (n=9) of admitted patients respectively. The use of validated screening/diagnostic tools or psychiatry consultation to diagnose agitation or depression was done only in 2 cases. NRL (62%) and ADP (80%) associated or not to BZD (47% and 25%) were the preferred agents for the treatment of agitation and depression during ICU stay respectively. There was a trend to longer ICU LOS in patients with previous history psychotropic medication (8.6±10.5 hours versus 5.0±6.4 hours; p=0.055). Compared to admission, more patients (44.5%) left the ICU receiving psychotropic medication (p < 0.001), at the expense of an increase on BZD prescription (44.0%) (p < 0.001). On the contrary, patients leaving the ICU with ADP (4.3%) were significantly fewer than those at admission (p < 0.001).

Conclusion

In this study patients with previous history of psycothropic medication were fewer than reported in literature(1). The rate of psychotropic agents used during ICU stay exceeded the diagnosis rate of psycho-affective disorders. The use of tools to screen and diagnose psycho-affective disorders in ICU could improve its detection and adequate treatment.

0424

ICU Sedation, Does It Matters?

Graca*, S. Fontes, P. Santos, E. Neutel, I. Aragao Unidade de Cuidados Intensivos Polivalente, Hospital de Santo António, Porto, Portugal

Introduction

Critically ill patients requiring mechanical ventilation are frequently treated with sedatives and analgesics. Important complications related to sedation practices in the intensive care unit (ICU) have been documented and efforts to modify sedation practices have begun. Administration of sedatives by continuous infusion has been identified as an independent predictor of a longer duration of mechanical ventilation as well as a longer stay in the ICU. Our aim is to know how sedation procedure is carried out in our unit and to propose alternative interventions.

Methods

Retrospective analysis of all patients who received mechanical ventilation and continuous intravenous infusion with sedative drugs in our adult polyvalent ICU. Study covers the period between June and December 2007. The exclusion criteria were less than 24 h in the ICU and less than 16 h of sedation. In M group (midazolam+morphine) and in P group (propofol+morphine) were the sedatives used. Variables analyzed: gender, age, weight, SAPS II score, admission diagnosis, duration of mechanical ventilation, length of ICU stay (LOS) and time of sedation. For statistical analyses we used median with 25th and 75th percentiles and mean±SD, t test and statistical significance p < 0.05.

Results

A total of 115 patients were studied. M group (64 patients): 70% men, median age 52 (38–67) years, median weight 72 (65–77)Kg, median SAPS II score 41 (33–50), the admission diagnosis was trauma in 25%, medical in 54,6% and surgical in 20.4%. P group (51 patients): 64.7% men, median age 50 (35–62) years; median weight 70 (64–75)Kg, median SAPS II score 41 (34–49), the admission diagnosis was trauma in 35.3%, medical in 47% and surgical in 17.7%. The mean LOS was 11.76±7.57 days for M group and 11.67±9.26 days for P group (p=0.48). The mean time of sedation was 5.96±5.48 days for M group and 4.27±3.31 days for P group (p=0.028); the mean duration of mechanical ventilation was 7.23±5.61 days for M group and 5.03±3.51 days for P group (p=0.008). The mean time of tracheal intubation was 8.52±5.82 days in M group and 6.22±3.94 days in P group (p=0.009). In M group the mean consume of midazolam was 0.107±0.179mg/Kg/h and morphine 0.007±0.204mg/Kg/h. In P group the mean consume of propofol was 1.79±0.70mg/Kg/h and morphine 0.0147±0.32mg/Kg/h.

Conclusion

In patients who received mechanical ventilation the use of propofol instead of midazolam decreased the time of sedation, mechanical ventilation and tracheal intubation. There was no statistical difference in LOS in ICU. There is a statistical difference between the two groups that should guide towards propofol use. Nevertheless most of the times probably the choice is based on the admission diagnosis and hemodynamic parameters.

0425

Implementation of Assessment and Management of Pain: Introduction of a Painsedation Treatment Algoritm

H. T. T. Nguyen*1, M. van Dijk2, H. A. Bruining3, J. F. Schoonderbeek1

1Intensive Care Unit, Ikazia Hospital Rotterdam, 2Department of Pediatric Surgery, Erasmus MC- Sophia, 3Department of Surgery, Erasmus MC, Rotterdam, Netherlands

Introduction

It is recommended to evaluate the levels of pain and agitation in critically ill patients, but only 43% of the ICU's use an analgeso-sedation scale. Implementation of assessment and management of pain also seems to be difficult. This investigation by questionnaire is performed to see whether it was desired to protocolize our pain and sedation therapy. After that, we introduced an algorithm, based on the Critically Ill Assessment score (CIA-score), and analyzed the compliance.

Methods

We investigated attitude versus behaviour towards a new clinical strategy. Attitude was analyzed by the results of the questionnaire, where behaviour was analyzed as compliance of a new analgeso-sedation algorithm. This algorithm implies a regular assessment every 4 hours of the CIA-scale and a treatment algorithm in case of high (CIA>9) (with reassessment within 1 hour) or low scores (CIA<7) (with reassessment within 4 hours).

Results

Attitude : all responders (29 nurses and 7 doctors) feel the need for a good management and assessment of pain. 83% of the responders agreed that medication can be administered indepentdently by nurses, when a good analgeso-sedation algorithm is available. Behaviour : the data of 73 patients were analyzed, in total 2596 CIA-scores. There was an adequate score (CIA 7–9) in 69.7% of the assessments. In 9.5% the CIA-score was <7, according to the algorithm reassessment should be performed within 4 hours and this was done in 64.6%. In 20.8% the CIA-score was >9, and reassessment within 1 hour was done in only 26.1% of the cases. Scores <7 indicate a high probability for oversedation, treatment only occurs when a repeated CIA-score is too low, which resulted in weaning in 14.1% and no intervention in 85.9% of the cases with a score<7. For the high scores(CIA >9) 49.7% did not result in a medical intervention within 1 hour. If an intervention was performed, it was more often increasing the sedative drugs (31.8%) than of the analgesive drugs (10.5%).

Conclusion

In almost 70% of the patients on our ICU the analgesosedative strategy is according to department-guidelines. This means that in the remaing 30% the compliance to the analgeso-sedative algorithm is not sufficient. Despite all promotion and analyzing of the attitude towards management and assessment of pain on our ward, the introduction of this new algorithm seems to have a limited effect on changing the behaviour of health care professionals.

0426

Mechanical Ventilatory Support and High Levels of Sedation and Analgesia are Poor Prognostic Factors to Critical Ill Oncologic Patients

D. R. Cardoso, M. S. Blanco, D. A. Guedes, A. C. H. Pereira, D. F. Rohrs, M. F. P. Q. Neta, F. S. Batista, P. D. C. S. Machado, O. H. C. Messeder, J. M. Teles*

Intensive Care Unit, Hospital Português, Salvador, Brazil

Introduction

The objective of this study is to available if mechanical ventilatory support and high levels of sedation and analgesia are related to ICU and hospital mortality to oncologic patients.

Methods

This study is a prospective cohort that was carried out in an Intensive Care Unit of an urban tertiary 300 bed hospital in Brazil. Data were collected prospectively from 200 consecutive patients who were admitted to the combined medical and surgical ICU. The outcomes studied were overall ICU and hospital mortalities. In the initial phase of the statistical analyses, relative risks of death to each independent variable and its respective 95% confidence interval were calculated. In a second phase, the Kaplan- Meier procediment was used. The final phase of the statistical analyses consisted of the Cox multiple regressions.

Results

Overall, 56, 2% of patients used invasive ventilatory support; 56, 7% used noninvasive ventilatory support and 24, 6% didn't use ventilatory support. There was significant difference between mortality of patients who needed ventilatory support and those who didn't need (p < 0,024). The survival of patients requiring invasive mechanical ventilation was 3 to 4-fold lower (p < 0,006) and the survival of patients requering non-invasive ventilation was 2 to 3-fold lower (p < 0,030). Additionally, those patients who needed high levels of analgesia (p < 0,001) and sedation (p < 0,009) (three or more medications) had higher ICU mortality rate. These variables are also related to hospitalar mortality: invasive mechanical ventilation (p < 0,006); non-invasive ventilation (p < 0,005); analgesia (p < 0,001) and sedation (p < 0, 0001).

Conclusion

Our study suggests that patients with high levels of sedation and analgesy and who had used mechanical ventilation during ICU stay had higher ICU and hospital mortality.

0427

Safety, Efficacy and Cost-Effectiveness of Acitrom for Deep Vein Thrombosis (DVT) Prophylaxis in Patients Admitted to ICU with Respiratory Failure- a Preliminary Experience

A. Azim*, R. Singh, A. Baronia

Critical Care Medicine, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India

Introduction

We present a safe and cost-effective strategy of using oral anticoagulant (Acitrom) for DVT prophylaxis in patients admitted to our ICU with respiratory failure requiring mechanical lung ventilation.

Methods

After getting Institutes ethical clearance between March 2007 and March 2008, thirty patients of respiratory failure requiring mechanical lung ventilation were randomly included prospectively in the study. Category of patients included in the study were Guillain Barre Syndrome, Myasthenia Gravis, Motor neuron disease, Status epilepticus, Community acquired pneumonia and Acute exacerbation of COPD requiring mechanical lung ventilation. From each patient the following data was collected at ICU admission:Age,Gender, hospital-admission diagnosis (Any co-morbid conditions),details of prior use of anticoagulant therapy, presence of other risk factors for DVT Exclusion criteria included: Age less than 16 yrs, patients with liver disease, contra-indication for use of Vit K antagonists(eg.Pregnancy)and any anticoagulation. Drug Protocol: After informed consent patients were randomized either to oral anticoagulant therapy (Acitrom) or for low molecular weight heparin (LMWH)group. In the Acitrom group LMWH was given along with oral anticoagulant for 4–5 days until we achieved the target INR of 2–2.5. Elderly, diabetics and hypertensive patients received 1mg/day and all other patients received 2mg/day. After achieving the INR LMWH was stopped and the patient was maintained on Acitrom. INR was done every fifth day or whenever any change in dosage of Acitrom was required. In the LMWH group daily prophylaxis was given based on international practice. Any complication related to anticoagulant therapy was taken into account in both the arms. Therapy was continued until the patient became ambulatory or was discharged from ICU. Surveillance Doppler of lower limbs was done every 15 days to rule out asymptomatic deep vein thrombosis and to see the efficacy of the drug therapy.

Results

The demographic profile and category of patients in two groups was identical. No significant drug related complication was observed in both groups. Surveillance Doppler did not show any evidence of occlusive or non occlusive deep vein thrombosis in both groups. There was statistically significant difference in the drug expenditure between the two groups. In the Acitrom group the average cost per patient was 2–3% of the cost of LMWH cost per patient.

Conclusion

Acitrom has been regularly used for prevention of reccurrent thromboembolism but there is paucity of literature regarding its use for DVT prophylaxis in Critically ill patients. We conclude from our limited experience that acitrom can be a safe and highly cost effective alternative to LMWH. However a larger trial is imperative to make any definitive conclusions.

0428

Effectiveness of a Sedoanalgesia Protocol in Non-Intubated Critically Ill Patients

J. Martínez Melgar1, N. Piñeiro Sande1, E. Alemparte Pardavila1, E. Merayo Macias1, E. Moreno Lopez*2

1Intensive Care Unit, 2Anesthesiology and Reanimation, Montecelo Hospital, Pontevedra, Spain

Introduction

To evaluate the experience and effectiveness of a protocol of sedoanalgesia in non-intubated critically ill patients.

Methods

A prospective and descriptive study was carried out on non-intubated critically ill patients between April 2006 and May 2007.A protocol of sedoanalgesia with Remifentanil (RF) was used in those patients who need pain control, propofol was used in those patients excluded from the previous group and remifentanil after extubation (RF-EX) in patients under mechanical ventilation sedated with remifentanil who are extubated keeping remifentanil at analgesic doses. Sedation for more than six hours, sedation for procedures and need for intubation in less than six hours (unless if it was a direct consequence of sedation) were excluded. Therapeutic failure was defined as: need for intubation for sedation, RASS or VAS out of optimal range (considering optimal sedation range as RASS between +1 and −2 and optimal analgesia as VAS between 0 and 3) or non-optimal doses of drugs (RF> 6µgr/kg/h for more than four hours or PF>2,5mg/kg/h for more than 4 hours). The following parameters were analyzed: age, APACHE score, reason for admission, indication for sedation, mean dose of sedatives, time spent under sedation, mean ICU stay, quality of sedoanalgesia, non-optimal doses and secondary effects.

Results

32 patients were included (26 males), 9 were sedated with RF, 12 with PF and 11 with RF-EX. Mean age: 45±14 years, APACHE score: 15±2, reason for admittance in the ICU: neurological (10 patients), respiratory (5 patients), politraumatized (15 patients), other pathologies (2 patients). Indications for sedation: in the RF group: pain control (100%), in PF group: agitation (58%) and neurological evaluation (18%) and in the RF-EX group: pain control (45%), neurological evaluation (18%) and pharmacological deprivation (27%). Mean sedative doses: RF group 3,52 ±0,5µgr/kg/h, PF group 1,59±0,3mg/kg/h and in RF-EX group 2,30±0,6µgr/kg/h. Sedation time: with RF 162±60 hours, with PF 100±43 and in RF-EX group 82±30 hours. Optimal sedation level: with RF 86%, with PF 75% and with RF-EX 91%; optimal analgesic level: with RF 89% and with RF-EX 100%. Optimal drug doses: with RF 78%, with PF 67% and with RF-EX 91%.

Conclusion

According to our sedoanalgesia protocol, in non-intubated critically ill patients, doses of 1, 59±0,3mg/kg/h of propofol achieved an optimal level of sedation in 75% of the patients. When using RF according to our protocol, an optimal level of sedation and analgesia is achieved between 80 and 90% using doses of 3, 52±0, 5µgr/kg/h.

Poster Sessions Fungal infections/Device-related infections: 0429–0442

0429

In Vitro Susceptibility of Intensive Care Unit (ICU) Candidemia Isolates By the Eucast and Commercial Methods — A Ten Year Survey

G. Dimopoulos1, A. Velegraki2, D. K. Matthaiou*3, E. C. Alexopoulos4, A. Zannos2, M. E. Falagas3

1ICU, University Hospital Attikon, 2Mycology Lab, Microbiology, Medical School, University of Athens, 3Biomedical Sciences, Alfa Institute of Biomedical Sciences, Athens, 4Public Health, Medical School, University of Patras, Patras, Greece

Introduction

In recent years a shift is reported in the frequency of candidemia due to non-albicans spp attributed to the increasing use of fluconazole in the ICU setting. This study reports data from Greece on the frequency of Candida spp. isolation from ICU candidemia episodes, their susceptibility to antifungals during the last decade and compares the performance of 3 commercial susceptibility testing methods with the EUCAST microdilution(MD) reference method.

Methods

We tested 137 Candida isolates ICU acquired candidemia episodes between 1997–2002 when increased use of fluconazole prophylaxis was engaged, and 2003–2007, when administration of empirical FL was reduced and the newer antifungals, such as caspofungin were gradually introduced. The conventional antifungals tested comprised fluconazoleFL), amphotericin B(AB), itraconazole(IT) and flucytocine(FC), whereas the newer antifungals included voriconazole(VO), caspofungin(CAS)and posaconazole(POS). The new candins, micafungin(MICA) and anidulafungin(AND) were also tested in vitro. Susceptibilities were tested by the EUCAST microdilution method the, Sensititre Yeast 9 panel (ST), Etest(Et), and by the CLSI M44 A(FL and POS).

Results

The frequency of the non-albicans species decreased significantly in the last 5 years of the study(from 82.3% to 46.6%, p < 0.001). Increased AB MICs were recorded for C. lusitaniae (1–4 mg/L) and C. tropicalis (0.5–1 mg/L) and increased candin MICs against C. parapsilosis (0.5–4 mg/L), C. guilliermondii (0.25–8 mg/L), C. rugosa (0.12–8) and C. kefyr (0.5–16). As expected high FL MICs were noted for C. glabrata and C. krusei. VO, POS and IT generally generated low MICs except against 2 C. glabrata strains (1 mg/L) respectively. Interclass correlation coefficient (ICC) for EUCAST versus sensititre vs. Etest was 0.98–0.99 (P<10–4) for all drugs. Pearson's correlation coefficient (measure of linear associations) for FLU and POS vs EUCAST MD vs Etest, vs disk diffusion was 0.944 (P<0.01).

Conclusion

Significant decrease (p < 0.001) in the non-albicans ICU candidemia isolates was recorded in the last 5 years, attributed to the rational use of fluconazole. FL MICs did not significantly decrease in the last 5 years (p>0.1). The overall agreement of commercial versus EUCAST was (94–98%). The candins performed well, except against the aforementioned Candida spp. Disk diffusion is an option for rapidly testing candidemia isolates. Standardized susceptibility testing can aid the selection of the most relevant antifungal therapy for the management of ICU candidemia.

Grant Acknowledgement

This study was supported in part by Gilead, Greece.

0430

Impact On Mortality of Empirical Treatment and Non Albicans Especies in Candidemia in Non Neutropenic Critically Ill Patients

R. Zaragoza*1, S. Sancho1, J. Camarena2, A. Artero3, C. Lloret1, R. González2, J. Nogueira2

1Intensive Care Unit, 2Microbiology, 3Internal Medicine, Hospital Universitario Dr. Peset, Valencia, Spain

Introduction

The aims of this stdy was: a) To know the prevalence of non-C.albicans species (NCA) as cause of Candidemia in non neutropenic critically ill patients (CNNCIP) and its influence on global and related mortality to infection, b) to describe the rate of inadequate empirical treatment (IEAT) in this setting and its consequences on outcome, and finally c) to analyse the impact on mortality of the different empirical antifungal therapies (broad spectrum vs Fluconazole).

Methods

A prospective and observational study was developed in a Spanish teaching hospital mixed-ICU (16 beds) during 12 years (1996–2007). All episodes of candidemia were collected. Clinical, microbiological and outcome variables were recorded. Uni and multivariate analysis were performed to assess the influence of inadequate empirical treatment, the antifungal agents used as empirical treatment and the species involved on mortality (SPSS 13.0).

Results

Among 397 bloodstream infections 7,8% (n=31) of the episodes were CNNCIP. The most frequent isolates were: Candida.albicans (51,7%), Candida parapsilosis (22,5%) and Candida glabrata (19,3%).There were no differences in age, sex and APACHE II score between C.albicans y non albicans species. Previous use of fluconazole was not more frequent in NCA candidemia (25%vs 26,6%; p= 0,91). Global and related mortality to candidemia were 67,7% and 29% respectively. The rate of (IEAT) was 70,9%, but it had no impact on mortality rates. The related to candidemia mortality rate was statistically higher in NCA (46,6% vs 12,5%; p= 0,03). The frequency of each antifungal drug used as empirical therapy was: Fluconazole (31,2%), Amphotericin B (31,2%), Caspofungin (18,7%) and Voriconazole (18,7%). The empirical antifungal therapy chosen had no impact on mortality neither global candidemias analysis nor NCA subanalysis. A multiivariate analysis confirmed the NCA isolate as the unique factor associated to related mortality to candidemia (OR=6.06; CI95%:1.01–36.6; p= 0.05).

Conclusion

We have observed an increase of NCA fungemia in non neutropenic critically ill patients with a mortality rate higher that those caused by C. albicans. This higher mortality was not associated neither IEAT nor the antifungal drug used as empirical therapy.

0431

Voriconazole Dosing in Critically ill Patients Undergoing Continuous Venovenous Hemofiltration

J. Radej*1, A. Krouzecky1, P. Stehlik2, R. Sykora1, J. Chvojka1, T. Karvunidis1, I. Novak1, M. Matejovic1

11st Medical Dept., ICU, 2 Institute of Clinical Biochemistry and Hematology, Charles University Medical School and Teaching Hospital, Plzen, Czech Republic

Introduction

Voriconazole (VRC) is a triazole potent broad-spectrum antifungal agent, an important part of antimicrobial therapy in critically ill patients. VRC is predominantly eliminated by metabolism. Continuous renal replacement therapy is one of the standard methods in the ICU. There is a lack information about pharmacokinetic alteration of VRC during these procedures to date.

Methods

We measured VRC concentration in serum and in ultrafiltrate by RP-HPLC method with UV detection in critically ill patients undergoing continuous venovenous hemofiltration (CVVH). We profiled 5-point pharmacokinetic concentration-time curve during a 12 hours interval of standard maintenance dosing 4 mg / kg. We derived an area under the curve (AUC), sieving coefficient (Sc) and total body clearance (CLt) and compared these with the literature. At the same time, we reviewed pharmacodynamics of VRC, minimal inhibitory concentration (MIC) for Candida spp. and for filamentous fungi in the literature.

Results

We are presenting preliminary data of three patients with septic shock, renal failure and suffering from an invasive mycotic infection. AUC was 22.8, 73.5 and 27.0 mg / h /l. CLt was 17.5, 5.4 and 14.8 l / h. Sc was 0.19, 0.06 and 0.16. Measured serum concentrations of VRC in every time point were higher than 1 mg / l which represents the MIC90 value for the all Candida spp. and all filamentous fungi.

Conclusion

Standard dosing of voriconazole was adequate during CVVH in our three critically ill patients and dosing adjustment was not required.

Grant Acknowledgement

Supported by MSM 0021620819: Replacement of and support to some vital organs.

0432

Changes in the Consumption of Antifungals in Critically Ill Patients Admitted to Spanish Icus (2002-2007 Period)

F. Alvarez-Lerma*1, M. Palomar2, P. Olaechea3, J. Insausti4, J. Sanchez-Godoy5, M. López-Pueyo6, M. Gracia1

1ICU, Hospital Universitario del Mar, 2ICU, Hospital Vall d'Hebrón, Barcelona, 3ICU, Hospital del Galdakao, Bilbao, 4ICU, Hospital de Navarra, Pamplona, 5ICU, Htal Nuestra Sra de laCandelaria, Santa Cruz de Tenerife, 6ICU, Hospital General Yague, Burgos, Spain

Introduction

To describe the changes in the use of antifungals in the Services of ICM (ICU) and the impact of the new antifungals.

Methods

Observational, prospective and multicenter study based on patients included in the National Surveillance Study of Nosocomial Infection (ENVIN) between 2002 and 2007. Prescriptions of nystatin, fluconazole, itraconazole, voriconazole, amphotericin B deoxycholate, liposomal amphotericin, amphotericin B lipid complex and caspofungin directed to prophylaxis, community-acquired infection, extra-ICU nosocomial infection, and ICUacquired infection ere analyzed. Type of treatment (empirical or specific) and duration were also assessed. Results are presented in percentages in relation to the total number of antimicrobials used for each indication and/or period.

Results

During the study period a total of 52,613 patients were controlled, 29,930 (56.9%) of whom received 66,024 antimicrobials, 3184 (4.8%) of which were antifungal agents. Absolute indications and percentages in relation to the total number of antimicrobials as well as the distribution by type of indication are presented in the tables. Major findings included a predominance of fluconazole (56.9% of all antifungals) with an increase of use during the study period (2.4% to 3.1% of all antimicrobials), decrease of amphotericin deoxycholate (0.5% to 0.03%), and increase of caspofungin (0% to 1.03%) and voriconazole (0% to 0.69%). The use of lipid amphotericin formulations did not vary. There was a predominance of fluconazole in all indications. Mean duration of treatment in the ICU ranged between 8 and 11 days. Empirical treatments were the most frequent (60%).

Conclusion

Limited use of antifungal agents in critically ill patients. Fluconazole was the most commonly used antifungal drug in all indications and showing an increasing trend. Caspofungin and voriconazole showed a progressive increasing use and a parallel decrease for the indications of amphotericin B deoxycholate. Predominance of the use of antifungals in empirical treatment

Grant Acknowledgement

(*) ENVIN-UCI sudy supported by Sanofi-Aventis.

0433

Human Tissue Concentrations of Voriconazole

S. Weiler, S. Dunzendorfer*, R. Bellmann-Weiler, M. Joannidis, R. Bellmann

Dept. of Internal Medicine, Division of General Internal Medicine, Innsbruck Medical University, Innsbruck, Austria

Introduction

The broad-spectrum antifungal voriconazole (VRC) is used against various invasive mycoses in critically ill patients. VRC exhibits a good penetration into body fluids, such as cerebrospinal fluid, epithelial lining fluid and pleural effusion. VRC tissue levels were determined in tissue samples obtained during routine autopsy from patients who have died during treatment with VRC.

Methods

VRC levels were determined in various tissue samples (kidneys, liver, heart, spleen, lung, brain) of 2 patients, who had been on treatment with VRC. Patient 1 had received a single dose of 200 mg, patient 2 had obtained a total dose of 3,800 mg. The intervals between the last administration and death were 36 and 12 hours, respectively. Both patients had been on vasopressor therapy and mechanical ventilation. Antimycotic therapy had been initiated for suspected or proven invasive fungal infections. VRC tissue levels were assessed by extraction of homogenized tissue samples, purification by solid phase extraction and measurement of VRC by high performance liquid chromatography.

Results

Tissue levels of VRC in patient 2 exceeded those achieved in patient 1 in all samples. Highest VRC concentrations (mean ± standard deviation) in VRC treated patients have been found in the organs of elimination — the liver (patient 1: 2.14 ± 0.40 µg/mL, patient 2: 4.21 ± 0.77 µg/mL) and the kidneys (patient 1: 1.97 ± 0.41 µg/mL, patient 2: 6.89 ± 0.06 µg/mL). Mean VRC lung concentration amounted to 1.30 ± 0.63 µg/ml. In the lung of patient 1, VRC levels reached 0.72 – 0.76 µg/ml after a single dose. In patient 2, lung concentrations ranged between 1.47 and 2.04 µg/ml. In different areas of brain tissue VRC was below the limit of detection (<0.25 µg/mL) in patient 1 and achieved levels of 3.34 ± 0.18 µg/mL in patient 2. There was no difference in VRC concentrations between different areas of the brain, such as cortex, hippocampus, nucleus caudatus, medulla oblongata and cerebellum. VRC concentrations amounted to 1.31 ± 0.03 µg/mL and 2.95 ± 0.05 µg/mL in samples of the spleen of patient 1 and 2, respectively. In the myocardium samples of patient 1 VRC was undetectable, but reached a mean concentration of 2,44 ± 0,25 µg/mL in patient 2.

Conclusion

VRC is detectable in the lung and other tissues already after the firs administration and thus appears to penetrate rapidly into human tissues. VRC accumulates in the liver and the kidneys, which are also the routes of elimination. Its penetration into brain and heart is probably slower.

Grant Acknowledgement

We thank Pfizer Austria for financial support.

0434

Candidemia in Critically Ill Patient in Tertiary Care Intensive Care Unit

K. A. Dalal*, S. K. Bindroo, I. P. Raut, T. Karamata, M. H. Patel, J. D. Sunavala

Critical Care Medicine, Jaslok Hospital and Research Centre, Mumbai, India

Introduction

Candidemia is frequently a life-threatening complication in critically ill patients admitted to intensive care units.

Methods

A 2-year retrospective observational study conducted in the open ICU of a tertiary care hospital at Mumbai, India. Patients whose blood cultures were positive for fungal growth were included in the study. The incidence, sub classification, risk factors and mortality were assessed from the data.

Results
From Feb 2006 to Jan 2008, 121 critically ill patients were found to have positive blood cultures for candida groups of fungi. Out of this, 80 (66.11%) patients had acquired the infection as Nosocomial and rest 41 (32.89%) patients had positive blood cultures at the time of admission to the hospital / ICU. The mean incidence of candidemia was 16.8/1000 admissions. Candida albicans and candida tropicalis were isolated in 87.6% and 8.26% of patients, respectively. The main risk factors for candidemia were malignancy (54.5%), diabetes (35%) and end stage renal disease (21%). Crude mortality was 40%.
Table 1

Subclassification of Candida Species

Subclassification

No. of patients

Percentage

Candida Albicans

106

87.6%

Candida Tropicalis

10

8.26%

Candida Glebrata

03

2.4%

Candida Parapsilosis

01

0.09%

Candida Krusei

01

0.09%

Table 2

Risk Factors Associated With Candidemia

Risk Factors

No. of patients

Percentage

Malignancy

66

54.54%

Diabetes

35

28.92%

End-stage renal disease

21

17.35%

Parentral Nutrition

16

13.22%

Chronic Liver Disease

08

09.68%

Others

12

14.82%

Conclusion

Candidemia is a common occurrence in critically ill patients admitted to Intensive Care Unit of our hospital. The main underlying risk factor is immuno-suppression due to various causes. If not promptly recognized and adequately treatment, mortality remains high. So, we should have a high index of suspicion of candidemia in critically ill patients.

0435

Efficacy and Safety of Liposomal Amphotericin B in Candidemia in Patients Admitted to Intensive Care Units: A Retrospective, Multicenter, Pharmacoepidemiological Study

F. Á lvarez-Lerma*1, F. Mariscal2, M. J. Pérez3, M. Nieto4, B. Á lvarez5, I. Jordán6

1Unidad de Cuidados Intensivos, Hospital del Mar, Barcelona, 2Unidad de Cuidados Intensivos, Hospital La Paz, 3Unidad de Cuidados Intensivos, Hospital La Princesa, 4Unidad de Cuidados Intensivos, Hospital Clínico San Carlos, Madrid, 5Unidad de Cuidados Intensivos, Hospital General Universitario de Alicante, Alicante, 6Unidad de Cuidados Intensivos, Hospital Sant Joan de Deu, Barcelona, Spain

Introduction

Invasive candidiasis, the most frequent fungal infection within high risk patients hospitalized in intensive care units (ICUs), is associated with a high mortality rate. The objective of this study was to analyse the efficacy and safety of liposomal amphotericin B (L-AMB) administered to ICU inpatients with candidemia.

Methods

Retrospective, multicenter study of ICU inpatients with candidemia treated with L-AMB in 2006.

Results

Forty-one patients were included in the study. Mean APACHE II score was 21.4 (SD: 7.7), mean time at ICU was 44.3 (SD: 32.9) days and mortality rate at ICU was 48.8%. Most common causes of admission at ICU were medical pathology (43.9%) and surgery (41.5%), and 65.9% of the patients had severe sepsis or septic shock. In most of the cases candidemia was due to Candida albicans (65.9%), follow by C. parapsilosis (9.8) and C glabrata (7.3%) and others (9.7%). Mean duration of treatment was 13.8 days and mean dose was 3.8 mg/kg/day. More than half of the patients had received previous antifungal treatment (63.4%), mainly fluconazole (34.1%) and caspofungin (22.0%). L-AMB most common indications were: non-stable disease (51.2%), infection localization (34.1%) and guidelines application (29.3%). Satisfactory clinical response and microbiological response was achieved in 58.5% (95% CI: 43.5, 73.6) and 63.4% (95% CI: 48.7, 78.2) of the patients, respectively. In evaluable patients, satisfactory clinical response was 66.6% (95% CI: 28.2, 60.7) and microbiological response was 78.8% (95% CI: 64.8%, 92.7%). None of the adverse reactions detected were reported as serious. There was no renal failure requiring a change in the antifungal treatment and no change in the mean creatinine value despite the fact that 58.5% of patients were receiving nephrotoxic drugs.

Conclusion

L-AMB was used in critically ill patients with candidemia previously treated or not. Satisfactory clinical response was very high. L-AMB was well tolerated even in patients taking concomitant nephrotoxic drugs.

0436

Candidaemia and in Vitro Sensitivity of Candida Isolates in A General Hospital and in Its Icu

Koteli*1, S. Tsingene1, S. Vasiliagkou2, E. Antoniadou2, M. Kriti1

1Biopathology Laboratory, 2Intensive Care Unit, G.Gennimatas General Hospital, Thessaloniki, Greece

Introduction

to investigate the production of slime factor among candida spp, isolated from blood cultures, to study in vitro activities of antifungal agents and compare these results with slime production, during 2003 –2006 period.

Methods

we studied a total of 28 Candida spp isolated from blood cultures. They were incubated in Bactek 9120 system (Becton Dickinson) in aerobic, anaerobic and mycosis vials. All positive cultures were Gram stained and re-cultured in blood agar, Mac Conkey agar, Sabouraud with TCC and chromagar candida (Becton Dickinson). The protocol for negative blood cultures included a seven days period. Identification of microorganisms and susceptibility test were performed with the Vitek 1 and 2 systems, API and ATP Fungus (BioMerieux, France).

Results

There were isolated 1 Aspergillus and 27 Candida strains: albicans 15, krusei 9,tropicalis 1, parapsilosis 1, zeylonoides 1. The locations were: ICU 18, medical dept 6, surgical dept 4. The susceptibility test results revealed a difference between 24 and 48 hours incubation time. The MIC for amphotericin B (AMB)for all candida species was 0.25 – 1 µg/ml and they were all characterized as sensitive to AMB as well as to 5-flurocytosin. 5 out of 15 C. Albicans strains have MICs for fluconazole 16–128 µg/ml and high MICs for itraconazole and ketoconazole.

Conclusion

Non-albicans Candida strains had no prevalence over candida albicans. Both candida species showed resistance to one or more antifungal agents and a progressive increase of it over the years studied. Assessment of the susceptibility test after a 48 hours incubation period is recommended.

0437

Pulmonary Artery Catheter Sleeve — An Effective Barrier?

T. B. Corcoran1, S. Grape*1, O. Duff2, R. Murray3

1Department of Anaesthesiology and Pain Medicine, 2Department of ICM, 3Department of Microbiology, Royal Perth Hospital, Perth, Australia

Introduction

Pulmonary artery catheterization is widely used in intensive care. Infectious complications associated with the use of pulmonary artery catheters (PACs) are a significant source of morbidity and mortality. Shielded (“hands-off”) PACs have been found to reduce PAC-related systemic infections. However, there is a risk that mobilization of a PAC within a contaminated sleeve may result in inoculation of organisms into the patient's bloodstream. We conducted a prospective observational trial to address the question whether it is safe to manipulate the PAC within the protective sleeve.

Methods

We prospectively enrolled 102 patients with PACs and collected the following data: APACHE II and SOFA scores, patient demographics, type, site, date and time of PAC insertion, type of surgery, antibiotic therapy at PAC insertion and removal. Upon PAC removal, 4 specimens were taken under sterile conditions: Saline reaspirated from the protective sleeve (1, “sleeve fluid”), the distal 5 cm of the PACs (2, “tip”), swab specimens from the introducer hub (3, “hub swab)” and from the skin at the PAC exit site (4, “exit swab”). In the microbiology laboratory, sleeve fluid and tips were inoculated onto horse blood agar and incubated. After 48 hours, colony counts were performed. Catheter hub and exit swabs were also inoculated onto horse blood agar and incubated. After 48 hours, growth of organisms was classified as “Nil” (no growth), “Colonization” (scanty growth) or “Infective” (moderate/abundant growth).

Results

Mean duration of PAC residence was 39.1 (24) hours. 92 PACs were internal jugular, 8 subclavian and 2 femoral. 11 patients had blood cultures (all negative), and there was no episode of catheter-related bloodstream infection. 6 patients had an infective shield fluid (5 coagulase-negative staphylococci (CNS), 1 mixed gram-negative bacilli (GNB)). 4 catheter tips were infective (3 CNS, 10 Escherichia coli), 6 introducer hubs were colonized (4 CNS, 1 GNB, 1 diphtheroid). Patients having PACs with an infective sleeve fluid did not differ significantly from those without an infective sleeve fluid in terms of APACHE/SOFA scores, age, gender or site of insertion. There was a highly significant association between a colonized introducer hub and growth of organisms in the shield fluid, OR 30 (4–213, p< 0.001). Neither a positive PAC tip nor a positive exit site swab was associated with an infective sleeve fluid culture.

Conclusion

In this cohort of 102 patients with a short duration of PAC residence, 6 patients had potentially infective cultures of sleeve fluid. This was significantly associated with a colonized introducer hub. These figures suggest that advancing a PAC within a “sterile” protective sleeve may have the potential to inoculate organisms into the patient's bloodstream.

0438

Reduction of the Catheter-Related Bloodstream Infections in an Intensive Care Unit

R. Peredo*, C. Sabatier, A. Villagrá, J. González, C. Hernández, F. Pérez, D. Suárez, J. Vallés

Intensive Care Unit, Hospital Parc Taulí, Sabadell, Spain

Introduction

To determine the utility of a multiple system intervention to reduce catheter-related bloodstream infections (CR-BSI) in an intensive care unit (ICU).

Methods

We carried out a prospective cohort study in a medical and surgical ICU. We determine the rate of CR-BSI per 1000 catheters days during the application of an evidence-based intervention used to decrease the CR-BSI in 2007 (March to December) compared with the rate during the same period in 2006 in which we just applied conventional measures of prevention.

During the intervention period we applied five measures: giving educational sessions about how to insert and maintain central catheters, cleaning the skin with chlorhexidine, filling in a checklist during the insertion of the catheter, using the subclavian vein as the preferred site and avoiding the femoral site if possible, and removing unnecessary catheters. CR-BSI was defined as the recovery of the same organism (same species, same antibiotic susceptibility profile) from the catheter tip and blood cultures.

Results

During the control and intervention period we registered 4289 vs 4174 patient-days and 3572 vs 3296 catheter-days respectively.

During the intervention period 8 CR-BSI were diagnosed compared with 24 CR-BSI in the control period. The mean incidence rate of CR-BSI was 6.7/1000 catheter days in the control period and 2.4/1000 catheter days in the intervention period (RR 0.3; 95% CI 0.1 to 0.7; p= 0.03).

A nursing intervention during the filling of the checklist was required in 17.7% of the insertions. The ratio of use of catheter was 81.5% during the control period and 80.6% in the intervention period without significant differences between periods.

Conclusion

The implementation of a multiple system intervention with evidence-based measures significantly reduces the CR-BSI in our intensive care unit.

0439

Intravascular Catheter-Related Bloodstream Infections — The Results After Implementation of A New Approach for an Old Problem

M. Lugarinho*, L. J. P. Peixoto, P. P. G. Castro, R. Beranger, P. C. P. Souza

Intensive Care Unit, Hospital de Clínicas Mario Lioni, Rio de Janeiro, Brazil

Introduction

Intravascular catheter-related infections are very critical in ICU environment, with elevated morbi-mortality and great impact on costs. In our unit, according to a quality political, it was established standards on prevention, diagnosis and treatment of nosocomial infections, with a periodic review of ours rates. We will describe one year of follow up results after a task force model chosen when we noticed an increase of the catheter related infections incidence: outcome management.

Methods

Prospective study, two-phases model, in a general ICU of 23 beds from December 2006 to January of 2008. First phase: multi-professional work group was created (4 physicians, 6 nurses and 2 respiratory therapists) who performed a meeting with the “brainstorm” technique. All the infections data were reviewed. The group identified main risk factors related to the problem using a diagram cause-effect. Then, it was established corrective measures, deadlines and ways for execution. The second phase was the implementation of the measures chosen: team for catheter insertion; full-barrier precautions for insertion of central venous catheters; semi-permeable and transparent dressings; avoiding the jugular and the femoral sites whenever possible; routine replacement of the catheters after ten days insertion; removal of the unnecessary catheters. The target was the return of catheter-related infection rates of the previously year.

Results

Sixty-one patients were followed during their stay in the ICU. A total of 118 intravascular catheters were used in the following sites: 61,9% subclavian vein (n=73), 19,5% internal jugular vein (n=23) and 18,6% in femoral vein (n=22). The median of catheter indwelling time was 10 days (SD de +/− 5,16). Regarding the rates or infection, there was a reduction in average rate from 13,08 to 7,43 per 1000 days of catheters (p< 0,05), with almost the same density of use (54,25 in 2007 versus 54,67 in 2006). The bacterial related to the infections were: BGN in 58.4% (n = 7), MRSA in 33.3% (n = 4) and staphylococcus coagulase negative 8.3% (n = 1). There was no statistical difference in time or site of the puncture between the infected group with the non-infected.

Conclusion

Catheter-related bloodstream infection is the nosocomial infection par excellence: costly, common, and frequently fatal. A program to improve patient safety must focus on simple and inexpensive interventions and prevention measures. Our target was achieved, although we were not able to identify the one factor that brought the biggest impact alone. However, we will keep our efforts to a new primary goal: zero rate of infection related to intravascular catheters.

0440

Simple Measures and the Reduction of Central Venous Catheter- Related Bloodstream Infections (Cr-Bsi)

M. P. C. D. Damasceno*1, P. Souza1, S. Vinhas1, M. Santos1, F. Henriques1, G. Addor1, A. Freitas2, L. Pinheiro1

1ICU, Hospital de Clínicas de Niteró i, 2Infection Control Team, INFECTO, Rio de Janeiro, Brazil

Introduction

CR-BSI is common, potentially lethal, and can indicate the poor quality and the inexistence of safety culture in an ICU. Some studies have demonstrated interventions to reduce this rate[1,2], which appears different when there is a comparison between United States of America (4.0) and developing countries (12.5)[3]. The aim of this study was to evaluate the impact of some simple measures that we used in our ICU to decrease CR-BSI.

Methods

After we had analyzed CR-BSI rate in our 10 bed medical-surgical ICU in 2006 using National Nosocomial Infection Surveillance (NNIS) methodology [4], we studied the problem and implemented simple and not expensive measures. Better control of catheter insertion with development of a central-line checklist, not use of femoral site if possible and the daily rounds discussing the removal of catheters were the main actions adopted. The analysis of 2007 rate was done and compared with 2006.

Results
Results are summarized in table.
Table 1

Periods Rates Using Nnis Methodology

 

2006

2007

Rate of Central Venous Catheter use*

0.61 (0.44 − 0.76)

0.54 (0.48 − 0.61)

Rate of CR-BSI per 1000 device days*

23.08 (8.33 − 33.52)

11.75 (0 − 20.59)

* Overall (pooled) and 10th to 90th percentile range for months of the years

Conclusion

Simple measures resulted in an expressive reduction (55.3%) in rates of CR-BSI. With this result, that is similar to the developing countries, we intend to maintain it and reach better rates.

0441

Age as A Risk Factor for Devices Related Infection

M. Palomar*1, F. Alvarez lerma2, P. Olaechea3, J. Otal4, M. Carrasco1, J. Insausti5, M. Lopez Pueyo6

1ICU, H Vall Hebron, 2ICU, H Mar, BCN, 3ICU, H Galdacano, Bilbo, 4Preventive M, H Vall Hebron, BCN, 5ICU, H Provincial, Pamplona, 6Preventive M, G Yague, Burgos, Spain

Introduction

Elderly people are more prone to infections.

Until recently there was a tendency to treat them in a more conservative way, but nowadays their admission in the ICU has become normal as well as the strong support they receive- Objective: to know the age- related infection rate (IR) in the critically ill patients.

Methods

Prospective multicenter study carried out during 3 months in the year 2007. Patients were allocated in three groups (A;B;C) according to their age: less than 60 (A), between 60 –74 (B) and equal or more than 75 years (C). The characteristics of patients were analysed as well as the use of devices. IR is shown per 100 patients and per 1000 days of device. Ventilator associated pneumonia (VAP), urine infection related to urinary catheter (UTI), and primary and vascular catheter related bacteremia (PB +CRB) were monitored.

Results
The survey included 12.453 patients admitted to the ICUs for more than >24h, with 38,8% in the group A, 35,2% in group B and 25,9% in group C, which represent an increase with regard to the previous years. The incidence rates were: use of device (days with device/days of stay) 0,49%, 0,48% and 0,43% for VAP; 0,82–0,81 y 0,80% for UC; and 1,33–1,33 y 1,21% for PB +CB, including arterial (AC) and venous catheter (CVC). Density and rates of infection are shown in table.
 

<60 y

60–75 y

>75 y

N

4.835

4.387

3.231

VAP × 100 pts

7,74

5,6

3,28

VAP × 1000 d MV

19,3

14,6

15,9

UTI × 100 pts

2,9

3,4

3,2

UTI × 1000 d UC

4,49

5,40

5,6

PB+CRB × 100 pts

3,6

3,2

2,3

PB+CRB × 1000 d CVC+AC

5,16

4,86

4,05

Conclusion

A quarter of the ICU’ populations are ≥75 years. Exposition to medical devices decrease with age, specially MV and vascular catheters. Infection Rates are lower in the eldest except for UTI. Differences are less significant using incidence density rather than by 100 patients.

Grant Acknowledgement

Aventis.

0442

Epidurals in Patients with Sepsis

P. F. Yoxall*, S. S. Swaraj, S. S. Nagaraja

Anaesthetic, University Hospital Aintree, Liverpool, United Kingdom

Introduction

Epidural analgesia has become widely regarded as the “gold standard” for post-operative pain relief after major surgery. Due to a lack of data in septic patients receiving epidurals, we have little knowledge of the true perioperative complication rates. The underlying knowledge of epidural abscess and haematoma rates are therefore unknown, making patient consent in septic individuals a difficult issue. The complications rate for patients (the majority being elective)receiving central nerve blockade is quoted as 0.1–0.001% for abscess formation 1. Following the MASTER trial there is now some doubt to whether or not there is any improvement in the postoperative mortality in high risk patients undergoing major surgery 2 despite the apparent reduction in respiratory morbidity.

Methods

Following an initial retrospective (2004–2006), 30 case note, review in 2006 we continued our audit prospectively through 2007 (22 case note) in septic patients receiving epidural having undergone major emergency abdominal surgery and required intensive care unit (ICU) post-operatively. The aims were to assess complications associated with epidural usage in this particular patient group. Post-operatively patients were regularly reviewed by our acute pain team until no further acute pain input was required.

Results

Of all the epidurals 38 were sited in theatre and 14 in the ICU for analgesia and as an aid to weaning from mechanical ventilation. The ratio of male to female was 27:25. The overall perioperative mortality was 42% (22/52). Two or more organ failure was seen in 33 patients. Positive blood cultures were identified in 5 cases. Of the epidurals 18 required two or more level attempts and the mean duration of in-situ catheter days was 4.7 (+/−1.3). At the time of insertion patient temperatures ranged from 35.6 to 38.9oC. Associated complications occurred in 6 cases with one epidural haematoma, one subdural catheter placement, two vascular punctures and two dural punctures despite the epidural insertions being made by senior trainees or consultants in 49/52 cases.

Conclusion

Due to the rarity of epidural complications in the general surgical population it is not possible to conclude whether the serious complications are greater in septic patients per se, however, it is a question of risk versus benefit in such a patient group. Unless a large scale, multicentre audit or research project is undertaken it is extremely difficult to evaluate both the true, and evidence-based, benefits and risks of epidural analgesia and anaesthesia in the septic patient population undergoing both major surgery and requiring postoperative ICU management.

Poster Sessions Advances in neuro-critical care IV: 0443-0456

0443

Elevated Cardiac Troponin I and Electro-Cardiographic and Echocardiographic Abnormalities After Aneurysmal Subarachnoid Hemorrhage

M. Hravnak*1, M. Frangiskakis2, E. Crago1, M. J. Gallek1, Y. Chang3, M. B. Horowitz3

1School of Nursing, 2Cardiology, School of Medicine, 3Neurological Surgery, School of Medicine, University of Pittsburgh, Pittsburgh, United States

Introduction

Patients with aneurysmal subarachnoid hemorrhage (aSAH) also experience myocardial injury at the time of rupture, but the prevalence and manifestations of neurocardiac injury are not well described. We investigated neurocardiac injury prevalence as quantified by elevated cardiac troponin I (cTnI) ≥0.3ng/ml within the first 5 days after aSAH and its relationship to electrocardiographic (ECG) and echocardiographic abnormalities.

Methods

Prospective longitudinal study of aSAH patients with Fisher grade≥2 and/or Hunt/Hess grade≥3 admitted to the Neuro ICU. Serum cTnI collected for 5 days following hemorrhage with peak value per patient utilized for cohort dichotomization (cTnI≥0.3ng/ml vs cTnI<0.3ng/ml. We evaluated the relationship between cTnI peak and demographics, bleed severity, admission 12-lead ECG, Holter monitoring days 1–5, and echocardiogram.

Results

Of 280 subjects, 31% had cTnI≥0.3ng/ml. Few patients in either group had a past cardiac disease history (10% cTnI≥0.3ng/ml vs 7.2% cTnI<0.3ng/ml, p=.448). A significant relationship existed between elevated cTnI and older age (mean 57yrs±11 cTnI≥0.3ng/ml vs 53yrs±11 cTnI<0.3ng/ml, p=.002), but not race or gender. A significant relationship existed between cTnIC0.3ng≥ml and greater bleed severity by Hunt/Hess (p=<.0001) and admission Glasgow Coma Scale (p<.0001). Patients with higher cTnI were significantly more likely to have initial 12-lead ECG findings of prolonged Qtc (mean 484ms±51 cTnI≥0.3ng/ml vs mean 452ms±47 cTnI<0.3ng/ml, p<.001), but not PR (p=.688) or QRS (p=.217) duration. On Holter monitoring, patients with elevated troponin were more likely to have ventricular dysrhythmias (21% cTnI≥0.3ng/ml vs 9.8% cTnI<0.3ng/ml, p=.030) but not atrial dysrhythmias (p=.473). Elevated troponin was also associated with lower cardiac ejection fraction (p < .0001) and wall motion abnormalities (40.6% cTnI≥0.3ng/ml vs 5.5% cTnI<0.3ng/ml, p < .0001) on echocardiogram, as well as pressor infusion (36.8% cTnI≥0.3ng/ml vs 18% cTnI<0.3ng/ml, p=.006).

Conclusion

Neurocardiac injury determined by elevated cTnI occurs commonly in young patients without cardiac history after aSAH, and is related to bleed severity. Elevated cTnI is associated with ECG abnormalities at admission and throughout the first 5 days after hemorrhage, as well with as poorer cardiac function on echocardiogram. Further study is needed to determine the mechanistic cause of neurocardiogenic injury and its contribution to neurologic outcomes.

Grant Acknowledgement

National Institutes of Health (NHLBI R01HL074316).

0444

Predictors of Brain Injury in Critically Ill Patients

R. D. Stevens*1, A. Pustavoitau2, V. Garnelo2, A. Zaky2, M. Johnson3, B. Crain3

1Department of Anesthesiology Critical Care Medicine, Neurology, Neurosurgery, 2Department of Anesthesiology Critical Care Medicine, 3Pathology, Johns Hopkins University School of Medicine, Baltimore, United States

Introduction

Critically ill patients are at high risk of developing brain dysfunction which may present clinically as delirium, coma, and long-term cognitive impairment (1). The mechanisms underlying this brain dysfunction are unknown. In order to evaluate the hypothesis that brain dysfunction is associated with histological markers of brain injury, we evvaluated the prevalence and predictors of histologically defined acute brain injury in patients who died of critical illness.

Methods

We conducted a case-control study of patients admitted to 2 medical and 3 surgical ICUs during the period 1996–2006 who died and underwent autopsy. Patients with prior neurological disease were excluded. Brain histology was reviewed for evidence of ischemic, hemorrhagic, or inflammatory injury. Clinical and pathological characteristics were compared between cases (histological injury present) and controls (injury absent) in a univariable and multivariable analysis.

Results

We evaluated histological and clinical data in 214 patients admitted to the ICU for respiratory failure (41% of patients), circulatory failure (25%), postoperatively (12%), and and other (12%). Histological evidence of acute brain injury was detected in 99 patients (46%). Lesions were most common in the cerebral cortex (38% of patients), followed by hippocampus (30%), cerebellum (20%), basal ganglia (15%), and brainstem (12%). Lesions were ischemic in 32% of patients and hemorrhagic in 25%, and inflammtory in 12%. Multivariable predictors of brain injury were female gender (OR 1.9, 95% CI 1.3–3), ARDS (OR 2.1, 95% CI 1.5–4.6), and septic shock (OR 1.4, 95% CI 1.1–4.8).

Conclusion

Histological evidence of brain injury was found in nearly half of critically ill patients without neurological antecedents who died in the ICU. An association was noted between brain injury and ARDS and septic shock. These results suggest a relationship between systemic inflammatory conditions and acute brain injury.

0445

Value of Serum Levels of Brain Injury Biomarkers in the Prognosis and Evolution of Acute Spontaneous Cerebral Haemorraghes and Their Relationship with Severity Scales

Serrano*, M. Juan, A. Mesejo, S. Martin, M. Blasco, S. Borrás, J. Pineda, M. Garcia, M. Argüeso

Intensive Care Unit, Hospital Clínico Universitario, Valencia, Spain

Introduction

Reliable biomarkers are being studied to predict the prognosis of Spontaneous Cerebral Haemorraghes (CH) like matrix metalloproteinses (MMP-9), D-dimer (DD), B-type natriuretic peptide (BNP) and S-100b protein. We aimed to: Evaluate the utility of systematic measurement of these brain injury biomarkers (BIM) to predict the prognosis of acute CH.

Methods

Prospective observational cohort study. Adult patients admitted in ICU with CH were included. Serum levels of BIM were measured at admission and sequentially (days 1,2,3,5,7 and 10) using the “ Triage Stroke Panel” of Biosite®. Severity scales (SS): APACHE II, SOFA and Glasgow were registred. Statistical analisis with Paired-Samples TTest (BIM serum levels), Independent-Samples T-Test (BIM and SS-mortality) and Bivariate Pearson Correlation (BIM-SS).

Results
Preliminary communication with the first 50 patients. Age 56±1,8 years; 66% men; APACHE II at admission 11±1. ICU stay 7,8±1 days and 5,4±1 days of mechanical ventilation (DMV). Mortality 20%. There was significant correlation between mortality and SS, specifically APACHE II (p 0,02), SOFA (p 0,001), Glasgow (p 0,001). Serum levels of MMP-9, BNP and DD were high at admission and significantly raised during evolution; pick level of MMP-9 was reached earlier than BNP and DD. Nevertheless no significant correlation was found between mortality and BIM.
Table 1

Evolution of BIM Serum Levels

 

Mean±Std.Error

p

BNP Day 0–5

154±41–159±36

0,02

BNP Day 0–7

154±41–177±34

0,02

MMP-9 Day 0–1

354±54–251±40

0,04

MMP-9 Day 0–3

354±54–201±38

0,04

MMP-9 Day 0–5

354±54–145±20

0,00

DD Day 0–3

1357±207–2295±279

0,00

DD Day 0–5

1357±207–2547±301

0,00

DD Day 0–7

1357±207–3715±342

0,00

Conclusion

The evolution of BNP, MMP9 and DD serum levels may reflect an estructural brain injury with a clear temporal profile. No significant correlation was found between mortality and BIM.

0446

Metabolic Characterization of Cerebral Contusions Following Head Injury Identification of A Transitional Penumbral Zone

M. G. Abate*1, V. F. J. Newcombe1, D. A. Chatfield1, J. G. Outtrim1, S. G. Harding2, J. P. Coles1, P. Hutchinson3, T. D. Fryer2, A. T. Carpenter2, D. K. Menon1

1Division of Anaesthesia, Wolfson Brain Imaging Centre, 2Wolfson Brain Imaging Centre, 3Department of Neurosurgery, University of Cambridge, Cambridge, United Kingdom

Introduction

Pericontusional tissue represents tissue at risk in traumatic brain injury (TBI), and a target for therapy. Diffusion tensor MR imaging (DTI) maps the apparent diffusion coefficient (ADC) that characterises tissue water. Higher ADC values are associated with vasogenic edema, and lower ADC with cytotoxic oedema. DTI shows regional differences in contusions, with a central core and a surrounding region of vasogenic oedema. Many contusions also demonstrate a region of cytotoxic oedema, peripheral to the vasogenic oedema. We have used oxygen-15 positron emission tomography (15O PET) to characterise physiology in these brain regions in TBI.

Methods

Six patients (median GCS 5) who required sedation and mechanical ventilation for ICP control underwent DTI at 3Tesla, and 15O PET imaging studies within 24 hours post injury. Four regions of interest were manually drawn around the contusion in order to quantify cerebral physiology respectively in the core of the lesion (low ADC core) in the surrounding vasogenic oedema (high ADC region), in the adjacent rim of tissue with low ADC (cytotoxic edema) and on a control region (structurally normal appearing tissue). ADC maps were coregistered to PET images of cerebral blood flow (CBF), oxygen extraction fraction (OEF), and cerebral metabolic rate of oxygen (CMRO2) in the four ROIs.

Results

The ADC showed statistically significant differences between all of the groups, confirming the validity of the tissue compartmentation (p<0.05 for all comparisons). Median CBF and CMRO2 were significantly lower (p<0.05) in both the core (15 ml/100g/min and 18 µmol/100g/min) and vasogenic oedema (22 ml/100g/min and 34 µmol/100g/min) than control regions (27 ml/100g/min and 67 µmol/100g/min). The rim of cytotoxic oedema showed intermediate CBF and CMRO2 values (25 ml/100g/min and 43 µmol/100g/min), with CMRO2 values that approached previous defined survival thresholds.1 OEF was not abnormally elevated in any of the regions.

Conclusion

These distinct patterns of physiology in contusions recapitulate a probable evolution of pathophysiology, with incorporation of the cytotoxic rim into the vasogenic oedema and, eventually into the contusion core. This mechanism may underlie the process of contusion expansion in TBI. The coexistence of low/normal OEF with cytotoxic oedema suggests that pericontusional ischemia may be microvascular in origin.>

0447

Arterial Blood Pressure Monitoring in Acute Stroke Patients: Invasive or Noninvasive?

R. H. Passos*1, J. M. Teles2, J. OliveiraFilho3

1Critical Care and Neurocriticalcare, Hospital Salvador and Espanhol, 2Critical Care, Hospital salvador, 3NeuroCriticalCare, Hospital Espanhol, Salvador, Brazil

Introduction

The optimal management of blood pressure in acute stroke remains poorly defined. Some studies note an association of poor outcome with patients with high BP on hospital admission. Others have noted a decreased risk of neurologic deterioration from stroke with higher BP and worse outcomes in patients who have a decrease in BP after admission. Therefore accurate measurement of blood pressure is crucial in the acute phase of stroke. Besides, the institution of suscessfuly treatments of blood pressure reduction or augmentation depends on it. Either non invasive or intraarterial blood pressure monitoring is commonly used in ICU. In different populations have been proved that non invasive method does not achieve right accuracy. In acute stroke population discrepancies between direct intraarterial blood pressure (IABP) and indirect blood pressure measurements (NIBP) can adversely affect therapeutic decisions and may have a negative impact on outcomes. The aim of this observational study was to test the hypotheses that IABP measurements are not accurately reflected by NIBP in a population of acute stroke patients.

Methods

A total of 52 first ever stroke patients underwent simultaneous noninvasive automatic oscillometric and intra arterial blood pressure monitoring. Each patient's height, weight, and arm circumferences at the mid-arm level were recorded. Intraarterial and oscillometric blood pressure measurements were obtained at least every 3 mins until 10 measurements from the arm into which the arterial catheter was inserted. Agreement between methods was assessed using Bland Altman analysis.

Results

Overall 520 pairs of simultaneous oscillometric and invasive blood pressure measurements were collected in 52 patients. The mean age was 57 +/−3 years. The mean NIHSS was 15 +/− 6. Thirty-five percent of the acute stroke patients were on intravenous anti hypertensive treatment. No patient was on vasopressor therapy. Mean arterial blood pressure ranged from 52 to 165 mmHg. The overall discrepancy between methods of measurement to mean blood pressure was 7,8 +/− 9,6 mmHg (p<0,001). The estimation of systolic blood pressure (SBP) using oscillometric recordings underestimated radial artery SBP by 12,7 mmHg (95% confidence interval :7,2–15, p<0,001 In contrast an upward bias of 6,7 mmHg (95% confidence interval: 4,5 +/− 7,2, p<0,001) was documented when noninvasive diastolic blood pressure (DBP) recordings were compared with intra arterial recordings. For SBP and DBP, the Pearson correlation coefficients between noninvasive and intraarterial recordings were 0,812 and 0,792 respectively.

Conclusion

The noninvasive automatic oscillometric blood pressure measurements underestimated mean arterial blood pressure. The SBP is underestimated and the DBP is overestimated by noninvasive automatic device. Therefore oscillometric blood pressure measurement does not achieve adequate accuracy in our acute stroke population. Physicians must consider it before institution of treatments of blood pressure reduction or augmentation.

0448

Haemodynamic and Metabolic States Predict Morbidity After the Occurrence of Subarachnoid Haemorrhage

F. Franchi*, S. Scolletta, P. Mongelli, M. Fiorillo, G. Salto, E. Casadei, P. Giomarelli

Anestesia e Rianimazione, University of Siena, Siena, Italy

Introduction

The incidence of subarachnoid haemorrhage (SAH) accounts for about 3% of all strokes with for 30-day mortality of approximately 40%. Initial treatment involves tabilization of vital signs and prevention or treatment of sequelae of SAH (rebleeding, hydrocephalus, seizures, vasospasm). Patients with SAH are often haemodynamically unstable with fluctuations of cardiac index (CI), blood pressure and arrhythmias being potential problems. Our aim was to identify the variables that could potentially be predictors of morbidity in patients with SAH.

Methods

Data of 140 adult patients with SAH (age mean 63,5±10) admitted at our Intensive Care Unit (ICU) were prospectively collected from June 2005 to December 2007. 70 hemodynamic, ventilatory, and metabolic parameters were evaluated within 3 hours after ICU admission. The CI and others haemodynamic derived parameters were obtained by PRAM (Pressure Recording Analytical Method). Accordingly to the GIVITI (Gruppo Italiano per la Valutazione degli Interventi in Terapia Intensiva — Italian Group for the Evaluation of Interventions in ICU) database definitions, complications were defined as one or more organ dysfunctions or failures occurred during the ICU stay. Statistically, univariate and multivariate analyses, and receiver operating characteristic (ROC) curve analysis were applied.

Results
Organ failure resulted in 49 patients (35%). 10 variables gained a significant level at the univariate analysis. Among these, a high blood lactate concentration (Lac), a high CO2 production index (VCO2i), a low central venous saturation (ScVO2), and a low CI achieved the statistical significance at the multivariate analysis (see Table). The ROC analysis showed a cut-off value of 2,2 mmol/l for Lac, 68% for ScVO2, 2,0 l/min/m2 for CI, and 113 ml/min/ m2 for VCOi. Sensibility vs specificity was 78% vs 69% for Lac, 75% vs 67% for ScvO2, 82% vs 77% for CI, and 72% vs 67% for VCO2i. The area under the ROC curve (AUC) resulted 0,8 for Lac, 0,78 for ScVO2, 0,79 for CI and 0,76 for VCO2i. The different ICU severity risk scores did not reach the statistical significance. The average of ICU stay was 21 ± 6 and 10,2 ± 3,4 days (complicated vs uncomplicated patients, p<0,001). Mortality was higher in patients with complications (25% vs 6%, p<0,001).
 

OR

CI 95%

p

ScVO2

1,2

1,1–1,4

0,05

Lac

3,2

1,3–8,2

0,01

VCO2i

1,1

1–1,2

0,05

CI

1,3

1,2–1,6

0,01

Conclusion

Our findings demonstrated that various ICU severity risk scores seemed not to predict the morbidity in patients with SAH. Conversely, the indexes of haemodynamic and metabolic states (CI, ScVO2, Lac, and VCO2i), early evaluated after ICU admission, resulted fine predictors of complications. Patients admitted at ICU after an occurrence of SAH should be accurately hemodynamically and metabolically monitored in order to improve their clinical management and outcome.

0449

Accuracy of Protein S-100B as a Marker of Brain Damage in Nonsevere Head Trauma

E. J. J. Egea-Guerrero*, J. Dominguez-Roldan, B. Pardo, A. Cayuela, M. Munoz-Sanchez, F. Murillo-Cabezas

Intensive Care Unit, H.H.U.U. Virgen del Rocío, Seville, Spain

Introduction

The protein S-100B is a brain-specific protein release from astroglial cells into the circulation after Traumatic Brain Injury. Research indicates that the S-100B serum level could be a useful indicator of Head Trauma severity, but there is not evidence enough about the role of S-100B after nonsevere Head Trauma (NSHT). The hypothesis that S-100B is a useful screening tool for the management of minor head injuries was tested. METHODS. Forty five patients with NSHT without decrease of conciousness but associated syumptons/signs, like amnesia, headache, dizziness, convulsion and vomits, were prospectively included. We recorded the clinical data on admission and a blood sample before tree hours after NHST, for S-100B inmunoluminiscence analysis. A routine cranial computed tomography scan (CT) was obtained within 24 hours after the injury. The diagnositc properties of S-100B serum levels<0.105 µg /L, for prediction of intracranial lesions revealed by CT were tested with receiver operating characteristic (ROC) analysis.

Results

S-100B levels were<0.105 µg /L, in eight of forty five patients (18%) without intracranial lesions. A Sensitivity of 100%, CI, 61–100%; a specifity of 21,6%, CI, 11,4%- 72,2%; a Negative Predictive Value of 100%, CI, 67.6%–100% was calculated for this S-100B value. ROC curve and cross-table analysis showed that above cut-off values of S-100B, improve specifity, without changes on sensitivity. In this study the S-100B levels were>0.105 µg /L, in all the patients with brain damage, six of forty five (12.5%).

Conclusion

Determination of serum protein S-100B seems to be a useful biochemical indicator of brain damage, in spite of the low specificity for the detection of intracranial lesions. These results, shows that cut-off value of S-100B<0.105 µg /L, with high negative predictive value, appear to be of substantial clinical relevance for the management of NSHT.

0450

Complications After Spontaneous Intracerebral Hemorrhage in the Intensive Care Unit

L. Marques Alvarez*1, L. Perez Belmonte2, D. Escudero Augusto1, L. Vinña Soria1, L. Forcelledo1, E. O Campos1, F. Iglesias Llaca1, R. Lopez Roger3, E. Fernandez Bustillo4, F. Taboada Costa1

1ICM, 2Ophtalmology, 3Radiology, 4Statistical Unit, Hospital Universitario Central de Asturias, Oviedo, Spain

Introduction

Spontaneous Intracerebral Hemorrhage (SIH) has the greatest mortality among strokes. During the stay in the intensive care unit (ICU) these patients develop some complications that often modify the evolution of the disease. The aim of our study was to determine these complications developed by the patients with SIH during their stay in ICU and its relation with mortality.

Methods

Prospective cohort study. Sixty six adults admitted to the ICU with SIH were studied during one year. SIH less than 24 hours,traumatic and subarachnoid haemorrhage, hemorrhagic transformation in acute ischemic stroke, coagulopathy and patients with preceding use of antiplatelet or anticoagulants agents were excluded. On admission we recorded: sex, age, high blood pressure (HBP), Apache II, Glasgow (GCS), location, volume (ABC/2). Complications during the stay in ICU were classified as neurologic, hemodynamic, infectious and metabolic. The average stay and mortality were determined. We studied the complications that influence more in mortality.

Results

Forty three patients were men. The mean age was 58±13 (35–79) years. Fifty four percent were hypertense. Mean value of Apache II was 16±6 and GCS 7.4. The most common location were basal ganglia (41%) and lobar (41%). The average hematoma volume on CT was 48±25 (8–140)mL. Fifty three percent exceeded 30 mL. The average stay in ICU was 15.4±15(2–78) days. ICU mortality was 44%. Most patients died within the first week. Encephalic death was the most common cause (90%) followed by infectious complications (10%). The most frequent complications were metabolic (30%), neurologic (27%), hemodynamic (26%) and infectious (17%). In the group of metabolic complications, high glucose levels (44%) was most common complication although only hyponatremia was associated with mortality in the ICU (p=0.01). In the group of neurologic complications, brain herniation (59%)in any of its variants was the most frequent and was statistically related to mortality (p=0.03) followed by edema (45.5%), ventricular hemorrhage (41%), rebleeding (35%), hydrocephalus (29%), ischemic (29%) and seizures (12%). The most frequent hemodynamic complication was HBP (53%). The supraventricular tachycardia and auricular fibrillation were the most common arrhythmias. The most frequent infectious complications were traqueobronchitis (36%), pneumonia (21%), urinary tract infection (18%), catheter-related infection (12%) and ventriculitis (1.2%). Only pneumonia was related to mortality (p < 0.001).

Conclusion

Mortality of patients with SIH in the ICU was high and usually happened during the first week. The main cause of mortality was brain death. Most frequent complications were metabolics. Neurologic complications, such as brain herniation, were related to higher mortality. Pneumonia was the infectious complication more frequently associated with mortality.

0451

Prognostic Factors for Neurological Outcome in Patients with Acute Ischemic Stroke Treated with Intravenous Recombinant Tissue Plasminogen Activator

M. Martinez*1, D. Canovas2, A. Carvajal3, J. Estela2, A. Villagra1, J. Perendreu4

1critical Care Centre, 2neurology, Hospital Of Sabadell, 3neuroradiology, 4interventional Radiology, Udiat, Sabadell, Spain

Introduction

IV administration of recombinant tissue plasminogen activator (rt-PA) within 3 hours of symptoms onset is the only approved therapy for patients with acute ischemic stroke. The identification of predictors of neurological improvement may help to improve patient selection. The aim of this study was to analyse prognostic factors of neurological outcome in our acute stroke patients treated with rt-PA.

Methods

Observational prospective analysis of all patients with acute ischemic stroke admitted in our ICU and treated with rt-PA (2004–2007). We analysed epidemiological data, stroke severity, CT scan findings, early clinical improvement, complications and neurological outcome at 3 months using a modified Rankin Scale (mRS). Results are expressed as mean (SD) or percentage, using a p<0.05 as significance level. We applied SPSS with the t test for continuous variables and Chi-square for categorical ones, and logistic regression tests, calculating odds ratio with their confidence interval for the significant variables.

Results

We have treated 64 patients (64% Male), age 66 yrs. The baseline NIHSS was 14 ± 5. History of diabetes was present in 13 patients (20%). Early signs of infarct on CT were found in 18 p (28%). We observed an early neurological improvement (>4 points compared to baseline in NIHSS at 24 hours) in 30 patients (47%). The rate of asymptomatic and symptomatic haemorrhages at 36 hours was 14% and 3.1%. The mRS at 3 months was good (0-2) in 34 p. (53.1%). Independent factors related to neurological outcome in multivariate analysis was an early clinical improvement (OR 151, CI 8-2552), while age and baseline NIHSS were associated with poor outcome.

Variables related with neurological outcome are shown in Table 1.
Table 1

Variables Related with The Neurological Outcome

 

Good outcome

Poor outcome

p

Diabetes

18%

82%

0.009

Age > 70 yrs

38,5%

61.5%

0.035

Early signs on CT

11.5%

88.5%

0.012

Baseline NIHSS

12

15

0.019

Improvement >4 points NIHSS at 24 hours

88.5%

11.5%

<0.001

Conclusion

In our series of acute stroke patients treated with rt-PA a good neurological outcome at 3 months was achieved in 53%. We found that age and stroke severity (baseline NIHSS and early signs of infarction on CT) were related with poor outcome, while an early neurological improvement was strongly associated with a good neurological recovery, probably related to an early recanalization.

Grant Acknowledgement

CIBER-enfermedades respiratorias.

0452

Clinical Characteristics and Outcome of Patients with Venous Cerebral Thrombosis (Cvt) Hospitalized in Icu

M. Rusca*1, J. Guichard2, I. Crassard3, D. Bresson4, J. Mateo1, A. Lukaszewicz1, D. Payen1

1Anesthesiology and surgical critical care, 2Neuroradiology, 3Neurology, 4Neurosurgery, Hôpital Lariboisiere, Paris, France

Introduction

CVT has a better prognosis than cerebral arterial thrombosis and is rarely hospitalized in ICU. In 624 patients, (Ferro, Stroke 2004; 35; 664–670) only 7 patients (1.1%) required mechanical ventilation. No information on clinic, outcome and strategy of care of these patients have been published.

Methods

We report our experience of care of patients with CVT admitted in ICU from January 2003 until December 2007.

Results

Twenty three cases of CVT (age: mean 43 yrs ± 15, ranging from 16 to 68yrs) were hospitalized in ICU in relation with: coma in 48%, status epilepticus in 22%; 13% in post-op. The median delay from the onset of symptoms to the CVT diagnosis was 5.63 days ± 5 SD days and from the diagnosis to ICU admission was 3.96 days ± 3SD days. Diagnosis of CVT was established by MRI/MR venography in 8 patients (35%), intra-arterial angiography in 3 (13%), CT venography in 12 (52%) Monitoring: intra-arterial catheter (20 patients, 87%), transcranial Doppler (14, 61%), capnography (13, 56%), continuous jugular oxygen saturation (4, 17%). Treatment: all patients received heparin at therapeutic dose; 18 patients (78%) were intubated and mechanically ventilated; 14 (61%) received antiepileptic drugs ventilation. Norepinephrine was used in 10 patients (43.5%) to maintain cerebral perfusion pressure. Within the first 24 hours in ICU, fluid balance was negative in 9 patients (39%) because of brain edema and in 10 patients at day 5 (43.5%). Intra-cranial hematoma was drained in 1 patient, another had a brain abscess drainage, 3 patients had a decompressive craniectomy (2 died). Outcome : 5 patients (22%) died in ICU, four in relation with refractory intracranial hypertension, and 1 for lethal pulmonary embolism. Comparison between survival and dead patients (non parametric tests and univariate analysis): except for sex (more death on female than in male group (p < 0.05), no difference were observed for medical context, chronic therapy; blood pressure, central venous pressure, heart rate, trans-cranial Doppler data and biological routine tests, except for platelet level at entry (240.103± 108 in survival vs 160.103±37 in dead patients; p< 0.0033) and initial glycemia (6.41± 1.4 in survival vs 14.24± 13.36; p<0.06). Localisation of thrombosis or intracranial complications did not differ between the 2 groups. The mean modified Rankin score for alive patients was 2,6.

Conclusion

Only 22% patients with CVT died in ICU under heparin treatment. Other instrumental therapeutic options did not seem to improve poor outcome, which is associated with low platelet level and higher glycemia.

0453

The Acute In-Hospital Mortality for Intracerebral Hemorrhage Patients with a Dedicated Neurocritical Care Team

T. Janjua, MD*, H. P. Lochungvu, P. Sinner, MPH1, A. Defillo,MD

Neurocritical Care Medicine, HealthEast Neurovascular Institute, St. Paul, United States

Introduction

Dedicated Neurocritical care units increase survival chances for neurologically unstable patients. We were interested in the effect of the Neurocritical care unit in a population with intracerebral hemorrhage (ICH) outcome. This study was done to review the mortality and treatment protocols in ICH. The national benchmark is 37–38% death rate within 30 days of admission.

Methods

The charts were retrospectively reviewed for the patients who were discharged from the Neurocritical ICU between January 1, 2006 and December 31, 2006. In-hospital mortality was assessed and information regarding gender, age, and length of stay was collected. Statistical analyses using chi-square tests and wilcoxon rank sum tests were performed using SPSS 15.0.

Results

There were 168 patients that were treated for ICH during this time period. Patients ranged in age from 22 to 94 years (mean = 65.2 years) and 58% were female. Eighteen patients (10.7%) died during this time period. Patients that died were significantly older than patients that survived (73.9 vs. 64.1, p = 0.02). The range of time to death was 1 to 21 days with an average of 5.3 days. Six of the 18 patients died within 1 day and an additional 7 died within 4 days. 3 patients survived longer than 10 days with one patient surviving 21 days. The cause of death in the later 3 patients was not directly related to the ICH diagnosis. The mortality prior to starting the Neurocritical care program was 26%.

Conclusion

The mortality rate with the addition of a dedicated neurocritical care team in our hospital was lower than the nationally reported range. Also this improvement was noted from our pervious critical care model. Based upon this study, we follow strict protocols for all intracranial hemorrhage patients.

0454

Brain Ct: Confirmation of Clinical Expectations and Consequences for Treatment

I. M. Purmer*1, E. Iperen, van1, L. F. M. Beenen2, J. M. Binnekade1, M. J. Schultz1, J. Horn1

1Intensive Care, 2Radiology, Academical Medical Center, Amsterdam, Netherlands

Introduction

Brain CT (BCT) scans are an important diagnostic tool for intracranial disorders. We analysed the ICU BCTs with regard to the expected findings, final results and consequences on treatment.

Methods

Prospective observational study, in a mixed medical and (neuro) surgical intensive care of a university hospital. All patients who went for BCT, between May 2007 and January 2008 were included. The physician requesting the BCT filled out a special radiological request form, in which the expectations for the results of this BCT and the consequences given these findings for treatment, were addressed. The BCT was evaluated by a radiologist. An “expectation” was radiologically confirmed if the finding was new or increased and negative if unchanged, decreased or not found. Treatment strategy after BCT was registered (unchanged, changed according to plan or otherwise).

Results
189 clinical questions in 117 BCT's were evaluated. The most important expected findings were shift, edema, ischemia/infarction, hematoma and hydrocefalus. In 20% these expectations were radiologically confirmed. After the BCT, the physician changed the treatment strategy in 53%. This was based on positive radiological confirmation in only 24%. Strikingly, in 76% the treatment was changed although no radiological confirmation of an expected finding was obtained. The chance of a therapy change after a positive CT scan finding is 1.5 times greater than an unchanged policy after a positive CT scan. However, non of the predefined expected abnormalities showed a significant change to be followed by a change in policy. The confidence limits of the crude overall risk tells us that the chance that the policy will be adjusted due to the CT scan can be as small as 1.04 (almost non) and as large as 2.4 (almost 2.5 times).
Table 1

Expected Abnormality

 

N

Ctscan Pos

Neg

Policy Pos

Neg

RR (95%CL)

Shift

16

1

15

11

5

undefined

Edema

30

7

23

18

12

4(0.6–29.2)

Ischaemia infarction

43

13

30

17

26

1.8 (0.7–4.4)

Haematoma

74

10

64

38

36

2.2 (0.6–7.9)

Hydrocephalus

26

7

19

16

10

0.5 (0.1–1.7)

Total clinical questions

189

38

151

100

89

1.5(1.04/2.4)*

* Crude RR over all strata (Mantel-Haenszel weighted RR):Chi2 4.24; p 0.04

Conclusion

Pre-emptive expectations of findings in a BCT are only radiologically rconfirmed in 20%. After evaluation of the BCT there is little correlation between the findings on the BCT and the following treatment, leading to more changes in treatment at negative findings on BCT.

0455

The Effects of Therapeutic Decompressive Craniectomy on Long-Term Outcome

S. Matano*, M. Bonizzoli, I. Guerri, V. Anichini, G. Cianchi, M. Ciapetti, A. Peris

Department of Emergency, Azienda Ospedaliera Universitaria Careggi, Florence, Italy

Introduction

High intracranial pressure (HIP) is still the most frequent cause of death and disability after severe head injury. The aim of the study is to investigate the prognostic impact of therapeutic decompressive craniectomy (TDC) in terms of overall survival and functional outcome in patients with HIP after traumatic brain injury.

Methods

Between 2005 and 2007 forty head trauma patients underwenting TDC (Group 1) were compared with an equal number of patients not treated with TDC during the years 2002–2004 (Group 2). Mean age was 40 years (range 15–65), 35 (43%) were woman and 45 (56.2%) men. The mean Glasgow Coma Scale [GCS] at ICU admission was 10 (range 7–13) in Group 1 and 7 (range 5–9) in Group 2. Injury Severity Score mean was 35 (range 30– 40) in Group 1 and 45 (range 43–47) in Group 2. SAPSII mean was 40 (range 35–45). Intracranial pressure was analysed with intraparenchimal catheter, and was considered like HIP without possibility of treatment when PIC was more than 30 mmH2O for more than 20 minutes during maximal medical therapies. In the two groups ICU mortality, Glasgow Outcome Scale (GOS) and GCS after 28 days and after 1 year was valuated.

Results
Mortality rate at 24 hrs was 12.5% in Group 1 and 20% in Group 2, mortality in ICU was 15% in Group 1 and 40% in Group 2, at 28 days 12.5% in Group 1 and 17.5% in Group 2, at 1 year 2.5% in Group 1 and 5% in Group 2. The GCS mean at 28 days was 12 (range 10–14) in Group 1 and 8 (range 6–10) in Group 2. At 1 year in Group 1 GOS was 5 in 10 patients (41.6%), 4 in 7 (16.6%), 3 in 5 (29.1%) and 2 in 2 patients (8.3%); in Group 2 GOS was 5 in 2 patients (22.2%), 4 in 1 (11,1%), 3 in 2 (22.2%), 2 in 4 (44.4%).
 

Group 1

Group 2

Mortality Rate at 24 hours

12.5

20

Mortality in ICU

15

20

Mortality at 28 days

12.5

17.5

Mortality at 1 year

2.5

5

Mortality Rate in two Groups

 

Group 1

Group

GCS at 28 days

12

8

GCS at 1 year

14

9

GOS at 1 year

4

3

GCS and GOS mean for two Groups

Conclusion

TDC treated Group seems to have a significantly better outcome then TDC-untreated head trauma Group in terms of functional outcome and overall survival.

0456

Timely Efficacy of Decompressive Craniectomy in the Management of Severe Traumatic Brain Injury

E. Anastasiou1, E. Geka*1, M. Giannakou1, E. Efthimiou1, A. Spiliotopoulos2, F. Drampa1, K. Matsi1, I. Papageorgiou1, E. Sofianos1

1Department of Anesthesiology and ICU, 2Neurosurgical Department, AHEPA Hospital, Thessaloniki, Greece

Introduction

Severe traumatic brain injury (TBI) is associated with high mortality and morbidity. Decompressive craniectomy (DC) has recently regained therapeutic interest; however, treatment guidelines consider it a last-tier treatment after failure of conservative therapy.

Methods

Retrospective chart review analysis was conducted to determine the timely efficacy of DC for the management of severe TBI. Out of 250 patients with severe TBI admitted to the ICU between the years 2000–2006, 30 underwent DC. The following parameters were recorded: age, Glasgow Coma Scale (GCS, accident site), Injury Severity Score (ISS), CT-Scan findings (Marshall scale), APACHE II and SOFA scores upon admission. Follow-up period was of at least 1 year after ICU discharge, the outcome being evaluated with the Glasgow Outcome Score (GOS). The patients were divided in 2 groups: 1) group A (n=17), when early DC within hours and 2) group B (n=13) when delayed DC (within days) after TBI was performed. The decision for early DC was based on clinical parameters, CT-scan and operative findings. The patients were treated with delayed DC when maximal treatment failed. Mann-Whitney was used for the statistical analysis between the groups. All values are expressed as mean (±SD).

Results
Patients that underwent DC comprised 12% of severe TBI admissions (n=250). Early DC was conducted within 3.6 (±1.9) hrs post TBI and delayed DC within 6.6 (±4.5) days. The groups presented no statistical difference regarding: age 41.4 (±20.7) vs 30.8 (±14.9) years, GCS 7.6 (±3.7) vs 7.3 (±3.7), ISS 28.4 (±8.3) vs 30.5 (±13.7), APACHE II 14.6 (±7.31) vs 14.9 (±5.5), SOFA 4.4 (±3.2) vs 5.5 (±2.9) and ICU stay 21.2 (±13.7) vs 23.6 (±17.8) days. CT-scan scale between groups was not quite statistically significant, 3.9 (±0.3) vs 3.4 (±0.7) (p=0,0945). Outcome upon ICU discharge and 1year later is shown in table.
Table 1

Outcome Upon ICU Discharge and 1Year Later

 

GOS 4–5 (ICU) n (%)

GOS 4–5 (1yr) n (%)

GOS 2–3 (ICU) n (%)

GOS 2–3 (1yr) n (%)

GOS 1 (ICU) n (%)

GOS 1 (1yr) n (%)

GroupA

8(47)

9(52.9)

7(41.2)

4(23.5)

2(11.7)

4(23.5)

n=17

      

GroupB

8(61.5)

8(61.5)

2(15.4)

2(15.4)

3(23)

3(23)

n=13

      
Conclusion

Decompressive craniectomy may be considered an efficient method not only as a ‘second tier’ therapy in patients suffering of severe TBI with surgically removable lesions.

Poster Sessions Sepsis: Experimental: 0457-0470

0457

Relationship Between Mitochondrial Enzyme Activity and Protein Expression in Critically Ill Patients

J. E. Carré*1, J. Orban2, M. Stotz1, P. Breen1, G. Bellingan1, M. Singer1

1Bloomsbury Institute for ICM, University College, London, United Kingdom, 2Département d'Anesthésie-Réanimation Est, Hôpital Saint-Roch, Nice, France

Introduction

Decreased functional capacity of mitochondrial complex I and compromised cellular energetic status are associated with non-survival in septic patients [1]. Mitochondrial enzyme activity largely reflects the level of expressed protein (balance between biogenesis and degradation), but also persisting modifications e.g. from oxidative/nitrosative damage. In an ongoing study in critically ill patients, we investigated respiratory complex activities and corresponding protein expression in skeletal muscle.

Methods

With ethics approval and appropriate consents, critically ill patients were recruited within 24h of ICU admission. Age-matched control patients were undergoing elective hip surgery. Muscle biopsies were taken from vastus lateralis. Mitochondrial enzyme activities were determined as previously [1]. Acid-precipitated protein extracts were neutralised and used for immunoblotting. Blots were semi-quantified by densitometry and normalised to the signal obtained with a standard sample. Data were analysed for significance using one-way ANOVA.

Results
Both mitochondrial respiratory enzyme activity and protein expression were decreased in critically ill patients. Expression of the mitochondrial oxidative stress protein MnSOD was increased.
Table 1

Protein Expression and Enzyme Activities in Patient Skeletal Muscle Biopsies

 

Control (7)

Survivors (6)

Non-survivors (4)

Protein expression:

 

Relative Density

 

Complex I

1.11 (0.68, 1.34)

0.33 (0.32, 0.34) †

0.33 (0.20, 0.50) †

Complex II

0.78 (0.60, 0.86)

0.50 (0.31, 0.61) †

0.38 (0.27, 0.49) †

Complex III

0.57 (0.31, 0.90)

0.15 (0.10, 0,27) †

0.09 (0.07, 0.13) †

Complex IV

0.73 (0.67, 0.92)

0.36 (0.26, 0.47) †

0.29 (0.20, 0.44) ‡

MnSOD

0.83 (0.70, 0.88)

0.98 (0.95, 1.03) †

0.91 (0.88, 0.92)

Enzyme activity:

 

nmol/min/mg protein

 

Complex I

13.3 (10.2, 15.5)

8.7 (4.5, 11.9)

10.9 (7.6, 31.1)

Complex IV

39.9 (23.4, 66.62)

49.0 (37.1, 54.8)

7.2 (5.8, 12.4)

Citrate synthase

106.2 (63.6, 117.7)

48.4 (39.7, 72.0)

40.4 (30.6, 50.5)

Median (IQ ranges); †p≤0.05; %p≤0.01 vs controls

Conclusion

The reduction in mitochondrial enzyme activity described in critical illness [1] can be partly explained by decreased levels of expressed protein. Elevated MnSOD expression indicates a response to oxidative stress. Together, the data suggest increased degradation of oxidatively-damaged proteins. We previously demonstrated preserved ATP content and transcript levels for markers of mitochondrial biogenesis in patients who eventually survived, but depressed levels in eventual non-survivors [2]. A decreased ability to replace damaged mitochondrial proteins may contribute to ATP depletion and mortality.

Grant Acknowledgement

MRC.

0458

Differential Regulation of Neutrophil Survival and Inflammatory Function by Endogenous Serine Proteases

J. L. Y. Tsang*, J. C. Parodo, J. C. Marshall

Critical Care Medicine, Saint Michael's Hospital, Toronto, Canada

Introduction

Activated neutrophils (PMNs) have been implicated in the pathogenesis of sepsis and ARDS. PMN serine proteases such as elastase are believed to contribute to tissue injury. Since increased PMN inflammatory activity is characteristically associated with prolonged survival through inhibition of constitutive PMN apoptosis, we sought to define the role of PMN serine proteases in PMN survival and inflammatory function.

Methods

We incubated PMN from healthy volunteers for 5 hours (2 hours for oxidative burst) with the serine protease inhibitor, diisopropylfluorophosphate (DFP, 2.5 mM). PMN apoptosis was quantified by flow cytometry as the nuclear uptake of propidium iodide by permeabilized cells (n=5). Caspase-3 (n=4), -8 (n=3) & elastase (n=5) activities were measured using specific substrate with a fluorescent (caspase-3 & elastase) or colorimetric (caspase-8) plate reader. Western blot was performed to characterize the processing of caspase-3 and caspase-8 (n=3). Oxidative burst was measured as the conversion of dihydrorhodamine 123 by flow cytometry (n=4).

Results
DFP almost completely ablated elastase activity. PMN apoptosis, and activity of caspases-3 were significantly inhibited by DFP. Caspase-8 activity was slightly inhibited by DFP (Table). The 12kDa form of active caspase-3 was reduced, and pro-caspase-8 expression significantly increased in DFP treated samples (Figure). DFP also inhibited PMN oxidative burst capacity (Table).

Samples

Apoptosis (n=5)

Oxidative Burst (n=4)

Caspase 3 Activity (n=4)

Caspase 8 Activity (n=3)

Elastase Activity (n=5)

Control

100

100

100

100

100

DFP

18.94

13.85

13.23

78.13

19.08

P-Value

<0.05

< 0.05

< 0.05

= 0.07

< 0.05

Conclusion

Endogenous PMN serine proteases such as elastase support PMN respiratory burst activity, but also promote PMN programmed cell death by enhancing caspase activity. This differential activity may in part explain the equivocal effects of elastase inhibitors in clinical trials in sepsis and ARDS.

Grant Acknowledgement

CIHR

0459

Selective Alpha7 Nicotinic Acetylcholine Receptor Stimulation Inhibits Cytokine Release in Human Blood Cells Stimulated with Various Toll-Like Receptor Agonists

M. Kox*1, J. C. Pompe1, P. Pickkers1, C. W. Hoedemaekers1, A. B. van Vugt2, H. G. van der Hoeven1

1ICM, 2Surgery, section Traumatology, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands

Introduction

The efferent vagus nerve can limit the inflammatory response via stimulation of the alpha7 nicotinic acetylcholine receptor (alpha7nAChR). Nicotine, a nonselective alpha7nAChR agonist has been shown to inhibit cytokine release in human PBMCs and monocytes as well as in animal models of inflammation. Due to its lack of selectivity however, nicotine has many unwanted side effects. Therefore, selective targeting of the alpha7nAChR may be promising for the treatment of inflammatory conditions. We compared the effect of the selective alpha7nAChR agonist GTS-21 and nicotine on LPS-induced cytokine release in human blood cells. Furthermore, because pattern recognition of the diverse classes of microbial products causing infection involves various TLRs, we investigated whether the immunomodulating effect of alpha7nAChR stimulation is dependent on the TLR stimulated.

Methods

Isolated monocytes and PBMCs of healthy male volunteers were incubated with the TLR4 agonist LPS in combination with GTS-21 (1 nM — 100 uM) and nicotine (100 uM — 1 mM). Furthermore, whole blood of healthy male volunteers was incubated with TLR2, TLR3, TLR4 and TLR5 agonists in combination with GTS-21, nicotine and the alpha7nAChR antagonist mecamylamine. TNF-alpha, IL-6, IL-1beta, IFN-gamma and IL-10 production was determined by ELISA and multiplex cytokine assays. All described changes are statistically significant.

Results

Stimulation of the alpha7nAChR by both GTS-21 and nicotine resulted in a dramatic dose-dependent inhibition of LPS-induced pro-inflammatory cytokine release in human monocytes and PBMCs. Likewise, pro-inflammatory cytokine release induced by stimulation of the various Toll-like receptors in human whole blood was dose-dependently inhibited by both GTS-21 and nicotine. Production of the anti-inflammatory cytokine IL-10 was not inhibited but stimulated by GTS-21. GTS-21 inhibited pro-inflammatory cytokine production stronger than nicotine at equimolar concentrations. Finally, inhibition of the alpha7nAChR did not have any effect on cytokine production.

Conclusion

Selective stimulation of the alpha7nAChR by GTS-21 has a profound antiinflammatory effect by inhibiting pro-inflammatory and stimulating anti-inflammatory cytokine release. In addition, GTS-21 proved to be more potent than nicotine in inhibiting proinflammatory cytokine release. The anti-inflammatory effect of alpha7nAChR stimulation was not restricted to a specific TLR. Therefore, this pathway appears to modulate the inflammatory response by a general mechanism. The lack of an effect of inhibition of the alpha7nAChR suggests that this receptor is not constitutively activated. Selective targeting of the alpha7nAChR using GTS-21 holds promise for future treatment options to modulate the innate immune response.

0460

Activation of Rxr Attenuates Cytokine and Chemokine Production in Human Monocytes

I. B. M. Kolseth*1, M. K. Dahle1, J. Å gren1, M. V. Tamburstuen2, S. P. Lyngstadaas2, A. O. Aasen1, J. E. Wang1

1Institute for Surgical Research, Rikshospitalet-HF and University of Oslo, 2Institute of Clinical Dentistry, Faculty of Dentistry, University of Oslo, Oslo, Norway

Introduction

Aberrant regulation of innate immune responses and uncontrolled cytokine bursts are hallmarks of sepsis and endotoxemia. Activation of the liver X receptor (LXR) have been demonstrated to suppress inflammatory genes. Our group has recently proposed that LXR is a key regulator of cytokine release in LPS-challenged human monocytes, possibly by interfering with posttranscriptional events (Myhre, 2008). LXR forms heterodimers with retinoid X receptor (RXR), which bind to LXR responsive elements in promoter regions of target genes. RXR is also a functional partner of several other nuclear receptors. We wanted to study the influence of RXR activation on endotoxin-induced cytokine release.

Methods

Peripheral venous blood was obtained from healthy volunteers, and mononuclear cells were isolated by Polymorphprep centrifugation and selective adherence. Adherent human monocytes were pre-treated with synthetic RXR (9 cis retinoic acid/9cisRA) agonists and subsequently challenged with lipopolysaccharide (E.coli LPS, 1 µg/ml). The amount of cytokines released from the adherent monocyte cultures were measured in the supernatants, and the intracellular phosphoproteins activated were measured in cell lysates by a multiplex antibody bead kit, measuring 30 different cytokines (Biosource) or 5 different phosphoproteins (Bio-Rad), according to the manufacturer's instructions. Differences between the groups were analyzed using one-way analysis of variance (ANOVA) for repeated measures with Newman-Keuls comparison test. P≤0.05 was considered significant.

Results

In this experimental model, using adherent human monocytes, activation of RXR by 9cis-RA (0,1 µM; 1µM), 1 hour before LPS stimulation, strongly decreased the LPS-induced levels of TNFalpha after 6 hours. Also, IL-6, IL-10, MIP-1alpha and MIP-1beta were significantly attenuated. The studies on intracellular signaling did not reveal any inhibition of LPS-mediated p38 MAPK, JNK, Akt, IkappaB and ERK phosphorylation after 20 minutes.

Conclusion

In this study we demonstrate that the nuclear receptor RXR has a strong anti-inflammatory effect in LPS-stimulated human adherent monocytes. The study indicates that RXR may be a target in the immunomodulation of sepsis.

Grant Acknowledgement

We would like to thank Rikshospitalet, the University of Oslo and the Health authority of Southern Norway for financial support.

0461

Chronic Sepsis Increases Local Anaesthetics Sensitivity of Muscle Voltage Dependent Sodium Channels

G. Gueret*1, L. Huard1, E. Guillard1, M. Gioux2, C. C. Arvieux1, J. Pennec2

1anesthesiology and critical care medicine, university hospital, 2UA FNST — Laboratoire de Physiologie, faculté de médecine, brest, France

Introduction

Loss of excitability of skeletal muscle is a central feature of critical illness myopathy (CIM). Electrophysiologic alterations of the voltage dependant sodium channel (NaV) were reported in animal models of CIM (1). As NaV are inhibited by local anaesthetics as ropivacaine, we wonder if sepsis modifiy the sensitivity of NaV to local anaesthetics. The aim of our study was to assess if the local anaesthetics sensitivity of NaV was modified by a chronic sepsis.

Methods

NaV properties were studied at different ropivacaine concentrations in control and septic rats. Chronic sepsis was induced by cecal ligation and puncture. At D8, peroneus longus was dissected then dissociated by collagenase treatment, and the sodium current recorded by macropatch patch clamp.

Results

Chronic sepsis induced a decrease of maximum sodium current (INamax): 20.68±0.17 vs 4.10±1.01 nA, mean difference = 16.58, IC95 = 14.94–18.21, p<0.0001. An increase of ropivacaine sensitivity was also observed in septic muscle channels (Figure 1). For a 1 mM ropivacaine concentration, INamax = 7.33±0.32 nA (35.46±1.53%) vs 0.36±0.05 nA (8.69±1.23%), mean difference 6.98 nA (26.77%), IC95 6.49–7.46 nA (23.99–29.55%), p<0.0001. The IC50 was 440 µM in control muscles vs 330 µM in septic ones. Open image in new window

Conclusion

Our results show that chronic sepsis induces a significant increase of skeletal muscle NaV sensitivity to ropivacaine.

0462

Mechanisms of Mrsa-Induced Endothelial Cell Death

R. Connelly*, P. Enkhbaatar, D. Traber

Anesthesiology, University of Texas Medical Branch, Galveston, United States

Introduction

Thermal injury leads to the loss of the natural cutaneous barrier to infection, which leads to microbial colonization by methicillin-resistant Staphylococcus aureus (MRSA). MRSA can induce apoptosis/necrosis in various cell types, but the specific molecular and cellular basis of MRSA-induced lung cell death is poorly understood. We have recently shown that MRSA induces a significant increase in endothelial microvascular permeability and lung NOS activity in our novel ovine model of MRSA-induced pneumonia and sepsis.

Methods

Human microvascular endothelial cells from the lung (HMVEC-L) were challenged with 105 cfu MRSA AW6 in the presence and absence of the NOS inhibitor LNAME or the NADPH oxidase inhibitor apocynin, then visualized for ROS/RNS generation via DCHF fluorescence. Real-time changes in mitochondrial membrane potential were tracked via JC-1 fluorescence. MRSA-challenged HMVECs were further analyzed via western blot and confocal microscopy, for 8-oxoguanine or poly (ADP-ribose) (PAR) polymer formation. Translocation of AIF from the mitochondrial compartment to the nucleus, activation of caspase-3, and consequent poly(ADP-ribose) polymerase (PARP) cleavage, all indicators of apoptosis, were also assessed.

Results

Compared to unchallenged controls, MRSA caused a 4-fold immediate nuclear ROS/RNS production that was dependent on NOS and NADPH oxidase activity, and a delayed 7-fold ROS/RNS response that was independent of NOS/NADPH oxidase activities. A significant shift in JC-1 fluorescence from 590nm to 520nm was observed within minutes, suggesting that MRSA causes a rapid depolarization of mitochondrial membranes. MRSA also caused 8-oxoguanine and PAR formation, indicators of oxidative DNA damage, within minutes of MRSA exposure. Furthermore, MRSA caused the translocation of AIF from the mitochondria to the nucleus as well as caspase-3 and PARP cleavage.

Conclusion

MRSA causes immediate NOS/NADPH oxidase-dependent nuclear oxidation. Mitochondrial depolarization also occurs within minutes, followed by a delayed mitochondrial oxidative burst and activation of apoptotic pathways involving AIF, caspase-3, and PARP.

Grant Acknowledgement

SHCG8450, SHCG8460 — NIH PPG Po12066312 — NIH RO160688.

0463

Midazolam Unmasks Blood Pressure Responses to The Katp Pore Blocker in Septic Rats

Y. Chan*1, R. Stidwill1, V. Taylor{ur1}, L. H. Clapp2, M. Singer1

1Wolfson Institute for Biomedical Research, 2Department of Medicine, University College London, London, United Kingdom

Introduction

Vascular KATP channel hyperactivity has been implicated in the pathogenesis of septic shock. However, glibenclamide, the channel's classic inhibitor, is ineffective in reversing septic shock clinically. In contrast, use of inhibitors acting on the channel's pore-forming subunit, rather than on the SUR regulatory subunit to which glibenclamide binds, can reverse sepsis-induced vascular hyporeactivity in vitro.1 We thus investigated the effect of pore blockade in an in vivo model, and the effect of sedation, a potential confounder that might alter vascular tone, neural reflexes, and sympathetic activity.

Methods

Male Wistar rats were subjected to faecal peritonitis and were fluid resuscitated for 6 hours. Changes in BP induced by norepinephrine (NE; 0.05 \mug/kg/min i.v. infusion), +/− the vascular KATP channel pore blocker PNU-37883A (PNU; 1.5 mg/kg i.v. bolus and 1.0 mg/kg/h i.v. infusion), were compared in awake and sedated animals (with midazolam 0.2 mg/kg i.v. bolus and 0.2 mg/kg/h i.v. infusion). The BP-lowering effect of the specific KATP channel opener levcromakalim (LEV; 150 \mug/kg i.v.), and the ability of PNU to reverse this, was also compared.

Results
Faecal peritonitis induced vascular hyporeactivity as early as 6 hours, as reflected by a substantially lower rise in BP to NE (Table 1). This was associated with accentuated responses to LEV, that were completely reversed by PNU. However, PNU did not reverse vascular hyporeactivity to NE, although it was able to cause a significant rise in baseline BP in sedated but not awake septic animals.

State

Treatment

Change in BP (mmHg) Sham (n = 4)

Change in BP (mmHg) Septic (n = 4)

Awake

NE

47.6±3.3

5.0±3.3

Sedated

NE

42.4±3.5

5.3±1.5

Awake

LEV

−29.3±2.3

−43.8±1.0*

Awake

LEV+PNU

0.1±2.6

−3.4±3.2

Awake

PNU

22.2±4.2

2.2±1.6

Sedated

PNU

15.3±1.5

11.0±2.5**

Awake

PNU+NE

31.8±1.6

5.9±2.7

Sedated

PNU+NE

36.0±0.5

5.5±1.3

* P <0.01 vs. sham animals; ** P<0.05 vs. awake animals

Conclusion

The vascular KATP channel is upregulated in sepsis, as evidenced by the greater hypotensive effect of pharmacological channel activation. KATP channel blockade is however ineffective as a pressor agent in awake septic animals. The increased effectiveness seen in the sedated state may be related to loss of neural control (baroreflex sensitivity) or decreased sympathetic stimulation. Thus KATP channel blockade might be of limited help clinically in reversing sepsis-induced vascular hyporeactivity.

Grant Acknowledgement

Chang Gung Medical Research Grant CMRPG260131.

0464

Phosphoinositide-3 Kinase Gamma Kinasae Activity Contributes to an Increased Coagulation Phenotypes During Sepsis

E. L. Martin*1, L. Del Sorbo1, V. Fanelli1, B. Assenzio1, V. Puntorieri1, L. DeSiena1, L. Mascia1, E. Hirsch2, V. M. Ranieri1

1Anesthesia and Intensive Care, 2Genetics, Biology and Biochemistry, University of Turin, Turin, Italy

Introduction

Sepsis is a systemic response to infection, which often leads to organ failure and death. An important aspect of the pathophysiology of sepsis is the systemic activation of coagulation and concurrent attenuation of fibrinolysis. Phosphoinositide-3 kinase gamma (PI3Kg), an isoform of the PI3K cell signalling family, has been shown to be involved in the development of sepsis. However, the role of this molecule in the dysregulation of the coagulation cascade during sepsis remains unknown. It is hypothesized that mice lacking PI3Kg or possessing a kinase dead form of this enzyme will be protected from septic-induced activation of coagulation and attenuation of fibrinolysis.

Methods

PI3Kg wild-type (WT), knockout (KO) and kinase dead (KD) mice were randomized to undergo either cecal-ligation and perforation (CLP)-induced sepsis or a sham laporatomy. After 18 hours, livers were harvested and levels of fibrinogen (pro-coagulatory), plasminogen activator inhibitor-1 (PAI-1) (anti-fibrinolytic), and tissue plasminogen activator (tPA) (fibrinolytic) were assessed.

Results
Following CLP-induced sepsis, WT mice developed a significant increase in fibrinogen and PAI-1 and a significant decrease in tPA compared to sham control. Both KO and KD mice also showed a significant increase in fibrinogen and PAI-1 and a significant decrease in tPA compared to sham controls, however these alterations were significantly less than observed in WT mice.
 

WT Sham

KO Sham

KD Sham

WT CLP

KO CLP

KD CLP

Fibrinogen (ng/mgPro)

805±35

807±40

779±33

1424 ±89 **

1043±57 *

1079±58 *

PAI-1 (pg/ώgPro)

2.9±0.2

2.7±0.2

3.1±.4

84.8±13.4 **

50.5±9.2 *

52.0±9.0 *

tPA (pg/ngPro)

404±17

369±53

423±23

1±1 **

104±36 *

61±31 *

Means ± SE. Significance is indicated by **,* = vs. all other groups, with p<0.05

Conclusion

PI3Kg knockout and kinase dead mice are partially protected from septic-induced activation of coagulatory molecules and attenuation of fibrinolytic molecules, which indicates that the kinase activity of PI3Kg plays a significant role in septic dysregulation of coagulation and may contribute to the development septic-induced injury.

Grant Acknowledgement

University of Turin.

0465

The Influence of Moxifloxacin on the Innate Immune System During Sepsis

C. Bode*1, D. Silivay1, P. Efferz1, P. Knuefermann1, H. Ehrentraut1, M. Velten1, S. Kim1, C. Grohé2, G. Baumgarten1

1Department of Anesthesiology and ICM, University of Bonn, Bonn, 2Department of Pulmonology, Lung Hospital, Berlin, Germany

Introduction

The innate immune system plays a central role for the pathogenesis of sepsis. It has been shown that anti-bacterial substances can modulate the immune system. However, it is not much known whether antibiotics modulate respectively stimulate the innate immune system. This study investigates whether Moxifloxacin (MXF) regulates the expression of the major innate immune receptors, the toll-like receptors (TLRs), and influences cytokine expression in an in-vitro sepsis model.

Methods

1) The monocytic cell line THP-1 was stimulated with Lipopolysacharide (LPS), MXF, and MXF in combination with LPS and phosphate buffered saline (PBS) as control for 24h. The TLR expression and the synthesis of the cytokines Tumour Necrosis Factor α (TNF-α), Interleukin-1 (IL-1) and Interleukin-6 (IL-6) were analysed on mRNA-level (RT-PCR). Results are presented as mean ± SEM. P-values ≤ 0.05 were considered to be significant. 2) To determine, whether MXF activates TLRs, a HEK293 cell line, which expresses a reporter gene driven by NFB, was cultivated MXF. Consequently, the activation of TLR activation was estimated by luminecence measurement.

Results

1) The mono-stimulation with MXF has no effect on TLR-expression and cytokine production. However, in contrast to the mono-stimulation with LPS, MXF in combination with LPS mediated significantly the up-regulation of TLR3 (15.08±1.81 Arbitrary Units (AU) vs. 0.36±0.38 AU), TLR4 (3.38±0.77 AU vs. 1.02±0.18 AU), TLR6 (3.62±1.40 AU vs. 0.76±0.29 AU), TLR7 (12.36±3.05 AU vs. 1.35±0.23 AU), TLR8 (56.30±3.52 AU vs. 6.20±0.75 AU; all p<0.001). In addition, the production of TNF-a (138.2±31.72 AU vs. 30.56±6.8 AU), IL-1b (2529±777.4 AU vs. 4.97±1.82 AU) and IL-6 (646.8±159.6 AU vs. 1.42±0.72 AU, all p<0.001) was raised after stimulation of MXF with LPS. In contrast, TLR2 expression is significantly higher for LPS stimulation in combination with MXF than LPS alone (2.25±0.47 AU vs. 1.46±0.28 AU; p<0.01). 2 MXF does not show any direct TLR activation in the TLR ligand screening.

Conclusion

MXF leads to a differential regulation of the TLR expression and cytokine production only in LPS-stimulated monocytes. However, MXF does not stimulate TLRs itself. Thus, it has been shown that MXF modulates the innate immune system in a gramnegative in-vitro model of sepsis.

0466

Anti-Inflammatory Effects of Drotrecogin Alfa (Activated) on Solid Organ Tissue in Murine Sepsis

T. Schuerholz*, R. A. Claus, E. Walther, T. Simon, M. Sieber, K. Reinhart, G. Marx

Anesthesiology and Intensive Care, Friedrich-Schiller-University, Jena, Germany

Introduction

Anti-inflammatory effects of drotrecogin alfa (activated) (DAA) have been identified. However, there is lack of data regarding direct organ-related effects of DAA. Since isolated murine heart cells are able to synthesize and release Macrophage Inflammatory Protein-2 (MIP2) upon stimulation with pro-inflammatory cytokines [1], we investigated whether DAA provides effects on murine heart in sepsis.

Methods

Sepsis was induced by cecal ligation and puncture (CLP) in male NMRI-mice (n=20, body weight 30±3g). Animals were randomly assigned to vehicle infusion (control), or CLP sepsis with DAA infusion (DAA; 24µg/kg/hr). A third group received only sham operation and vehicle infusion (Sham). 48 hrs prior to CLP all mice were given a permanent central i.v.-line and an arterial transmitter to measure heart rate (HR) and mean arterial pressure (MAP). CLP was adjusted to survive 24hrs. After 12hrs of sepsis, heart tissue was snap frozen in liquid hydrogen and plasma samples were collected and frozen following manufacturers guidelines until measurement. RNA was extracted of frozen tissue (RNeasy, Qiagen, Germany) and analysed by real-time-PCR with mouse specific primers to CD14 and MIP2 normalized to β-Actin (all: biomers, Ulm, Germany) using an iCycler (Biorad, Hercules, USA). Determination of Interleukin (IL)-6, IL10, and monocyte chemoattractant protein-1 (MCP-1) was performed using a cytometric bead array (CBA, BD Biosciences, Heidelberg, Germany). Differences between groups were evaluated by 1-way ANOVA and Pair-wise fixed reallocation randomisation test© for PCR diagnostic. All data are given as mean±SD. P<0.05 was considered significant.

Results

There were no significant differences in HR between the groups (Sham 650±38/min; DAA 547±175/min; Control 530±143/min). MAP was significantly higher in sham group (p=0.031) and non-significantly higher in DAA group when compared to control (Sham 137±15mmHg; DAA 111±26mmHg; Control 97±4mmHg). Normalized expression ratio of CD14 was 28.7 fold higher in control group compared to sham group. DAA treatment resulted in an 18.1 fold expression ratio (p=0.001). MIP2 expression ratio was 25.4 fold higher in control and 11.0 fold higher in DAA group. IL-6 (DAA 4663±6229 pg/ml, p=0.027; Control 10906±8324 pg/ml; Sham 54±32 pg/ml, p<0.001), IL-10 (DAA 220±338 pg/ml, p=0.072; Control 2539±5251; Sham 7±11, p=0.049) and MCP-1 (DAA 2708±3401 pg/ml, p=0.031; Control 7010±7095; Sham 75±107, p=0.001) were significantly lower after treatment with DAA compared to control.

Conclusion

Our data provide evidence of DAA-induced direct anti-inflammatory effects on murine heart cells, suggesting that DAA plays an important role in the anti-inflammatory response to septic injury.

0467

Cervical Chordotomy Increases Survival and Decrease the Degree of Systemic Inflammatory Response, Acute Lung Injury in Rats with Endotoxemia Induced by Lipopolysaccharide

X. Zhu*, Y. Lin, Y. He, Z. Wang, H. Li

SICU, Peking University 3rd Hospital, Beijing, China

Introduction

Study design: In vivo study using a chordotomy of C7 spinal cord model in adult male rats.

Objective: To investigate the effects of cervical chordotomy on systemic inflammatory response, acute lung injury and the outcomes in rats with endotoxemia induced by LPS.

Methods

The experiment was made up of two parts: Part one :Animals (N=72) were randomly divided into 3 groups: NC,E and E+C group, and those latter groups were divided into 4 subgroups respectively according to different time interval (n=8). Rats in E and E+C group were injected LPS intraperitoneally; and the C7 spinal cord of rats in E+C group were disconnected. Samples of blood and lung were collected at different time interval. The plasma levels of NE, IL-10, IL-6 and PaO2 were determined. The changes of histopathology and lung wet/dry weight ratio in every group were observed. Part two: Another 20 animals (N=20) were randomly divided into 2 groups: E and E+C group, each 10 animals. Rats in E and E+C group were injected LPS intraperitoneally;and the C7 spinal cord of rats in E+C group were disconnected. We observed the 48-hour survival rates of rats in E and E+C group.

Results
Rats in E+C group demonstrated decreased levels of NE and IL-6 with increased production of IL-10. The degree of lung injury was also alleviated, and PaO2 was improved. Moreover, the 48-hour survival rate was increased from 20%(E) to 70%(E+C).
Table 1

Plasma NE,IL-6,IL-10 Concentration and PaO2 and W/D Ratio in Each Group (N=8)

 

NE (pg/ml)

IL-6 (pg/ml)

IL-10 (pg/ml)

PaO2 (mmHg)

W/D

NC

209.5±77.0

163.1±31.9

188.7±58.0

100.8±17.0

4.1±0.2

E group 3h

1597.6±526.4*

341.0±41.3*

54.0±10.0*

90.2±3.9

4.3±0.3

E group 6h

1825.9±719.5*

297.3±74.2*

48.0±17.7*

62.4±8.0*

5.0±0.3*

E group 12h

2357.2±1271.5*

342.4±62.9*

112.4±17.1

72.6±6.0*

5.0±0.2*

E group 48h

3036.1±1103.7*

430.8±198.8*

252.5±18.6

78.1±5.5*

5.0±0.3*

E+C group 3h

1286.9±384.6*

228.5±62.5#

151.7±32.6#

94.7±4.9

4.2±0.4

E+C group 6h

874.3±361.4*#

202.9±63.2#

277.9±75.9#

92.2±4.0#

4.6±0.5*#

E+C group 12h

496.0±282.7*#

182.9±71.3#

752.2±33.5*#

94.3±10.5#

4.5±0.5#

E+C group 48h

467.0±128.4*#

276.2±10.0*

1236.6±247.1*#

96.8±13.5#

4.4±0.4#

*: campared with the NC group, p <0.05; #campared with the E group, p <0.05

Conclusion

It suggests that cervical chordotomy can inhibit the secretion of proinflammatory cytokines, alleviate the degree of acute lung injury and improve the outcomes significantly.

0468

Effects on Proinflammatory Response of Volume Replacement Fluids in a two Hit Model of Porcine Sepsis?

T. P. Simon*1, T. Schuerholz1, M. Sasse1, F. Heyder1, B. Schmidt1, W. Pfister2, G. Marx1

Department of Anästhesiology and Intensiv Care, 2Department of Microbiology, FSU Jena, Jena, Germany

Introduction

Using a porcine two hit model combining a haemorrhagic and septic shock we tested the effects of 6%HES 130/0.42 (HES130), Gelatine 4% (Gela) and 10%HES 200/0.5 (HES200) compared to Ringer Acetate (RAc) on inflammation and systemic haemodynamics.

Methods

Prospective randomised, controlled animal study. 23 anaesthetised, ventilated pigs (28.2 ± 1.7 kg) were randomised (5 in each shock group, 3 in control group) to volume replacement therapy with colloids, RAc or a non-septic control group (control) receiving RAc. Animals were bled in the shock groups until reaching half of their baseline mean arterial pressure (MAP) or cardiac output (CO) for 45 minutes. After hemorrhagic shock Volume resuscitation started until baseline MAP was accomplished. As second hit sepsis was induced using an E. Coli bacterial clot into abdominal cavity 6 h after hemorrhagic shock. Infusion rate was titrated to maintain a central venous pressure of 12 mmHg. Systemic haemodynamics and cytokines were obtained before (bl) hemorrhagic and septic shock and every 2 hours after induction of septic shock. Cytokine levels were assessed with commercially available enzyme linked immunosorbent assay kits specific for pigs and used according to manufacturer's guidelines. Statistics were performed using ANOVA.

Results

12h after sepsis induction MAP (mmHg) was lower in RAc group (59±5) compared to all other groups (HES200 70±21; HES130: 86±11, Gela: 71±26; Control: 96±7). CO [ml/kg] increased in all groups and was higher in colloid groups at study end (HES200: 157±61; HES130: 171±47; Gela: 160±42; RAc: 137±32; Control: 120±28). TNF-a Levels [pg/ml] was significantly higher in the HES200 group (4122±1417) after 2 hours of sepsis compared to HES 130, RAc and control at study end (HES130: 1199±1238; RAc 1101±498; Control 285±386; Gela 1939 ± 1961). The IL-10 [pg/ml] Levels increased after 12 hours of sepsis in all sepsis groups but only significantly in the HES 200 group (HES200: 411±245) compared to HES130, RAc, and control group (HES130 88 ± 197; RAc 14 ± 31;control 0±0). 2 hours after sepsis induction the IL-6 Levels [pg/ml] increased in all sepsis groups (HES130 1543±1521; RAc 874 ± 363; Gela 1623 ± 1242) and significantly in the HES 200 group (3845 ± 1472) compared to RAc (864 ± 363) and control (91 ± 157).

Conclusion

In this porcine two hit model 10%HES 200/0.5 induced a significantly higher proinflammatory response compared to 6%HES130/0.42, RAc and control. A more marked endothelial damage could be a potential mechanism explaining this result.

Grant Acknowledgement

Supported by a restricted grant of Bernburg AG, Germany.

0469

Infusion of 10% Hes 200/0.5 is Linked to Tubular Injury and Renal Dysfunction in Ovine Endotoxemic Shock

C. Ertmer*1, G. Köhler2, S. Rehberg1, B. B. Pinto1, A. Morelli3, M. Lange1, H. Van Aken1, M. Westphal1

1Department of Anaesthesiology and Intensive Care, 2Department of Pathology, University Hospital of Muenster, Muenster, Germany, 3Department of Anaesthesiology and Intensive Care, University Hospital of Rome “La Sapienza”, Rome, Italy

Introduction

The purpose of the present study was to compare the impact of 10% HES 200/0.5, 6% HES 130/0.4 and crystalloids on renal function and integrity in ovine endotoxemic shock.

Methods

Thirty sheep were subjected to incremental doses of endotoxin (started at 5 ng/kg/min and doubled every hr) until MAP fell below 65 mmHg. Thereafter, fluid resuscitation was initiated to maintain CVP at 8–12 mmHg and PAOP at 122-15 mmHg. In sheep allocated to the colloid groups, HES was infused up to a max. dose of 20 mL/kg. Thereafter, only crystalloids were infused in both HES groups. In the crystalloid group, a balanced, isotonic crystalloid (Sterofundin® ISO) was infused to achieve goal values.

Results

Creatinine concentrations were significantly elevated in sheep treated with 10% HES 200/0.5 as compared to the other two groups. Light microscopy revealed acute tubular injury characterized by necrotic tubular cells in the latter group. Likewise, electronic microscopy revealed tubular damage (Fig. 1) with destruction of epithelial cells (open arrows), vacuolar cell degeneration (dashed arrow), “naked” basal membranes (closed arrow), and intratubular protein precipitation (arrowheads) only in sheep treated with 10% HES 200/0.5. Open image in new window

Conclusion

The present study provides evidence that 10% HES 200/0.5 impairs renal function secondary to acute tubular injury in ovine endotoxemic shock.

0470

Peroxynitrite Decomposition Catalyst Ameliorates Severity of Methicillin-Resistant Staphylococcus Aureus (Mrsa) Pneumonia and Sepsis

P. Enkhbaatar*, A. Hamahata, M. Lange, C. Jonkam, R. Cox, L. D. Traber, D. Herndon, D. L. Traber

Anesthesiology, University of Texas Medical Branch, Galveston, United States

Introduction

Methicillin-resistant staphylococcus aureus (MRSA)-related pneumonia and/or sepsis are a frequent serious menace. We have previously shown that excessive nitric oxide mediates MRSA-induced cardiovascular collapse. In the present study, we tested the hypothesis that resultant reactive nitrogen species (RNS) play a major role in the MRSA-related cardiovascular morbidity.

Methods

Sheep were operatively prepared for chronic study. After 5 days recovery, tracheostomy was performed under anesthesia and pneumonia was induced by instillation of live MRSA (AW6) (5X1011 CFU) into the airway by a bronchoscope. After the injury, animals were awakened and maintained on mechanical ventilation by 100% O2 for first 3h and thereafter O2 concentration was adjusted according to blood gases. The sheep were resuscitated by Lactated Ringer's solution with initial rate 2ml/kg/h that was further adjusted according to hematocrit. Study groups: sham (non-injured, non-treated, n=6); Control (injured, not treated, n=4); and treated (injured, treated with peroxynitrite catalyst, INO-4885 [0.1mg/kg bolus followed by 0.02mg/kg/h continuous infusion], n=3). Experiment lasted 24 h.

Results

Injured animals showed the signs of severe sepsis-related multiple organ failure 3h after insult. Cardiovascular morbidity was evidenced by severe hypotension (MAP), with increased heart rate, cardiac output, left atrial pressure and severely decreased systemic vascular resistance index. Microvascular hyper permeability was evidenced by severe hemoconcentration, decreased plasma protein with decreased plasma oncotic pressure, and increased fluid retention. All these changes were attenuated by intravenous administration of peroxynitrite catalyst. MAP was 94±3 at baseline and 66±3 at 24 h in control vs. 93±1 at baseline and 82±10 at 24 h in treated groups. Fluid retention was 163±63 and 134±54 in control and treated groups at 24 h respectively.

Conclusion

RNS play a profound role in cardiovascular collapse due to MRSA sepsis. RNS inhibitors should be considered for the future therapy for septic patients.

Grant Acknowledgement

AHA CON14589

NIH GM066312, GM060688

Shriners of North America 8954, 8630.

Poster Sessions Perioperative infections: 0471–0477

0471

Deep Sternal Infection in Cardiac Surgery Patients

M. D. Carrasco González*, F. Roma

Postoperative Cardiac Surgery Unit, Hospital Vall de Hebron, Barcelona, Spain

Introduction

Deep sternal infection (DSI) is a severe complication of cardiac surgery. The incidence of DSI ranges from 0,5 to 5% percent, with the incidence in most centres being between 1 and 2%. DSI increases the duration and costs of hospitalization, and has a dramatic impact in the prognosis and outcome. The purpose of this study was to determine its exact magnitude in our patients and the prognosis.

Methods

Prospective, observational study carried out during a year (November 2006 – October 2007) in a tertiary-care teaching hospital. Inclusion criterion: adult patients undergoing open heart surgery via midline sternotomy. Surgical wound infections (SWI) were defined and classified according to the Centers for Disease control and Prevention (CDC) criteria.

Results

Five hundred and thirty eight patients underwent surgery. The mean age was 63 years (41–81), with 65,4% male. Coronary artery bypass grafting (CABG) followed by valve procedures were the main interventions with 38,5%, and 26,9% respectively. A total of 26 (4,83%) patients had a SWI in the chest area. 52,4% had history of tobacco use; 37,5% had prior respiratory disease and 68,6% overweight or obesity. SWI was considered superficial in 76,9% (20). DSl occurred in 1,1% (6/538). Overall, the most common infecting microorganism was coagulase-negative staphylococci (72,7%), with gram-positive cocci accounting for 86,4%, including Staphylococcus aureus (MSSA) in 13,6%. Gram-negative-bacilli were documented in 4,5%. In addition to antibiotic treatment, vacuum-assisted closure (VAC) was initiated as a therapy in 34,6%, before surgery. Mortality in patients with clear mediastinitis reached 100% (2).

Conclusion

Our incidence rate of deep sternal infection is within the standards with a high mortality rate for mediatinitis.

0472

Infectious Complications of Pulmonary Artery Catheter Use in Cardiac Surgery

I. S. Ilic*, D. Mijuskovic, D. Unic Stojanovic, M. Lukic, S. Jerotic, B. Krivokapic, M. Jovic

Clinic for anesthesia and intensive care, Dedinje Cardiovascular Institute, Belgrade, Serbia

Introduction

In critically ill patients, the impact of pulmonary artery catheter (PAC) use on outcome is debatable (1,2). Patients undergoing heart surgery show a high risk of catheter colonization and catheter-related bloodstream infections (3). The aim of this study was to assess the incidence and etiology of colonization and infection of pulmonary artery catheters inserted in cardiac surgery patients.

Methods

Design: prospective, observational study. Settings: A 16-bed heart surgery intensive care unit in a tertiary university hospital.Patients: heart surgery patients. Interventions: none. Criteria for catheter-related bloodstream infection (CR-BSI): a bacteremia or fungemia in a patient with an intravascular catheter with at least one positive blood culture obtained from a peripheral vein and clinical manifestations of infection (i.e., fever, chills, and/or hypotension) and no apparent source for the BSI except the catheter (3).

Results

Total number of patients included in the study was 183, 132 males and 51 females, average age of 63.7 years old. The average ejection fraction (EF) was 34.4% and Euro Score 6.8. The total mortality in the study group was 8.2% The PAC was left in place from 1 to 13 days, average 4.4 days. Twelve patients had positive blood cultures (6.6%), while 24 patients had positive catheter tip cultures (13.1%). The average number of days catheter was left in place in the group with positive catheter tip culture was 5.8 days. Six patients fulfilled criteria for catheter related sepsis (3.3%) of which two patients died. The study group had lower mortality compared to expected values based on low EF and high Euro Score. The number of days catheter was used presents as a risk factor for catheter colonization and blood stream infection.

Conclusion

The incidence of colonization of catheter tip and CR-BSI in the study group is similar to other studies and meta analysis. The heterogeneous characteristics of the group and lack of data related to number of days on mechanical ventilation, incidence of postoperative myocardial infarction, bleeding and re-interventions could decrease the value of results obtained. Further, randomized studies should evaluate use of pulmonary artery catheter and infection in cardiac surgery.

0473

Postoperative Procalcitonin Kinetics: An Indicator for Therapeutic Strategy in Peritonitis?

C. Lepouse*, O. Murat, F. Nicolai, F. Barbieux, E. Bankole, T. Floch, J. Petit, A. Leon

Anesthesiology and ICU, Robert Debre Hospital, Reims, France

Introduction

Differentiation between acute bacterial infection from other types of inflammation is often difficult in particular in postoperative period. The aim of this study was to assess the profile of time course of procalcitonin (PCT) in peritonitis patient after surgical and anti-effective therapies in ICU.

Methods

Every patient admitted to the ICU consequently to surgical procedure for peritonitis over a 18 months period, from July 2006 to December 2007, was included in our study. PCT, CRP and cytokines (TNF, IL-6, IL-8) were measured at admission (J0), on second day (J2) and on fourth and seventh day (J4, J7). Prognostic indicators of severity were recorded on admission: Apache II, IGS II and SOFA score.

Results

Sixty-one patients were included in the study. A second look operation was performed for 10 patients (group I). Fifty one patients (group II) were not surgically reinvestigated. Second look discovered persistent peritonitis 6, peritoneal abscess 1,necrotic pancreatitis 1, ileal fistula 1 and mesenteric ischemia 1. The day of ICU admission, the concentrations of PCT (53.9 ± 42.6 ng/ml) were not significantly different in the group I than in the group II (46.4 ± 51.4 ng/ml). There is a remarkable and significant decrease in concentrations of procalcitonin on day 2, 4 and 7 in group II whereas the concentrations of procalcitonin increase until day 4 then plateau in group I (p < 0.03).

Changes in procalcitonin patients with (group I) and without (group II) second look.

PCT (ng/ml)

Group I

Group II

p

Day 0

53.95 (n=10)

45.53 (n=52)

0.74

Day 2

68.08 (n=10)

18.08 (n=39)

0.00002

Day 4

82.21 (n=9)

5.38 (n=26)

0.0004

Day 7

38.66 (n=9)

2.75 (n=12)

0.03

Conclusion

During postoperative period after initial surgical procedure for peritonitis, persistence of high PCT levels may be related to inefficacy of anti-infective therapies. Therapeutic strategy must be discussed including second look.

Grant Acknowledgement

University of Reims Champagne Ardenne.

0474

Comparative Analysis of the Plasma Levels of Markers for Acute Inflammation in Patients Receiving Aprotinin and Corticosteroids During Coronary Artery Bypass Grafting with Extracorporeal Circulation

W. Homena*, E. Bastos, M. Pantoja, B. Santos, P. Resende, F. Oliveira, A. Pyramides, A. Siqueira, I. Pinheiro, V. Carreira

Surgical Intensive Care Unit, Barra D'Or Hospital and Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

Introduction

Approximately 20% of low-risk patients have postoperative complications. Several therapeutic measures aiming at reducing the SIRS after CPB, have been studied; huge difficulties, however, need to be overcome. So, our objective was to compare the plasma levels of markers for acute inflammation in patients undergoing extracorporeal circulation (CPB) receiving either aprotinin or corticosteroids.

Methods

The study comprised 40 patients CABG with CPB divided into the following 3 groups: G0 (control = 12); GA (aprotinin = 12); and GC (methylprednisolone = 16). Blood samples were collected as follows: T0, 24 hours prior to surgery; and T1 and T2, 3 and 18 hours after, respectively. Were quantified: TNF-alpha; IL-6; PCR; fibrinogen; haptoglobin; leukocytes and neutrophils.

Results

The groups did not differ in regard to their demographic, clinical and surgical characteristics. The IL-6 levels were lower in G0 as compared to those in GA and GC (374.4 ± 341.8; 260.2 ± 373.3; 249.2 ± 311.3 pg/ml, respectively), but not significantly (P = 0.2). The TNF-alpha levels in groups 0, A and C did not differ. The differences in the greatest postoperative serum levels (T1 and T2)for fibrinogen, haptoglobin and PCR were equivalent (P = NS). A difference was observed between the greatest levels in T1 and T2 and the baseline levels of leukocytes in group C as compared to those in groups 0 and A (15.0 ± 5.9; 8.4 ± 4.1 and 9.2 ± 3.4, respectively; P = 0.02 and P = 0.003). The relative levels of neutrophils did not differ (P = 0.3).

Conclusion

This study could not show a reduction in several markers of the acute phase of inflammatory response in patients undergoing CABG with CPB and receiving aprotinin and methylprednisolone.

0475

A Risk Factors Associsted with Arterial Puncture Following the Internal Jugular Vein Access Procedures

R. J. Jankovic*, M. S. Pavlovic, M. S. Andjelkovic, A. N. Bogicevic, D. R. Djordjevic, J. C. Ristic, Z. R. Stevanovic

Department for Anesthesia and Intensive Care, Clinical Center Nis, Nis, Serbia

Introduction

The most frequent complication with the internal jugular vein cannulation access is the carotid artery puncture. The objective of this 1-year prospective, singlecenter study was to illuminate the incidence of carotid artery puncture following more than one cannulation attempt with the internal jugular vein access procedures (IJVAP).

Methods

We analyzed all landmarks guided IJVAP that were performed by either the anterior or the posterior approach, using the Seldinger technique in both: operating theatre and intensive care unit, during 1-year period. The side of the puncture site was chosen according to clinical necessities. Age, gender, puncture side, number of cannulation attempts and number of arterial puncture after cannulation were recordered. Correct placement of central venous catheter was confirmed by free venous blood return, free flow of fluid through all ports of catheter and post insertion chest X-ray.

Results

During the study period, we studied 112 IJVAP attempts (74 right-sided and 38 left-sided). The patient s age and gender were not associated with increased risk for arterial puncture. The arterial puncture occurred in nine IJVAP attempts (8.03%). Arterial punctures occurred more often although the difference was not statistically significant in the left-sided IJVAP (left-sided IJVAP=13,1%; right-sided IJVAP=5,4%, p=0.12). The single attempt successful rate was 85, 7%, with low risk of arterial puncture (3,1%). Of 16, more than one cannulation attempt IJVAP, six (37.5%) were accompanied by unintentional carotid route puncture. More than one cannulation attempt frequently occurred during left sided IJVAP (left-sided IJVAP=26, 31%; right-sided IJVAP=8,1%, p<0.05), and was associated with a significantly increased arterial puncture rate(p<0.001).

Conclusion

Our results imply that more than one cannulation attempts frequently occurred during left-sided IJVAP, and it was considerably associated with increased arterial puncture rate. Generally, carotid puncture commonly occurred during left-sided IJVAPs, but this difference was not statistically significant.

0476

Serious Postoperative Complications after Esophagectomy for Esophageal Carcinoma: Analysis of Risk Factors

F. Arméstar1, E. Mesalles*1, A. Font2, A. Arellano3, J. Roca4, J. Fernández Llamazares5, J. Klamburg1

1Intensive Service, 2Oncology Service, 3Radiotherapy Service, 4Epidemiology Service, 5Surgery Service, Germans Trias i Pujol Universitary Hospital, Badalona, Spain

Introduction

Patient survival after surgery for locally advanced esophageal carcinoma has improved in recent years. The advances in surgical technique and perioperative management are the main reasons for this improvement. Nevertheless, these patients continue to present a high incidence of postoperative complications. To identify risk factors for postoperative morbidity and mortality in patients undergoing esophagectomy for esophageal cancer.

Methods

A series of 159 consecutive patients who had undergone esophagectomy for locally advanced esophageal cancer were analysed retrospectively. The study was performed at the Germans Trias i Pujol University Hospital in Spain. Patients were operated on between January 1985 and December 2004 by the same team of surgeons. In 1991, through the study period, preoperative chemotherapy and radiotherapy were introduced for the treatment of squamous carcinoma (SC). The data were obtained through the revision of clinical histories.

Results
Infections were the main cause of both complications and postoperative mortality. The 54% of our patients presented serious complications. The mortality of the series was 12.5%. Multiorganic failure secondary to sepsis was the more frequent cause of death. The postoperative complications showed a significant association with: alcoholism (p<0.04), hepatic cirrhosis (p<0.03), the location of the tumor in middle third of the esophagus (p<0.04), and the APACHE II score greater of 10 (p=0.003). Mortality was associated significantly (Table 1) with the presence of chronic pulmonary disease (p = 0,03) and with an APACHE II score superior to 10 (p = 0,02).
Table 1

Factors Associated with Mortality

Parameter

Odds Ratio

95% Wald Confidence Lower/Upper

p

APACHE II > 10

3.034

0.958 / 9.487

0.0228

COPD

3.225

1.132/9.687

0.0302

Conclusion

The APACHE II score can be used so much as prognostic factor of mortality like of serious complications. Chronic obstructive pulmonary disease is a risk factor for postoperative mortality. Alcoholism, hepatic cirrhosis and the location of the tumor, are factors associated to postoperative serious complications.

Grant Acknowledgement

Hospital's interdisciplinary committee for esophageal surgery.

0477

Vomiting Induced Spontaneous Diaphragm Rupture in a Patient with Ehlers Danlos Syndrome

M. K. Narayanan*1, R. S. Sardar2, A. B. Ratnaparkhi3

1Anaesthetics, St Thomas hospital, London, 2Anaesthetics, Southern general hospital, Glasgow, 3Anaesthetics, Royal Victoria Infirmary, Newcastle, United Kingdom

Introduction

Ehlers Danlos syndrome(EDS)is an inherited connective tissue disorder caused by the production of abnormal collagen. The major manifestations of EDS are skin fragility, skin hyper extensibility, and joint hypermobility (1). EDS typeIV can present as a surgical emergency with organ or arterial rupture. We present the spontaneous occurrence of diaphragmatic rupture with gastric herniation due to violent vomiting in a patient with EDS typeIV.

Methods

35 year old male known case of EDS presented with a 2 day history of non colicky lower abdominal pain; he had a negative laparotomy for suspected appendicitis. Post operative period was complicated by refractory nausea and vomiting and then developed respiratory distress with subcutaneous emphysema. Chest Xray and CT scan revealed a diaphragmatic rupture and gastric herniation. The patient underwent a diaphragmatic hernia repair but required ventilatory support for 24 hours post operatively. The patient was subsequently discharged from intensive care.

Results

We succesfully report the occurence of spontaneous diaphragmatic rupture due to projectile vomiting in a patient with Vascular type Ehlers Danlos syndrome, though transient respiratory support was required, the patient made an uneventful recovery. Open image in new window

Conclusion

Ehlers Danlos syndrome is characterised by tissue fragility, these cases can present as surgical emergencies due to bowel, organ or vascular rupture. The success of the management lies in the recognition of the syndrome and its associated complications.

Poster Sessions Tracheostomy in the ICU: 0478–0488

0478

Better Timing of Tracheostomy in Critically Ill Medical Patients Who Were Expected to Require Prolonged Mechanical Ventilation

H. Jang*, E. Jo, S. Hong, Y. Koh, C. Lim respiratory and critical care medicine, Asan medical center, Seoul, South Korea

Introduction

Tracheostomy is one of the most commonly performed procedure in the ICU. But there're no established recommendations about deciding the appropriate point of time. In addition to the effect of early and late timed tracheostomy on patient's outcome. We also tried to investigate whether the proper timing of tracheostomy would be different according to the patients’ different type of underlying lung disease.

Methods

From August 2006 to February 2008, This study has been conducted prospectively in medical ICU of one tertiary teaching hospital. Patients on their third to fifth day after translaryngeal intubation have been screened daily. 74 patients who were expected to require prolonged mechanical ventilation and satisfied our inclusion criteria were randomized. tracheostomy was performed surgically within 7 intubation days in early group and 14∼21 intubation days in late group.

Results

38 and 36 patients were separately randomized as early group and late group. Each 19 and 18 patients were excluded Because of final refusal to tracheostomy and inappropriate timing of procedure. Mean translaryngeal intubation days of each group were 6.6±1.5 and 16.2±3.4 days. Early group showed significant short mechanical ventilation duration(18.6±13.6 vs. 36.1±27.0 days, p=0.017). and short length of ICU stay(22.7±14.3 vs. 38.3±26.5 days, p=0.032). There are no significant differences in ICU mortality(26.3% vs. 38.9%, p=0.479) and hospital mortality(42.1% vs. 44.4%, p=0.615).

Conclusion

Our study showed that early tracheostomy can shorten the duration of mechanical ventilation and ICU stay in patient requiring prolonged mechanical ventilation. But possible differences according to the baseline lung characteristics need to study much more patients.

0479

Endoscopic Injuries and Laryngeal Dyspnea after Extubation. Incidence, Risk Factors

J. M. Tadié*1, E. Behm2, E. Guerot1, J. L. Diehl1, J. Y. Fagon1

1Medical Intensive Care Unit, 2Otorhinolaryngology, Ho^pital Européen Georges Pompidou, Paris, France

Introduction

Endotracheal intubation may generate laryngeal or tracheal complications responsible for respiratory distress resulting in extubation failure. This respiratory distress, with symptoms of upper airway obstruction, is commonly called laryngeal edema. Further studies have shown that different types of laryngeal or tracheal injuries and not only edema could be responsible for upper airway obstruction[1]. Our study has evaluated the incidence of postextubation laryngeal injuries in ICU and their relationship with postextubation laryngeal dyspnea and extubation failure.

Methods

After extubation a laryngeal nasofibroscopy was performed in 136 patients within 3 hours by otolaryngologists and damages (ulceration, granulation, edema and vocal cord mobility) were grading (0,1,2). Patients with audible inspiratory wheeze after extubation were classified as postextubation stridor (PES). PES with laryngeal endoscopic lesions were defined as Laryngeal dyspnea (LD).

Results

Thirty-six patients had a normal exam with a shorter duration of intubation (3.7±4.9vs8.0±9.3; p < 0.01), a lowest APACHE II at admission (23.1±7.2vs 26.3±7.8;p=0.03), and were intubated more frequently with neuromuscular blockade than patients with laryngeal injuries (97%vs81% p=0.04). We found 45 patients with ulceration, 74 with edema, 15 with granulation and 26 patients had vocal cord mobility altered.18 patients (13.2%) presented PES. Five patients did not have laryngeal injury. Two of them had postextubation tracheal injuries : sub-glottic stenosis and obstructive fibrinous tracheal pseudomembrane. Three patients had saliva over the larynx making obstruction associated with poor cough reflex. 13 patients developed LD (9.6%). Endoscopic findings were: 12 patients with grade 2 vocal cord edema and altered mobility. One patient had a grade 2 vocal cord granulation with obstruction of the laryngeal lumen. In unvaried analysis, risk factors of LD were: gender female, duration of translaryngeal intubation, number of intubation and a smaller height/ETT diameter ratio. Height/ETT diameter ratio was the only clinical independent risk factor of LD (OR 0.90 [0.85–0.96]; p<0.001). 17 patients (12.5%) needed to be reintubated in the 48 hours following extubation. In unvaried analysis, these patients had more frequently: a GCS<15 at extubation (11/17 vs 27/119; p<0.01), a longer duration of intubation (13.7 ±13.8 days vs 5.9 ± 7.2; p<0.01), a LD (5/17 vs 8/119 LD; p=0.01), they presented more frequently abnormal vocal cord mobility (10/17 vs 16/119; p<0.01) and grade 1 or 2 granulation (5/17 vs 10/119; p=0.02). Independent risk factors of extubation failure were: GCS<15 at extubation, grade 1 or 2 granulation of vocal cord and abnormal vocal cord mobility.

Conclusion

Laryngeal injuries following intubation are frequent and could be associated with a use of large diameter of ETT, a long duration of intubation, and the absence of using curare during intubation. These injuries could have several localization and different patterns, but all of these injuries can potentially result in significant airway obstruction. It not seems possible to recognize each type clinically, only endoscopic examination could determine it and the best treatment to avoid recurrences and extubation failure.

0480

Patient Outcomes after Extubation in an Adult Intensive Care Unit

P. A. Hampshire1, G. A. Iyer*2, G. George2, J. Walker3

1Department of Anaesthesia, Whiston Hospital, 2Department of Anaesthesia, University Hospital Aintree, 3Department of Intensive Care, Royal Liverpool University Hospital, Liverpool, United Kingdom

Introduction

Reintubation following extubation may be associated with higher mortality for intensive care patients. Risk factors for reintubation include pneumonia and high rapid shallow breathing index (RSBI)1. Weaning, extubation, and the use of continuous positive airways pressure (CPAP) after extubation are not protocolised on our unit. An observational study over six months was performed to establish the rate of reintubation on our unit and to describe the characteristics of the patients who were extubated.

Methods

Between January and July 2007 100 consecutive patients who were extubated on the intensive care unit were studied using a standardized questionnaire. Only data from the first episode of extubation was collected if patients were extubated more than once. Separate admissions to intensive care were counted as separate episodes. Statistical analysis was performed on contingency tables using two-tailed Chi-squared tests with Yates’ correction. All analysis was done using Microsoft Excel spreadsheet software.

Results

Ninety-seven patients were included in the analysis. Three patients were excluded from the analysis due to incomplete data. Mean age (SD) was 57.2 (17.2) years. 47.4% of admissions were medical, and 52.6% of admissions were surgical. The duration of intubation prior to extubation was a median of 3 days (range 4 h to 24 days; IQR 1–5 days). There were 16 reintubations (16.3%) and 14 tracheostomies were performed (14.4%). Mean time to reintubation was 16.9 h (range 10 mins – 96 h). The mean (SD) duration of the first intubation period in reintubated patients was 6.4 (5.1) days. Patients who were reintubated were admitted with pneumonia (3 patients, 18.8%); post-cardiac arrest (3 patients, 18.8%) or after aortic aneurysm repair (2 patients, 12.5%). 69 (70.4%) patients were extubated after being weaned onto CPAP (with or without additional pressure support). 28 patients (28.8%) were extubated from biphasic positive airways pressure ventilation. There were 84 (85.7%) planned extubations, of whom 13 patients (15.5%) were reintubated. There were 9 (9.2%) unplanned extubations, of whom 3 (33%) were subsequently reintubated. Unplanned extubation was not significantly associated with reintubation (p=0.37). Prior to extubation, mean (SD) fraction of inspired oxygen was 0.38 (0.08), mean (SD) P-F ratio was 29.8 (13.4) kPa and median respiratory rate was 10.5 breaths per minute. Non-invasive CPAP was used in 26 (26.8%) of patients after extubation. Five (19.2%) of these patients were re-intubated. There were only 7 deaths on the unit among the 97 patients studied, giving a unit mortality rate of 7.2%. Mortality among reintubated patients was 12.5% compared to 6.2% in the non-reintubated group, but this was not statistically significant (p=0.72). 75% of those patients who were reintubated went on to have a tracheostomy.

Conclusion

There was no association between reintubation and unit mortality in this population of patients. There was no association between unplanned extubation and reintubation. 75% of patients who were reintubated went on to have a tracheostomy to facilitate weaning.

0481

Comparison of Two Techniques of Percutaneous Dilatational Tracheostomy (PDT): Griggs'; Dilating Forceps Vs Ultraperc® Single Stage Dilator

D. Govil*, S. Gupta, O. Prakash, A. Malhotra, P. Mehta, P. Govil, D. Arora, S. Das, R. Tobin, P. N. Kakar

Anesthesiology, Max Super Speciality Hospital, New Delhi, India

Introduction

The percutaneous dilatational tracheostomy is a minimally invasive, safe, reliable and widely accepted procedure. PDT using Griggs’ dilating forceps over the guidewire is still practiced in many ICUs. Recently, a single stage dilational tracheostomy kit, Ultraperc® having a gradually curved dilator and an ergonomic tube introducer has gained popularity. The aim of our study was to compare the two techniques of PDT in terms of procedural time, amount of blood loss and immediate complications.

Methods

Thirty consecutive patients on mechanical ventilation needing tracheostomy during the course of treatment were randomly divided into two equal groups. PDT was performed in either group under sedation, as per the technique. Procedural time was defined from incision to successful placement of tracheostomy tube. Amount of blood loss was calculated by percentage soakage of a 12inch × 9inch gauze piece. 5ml blood completely soaks such a gauze piece. Immediate complications such as bleeding from the stoma, transient hypoxemia (SpO2<95%) and hypercarbia (>5 mm increase in ETCO2 from the baseline) were recorded. Bleeding was classified in four grades: I :<5 mL, II : 6–10 mL, III : 11–20 mL and IV :>20 mL. Chest x-ray was done after the procedure to look for pneumothorax, lung collapse and position of tracheostomy tube.

Results
The procedure with Ultraperc® single stage dilator technique took slightly longer time but had reduced amount of blood loss as compared to Griggs’ dilating forceps technique. The incidence of hypoxemia and hypercarbia was higher with Ultraperc® single stage dilator technique attributed to longer procedure time.
 

Griggs’ Dilating Forceps

Ultraperc® single Stage Dilator

Procedural time (rain)

2.5 ± 0.5

4 ± 0.7

Blood loss Grade I

2

9

Blood loss Grade II

3

5

Blood loss Grade III

9

1

Blood loss Grade IV

1

0

Hypoxemia

3

5

Hypercarbia

2

4

Conclusion

Although Griggs’ dilating forceps technique is an easier and quicker technique to perform but it causes more tissue trauma than Ultraperc® single stage dilator technique.

0482

Use of Portable Ultrasound to Minimise Risks Associated with Percutaneous Dilational Tracheostomy

R. M. Mahroof*1, R. Rajendram2, G. Morgan3

1Anaesthesia and Intensive Care, John Radcliffe Hospital, Oxford, 2Anaesthesia and Intensive Care, London Deanery, London, 3Anaesthesia and Intensive Care, Queen Alexandra Hospital, Portsmouth, United Kingdom

Introduction

Percutaneous dilational tracheostomy (PDT) has become the commonest surgical procedure performed within the ICU, since its description by Ciaglia in 1985. PDT performed in selected patients by experienced and trained individuals is comparable in safety to that of surgical tracheostomy (ST)1. PDT is easier and quicker to perform with less periprocedural bleeding, is cost effective and is cosmetically more desirable than ST. However, it is associated with a 5–11% risk of complications, including catastrophic haemorrhage2, pneumothorax, surgical emphysema, misplacement, cartilage damage, and death2. Bleeding is the commonest complication associated with PDT. MRI assessment of cervical anatomy suggests considerable variation in the relation of blood vessels to the anterior aspect of the trachea3. Portable ultrasound (P-US) prior to PDT has demonstrated the presence of a vulnerable vessel in up to 27% of patients4. NICE guidelines on the placement of central venous catheters encourage the use of P-US to identify blood vessels and to enhance the safety of the procedure.

Methods

The project was submitted to the Local Research and Ethics Committee for approval. P-US was used in 97 serial adult patients who had no history of surgery to the neck, prior to PDT as a planned procedure in the general ICU. US examination of the neck was undertaken using a SonoSite iLook25 with a linear 10MHz probe with the patients’ neck prepared in position for a standard PDT(SonoSite, Inc, Bothell, WA, USA). The focussed PUS study examined the ease of identification of the tracheal rings and their relation to the midline, the presence of arteries or veins within a 2cm radius of intended point of entry of seeker needle, and the position of the thyroid isthmus or lobes. Subsequently, P-US was not used to further facilitate the PDT procedure.

Results

Ninety seven patients (M:F,45:52) between the age of 26–74yrs (mean 58yrs) were studied. The tracheal rings and their midline, thyroid lobes and isthmus were identified in all patients. Thirty one patients (32%) had a clinically undetected vessel within a 2cm radius of intended point of entry (6 art:25 veins). Of these 4 and 11 respectively were more than 4mm in diameter. The latter 15 patients did not undergo a PDT, and instead, ST was performed safely and without immediate complications. The remaining 82 patients underwent safe and uneventful PDTs without any immediate complications.

Conclusion

P-US is a strong and useful non-invasive tool which may enhance informed patient selection and risk stratification, with resultant improvements in the safety and outcome of PDT.

0483

Routine Chest Radiography after Uncomplicated, Bronchoscopically Directed Percutaneous Dilatational Tracheostomy is not Required

D. McIntosh*, I. Gonzalez, S. Bonner

Intensive Care Department, The James Cook University Hospital, Middlesbrough, United Kingdom

Introduction

Percutaneous dilatational tracheostomy (PDT) is an established technique used to facilitate ventilatory weaning and avoid the complications of long term translaryngeal intubation. Post-procedure chest radiography is common however its value after a routine PDT has been questioned [1,2].

Methods

Our institution's practice is a bronchoscopically directed PDT technique using a Ciaglia ‘Blue Rhino’ dilator for tube insertion in adults. A retrospective audit of case records and radiological images was undertaken from 1996 to 2001. This was followed by a prospective cohort study from 2002 to 2005 where every PDT was followed by a chest radiograph (CXR) and compared to the most recent CXR taken as part of routine clinical practice. The responsible physician commented as to whether the PDT was routine and if the CXR changed post-procedure management. A review in January 2006 changed practice and a post-procedure CXR was required only if there were technical difficulties or concerns with ventilation.

Results

A total of 221 patients were identified in the retrospective study from 1996–2001. The overall complication rate was 8.6% subdivided as follows: bleeding (5.4%), mispositioned tube (0.9%), surgical emphysema (0.9%), pneumomediastinum (0.45%). All cases with radiographic abnormalities had been technically difficult with a high index of clinical suspicion for injury. The prospective cohort studies from 2002–7 identified a total of 478 cases − 218 cases had a mandatory post-procedure CXR (before the change in practice in 2006) and 260 with CXR only if clinically indicated. Of the mandatory post-procedure CXRs following routine PDT, there was no evidence of barotrauma or tube misposition. Following routine PDT the CXR altered management in only one case where a higher level of positive endexpiratory pressure (PEEP) was used to treat radiographic atelectasis. From 2002–7 there were 2 cases of surgical emphysema with immediate clinical manifestation. Both cases required no specific treatment and had been difficult with multiple tracheal cannulations. CONCLUSION. We have studied a total of 699 PDTs over 11 years and after routine PDT no evidence of barotrauma or other problem requiring urgent intervention was identified on CXR. On one occasion the post-procedure CXR affected management, treating radiographic atelectasis, however derecruitment of bronchopulmonary segments is common after PDT and can be anticipated [2]. In uncomplicated, bronchoscopically guided PDT a routine postprocedure CXR is unnecessary in the absence of clinical deterioration. We recommend that a chest radiograph is mandatory in the following circumstances — anticipated postoperative complications secondary to technical difficulties, visualised tracheal damage or pre-procedure pneumothorax.

0484

Safety and Complications of Percutaneous Tracheostomy in a Cohort of 800 Mixed ICU Patients

G. Díaz-Reganñó, M. A. Ballesteros*, A. Ruíz, S. González-Herrera, M. Holanda-Penã, F. López-Espadas, E. Miñambres

ICM, Hospital Universitario Marqués de Valdecilla, Santander, Spain

Introduction

Percutaneous tracheostomy (PCT) is a well established technique used primarily to assist weaning from mechanical ventilation on many intensive care units (ICU).

Objective: The aim of this study is to present our experience with PCT performed at ICU bedside by critical care staff and residents in the last two years of fellowship training in order to evaluate its efficacy in terms of safety and intraoperative and postoperative complications. We also aimed to quantificate the learning curve for PCT.

Methods

We analysed our experiences of a total of 800 PCT performed in our ICU by a collaborative team (critical care and otolaryngologist physicians). Statistical significance was assessed by the Chi-squared test with significance interpreted as a p value < 0.05.

Results

Most of the procedures (n = 685, 85.6%) were performed at the bedside by residents in the last two years (fourth and fifth) of fellowship training. We observed complications in 32 patients (4%). Intraprocedural complications occurred in 17 patients (2.1%), early postprocedural in 6 (0.75%), and late postprocedural in 9 (1.1%). No case of death was directly related to PCT. The number of complications was superior in PCT performed by the residents in their initial five attempts than in the rest of attempts (9.2 % vs 2.6%; p<0.05).

Conclusion

PCT is a simple procedure to be learned and to achieve adequate training. The low incidence of complications indicates that bedside PCT can be performed safely, and can be considered a routine procedure in the management of ICU patients in daily practice.

0485

Early Vs Late Percutaneous Dilatational Tracheostomy (PDT) in an Italian Cardiac Intensive Care Unit

G. Merli*, S. Gregu, C. Beverini, S. Salis, C. Brambillasca, E. Sisillo

Anesthesia and Intensive Care, Centro Cardiologico Monzino IRCCS, Milano, Italy

Introduction

Early tracheostomy is considered useful in improving ICU patients survival. However some controversies exist when considering cardiac surgery patients because PDT, when performed soon after sternotomy, could be associated with an increase in deep sternal wound infections and in overall morbidity and mortality.

Methods

86 cardiac patients, both medical and surgical, admitted to our ICU, underwent PDT during the last three years. A retrospective analysis of their clinical data was performed in order to study the possibile correlations between early or late tracheostomy and outcome in both classes of patients.

Results

45/86 patients (52.3%) were surgical patients, admitted to ICU after Coronary Artery Bypass Grafting or Valvular surgery. Most of them (75,6%) underwent combined or emergent surgery and in several cases were reoperated.

41/86 patients (47.7%) were admitted to medical ICU after complicated myocardial infarction.

Both groups had similar mortality rates: 26/45 (57.8%) in surgical patients and 21/41 (51.2%) in medical patients.

Medical patients showed an increase in mortality when late PDT was performed while surgical patients showed the highest incidence of mortality after early PDT.
Table 1

Mortality Rates According to Different PDT Timings

PDT timing

< 7 days

> 7 days

> 10 days

> 14 days

Medical pts

19/41 (46.3%)

22/41 (53.7%)

12/41 (29.3%)

8/41 (19.5%)

Mortality rate

9/19 (47.4%)

12/22 (54.5%)

8/12 (66.7%)

6/8 (75%)

Surgical pts

16/45 (35.6%)

29/45 (64.4%)

23/45 (51.1%)

16/45 (35.6%)

Mortality rate

10/16 (62.5%)

16/29 (55.2%)

11/23 (47.8%)

9/16 (56.2%)

Conclusion

According to recent literature, our data suggest that tracheostomy itself is not able to improve the outcome in unselected patients. Different characteristics and particular requirements of the patients seem to play an oustanding role in determining the indications to tracheostomy, in chosing the correct timing and in showing the real effects on outcome.

0486

Legs Down Improve Ventilation and Oxygenation in Supported Invasive Ventilation

F. van Beers*, B. Speelberg

Intensive Care, St.Elisabeth Hospital Tilburg, Tilburg, Netherlands

Introduction

Body position is of great influence in ventilated ICU patients. According to the surviving sepsis guidelines, patients should be ventilated in the 45 degree body position. In a previous abstract we demonstrated the role of the 30 degree body position on ventilation and oxygenation parameters in patients who were weaned from the ventilator. In this study we investigated the potential benefit of legs horizontal and legs down position on end tidal CO2 (%), rapid shallow breathing index and oxygen saturation in patients who were on pressure support ventilation.

Methods

16 patients on pressure support ventilation were studied. Patients were ventilated with a Servo 300(A)® and Servo-i® ventilator (Maquet). First, patients were placed in the 30 degree position with the legs horizontal. Half an hour after stabilisation, end tidal CO2, tidal volume, breathing frequency and oxygen saturation were measured. After that legs were lowered as much as possible. Rapid shallow breathing index was calculated as :respiratory rate divided by tidal volume. Values were compared using paired samples T-test. A significance level of < 0,05 was considered significant.

Results

Oxygen saturation improved with 0,63% +/− 0,96 when legs were lowered (p< 0,05). End tidal CO2 improved with 0,16%+/− 0,26 when patients were placed in legs down position (p<0,05). Rapid shallow breathing index did not show significant changes.

Conclusion

We demonstrated an improvement in oxygen saturation and a decrease in end tidal CO2 when the legs of patients were lowered during pressure support ventilation. Perhaps relieve of abdominal pressure plays a role in an increased ventilation. In this study patients were examined who were weaned from the ventilator. We think, positional therapy in this group of patients is very important. Further study is necessary to confirm our observations.

0487

Tracheostomy Practice in The ICU of Two Large Tertiary Medical Centers

E. Paramythiotou*1, J. Papanikolaou2, A. Karabinis3, A. Armaganidis1

1ICU, Attikon University hospital, 2ICU, George Gennimatas general hospital, 3ICU, G, Athens, Greece

Introduction

Tracheostomy is a common procedure in critically ill patients undergoing prolonged mechanical ventilation. The proper time and indications remain controversial and different practices may exist between hospitals. Our aim was to record practice concerning tracheostomy in the Intensive care units of two large tertiary centers in Greece.

Methods

Our study was conducted in two large tertiary care hospitals. “Attikon” university hospital has an 18 — bed Intensive care unit and “George Gennimatas” general hospital has a 12 bed Intensive care unit. Patients hospitalized in these two medical — surgical ICUs between 1st January 2007 and 31st December 2007 who were submitted to tracheostomy were included in the study. Demographics of patients, cause of admission, severity of admission as measured by Apache II score, timing, duration of mechanical ventilation, and outcome were recorded.

Results

Four hundred patients were hospitalized during the study period (225 and 175 for each hospital). One hundred twenty three patients were submitted to tracheostomy (31%, 61/ 225 for “Attikon” and 62/175 for “G Gennimatas”). There were 78 men and 45 women. Percutaneous tracheostomy was performed in 98% of cases. Surgical tracheostomy was performed in the rest because of nodular thyroid gland. Complications were noticed in two patients (one case of hemorrhage and one case of trachea trauma ending to death). Median age of patients was 59 years, (range 17 – 97). Median length of ICU stay was 47 days (range 5 – 230 days). Mean Apache II severity score was 19 (range 15 – 25). Overall number of deaths was 47 (66%, 27/ 62 and 15 / 61). Seventy — three patients were surgical (59%) and the rest were medical. There were 17 cases of abdominal surgery, 10 cases of cardiac / thoracic surgery, 28 neurosurgical patients, two burn patients,12 multi-trauma patients, one p with tracheal trauma and 3 orthopedic cases. Mean time for tracheostomy performance was 18 days (range 12 – 30).15/ 61 and 12/ 62 patients needed continuous venovenous hemodiaflitration. Mean duration of mechanical ventilation was 35 days (range 12 – 230 days).

Conclusion

Although the severity of hospitalized patients is similar in the two hospitals, tracheotomies are performed with a different frequency in each one of them. Timing and indications of tracheostomy are a matter of controversy and although it can help weaning from mechanical ventilation, it remains a surgical procedure with possibly severe complications.

0488

“Blue Dolphin”® Percutaneous Dilative Tracheostomy: Evolution of Ciaglia Tecnique is Going on. Preliminary Results

G. Gambale*1, C. Coniglio2, R. M. Corso1, F. Cancellieri2

1Anaesthesia and ICU Dpt., Morgagni-Pierantoni Hospital, Forlì, 2ICU Emergency Dept., Maggiore Hospital, Bologna, Italy

Introduction

23 years ago the Authors of first Ciaglia Percutaneous Dilative Tracheostomy (PDT) published the preliminare results. This I.C.U. bedside procedure introduced one of the most useful innovations for intensivists. Fifteen years after Ciaglia Blue Rhino (BR) represented a very good development of the PDT with one-step dilation. “Blue Dolphin” (BD) PDT is the last evolution to the one-shot procedure.

Methods

Seventeen patients admitted to I.C.U. (of two Italian Hospital) with acute diseases conditioning long-term mechanical ventilation were undergone consecutively to PDT with “Blue Dolphin” (BD) technique. The first part of the procedure was quiet similar to the well known one of previous abovementioned methods. After the use of first 14 Fr. introducer dilation and insertion of the cannula was obtained with a balloon-tipped catheter loading dilator assembly where, over the loading dilator, had at first preassembled the tracheal cannula. The dilative balloon is very like to a balloon for Percutaneous Transluminal Coronary Angioplasty obviously modified and bigger. The balloon was inflated at 11 atmospheres of pressure for 5 or 6 seconds. For every procedure was performed an inside and outside video to check every aspect of the “modus operandi”. Vital signs of the patients was of course monitoring continuously.

Results

The BD was performed in the group of patients without choosing “best necks”. The operating doctors were the same of the common approach (“Blue Rhino” PDT). Sixteen on seventeen procedures were uneventfully completed. Median time of BDs was 2,5 minutes (from the 14 Fr. introducer removal to the insertion of the cannula). Vital signs were unmodified during the procedures. The single case in which BD was not completed was related to the choice of the level of dilation (between cricoid and first tracheal ring) and consequent failing balloon dilation. The PDT was however completed with BR rescue. Neither major nor minor complications occurred all over the study.

Conclusion

First operating experiences with BD show that the introduction of a dilation ballon is effective, quick and safe. The assembly of balloon together the cannula loading dilator optimizes the kit reaching just about a one-shot PDT. Further studies need to be done for a more definitive evaluation.

Poster Sessions Infection and sepsis: 0489–0498

0489

Risk Factors Predicting Candidaemia Species Potentially Resistant to Fluconazole: Implications for Empiric Therapy

A. Diaz-Martin*1, J. Garnacho-Montero1, M. Ruiz-de Piapon2, J. A. Marquez-Vacaro1, J. Aznar-Martin2, C. Ortiz-Leyba1

1Intensive Care Unit, 2Microbiology, H. U. Virgen del Rocio, Seville, Spain

Introduction

C. albicans is the species responsible for the majority of candidaemia in critically ill adult patients, followed by C. tropicalisand C. glabrata. In the last few years there has been an increase in the number of Candida non-albicans isolates, with special attention being paid to C. glabrataand C. krusei in view of their diminished sensitivity or resistance to fluconazole. As fluconazole is still a main line therapy for candidaemia, risk factors identifying infection by the more resistant species would be very useful in order to provide early and appropriate therapy. Objective of the study: To identify risk factors for candidaemia species potentially resistant to fluconazole.

Methods

Observational study, including all patients with Candida isolates in blood cultures in adult patients in our hospital during the period January 2003 to December 2007. Two groups were compared: those with fluconazole-sensitive Candida isolates (ATP group: albicans, tropicalis, parapsilosis), and patients with potentially resistant to fluconazole Candida isolates (GK group: glabrata, krusei). We analyzed demographic risk factors, underlying diseases (diabetes, neutropenia, cancer, HIV, severe trauma, solid organ transplantation, bone marrow transplantation, acute renal failure), APACHE II on the day the positive blood culture was taken, and prior treatment with antibiotic, antifungals and/or steroids.

Statistical analysis: chi-square or Fisher's exact test, Student's t or Mann-Whitney U-test as needed, and multivariate analysis including variables with p<0.05 in the univariate analysis.

Results

154 patients were included. The average age was 58±18 years old. There were a similar distribution in gender (female 63 cases, 40.6%), severity score measured by APACHE II on the candidaemia day was 12.1±5.9. 62 patients (40%) were in the ICU with an average stay in the ICU prior to candidaemia of 11.7±9.8 days. Candida sensitive to fluconazole accounted for 79.9% of cases (n=123: C. albicans 70, C. tropicalis 26, C. parapsilosis 25, other 2), while 20.1% were Candida species potentially resistant to fluconazole (31 cases: C. glabrata 20, C. krusei 11). Variables associated with Candida species isolated from the GK group were neutropenia, acute renal failure and azole pretreatment. Multivariate analysis showed that azole pretreatment, OR 8.23 (95% CI 2.33–29, p = 0.01) and renal insufficiency OR 3.7 (95% CI 1.5–9.0 p = 0.004) were the only risk factors independently associated with this group. We performed the same multivariate analysis in the 62 patients who required ICU admission, and found azole pre-treatment as the only risk factor, OR 5.86 (95% CI 1 34, p = 0.05).

Conclusion

Prior azole treatment is a major risk factor for fungaemia involving species potentially resistant to fluconazole. In such patients, a non-azole antifungal agent may be a superior option.

0490

Pulmonary Infection Verses Acute Chest Syndrome/Vaso-Oclusive Crisis (VOC) in Sickle Cell Disease (SCD) At an Intensive Care Unit (ICU)

A. Hameed*, H. Mohamed

Intensive Care Unit, Salmaniya Medical Complex, Manama, Bahrain

Introduction

Sickle cell disease includes a group of diseases where the red blood cells produce a mutant form of β hemoglobin called sickle hemoglobin. Sickle cell anemia is the best known example of sickle cell disease. We report our experience with SCD patients admitted to ICU either with Acute Chest Syndrome (ACS)/Vaso-oclusive crisis (VOC) or pulmonary infections and outcome. Individuals with sickle cell disease generally fare well, but suffer from periodic sickle cell crises. These crises occur when red blood cells become stuck in blood vessels, and cause ischemia or infarction. Such crises can cause pain and organ damage. Patients with sickle cell disease are also prone to infections and have an average life expectancy of 40 years.

Methods

Data was collected from patients file on day-to-day manner from October 2002 to December 2006. This is 11-bed adult medical/surgical ICU with admission rate 45–50 patients per month. The average APACHEE II is 22 and mortality is 17%.

All Acute Chest Syndrome (ACS)/Vaso-oclusive crisis (VOC) or pulmonary infections were studied in 157 adult SCD patients admitted to the intensive care unit (ICU). All SCD whose haemoglobin S was more than 30% has received exchange transfusion. The average APACHEE II score in chest infection group was 21.3 in ACS/VOC group was 20.7.

Results
Out of 157 SCD admitted to ICU 98 patients were admitted with chest infections and 59 patients were admitted with chest infections, in chest infection group 41%(31) patients required ventilator support as compared to patients admitted with ACS/VOC 12%(7) p-value 0.0003 and 32% patients required blood exchange transfusion and 85%(50) in ACS/VOC group p-value<0.0001, the intotrpes require in chest infection group was 39% (38) and ACS/VOC group was 0% (0) the p-value <0.0001.The mortality in chest infection group was 19% where as in ACS/VOC group 5%, the p-value 0.0258 (Table 1).
Table 1

Comparison SCD* Patients Admitted to ICU With Chest Infections Verses ACS/VOC

 

On ventilator

X-change

Inotrops

Mortality

Chest infections

41%

32%

39%

19%

98 Patients

    

ACS/VOC

12%

85%

0%

5%

59 patients

    

p-value

0.0003

<0.0001

<0.0001

0.258

Conclusion

SCD patients admitted to ICU with severe chest infection need extra precautions as compare to SCD admitted with ACS/VOC. Early exchange transfusion if haemoglobin S percent is >30 is beneficial and improve the mortality.

0491

The Diagnostic Role of Biomarkers in Meningitis for Neurosurgical Patients in ICU

C. Diakaki*, E. Karetsi O. Apostolopoulou, P. Zigoulis, I. Tsilioni, T. Kiropoulos, E. Zakynthinos

Critical Care Department, 2Pulmonary Department, University hospital of Larissa, Larissa, Greece

Introduction

The diagnosis of meningitis in neurosurgical patients is difficult, as the standard markers in cerebrospinal fluid (CSF) are not always diagnostic. In order to find a reliable marker for the diagnosis of meningitis in these patients we measured interleukin-6 (IL-6) and vascular endothelial growth factor (VEGF) in CSF and we compared them with standard markers (CSF leucocytosis, glucose, proteins).

Methods

17 neurosurgical patients (13men) with intracerebral or/and subarrachnoid hemorrhage, aged 52±15 years, septic with clinical suspicion of meningitis were enrolled. All patients had intraventricular catheters for external drainage of CSF. 8 had proven meningitis. Samples of CSF and blood were taken at time of suspected meningitis and in meningitis group at time of clinical improvement also. Leucocytes, glucose and proteins were measured in blood and CSF. The samples were centifigurated and frozed in − 4oC and IL-6 and VEGF were measured in CSF and serum (the upper limit measured for IL-6 was 5000 pg/ml).

Results

IL-6 concentrations in CSF were extremely high in all patients (3649±1914pg/ml). IL-6 levels were higher in CSF from patients with meningitis compared with those from patients without meningitis (5000±0 versus 2449±1971 pg/ml, p=0.002). In all patients CSF levels of IL-6 were higher than plasma levels (p<0.0001) and in meningitis group were over 100fold higher, (p<0.0001). In meningitis group, CSF IL-6 levels decreased significantly as patients improved (p=0.03). In contrast VEGF levels performed high in serum as well as in CSF (842±166pg/ml vs 130±47pg/ml), but did not differ in two groups. Between two groups (meningitis and no meningitis group) there was no difference in CSF pleocytosis (523±318,7 vs 273±137,9) CSF glucose (57,1±14,7 vs 85,2±11,6mg/dl) and CSF proteins (275±49,5 vs 225,7±59,4gr/dl).

Conclusion

It seems that CSF pleocytosis and glucose are not efficient markers to establish meningitis in neurosurgical patients. CSF IL-6 increases in meningitis, much more than plasma levels, suggesting local production, predicts meningitis with 100% sensitivity 64% specificity and decreases with clinical improvement. More data may help to consider IL- 6 as a sensitive and specific marker for postsurgical meningitis in the difficult neurosurgical patient.

0492

Microbial Diagnosis in Severe Human Sepsis: Multiplex PCR Vs. Blood Culture — a Prospective Study

S. Sachse*1, K. Schmidt1, R. Schmitz2, M. Lehmann2, F. Bloos3, S. Russwurm2, K. Reinhart3, E. Straube1

1Institut of Medical Microbiology, Friedrich Schiller University, 2-, SIRS-Lab GmbH, 3Department of Anaesthesiology and Intensive Care Therapy, University Hospital, Jena, Germany

Introduction

Sepsis is a life-threatening disease that results from excessive host responses to microbial infections with a rapidly increasing mortality per hour after infection. Therefore, the causative pathogens and/or antibiotic resistances are to be determined as early as possible. The traditional gold standard blood culture (BC) usually takes at least 2–3 days. Facing the time-to-result, nucleic acid based technologies (NAT) are increasingly desirable. The aim of this study was to compare the sensitivities and specificities of BC with a novel molecular-biological pathogen detection method. The NAT-system includes a sample preparation technology, which enables the enrichment of bacterial and fungal DNA out of whole DNA-samples. Cell lysis within whole blood and subsequent DNA preparation are followed by multiplex PCR to rapidly identify 99% of clinically relevant sepsis causative pathogens. At the same time, five of the most common antibiotic resistance genes are detected.

Methods

In a prospective study (6/2006–7/2007), a totality of 35 peripheral blood samples from 25 patients (septic group) were consecutively drawn for the evaluation of possible sepsis by culture (e.g. BC) and molecular diagnosis. 32 whole blood samples were taken from 22 patients (control group) without the clinical indication of sepsis. The DNA preparation was performed according to a standard DNA extraction protocol combined with protein-based affinity chromatography (LOOXSTER®, SIRS-Lab). All purified DNA specimens were analyzed by multiplex PCR (VYOO®, SIRS-Lab). The blood samples were incubated in an automated blood culture device (BacTAlert®, BioMerieux). Positive isolates from BC were determined for identification and antimicrobial susceptibility by standard laboratory procedures.

Results

In the septic group, 20% of the patients were tested BC positive with none of them belonging to the positive 28 day mortality group. In the 58% NAT positive patients, four non-survivors (80%) were included. The NAT method was negative as expected for all samples of the control group patients. In contrast, 23% BC were positive in this group. Concerning the clinical microbiological evaluation, the BC positive controls can be considered as false positive results. The positive results from blood cultures were not available before 72 hours whereas the NAT method delivered results within 6–8 hours.

Conclusion

Compared to the gold standard and other culture-based methods, the NAT method was in significantly less time more sensitive and specific. This might be a milestone allowing “same day diagnosis” of sepsis and SIRS as well. Further, the results revealed the expanding potential of NAT-based methods for the diagnosis of systemic microbial infections.

Grant Acknowledgement

This study was supported by grants from the TMWFK.

0493

IL-6 Gene Transcription in Infection and Severe Sepsis Displays Distinctive Characteristics

M. White*1, M. O. Dwyer1, R. Grealy1, P. Stordeur2, D. P. Kelleher3, R. McManus3, T. Ryan1

1Department of Anaesthesia, St James Hospital, Dublin, Ireland, 2Service d'Immunobiologie-Hemobiologie-Transfusion, Hopital Erasme, Universite Libre de Bruxelles, Brussels, Belgium, 3Department of Clinical Medicine, Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland

Introduction

Interleukin 6 (IL-6) is a cytokine with diverse biological activities. IL-6 has been reported as a possible diagnostic marker for the presence of bacteraemia and serves as a marker of disease severity in sepsis [1]. Indeed recently adverse outcome in community-acquired pneumonia has been reported in the context of elevated IL-6 levels [2]. We investigated IL6 gene transcription in peripheral blood leukocytes in three study groups, 58 patients with severe sepsis, 15 patients with a gram-negative infection but without sepsis and 10 healthy controls.

Methods

This is a prospective observational study. Blood samples were collected from healthy controls at 1 time point. In bacteraemic patients, blood sampling was carried out within 24 hours of the positive blood culture being reported. In 58 patients presenting with severe sepsis blood sampling was carried on day 1 of intensive care admission and on day 7 in survivors. Mononuclear cells were isolated, and total RNA was extracted. IL-6 mRNA was quantified using the technique of quantitative real-time polymerase chain reaction (qRTPCR). All values are stated as median and inter-quartile range. Between group comparisons was performed by Wilcoxon rank sum test.

Results

In the sepsis group 19 patients died. IL-6 mRNA copy numbers was significantly reduced in the bacteraemic group (2.96*106; 3.59*106–2.66*106) compared to controls (3.73*106; 4.12*106 – 3.48*106), (p = 0.012), and was significantly reduced in the sepsis group (3.73*106; 4.47*106 – 3.28*106) compared to the bacteraemic group (2.96*106; 3.59*106–2.66*106) (p = 0.0032). There was no association between IL6 mRNA copy numbers and outcome measures such as mortality, the presence of shock after prolonged sepsis, duration of vasopressor support, duration of mechanical ventilation and duration of intensive care stay.

Conclusion

The human host response to infection is related to a distinct pattern of IL-6 gene transcription. There is deficient IL-6 gene transcription in patients who tolerated infection with impunity. Down-regulation of IL6 in the setting of infection may be reflective of an appropriate bactericidal response.

Grant Acknowledgement

This study was in part funded by the Intensive Care Society Ireland.

0494

Procalcitonin (PCT) Serum Level in Severe Sepsis and Septic Shock Induced by Multiresistant Pathogens

B. Tamowicz*1, K. Stefanska-Wronka1, J. Tyczka2, T. Albrecht1, A. Mikstacki1

1Department of Anaesthesiology and Intensive Therapy, Karol Marcinkowski University of Medical Sciences in Poznan, 2Department of Anaesthesiology and Intensive Therapy, Pulmonary Diseases and Thoracic Surgery Center in Poznan, Poznan, Poland

Introduction

Procalcitonin (PCT) as a marker for bacterial infections facilitates an early diagnostics and appropriate therapeutic decisions, as well as informs about the course and prognosis of the disease1. The aim of the study was to evaluate the usefulness of the PCT serum level test in sepsis induced by multiresistant pathogens.

Methods

A prospective study included 35 patients aged 18–75 years, meeting the criteria of severe sepsis or septic shock, treated between July and December 2007, divided into two groups: I — patients with sepsis induced by multiresistant pathogens, II — patients without isolation of multiresistant pathogens. The following parameters were assessed: APACHE II and SOFA scores; lactate; PCT levels (VIDAS BRAHMS PCT) on the 1st, 3rd, 5th day of sepsis therapy; the length of stay (LOS) and mortality. A microbiological analysis of material obtained from the blood, central venous catheters, lower respiratory tract, pleural and abdominal cavity was performed. Data were analysed using the nonparametric Mann-Whitney test and chi-square test.

Results

Demographic data in both groups were comparable. The following pathogens were isolated in group I (n=19): MRCNS, MRSA, VRE, ESBL-producing Gram-negative rods; and in group II (n=16): Gram-negative bacteria, MSSA, MSCNS, E. faecalis. The PCT serum level in group I was higher than in group II (88.3±62.13ng/mL vs.12.0±8.78ng/mL; p <0.05). A statistically significant difference in LOS and mortality between group I and II was noted (21.0±8.56 vs.14.2±4.77 days, p<0.05; 41.6% vs.25.0%, p<0.05; respectively). The PCT serum level was found to be significantly higher in septic shock vs. severe sepsis patients (p<0.05).

Conclusion

The occurrence of multiresistant pathogens significantly affects the PCT serum level in severe sepsis and septic shock. PCT is a sensitive marker of the severity of sepsis.

0495

Temporal Changes in Tissue Oxygenation in Predicted Survivors and Non-Survivors of Faecal Peritonitis

A. Dyson*, A. Rudiger, M. Singer

Department of Medicine and Wolfson Institute for Biomedical Research, University College London, London, United Kingdom

Introduction

We recently reported that tissue oxygen tension (tPO2), the balance of local O2 supply/demand, is variably affected in four different organs (kidney cortex, liver, muscle, bladder) during endotoxaemic sepsis (1) and early (6h) faecal peritonitis (2). We sought to measure temporal changes in circulatory pathophysiology and tPO2 in these organs in predicted survivors and non-survivors of faecal peritonitis.

Methods

Fluid-resuscitated male Wistar rats received an intraperitoneal injection of faecal slurry, modelling faecal peritonitis. At 6h post-insult rats were anaesthetised and echocardiography performed. Rats were divided into predicted survivors (mild sepsis) and non-survivors (severe sepsis) based on a stroke volume greater or less than 0.14 ml, respectively (3). At either 6h or 24h post-sepsis, animals received a tracheostomy and tPO2 probes placed in muscle, between two liver lobes, within the bladder lumen and left renal cortex. After 30-min stabilisation, recordings were made of tPO2, and cardiac output (CO) by echocardiography (Vivid 7, GE Healthcare, UK). Blood gas analysis was performed to determine lactate levels. Comparisons were made against sham-operated animals that underwent instrumentation but received no septic insult.

Results
Data shown as mean (± SE), *p<0.05 between sham (n=8) & sepsis (n=8). Statistics: 2-way RM-ANOVA and post-hoc Tukey's test.

Time/Group

CO (ml/min)

Global DO2 (1/min)

Muscle tPO2 (kPa)

Bladder tPO2 (kPa)

Liver tPO2 (kPa)

Kidney tPO2 (kPa)

Lactate (mmol/1)

6h Sham

114 (5)

16 (1)

5.3 (0.7)

6.9 (0.4)

2.7 (0.3)

2.2 (0.4)

1.6 (0.2)

6h Mild

77 (3)*

15 (1)

4.5 (0.5)

7.5 (0.2)

2.0 (0.5)

1.6 (0.2)

2.1 (0.4)

6h Severe

55 (2)*

11 (1)

3.2 (0.6)*

7.5 (0.4)

0.5 (0.1)*

1.1 (0.3)*

3.5 (0.2)*

24h Sham

115 (5)

18 (1)

5.1 (0.7)

8.0 (0.3)

2.4 (0.4)

2.5 (0.3)

2.0 (0.2)

24h Mild

120 (8)

18 (1)

4.7 (1.0)

8.3 (0.6)

2.5 (0.7)

3.0 (0.7)

3.9 (0.4)*

24h Severe

119 (13)

16 (2)

4.7 (1.2)

6.7 (0.5)

2.0 (0.7)

2.1 (0.5)

4.7 (1.0)*

Conclusion

Early (6h) sepsis caused circulatory dysfunction and an oxygen supply/demand imbalance that varied between organ beds, with liver most affected, followed by muscle and kidney. During established (24h) sepsis, circulatory function and tPO2 were restored although animals showed signs of clinical illness and continued hyperlactataemia. The data support the hypothesis that early circulatory impairment followed by later mitochondrial dysfunction may combine to produce multi-organ failure in sepsis.

Grant Acknowledgement

This work is supported by the MR Council (UK).

0496

Differential Gene Expression Profile in a 24-Hour Porcine Model of Fluid-Resuscitated Fecal Peritonitis

J. Wauters*1, H. van Malenstein2, P. Van Hummelen3, H. Cauwenberghs1, K. Reynders1, M. Van Wambeke1, L. Langouche1, A. Wilmer1, J. van Pelt2

1ICM, 2Liver Facility, 3VIB-MAF, UZ Gasthuisberg, Leuven, Belgium

Introduction

The molecular and pathophysiological mechanisms underlying sepsisrelated liver dysfunction are not yet fully understood. To elucidate the pathways involved we investigated gene expression by microarray in a clinically relevant porcine model of fluid-resuscitated septic shock.

Methods

Anesthetized and ventilated pigs (40±3kg) were randomly assigned to fluid-resuscitated septic shock by fecal peritonitis (S, n=6) or control (C, n=6). In vivo liver samples were collected at baseline (BL) and 24h for analysis of mRNA expression by Affimetrix microarray (S:n=2, C:n=2). Changes in gene expression (ratio in base 2) between BL and 24h and at 24h were analyzed by GeneMath 3.5. To identify the affected molecular pathways, gene expression data were mapped on a pathway database using MapFinder 2.1.

Results

Septic pigs developed a normotensive, hyperdynamic circulation. In sepsis 2495 genes had significantly changed (>2-fold up or <0.5-fold down) between 24h and BL versus only 1955 genes in the control group. After pathway mapping, we identified pathways of inflammation, apoptosis and cell death being upregulated in sepsis, which were not altered in controls (p<0.05). 24 h after induction, sepsis had limited effect on expression of mitochondrial genes. Detoxification by P450 cytochrome enzymes was upregulated in controls while septic pigs showed no change or even downregulation (p<0.05).

Conclusion

Fluid-resuscitated sepsis is accompanied by an upregulation of inflammation, apoptosis and cell death, while hepatic detoxification is suppressed.

0497

Oxygen Challenge Test Reveals Macro- and Microcirculatory Dyssynchrony in Predicted Survivors and Non-Survivors of Faecal Peritonitis

A. Dyson*, M. Singer

Department of Medicine and Wolfson Institute for Biomedical Research, University College London, London, United Kingdom

Introduction

An increment in transcutanoeus tissue oxygen tension (tPO2) of <21 mmHg following 100% O2 administration was associated with a poor outcome in septic patients (1). We applied this oxygen challenge test (OCT) in our rat model of faecal peritonitis in peripheral (muscle, bladder) and deep (liver, renal cortex) organ beds.

Methods

Fluid-resuscitated male Wistar rats received an intraperitoneal injection of faecal slurry, modelling faecal peritonitis. At 6h post-insult rats were anaesthetised and echocardiography performed. Rats were divided into predicted survivors and non-survivors based on a stroke volume greater or less than 0.14 ml, respectively (2). Animals then received a tracheostomy and had tPO2 probes placed in muscle, between two liver lobes, within the bladder lumen and left renal cortex to measure the regional oxygen supply/demand balance. After confirmation of normovolaemia, the fraction of inspired oxygen (FiO2) was increased to 0.3, 0.6 and 1.0 at 15 minute intervals. In separate experiments, this was repeated at 24h postinsult. Comparisons were made at both timepoints against sham-operated animals that underwent instrumentation but received no septic insult.

Results

Data shown as mean (± SE), *p<0.05 between sham (open circles; n=8), good-prognosis sepsis (light grey triangles; n=8) and poor-prognosis sepsis (black diamonds; n=8). Statistics: 2-way RM-ANOVA + Tukey's test. Open image in new window

Conclusion

The OCT revealed a reduced response to hyperoxia in predicted nonsurvivors after 6h of sepsis, suggestive of microcirculatory dysfunction in this early phase of sepsis, Apart from the renal cortex, this had recovered by 24h even though clinical status, organ function and lactataemia had significantly worsened (data not shown). The OCT could provide a simple method to assess microvascular perfusion and its responsiveness.

Grant Acknowledgement

This work is supported by the MRCouncil (UK).

0498

Procalcitonin as an Early Diagnostic Marker for Septic Shock

C. Montoya*, E. Monares, C. Sanchez, M. Poblano, J. Aguirre, C. Olvera, J. Martinez, J. Franco

Intensive Care Unit, The American British Cowdray Medical Center, Mexico City, Mexico

Introduction

Procalcitonin (PCT) is more a “follow-up” prognosis marker than an early diagnostic tool. In early stages of shock, it may differentiate between septic and other causes of shock allowing an early treatment according to it. The objective of our study was to determine PCT usefulness as an early diagnostic marker for septic shock. METHODS. Prospective, longitudinal, observational study from October 2007 to March 2008. 108 patients with an initial diagnosis of shock were included. PCT was determined at admittance and 5 days later. Sedimentation Rate (SR), C-reactive protein (CRP), leukocytes, APACHE II Score and SOFA score were also obtained. Results are expressed in median (25–75th interquartile interval) and analysis was made with U Mann-Whitney and Pearson chi square as required and ROC curves created to determine predictive values.

Results

At admittance, patients with septic shock presented higher PCT, SR, CRP and leukocyte values than other shock from other causes: PCT 4 (3–5) vs. 0 (0–0) of septic shock vs. others respectively (p<0.01); SR 20(15–25) in septic shock vs. 9 (7–14) and p value <0.01; CRP 15 (10–19) vs. 7(4–10) respectively (p<0.01); Leukocytes 11.3 (9–13.3) vs. 6 (4.3–8.1) and p<0.01. APACHE II score for both groups was different: 22 (18–24) of septic shock vs. 12 (10–15) (p<0.01) but SOFA score for both patients was identical: 10(10–12). As a diagnostic tool for septic shock, when compared with other markers, PCT resulted to be 100% sensitive, 89% specific with positive predictive value of 86%, negative predictive value of 100% and an area under the curve of 0.996.

Conclusion

As expected, this study demonstrates that elevations of PCT in shock patients at admittance, is the best way to differentiate between septic shock and other causes of shock.

Poster Sessions Creative nursing: 0499–0511

0499

Successful Implementation of a Delirium Assessment Tool in the ICU

M. v. d. Boogaard*1, L. Schoonhoven2, G. Roodbol3, T. v. Achterberg2, H. v. d. Hoeven1, P. Pickkers1

1Dept. of Intensive care, 2Dept. Quality of Care, 3Dept. of Psychiatry, Radboud University Medical Centre Nijmegen, Nijmegen, Netherlands

Introduction

In critically ill patients delirium is a serious and frequent disorder. As early recognition of the delirium is important for adequate treatment routine screening of the patients is necessary. To be meaningful for practice this screening should be performed frequently and interrater reliability (IRR) should be high. The aim of our study was to implement the validated confusion assessment method-ICU (CAM-ICU) in our ICU, aiming at a scorings rate of >80%, an IRR score of >0.80 and a high level of knowledge concerning delirium.

Methods

The implementation strategy focused on potential barriers for screening with the CAM-ICU, i.e. lack of knowledge and availability of the assessment tool. We integrated the CAM-ICU in our patient data management system (PDMS). Necessary testing tools became available at every bed. Every patient had to be assessed once a shift, if not a pop-up on the PDMS-screen appeared. We used ‘delirium key-nurses’ for further instruction and introduction in the ICU. In addition, instruction posters for nurses and the medical staff were distributed. Feedback about results and performance of the CAM-ICU was supplied weekly by mail and during clinical meetings. Scoring rate was calculated by the percentage scored patients per day. IRR was assessed by calculating the Cohen's Kappa between the score on the CAM-ICU by the ICU nurse with the score by an expert psychiatric nurse. To assess the knowledge level (on a scale for 0–10), all ICU nurses performed a test prior to group training and a post-test 4 months later.

Results

During 4 months, 78 patients were screened for IRR. In the first month this was 0.78 (N=25, 95%CI: 0,5–1,0). After intensive training on the job and feedback twice a week, the IRR increased relevantly to 0.89 (N=47, 95%CI: 0,75–1,0). The scoring rate of the CAM-ICU increased from 77% to 92%. The delirium knowledge level of the nurses increased from 6.2±1.7 to 7.4±1.2 (p=0.0001).

Conclusion

Tailoring our implementation strategy to the ICU was successful. The main goals were achieved in a relatively short time. Early recognition of delirium with the CAM-ICU has become a standard component of the daily care of our ICU nurses and contributes to the quality of care.

0501

Utilization of a Nursing “journal Club” to Disseminate Evidence-Based Practice

L. DeStefano*

Critical Care Services, Saddleback Memorial Medical Center, Laguna Hills, United States

Introduction

Nursing practice based on evidence improves patient outcomes. Members of an American Association of Critical Care Nursing (AACN) local California chapter in the United States reported significant barriers that prevent them from incorporating research findings into their practice. Subsequently, a masters-prepared advance practice nurse developed a unique “Journal Club” to reduce these barriers and increase awareness of Evidence- Based Practice (EBP).

Methods

In 2002, a monthly Journal Club was initiated to engage acute and critical care nurses from local regional hospitals in discussion of current, recently published research findings. The monthly topics were selected based on clinical relevance, educational needs assessment, and requests by members. The monthly Journal Club meetings take place at the same time and place in a local restaurant, usually with vendor sponsored dinners. Methods to decrease the inherent barriers related to critiquing academic research included a chapter led critique and review of research findings. This method decreased anxiety related to lack of skills to critique or synthesize the literature and difficulties interpreting the findings by many nurses. Chapter facilitated discussion of complex research by placing less emphasis on interpretation of statistics and research methodologies and focused on adaptation of findings into real clinical practice. The confidential, non-threatening, supportive environment promoted open, frank discussion and diminished intimidation.

Results

Over the past five year period, the monthly Journal Club attendance increased from 4–6 members in 2002 to 40–50 members currently and continues to grow. Several members travel over 100km to attend. Community subgroups of interest have also been established to review sedation protocols and other hot topics prompted by an initial Journal Club discussion. Many hospitals are willing to “share” policies, guidelines, and tools that they have created with other nurses/institutions in the community to avoid having to “reinvent the wheel” to get things rolling in their hospital of practice.

Conclusion

The regional Journal Club has grown to be a huge success. The creator of the Journal Club and the Board of Directors of the AACN chapter have received numerous member anecdotal accounts of the influence meetings have played in their professional lives by sharpening their professional interest and increasing awareness. The use of research in everyday practice is promoted by simplification of complex research topics and promotes replacing practice based on tradition with practice based on evidence and results in positive patient outcomes.

0502

Improving Postoperative Pain Relief by Extending Services Beyond the ICU Walls

M. Ulstein*, M. Skagestad, G. Bjømå

Intensive care department, Stavanger University Hospital, Stavanger, Norway

Introduction

Postoperative pain treatment is often initiated during operation, followed up in the recovery room/ICU, and to be continued in a ward. Information, observation and understanding of pain treatment protocols has become a multidisciplinary matter (1,2). In order to secure uniform postoperative pain treatment, we have implemented a system based on extending the ICU protocols to the wards: A) Dedicating a “pain nurse” from the ICU to follow up on patients after they left the ICU/recovery room. B) Developing uniform procedures for treatment of postoperative pain. C) Introducing the “pain nurses” in a liaison role between doctors and nurses in all wards. D) Implement a uniform pain scoring form in all wards. The objective of this investigation is to evaluate the system we have implemented. Since epidural pain relief is associated with both technical problems and potentially serious side effects (3,4), we used this group of patients in this study.

Methods

Between 2002 and 2006 2062 patients received an epidural infusion as postoperative pain relief. A standardized mixture for epidural pain relief consisting of Bupivacain, Adrenaline and Fentanyl was given to all patients. All data was entered into the form described above by the nurses responsible for the patient on the wards. In this study we have retrieved the following information: Age, type of operation, level of catheter insertion, duration of treatment, average dose (ml/hour), total dose, reason for unscheduled termination of treatment, side effects and patient satisfaction — whether the overall treatment was perceived as poor, acceptable, good or excellent.

Results

From 2002 to 2006 the number of patients indicating that their pain relief was excellent rose from 32% to 73%. The number of patients reporting that their treatment was poor fell from 12% to 3%. In 2002 13% of the patients had to terminate the treatment due to technical problems, in 2006 only 5% of them had to do the same. In 2002 12% of the patients had to terminate the treatment due to side effects. In 2006 only 5% had the treatment stopped due to this.

Conclusion

We conclude that the improvement we have seen in the postoperative treatment with epidural infusion is due to the organizational changes we have introduced. We believe that informing those closest to the patients — the nurses on the wards — is the single most important initiative we introduced in order to improve the postoperative pain treatment. The effect of this has been a better communication and understanding of the patients needs. A more vigilant set of observations and hence swifter response from the staff when the pain treatment was not satisfactory, or when side effects occur. It is also likely that similar improvements can be obtained in other areas where ICU treatment strategies are supposed to continue after the patients have left the ICU. Well-established and documented treatments may result in sub optimal results due to lack of communication and knowledge, and poor implementation of protocols between the ICU and all other involved personnel and departments.

0503

Service Gap: Pregnant and Postnatal Women in ICU

W. E. Pollock*1, N. Harley2, S. Nelson3

1School of Nursing and Social Work, The University of Melbourne, Carlton, 2Intensive Care Unit, Royal Melbourne Hospital, Parkville, Australia, 3Faculty of Nursing, University of Toronto, Toronto, Canada

Introduction

Pregnant and postnatal women in ICU require specialist care that may not be readily available within the critical care team, with midwifery increasingly recognised as a separate profession to nursing. The aim of this study was to examine the health services provided to parturient women in ICU to identify if a service gap existed.

Methods

A prospective cross-sectional study was conducted in four tertiary-level intensive care units (ICU) in Melbourne, Australia. Only one ICU was in a hospital that provided obstetric services. All pregnant and postnatal women admitted to ICU were eligible for the study. Nurses formed core research teams and assisted with the auto-enrolment of eligible women, time-critical data collection and the obtaining of consent. Additional data were gathered from the medical history following discharge. Data were entered in SPSS (v12) and analysed.

Results

Of the 43 known eligible women, 35 were enrolled, and 33 gave consent; 8 were pregnant. Two of the eight pregnant women were 8 weeks’ and 14 weeks’ gestation respectively. The gestation of the other six pregnant women ranged from 27 to 36 weeks'. Of the six women with a viable pregnancy, three had a fetal ultrasound when in ICU, five out of six had at least one cardiotocograph and one woman had no fetal assessment in ICU. None of the pregnant women gave birth during the ICU admission. Of the 25 postnatal women, 92% were admitted in the first 24 hours following birth. Four women saw their baby whilst in ICU, with the remaining women needing to wait until they were transferred out of ICU. An obstetric condition was the principle diagnosis for two thirds of women. However, pregnant and postnatal women were seen by 23 different types of medical specialties whilst in ICU. The mean number of specialties seen by individual women was 3.6 (median 4.0). 53% of women were seen by an obstetrician when in ICU: 100% of women with private health insurance (n=11) compared with only 29% of public patients (p<.05). Having the ICU and obstetric services on the one-site increased the chance of the woman being seen by an obstetrician, but did not guarantee it. Only 30% of the pregnant and postnatal women in ICU were seen by a midwife. Whether the woman was pregnant or postnatal did not affect her chance of being seen by an obstetrician or midwife in ICU. 70% (n=23) of women were transferred by ambulance between hospitals to access the ICU bed.

Conclusion

Women received fragmented care from many health professionals and service providers. A service gap existed in the care provided to the pregnant and postnatal women admitted to ICU and articulated processes are required if pregnant and postnatal women are to have access to maternity care in ICU.

0504

Fever Magnitude As Predictor of Patient Acuity in the ICU

P. Kiekkas*1, K. S. Filos1, M. Karanikolas1, D. Aretha1, G. I. Baltopoulos2

1Anaesthesiology and Critical Care Medicine, Patras University Hospital, Patras, 2School of Nursing, University of Athens, Athens, Greece

Introduction

Prediction of ICU patient (pt) acuity is very important because inadequate coverage of pt care needs is associated with adverse outcomes. Since clinical severity is highly correlated with pt care needs1, this study compares fever magnitude vs. pt clinical severity as predictors of ICU pt acuity.

Methods

We prospectively studied 89 pts with fever (defined as tympanic temperature >38.3°C) admitted to the medical-surgical ICU of an academic hospital, in a 6-month period. Pt acuity was assessed daily with TISS-28. Clinical severity on admission was assessed with APACHE II score. Receiver operating characteristics (ROC) curve analysis was used to evaluate the discriminative power of fever magnitude on high pt acuity (mean daily TISS-28 >34 points).

Results

67 of 89 pts were male, and 48 were medical. Mean (±SE) age was 50.4±2.0 years and mean APACHE II score was 14.3±0.6 points. 35 pts were of high acuity. ROC analysis showed that, compared to APACHE II, fever magnitude was a better predictor of high pt acuity, although the difference did not reach significance (area under the curve: 0.73±0.06 vs 0.61±0.06, p=0.114). Temperature >39.1°C was selected as fever magnitude threshold by ROC analysis (sensitivity 71.4%, specificity 75.1%). Open image in new window

Conclusion

Fever magnitude is a good predictor of high pt acuity in the ICU, and can be used instead of APACHE II for properly allocating nursing manpower among febrile pts.

0505

Nursing Workload Based on Nursing Activities Score in Groups of Surgical and Non-Surgical Patients in Polish Intensive Care Units

E. K. Cudak*, D. Dyk, K. Kabacinska

Department of Anaesthesiological and Intensive Care Nursing, Poznan University of Medical Sciences, Poznan, Poland

Introduction

Using nursing time in intensive care units is connected with measuring nursing workload. Its aim is to determine optimal level of staffing in order to ensure high quality of medical services.

The purpose of the paper is to present nursing workload in two patients groups: surgical and non-surgical.

Methods

The studied group comprised 314 patients treated in 5 intensive care units with different treatment profiles [1]. The study applied prospective analysis of nursing workload among surgical and non-surgical patients. The methodological tool was NAS — Nursing Activities Score [2]. For evaluation of the severity of the patient's clinical condition APACHE II was used. Research techniques included participant-observation and the analysis of patients’ medical documentation and the analysis of medical staff documentation.

Results

In the studied group there were 176 patients treated with various surgical methods (56%) oraz 138 non-surgical patients (44%). Clinical severity condition among surgical patients and non-surgical patients was slightly different and did not show correlation with nursing workload (p>0,005). The median in APACHE II for surgical patients was 20 (1 – 36) and for non-surgical patients it was 21 (3 – 42). Nursing workload expressed with a median of scores in NAS for surgical patients was 69,7; 43,6–130,9, while for non-surgical patients it was 74; 12–150,3 (p>0,05). Optimal level of nurse staffing in the studied units expressed by nurse:patient ratio based on NAS was 1:1,2, while the real ratio nurse:patient was 1:1,3 – 1:2.

Conclusion

No correlation between the patient's clinical condition and nursing workload was observed. No statistically significant differences were found in nursing workload in surgical and nonsurgical group.

0506

Professional Autonomy and Job Satisfaction: A Questionnaire Survey of Hospital Critical Care Nurses in Mainland Greece

K. Iliopoulou*

Intensive Care Unit, 401 Military Hospital, Athens, Greece

Introduction

Professional autonomy is generally considered a highly desirable nursing attribute and a major factor of nurse job satisfaction. In the critical care environment a high level of accountability, responsibility and autonomy are required in order to optimize outcomes of critical unstable patients.

Methods

A cross-sectional survey was conducted in 16 public hospitals located in Athens, from mid Apr to the end of July 2007. 302 critical care nurses completed questionnaires, yielding a response rate of 70%.

Results

Overall Greek critical care nurses reported acting moderately autonomously with a mean score of 165.4 (SD=24.6) compared with the possible range from 60 to 240. Respondents’ age, gender, level of appointment, completion of clinical specialty, length of critical care experience, shift and membership of a professional organization were found to be significantly associated with reported autonomy (p<0.005). A positive moderate association was found between reported autonomy, job satisfaction and role conflict and role ambiguity (r=0.331, p<0.001 r=0.324, p<0.001 respectively). No significant association was found between job satisfaction and reported role conflict and role ambiguity (r=0.047,p=0.411).

Conclusion

Greek critical care nurses’ autonomy appears to allow further support and enhancement if the nurses are to realize their full professional role.

Grant Acknowledgement

This project was funded by the Military Medical Corps of Greece.

0507

The Nurses Role in Detecting and Managing the Deteriorating Ward Patient: A Systematic Review

M. Odell*

Critical Care, Royal Berkshire NHS Foundation Trust, Reading, United Kingdom

Introduction

Ward patients can experience unexpected physiological deterioration that can lead to critical outcomes and death. This deterioration can be signalled in the patients physiological signs that are monitored by ward nurses as part of routine practice. In many cases these physiological signs have been missed, neglected or poorly managed (12). Systems designed to track patients physiological signs and trigger referral to an expert team have been widely implemented. Proving effectiveness of these systems has been problematical, partly due to the reliance on a timely and robust system of observation monitoring, plus compliance by the ward teams in the referral process to response teams. It is vital that we understand these ward processes. A systematic literature review is presented that critically evaluates research describing or appraising the nursing practice of taking and recording observations in the detection and management of the deteriorating ward patient.

Methods

Eight data bases were systematically searched from 1990 to 2006, plus citations from key articles and reports. Experts in the field were contacted. All study designs were incorporated. Eligible studies were assessed for quality using an assessment tool for both qualitative and quantitative studies. The findings from the studies were critically evaluated using a narrative technique.

Results

Fifteen studies were eligible from 740 titles. Study quality varied. Findings are presented using the Record, Recognise, Report and Rescue (4R) model that outlines a series of actions that characterises the optimal process of detection, recognition and appropriate management of the deteriorating ward patient. Complex dichotomies exist in each of the 4R model categories that show ward nurses struggling to adequately detect and manage the deteriorating ward patient, hampered by inexperience, lack of skill and excessive workloads. Where support systems are in place, they require skilled and complex collaboration with medical staff and can be inconsistently activated and utilised.

Conclusion

Effective detection and mangement of patient deterioration relies on robust patient observation practice and adherence to track and trigger referral triggers. In order to see the benefit of these systems we need to understand the issues that enhance and inhibit the practice of recording, recognising, reporting and rescuing (4Rs) by ward nurses. We can then find ways to better equip wards to deal with these issues and improve patient outcome.

0508

Medical Art Therapy: Influence on Quality of Life and Compliance of Patients with Advanced Heart Failure

N. S. Sela*1, N. B. Baruch2, R. S. Stein1, V. Yaari1, A. P. Pinchas1, A. Asaly1, M.Vaturi1, A. Battler1, T. Ben Gal1

1cardiology Department, Rabin Medical Center Beilinson Hospital, 2cardiology Department, Rabin Medical Center, Petach Tikva, Israel

Introduction

Medical Art Therapy (MAT) enables medically ill people to express their feelings of frustration, fear, anxiety and hopes in art. In heart failure (HF) depression and anxiety are common.

Aim: To evaluate the influence of guided group MAT on the quality of life and compliance to therapy of patients with advanced HF.

Methods

Methods: Between April and July 2006, 20 advanced HF patients were randomly divided: ten in the intervention group (group A) and ten in the control group (group B). First and last visits were individual during which the Ulman Personality Assessment Procedure (a MAT diagnostic tool), the Minnesota Living with HF and compliance questionnaires were recorded. Both groups met weekly for 6 weeks. A MAT therapist (N.T.) guided group A to express their feelings using art material. Group B met for a routine clinical visit only.

Results

Results: Baseline Ulman, compliance and Minnesota scores were similar amongst the two groups. By the end of the study, Ulman score improved significantly in group A compared to group B (95±10 to 82± 14 p=0.0006 vs. 86±10 to 81±12 p=0.5) as was the compliance score (29±11 to 33±13 p=0.05 vs. 36±5 to 36±6 p=0.9). In group A, Minnesota score improved significantly in 7 patients and did not change in 3 while in group B it improved in 2, did not change in 6 and worsened in 2.

Conclusion

Conclusions: MAT improves quality of life and compliance in advanced HF patients and should be a part of the therapeutic arsenal in every HF clinic.

0509

Quantification of Nurse Workload and Discharge Criteria — a Pilot Study in Postoperative Patients

H. B. Alfheim*, B. K. Moskaug, J. H. Laake

Dept. of Anaesthesia and ICM, Rikshospitalet Medical Centre, Oslo, Norway

Introduction

Efficient and humane care of postoperative patients in the postanaesthesia care unit (PACU) calls for objective measures to forecast patient acuity, and enable prediction of nurse workload related to different patient categories. Additionally, discharge criteria will ensure that length of stay does not exceed what is needed.

Methods

Prospective evaluation of two scoring systems; one that measured nurse workload, developed by Kaiser Permanente, USA (1) and another that rated objective discharge criteria, developed by the Danish anaesthesia society (2). Patients were scored upon admission and discharge, and during their stay in the PACU.

Results

Thirty-seven postoperative patients were included in this pilot study. Nurse workload was correlated to length of stay and was primarily determined by the number of drugs and/or doses of drugs administered (p < 0.01, Rsq = 0.37). Twelve patients (32%) fulfilled criteria for discharge upon admission to the PACU. Pain was the major reason why patients did not fulfil discharge criteria at any time. Open image in new window

Conclusion

These scoring systems seem to provide relevant information about nursing workload for postoperative patients. Also, objective discharge criteria may ensure that patients will receive care better adapted to individual needs. This pilot investigation precedes a study in 750 patients, data from which will be available at the ESICM congress.

0510

Educating Intensive Care Nurses in End of Life Issues and Skills

M. Ben-Nun*1, O. Raanan2, J. Benbenishty3

1General Intensive Care, Kaplan Medical Center, Rehovot, 2Nursing School, Sheba Medical Center, Tel Aviv, 3Nursing School, Hadassah, Jerusalem, Israel

Introduction

Intensive Care nurses are well trained in the skills needed for the battle for life. However less education is available in the skills needed when the patient is dying. This presentation describes a workshop developed by nurses for nurses as part of post basic studies for a diploma in ICU nursing.

Methods

A two day seminar was developed to give nursing staff insight, theoritical knowledge and skills to use when dealing with the dying patient and the grieving family. The setting was a nursing school situated in a thousand bed general hospital. The participants were twenty-five trained nurses working in adult intensive care, neurosurgical, internal medical and recovery departments. The facilitators were the nurse tutor intensive care course and two nurses trained in group dynamics and with expertise in End of Life and ethics. Two six hour sessions were held. Use of a group dynamic allowed each nurse to express how they view EOL care in their own personal practice. This opening was followed by a group exercise using collage, an art therapy technique adapted to allow exploration of painful feelings on death and dying. By promoting self awareness into individual attitudes on death the caregiver is strengthened when caring for others. Theoritical knowledge was given next via frontal lectures on ethics, attitudes on EOL and law regarding the rights of the dying. The care of the grieving family was addressed using the Family Needs Inventory(CCFNI). Techniques of breaking bad news were discussed and theories of other disciplines such as Oncology were examined for models of palliative care that could be used in intensive care settings.

Results

A pre-workshop questionaire showed that all nurses felt they had received inadequate training in EOL care. The collage exercise demonstrated that nurses experience considerable distress from the death of patients. Post workshop feedback confirmed that nurses felt they had been given tools to use while dealing with death and dying.

Conclusion

Composure has been described to be the ability of the nurse to be present for the patient at the end of life, and demonstrate concern and compassion whilst changing the aim of saving life to the preservation of the dignity of the patient and the stability of the family. This workshop was given by nurses to help nurses achieve composure, that most difficult of skills, and allow them to be there for their patients when doing is no longer possible.

0511

Patient Diaries in Danish ICUs — Constructing National Guidelines

I. Egerod*

UCSF, department 7331, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark

Introduction

Patient diaries have been used in intensive care units (ICUs) in the Scandinavian countries since the 1980s and in other European countries through the 1990s–2000s as a means to help patients to come to terms with their critical illness, which is often accompanied by distorted memory, lack of recall and nightmares (1). Studies are emerging into the effect of patient diaries as a debriefing instrument in relation to symptoms of posttraumatic stress disorder (PTSD) (2, 3). The purpose of this study was to describe the use of patient diaries in Danish ICUs and to construct a National Guideline for Patient Diaries in the ICU (4).

Methods

The study is descriptive and interactive. The first part of the study was to establish the use of diaries through telephone interviews with nurses at all ICUs using patient diaries, and the second part was to construct a National Guideline by consensus among Danish ICU nurses from the five administrative regions of Denmark. The guideline was constructed using the AGREE-instrument and was submitted to the Department of Health for discussion of ethical and legal issues, as Danish law is very restrictive regarding patient protection.

Results

Patient diaries are used in half of the Danish ICUs and only few local guidelines have been written. The use of patient diaries has been an unsystematic bottom-up initiative by ICU nurses with little management support. On an initial workshop on patient diaries it was clear that consensus could not be reached regarding the application of the intervention. Five regional groups were established and consensus was gradually achieved by negotiation among the groups.

Conclusion

Patient diaries are emerging as an effective instrument for debriefing and prevention of PTSD in patients discharged from ICU. The practice of writing diaries is not homogeneous and needs more conformity in order to be studied further. More studies are needed to establish the role of the family in patient diaries and of follow-up regimes for this group of patients. non|Poster Sessions Alternative strategies for oxygenation and CO2-removal: 0512-0523

0512

Blood Acidification Enhances Carbon Dioxide Removal of Membrane Lung: Preliminary Experimental Data

A. Zanella*, N. Patroniti, S. Isgrò, B. Vergnano, V. Scaravilli, D. Bonacina, L. Castagna, P. Brescia, S. Redaelli, A. Pesenti

Experimental medicine, Milano Bicocca University — San Gerardo Hospital, Monza, Italy

Introduction

Extracorporeal CO2 removal is an effective technique to limit minute ventilation, thus allowing a protective ventilatory strategy even in the most severe ARDS patients. However, since CO2 is driven through the membrane lung only by the partial pressure of the dissolved CO2, which is less than 10% of the total CO2 content, extracorporeal blood flow as high as 2 liters/min are required to remove the total CO2 production of an adult patient. We hypothesized that acidification of the blood entering the ML could convert the CO2 contained in blood as bicarbonate ions to dissolved CO2, hence increasing the partial pressure of CO2 (pCO2). The consequent increase in CO2 transfer though the ML should allow for a reduction of extracorporeal blood flow.

Methods

Six pigs (37±2 Kg) were sedated, intubated, mechanically ventilated and connected to a veno-venous extracorporeal circuit comprising a ML (Quadrox D). Minute ventilation was adjusted during the experiment to maintain a constant arterial pCO2. The extracorporeal blood flow was set at 500 ml/min while the gas flow was 10 l/min of oxygen. A continuous infusion of 0.5 N lactic acid was added to the extracorporeal blood flow before the ML at a rate of 1, 2 and 5 mEq/min for 15 minutes each. After each infusion we waited at least 30 minutes to allow the acid to be metabolized. Each infusion was repeated two times. At each step we obtained the amount of CO2 removed, free hemoglobin and gas analysis from different sites of interest.

Results
The pH of the acid loaded blood decreased as expected, while blood pCO2 increased (see Table). ML CO2 removal (VCO2 ML) increased 11, 23, 70% during acid infusion of respectively 1, 2 and 5 mEq/min. Blood free hemoglobin, a marker of hemolysis, was constantly within normal values.

Acid infusion (mEq/min)

0

1

2

5

Venous drainage

52±9

53±12

55±8

57±8

pCO2 (mmHg)

    

Venous drainage pH

7.440±0.045

7.426±0.055

7.402±0.025

7.355±0.028

Pre ML, acid loaded

52±9

66±15

82±13

132±16

pCO2 (mmHg)

    

Pre ML, acid loaded pH

7.440±0.045

7.310±0.05

7.214±0.048

6.926±0.044

ML VCO2 (ml/min)

108±5

123 ±21

137±21

173±24

Conclusion

Blood acidification at the inlet of a ML, with infusion of 1, 2 and 5 mEq/min of a lactic acid, can increase up to 70%, the CO2 removal capacity of the ML, allowing to remove the equivalent of 80% (173±23 ml/min) of the total CO2 production of an adult man from as low as 500 ml of blood.

Grant Acknowledgement

Supported by MIUR.

0513

Pumpless Arterio-Venous Extracorporeal Interventional Lung Assist in Post-Traumatic Acute Respiratory Distress Syndrome

M. Zimmermann*1, S. Wittmann1, A. Philipp2, M. Nerlich3, T. Bein11Anesthesia and Intensive Care, 2Cardiothoracic Surgery, 3Trauma Surgery, University Hospital, Regensburg, Germany

Introduction

ARDS due to trauma is associated with a high mortality rate. The use of pump-driven extracorporeal membrane oxygenation (ECMO) is reported in severe posttraumatic ARDS, but the complication rate (bleeding, shock) is high. We report on the use of a pumpless interventional lung assist (iLA, Novalung, Hechingen, Germany) establishing an arteriovenous shunt, in which a single-use ultracompact gas exchange system is integrated.

Methods

iLA was inserted in 36 trauma patients suffering from severe ARDS with lifethreatening hypoxemia/hypercapnia. Contraindications were hemodynamic instability (cardiac index <3 l/min/m2) or severe coagulation disorder (platelets <60.000, aPTT >60 sec). The system was started after careful cannulation of femoral artery (15–17 F) and vein (17–19 F).

Results
The median (interquartil ranges) age of patients was 23 (20–31) years. The Injury Severity Score (ISS = 42 [34–59]) indicated a most severe trauma. The implementation of iLA induced a moderate increase in oxygenation and a marked and rapid correction of hypercapnia, allowing a significant decrease in tidal volume and plateau inspiratory pressure (“lung protective strategy”). In 7 patients (19%) moderate complications (transient ischemia of lower limb, bleeding) occurred. The hospital mortality rate was 36%.
Table 1

Effects of iLA in Post-Traumatic Ards (Median Values + Interquartil Ranges)

 

pre iLA

2 hrs after iLA–start

24 hrs after iLA–start

PaO2/FIO2 (mmHg)

69 (49–84)

91 (64–136) **

116 (79–174) **

PaCO2 (mmHg)

66(52–78)

40(35–48) **

39 (31–46) **

art. pH

7,23 (7,17–7,32)

7,39 (7,28–7,46) **

7,45 (7,35–7,48) **

mean art. pressure (mmHg)

71 (65–79)

81 (71–90)

83 (75–98)

norepinephrine (ug/kg/min)

0,2 (0,03–0,4)

0,2 (0,03–0,4)

0,1 (0,04–0,5)

tidal volume (ml/kg/IBW)

6,6 (6,1–7,3)

6,1 (4,7–6,7) *

5,5 (4,6–6,8) **

iLA–flow (l/min)

 

2,0 (1,6–2,2)

1,9 (1,7–2,2)

* = p <0.05; ** = p<0.01; IBW = ideal body weight

Conclusion

iLA might find a place in the scenario of trauma management, since it provides a useful technique in posttraumatic ARDS. Due to effective CO2-removal, iLA is an important tool in realising lung protective ventilation. In our patients, the mortality rate was lower than expected from the ISS.

0514

High Effectiveness of a New Portable Mini-ECMO System for Out-of-Centre ECMO Support

M. Arlt*1, A. Philipp2, S. Voelkel1, M. Hilker2, C. Schmid2, K. Taeger1 p ]1Anesthesiology, 2Cardiothoracic Surgery, University Hospital Regensburg, Regensburg, Germany

Introduction

Severe respiratory and cardiac failure resistant to critical care treatment leads to hypoxemia and death of hypoxic organ failure. New treatment options for hypoxemia are necessary, even for patients primary located in outlying medical facilities. We report our experience with a new developed, portable Mini-ECMO (Extracorporeal Membrane Oxygenation) system for emergency treatment of hypoxemia caused by ARDS, cardiogenic shock or septic shock. We describe out-of-centre emergency percutaneous ECMO implementation and additional air- and ground ambulance transfer for referral centre treatment.

Methods

Between March 2006 and April 2008, we treated 15 adult patients located in outlying medical facilities with the new portable Mini-ECMO system and carried out interhospital transfer on the running system. Diagnosis included ARDS (n = 6), cardiogenic shock (n = 6) and septic shock (n = 3). Bedside cannulation was achieved percutaneously. In respiratory failure, we used a femoro-jugular veno-venous vessel access, in cardiocirculatory failure we implementated the Mini-ECMO system using a femoro-femoral veno-arterial vessel access. The new portable Mini-ECMO system consists of a centrifugal pump and a membrane oxygenator. Because of tip-to-tip heparin bonded circuits, we need no full dose heparinization. The system (27kg) is capable for hand-held use and suitable for storing on a standard gurney. With its inbuilt battery pack and oxyen-supply the hole system can act as a stand-alone device (e.g. during patient transfer from ICU to the ambulance).

Results

Bedside cannulation was uneventful, bleeding complications did not occur. After start of the extracorporeal membrane oxygenation the systemic blood-flow and oxygenation were restored. During extracorporeal assistance, including air (n = 12) and ground (n = 3) ambulance transfer, no technical complication arrived. Limb ischemia due to the arterial cannula was observed in two cases. Hospital survival rate was 66%.

Conclusion

The use of this new portable MINI-ECMO system is safe and high effective for management of hypoxemic patients with severe respiratory and cardiocirculatory failure, resistant to conventional critical care treatment. Especially patients in outlying medical facilities can now be first time treated with bedside ECMO support without extended technical or personnel support. Interhospital transfer for ARDS centre treatment or urgent coronary interventions can now be handled in re-established cardiac output and oxygen delivery. The new hand-held MINI-ECMO device enables emergency treatment of hypoxemia and improves survival rate in critically ill patients.

0515

Major Thoracic Surgery During Extracorporeal Life Support In Ards

M. Bombino*1, N. Patroniti2, S. Isgrò3, G. Foti1, R. Marcolin1, A. Benini1, P. Tagliabue1, R. Fumagalli2, A. Pesenti2

1Perioperative and ICM, San Gerardo Hospital, 2Perioperative and ICM, Experimental Medicine, San Gerardo Hospital, Milano Bicocca University, 3Experimental Medicine, Milano-Bicocca University, Monza, Italy

Introduction

major surgical procedures in patients with ARDS during extracorporeal respiratory support are potentially lifethreatening due to anticoagulation. Moreover, during thoracothomies the surgeon must face the not-collapsable ARDS lung that can limit the accessibility to the thoracic cavity.

Methods

retrospective analysis of indications, complications and outcome of thoracothomies performed during veno-venous ECLS in patients with ARDS.

Results

From 1980, 132 patients with ARDS have been supported with ECLS in our institution. In 21 patients (16%) [13 females, 8 males, mean age 30±12 (8–53 years)] we performed a total of 32 thoracic surgical procedures.

In 15 patients, thoracothomy indication was to control major chest tube bleeding or massive hemothorax, 4 lobectomies were performed. 6 patients underwent revisions within 48 hrs due to persistent bleeding, in other 3 patients 4 delayed reopening were required due to thoracic cavity tamponade with mediastinal shift. Of these 15 patients, 5 (33%)were successfully deconnected from ECLS, and discharged alive from the hospital, 2 remained on ECLS for more than 2 weeks after the first thoracotomy. In one patient a sternotomy with pericardiectomy, following three episodes of cardiac tamponade, was performed; the patient was deconnected from ECLS after operation and survived. In 4 patients thoracothomy indication was to control the source of septic shock and concomitant bronchopleural fistula, 3 pneumonectomy and 1 lobectomy were performed, none survived. One patient underwent a first thoracotomy for bleeding and subsequentely a pneumonectomy due to lung colliquation, and died. Packed red cell and fresh frozen plasma transfusions during surgical procedures were 2.58±1.53 lt and 2.37±1.56 respectively.

Conclusion

surgical control of hemorrhage from the thoracic cavity in ARDS patients anticoagulated for ECLS is a demanding task. Since 1991, the use of heparinized circuits allowed us to perform the surgical procedures with normal or close to normal coagulation status while on bypass. Nevertheless, multiple procedures, as reported in literature, are frequently required to control bleeding. Other thoracotomy indications had an unfavorable outcome in our experience.

Grant Acknowledgement

Supported by MIUR.

0516

Physiologic Effects of Intrapulmonary Percussive Ventilation in Patients At Risk for Respiratory Distress After Extubation

F. Vargas*1, S. Dimassi2, J. Dellamonica2, A. Lyazidi2, F. Roche-Campo2, L. Brochard21Departement de Reanimation Méadicale, Hopital Pellegrin — Tripode, Bordeaux, 2Reanimation Medicale — Unitéa INSERM U841, Hopital Henri Mondor, Creteil, France

Introduction

Intrapulmonary percussive ventilation (IPV) is a technique that delivers small burst of gas at high rate, intended for mobilization of secretions, and which can be superimposed on spontaneous breathing (SB). Surprisingly little studies assessed the physiological changes induced by IPV. In stable COPD patients, IPV induced a significant unloading of the diaphragm (1). In a bench study, we demonstrated that IPV added to a conventional ventilator, increases positive end expiratory pressure and tidal volume (2). The aim of this prospective study was to assess the short term physiologic effect of a session of IPV in patients at high risk extubation failure treated by non invasive pressure support ventilation (NPSV).

Methods

In 10 patients, 1 hour after extubation, we evaluated gas exchange, respiratory rate and inspiratory effort during a 20 minute-period of NPSV and a 20 minute-period of IPV superimposed on SB (IPV-SB). PSV was gradually increased until expired tidal volume was 6 to 8 ml per kilogram of body weight. Positive end-expiratory pressure was set at 4–5 cm H2O. The setting for IPV were as follows: frequency of the percussion = 250/min, driving pressure 1.2 bar. The interface was a facial mask. Initial measurements were performed while the patient was in SB(initial) with return to a final SB condition.

Results
Main results are shown in Table 1. Data are reported as the median and interquartile range [IQR]. § p < 0.05 vs SB; # p < 0.05 vs IPV. IPV-SB and NPSV improved inspiratory effort (p < 0.05). IPV was less efficient than NPSV in reducing inspiratory effort (p < 0.05). Neither pH, PaCO2 nor PaO2/FiO2 changed significantly with IPV-SB compared to SB.
 

SB initial

IPV-SB

NPSV

SB final

PTPdi/breath

9.1 [7.0–15.0]

7.9 [5.9–12] §

5.8 [4.5–7.4] §,#

8.9 [7.0–15]

PTPdi/min

250[156–288]

156 [114–196] §

118[104–131]

189 [149–234]

RR (breath/min)

23 [18–31]

17 [16–30] §

19 [17–23] §

20 [17–32]

pH

7.41[7.38–7.44]

7.43[7.39–7.45]

7.43[7.39–7.45]

7.42[7.38–7.45]

PaCO2(mmHg)

44.0 [41–52]

43 [38–50]

43 [40–46] §

43 [40–50]

PaO2/FiO2(mmHg)

261 [210–280]

262 [233–293]

299 [222–303] §

260 [216–348]

Conclusion

This physiological study, performed in patients at high risk for extubation failure confirmed that the application of IPV is associated with a reduction of the inspiratory effort, and interestingly, with a marked reduction in respiratory rate. The unloading effect is moderate and less important than with NPSV.

0517

Effect of Frequency on Lung Protection During High Frequency Oscillation Ventilation in Sheep Acute Respiratory Distress Syndrome Model

S. Liu, H. Qiu*, Y. Yang, Q. Chen, Y. Huang

Intensive Care Unit, Southeast university medical college, Nanjing, China

Introduction

To evaluate the effect of frequency on the prevention of Ventilation- Induced Lung Injury (VILI) during High Frequency Oscillation Ventilation (HFOV) in a Sheep Acute Respiratory Distress Syndrome (ARDS) Model.

Methods

Twenty-four adult sheep (38.3± 2.3 kg) were randomly divided into four groups (n=6): 3 HFOV groups (3, 6, 9Hz) and conventional mechanical ventilation (CMV) group. After induction of ARDS model (PaO2 400mmHg. After this recruitment procedure, optimal mean airway pressure was selected by decreased in 2mmHg every 5 minuets until PaO2 decreased blow 400mmHg, and ventilation was continued for 4 h. Hemodynamics, respiratory mechanics and gas exchange were measured throughout the experiment, and lung histopathological changes, lung wet/dry weight ratio, lung IL-6 expression (ELISA) were determined.

Results

Heart rate (HR), Mean arterial pressure (MAP), cardiac output (CO), Central venous pressure (CVP) and pulmonary arterial wedge pressure (PAWP) did not differ among the four groups in experiment (p > 0.05). After lung recruitment, sustained improvements in the oxygenation index were observed in all groups. Histologically, lung injury score was significantly lower in 9Hz HFOV group than other groups (p < 0.05). Lung wet/dry weight ratio did no differ among the 4 groups (p < 0.05). The expression of IL-6 in lung tissue significantly reduced in 6Hz and 9Hz HFOV-treated animals (p < 0.05).

Conclusion

Compared with CMV and low frequency in HFOV, the higher frequency in HFOV result in less lung injury. High Frequency Oscillation Ventilation (HFOV) may be an optimal lung-protective strategy.

0518

Acute Effects of High Frequency Percussive Ventilation on Oxygenation in Patients with Acute Respiratory Failure

M. Poriazi, C. Routsi*, E. Angelopoulos, D. Zervakis, C. Roussos, S. Nanas ICU, Evangelismos Hospital, Athens, Greece

Introduction

High Frequency Percussive Ventilation (HFPV) has been used in the past mainly as a rescue therapy in patients with acute respiratory failure and persistent hypoxemia with promising results. However the role of HFPV has not been completely confirmed yet because of a relatively small number of published data. We tested the efficacy of the HFPV (VDR 4 ventilator) in patients with hypoxemia of various etiologies in comparison with the conventional mechanical ventilation.

Methods

We studied 8 mechanically ventilated patients (APACHE II score 20 ± 6, SOFA score 10 ± 3, Lung Injury Score 2.4 ± 0.8). Oxygenation, respiratory parameters and hemodynamics were monitored on conventional mechanical ventilation (volume control ventilation, VCV), during a short course of 2 hours on HFPV and finally 30 minutes after reconnection to conventional ventilation.

Results

Values of PO2/FiO2, PO2, PCO2, pH on conventional mechanical ventilation, before the onset of HFPV, after 2hours on HFPV and 30min after reconnection to conventional mechanical ventilation were analyzed with repeated-measurements method (Anova) and resulted in an overall significant effect. All p-values were Bonferroni adjusted. PO2/FiO2 before the onset of HFPV and 2 hours after was significantly increased (128 ± 68 vs. 281 ±110, respectively, p=0.02). In contrast, it was decreased 30min after reconnection to VCV (202 ± 81, p=0.04). pH values were increased 2hours after HFPV compared to those before (7.47 ± .09 vs. 7.37 ± .09 respectively, p=0.01) and 30min after reconnection to VCV (7.42 ± .09, p=0.03). Changes in PCO2 were also significant between VCV and HFPV (44.7 ± 15.3mmHg vs. 32.5 ± 5.7mmHg respectively, p=0.049) and 30min after reconnection to VCV (39.07 ± 6.7mmHg, p=0.01). Changes in PO2 were also significant before and after HFPV(91.6 ± 25.1mmHg vs. 178.7 ± 66.1mmHg respectively, p=0.01) and 30min after reconnection to VCV (123.5 ± 39.2, p=0.03). There were no statistically significant changes in hemodynamics during the study period.

Conclusion

These results indicate that a short course of HFPV acutely improves oxygenation without affecting hemodynamics. Further study should focus on longitudinal effects of HFPV as an alternative type for ventilation in patients with hypoxemia.

0519

Biological Markers of Lung Injury Before and After the Institution of High Frequency Percussive Ventilation in Patients With Smoke Inhalation Injury

P. A. E. G. Reper*, K. van Loey

Critical Care Department, Queen Astrid Hospital, Brussels, Belgium

Introduction

Several biological markers of lung injury are predictors of morbidity and mortality in patients with acute lung injury (ALI). The low tidal volume lung-protective ventilation strategy is associated with a significant decrease in plasma biomarker levels compared to the high tidal volume ventilation strategy. The primary objective of this study was to test whether the institution of high frequency percussive ventilation in spontaneously ventilating patients with severe smoke inhalation injury exacerbates pre-existing lung injury by using measurements of biomarkers of lung injury before and after starting HFPV.

Methods

A prospective observational cohort study was conducted in the burn intensive care unit. Twelve intubated, mechanically ventilated patients with smoke inhalation injury were enrolled. Physiologic data and serum samples were collected within 6 hours before intubation and at two different time points within the first 24 hours after intubation to measure the concentration of interleukin (IL)-6, IL-8 and TNFalpha. The differences in biomarker levels before and after intubation were analysed using repeated measures analysis of variance and a paired t test with correction for multiple comparisons.

Results

Before endotracheal intubation, all of the biological markers (IL-8, IL-6 and TNFalpha) were elevated in the spontaneously breathing patients after smoke exposition. After intubation and the institution of HFPV monitored with Bicore device (tidal volume 7 to 8 ml/kg per ideal body weight), none of the biological markers was significantly increased at either an early (3 +— 2 hours) or later (20 ± 3,5 hours) time point. However, the levels of IL-8 were significantly decreased at the later time point (20 ± 3,5 hours) after intubation. During the 24-hour period after intubation, the PaO2/FiO2 (partial pressure of arterial oxygen/fraction of the inspired oxygen) ratio significantly increased and the plateau airway pressure significantly decreased.

Conclusion

Levels of IL-8, IL-6, and TNF are elevated in spontaneously ventilating patients with ALI following smoke inhalation prior to endotracheal intubation. The institution of HFPV does not further increase the levels of biological markers of lung injury; the results suggest that the institution of HFPV does not worsen the pre-existing lung injury in most patients with smoke inhalation injury.

0520

Ventilatory Parameters Not Affected by High Frequency Chest Wall Oscillation and Manual Hyperinflation

A. van Hees*, B. Speelberg

Intensive Care, St.Elisabeth Hospital Tilburg, Tilburg, Netherlands

Introduction

The Vest® System is a new device used for mobilising secretions in ventilated patients. The system operates by means of external oscillations of the chest wall.

Methods

6 Patients were studied in the standard treated group and 6 in the group treated with the Vest® System. Standard therapy consisted of manual hyperinflation with 100% oxygen 3 times a day with suctioning of the airways and exprimation. Patients received an external vest around the thorax. The external vest contained an air chamber which could be inflated with pulsatile oscillative flow. In this way external compressions were done during 20 minutes. SpO2, breathing frequency, tidal volumes, PeCO2 and dynamic lung compliance were recorded at baseline, direct after and 1 hour after therapy. Differences between the measured values compared to baseline were calculated on the time point :direct after treatment and after 1 hour of therapy. These differences were used to investigate if there was a statistical difference between the standard care group and the group treated with The Vest® System. SPSS version 15.0 for Windows was used to investigate the statistical difference with nonparametric Mann Whitney U testing. A significance level of <0.05 was considered significant.

Results
The difference in breathing frequency directly after treatment was significant lower in the group treated with The Vest® System. Oxygen saturation directly after treatment was significantly higher in the standard treated group.
Table 1

Differences Compared to Baseline in the 2 Groups of Patients

 

Standard therapy

The Vest® system

Tidal volume difference direct mean +/− SD

− 9.0 +/− 66.0

− 23.3 +/− 91.3

Tidal volume difference after 1 hour mean +/− SD

0.6 +/− 57.6

14.6 +/− 108.3

SpO2 difference direct mean +/− SD

0.7 +/− 1.8

− 0.02 +/− 1.73

SpO2 difference after 1 hour mean +/− SD

0.5 +/− 1.8

0.1 +/− 1.8

PeCO2 difference direct mean +/− SD

0.16 +/− 0.51

−0.02 +/− 0.53

PeCO2 difference after 1 hour mean +/− SD

0.06 +/− 0.36

0.03 +/− 0.62

Dynamic lung compliance difference direct mean +/− SD

− 0.3 +/− 5.7

− 1.5 +/− 7.5

Dynamic lung compliance difference after 1 hour mean +/− SD

0.2 +/− 5.6

1.5 +/− 7.6

Breathing frequency difference direct mean +/− SD

4.1 +/− 6.7

− 0.2 +/− 6.0 **

Breathing frequency difference after 1 hour mean +/− SD

− 1.7 +/− 3.1

− 1.6 +/− 4.0

*=p < 0.05; ** = p < 0.005

Conclusion

We did not demonstrate an advantage in the group treated with the Vest® System. The SpO2 improvement in the standard treatment group is probably the result of manual hyperinflation with 100% oxygen. The demonstrated increase in breathing frequency direct after treatment with manual hyperinflation and exprimation could have been the result of waking up the sedated patient. The value of the Vest® System needs still to be demonstrated but also the value of conventional therapy as manual hyperinflation needs to be proven.

0521

Reduction in Complication and Mortality Rates with a New Algorithm for Extracorporeal Pumpless Lung Assist — a Prospective Single-Center Report

M. Zimmermann*1, A. Philipp2, T. Mueller3, M. Ramming1, T. Bein1

1Anesthesia and Intensive Care, 2Cardiothoracic Surgery, 3Internal Medicine, University Hospital, Regensburg, Germany

Introduction

Pumpless arterio-venous interventional lung assist (iLA, Novalung, Hechingen, Germany) is used for extracorporeal gas exchange in severe ARDS. iLA is a single-used ultracompact gas exchange system perfused by the heart after arterial and venous cannulation. We report on a reduced complication and mortality rate using a new indication and implementation algorithm in comparison with a published cohort(1).

Methods

Between 10/2004 and 3/2008 iLA was used in 51 patients suffering from severe ARDS (PaO2/FIO2 75 +/− 27 mmHg, PaCO2 73 +/− 19 mmHg) due to pneumonia, trauma, sepsis, pancreatitis/peritonitis or postoperatively. Insertion was performed mainly for extracorporeal CO2-removal after careful evaluation including a stabilisation period only when certain hemodynamic (norepinephrine < 0,4 ug/kg/min), pulmonary (PaO2/FIO2 70–200 mmHg) and coagulation (platelets > 60.000, aPTT < 60 sec) conditions were fulfilled. A new implementation algorithm was developed including use of smaller arterial cannulae and improved material/insertion technique.

Results
Using a modified algorithm with prior stabilisation, smaller cannulae and improved technique, the incidence of complications and the mortality rate were significantly reduced, although the severity of disease (SOFA-score) was similar between the cohorts. iLA induced a marked CO2-removal allowing a lung protective ventilation (tidal volume < 6 ml/kg).
Table 1

Complication and Mortality Rate (Pat/%) Following Ilaimplementation * = P < 0.05)

 

cohort 1 (n = 90) 9/1996-9/2004 (1)

cohort 2 (n = 51) 10/2004-3/2008

ischemia of lower limb

13 (14,4%)

4 (7,8%) *

total complication rate

22 (24,4%)

6 (11,8%) *

cannula size (arterial)

17,8 +/− 1,3 F

16,5 +/− 1,3 F *

mortality rate

53 (58,8%)

25 (49,0%) *

Conclusion

Pumpless arterio-venous extracorporeal lung assist is a safe and effective measure for CO2-removal allowing a strict lung protective strategy in ARDS.

0522

To Evaluate the Role of Continuous Tracheal Gas Insufflation (C-TGI) with Protective Lung Ventilation in Patients with Respiratory Acidosis of Varied Etiology

Y. Javeri*, T. Rawat, S. Bal, R. Uttam, A. K. Mandal, H. Tewari, R. Kumar, D. Nama, P. Arora, R. Mani

Pulmonology-Thoracic Surgery and MICU, Fortis Flt. Lt. Rajan Dhall Hospital, New Delhi, India

Introduction

CO2 elimination with TGI is dependent on flow rate as fresh gas flushes higher proportion of proximal dead space. High flow rate causes tubulence at tip and increases distal gas mixing. The aim of present study is to confirm effect of TGI throughout the respiratory cycle on alveolar ventilation at various catheter flows and as an adjunct to protective lung ventilation for various pulmonary diseases. We measured effect of TGI on pH and PCO2 in cases of severe respiratory acidosis.

Methods

20 patients (13 males and 7 females) with varied etiologies were studied. ALI/ARDS(n= 9), Pneumonia with / without sepsis and MODS (n = 8), Post LVRS (n = 1), Post lobectomy (n=1), Disseminated CMV with MODS (n=1). All the patients were on protective lung ventilation strategies and had respiratory acidosis. The ventilatory parameters were Tidal volume 4–10 ml/Kg, FiO2: 0.4–1, PEEP : 5–15, RR: 12 – 30/min), Inspiration: Expiration (I:E) : 1:2 – 1:4. c- TGI was given with a tool using infant feeding tube through a catheter mount and with a flow rate 4 – 10 litres / min. Data collected were arterial blood gas (ABG) at 0, 1, 4 hour and at every 24 hours after establishing TGI. The following parameters in ABG were noted: pH, PaCO2, PaO2, HCO3 and compared.

Results

After the use of c- TGI with protective lung ventilation we noted Respiratory acidosis with a baseline pH 7.123±0.11 with improvement in pH at 1 hour (7.225±0.1086 (p=0.000)) and 4 hour (pH 7.261±0.0941 (p= 0.000)). The PaCO2 at baseline was 82.9 ± 23.79 mmHg and there was fall in PaCO2.

PCO2 levels were 64.53±18.79(p = 0.00) at 1 hr and 60.4±17.95 (p= 0.00) at 4 hour which are significant.
Table 1

Statistics

 

pH (0)

pH (1 hr.)

pH (4 hr)

PCO2 (0)

PCO2 (1 hr)

PCO2 (4 hr)

Mean

7.1235

7.2255

7.261

82.9000

64.53

60.48

Std Error of Mean

2.520 E-02

2.428 E -02

2.115 E-02

5.3188

4.2023

4.0138

Median

7.1400

7.2300

7.2700

73.600

63.600

57.0000

Standard Deviation

0.1127

0.1086

0.094

23.7966

18.7932

17.9503

Conclusion

In a cohort of 20 patients with respiratory acidosis of varied etiology, cTGI appears to have a significant effect on lowering PaCO2 and pH values in initial 1 hour of TGI with protective ventilation. There was significant improvement in PaCO2 and pH at 4 hours and correction in respiratory acidosis.

0523

Treatment of Broncho-Pleural Fistula Using Endobronchial Watanabe Spigots in Mechanical Ventilated Patients

S. Lachkar*1, G. Beduneau2, F. Corne1, B. Veber3, J. C. M. Richard2, L. Thiberville1

1Clinique pneumologique, 2Medical ICU, 3Surgical ICU, Charles Nicolle teaching hospital, Rouen, France

Introduction

Bronchopleural fistula (BPF) is a challenging problem in patients under mechanical ventilation (MV), where it is associated with a significant mortality. Surgery is frequently not possible because of the poor patient's general conditions and high operative risk. The treatment of BPF using Endobronchial Spigots was recently described. This perendoscopic technique makes use of silicone stoppers to obstruct segmental bronchi in the area of a BPF. We report our experience of therapeutic bronchial obstruction for persistant BPF mechanically ventilated patients.

Methods

Indications for endobronchial spigots were association of (1) distal BPF in a mechanically ventilated patient with weaning failure (2) persistent pleural leakage after several days of drainage (3) patient not suitable for surgery. The technique comprises 2 steps: 1. Localization of one or several bronchi related to the fistula, using a Fogarty balloon catheter introduced through the bronchoscope working channel. The leakage is localized by inflating the balloon catheter in the selected bronchi until the bubbling stops in the pleural drainage system. 2. Introduction of the silicone stopper (Novatech ® “EWS” or Endobronchial Watanabe Spigots) previously attached to a biopsy forceps through the bronchial tree, down to the localized segmental bronchi. Complete success of the technique was defined by an absolute control of the air leakage, allowing pleural tube removal. The partial success of the technique was defined by the persistence of a reducted air leakage associated with patient's improvement.

Results
Between January 2003 to November 2007, 9 mechanically ventilated patients with untractable BPF were treated by therapeutic bronchial occlusion using EWS. The cause of BPF was an empyema in 6 patients, traumatic pneumothorax in 2 patients and complication of lung surgery in one. In one case, a spigot migration produced a tracheal tube obstruction which obliged to emergency reintubation. In 2 cases (3,4) the air leakage reoccured due to collateral ventilation leading to a second EWS positioning. Results are detailed in table.

Patient

Day's drainage before EWS

Withdrawal chest tube

MV weaning

EWS success

Evolution

Etiology of death

1

4

0

0

partial

Death

Sepsis

2

26

1

1

1

Cured

0

3

19

0

0

0

Death

ARDS

4

50

0

0

0

Death

Sepsis

5

30

0

0

partial

Death

Sepsis

6

30

1

1

1

Cured

0

7

8

1

1

1

Death

Aspiration

8

30

0

0

0

Death

ARDS

9

35

1

1

1

Cured

0

Conclusion

Endoscopic treatment of EBF using EWS is an efficient technique in mechanically ventilated, non operable patients. Several EWS positioning could be necessary. The possibility of a spontaneous migration of the spigot should be taken into account in patients under mechanical ventilation through an orotracheal tube.

Poster Sessions Prognosis and outcome: 0524-0537

0524

An Appraisal Of Multivariate Logistic Regression Models In The Critical Care Literature Anno 2006: Preliminary Results

P. Lambrecht1, P. Verslype1, M. Bekaert2, P. Depuydt3, S. Vansteelandt2, J. Decruyenaere3, D. Benoit*3

1Faculty of Medicine and Health Sciences, 2Department of Applied Mathematics and Computer Science, Ghent University, 3ICM, Ghent University Hospital, Gent, Belgium

Introduction

Improper reporting of multivariate logistic regression models (MLR) can potentially make the results of a study inaccurate, misleading, or difficult to interpret (1). The aim of this study is to evaluate the adequacy of (reporting) MLR in two major critical care journals anno 2006; Critical Care Medicine and ICM.

Methods

At the present 129 articles published in Critical Care Medicine in 2006 were manually reviewed for 10 potential limitations in the reporting of MLR. Potential limitations were essentially based upon previous recommendations (1,2). Values are expressed in median (25th – 75th percentiles) and percentages when appropriate.

Results

Of the 129 articles, 29 (22%) reported the use of MLR and the results were shown and could subsequently be comprehended fully in 25. A median of 3 (2–21) models per article were reported. A total of 79 models could be checked for potential limitations. The selection of variables initially included in the models was clearly specified or could be deduced from the manuscript in 91% of the models. Testing for collinearity was properly reported in only 10%. The type of MLR modelling (stepwise, forward, backward..) was not specified in 6%. Nonconformity to the linear gradient and testing for an interaction term or effect modification were checked in only 20% and 25% of all models respectively, and in only 29% and 32% of models including[500 patients. Units of continuous variables were not provided in 33%. Criteria for overfitting data was violated in 35% and Hosmer-Lemeshow statistics were not provided in 84% of the models. Discrimination statistics by mean of area under ROC curves were provided in 3% and in 2/18 (11%) of models which had the primary intention in predicting outcome respectively. The recommendation regarding the coding of the end-point according to the Cochrane Collaboration was violated in 9%. A median of 4 (3–5) limitations were found per model.

Conclusion

According to these preliminary results only a minority of the reports published in critical care literature include enough information to allow the critical reader to fully comprehend the MLR model and the subsequent conclusion of the study. Critical care journals should pay more attention to the statistical review of manuscripts using MLR.

0525

Reason for Intensive Care Unit Admission as a Determinant of Clinical Course and Outcome of Very Elderly Patients

M. Giannakou, K. Matsi, F. Drampa, G. Tsaousi, E. Anastasiou, E. Geka*, E. Efthimiou, I. Papageorgiou, E. Sofianos

Department of Anesthesiology and ICU, AHEPA Hospital, Thessaloniki, Greece

Introduction

As population demographics change and as new technologies and interventions prolong life expectancy, the proportion of very elderly patients admitted to and surviving ICU will constantly increase. This poses complex challenges for the benefit of their management and the ultimate use of ICU resources. The objective of this study was to review the impact of reason for admission on the ICU course of the very elderly patients in a multidisciplinary ICU.

Methods

All records of patients >=75 years admitted to the ICU between Jan.2003–Dec.2006 were retrospectively evaluated from our electronic database. The assessed data included: age, previous medical history, reason for admission (elective/emergency surgery, medical, stroke and trauma), severity of illness scores (GCS, APACHE II, SOFA1 upon admission and SOFA2 upon discharge), ventilator associated pneumonia (VAP), days of mechanical ventilation (MV), use of inotropes, ICU length of stay (LOS), ICU mortality according to reason for admission. Data are presented as mean values ±SD. For statistical purposes One way ANOVA and chi-square test were applied.

Results
The study population (n = 167) accounted for 11.65% of total ICU admissions (n = 1433). Patient data according to reason for admission and statistical significance are referred to in Table 1. Total duration of MV in all categories of survivors was 8.15±10.3 days. Previous medical history (cardiac, pulmonary, neurological disease, hypertension, diabetes) did not seem to influence mortality.
Table 1

Patient Data and Statistical Analysis

Admission reason n(%)

El. surgery 56 (33.5)

Em. surgery 38(22.8)

Medical 40(24)

Stroke 22(13.2)

Trauma 11(6.6)

p value

Age (years)

77.5±3.2

79.9±3.5

79.35±4.8

77.8±2.9

78±3.3

.015

APACHE II

12.2±5.9

17.6±7,7

22.8±8.1

22±7.3

24±5.6

.000

GCS

14.7±1.2

12.6±3.9

9.4±4.7

6.3±2.4

9.7±9.3

.000

SOFA1

2.3±2.4

4.4 ±3

5.6±2.9

5.09±2.4

62 ± 2.9

.000

SOFA2

2.4±1.9

3.5 ±2.4

4.8±2.9

5.2±2.6

4.7±1.5

.000

ICU LOS (days)

2.9±7.3

9.7±13.2

18.5 ±21.08

16±15.1

17.4±18.4

.000

Inotropes n(%)

14(25)

19(50)

21(152.5)

3(13.6)

6(54.5)

.002

VAP n(%)

3(5.7)

8(21.6)

18(45)

5(22.7)

4(36.4)

.000

Nonsurvivors n(%)

5(8.9)

7(18.4)

21(52.5)

10(45.4)

6(54.5)

.000

Conclusion

Patients who underwent elective surgery clearly displayed an uneventful course and a satisfactory survival rate. Acute medical and surgical conditions predisposed to a higher mortality rate due to initial acuity of illness, a longer ICU LOS and an increased incidence of VAP. Nonetheless ICU stay proved to be beneficial in the above category of patients as shown by the discharge SOFA score. Our results reinforce the aspect that age alone should not hamper the decision for ICU admission.

0526

Long-Term Prognosis of Patients with Lung Cancer Admitted to the Intensive Care Unit

S. Roques*, A. Parrot, A. Lavole, V. Gounant, M. Djibre, M. Fartoukh

Service de pneumologie et reanimation, hopital Tenon, Ap-Hp, paris, France

Introduction

Benefit of ICU admission for lung cancer patients remains unclear. ICU triage decision of such patients may be affected by both the estimated late outcome and the possibility to administer specific cancer-related treatments following ICU stay.

Methods

Retrospective study of critically-ill lung cancer medical patients admitted to a 10-bed respiratory intensive care unit in a tertiary care university hospital in order to identify early predictive factors of long-term mortality and outcome (six months) and to assess whether the ICU survivors received afterward any cancer related treatments.

Results
105 consecutive lung cancer patients were included over a nine-year period. Of the 105 patients (mean age 64.8 yrs), 87 patients (83%) had non-small cell lung cancer (NSCLC). An advanced disease was diagnosed in 83 patients (79%) [NSCLC stage IIIB and IV or disseminated small lung cell cancer (SCLC)]. The main reason for ICU admission was acute respiratory insufficiency (59%). The simplified acute physiologic score (SAPS II) score was 40 ± 21. Forty-one percent required mechanical ventilation. The ICU, hospital and 6-month mortality rates were 44%, 55% and 72%, respectively. A performance status (PS)[2 or =2 [odds ratio OR=6.4 (95% confidence interval CI (2–21)] and mechanical ventilation [OR=5 (95%CI (1.7–14)] were independently associated with an increased 6-month mortality. Thirty (70%) of the ICU survivors received specific cancer treatments after their ICU stay.
Table 1

Factors Associated with 6-Month Mortality. Univariate Analysis

Variables

All

Survivors

Non Survivors

p

 

n = 105

n = 29

n = 76

 

Age, yr

64.8 ± 10.6 (39–86)

66.0 ± 8 (42–81)

64.2 ± 11 (39–86)

0.41

SAPS II

40 ± 21 (13 to 112)

30.7 ± 10 (13–54)

44.2 ± 23 (13–112)

0.009

Performance Status 0 or 1

56

23

33

0.001

Advanced Cancer

83

19

64

0.015

Admission for Acute respiratory failure

62

12

50

0.02

Admisson for Haemoptysis

47

14

33

0.65

Admission for Septic Shock

10

1

9

0.19

Admission for Neurological reason

10

3

7

0.85

Mechanical Ventilation

43

4

39

0.0005

Vasopressors

33

3

30

0.004

Complete clinical information was available for 103 patients at 6 months

Table 1

Prediction of 6 Month Hospital Mortality Using Multivariate Analysis

Variables

p value

Odds ratio (95% interval confidence)

Performance Status 0 or 1

p = 0,021

OR 6,4 (2–21)

Mechanical ventilation

p = 0,031

OR 5 (1,7–14)

Complete clinical information was available for 103 patients at 6 months

Conclusion

ICU admission of non-surgical lung cancer patients may be beneficial for selected patients (PS 0 or 1) in terms of the administration of cancer-related treatments after discharge and long-term survival.

0527

Risk Factors for Mortality Following Cardiac Valve Re-Operations for Prosthetic Valve Dysfunction

J. J. Massaut*1, P. Reper1, P. Wauthy2, F. Deuveart2

1Intensive Care, 2Cardiac Surgery, CHU Brugmann, Bruxelles, Belgium

Introduction

The purpose of the study was to identify independent risk factors for intensive care mortality in patients undergoing cardiac valve re-operations in our institution.

Methods

We retrospectively analyzed the data concerning 235 patients admitted in our intensive care unit following valve re-operations during a ten years period (1997–2006). Multivariate stepwise logistic regression was used to analyze preoperative, intra-operative and intensive care variables and determine risk factors for intensive care mortality. Statistical analysis was performed using Stata 8 for UNIX.

Results

The overall hospital mortality was 7.89% (19 patients /238) for a predicted mortality by Euroscore of 10.12%. Intensive care mortality was 6.8% (16/235). Univariate analysis showed that the following variables were associated with higher mortality rates: age, urgent intervention, advanced New-York Heart Association functional class, increase preoperative urea, creatinine and lactic dehydrogenase (LDH) plasma levels, low preoperative hemoglobin level, presence of paravalvular leak, number of re-operation (Redo), length of intervention and extracorporeal circulation, necessity for transfusion and length of administration of inotropes. Multivariate logistic regression identified high preoperative LDH, urea plasma levels and Redo as independent predictors of intensive care mortality (Table I) with a high specificity and sensitivity (Area under the ROC curve = 0.94). A score was computed using these factors and the coefficients to predict mortality. No mortality was observed for a score value < 3.5.

For postoperatives factors, the prolonged use of inotropes was independently associated with intensive care mortality. Failure to wean patients from inotropes during the first 48 hours was associated with very significant higher mortality rate (odds ratio = 12).
Table 1

Logistic Regression Analysis of Indepedant Risk Factors For IC Mortality

 

Coefficients

Odds Ratio

CI 95%

p values

LDH (UI/L)

0.00087

1.00087

1.00033–1.00141

0.002

Urea (mg/dL)

0.02918

1.02961

1.00284–1.05708

0.030

Redo

1.00978

2.99952

1.05329–8.53051

0.040

Constant

−7.7172

  

<0.001

Conclusion

Number of re-operations, higher urea and lactic dehydrogenase preoperative plasma levels were independent predictors of intensive care mortality in patients admitted to our intensive care unit following cardiac valve re-operations. Failure to wean patients from inotropes within 48 post-operative hours was associated with a high increase in mortality.

0528

Influence of Body Mass Index on ICU Outcome

B. Schildt*, S. Lehtipalo, J. Claesson

Anesthesiology and Intensive Care, Umeå University Hospital, Umeå, Sweden

Introduction

Overweight is an increasing issue in the society, but the influence of body mass index (BMI) on intensive care unit (ICU) outcome is uncertain. Our objective was to evaluate the impact of BMI on ICU outcome and test the hypothesis that underweight or obese patients will have a higher morbidity and mortality compared to patients with a normal BMI.

Methods

Retrospective study of 930 consecutively admitted ICU patients at Umeå University Hospital, Sweden, from 2006 to 2007. 238 patients fulfilled the inclusion criteria age ≥ 20 years, ICU stay ≥ 72 h (main study group). A group of 52 nonsurvivors, ICU stay < 72 h constituted the separate early death group. Patients were compared by age, APACHE II score, BMI, admission diagnosis and ICU outcome (mortality, ICU length of stay (LOS), time on ventilator, and SOFA score. Using an individual matching procedure, the obese groups (BMI > 30) and underweight group (BMI < 20) were compared with the normal weight group (BMI 20.0–24.4). The relationship between BMI and ICU outcome was determined by multiple linear regression analysis.

Results
In the main study group, duration of ventilation, ICU LOS (Figure 1), SOFA score and mortality showed no differences among BMI groups (p= 0.55, p=0.06, p=0.40, and p=0.71) and admission diagnoses were not related to BMI groups (p=0.06). Multiple regression analysis and individual matching (Table 1) showed that BMI was not an independent predictor of ICU outcome.
Table 1

Outcome Variables for Matched Pairs

  

BMI<20.0

BMI 20.0–24.4

p

BMI >30.0

BMI 20.0–24.4

p

ICU lenght of stay

(hours)

233 ± 139

217 ± 81

0.82

261 ± 220

211 ± 128

0.61

Time on ventilator

(hours)

162 ± 119

165 ± 92

0.75

177 ± 190

154 ± 117

0.69

Creatinine day 1

mumol/L

77 ± 46

87 ± 24

0.09

134 ± 86

87 ± 52

0.01*

Delta SOFA

 

0.3 ± 2.1

0.6 ± 2.3

0.39

0.7 ± 1.9

1.1 ± 2.0

0.33

Conclusion

ICU outcome (mortality, ICU LOS, time on ventilator, SOFA score) was not influenced by BMI.

Grant Acknowledgement

This study was supported by grant from Medical Faculty, Umeå University.

0529

Patients with Cardiac Arrest Transferred From Aged People Residence

Y. Moriwaki*, M. Iwashita, Y. Tahara, S. Matsuzaki, N. Harunari, S. Arata, H. Toyoda, T. Kosuge, N. Suzuki, M. Sugiyama

Critical Care and Emergency Center, Yokohama City University Medical Center, Yokohama, Japan

Introduction

Today, in some developed countries, many aged people stay in aged people residence rather than private home with their families by several reasons. Most aged people residence have some relationship with supporting hospitals or clinics because most aged people have some medical problems, who usually regularly consult a doctor of these hospitals or clinics. However, in emergency condition, particularly in the night or holly day, these inhabitants can not be transferred to these hospitals and are transferred near emergency department in which hospital doctors have no information of the patients’ medical condition and their philosophy for living and dying process. We often perform aggressive resuscitation against the patient's mind. The aim of this study is to clarify the circumstances of management of inhabitants of aged people residence in emergency condition.

Methods

We reviewed patients with OHCPA who were transferred to our center from aged people residence for past 2 years. In Yokohama, CPA patient is basically transferred to the nearest ED of selected 11 hospitals with adequate ability of proper CPR except special cases in which families can immediately tell the patient's mind and hope to EMS.

Results

27 patients were transferred to our CCE Center. Although in all 27 patients, ADL were restricted and death could be expected, all patients and their families did not consulted concerning their dying process. We blindly performed agressive resuscitation. 12 patients reached ROSC and only 2 survived, 1 of them could returned to the previous aged people residence with neurologically normal condition and the other transferred to other hospital with vegetative state. Although both were witnessed by nurse in the residence, both showed asystole as a first cardiac rhythm on the scene. Of all 27 patients, 16 were witnessed by staff. Of all patients, 7 were witnessed in the dining room, 1 in the bath room, and 1 found in the labatory without witness. In 18 patients who found in the private room, only 7 witnessed by staff and 1 by emergency life saving technicien (ELST) after the scene. In 7 cases bystander voluntarily performed CPR, in 11 cases bystander performed after telephone CPR advice, and in 8 cases any CPR was performed.

Conclusion

Most of patients in aged people residence have some ill condition such as malswallowing, low ability and actibity in daily life. They can be expected to die from these conditions. In some countries including Japan, death is put under taboo and most staffs of these residences do not discuss a death with patient or their families. However, to prepare for emergency condition including CPA and to avoid confusion and unwilling feeling in these conditions, we should consult the patient and his families about his dying process.

0530

Readmition to the ICU: One Year Study

R. P. Oliveira1, M. P. Hetzel1, D. M. Dallegrave1, R. C. Santos1, J. A. Höher1, G. Friedman*2

1icu, Complexo Hospitalar Santa Casa, 2icu, Universidade Federal Ciências Da Saúde De Porto Alegre, Porto Alegre, Brazil

Introduction

ICU readmission during the same hospital stay is a quality indicator. Readmission soon after ICU discharge is considered a premature discharge. Readmissions are associated with worsening of the patients’ disease process, higher hospital costs, and poor outcome. The aim of this study was to describe patients characteristics readmitted to the intensive care unit (ICU) during the same hospital stay.

Methods

Prospective cohort study of all patients (n = 604) consecutive admitted from June-2005 to June-2006 to a mixed university ICU with 18 beds. Data collected were: demographics, diagnosis at admission, APACHE II, and SOFA score at admission and at discharge of ICU, TISS score at 24h, 72h and at discharge ICU; outcomes recorded were: ICU and hospital length of stay, ventilator-days, and ICU and hospital mortality.

Results

During the study period, 604 patients were admitted in the ICU; their mean age was 61±17 yrs and APACHE II was 18.5±9. Forty-one patients (6.78%) required ICU readmission, 31 from ward, 7 from operating room, 2 from hemodialysis and 1 from another ICU. Patients were classified at admission in clinical (83%) and postoperative (17%). Patients requiring ICU readmission had more co-morbidities (3.2±0.8 vs. 2.3±1.0; p = 0.07). There were no differences in age (63±16 yrs vs. 61±17 yrs), APACHE II (19.8±7.3 vs. 18.3 ± 9.1), Glasgow coma scale (12.9±3.7 vs. 12.5±4.2;), SOFA at admission (4.17±3.4 vs. 3.7 ± 3.4) and TISS at 24h (22.2±6 vs. 21.8±7.4) or 72h (21.7±6.1 vs. 21.9±7.7). Twenty seven patients (65.8%) required mechanical ventilation (MV); 25 required invasive MV, 1 non-invasive MV and 1 both techniques. The duration of MV (11.6±13.5 days vs. 9.8 ± 12.1 days) and ICU length of stay (12±13 days vs. 10±21 days) were similar for both groups. Readmitted patients had higher SOFA score at ICU discharge (5.8±5.6 vs. 4.0 ± 4.8; p=0.028). Mean TISS at ICU discharge was slightly greater in readmitted patients (19.2 ± 8.8 vs. 16.8±9.8; p=0.14). The pressure ulcers proportion was higher in readmitted patients (26.8% vs. 12.2%; p=0.016). Although ICU mortality was not different (41.4% vs. 31.6%), hospital mortality rate was significantly higher in readmitted patients (58.5% vs. 41.4%; p=0.049).

Conclusion

Patients who required ICU readmission had more organ dysfunctions and higher hospital mortality risk.

0531

Descriptive Analysis and Epidemiology of Patients Admitted to an Intermediate Care Unit Based on Conventional and Independent Model

M. Andrade, V. Guedes*, N. Dassa, A. Moraes, L. Viana, R. Outeiro Jr, S. Gomes, M. Kandelman, J. Castro, F. Costa

Intensive Care, Hospital Copa D'Or, Rio de Janeiro, Brazil

Introduction

Intermediate Care Areas (ICA) are ways to provide health care services to potentially critical patients that allow for improved cost-benefit ratio of the care offered by Intensive Care Units (ICU) and the conventional and independent model can facilitate utilization of critical care resources.

Methods

Restrospective analysis in 1,090 consecutive admissions from January 2007 to December 2007 in a 18 ICA bed unit. To evaluate specific analysis of mortality, 204 patients were excluded due to readmission. Demographic data were (gender and age, type of patient, origin on admission, and disease group), occupancy rate, severity score (SAPS II and APACHE II), ventilatory parameters, frequency of transfer to the ICU and analysis of the lenght of stay (LOS) and mortality were included.

Results

In the 1090 admissions 549 were male and 542 were female. The mean age was 66,5 ± 20 years. The mean APACHE II and SAPS II score were 10,2 ± 5,1 and 25,9 ± 11, respectively. The mean occupancy rate was 95%. A total of 222 patients were under ventilatory support (20,36%) on admission. When stratified by disease group 228 patients were gastrointestinal (20,8%) and 155 patients (14,1%) were respiratory. In total of admissions 60,7% were nonoperative and 39,3% were operative reason. In relation to origin of admission 70,6% were from ICU, 10,9% from general ward, 10,8% from emergence department, and 7,7% from operating room. Discharge profile was 733 to general ward (67,2%), 178 to home (16,3%) and a total of 159 patients (14,6%) needed to be transferred to the ICU. Invasive mechanical ventilation was used in 48,4% of patients. The most frequent ventilatory mode was continuous positive pressure (CPAP) in 65,2% of patients. The mean positive end expiratory pressure (PEEP) was 9,17 ± 1,04 cmH2O, the mean time under ventilatory support was 7,7 ± 7,8 days. Nocturnal mechanical ventilation was used in 3,6% of patients. The mean LOS was 5,3 ± 6,3 days. The mortality rate intra ICA was 1,8% and post ICA discharge was 9,1%.

Conclusion

When confronted with a growing demand for ICU services, developing intermediate care beds for low-risk patients represent a safe and effective strategy for providing critical care at a reduced cost. A proportion of ICU admissions can be considered low risk for receiving services that are unique to or best provided in an ICU. Despite the widespread notion that larger ICA (above 10 beds) can show low occupancy rates, our data is in contrast and shows that a high occupancy rate is feasible and leads to a better utilization of critical resources.

0532

Predictive Factors of Invasive Procedures in a Cohort of 407 Severe Postpartum Haemorrhage

E. Gayat*1, M. Resche-Rigon2, P. Brechat3, D. Castiel4, H. Khouadja1, M. Rossignol1, O. Morel5, Y. Fargedou6, E. Barranger5, M. Laisné1, P. Schurando1, D. Payen1, A. Mebazaa1

1Department of Anesthesia and Intensive Care, Lariboisiere University Hospital, 2Department of Biostatistic and medical informatic, Saint Louis University Hospital, 3Unité de médicalisation du systéme d'information, service de santé publique et économie de la santé, Lariboisiere University Hospital, 4UFR SMBH, Université Paris-Nord, 5Department of Obstetric and Gynecology, 6Department of Visceral Radiology, Lariboisiere University Hospital, Paris, France

Introduction

Despite the overall decline in maternal mortality in high-outcome countries, postpartum haemorrhage (PPH) still makes a major contribution to maternal mortality and severe morbidity in Europe. The goal of this study was to identify the predictive factors for invasive procedures (haemostatic surgery and arterial embolisation) in case of postpartum haemorrhage already treated with sulprostone infusion.

Methods

All PPH patients were admitted in our tertiary centre transferred from other hospitals, over a 3-year period.Wehave compared the patients characteristics in two groups : 1/a « Invasive Procedure » (IP) group including patient managed with haemostatic surgery and/or arterial embolisation and 2/a « Medically Managed » (MM) group including the remaining patients for which no intervention was needed. A multivariate analysis led to identify independent predictors of invasive procedure, performed in the 4 hours following admission.

Results

Patients characteristics confirm the occurrence of severe bleeding in our 407 patients before admission in our centre (Table 1).

The five identified predictors are represented on the Table 2.
Table 1

Parameters at Admission

 

Overall n = 407

“MM” group n = 234

“IP” group n = 173

p value

RBC before transfer (units)

2 ± 3

1 ± 2

3 ±4

< 0.0001

SBP (mmHg)

113 ± 25

119 ± 23

105 ± 25

< 0.0001

HR (bpm)

102 ± 22

96 ± 19

111 ± 24

< 0.0001

Hemoglobine (g/dL)

8.9 ± 1.9

9.3 ± 1.7

8.4 ± 2.0

< 0.0001

PT (%)

66 ± 19

72 ± 15

58 ± 21

< 0.0001

Fibrinogen (g/L)

2.5 ± 1.1

2.8 ± 0.9

1.9 ± 1.1

< 0.0001

Troponin (pg/L)

0.4 ± 2.4

0.2 ± 1.7

0.6 ± 3.1

< 0.0001

(RBC : red blood cell, SBP : systolic blood pressure, HR : heart rate)

Table 2

Items Present at Admission

 

OR

confidence interval 95%

p value

Placentation's abnormality

13.5

3.6 – 51.3

0.0001

PT < 50%

2.8

1.1 – 6.8

0.02

Fibrinogen < 2 g/l

3.0

1.5 – 5.5

0.001

Troponin I detectable

3.0

1.8 – 5.2

0.0001

HR > 115 bpm

3.8

3.0 – 7.0

< 0.0001

Conclusion

Patients with a persistent or worsening PPH are often difficult to evaluate, because of the importance of the bleeding and/or the treatment already administrated. The delay in the management of those patients directly influences outcome. We described five predictive factors that allow to classify patients in high, intermediate or low risk of needing a haemostatic invasive procedure.

0533

Prognostic Value of Antithrombin III in Critical Care Patients

E. Ioakimidou*1, A. Theodotou1, A. Amigdalou1, N. Panagiotou2, K. Mandragos1

1ICU, 2Haematology, Red Cross Hospital, Athens, Greece

Introduction

Changes in biomarkers of coagulation, anticoagulation, fibrinolysis and inflammation occur in patients with systemic inflammatory response syndrome. The purpose of this prospective study is to evaluate the prognostic value of antithrombin III activity in severe ill patients.

Methods

Baseline levels of antithrombin III (AT III) in blood sample, of 45 critically ill patients was correlated with acute disease severity, as measured by Acute Physiology and Chronic Health Evaluation Scores (APACHE II), and with organ failure severity, as measured by The Sequential Organ Failure Assessment (SOFA) score. The quantitative determination of the antithrombin activity level in plasma was measured by the synthetic chromogenic substrate method. Correlation between parameters was made by linear regration analysis and comparison of antithrombin levels in all groups was made by t-test (p < 0.05).

Results

Trauma, lung infection, brain injury and abdomen infection was the main underling patients’ diseases. On the time of their admission in ICU the patients had mean APACHE II score 22.22±5.68, mean SOFA score 9.11±3.10 30 and mean baseline levels of antithrombin III 71% ± 22%. 30 of 45 patients (64.23%) died in ICU after 21.02±12.56 days of hospitalization. The mortality of patients with baseline levels of antithrombin III less than 60% (14 patients) was 78.58% and the mortality of patients with baseline levels of antithrombin III more than 60% (31 patients) was 58%. The characteristics of the four groups of patients are shown in the Table 1.

There is not a statistical important correlation between baseline levels of AT III and the outcome of the patient. Also there is not a statistical important correlation between baseline levels of AT III and APACHE II severity score. There is statistical important correlation between baseline levels of AT III and the SOFA score in the group of patients with multiple trauma. The baseline levels of AT III in patients with brain injury is higher compared with the other groups of patients (p < 0.05).
 

PATIENTS

AGE

AT III

APACHE II

SOFA

Lung infection

14

67.92±14.84

64%±22%

23.28 ±5.01

10.07 ±2.52

Abdomen infection

7

71.51±11.94

59%±12%

22.14±6.56

9.14±4.48

Brain injury

14

60.5±10.93

89%±23%

24.42±5.59

10.07 ±2.01

Multiple trauma

10

47±29.59

62%±16%

17.7±15.28

6.4±2.57

Conclusion

Although a higher mortality rate is found in patients with baseline AT III activity of less than 60% AT III activity measured as a single blood marker cannot be used as a prognostic parameter. APACHE II is a superior to AT III activity prognostic parameter.

0534

Plasma Brain Natriuretic Peptide (BNP), C-Reactive Protein (CRP) and High Density Lipoprotein (HDL) as Prognostic Indicators in Long Term ICU Patients

M. Karvouniaris*1, A. Tefas1, M. Stougianni1, S. Xitsas2, P. Kasviki2, D. Lagonidis1

1ICU of Giannitsa, 2Laboratory, General Hospital of Giannitsa, Giannitsa, Greece

Introduction

Biomarkers have recently showed promise as a complementary prognostic tool in ICU patients. Both CRP and HDL have a role as indicators of inflammation and BNP levels rise not only in cardiac dysfunction but as a response to proinflammatory stimuli.

Methods

Methods: Prospective study of 22 patients that stayed in the ICU for at least 7 days and had the following characteristics: Age 63.4±12.9 years, length of stay 23.2±11.0 days, APACHE score 20.7±7.8.

CRP, HDL and BNP were measured both in admission and on day 7.

First, we correlated the above parameters with the length of stay (LOS) in the ICU by using Pearson's correlation test. Secondly, we compared the means between survivors and nonsurvivors after 6 months with independent samples’ t-test. We finally performed receiver operating curves (ROC curves) of the above parameters according to mortality.

Results

Mortality in the ICU was 22.73% and six-month mortality 63.63%. The values of the parameters are showed in Table 1.

No correlation could be found between the LOS and the aforementioned parameters, neither a difference could be found between survivors and non-survivors with the independent samples’ t-test (p>0.05).

ROC curves are shown in Table 2.
Table 1

Values of the Parameters

 

BNP (pg/ml)

CRP (mg/dl)

HDL (mg/dl)

Day 1

488.28±806.76

14.21±15.13

29.83±15.57

Day 7

676.62±872.82

11.16±7.93

21.04±13.85

Table 2

ROC Curve Values

BNP

CRP

HDL

 

Day 1

0.580

0.545

0.491

Day 7

0.598

0.652

0.696

Conclusion

Measuring these biomarkers on admission does not predict a long ICU stay. Only CRP and HDL on day 7 might add prognostic information in this group of patients.

0535

Long-Term ICU Complications after Prolonged Critical Illness: Post-ICU Assessment in an Outpatient Clinic Setting

D. H. T. Tjan*, L. E. M. van Lelyveld-Haas, C. W. Buiter, A. R. H. Zanten van Intensive Care, Gelderse Vallei Hospital, Ede, Netherlands

Introduction

Survivors of critical illness may experience a spectrum of morbidities that may reduce quality of life and prolongs the recovery after discharge from the hospital. p ]In November 2007 our ICU started a post ICU outpatient clinic to evaluate specific problems of patients experienced after discharge to improve their medical and psychological status.

Methods

Survivors after severe sepsis with than 10 days mechanical ventilation were identified and invited to visit the clinic. A questionnaire was mailed to the participants with questions regarding daily living activity (ADL and Barthel). Medical and psychological aspects were discussed during the visit.

Results

We evaluated the first 24 patients. Preliminary analysis shows that survivors experience a significant range of especially psychological and medical complaints within the first three months after discharge. Problems frequently encountered were weight loss > 10% (38%), critical illness polyneuropathy (42%), sleep disturbances (29%), post-traumatic stress disorder (29%), and signs of emotional imbalance (20%). New medical problems were less common and not severe. Only nine patients (38%) were discharged with maximum IADL and Barthel score. Duration of complaints varied between 2 weeks up to one year after discharge.

Conclusion

After prolonged critical illness significant morbidity may influence recovery and rehabilitation following discharge from hospital. Quality of life including ADL must be considered when assessing ICU outcome. At present this problem may be underestimated and more studies are warranted to improve post-ICU care.

0536

The German Translation of the Cam-ICU to Monitor Delirium in Intensive Care Patients — Applicability and Time Consumption

U. Guenther*, H. Wrigge, T. Muders, C. Putensen

Anaesthesiology and ICM, University Hospital, Bonn, Germany

Introduction

Delirium is associated with longer stay in ICU, higher treatment costs and increased mortality up to six months [1,2]. Though recommended by guidelines [3], very few ICUs have implemented a daily monitoring for delirium yet. When introducing a delirium-monitoring tool, objections such as “too time-consuming” and “too complicated” will be encountered, and it is often stated, that clinical judgment alone would prove satisfactory to detect delirium. We tested the German translation of the “Confusion Assessment Method for the Intensive Care Unit (CAM-ICU)” [4] in a 20 bed surgical intensive care unit (ICU) for applicability, time-consumption, and compared clinical judgment of nurses to detect delirium with CAM-ICU assessment.

Methods

Every patient admitted to our ICU in June 2007 was monitored for delirium with the CAM-ICU on a daily basis. (1) Nurses were interviewed for their delirium criteria, and they assessed their patients for delirium based on their clinical judgment. Patients with severe neurological impairment (stroke or dementia) and inability to understand German language were excluded. (2) To assess applicability of the CAM-ICU, nurses familiar with the CAM-ICU were stop watched when examining patients and interviewed for ease of use of each of the four CAM-ICU features.

Results

69 patients were eligible for analysis (mean±SD; 68 ±12,2 yrs; 79 ±17 kg BW; 40 male, 29 female). (1) 43% of patients developed delirium at some time during their ICUstay, 58% were intubated. 13% of patients were deemed non-delirious if assessed only by clinical judgment, but were in fact delirious. Nurse's subjective criteria for delirium included confusion, disorientation, altered level of consciousness, inappropriate speech and inability to concentrate on a brief communication. (2) 18 Nurses familiar with the CAM-ICU rated the CAM-ICU as easy to understand and use, and required a mean 2:30 min (range; 1:15 – 3:00 min) to complete all 4 features in delirious patients.

Conclusion

More than a quarter of delirious patients were deemed non-delirious by mere clinical judgment. These patients had predominantly a hypoactive subtype of delirium as assessed with the Richmond-Agitation-Sedation-Scale. No more than 3 minutes were required to complete all four features of CAM-ICU. Mostly, however, CAM-ICU can be cut short because the diagnosis is made with the first one to three features. A daily delirium monitoring with the CAM-ICU is not time-consuming, can be performed by nurses, and detects a significant number of delirious patients, which go undiagnosed otherwise.

0537

The Incidence of Delirium in Intensive Care Patients and Association with GABA Agonist Administration

J. Snell*, D. Thorburn, R. Wenstone, J. Walker

Intensive Care Unit, Royal Liverpool University Hospital NHS Trust, Liverpool, United Kingdom

Introduction

The reported incidence of delirium in non-comatose intensive care patients ranges from 19% to 80%1,2 and has been associated with GABA agonist (propofol and benzodiezepine) use3. Delirious patients may not be agitated, so signs of delirium must be actively sought. The Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) is a validated screening tool4. The aim of this study was to use the CAM-ICU in a 13 bed mixed surgical and medical ICU to determine the incidence of delirium and whether there was any correlation with sedative agent use.

Methods

JS and RT (trainee ICU doctors who had had a brief tutorial on CAM-ICU) reviewed each ICU patient daily and performed a CAM-ICU if assessable (Richmond Agitation and Sedation Score more than minus 4 = rousable enough to assess). Sedative agents given in the preceding 24 hours were recorded. Statistical significance was determined with Fisher's exact test.

Results
186 reviews of 45 patients were made over 6 weeks; in 117 of these (63%) the patients were rousable enough to perform a CAM-ICU. On 16 of the 117 CAM-ICU assessments (14%) the patients were delirious, and on 12 of these (75%) the delirium was of the hypoactive type (patient calm or drowsy). 11 of the 45 patients (24%) were delirious at least once. 7 out of 24 assessments (29%) on patients who had received a GABA agonist in the previous 24 hours were positive for delirium, compared to 7 out of 66 assessments (11%) on patients who had received no drug in the previous 24 hours (p=0.047), and 2 out of 27 assessments (7.4%) on patients who had received any sedation that did not include GABA agonists (p=0.066) in the previous 24 hours.
Table 1

Incidence of Delirium When Assessed, Categorised by Sedative Administration

 

Delirious at time of assessment n (%)

Not-delirious at time of assessment n (%)

All patients

16(14)

101(86)

No sedative drug given in previous 24 hours

7 (11)

59 (89)

GABA agonist given in previous 24 hours

7 (29)

17 (71)

Any non-GABA agonist sedative drug given in previous 24 hours

2 (7)

25 (93)

Any sedative drug given in previous 24 hours

9 (18)

42 (82)

Conclusion

The incidence of delirium on this intensive care unit was within the range previously reported. Significantly more patients who had received a GABA agonist were delirious compared with those who were sedative free.

Poster Sessions Technology assessment organ support I: 0538-0551

0538

Sedation and Analgesia Practice in European Intensive Care Units

D. Nethirasigamani*, V. Mani, A. Abasalom

University Department Anaesthesia, Addenbrookes Hospital NHS Trust, Cambridge, United Kingdom

Introduction

We conducted an e survey among ESICM members to identify and characterise the practice of sedation and analgesia in the critical care units.

Methods

After obtaining the approval of the Research Committee of ESICM, the e-survey questionnaire was published on the ESICM website for participation of the ESICM members from 1st December 2007 to 31st January 2008. There were 32 questions in the survey.

Results

92 members responded, of whom 80% were senior medical staff, 50% of the respondents were Anaesthesiologists and 35% with ICU as their primary speciality. Around 90% of the respondents worked in the ICUs which admitted both medical and surgical patients. Only 50% of the respondents had a written protocol for sedation in their units and assessed sedation level every 4-6 hours. A sedation scoring system was used by 90% of the respondents, with Ramsay scoring being the most frequently used system (60%). 21% of the respondents use Bispectral monitor to assess the sedation level, of which 50% were Neurocritical care units. There were only 13% of the respondents who preferred to use propofol for sedation in children, and the youngest age they would use ranged from 6 months to 16 years in children. Only 50% of the respondents admitted that they would adjust the calorie requirement when patients are on propofol infusion. 22% of the respondents have encountered propofol infusion syndrome in their clinical practice. 53% used rise in lactate and 3% used Acyl or malonyl carnitine levels as a marker, if they suspected propofol infusion syndrome. 96% of the respondents used continuous infusion for sedation, with propofol being the preferred drug used by 83% of them for short term sedation (less than 48 hours). Midazolam was the second common drug used for short term sedation (around 58%) but it was the preferred drug (85%) for long term sedation (more than 48 hours). The respondents felt that haemodynamic status (90% of them) and cost (63% of them) did influence their choice of sedation. Only 60% of the respondents were practising sedation hold regularly and around 30% hardly practised sedation hold. Haloperidol was the preferred drug (70%) for treating agitation in the intensive care unit, with clonidine being the second choice (30%).

Atracurium was the neuromuscular blocker preferred by the respondents, with approximately 80% of them used it only intermittently. One third of the respondents were not monitoring neuromuscular blockade and only 50% used peripheral nerve stimulator to monitor the neuromuscular blockade.

Conclusion

There is wide variation in sedation and analgesia practice. There is also substantial difference in the way sedation and neuromuscular blockade are being monitored. Considering the potential benefit in the patient outcome, it might be worthwhile implementing evidence based protocol in the clinical practice. Open image in new window

0539

Effects of Gas Scavenging Systems on the Functioning of ICU Ventilators

J. Arnal*1, M. Bisbal1, M. Wysocki2, D. Demory1, J. Durand-Gasselin1

1Réanimation Polyvalente, Hopital Font Pré, TOULON, France, 2Medical Research, Hamilton Medical, Bonaduz, Switzerland

Introduction

When inhaled sedation is used with intensive care ventilators, gas scavenging systems are recommended in order to minimize occupational exposure (1). Theses devices are dedicated for anaesthesia ventilators and may disturb the functioning of ICU ventilators. This prospective cross-over study compared the effects of charcoal adsorption and active suctioning system on PEEP level and stability during mechanical ventilation.

Methods

20 patients (age: 64 [50 – 76], SAPS II: 54 [43 – 62]) requiring deep sedation (Ramsay score 5 or 6) were included. Patients were ventilated with a Galileo Gold or G5 ventilator (Hamilton Medical) in Adaptive Support Ventilation (ASV). The AnaConDa® filter (Sedana Medical) was used for heat and moisture and to deliver Sevoflurane. Measurements were done without scavenging system (w/o scavenging) and after 30 minutes of Sevoflurane infusion, using in random order a charcoal adsorption canister (w/charcoal) (Cardiff, Smiths Industries) or an active suctioning system (w/suction) (Scavenging tubing system, Hamilton Medical) connected to the gas outlet. Respiratory mechanics, tidal volume and PEEP delivered were measured at 67 hertz using a proximal pneumotachograph. Comparison between PEEP delivered and PEEP set was assessed by the average of 5 expiratory cycles. PEEP oscillations were evaluated by the standard deviation from the mean. Autotriggering was assessed by visual inspection and flow-pressure traces analysis. Results are given in medians [25th – 75th quartile] and comparison used Signed Rank Test.

Results

PEEP delivered was decreased w/charcoal as compare to w/o scavenging and w/suction (6.0 [4.9 – 10.0], 6.8 [5.2 – 10.1], 6.6 [5.2 – 10.2] cmH2O respectively; p = 0,001). PEEP oscillations were increased w/charcoal as compared to w/o scavenging and w/suction (0.8 [0.5 – 1.8], 0.3 [0.3 – 0.4], and 0.3 [0.3 – 0.4] cmH2O respectively; p < 0,001). Autotriggering occurred in 10 patients w/charcoal and did not occured w/o scavenging and w/suction. The expiratory time constant was decreased w/charcoal as compared to w/o scavenging and w/suction (0.57[0.50 – 0.82], 0.75[0.66 – 0.93], and 0.77[0.64 – 0.92] s respectively; p = 0.02), which resulted in different tidal volume delivered (499 [405 – 578], 523 [427 – 594], and 541 [436 – 587] mL respectively, p = 0.02). Plateau pressure, inspiratory resistances, static compliance and inspiratory time constant were not modified.

Conclusion

Charcoal adsorption disturbs the functioning of ICU ventilators. For anaesthesia gas scavenging in ICU, an active suctioning system should be preferred.

0540

Application of Ultrasound-Guided Pigtail Catheter for Drainage of Pleural Effusions in The ICU

W. Chen*, S. Liang

Department of Internal Medicine, China Medical University Hospital, Taichung, Taiwan

Introduction

Little is known about the effectiveness of the pigtail catheter for drainage of pleural effusions in the intensive care unit (ICU).

Methods

We conducted a retrospective review of adult patients (≥18 years) who underwent ultrasound-guided pigtail catheter drainage of pleural effusions in the medical and surgical ICUs from January 2005 to July 2007 in a university hospital.

Results

Among the 133 enrolled patients, there were 93 (70%) males and 40 (30%) females, with the mean age of 63.7±15.4 years old. The reasons for pigtail drainage were as follows: complicated parapneumonic effusion or thoracic empyema (n=59, 44%), massive transudative pleural effusions (n=33, 25%), postoperative pleural effusion (n=29, 15%), malignant pleural effusion (n=18, 14%) and traumatic hemothorax (n=3, 2%). In comparing the total amount of fluids drained, the duration of drainage, success rate and complication rate among these different causes of pleural effusion, pigtail drainage for massive transudative pleural effusion yielded the largest amount of pleural fluids (5382±4844 ml), provided the longest duration of drainage (9± 7 days), and had the highest complication rate (18%). The success rate was highest when used to treat traumatic hemothorax (100%) and postoperative pleural effusions (85%); drains inserted for empyema were more likely to fail (overall success rate, 42%). No significant insertion complication, such as hollow organ perforation, was caused by this procedure.

Conclusion

The ultrasound-guided pigtail catheters are useful in the drainage of a variety of pleural effusions in critically ill patients. Using ultrasonic guidance, insertion complications, such as hollow organ perforation and pneumothorax, can be reduced significantly. Drainage for massive transudative pleural effusions should be of serious concern, because it may lead to prolonged drainage and infection. Thoracic empyema remains a difficult clinical problem, and pigtail catheter drainage of this empyema was only partly effective in our study.

0541

Higher Peep Levels Results in Small Increases in Intrabdomial Pressure in Critical Care Patients

C. Ferrer*, E. A. Piacentini, E. Molina, J. Trenado, B. Sanchez, J. M. Nava Critical Care, Hospital Mutua Terrassa, Terrassa, Spain

Introduction

In recent years, the measurement of intraabdominal pressure (IAP) and abdominal perfusion pressure (APP) in critically ill patients has become important because of the high risk of developing abdominal compartment syndrome (ACS). These patients are frequently affected of acute lung injury (ALI) or acute distress respiratory syndrome (ARDS)(1). Recent studies have proved that higher PEEP levels are better than lower PEEP levels to treat patients with ALI/ARDS (2). Transmission of IAP to the thorax has an impact on the respiratory system, and changes in intra-thoracic pressure may affect IAP, but the effects of PEEP on IAP and APP measurement are not clear. OBJECTIVE: To evaluate the effect of increased PEEP levels on IAP measurement.

Methods

Mechanically ventilated and hemodynamically stable patients were included. Patients with high intracranial pressure, in prone position or with open abdominal wall were excluded. IAP and mean arterial pressure (MAP) were measured at different PEEP levels (5, 10, 15 and 20 cmH2O). After each change in PEEP a set of measurements was recorded after a stabilization interval of 5 minutes. After the last measure, PEEP was returned to basal level. IAP was measured according to the method described in The International Conference of Experts on Intra-abdominal Hypertension and Abdominal Compartment Syndrom (WSACS)(1).

Results

12 patients were studied. MAP and APP were stable throughout the study in all cases. Plateau pressure increases proportionally with PEEP. IAP increased with PEEP 15 and PEEP 20 cmH2O. No complications or adverse events were observed.

* p < 0.05 vs previous value. Open image in new window

Conclusion

Increases in PEEP levels were associated with significative increases in IAP without changes in APP measurements. These changes had no clinical relevance when IAP was within the normal or mild higher range. The main limitation of our study was the abs ence of patients with severe intra-abdomial hypertension or abdominal compartment syndrome.

0542

Evaluation of the User-Friendliness of New Generation ICU Ventilators

L. Vignaux*, P. Jolliet, D. Tassaux

Intensive Care Unit, Hôpital Cantonal universitaire de Genève, Genève, Switzerland

Introduction

In the ICU setting, incidents are often caused by human errors, often related to mechanical ventilation (1). ICU ventilators are becoming increasingly sophisticated and complex, which can lead to errors, particularly in emergency situations. The aim of this study was to evaluate the user-friendliness of 7 new generation ICU ventilators.

Methods

10 Physicians well trained in mechanical ventilation, but without knowledge of the 7 ICU ventilators tested, performed 8 tasks per machine. Time performances were compared between each other and with a reference time established by a trained respiratory therapist (RT). Tasks were to successively: switch on the ventilator, recognize the already adjusted mode, recognize and set alarms, change mode, find the pre-oxygenation command, adjust parameters for pressure support mode, stand-by and find NIV mode. Physicians rated their subjective assessment of task difficulty on a visual analogue scale (0: very easy-10 very difficult).

Results

Results are expressed as medians and interquartile range (IQR: 25th–75th percentile). For each task physicians were slower than the RT: 44 (25–89) s for the most rapid, vs. 14 (5–22) s for the RT. A mean of 9 (7–13) failures was observed by ventilator; on 3 machines more than 10 failures occurred. The most rapid task with the fewest failures was pre-oxygenation (5, 3–17 s). Tasks with the most failures were adjusting pressure support mode and finding NIV mode. The longest task was mode recognition (106, 74–146). The most common errors were: confusion between adjusted and measured parameters, type and level of trigger and plateau pressure assessment. Visual analogue difficulty scores ranged from 3.8 to 7.3.

Conclusion

Physicians without prior experience with specific ICU ventilators perform poorly when confronted with specific tasks. These results suggest a need for standardization between machines and improved interface user-friendliness in the design of ICU ventilators.

0543

Ultrasonography of the Neck Prior to Percutaneous Tracheostomy

M. F. Pollard*, J. Naisbitt, H. N. Thomson, S. Laha, P. Bunting

Critical Care Unit, Royal Preston Hospital, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, United Kingdom

Introduction

Percutaneous tracheostomy is a common but potentially hazardous procedure. Major haemorrhages have been reported requiring emergency surgical intervention to prevent significant morbidity or mortality 1. Aberrant vessels may be missed despite adequate surface assessment before the procedure. Portable ultrasound machines are available in all intensive care units in compliance with NICE guidelines to assist with central venous access. It has been suggested that the incidence of major bleeding may be reduced by routine ultrasonography of the anterior neck prior to the procedure2, 3.

Methods

A prospective audit was performed in our critical care unit. A questionnaire was completed for every tracheostomy performed over a three month period. Recorded information included, any abnormalities observed during surface examination, the ultrasonographic findings including tracheal depth, complications during the percutaneous procedure and whether an open surgical tracheostomy became appropriate as a result of clinical or ultrasonographic examination. Coagulation abnormalities were corrected prior to any procedure in line with standard practice.

Results

39 tracheostomies were performed within the study period. Percutaneous tracheostomies were performed using the single “rhino” dilatation technique. 3 patients were referred for surgical tracheostomies on the basis of their history. In 11% of the remaining cases ultrasonography was unavailable. Of the remaining 30 patients, 12 patients had abnormal anatomy detected on ultrasonography resulting in 8 changes to technique or practice. No patient received a transfusion as a result of any procedure, however 1 patient developed post tracheostomy surgical emphysema. The mean tracheal depth was 1.88 86cm (range 4cm-1cm).

Conclusion

We describe a safe non-invasive technique using available resources at little extra cost and may reduce complications associated with percutaneous tracheostomy. We have demonstrated a change in practice in 26.6% of patients and 3 patients had a surgical tracheostomy as a result (10%).

0544

Delivering Humidified High Flow Therapy at Increasing Gas Flow Rates Generates Higher Airway Pressure

R. L. Parke*1, S. P. McGuinness1, M. L. Eccleston2

1Cardiothoracic and Vascular Intensive Care Unit, Auckland City Hospital, 2Respiratory Humidification, Fisher and Paykel Healthcare, Auckland, New Zealand

Introduction

A recent study in cardiac surgical patients demonstrated that a positive pressure is delivered during humidified high flow oxygen therapy (HHFT)(1).

This study set out to determine the relationship between flow and pressure in cardiac surgical patients receiving HHFT via a nasal interface.

Methods

Following ethical approval, 12 patients scheduled for elective cardiac surgery were consented and enrolled into this study. Whilst sedated and ventilated post-operatively a 10French catheter was inserted into the nasopharynx via the nose. Pressure measurements were performed once the participant was awake and extubated. Placement of the catheter was first confirmed using end tidal CO2 monitoring. The Fisher and Paykel Healthcare Optiflow TM system was used to deliver humidified nasal oxygen and measurements were performed with gas flow rates of 30, 40 and 50 lpm. Measurements were performed with patients mouth open and mouth closed. Pressure was recorded over one minute of breathing. The mean nasopharyngeal airway pressure was determined by averaging the pressure over one minute. This allowed the entire pressure profile of each breath to be included within the mean airway pressure calculation.

Results

n = 12. Average age = 64.6years (40 – 84); average height = 174.3cms (163 – 205); average weight = 86.3kgs (55 – 121). 75% (n = 9) male.

At 50 lpm with mouth closed, the Optiflow TM system with nasal interface delivered a mean nasopharyngeal airway pressure of 3.3 cmH20.

Other results were:
 

30 lpm

40 lpm

50 lpm

Mean Airway Pressure

1.0

1.3

1.7

Mouth Open (cmH2O)

   

Mean Airway Pressure

1.9

2.6

3.3

Mouth Closed (cmH2O)

   
Conclusion

This study demonstrates that the pressure delivered by the Optiflow TM Humidified High Flow therapy system does increase as flow rate is increased in a group of cardiac surgical patients.

Grant Acknowledgement

Research in the Cardiothoracic and Vascular Intensive Care Unit is supported in part by Fisher and Paykel Healthcare.

0545

Lung Sound Distribution Shifts to Lower Lung Regions with Increased Peep

S. Lev*, J. Cohen, I. Kagan, M. Grinev, P. Singer ICU, Rabin Medical Center, Petach Tiqva, Israel

Introduction

Vibration response imaging (VRI) maps the lung sound distribution during the respiratory process. VRI measurement has been proved sensitive to changes in PEEP. In the present study, we compared the lung sound distribution at different PEEP levels. Our hypothesis was that elevation of PEEP is associated with better recruitment of dependent lung areas and should be reflected by increased lung sounds in these areas.

Methods

Lung sounds distribution was assessed at peak inspiration in three lung regions (apex, central and base). Maps of lung sounds obtained at PEEP 0, 5 and 10 cm H2O were compared in 34 mechanically ventilated patients. Furthermore, maps of two repeated measurements at each PEEP level were also evaluated.

Results

Lung sound distribution in the base areas was significantly increased when increasing PEEP (paired t-tests). Furthermore, lung sound distribution in the apex areas was significantly decreased. Central lung segments did not change with PEEP. Furthermore, maps of repeated measurement at same level of PEEP were not significantly different. Open image in new window

Conclusion

Lung sound distribution shifted from apex to base when increasing PEEP probably reflecting increased flow in these areas.

0546

Visual Tracking System of Emergency Room Patient Flow Significantly Reduced Waiting Times at Emergency Department

K. Vanstraelen*1, C. De Deyne1, G. Vundelinckx2, E. Vandermeersch3, R. Heylen4

1Anaesthesiology, 2Emergency Medicine, Ziekenhuis Oost-Limburg, Genk, 3Anaesthesiology, University Hospital Louvain, 4Anaesthesiology, Ziekenhuis Oost-Limburg, Louvain, Belgium

Introduction

Patient satisfaction at emergency departments (EDs) can be improved by reduction in patient waiting times. In this paper, we evaluated the impact of an ED informatisation system, available for all ED (medical and nursing) staff and displaying a visual tracking system of all pts admitted to the ED.

Methods

We retrospectively analysed all different waiting times for all patients admitted to the ED in a 3months periods before and after the installation of the visual tracking system. Different waiting times were : wR : wait for registration, wX : wait for initial triage, wC : wait for consulting room, Wi : wait for technical investigation, divided in time of technical investigation and time to result of technical investigation available, wT : wait for diagnosis/treatment and finally wD : wait for discharge from ED (home or hospital).

Results

A total of 4720 pts were included in the first 3months period (2005), compared to 4910 pts for the second period (2006). We observed a significant decrease in all, except one (wR), waiting times after the installation of the visual tracking system. Analysis of different waiting times revealed that largest reductions were obtained in the wait for initial triage and consult (wX : m28min reduced to m19min and wC : m43min reduced to m27min). We also observed a significant reduction in outliers, i.e. extremely long waiting times, mostly occurring for technical investigations (wI) (as well for the investigations as for the results of the investigations) as for final diagnosis (wT).

Conclusion

The implementation of an ED informatisation system, providing continuously real-time updated visual tracking screens displaying the flow of all ED pts, resulted in signicificantly reduced waiting times and increased patients satisfaction at the Emergency Department.

0547

Critical: Common Resources for Interactive Training of Intensive Care Staff at All Levels

J. A. Wright*1, D. E. F. Newton2

1Research and Development, North West London Hospitals NHS Trust Northwick Park Hospital, 2Anaesthesia and Critical Care, Northwick Park Hospital, Harrow, United Kingdom

Introduction

The practice of teaching and training medical and nursing staff in the intensive care domain currently consists of bedside teaching with real-life patients. Presently, simulators available are expensive or difficult to access for the majority of staff. The search for a “virtual patient” suitable for the teaching needs of senior intensive care educators is still not available. While high fidelity simulators for anaesthesia and trauma are available, and could be adapted, their cost and location means that only small numbers of learners can participate.

Methods

The 4 European partners of the CRITICAL project aim to deliver a screenbased, digital Virtual Intensive Care Patient (VIP) with a wide range of physiological attributes, capable of being reconfigured to treat the variety of pathological states which result in different patient symptoms and outcomes.

The project uses a software toolset called EduCAT which allows teachers who are not computer specialists the flexibility to create their own interactive educational material. Via the web, these intensive care tutors in hospitals and universities worldwide can provide trainees with a readily available, low cost approach to medical management of different disease states. They will be able to try alternative treatment options and observe patient outcomes faster than in “real life” without any danger to patients in the “real” clinical environment.

Prominent international leaders of ICM will head the project and develop medical-based scenarios which have been rigorously tested, evaluated and refined throughout the project lifecycle.

Results

By the project end in August 2009, the result will be a developed software simulation of an intensive care patient, available in 4 languages (English, French, German and Spanish) which will be configured by the teacher and used directly for diagnostic exercises set by the teacher.

Conclusion

The VIP will be accessed over the internet from a website which will also provide a collaborative forum for interested education groups to further develop the model and share latest research of best practice bringing intensive care teaching into the 21st Century.

Grant Acknowledgement

We gratefully acknowledge the assistance of ECOTEC UK (Leonardo da Vinci Programme) who have provided 75% of total funding for this project.

0548

High Frequency Chest Wall Oscillation is not Better Than Manual Hyperinflation in Atelectatic Ventilated Patients

A. van Hees*, B. Speelberg

Intensive Care, St.Elisabeth Hospital Tilburg, Tilburg, Netherlands

Introduction

Atelectasis is a frequent occurring disease in ventilated patients. There is no standard treatment for this disease. The treatments of choice are :suctioning, positional therapy, bronchoscopy and conservative treatment. The patient with atelectasis suffers from impaired ventilation capacity and impaired oxygenation in most cases. Therefore the atelectasis should be cleared. We investigated the role of a new external pneumatic oscillation device, The Vest® system compared with suctioning in combination with exprimation and hyperinflation, which is the standard therapy for atelectasis in our intensive care unit.

Methods

We performed a single centre randomised open study to investigate the effects of The Vest® system in atelectatic ventilated patients who were hospitalised in our ICU in 2007. After approval of our ethical committee, 6 patients were studied with the Vest® system. They were treated 3 times a day with the use of an external vest, which contained an air chamber. A pulse generator was connected with the airspace in the Vest®. In this way pulsatile, oscillative compressions were made around the thorax, during 20 minutes. Standard manual hyperinflation in combination with exprimation and suctioning was done 3 times a day in 6 patients. After maximal 12 treatments, the study was stopped. Complete atelectasis of one lobe was scored as 3 points, major atelectasis of one lobe was 2 points and minor atelectasis of one lobe was scored as one point. Complete atelectasis of the right lung should be scored as 9 points. One independent radiologist reviewed the X rays. Atelectasis score and time to reach the lowest score were compared with Mann Whitney U testing. A significance level of < 0,05 was considered as significant.

Results

Atelectasis score just prior to treatment minus lowest atelectasis score was 1,3 +/− 1,5 in the Vest® group and 2,5 +/− 2,4 in the conservative treated group. Time to reach the lowest atelectasis score was 77,3 +/− 23 hours in the Vest® group and in the standard treated group 58,7 +/− 31,1. No significant changes in both values were seen in both groups.

Conclusion

In this study no benefit in resolution in atelectasis has been demonstrated when the Vest® system was used compared to manual hyperinflation in combination with exprimation and suctioning. The sample size in this study was small. Treatment duration and frequency of chestwall oscillations could also be of influence on our results.

0549

Breath Sound Analysis Detects Injury and Recruitment in the Lung During Mechanical Ventilation

A. Vena*1, G. Perchiazzi2, C. Rylander3, R. Giuliani2, T. Fiore2, A. Magnusson4, G. Hedenstierna5

1Intensive Care Unit, SS Annunziata Hospital, Taranto, 2Emergency and Organ Transplant, Bari University, Bari, Italy, 3Anaesthesia and ICM, Sahlgrenska University Hospital, Göteborg, 4Radiology, 5Medical Sciences — Clinical Physiology, Uppsala University, Uppsala, Sweden

Introduction

Acute lung injury may change the acoustic properties of the respiratory system. Pathologic phenomenons, such as “crackles”, are usually heard on auscultation. The aims of this study were to evaluate whether computerized analysis of breath sounds can help in 1) localizing lung injury and 2) detecting the recruiting effect obtained by raised positive end-expiratory pressure (PEEP) and confirmed by computed tomography (CT).

Methods

Microphones were applied to four locations on the chest of six anaesthetised pigs, ventilated in volume-control mode with a PEEP of 5 cmH2O applied. Breath sounds were recorded before and after unilateral oleic acid-induced lung injury. Then PEEP was increased in steps of 5 cmH2O from 0 to 20 cmH2O. Breath sounds as well CT scans were recorded at each PEEP level. Pathological changes detected by Fast-Fourier-Transform (FFT) frequency analysis of the sounds were compared to the quantity of non-aerated and very poorly aerated (+100/−300 HU) lung tissue present in the 9 mm CT images.

Results
The acoustic spectra recorded in healthy condition were mainly within 400−500 Hz for both lungs. After unilateral oleic acid injury, the breath sounds recorded over the injured lung, but not over the non-injured lung, turned highly pathologic in all animals, showing increased spectral frequency content above 500 Hz and −70 dB. During stepwise increase of PEEP, the proportion of pathologic sounds gradually decreased and with a PEEP of 20 cmH2O, the breath sounds recorded over the injured lung were significantly decreased. The gradual decrease in acoustic pathology corresponded to a gradual transition of lung tissue from non-aerated or very poorly aerated to aerated in end-expiratory CT images (Table).

INJURED LUNG

PEEP 0

PEEP 5

PEEP 10

PEEP 15

PEEP 20

Pathologic sounds (FTT % > 500 Hz)

32.0 ± 7.3

36.2 ± 7.0

33.6 ± 9.0

27.6 ± 14.5

12.5 ± 9.5*

Non/very poorly aerated lung (ml)

22.7 ± 8.2

19.5 ± 8.0

15.4 ± 5.8*

6.1 ± 3.8*

2.2 ± 1.2*

Data are mean +/− SD * = p < 0.05 compared to PEEP 0 (cmH2O) by Wilcoxon test

Conclusion

Quantification of pathological breath sounds by computerized analysis can be helpful in localizing lung injury and recruitment obtained by raised positive end-expiratory pressure in volume-control ventilation, thus positioning the tidal inflation above the level of end-expiratory collapse of injured lung.

Grant Acknowledgement

The Swedish Medical Research Council (5315).

0550

Circuit Lifetime Could be A Quality Indicator In Continuous Renal Replacement Therapy in the Critically Ill

E. M. I. Fässler*1, G. Kleger2

1Medical ICU, Kantonsspital St. Gallen, St. Gallen, Switzerland, 2,

Introduction

Continuous renal replacement is frequently used in critically ill patients with acute renal failure and sepsis. A frequent change of clotted circuits increases the workload and costs substantially. Circuit down time is the most important factor, compromising the cumulative filtration goal (35 ml/kg/h). Cumulative filtration rate, especially with the predilution mode is difficult to measure. Alternatively, circuit lifetime might serve as an easy to measure indicator to assess quality.

Methods

In our 12 bed medical ICU, we used CVVHDF in the pre- and postdilution mode (Prismaflex system, Hospal®). Circuit lifetimes were prospectively collected during a 12 month period. Unfractionated heparin (UFN) was the first choice. No anticoagulation was used in patients with severe coagulation abnormalities or hepatic failure, while regional citrate anticoagulation (CBA) was used in patients with recurrent circuit clotting or with bleeding predispositions. Regular change of circuits was planned after 72 h. We performed survival analysis. Circuit lifetimes were censored when changed because of interventions outside the ICU, a switch to palliative treatment or a 24 h change due to severe septic shock, but not when changed because of access or technical problems.

Results

38 consecutive patients and 167 circuits were observed. No bleeding or major metabolic complications were seen. There were no differences concerning vascular access site, the proportion of sepsis and vasopressor dependency between the anticoagulation groups. Consistent with the literature, circuit lifetime was longer and circuit patency rate higher in CBA. Our circuit lifetimes are higher than generally reported. Open image in new window

Conclusion

CBA is safe and has superior circuit lifetime and patency, compared to UFH. Beneath monitoring of the complication rate, measuring of circuit lifetime, processed by survival analysis tools is easy and feasible to assess quality of a highly complex procedure in critically ill patients.

0551

Effectiveness of Combined Acute Blood Purification Therapy by Direct Hemoperfusion Using a Polymyxin B-Immobilized Fiber Column and Continuous Venovenous Hemodiafiltration Using a Polymethylmethacrylate Membrane Hemofilter for Septic Shock Patients

Y. Sakamoto*1, K. Mashiko1, T. Obata2, H. Matsumoto1

1emergency And Critical Care Medicine, Chiba Hokusou Hospital, Nippon Medical School, Chiba, 2molecular Cell Biology, Institute Of Dna Medicine, Jikei University School Of Medicine, Tokyo, Japan

Introduction

Septic shock is a condition associated with diffuse coagulopathy and multiple organ failure, and frequently leads to death. Direct hemoperfusion using a polymyxin B-immobilized fiber column (DHP-PMX) was first developed in 1994, Japan and has been used for the treatment of septic shock. In particular, we usually start DHP-PMX within three hours at the diagnosis of septic shock. On the other, there is another kinds of hemofiltration column, the continuous venovenous hemodiafiltration using a polymethylmethacrylate membrane hemofilter (CVVHDF using PMMA), we have reported with the effectiveness of clinical treatment for crinically ill patients in both therapies.

Methods

We treated 34 septic shock patients with Acute Physiology and Chronic Health Evaluation (APACHE) II scores of over 20 by DHP-PMX. We reviewed the effectiveness of DHP-PMX therapy by comparison of the improvement rates of the APACHE II score, blood pressure. In another examination, the patients were divided into three groups: namely, group-1 in which CVVHDF using PMMA therapy following DHP-PMX (17 cases), group-2 in which CVVHDF using a polyacrylonitrile membrane hemofilter (PAN) therapy following DHPPMX (6 cases), and, group-3 in which CVVHDF was not performed after DHP-PMX (11 cases). The outcomes and improvement rates of the laboratory parameters (IL-6, protein C, PAI-1, endocannabinoids, HMGB-1 and oxidative stress) in the three groups were compared.

Results

The average of APACHE II score and sepsis-related organ failure assessment (SOFA) score were 30.2 and 12.7, respectively. The overall survival rate was 55.9% (good outcome judging from the APACHE II score). The improvement rates of the blood pressure (increased by more than 30 mmHg) were 58.8% (good effect judging from systematic review (1)). For another examination, only group-1 showed a better outcome (survival rate of 76.5%) compared with the other groups (survival rate of group-2 for 33.3%, group-3 for 36.4%). In addition, only group-1 showed significant improvements of the serum PAI-1 (155.6ng/ml before therapy vs. 90.5ng/ml on day 3; p=0.0041), protein C (34.2% before therapy vs. 40.0% on day 3; p=0.0433), IL-6 (13915.6pg/ml before therapy vs. 1052.1pg/ml on day 3; p=0.0477), N-arachidonoylethanolamine (AEA) levels(553.0pg/ml before therapy vs. 477.6pg/ml on day 3; p=0.0235) and oxidative stress marker (F2-isoprostane) (552.8pg/ml before therapy vs. 370.8pg/ml on day 3; p=0.0368).

Conclusion

Our study suggests that the early introduction of DHP-PMX treatment yielded good outcomes. In addition, therapies aimed at blood purification, CVVHDF and continuous hemofiltration (CVVHF) have been reported to be effective for the removal of inflammatory cytokines and various mediators. Although several types columns have been used for CVVHDF, few reports have shown the influence of the column using CVVHDF on the removal rate of the above-mentioned factors. Our findings suggest that CVVHDF using PMMA following DHP-PMX treatment have removed of various factors, and has improved survival rates in severe septic shock patients.

Poster Sessions Pediatric and neonatal care: 0552–0563

0552

Parents That Stay Continuously with Their Child in the Picu Reduces Their Distress

A. P. Serafim*1, M. F. Lira1, V. Coradini1, D. C. Junior2

1Pediatric Intensive Care Unit, Hospital Regional da Asa Sul, 2Department of Pediatrics, University of Brasilia, Brasilia, Brazil

Introduction

The aim of this study was to evaluate whether the distress presented by children in the Pediatric intensive care unit (PICU) was affected by the parents staying continuously with their child or not.

Methods

Design: Prospective, observational study. Setting: PICU at Hospital Regional da Asa Sul (Brasilia, Brazil).

Patients: A convenient sample of children admitted to the PICU from August 2002 to January 2003 aged 1 to 153 months.

Study Design: The sample was divided in two groups. One group was constituted of children of parents that stayed continuously with their child while in the other group parents didn't stay. The decision of staying or not with the children was made freely by the parents. Informed written consent was obtained in all the cases.

Measurements: Demographic data was collected for all the patients. Distress was assessed by obtaining COMFORT scale scores for each patient in the morning. Blood samples for measurement of total cortisol and glycemic levels were also obtained each morning. The risk of mortality was assessed through PRISM scores. Statistical analysis was performed using Mann-Whitney's test. The chosen level of significance was P<0.05.

Results

Twenty children were enrolled, divided in two groups of 10. The medium age was 6.5 months, with no statistical significant difference between the groups. The average PRISM scores were equivalent for both groups (14 and 13.9). The therapeutical interventions for each group were similar, including the total amount of analgesics and sedatives administered. There was statistical significant difference between the COMFORT scale scores of the children which the parents stayed continuously and the other group (7.6 and 13.4, p= 0.0423). There was no significant difference between the cortisol of the children that stayed with parents and the children that stayed without parents (25.5 ± 16.2 and 52.2 ± 42.8, p= 0.07). There was no significant difference between the measured glycemia in the two groups (99.3 ± 20.7 and 111.1 ± 34, p= 0.21). The average length of stay in PICU was significantly different (7 days in the staying parents group vs. 11 days in the other group).

Conclusion

The patients of both groups showed signs of distress. Comparison between groups suggests that parents staying continuously with their child in PICU can reduce their level of distress and also could shorter their lenght of stay.

0553

Impact of Nonpulmonary Organ Failure on Outcome in Children Treated with High Frequency Oscillatory Ventilation

A. Khaldi*, A. Bouziri, K. Kazdaghli, A. Hamdi, S. Belhadj, K. Menif, N. Ben Jaballah Pediatric Intensive Care, Children's Hospital, Tunis, Tunisia

Introduction

High Frequency Oscillatory Ventilation (HFOV) have been widely used in pediatric and neonatal patients with acute hypoxemic respiratory failure (AHRF) to limit lung injury and to improve oxygenation and CO2 epuration. Mortality associated with pediatric AHRF is also dependant on other organ failure rather than lung dysfunction. The objective of this study is to quantitate contribution of the nonpulmonary organ failure to poor outcome of pediatric patients with AHRF managed with HFOV.

Methods

Forty two consecutive pediatric patients (median age : 4 months, IQR:2–10) with AHRF (pneumonia: 35, sepsis with ARDS: 3, Others: 4), failing conventional ventilation (alveolar-arterial oxygen difference (P(A-a)O2) of 580 torr (453–645), Oxygenation index (OI) of 30 (22,5–37)) were ventilated with HFOV. A prospective and repeated record of oxygenation parameters and ventilator settings was made. Assessment of hemodynamic, renal, hepatic, neurological and haematological organ function was performed at regular times.

Results

Thirty four patients (81%) survived to hospital discharge without any oxygen dependency (group 1). Eight patients died (group 2). Nine patients (21%) had isolated respiratory failure and their mortality was 0%. Percentages of patients with 2, and 3 or more organ failure were 31%, 48% and their mortality was significantly higher, 7,6%, and 35% respectively. Patients with isolated respiratory failure demonstrated a significantly rapid and sustained improvement in oxygenation then patients of group 2. Severe shock requiring epinephrine or norepinephrine was associated to death (RR ratio of 8,5, CI[3,2–15,5]). Patients with 3 or more organ failure had a higher length of hospital stay than the other patients (13 vs 7 days, p:0,03).

Conclusion

Patients managed with HFOV for AHRF and with non pulmonary organ failure were significantly less likely to improve oxygenation on HFOV and had a significantly higher length of stay and mortality than patients with isolated respiratory failure. Children with such conditions must be identified early for other therapeutic considerations.

0554

Leukopheresis for Severe Pertussis: A Lifesaving Therapy

M. Grzeszczak*, F. E. Barr

Pediatric Critical Care, Children's Hospital at Vanderbilt, Nashville, United States

Introduction

The rate of severe pertussis infections has increased in recent decades. It mostly affects young children with the most severe cases and a highest mortality occurring in infants less than 6 months of age. Severe pertussis infections can cause respiratory and cardiovascular failure with an extremely high mortality rate (above 70%). The primary cause of death in severe pertussis is severe refractory pulmonary hypertension (PHT). PHT can progress very rapidly and in general is not responsive to any treatment modalities including extracorporeal support. The exact mechanism of PHT in pertussis is not completely known but hyperleukocytosis with leukostasis and leukocyte thrombi in pulmonary vasculature are thought to be contributing mechanisms.

Methods

After IRB approval, the medical records of all patients admitted to our PICU with diagnosis of pertussis between 1997 and 2007 were reviewed. Patients with severe pertussis requiring intubation and mechanical ventilatory support and with a echocardiographic diagnosis of pulmonary hypertension were analyzed.

Results

Six patients were identified who met the above criteria. We have previously described 4 of these patients who required ECMO support for severe pertussis and who all died from severe refractory pulmonary hypertension (Pediatrics 2003; 112:1274–78). Since 2003, we identified 2 subsequent cases of severe infantile pertussis with good outcome. Both patients were treated with leukopheresis to achieve leukoreduction. The first patient had significant respiratory and cardiovascular instability and was initially placed on ECMO for stabilization and then underwent leukopheresis. We previously reported this case in detail (PCCM 2006; 7:580–582). A second more recent unreported patient is now described in more detail. The patient was a 5 wk old male infant who was admitted to PICU on his 5th hospital day. Two days later he was intubated and placed on mechanical ventilatory support for respiratory failure and a presumed diagnosis (subsequently confirmed) of pertussis was made. Within following 2 days he had ECHO evidence of PHT and his WBC rose to 90.8K. He underwent leukopheresis. His WBC was reduced to 28.4K and follow up ECHO had no evidence of PHT. He survived and was discharged home.

Conclusion

Based on our recent experience and other previously published cases we propose an algorithm for approach in critically ill infants with severe pertussis. Pertussis induced respiratory failure on mechanical ventilatory support and Evidence of pulmonary hypertension.

Or WBC >100K if YES and cardiac and respiratory status stable than Leukopheresis or double volume exchange transfusion (DVET). If unstable place on ECMO (if available) to bridge for leukopheresis or DVET. If NO evidence of PHT and WBC< 100K continue support and monitor for PHT and WBC trend. PHT defined as RV pressure equal or greater than 1/2 sytemic.

0555

Milrinone in Paediatric Septic Shock

C. Mendes*1, B. Robalo1, J. Fermeiro1, F. Abecasis1, M. Vieira1, J. Pereira1, R. Anjos2, M. Correia1, G. Rodrigues1

1Paediatric Intensive Care Unit, Hospital Santa Maria, 2Paediatric Cardiology, Hospital Santa Cruz, Lisbon, Portugal

Introduction

Severe paediatric septic shock is frequently characterized by depressed myocardial function associated with high systemic vascular resistance. In these circumstances, milrinone may be a valuable tool for cardiovascular support of children with septic shock, owing to its inotropic effect (with low myocardial oxygen consumption) and vasodilator properties.

Methods

We reviewed the clinical records of patients with septic shock admitted to our PICU from January 2005 to March 2008 who were treated with milrinone. Demographic, clinical and laboratory data were analysed.

Results

Twelve patients received milrinone during the study period, but one patient was excluded due to infra-therapeutic milrinone dosage. The median age of the 11 patients included was 3 years and 3 months (3 months to 7 years). All patients were volume resuscitated to CVP>10 cmH2O. Dopamine was the first inotropic agent used in all patients. Milrinone was the second agent used in 4 patients. The average interval between admission and milrinone administration was 15 hours. Only one patient received a milrinone loading dose (50mcg/Kg). The initial infusion dose varied from 0.25–0.75 mcg/kg/min; the maximum infusion dose was 0.82mcg/kg/min. The median duration of milrinone infusion was 72 hours. Nine patients needed 3 or more inotropic agents (7 received norepinephrine). Cardiac function was evaluated by echocardiography in 8 patients before and after milrinone was started: 5 improved, 2 had no change and 1 worsened cardiac contractility. Blood lactate level increased in 2 patients and decreased in 7 patients (a mean decrease of 20.1 mg/dL) after starting milrinone. Concerning the adverse effects of milrinone: mean arterial pressure decreased in 4 patients (a mean decrease of 10 mmHg); 5 patients needed norepinephrine or an increase of dosage after milrinone infusion started; heart rate increased in 7; none had disrythmia; 2 patients developed secondary thrombocytopenia. Seven patients needed mechanical ventilation, 4 needed renal replacement techniques, none had ARDS and 1 patient had severe ischemic sequelae leading to lower limb amputation. The median PRISM scale value was 23. Mortality rate was 18% (2/11).

Conclusion

Milrinone has shown promising results with very few adverse reactions. In our patients, despite insufficient haemodynamic monitoring, the results were encouraging in terms of cardiac function and peripheral perfusion. A randomized controlled trial of milrinone use in paediatric septic shock is needed.

0556

Percutaneous Dilational Tracheostomy in Children

A. K. Baronia*

Critical Care Medicine, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India

Introduction

Purpose of the study: To study safety and feasibility of percutaneous dilational tracheostomy (PDT) in children.

Methods

PDT performed over the past ten years in 21 children, requiring prolonged mechanical ventilation admitted to the single intensive care unit (ICU), were analyzed. Institutional ethic approval was obtained.

Results

Age of the children varied from 08 month to 12 years. Average mechanical ventilation and ICU days, respectively were 26 and 36. Seven children were below 06 years, smallest child was only 08 month old. All procedures were performed by a single operator experienced in adult PDT.

In all the children, modifications were required in the steps of Ciaglia technique namely 1) airway control, 2) identification of trachea, 3) stoma formation, and 4) tracheostomy tube insertion.

All children tolerated the procedure without any mortality or significant morbidity. In three children airway was lost transiently. No other complication was observed.

Average procedure time was 25 minutes and ranged between 15 to 40 minutes. Smaller children (less than 06 years)required more time as compared to older children.

Conclusion

Like adults, children requiring long term mechanical ventilation need tracheostomy. However, all ICUs providing care to children may not have surgical support to perform surgical tracheostomy.

No custom made introducer and dilational set or wire guides could satisfy the technical needs of procedure. Assorted items were selected from the broad range of products used for adult percutaneous tracheostomy, vascular access, and radiological interventions.

In the hands of experienced operator supported by an expert controlling airway, PDT can be performed safely in the ICU.

0557

Respiratory Dengue Disease in a Brazilian Paediatric ICU Population

A. Guizzardi1, H. Falcão2, B. Araújo*3, C. Motta1, S. Victal4, J. Luiz4

1Serviço de Fisioterapia, NEOVIDA — Centro Infantil de Terapia Intensiva, COTEFIL Hospital Geral, Duque de Caxias, 2Intensive Care Physician, Rio de Janeiro, 3Paediatric Intensive Care Physician, NEOVIDA — Centro de Terapia Intensiva / COTEFIL Hospital Geral, Duque de Caxias, 4Paediatric Intensive Care Physician, NEOVIDA — Centro de Terapia Intensiva / COTEFIL Hospital Geral, Duque de Caxias, Rio de Janeiro, Brazil

Introduction

Dengue epidemics results in new admission in paediatric critical care units in Rio de Janeiro. Low age is one risk-factor to severity in dengue fever and the occurence of secondary pleural effusions is a common alteration in a group of pediatric critical care patients.

Methods

We reviewed the medical records of children admitted to the Pediatric Intensive Care Unit, with serologically confirmed dengue diagnosis, from march to april 2008. A specific protocol was filled out by physical therapists to obtain the following data: age, sex, vital signs, utilization of non-invasive ventilatory support, occurrence of pleural effusions, and Lenght-of-stay (LOS).

Results

Among 15 patients, 7 were boys. Age was between 3m and 10y. All children came from Baixada Fluminense area, in Rio de Janeiro state. Sixty-seven percent of them presented pleural effusion on ICU admission, and 40% were submitted to non-invasive ventilatory support (BiPAP) during ICU. The mean LOS was 5,3 days (+− 1.45 days). Two patients were submitted to thoracentesis. The occurrence of secondary complications, necessity of invasive ventilatory support and death in this cohort were equal to zero. Open image in new window

Conclusion

Respiratory findings were common in this cohort, but were not associated to a LOS > 7 days, higher morbidity and mortality rates. There was an apparent trend in more consistent respiratory frequency reduction BiPAP group, in contrast to no-BiPAP. Further investigations are needed to study the effects of BiPAP in dengue respiratory manifestations.

0558

Telephone Information Provided to Parents in Italian Pediatric ICUs

A. Giannini1, G. Miccinesi2, S. Leoncino1, E. Prandi*1

1Pediatric Intensive Care Unit, Fondazione IRCCS Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena, Milano, 2Unit of Epidemiology, Centro per lo Studio e la Prevenzione Oncologica, Florence, Italy

Introduction

Families of ICU patients need information, proximity to their loved ones and assurance [1]. This leads them to telephone frequently for news [2]. To date, no published data are available on telephone information provided to parents in Italy's pediatric ICUs (PICUs). We investigated this issue in the course of a national survey on visiting policies in Italian PICUs.

Methods

An email questionnaire on visiting policies was sent to the heads of all 34 Italian PICUs, including questions about their policy on providing telephone information to parents.

Results

The response rate was 100%. Daily meetings of doctors with parents were held systematically in almost all ICUs (97%). Information was also given by phone (often or always, 70%; sometimes, 23%; never, 6%). Those authorized to give this information were mainly physicians (doctor on duty, 94%; charge nurse, 18%; nurses, 35%). Frequently (often or always, 85%) the family was given the ICU's extension number and 23% of ICUs had a specific time slot for taking relatives’ phone calls. Not only reassurance (59%) and logistical information (44%) were given over the phone, but also generic clinical information (79%), e.g. regarding temperature or sleep. However, even detailed clinical data, e.g. on diagnosis, prognosis and treatment, was given in 23% of ICUs. To ensure confidentiality, 41% of ICUs arranged with the family for a single interlocutor to call at set times, 47% provided only generic information, and 6% gave the family an ID code.

Conclusion

Our findings suggest that in Italian PICUs the telephone plays a relatively important role in giving parents information, and that it is more widely used than in Italian adult ICUs [3]. Day-to-day information must be based on direct meetings between doctors and families; however, despite possible problems of confidentiality and disruption to the work of the ICU staff [2], the phone can represent a complementary tool in providing certain information to parents and, above all, in addressing their considerable need for reassurance [1].

Grant Acknowledgement

This study was supported by ABN (Associazione per il Bambino Nefropatico, Milan, Italy).

0559

The Improvement of Dynamic Compliance Using Furosemide / Natrium Chloride Aerosols in Mechanical Ventilated Children

B. M. Neamtu*1, S. S. I. Iurian1, P. P. N. Nicolcescu2, C. C. D. Dima2, M. M. L. N. Neamtu1, C. C. S. N. Berghea Neamtu3

1Pediatric Clinic, 2Anesthesiology and Intensive Care Unit, 3Pediatric Surgery Clinic, Pediatric Hospital, Sibiu, Romania

Introduction

Inhaled aerosols with Furosemide were used in many studies regarding mechanical ventilated preterm babies with chronic lung diseases (to reduce lung edema) and even in adults in order to improve dyspnoea. The proposed effects for Furosemide were related to lung mechanics improvement.

Methods

The study included 30 children with normal lung function (FEV1 / FVC > 80%) aged between 8–12 years; sex ratio 1:1; no previous pulmonary disease; the patients have received during surgery anesthetics agents (Propofol 2.5 mg/kg.body weight, Fentanyl 1–2 µg/kg.body weight, Esmeron 0.6 mg/kg). After surgery, they were mechanical ventilated in BIPAP mode (with PIPmaximum 15 cm H2O, PEEP 3 cm H2O, FiO2 =0.3–0.5, Ti/Te = 1 / 2, Tidal volume= 8–10ml/kg body weight) receiving constantly through a pressure nebuliser aerosols with Furosemide / Natrium chloride (2 mg Furosemide / ml). Dynamic compliance (Cd) and airflow resistance (Raw) were measured every minute for the first 10 minutes (M0 to M10) of ventilation with consequent measurements at minutes 15′(M15) and 20′ (M20)respectively. The monitored parameters: SaO2>96%; pCO2=35–38 mmHg; pH= 7.35–7.45. To estimate potential electrolytes, ureea and creatinine imbalances,3 blood samples have been taken for each patient: the reference probe before commencing the ventilation, the second at minute 10′ and the third one at minute 30′. The urine output was also measured.

Results

We have noticed a decrease in Raw between 8–10% and an increase in Cd between 43 – 72% as compared to the baseline values of the patients for each parameter. The difference for each patient between the medium value of the starting moment (M0) and the medium value for M1 to M20 had a statistical significance (p value<0.001).

The blood natrium, kalium, creatinine and ureea levels were maintained in normal ranges before and after ventilation with Furosemide aerosols.

Conclusion

The results were temporary related to the period of the administration of aerosols. The data are encouraging in the treatment of acute respiratory distress syndrome in order to ameliorate, even temporary, the specific pulmonary mechanics in this disease.

0560

Use of Mechanical in-Exufflator in Long Term Ventilated Paediatric Patients: Preliminary Data From Italian Survey

F. Racca*1, G. Ottonello2, P. Banfi3, A. Wolfler4, V. Landoni5, I. Tardivo6, G. Berta1, E. Calderini7, I. Salvo4, V. M. Ranieri1

1Anestesia e Rianimazione, Università di Torino, Torino, 2Anestesia e Rianimazione, Istituto G. Gaslini, Genova, 3Pneumologia, O. Mellini Mellini, Chiari, 4Anestesia e Rianimazione, O. Buzzi, Milano, 5Medicina Riabilitativa, O. Valduce, Costamasnaga, 6Pneumologia, OIRM, Torino, 7Anestesia e Rianimazione, O. De Marchi, Milano, Italy

Introduction

Children with neuromuscular disease (NMD) exhibit progressive muscle weakness and impaired cough. Ineffective cough can cause serious respiratory complications. Mechanical In-Exufflator (MI-E) can be used to improve clearance of airway secretions and therefore to reduce respiratory morbidity in children with NMD. The aim of the present study was to identify the number of children requiring MI-E in Italy, and to establish their underlying diagnosis, and the reason for MI-E use.

Methods

Surveys were sent by mail to all centers thought to be involved in pediatric long term mechanical ventilation (LTV). The study included all patients aged less than 18 years on LTV on January 1, 2007.

Results
Detailed informations were obtained on 67 children requiring LTV and MI-E use. Eighty seven % (n=58) of MI-E users were patients with NMD. Thirty six patients had spinal muscular atrophy. The most common indication for MI-E application are listed in Table 1. The distribution of the age of MI-E users is shown in Table 2. MI-E was applied via a facemask (88% of children) or via a tracheotomy (12% of children). In 96% of patients in our study population, MI-E was used at home.
Table 1

Methods Used to Guide Decision to Initiate MI-E

Methods

Number of children

Clinical symptoms

56

Pulmonary function and respiratory muscle strength measures

10

Peak expiratory flow measure

7

Table 2

Number of Mi-E Users According to Age

 

Age interval

 

< 12 month

1 to 5 years

6 to11 years

12 to 17 years

N° of children

5

21

14

27

Conclusion

In the Italian pediatric population requiring LTV, identified in this study, the majority of patients requiring MI-E had NMD, more than 5 years of age, and applied MI-E via a facemask. The most common reason for MI-E application was the presence of clinical symptoms suggesting ineffective cough.

0561

Changes in Infectious Disease Mortality Among Children in the Netherlands

E. S. Veldhoen*1, T. F. W. Wolfs2, A. J. van Vught1

1Paediatric Intensive Care Unit, 2Department of Paediatric Infectious Diseases, University Medical Center Utrecht, Utrecht, Netherlands

Introduction

We regularly admit children with life threatening infections. In this study we examine the changes in mortality due to infectious diseases in childhood during the last decades.

Methods

We analysed mortality data among those aged less than 20 years due to infectious diseases from 1969 till 2006, obtained from the Central Office of Statistics (CBS) in the Netherlands. We reviewed mortality data of all Dutch Paediatric Intensive Care Units (PICU's) in 2005 and infectious causes of death in all children, admitted to our PICU from 1997 to 2006.

Results

This study shows a steep decline of infectious disease mortality in the seventies, followed by a relative stabilisation in the years thereafter. This steep decline was caused by an isolated decline in infectious disease mortality in infants (40 to 10 deaths per 100,000 children) and children between 1 and 5 years of age (7.9 to 2.6 per 100,000). In children over 5 years of age the infectious disease mortality remained stable over the whole study period. Analysis of mortality data of our PICU shows an increasing trend in mortality due to infectious diseases in children with underlying illnesses over the last ten years.

Conclusion

Infections in childhood remain a stable burden of mortality during the last decades, despite major improvements in therapeutic and preventive measures. This might be explained by increasing numbers of fatal infections in children with an underlying medical condition, known to increase the risk of severe infections.

0562

Epidemiological Surveillance of Poisoning in Children

I. Klironomi*, S. Sallabanda, E. Kola, R. Lluka, F. Zavalani, I. Kasmi, K. Marku, G. Sallabanda, A. Kola

Pediatric, University Hospital Center Mother Theresa, Tirana, Albania

Introduction

Poisoning represents one of the most common medical emergencies encountered in paediatric ages.

Methods

We report the results of a ten years retrospective study for poisoning in children. Are included 218 children with poisoning exposures admitted at our PICU for the period 1998 – 2007. It has been evaluated the frequency the age the type of exposure route of exposure and the circumstances of poisoning iatrogenic intentional or unintentional. Patients are divided in three groups: 0–5 years 6–12 years over 12 years and variables were analyzed for each group.

Results

Poisoning constituted 4.5 per cent of all admissions. The number of poisoning is increased after 2002 from 3.5 per cent to 5.5 per cent. The peak age of poisoning occur under 5 years old 70.1 per cent. 47.7 per cent of cases were related to medicaments poisoning, pursued par the pesticides 22.4 per cent of cases. Antidepressants represent the most common poisoning. Significant was in our study poisoning from pseudoephedrina 13.6 per cent of medicaments poisonig in children 1 year old and methoclopropamid iatrogenic. Poisoning from acetaminophen is increased after 2002. Poisoning in 84.7 per cent has been accidental. Suicidal poisoning 5.5 per cent are found only in girls older than 12 years old. The mortality rate is reduced less than 10 per cent. Lethal have been poisonings from Fostoksina and mushrooms.

Conclusion

Given the increased number of poisoning in children during last years it is necessary to intensify the level of prevention for poisonings.

0563

Long Term Ventilation of Children in Italy: Preliminary Data from Questionnaire Survey

F. Racca*1, M. Bonati2, G. Berta1, R. Testa3, F. Benini4, M. Benedetti5, E. Bignamini6, M. Maspoli7, I. Salvo8, V. M. Ranieri1

1Anestesia e Rianimazione, Università di Torino, Torino, 2Istituto di Ricerche Farmacologiche, Mario Negri, Milano, 3Anestesia e Rianimazione, O. Santobono, Napoli, 4Pediatria, Università di Padova, Padova, 5Anestesia e Rianimazione, O. Civile Maggiore, Verona, 6Pneumologia, OIRM, 7Assessorato Sanità, Piemonte, Torino, 8Anestesia e Rianimazione, O. Buzzi, Milano, Italy

Introduction

Long term mechanical ventilation (LTV) for children with chronic respiratory failure is an established supportive therapy that reduces morbidity and mortality. Home ventilatory support is the optimal option to satisfy child's psychological needs and therefore enhance quality of life. The aim of the present study was to identify the number of children requiring long term ventilation in Italy, and to establish their underlying diagnosis, ventilatory needs and hospital discharge rate.

Methods

Surveys were sent by mail to all centers thought to be involved in pediatric LTV. The study included all patients aged less than 18 years on LTV on January 1, 2007.

Results
Of the 611 initial surveys posted, 189 were returned, identifying 407 children requiring LTV. Detailed informations were obtained on 305 children. LTV users were classified in three disorder categories: neurological (n=229; 75%), thoracic (n=13; 4%), lung/ upper airway (n=63; 21%). The age of institution of LTV according to disorder category is shown in Table 1. One hundred twenty two (40%) children were ventilated via a tracheotomy with the highest percentage (89%) being neurological patients (Thor 2%, Lung/airway 9%). The majority of non invasively ventilated children used nasal masks (87%). All of the LTV users had positive pressure ventilators. Pressure-limited and volume-limited modes were equally distributed among patients with neurological and thoracic disorders. Patients with lung/upper airway disorder preferred pressure-limited ventilation. Twenty one % of LTV users required mechanical ventilation for more than 20 hours per day, 20% were ventilated 12–20 hours per day, and 59% received ventilation only when asleep. The majority of LTV children requiring mechanical ventilation for more than 20 hours per day were neurological patients. Three hundred (98%) children were cared at home. Only five neurological patients remained in hospital.
Table 1

Age of Institution of LTV According to Diagnostic Category

Disorder category

Number of children

 

<12 Month

1 to 5 yers

6 to 11 years

12 to 17 yers

Neurological

11

75

74

69

Thoracic

0

3

5

5

Lung/Upper airway

3

21

12

27

Conclusion

In the Italian pediatric population requiring LTV, identified in this study, neurological patients represented the category that needs more health care resources. Successful discharge home was possible for the majority of LTV children despite the very young age, the severity of disease and the need for technological care.

Poster Sessions Clinical outcome I: 0564–0577

0564

Percutaneous Tracheostomy in a District General Hospital

S. Mathieu*, K. Wilkinson, T. Hanham, N. White

Critical Care Unit, Royal Bournemouth Hospital, Bournemouth, United Kingdom

Introduction

Percutaneous tracheostomy is a frequently performed procedure on the Intensive Care Unit. It's main indication is when mechanical ventilation is expected to be prolonged. However, optimal timing for tracheostomy insertion is unclear, and whilst it has traditionally been viewed as a minimally invasive procedure, serious complications can occur. The requirements for anaesthesia and necessity to teach trainees to become familiar with this procedure are further potential hazards.

Methods

Retrospective data for all tracheostomies inserted in our intensive care unit at the Royal Bournemouth Hospital from 2006 – 2007 were analysed. The following data was collected: indication for tracheostomy, period from endotracheal intubation to insertion of a tracheostomy, quality of documentation of the procedure, level of supervision, use of fibreoptic bronchoscope, and early and late complications.

Results

36 patients had percutaneous tracheostomies in our unit between 2006 and 2007. The majority were indicated for respiratory and neuromuscular problems. 45% were inserted within 5 days of endotracheal intubation and 86% by day 10. Complete documentation of the procedure occurred in 91% of cases with bronchoscope guidance and full monitoring in 100%. A consultant (64%) and/or Specialist Registrar was directly involved in the procedure in all cases. If a consultant was not directly involved, they were present at the bedside. We had one early complication of minor bleeding and late complications (at 1 year) in 8% cases (2 stomal infections and 1 symptomatic tracheal stenosis). There was no case of mortality related to tracheostomy insertion.

Conclusion

The results from our audit compare favourably with previous studies. A sound level of supervision was achieved in all cases and our complication rate was low. All procedures were successful, and subsequently during this period there were no surgical tracheostomies performed. The one case of tracheal stenosis did not require any intervention and continues to be followed up with no evidence of progressive sequelae. Our level of documentation was below the gold standard as stipulated by the Royal College of Anaesthetists. Since this audit, we are in the process of addressing this, by producing a self-adhesive tracheostomy label to be inserted into the notes. We hope this will not only encourage a more complete documentation of this procedure, but also make identifying insertion details easier.

0565

Bedside Endoscopically Guided Percutaneous Tracheostomy: A Five Year Routine Use of Percutwist (PDT) Technique

P. Santos*, S. Fontes, A. Graça, E. Gomes, T. Cardoso, A. Carneiro, E. Neutel

Polyvalent Intensive Care Unit, Hospital de Santo Antoónio, Porto, Portugal

Introduction

Tracheostomy is one of the most usual procedures in intensive care units. Percutaneous tracheostomy offers advantages over operative tracheostomy, namely requiring less time to be performed and being less expensive.1 In addition, complications may be more frequent with surgical tracheostomy.2,3,4,5 Our objective was to determine the relative cost-effectiveness and outcome of PDT.

Methods

Cohort study of 99 patients requiring an elective tracheostomy in a General Intensive Care Unit of a tertiary University Hospital from 2003 until January 2008. Percu-Twist was performed in all cases who had no contraindication. All the procedures were endoscopically guided and held under general anesthesia. Patients were followed-up at six months after discharge or until death. Time of execution, early and late complications were registered.

Results

Sixteen women and 83 men with a median age of 52 (25th–75th percentile: 35–70) years were included. As far as we know this is one of the biggest cohorts using this technique. The admission diagnosis were: trauma in 55% of these patients, medical in 32%, and emergency post-operative in 13%. Mean ± SD SAPS II score was 45 ± 12. Mean ± SD ICU length of stay was 23,6 ± 19,2 days. The motive of tracheostomy was airway protection in 50% and prolonged ventilation in another 50%. Procedures took 5–10 minutes in 36% of cases, 10–15′ in 45%, 15–20′ in 13% and 20–25′ in 6%. Early complications occurred in 20% of the study population, with tracheal ring fracture being the most frequent (17%), followed by minor bleeding (2%) and loss of airway (1%). Late complications were seen in 3% of the group with one tracheal stenosis and two hoarseness cases. No stomal infection was registered and no mortality was associated with the technique.

Conclusion

PercuTwist technique is a quick execution, safe method, with low early and late complications.

0566

ICU and Hospital Mortality in Elderly Patients with Severe Sepsis and Septic Shock

T. Schuerholz*, M. Tondt, T. Simon, K. Reinhart, G. Marx

Anesthesiology and Intensive Care, Friedrich-Schiller-University, Jena, Germany

Introduction

Severe Sepsis and septic shock are an important problem in the growing group of the elderly. Although the higher incidence of sepsis in elderly patients was described [1], there is a lack of data in the intensive care unit. We conducted this study to gain more detailed information on this population.

Methods

In a retrospective study, 563 patients of a university intensive care unit were included to analyze data between March 2003 and November 2006. Inclusion criteria were sepsis or severe sepsis as defined by ACCP/SCCM consensus conference. Patients were divided in two groups of age more than and less than 65 years. After verification of skewness, patient data were analyzed using Mann-Whitney-U-test and χ2-Test where appropriate. P<0.05 was considered significant.

Results

Of 563 patients 52.2% were above 65 years. APACHE II score was significantly higher in elderly patients compared to those less than 65 years (28±10 (median±IQR) vs. 25±11; p=0.01). TISS score was similar in both groups (49±12 vs. 47±11). The incidence of peritonitis was non-significantly lower in the elderly (27.2% vs. 32.7%; p=0.154), whereas pneumonia was more frequent in elderly patients (45.2% vs. 30.9%; p=0.0001). In elderly patients acute renal failure (ARF) occurred more frequently (60.7% vs. 39.3%; p=0.002). ICU mortality and hospital mortality was higher in patients >65 years compared to <65 years (37.4% vs. 24.5%; p=0.001 and 45.2% vs. 33.8%; p=0.006, respectively).

Conclusion

Elderly septic patients are more severely ill, but were treated to the same extend as patients of less than 65 years. Sepsis origin may change in the elderly and complicating ARF in the ICU is more common in this population. These factors may contribute to a higher mortality in the ICU and in hospital of patients >65 years.

0567

Norepinephrine Dose as a Marker of Septic Shock Severity

S. R. Bauer1, J. J. Aloi1, J. A. Guzman*2

1Pharmacy, 2Pulmonary and Critical Care Medicine, Cleveland Clinic Foundation, Cleveland, United States

Introduction

The hallmark of septic shock is hypotension refractory to fluid administration. Vasopressors are routinely administered to restore mean arterial pressure. Although it may be intuitive, limited data exists relating the dose of vasopressor to the severity of shock. Consequently, doses of vasopressors have been picked arbitrarily to classify septic shock severity. The present study was conducted to evaluate the prognostic value of the maximal dose of norepinephrine (NE) during the first day of septic shock in comparison to the APACHE II, SAPS II and SOFA scores from the same day.

Methods

Retrospective analysis of a septic shock database compiled from three large academic hospitals. Patients were included in the study if they received NE as the main, but not the only, vasopressor agent to maintain goal MAP. An empirical receiver operating curve (ROC) of mortality was created for selected outcome predictors. The Hanley and McNeil nonparametric method was used to estimate the area under ROC curve. Additionally, correlation analysis was performed between NE dose and the other outcome predictors.

Results
Ninety six patients were included. Mean APACHE II, SAPS II and SOFA scores were 29.6 ± 8.3, 63.2 ± 20.3 and 12.9 ± 3.7, respectively. Overall survival rate was 45% with a mean maximum dose of NE during the first 24 hrs of shock of 35.4 ± 32.1 mcg/min. Correlation analysis between APACHE II, SOFA, SAPS II, and NE revealed significance only for the NE-APACHE II pair (R2 0.04045, P 0.049).
Table 1

Roc for Each Outcome

Predictor [n]

Estimate of ROC Area (SE)

95% CI

P Value

Dose of NE (mcg/min) [96]

0.614 (0.058)

0.499–0.728

0.026

APACHE II [96]

0.683 (0.055)

0.576–0.79

< 0.001

SOFA [96]

0.670 (0.055)

0.562–0.778

0.001

SAPS II [73]

0.741 (0.058)

0.628–854

< 0.001

Table 2

Differential Thresholds

Predictor

Differential threshold

Sensitivity (95% CI)

Specificity (%95 CI)

Dose of NE (mcg/min)

29.9

0.547 (0.404–0.684)

0.651(0.491–0.790)

APACHE II

29

0.717 (0.576–0.832)

0.581(0.421–0.730)

SOFA

14

0.604 (0.460–0.736)

0.721(0.563–0.847)

SAPS II

68

0.641 (0.472–0.788)

0.735(0.556–0.871)

Conclusion

The maximal dose of NE during the first day of septic shock may help predicting outcome. Our data suggest a dose >29.9 mcg/min indicative of higher mortality although the sensitivity is relatively low. A validation study with prospective data collection is warranted.

0568

The Protocol in the Massive Haemorrhage at a Third Level Hospital

L. Alvarez-Montero*1, R. Aragones-Manzanares1, M. Fernandez-Zamora1, E. Mingot-Castellano 2, A. Medina Arteaga3, M. Delgado-Amaya1, J. Muñoz-Bono1, A. Cabrera-Calandria1, J. Galeas-Lopez1, G. Quesada-Garcia1

1Critical Care, 2Haematology, 3Anesthesiology, Hospital Carlos Haya, Malaga, Spain

Introduction

The massive haemorrhage is a vital emergency in which the multidisciplinary treatment presents a recommendation grade 1A (1). Our objective is to analyse the fulfilment grade of the clinical recommendations, just as establishing the improvement points, to be able to establish a suitable protocol for our hospital. Design of the intervention protocol, as well as the formative action which allows the homogeneity of knowledge among the different services.

Methods

A collection data was designed in order to evaluate the admitted patients in our hospital with massive haemorrhage resistant to conventional treatment. These patients needed a specific treatment in 2004–2008. Items related to the quantity of transfused products, the specific products used, the hemodynamic parameters, the period of diagnosis and treatment, the monitorization, the coagulopathy, the complications and other analytic parameters were gather for that purpose. After analysing the improvement points, a formative programme was designed, which was based on the latest bibliographical recommendations. It included the designed intrahospitalary protocol, with a calendar which contained the different implicated services.

Results

We identified 14 patients whose demographic data with a medium age of 50,5 years, the mortality per month was 28%. The average quantity of hemoderivated pre and post rFVIIa administration was: Concentrated red blood cells pre 3229 ml, post 1022 ml; Plasma pre 1446 ml and post 591 ml; Platelet pre 414 ml and post 167 ml. In the diagnosis of haemorrhage the 81,8% presented coagulopathy, the 42,8% had a shock, the 71,4% presented a temperature of < 35,5 °C and the 35,7% had a heart rate >100. The 71,4% presented belated complications. Factor VII was administered in the 100% of the cases, fibrinogen and vitamin K in the 28,5% of the cases and amchafibrin in the 14,2%. The rewarming up with an electrical blanket was of 35,7%. The 64,2% and the 78,5% required inotropic and vasoactive drugs, respectively. The 50% needed continuous venovenous hemodiafiltration techniques.

The average of SOFA in the admission was of 9. The rewarming of fluids, the contribution of ions according to the protocol, the availability of specific products (Factor VII recombinant, fibrinogen, protromplex), the speed up of the analytical results, the monitorization of the shock with lactate levels were considered improvement points.

The reduction hemoderivates after the administration of specific products lowers the chances of having soon and belated complications.

Conclusion

The establishment of an intrahospitalary protocol of critical haemorrhage is necessary, the same as the analysis of the intervention and the evaluation of the improvement points.

The utilization of specific products to control the massive bleeding as the rFVII seems to diminish the cost of hemoderivates and the complications, from which we consider it should be protocolised.

0569

Severe Acute Pancreatitis: Clinical Study of 50 Patients with Multiorgan Dysfunction

A. K. Baronia*, R. K. Singh, A. Azim

Critical Care Medicine, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India

Introduction

Severe acute pancreatitis (SAP) with multiorgan dysfunction is associated with high morbidity and mortality. Management of these patients require multidisciplinary approach. We present here the results of a clinical study of 50 patients of SAP with multiorgan dysfunction.

Methods

Records of all SAP patients admitted in ICU between June 2002 to July 2007 were analysed retrospectively. Data included important laboratory and clinical values. Patients were grouped as survivors and non-survivors. The values were expressed as mean +/− SD, median with range, and percentage. Statistical analysis was performed using SPSS 14.

Results

Multiparametric scores like APACHE II and serial SOFA, referral pattern, CT severity index (CTSI), intra-abdominal pressure (IAP), interventions like percutaneous drainage and surgical debridement, renal and cardiovascular support, nutrition, mechanical ventilation, microbiology and transfusion requirements were analysed. Out of fifty, twenty patients survived. Survival was found to be influenced by (1) APACHE II at admission and 24 hour, (2)SOFA at admission, and day 3,7 and 14, (3) IAP, (4) CTSI and (4) referral pattern.

Conclusion

Early transfer of SAP patients to ICU and routine monitoring of IAP with timely management of intra-abdominal hypertension were found to be readily modifiable factors to improve outcome of SAP patients in the hospital setting where the study was conducted.

0570

Family Participation in Care of Critically Ill Patients: Opinions of Families, Staff, and Patients

F. Philippart*1, V. Willems1, A. Tabah1, I. Coquet1, C. Bruel1, C. Jeannot1, F. Diaw1, J. T. Timsit2, J. Carlet3, B. Misset1, M. Garrouste-Orgeas1

1Medical-Surgical ICU, Groupe Hospitalier Paris Saint Joseph, Paris, 2Inserm U-823, Institut Bonniot, La Tronche, 3HAS, Groupe Hospitalier Paris Saint Joseph, Paris, France

Introduction

Participation of family members in the care of ICU patients, suggested as a component of family-centred care, has met with reluctance among families (1). Patients’ opinions are not known. The aim of this study was to determine opinions of families, staff, and patients.

Methods

Single-centre prospective study (March 1 to July 17 2006 and September 1 to December 31 2006) of consecutive patients with ICU stays more than 3 days. We evaluated opinions about 13 items of care: wiping the eyes, cleaning the oral cavity, moistening the oral cavity, hydrating the lips, aspirating secretions, cleaning the nose, preventing pressure sores, helping staff change patient's position, helping with bed baths, shampooing, washing the patient's feet, manicuring, and applying hydrating agents. We collected patient characteristics and ICU mortality. Between days 3 and 5, questionnaires were given to the nurse, physician, nursing assistant, and family. Patients were interviewed by phone after hospital discharge. Satisfaction and symptoms of anxiety/depression in families were measured using the CCFNI and HADS scores. Care items actually performed by families were collected by the nurses throughout the ICU stay.

Results

Of 220 admitted patients, 129 were included, among whom 28 were not analysed (no family, n=3; refusal, n=3; not fluent in French, n=4; died on day 4, n=2, second admission, n=5; questionnaire not given, n=1). The remaining 101 patients (age, 64.4 y±16.1; SAPS II, 36.0±14.3) had ICU and hospital mortality rates of 24.7% and 30.6%, respectively. Questionnaires were returned by all staff members and 98% of families. 44 survivors were interviewed and 27 not (refusal,n=4; dementia, n=1; no fluent in French, n=4; current hospitalization, n=4, died, n=10; lost of follow up, n=4 after ICU discharge). Family participation in care was deemed desirable by all physicians, 95% of nurses, 91% of nursing assistants, 95% of families, and 77.2% of patients. Only 14 (13.8%) families actually participated in care (412 care items in all). The items viewed most favourably by families and patients were wiping the eyes (74.7 vs. 72%), moistening the oral cavity (87.8 vs. 76.7%), hydrating the lips (85.8 vs. 74.4%), preventing pressure sores (76.7 vs. 72.0%), and applying hydrating agents (77.6 vs. 72.0%). Family satisfaction was high (11.0±2.5 on a 0–14 scale). Symptoms of anxiety and depression were present in 58/101 (58.5%) and 26/101 (26%) families, respectively.

Conclusion

Families and ICU staff were strongly in favour of participating in care aimed at improving patient comfort. Family satisfaction was high, symptoms of anxiety were less than in many other studies, and actual participation in care by families was low. The willingness of patients to receive care from their families warrants an interventional study aimed at encouraging families to participate in care.

0571

What are the Causes of Discomfort in Intensive Care?

M. Lefevre1, B. Bourgois2, S. Jaffuel1, G. Prat1, J. Tonnelier1, E. L'Her1, A. Renault1, J. Boles*1

1Intensive Care, 2anesthesiology, CHU Brest, Brest, France

Introduction

In order to improve the well-being of patients in our intensive care unit, we have evaluated the sources of discomfort that were reported within discharge.

Methods

This prospective study was approved by the local ethics committee. It was performed within a 15 beds intensive care unit of a teaching hospital. On the day of discharge, a questionnaire including 16 items was submitted to patients whose Glasgow score was higher than 13, and who spent more than 48 hours in the unit. For some items, we tried to precise if patients were more bothered at night. Besides, we looked for the main sources of pain and anxiety, and how they were relieved.

Results

50 patients were interrogated (36 men and 14 women, 53.9 ± 14.5 years old; SAPS II 34.7 ± 21.9; mean length of stay 7.7 ± 4.9 days), corresponding to 25.3% of the 197 survivors, over a time period of 154 days. Among these, 32 (64%) had been intubated, for a period of 5.9 ± 6.1 days and 24 (48%) had been sedated for a period of 4 ± 3.4 days. PHYSICAL DISCOMFORTS: the most frequently reported physical discomfort was thirst, mentioned by 56% of patients. 50% of patients complained about a lack of sleep, 48% about pain, most frequently back pain and throat pain caused by intubation (25%). 44% of intubated patients were bothered by the endotracheal tube, and 40% by the restraint. 40% of patients complained about infusion tubing and cables. The next cited sources of discomfort were cold (38%), noise (36%) and light (26%), mainly at night (61% for noise and 84.6% for light). Alarms were responsible for 72.2% and staff for 22.2% of the noise discomfort. Heat, saturometer sensor and hunger were less frequently reported (20%, 18% and 14%). PSYCHIC DISCOMFORT: The main psychic discomfort was anxiety (38%), which was relieved for 66% of the patient, mainly by the staff presence and support (64.2%), more rarely by families and drugs (14.2%). 28% of the patients suffered from a lack of information, 20% felt isolated from their family and 12% complained from a lack of privacy.

Conclusion

43 patients reported at least 3 sources of discomfort, and 7 report 2. Thirst concerned more than half of the patients, despite the use of water sprays and wet compresses. The pain relief protocol is still insufficient, and analgesia alone does not resolve discomfort caused by decubitus or by the endotracheal tube. Noise and light, which are more bothering at night, participated to a lack of sleep that is often complicated by confusion and agitation. Different measures have to be taken in order to improve the well-being of intensive care patients: early screening of physical and psychological pain, changes in the patients environment and in our habits to preserve their nycthemeral rhythm. Once these measures will be built-up, we will evaluate their effect by further surveys.

0572

Post-ICU Visit: Feasability and Results in Patients with Multiple Organ Failure

R. Hamidfar-Roy1, C. Schwebel*1, A. Bonadona1, D. Barnoud1, L. Hammer1, A. Tabah1, J. Remy1, A. Francais2, C. Ara Somohano3, J. F. Timsit1

1Medical Intensive Care Unit, Grenoble University Hospital,2INSERM U823, Institut Albert Bonniot, 3Emergency Department, Grenoble University Hospital, Grenoble, France

Introduction

Follow-up and quality of life after ICU are not routinely integrated in the process of care. We present the results of the outpatient Post-ICU visit. It was systematically given for mechanically ventilated (MV) patients of more than 48 hours with at least another organ failure in our ICU.

Methods

All consecutive Multiple organ failure (MOF) survivors from January 2007 to March 2008 were seen 3 months after ICU discharge. Psychiatric or bedridden patients were not invited to the visit. A team-dedicated attending physician performed clinical examination, chest-Xray and functional respiratory tests. Psychological symptoms and nutritional parameters were checked and quality of life was evaluated by SF-36 questionnaire.

Results

89 patients were asked to participate. 57 patients (40 men, 17 women, 51 medical, 5 unscheduled surgery, 58.4 yo +− 14.5, SAPS II 36.1 +− 17), ventilated for 7 days in median, who stay 13 days in median in the ICU went to consult. Before ICU stay, 13 were retired, 11 were severely disabled.

After a median delay of 104 days, 55 were at home and 2 in long term care facilities. The weight gain was 3kg in median. Only 2 patients returned to work. SF-36 was 66 in median. 23 present sleep disorders, 3 patients suffered severe anxiety/depression, 26 patients had no ICU recall, 31 remembered unpleasant recall (fear 7, hallucination 8, constraints 4, isolation 4). The main complain was weakness (11) and pain (13) and autonomy lost (11). Overall 67% patients felt their QOL good, but it decreased in 43% cases as compared to the last year. 4 patients suffered dyspnoea but functional respiratory tests revealed 19 obstructive, 8 restrictive syndromes and 17 CO diffusion abnormalities. Physiotherapy was achieved in 34 patients, ongoing in 19 and refused in 4. The visit lead to new therapy or specific care in 12 cases.

Conclusion

Follow-up of ICU patients is routinely feasible but require specific organisation. It unmasked frequent symptoms that required specific care. The information feedback to the ICU team could benefit in improving our routine ICU care.

0573

Long Term Prognosis of People Aged 75 Yrs Admitted to a Medical ICU

J. Boles*, M. Lefevre, A. Renault, S. Jaffuel, G. Prat, J. Tonnelier, E. L'Her Service de Réanimation médicale, Hôpital de la Cavale Blanche — CHU, Brest, France

Introduction

The mean age and the number of people >75 yrs in ICUs is constantly increasing. Their long-term outcome is not well known.

Methods

We performed a retrospective study in our 15-bed medical ICU in a university hospital : all patients aged >75 yrs admitted in 1996 and between 2001 and 2006 included were eligible. For the latter, we recorded the number of deaths in the ICU, in the hospital, and survival up to April 2007 by asking for certificates of deaths in towns of residence.

Results
All results are summarized in the two following tables.
Table 1

1) Admittance of People >75 YRS

 

total pts/died

pts > 75 yrs (%)

ICU deaths [75 (% all deaths)

hospital death

total deaths > 75 yrs (%)

1996

461/128(27,8%)

48 (10,4%)

21 (16,4%)

13

34 = 70,8%

2001

331/80 (24,2%)

43 (13,0%)

18 (22,5%)

4

22 = 51,2%

2006

639/147(23,0%)

107 (16,7%)

40 (37,4%)

9

49 (45,8%)

Table 2

2) Survival Upon 04/20/2007

year

pts > 75 admitted

hospital survivor (% of admitted)

2007 survivors (% of admitted)

2001

43

21=48,8%

4=9,3%

2002

50

17=34%

5=10%

2003

74

34=45,9%

16=21,6%

2004

87

41=47,1%

22=25,3%

2005

90

39=43,3%

26=28,9%

2006

107

58=54,2%

44=41,1%

Conclusion

The absolute number and percentage of patients >75 yrs have increased over the past ten years. ICU and immediate intrahospital mortality have decreased. The percentage of survivors after ICU discharge decreased from about 40% 4 to 15 months after to 25% 27 to 39 months after and to 10% 51 to 75 months after. The annual “attrition rate” is about a third of survivors 3 years after ICU discharge. There is no sex- or age bracket-related mortality difference. This unexpected long term result leads to conclude that ICU admittance should not be performed on an age basis per se. Quality of life and functional status of surviving patients is under investigation to ascertain their real autonomy.

0574

Long-Term Prognosis in Critically Ill Patients with Liver Cirrhosis

P. Piringer*, F. Firlinger, R. Buder, B. Überall, T. Steinmaurer, M. Binder, C. Jocher, C. Kapral, F. Wewalka, K. Lenz

Department of Internal and ICM, Hospital St. John of God, Linz, Austria

Introduction

The mortality among patients with cirrhosis, who require ICU treatment, is high. Depending on the degree of hepatic insufficiency and dysfunction of extrahepatic organ systems, mortality rates are reported up to 70%. ICU admission is frequently questioned because of poor prognosis and limited resources.

Methods

131 cirrhotic patients (118m, 13f) admitted to our ICU between 2002–2006 were retrospectively evaluated. ICU, in-hospital and 1 year mortality were recorded.

Results

On admission, 60 Patients (45.8%) primarily presented with upper gastrointestinal bleeding (GI), 13.7% with hepatic coma, 9.9% with hepatorenal syndrome (HRS). Alcohol was the main reason for cirrhosis. The ICU mortality was 29.77% (overall ICU Mortality 11.3%), in-hospital mortality 34.35% and cumulative 1 year mortality 51.15%. MELD Score (30.89 vs. 16.05), APACHE II Score (13.22 vs. 25.04) and SAPS II Score (27.71 vs. 54.53) were significantly higher for hospital non-survivors than for hospital survivors (p < 0.01 — Mann-Whitney U-Test). Using the area under ROC curves, the three scores discriminated well between hospital survivors and non-survivors (AUROC: MELD 0.826; APACHE II 0.869; SAPS II 0.858) (Fig. 1). Patients admitted with upper GI bleeding had the lowest hospital mortality rate with 15%, whereas 69.2% died from HRS. Open image in new window

Conclusion

Cirrhotic patients admitted to our ICU have a higher in-hospital mortality rate (34%) than the average patient treated on an ICU. 22 patients died during the 1 year follow-up, hence the 1 year survival rate was less than 50%. High scores identified patients with a poor outcome and a high probability of death during treatment in hospital, e.g. nobody survived with a MELD score higher than 42. All patients that exhibited an APACHE II score lower than 11 or a SAPS II score lower than 21 were discharged alive. Furthermore, the MELD score discriminated significantly one year survival (18.36 in non-survivor vs. 14.52 in non-survivor; p=0.034-Mann-Whitney U-Test), APACHE II (15.14 vs. 12.44) and SAPS II (29.91 vs. 26.58) could not statistically differentiate in long-term survival.

0575

Health-Related Quality of Life 1 Month After Discharge from More Than 3 Days Intensive Care

C. B. Kancir*, E. Iversen, K. Damborg

Intensive Care Unit, Dept. Of Anaesthesia, Holstebro Hospital, Holstebro, Denmark

Introduction

Quality of life is one of the most important outcome measures of survivors after critical illness. But physical and psychological sequelae have been described even 12 months after ICU discharge. Rehabilitation is therefore mandated as soon as possible after ICU to restore optimal health status. The Short Form 36 (SF 36) is a robust tool validated for quality of life assessment following critical illness (1). Therefore we studied SF 36 as soon as possible after critical illness and before attending the follow-up clinic.

Methods

In a 6 beds mixed (predominantly medical) ICU in a community hospital, all adults (>18 years) ICU survivors during a 14 months period and with stay >3 days were included. One month after ICU discharge or soon after hospital discharge they were sent the SF 36 using regular mail. Results were compared with an age material matched control group drawn from Danish normative data. Values are expressed as mean (SD).

Results
Among the 76 eligible patients, 49 (64,5%) answered the SF 36. The mean age was 61 years. Their mean SAPS II score was 38,1 and mean length of ICU stay was 6 days. 39 (79,6%) patients had mechanical ventilation during a mean duration of 7 days.
 

GH

PF

RP

SF

BP

VT

MH

RE

1 Month

50 (20)

38 (29)

13 (25)

68 (29)

60 (30)

46 (23)

71 (19)

43 (43)

Population

68 (22)

81 (24)

75 (36)

90 (19)

75 (25)

69 (23)

82 (18)

85 (24)

(GH General Health; PF Physical Functioning; RP Role Physical; SF Social Functioning; BP Bodily Pain; VT Vitality; MH Mental Health; RE Role Emotional. 0 = worst score, 100 = best score)

Conclusion

The comparison between the 1-month study sample and a control population of apparently healthy Danes of similar age revealed significant impairment in all items of the SF 36 with the greatest limitation for physical role.

0576

Routine Delirium Monitoring in a UK Critical Care Unit Is Feasible and Identifies a High Incidence of Unrecognised Delirium

V. J. Page*, S. V. Navarange

Anaesthetics and Critical Care, Watford General Hospital, Watford, United Kingdom

Introduction

Delirium is a common neurobehavioural syndrome associated with increased morbidity and mortality. Delirium goes unrecognised unless a screening tool is used as the majority of delirium is either motoric mixed or hypoactive. The confusion assessment method for the intensive care unit (CAM-ICU) is a validated delirium screening tool for critically ill intubated patients with sensitivity and specificity over 93%. The aim of this clinical audit was to determine the incidence and outcome of delirium using the CAM-ICU in a UK critical care unit.

Methods

Routine CAM-ICU monitoring was implemented in our mixed critical care unit in January 2007 following two months educational and promotional campaign. Management of delirium guidelines were introduced but there was no change to sedation guidelines. During September and October 2007 the daily CAM-ICU was recorded by the bedside nurse for 50 consecutive level 2 and level 3 patients admitted to the mixed medical/surgical critical care ward in a district general hospital. This was repeated in January 2008 for 30 consecutive patients. Patient outcome was recorded.

Results
It was not possible to assess for delirium in ten patients due to coma. The overall incidence of delirium was 26% (21 CAM-ICU positive). If elective post-operative patients were excluded the incidence was 37%. The mortality in the CAM-ICU positive group was 38% compared 26% for all patients. Nine of the ten patients who were unable to be assessed for delirium died.
 

Delirious

Not delirious

Unable to assess

Number of patients

21

49

10

Deaths

8

4

9

Mean % of Deaths (95% CI)

38% (20%–60%)

8% (3%–20%)

90% (55%–99%)

Conclusion

We have demonstrated that delirium screening is feasible in a UK intensive care population. The high incidence of delirium and the impact on outcomes in this UK cohort of patients is in line with previous reports from other countries studying a similar case mix. There is an urgent need to develop strategies to treat delirium in the critically ill.

0577

Long-Term Outcome and Prognostic Factors for Hematological Patients Admitted to An ICU

T. S. Revenfeld*1, H. M. Betsch1, N. K. Schønemann1, S. E. Dyrskog PHD2

1Department of anesthesiology, Aarhus university hospital, Aarhus sygehusTHG, Aarhus C, 2Department of anesthesiology, Aarhus university hospital, Skejby, Århus N, Denmark

Introduction

The goal of our study was to find predictors for both ICU and long term outcome in patients admitted to the ICU with a hematological diagnosis. The possible predictors chosen were: ventilator treatment, CRRT, secondary infection during ICU stay, ICU length of stay (LOS), APACHE II score and in-hospital length of stay prior to ICU admission.

Methods

Retrospective unicentric analysis of patients with a hematological disease admitted to the multidisciplinary ICU (6 beds) at a tertiary university hospital. The ethical issues involved in the individual ICU stay were also studied. We looked for decisions in the patient journals to withdraw or withhold resuscitation, ventilator, CRRT and inotrope therapies.

Results

In this one-year period 65 patients with a hematological diagnose were admitted to the ICU. The decision to admit was taken by both a hematologist and an ICU-specialist in cooperation. The mean age was 56,9 years (range 16 – 84), mean LOS on the ICU were 6.4 days (range 1 – 52) and the mean APACHE II-score 26,4 (range 12–39).

In-ICU mortality was 30,7%, in hospital mortality 52,3% and one year mortality 70,7%.

A total of 20 patients died in the ICU. We found that secondary infection (10/20 50%), CRRT (11/20 55%), ventilator therapy (15/20 75%), ICU stay ≥ 7 days (11/20 55%) and a active ethical decision about therapy level (15/20 75%) all seem to correlate with a higher ICU mortality compared with a baseline mortality of 30.7%.

Only two out of the nineteen patients that had had a secondary infection were alive at 1 year. Twenty-eight patients admitted to the ICU had had a prior stay at the hematological ward of ≥ 7 days. Only 5 of those were alive after 1 yr.

Secondary infection on the ICU ward and a long pre-ICU stay in the hematological ward both appear to be predictors for long term survival.

Conclusion

Even with a moderate ICU-mortality, the one year mortality of hematologic patients seems very high.

We find the referral of critically ill patients with hematological disorders to the ICU an increasingly complex task. Improved possibilities for treating malignancies which previously would have been incurable, now renders the task of an initial triage for the ICU difficult.

We suggest that the long term survival rate (1 yr) of the hematological disease be taken into account when considering ICU admission. A long ICU stay with very poor prognosis is possibly both unethical and costly, if we take long term outcome into consideration.

Another suggestion could be a critical evaluation after e.g. 7 days of ICU treatment. This suggestion is based on the high 7+ days ICU stay mortality.

Poster Sessions Microcirculation in the critically ill: 0578–0591

0578

Microcirculation is Compromised in Extreme Exercise: A (Pilot) Observational Study in Marathonrunners

J. Weigel*1, A. Lima1, E. Klijn1, P. Goethart2, J. Hazelzet3, J. van Ochten4, J. Bakker1, C. Ince1

1Intensive Care, Erasmus Medical Center Rotterdam, Rotterdam, 2Clinical Physiology, Academic Medical Center, Amsterdam, 3Intensive Care, ErasmusMC-Sophia Children's Hospital, 4Erasmus University Rotterdam, Erasmus Medical Center Rotterdam, Rotterdam, Netherlands

Introduction

Microcirculatory dysfunction is thought to play a role in the development from shock to multiple organ dysfunction syndrome. The response to extreme exercise, such as marathon running, is in many aspects similar to the hemodynamic and inflammatory response seen in critical illness (1). In this study therefore we hypothesized that microcirculatory alterations would occur following a circulatory insult such as running a marathon.

Methods

We studied 5 marathonrunners (2M/3F) participating in the marathon of Rotterdam. A microcirculatory assessment was performed before, on the finish-line and the following day after the marathon. Microcirculation was measured by sublingual sidestream dark field (SDF) imaging (2). Sublingual capillary blood flow was estimated using semi-quantitaive microvascular flow index (MFI) in small (diameter 10–25 ώm), medium (25–50 ώm) and large (50–100 ώm) sized microvessels (0 = no, 1 = sluggish, 2 = intermittent, 3 = continous flow).

Results
MFI decreased in all groups of vessels after completion of the marathon race. These alterations persisted the following day in all subjects (P<0.05) (Table 1). We found a difference in weight (pre-race minus post-race) of −1.28 +/− 0.98 kg.
 

Before

After

>20 hours

Small

2.19 +/− 0.26

1.22 +/− 0.27

1.14 +/− 0.46

Medium

2.48 +/− 0.35

1.69 +/− 0.23

1.52 +/− 0.35

Large

2.88 +/− 0.14

2.20 +/− 0.15

2.43 +/− 0.56

Conclusion

Sublingual microcirculatory blood flow alterations were present in all participants after completion of the marathon race and persisted the following day. These alterations are consistent with the expected affect on the microcirculation of hypovolemia due to fluid loss and cytokine activation in marathonrunners.

0579

Does Increased Perfusion Pressure Improve Gastric Tube Microvascular Perfusion?

J. Van Bommel1, M. Van Genderen*1, J. De Jonge2, C. Ince1, D. Gommers1

1Dept of Intensive Care, 2Dept of Surgery, Erasmus MC, Rotterdam, Netherlands

Introduction

Esophagectomy with gastric tube reconstruction is the surgical treatment for cancer of the esophagus. Following reconstruction perfusion of the distal part of the tube depends exclusively on the microcirculation, making it susceptible to hypoperfusion and anastomotic leakage. It is unknown whether an increase in arterial perfusion pressure can exert a beneficial effect on gastric tube tissue perfusion. For this purpose we developed an experimental model of gastric tube reconstruction.

Methods

In 6 anesthetized and mechanically ventilated pigs, bodyweight 32 ± 2 kg (mean±SD), a midline laparotomy was performed. Formation of the gastric tube comprised ligation of all gastric arteries except the right gastro-epiploic artery. A gastric tube was formed by dissection of the lesser curvature and the gastro-esophageal junction.

Systemic hemodynamic monitoring was performed with an arterial line and a PA catheter. Microvascular blood flow (MBF) was assessed with laser Doppler flowmetry and microvascular HbO2 saturation (mHbSO2) with spectrophotometry (O2C, Lea Medizintechnik Giessen Germany). Measurements were made in antrum, corpus and fundus at increasing mean arterial blood pressures from 50 to 110 mmHg. For this purpose blood pressure was increased in steps of 10 mmHg with incremental infusion of noradrenaline.

Results

Heart rate, cardiac output and PAOPs did not change significantly throughout the experiment. MBF measurements are shown below (BL-baseline, DV-devascularisation). mHbSO2 was significantly lower only in the fundus, and increased significantly at a higher MAP. Open image in new window

Conclusion

We were able to establish a animal model for gastric tube reconstruction. In this model we demonstrated that although gastric perfusion is severely compromised following creation of the gastric tube, tissue perfusion could be improved significantly by increasing perfusion pressure with noradrenaline. On the other hand, this impairment in tissue blood flow was aggravated by hypotension. Whether these results should have consequences for the clinical setting remains to be investigated.

0580

Early Microvascular Changes in Sepsis, Severe Sepsis and Septic Shock

A. Spanos*1, S. Jhanji1, T. Harris2, R. M. Pearse1

1William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, 2Dept of Emergency Medicine, Barts and The London NHS Trust, London, United Kingdom

Introduction

Improvements in the early management of sepsis may result in improved outcomes. This has led to increased interest in the pathophysiology of the early stages of this disease process. Microvascular derangements are well described in patients with established sepsis but little data is available from patients in the early stages of hospital care.

Methods

Following local research ethics committee approval, observational data were collected in patients with sepsis, severe sepsis and septic shock within six hours of presentation. Patients received usual clinical care (not including early goal directed therapy). Microvascular flow index (MFI) was calculated from video images of the sublingual microcirculation (sidestream darkfield imaging) whilst cardiac index (CI) and oxygen delivery (DO2I) were measured non-invasively using a supra-sternal Doppler method. Additinal data included mean arterial pressur (MAP) and serum lactate. Data are presented as mean (SD) or median (IQR). Data were tested with the t-test where normally distributed and the Mann-Whitney U test where not normally distributed.

Results
48 patients were recruited. Data are presented in Tables 1 and 2.
Table 1

Patient Data for Study Groups

 

Sepsis n=20

Severe Sepsis n=18

Septic Shock n=10

Age (years)

42 (28–64)

50 (30–70)

52 (30–68)

Gender

12 female

10 female

2 female

APACHE II score

6 (4–9)

13 (9–16)

21 (17–26)

Lactate (mmmol/l)

2.0 (1.7)

2.1 (0.9)

3.1 (2.2)

Mortality (%)

0 (0%)

2 (11%)

5 (50%)

Table 2

Haemodynamic and Microvascular Data

 

Sepsis

Severe sepsis

Septic shock

MAP t=0 (mmHg)

90 (16)*

79 (15)*

68 (16)

MAP t=4 (mmHg)

82 (16)*

81 (15)*

65 (16)

DO2I t=0 hours

620 (187)*

519 (224)

349 (244)

DO2I t=4 hours

617 (184)*

408 (186)*

266 (142)

MFI (<20µm) t=0

3.0 (2.5–3.0)*

2.9 (2.5–3.0)

2.5 (1.9–3.0)

MFI (<20µm) t=4

3.0 (2.8–3.0)

3.0 (2.5–3.0)

2.8 (2.5–3.0)

* p < 0.05 vs septic shock. DO2 units: ml/min/m2

Conclusion

Characteristic sepsis related derangements of both global haemodynamics and microvascular flow can be readily identified early after presentation. The severity of these abnormalities correlates with the severity of sepsis. This data supports early commencement of treatments intended to restore microvascular perusion deficits in septic patients.

Grant Acknowledgement

This research was supported by a European Union Leonardo Da Vinci Award.

0581

Effect of Increasing Doses of Norepinephrine on Microvascular Perfusion and Tissue Oxygenation in Septic Shock

S. Jhanji*1, S. Stirling2, N. Patel3, C. Hinds1, R. M. Pearse1

1William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, 2Intensive Care Unit, National Hospital for Neurology and Neurosurgery, 3Intensive Care Unit, Barts and The London NHS Trust, London, United Kingdom

Introduction

Vasopressor therapy is routinely used to maintain an adequate mean arterial pressure (MAP) in patients with septic shock. Previous studies suggest that a MAP of 65mmHg provides adequate tissue perfusion and oxygenation. However, these studies utilised only indirect measures of microvascular flow and did not evaluate tissue oxygenation. The aim of this study was to determine the effects of various doses of norepinephrine (NE)on microvascular flow and tissue oxygenation.

Methods

Following approval by the local research ethics committee and the medical and healthcare products regulatory agency, data were collected from patients receiving NE infusion for the treatment of septic shock. The dose of NE was adjusted to achieve a MAP of 60 mmHg in the first instance and then MAPs of 70, 80 and 90 mmHg. Patients otherwise received usual clinical care. The following were determined after a 45 minutes stabilisation period at each MAP: cutaneous PtO2 (Clark electrode),mean red cell flux (laser Doppler), Microvascular Flow Index (MFI) from sublingual microcirculation images (sidestream darkfield imaging) and DO2I (lithium indicator dilution). Significance was tested using repeated measures ANOVA for parametric data or Friedman's test for non-parametric data. Data are presented as mean (SD) or median (IQR).

Results
16 patients were recruited (9 males, age 67 [55–72], APACHE II 23 [17–30]) of whom ten (63%) subsequently died. Significant changes were identified in global haemodynamics, red cell flux and tissue oxygenation. There was no difference in MFI (Table). No adverse effects of NE were noted.
Table 1

Changes in Haemodynamics and Microcirculation with Increasing Doses of Norepinephrine

MAP (mmHg)

60

70

80

90

p

DO2I (ml/min/m2)

487 (418–641)

536 (446–720)

550 (474–800)

662 (498–829)

0.0055

ScvO2 (%)

71 (6)

72 (7)

73 (7)

74 (7)

0.03

MFI (<20µm)

2.3 (1.9–2.7)

2.5 (2.5–2.8)

2.5 (2.3–2.7)

2.4 (2.0–2.7)

0.45

PtO2 (kPa)

5.9 (1.7)

6.7 (2.4)

7.1 (2.0)

7.3 (2.0)

<0.0001

Red cell flux (flux units)

26 (16–42)

27 (18–44)

27 (20–47)

33 (20–47)

0.04

NE (µg/kg/min)

0.18 (0.18)

0.25 (0.22)

0.35 (0.27)

0.41 (0.26)

<0.0001

Conclusion

In patients with septic shock, significant improvements in global haemodynamics, tissue oxygenation and cutaneous blood flow assessed by laser Doppler were seen when the dose of NE was increased to achieve higher MAPs. No differences in sublingual microcirculation were observed. These findings appear to contradict previous findings. The clinical significance of our observations is uncertain.

0582

Alterations in Tissue Oxygenation and Microvascular Flow During and After Major Abdominal Surgery

S. Jhanji*1, C. Lee2, D. Watson1, C. Hinds1, R. M. Pearse1

1William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, 2Intensive Care Unit, Barts and The London NHS Trust, London, United Kingdom

Introduction

Deranged microvascular flow and tissue oxygenation have been demonstrated in patients undergoing major surgery. However, there is little data describing how these parameters relate to each other or to clinical outcome.

Methods

Following approval by the local research ethics committee, observational data were collected prior to and for eight hours after major elective abdominal surgery. Patients received routine clinical care. Data included cutaneous tissue PtO2 (Clark electrode), sublingual microvascular flow index (MFI) and the proportion of perfused vessels (PPV) in small (<20lm) and large (>20µm) vessels (sidestream darkfield imaging), mean red cell flux (laser Doppler flowmetry), systemic oxygen delivery index (DO2I) (lithium indicator dilution) and mean arterial pressure (MAP). Data are presented as mean (SD) or median (IQR).

Results
25 patients were recruited (11 males; age 69 years [63–72]; P-POSSUM score 35 [31–40]) with two deaths. Following surgery, significant reductions were seen in PtO2:FiO2 ratios but there were no significant changes in mean red cell flux, PPV or MFI (Table 1). Postoperatively, small vessel MFI and PPV were significantly lower in patients who developed complications (Table 2) despite there being no difference in PtO2:FiO2 ratio or global haemodynamics between those with or without complications.
Table 1

Microvascular Changes Before and After Major Abdominal Surgery

 

Before surgery

After surgery

MAP (mmHg)

90 (10)

83 (15)

MFI (<20µm)

3.0 (2.5–3.0)

2.1 (2.5–2.8)

PPV (<20µm) (%)

85 (74–93)

74 (61–90)

PtO2:FiO2 ratio

30 (27–35)

22 (18–27)*

*p < 0.0001. Wilcoxon matched pairs test

Table 2

Post-Operative Do2I and Microvascular Haemodynamics According to Outcome

 

Complications (n=14)

No complications (n=11)

DO2I (8 hour mean)

384 (289–440)

360 (291–390)

MFI (<20µm) t=0 hrs

2.50 (2.00–2.75)

2.63 (2.13–3.0)

MFI (<20µm) t=4 hrs

2.5 (2.0–3.0)

3.0 (2.3–3.0)*

MFI (<20µm) t=8 hrs

2.75 (2.38–2.78)

3.0 (2.8–3.0)*

PPV (<20µm) t=0 hrs (%)

76(50–100)

89 (50–100)*

PPV (<20µm) t=4 hrs (%)

75 (53–100)

90 (67–100)**

PPV (<20µm) t=8 hrs (%)

100 (78–100)

100 (82–100)

*p < 0.05, **p < 0.01 Mann-Whitney test. DO2I units = ml/min/m2

Conclusion

After major abdominal surgery tissue oxygenation was impaired. Heterogeneity of small vessel flow was more marked in those patients who developed complications although there was no significant difference in tissue oxygenation. Large vessel flow remained unaffected suggesting microvascular shunting.

Grant Acknowledgement

ESICM Spacelabs Intelligent Monitoring Award. Open image in new window

0583

Oxidative Stress Generated by Iron Supply in Micro Circular Endothelial Cells

M. Mesnil*1, O. Huet1, A. Harrois1, L. Dupic2, D. Benhamou1, E. Vicaut3, J. Duranteau1

1Département d'anesthésie réanimation, CHU Bicêtre, Le Kremlin Bicetre, 2Réanimation pédiatrique polyvalente, CHU Necker, 3Biophysique et médecine nucléaire, CHU Lariboisiere, Paris, France

Introduction

Anemia occurs in more than 70% of intensive care patients, and is an independent risk factor of morbidity and mortality. Intra venous iron supply can be suggested, but iron may induce reactive oxygen species (ROS) production, in particular in inflammatory context. The aim of the study was to assess the effect of iron supply on ROS production with a model of human pulmonary micro circular endothelial cells (HPMEC), and to evaluate a possible potentialization of this oxidative stress first induced by a prior exposition to TNF/alpha.

Methods

Confluent HPMEC in monolayer conformation have been placed in a perfusion chamber mounted on an inverted microscope equipped with a digital camera. Cells were perfused with Krebs buffer containing a fluorescent probe (2′7′ DCFH), which permitted to assess the ROS production, mainly H2O2. 4 groups have been studied: control, iron (iron hydroxide saccharose), TNF/alpha, TNF/alpha + iron. Oxidative stress level was evaluated with DCFH fluorescence variation compared to level at TNF/alpha introduction, after 40 minutes of stabilization. Statistics analysis was performed with ANOVA for repeated measure.

Results

Exposition of HPMECto iron does not induce intracellular ROS production. On the other hand, after 20 minutes of stimulation with TNF/alpha, iron supply leads to a significant increase of ROS production (p=0.03), which persists despite iron exposition ending (Fig 1). Open image in new window

Conclusion

Iron exposition of endothelial cells first exposed to TNF/alpha potentializes ROS production induced by this inflammatory mediator. Intravenous iron supply could be deleterious in patients in inflammatory state.

0584

Laser Speckle Imaging of Nailfold Microcirculation During Dobutamin Infusion in Patients Suspected of Cardiac Ischemia

E. Klijn*, C. den Uil, C. Ince, J. Bakker

Intensive Care, Erasmus Medical Center, Rotterdam, Netherlands

Introduction

The aim of the study was to investigate whether peripheral microcirculatory flow increases by administration of dobutamin, measured with Laser Speckle Imaging (LSI). LSI is a new technique which exploits the fact that the random speckle pattern that is generated when tissue is illuminated by laser light changes when blood cells move within a region of interest. The contrast image is processed to a color-coded image that correlates with blood flow in tissue.

Methods

We used LSI for evaluation of the microcirculation in the nail fold area. 12 Patients scheduled for dobutamin stress scintigram in order to evaluate cardiac ischemia in the out patients clinic of the cardiology department were included in the study. Patients received increasing dosages of dobutamin (10–20–30 mcg/kg/min) or until target heart rate was reached. Microcirculatory flow was measured during the whole experiment. We determined the change in flow for each step compared to baseline flow (delta flux) in arbitrary units (AU).

Results

Data are presented as median [interquartile range]. Heart rate increased from a baseline value of 70 bpm [62–85] to 79 bpm [68–85], to 95 bpm [80–119], to 114 bpm [98–126] after the subsequent dobutamin dosages. There were no significant changes in blood pressure observed. At a dobutamin dosage of 20 mcg/kg/min there was a significant increase in peripheral flow seen as compared with baseline. 2 patients reached the target HR at 20 mcg/ kg/min and no subsequent dose was administered. Open image in new window

Conclusion

We found that dobutamin in a dosage of 20 mcg/kg/min increases peripheral blood flow measured with LSI at the nailfold area.

0585

Low-Dose Nitroglycerin Improves Sublingual Microcirculation in Acute Heart Failure

C. den Uil*1, W. Lagrand2, E. Klijn3, L. Jewbali3, C. Ince3, P. Spronk4, M. Simoons1

1Cardiology, Erasmus MC, Rotterdam, 2Intensive Care, LUMC, Leiden, 3Intensive Care,

Erasmus MC, Rotterdam, 4Intensive Care, Gelre Hospitals, Apeldoorn, Netherlands

Introduction

Nitroglycerin (NTG) is frequently administered to patients with acute heart failure to reduce ventricular afterload. We tested the hypothesis whether NTG promotes microcirculatory blood flow due to its vasodilating action.

Methods

We included acute heart failure patients who were admitted to the Intensive Cardiac Care Unit. In each patient, an intravenous loading dose of 0.5 mg NTG was given followed by continuous NTG infusion (2 mg/h). Using Side-stream Dark Field imaging, sublingual micro-vascular perfusion was evaluated before NTG administration (T0) and 15 minutes after initiation of NTG (T1). At least three video sequences of the microcirculation were recorded and analyzed. Microscan Analysis Software was used to measure Functional Capillary Density (FCD), an indicator of tissue perfusion. Capillaries were defined as the micro-vessels with a diameter of <25 mum. Values are expressed as median [P25–P75].

Results

Fourteen patients were included in this study. Mean arterial pressure (MAP) decreased after NTG administration (81 [73–85] mmHg at T0 vs. 76 [65–84] mmHg at T1, p=0.02), whereas central venous pressure (17 [14–19] mmHg at T0 vs. 16 [13–18] mmHg at T1) and heart rate (89 [62–108] bpm at T0 vs. 85 [65–109] bpm at T1) did not change significantly. Nitroglycerin improved FCD (figure: 11.2 [9.6–12.6] mm−1 at T0 vs. 12.4 [11.2–13.9] mm−1 at T1, p=0.03). Open image in new window

Conclusion

Micro-vascular perfusion can be improved by low-dose NTG in patients with acute heart failure.

0586

The Effects of Red Blood Cell Transfusions on Sublingual Microvascular Perfusion in Patients Undergoing on-Pump Cardiac Surgery

K. Yuruk, E. Almac*, C. Ince, P. T. Goedhart

Clinical Physiology, Academic Medical Center, Amsterdam, Netherlands

Introduction

Red blood cell transfusions are often given in a number of clinical conditions in order to correct the diminished tissue oxygenation by primarily increasing systemic hemoglobin concentration. However it is controversial whether the increase in systemic hemoglobin concentration also leads to an improvement of microcirculatory oxygenation and corrects tissue hypoxia. In this study we investigated the efficacy of red blood cell transfusions in patients undergoing on-pump cardiac surgery and whether blood transfusions improve microcirculatory parameters and oxygen delivery to the tissues.

Methods

In 8 patients (group A) sublingual microcirculation and perfusion was imaged using Sidestream Dark Field (SDF) (Microscan, Microvision Medical, Amsterdam, Netherlands). This device is a handheld microscope using green light-emitting diodes (LEDs) of a wavelength of 530 nm. This wavelength of light is absorbed by the hemoglobin in erythrocytes, so these cells can be clearly observed as flowing cells. In 11 patients (group B) sublingual reflectance spectrophotometry (O2C®; Lea Medizintechnik, Germany) was used to measure microcirculatory hemoglobin concentration and hemoglobin oxygen saturation. Both measurements were performed 15 minutes before and 30 minutes after blood transfusion during on-pump cardiac surgery.

Results

In group A (n=8) 5 male, 3 female, 64±11 red blood cell transfusions increased hemoglobin concentration from 4.4±1.03 mmol/L to 5.3±0.82 mmol/L (p < 0.01), mean arterial pressure from 60.1±10.9 mmHg to 66.1±7.7 mmHg (p<0.03), vessel density from 10.5±1.1 mm/mm2 to 12.8±1.1 mm/mm2 (p<0.001). Microvascular flow index (MFI) was analyzed according to the method described before and was not significantly altered by red blood cell transfusion. Red blood cell storage time did not have a statistically significant influence on microvascular perfusion and oxygenation.

In group B (n=11) 10 male, 1 female, 64±12 years, following red blood cell transfusions systemic hemoglobin concentration increased in each patient, from 4.9±0.7 mmol/L to 5.6±1.0 mmol/L. Similarly, sublingual microcirculatory hemoglobin concentration increased in each patient following blood transfusion from an average of 62.3±10.8 units to and 70.2±8.9 units (p < 0.001) whereas microcirculatory hemoglobin oxygen saturation increased from 64.9±11.6% to 67.7±10.2% (p < 0.05) respectively.

Conclusion

Microcirculatory perfusion and oxygenation, parallel to systemic variables, increased following transfusion of red blood cells. Combining these two methods together gives an integrative and comprehensive monitoring of the functional activity of the microcirculation at the bedside by providing information about microcirculatory oxygen delivery by the RBC and oxygen availability.

0587

Low Incidence of Rectal Microcirculatory Flow Alterations after Elective Cardiac Surgery

E. C. Boerma*1, A. Konijn1, K. Kaiferova1, C. Ince2

1Department of ICM, Medical Center Leeuwarden, Leeuwarden, 2Department of Physiology, Academic Medical Center Amsterdam, Amsterdam, Netherlands

Introduction

During on-pump cardiac surgery patients are subject to a number of hemodynamic changes such as hypotension, hemodilution, hypothermia, shifts from pulsatile to continuous bloodflow and inflammation. These insults have been associated with microcirculatory alterations especially in the gastro-intestinal tract. We tested the hypothesis that cardiac surgery induces microcirculatory alterations in the rectal mucosa, directly observed by Sidestream Darkfield (SDF) imaging.

Methods

In a single center observational study in 26 elective on-pump cardiac surgery patients SDF imaging of the rectal mucosa was performed in the first hour after ICU admittance. Fecal contamination of the rectal pouch was a contra indication for enrolment. Semi-quantitative analysis was performed as described in detail elsewhere1. The percentage of well-perfused capillaries was predefined as the number of rectal crypts with a flow score >1 divided by the total number of crypts × 100%. Data are expressed as medians and inter quartile ranges (IQR,[P25–P75]).

Results
Baseline characteristics of the study population are summarized in Table 1. 23 patients underwent a CABG procedure, 2 heartvalve surgery and 1 CABG + heartvalve surgery. All patients survived to hospital discharge. Microvascular Flow Index was 3[3–3] and the percentage of well-perfused capillaries 84.5%[72–92.5].
Table 1

Baseline Characteristics Study Population (N=26)

Variable

Median [IQR]

Age

63 [59–73]

EuroSCORE

3.5 [2–6]

Cardio Pulmonary Bypass duration (min)

127 [88–143]

Aortic Cross Clamping time (min)

82 [59–90]

Lowest Hematocrit (%)

25 [23–27]

Mean Arterial Pressure (mmHg)

82 [75–91]

Central Venous Pressure (mmHg)

10 [7–12]

Central Venous Oxygen Saturation (%)

72 [67–75]

Lactate (mmol/L)

2.4 [1.7–3.2]

Conclusion

Despite significant hemodynamic insults during elective on-pump cardiac surgery, rectal microcirculatory flow alterations during the first hour of ICU admittance could not be demonstrated.

0588

Low Oxygen Delivery is Associated with Cellular Damage of the Gut Lumen as Shown by Elevated Intra Luminal Glycerol Levels

N. Al-Subaie*1, R. Hagger2, T. Addei3, E. D. Bennett1, A. Rhodes1, R. M. Grounds1, G. M. Hall3

1General intensive care, 2Dept. of Colorectal surgery, 3Dept. of Anaesthesia, St. George's Hospital, London, United Kingdom

Introduction

Intra operative goal directed therapy is associated with better outcome. This maybe related to splanchnic flow preservation which prevents gut wall damage and bacterial translocation. This study utilises microdialysis to measure intra luminal glycerol, which was shown previously to be associated with the degree of cellular damage [1], to assess the impact of changes in oxygen delivery during major surgery on large bowel.

Methods

An observational study conducted on patients undergoing major abdominal surgery. Intra operative oxygen delivery is derived from oesophageal Doppler monitoring (Deltex Doppler). A CMA 61 microdialysis catheter is placed in the ascending colon and continuously flushed with 2uL/min. Samples are obtained every 6 to 8 minutes if clinically permitted.

Results

Data from eleven adult patients undergoing major abdominal surgery were collected. The number of samples obtained from microdialysis was between 4 and 14 depending on the duration of surgery. Maximum intra luminal glycerol level attained correlates inversely to the median oxygen delivery of each patient (rho r = −0.709, p=0.015 graph 1). Open image in new window

Conclusion

Low oxygen delivery is associated with high intra luminal glycerol level indicating cellular damage.

0589

Transcutaneous Oxymetry in Septic Shock : Pronostic Value of an Oxygen Challenge Test

F. Vallée*, A. Mari, J. Bedel, J. Paraguette, S. Silva, M. Génestal, O. Fourcade

Intensive Care Unit, Universitary hospital of Purpan, Toulouse, France

Introduction

Alterations of tissue oxygenation and perfusion during septic shock (SS) is a crucial issue. Monitoring peripheral perfusion could be an early and sensitive way to detect vital organ hypoperfusion (1). Transcutaneous oxymetry (TO) allows a non invasive measurement of oxygen trancutaneous pressure (PtcO2) which is related to arterial oxygenation pressure (PaO2) and circulatory status (2). A recent study emphasizes the pronostic value of parameters from TO in patients in SS (3). We hypothesized that variations of PtcO2 (DeltaPtcO2) induced by variations of oxygen arterial content estimated by PaO2 (DeltaPaO2) during an oxygen challenge test could be related to outcome in SS.

Methods

We conducted a prospective, observational study, in a universitary hospital. From January to August 2007, all consecutive patients admitted for severe SS requiring mechanical ventilation and vasopressor therapy after early hemodynamic optimisation were included. PtcO2 was measured (Tina 4, Radiometer®, Copenhagen, Denmark) in basal condition and at the end of an oxygen challenge test (OCT: FiO2 1.0 during 15 min.). Arterial blood gaz were performed before and after OCT. DeltaPtcO2/DeltaPaO2 (%) during OCT was calculated 3 times during the first 2 days of hospitalization (H0, H24, H48). The time course of DeltaPtcO2/DeltaPaO2 was analysed and compared between survivors (S) and non survivors (NS) at day 28th (D28), as well as usual hemodynamics and metabolic parameters: mean arterial pressure (MAP), cardiac index (CI), central venous oxygen saturation (ScvO2), norepinephrine infusion amount and lactatemia.

Results

Twenty one patients were included. Eight of them had died at D28 (NS: n=8; S: n=13). Hemodynamic parameters (MAP, CI, ScvO2), infusion of vasopressor and lactatemia were similar between S and NS at H0, H24 and H48. There was no significant difference in the oxymetric index between the 2 groups at H0 (DeltaPtcO2/DeltaPaO2: S:57±31% vs NS:36±38%, p=0.17). However, at H24 and H48, patients showed a DeltaPtcO2/DeltaPaO2 index significantly higher in group S compared with NS (S:83±40% vs NS:37±22%, p=0.01 at H24; S:105±54% vs NS :43±34%, p=0.02 at H48).

Conclusion

These preliminary results suggest that increase in DeltaPtcO2/DeltaPaO2 seems to predict better outcome in severe septic shock while circulatory and metabolic parameters such as ScvO2 or lactatemia during the first 2 days of resuscitation remain similar. OCT could help us to early detect patients who will undergo pejorative outcome in septic shock but further analysis is needed.

0590

The Effect of Vasoactive Amines on the Gut Microcirculation of Rats with Endotoxemia

F. E. Nacul*, I. L. Guia, B. Sabino, E. Tibiricçá

Laboratory of Neurocardiovascular Pharmacology, FIOCRUZ, Rio de Janeiro, Brazil

Introduction

The choice of catecholamines in septic shock patients has been an ongoing debate for several years. Because vasopressors can accentuate mesenteric hypoperfusion secondary to circulatory shock, gut circulation is a major consideration in supporting hemodynamics. The objective of our study was to determine the influence of dopamine, norepinephrine, phenylephrine, epinephrine, and an association between norepinephrine and dobutamine on mesenteric microcirculation in a rat model of endotoxemia.

Methods

Sixty Wilstar rats (250–350 g) were divided into 6 groups of 10 rats. The animals in groups 2, 3, 4, 5, and 6 received 2 mg/Kg of LPS from Escherichia coli — LPS serotype 055:B5. After a 40% reduction in MAP, the animals were treated by the administration of dopamine, norepinephrine, phenylephrine, epinephrine, or the association of norepinephrine with dobutamine. The functional capillary density (FCD) of the intestinal mucosa was estimated with intravital fluorescence videomicroscopy using an epifluorescent microscope. The FCD was estimated before the administration of LPS, after a 40% decrease in MAP, and after the normalization of the blood pressure with vasoactive amines.

Results

The results are shown in the figure (* = p < 0.005). Open image in new window

Conclusion

1) Vasoactive amines normalized blood presure of endotoxemic rats; 2) Vasoactive amines reduced FCD of gut mucosa; 3) When used in association with norepinephrine, dobutamine might have protective effect on gut microcirculation.

0591

Hypercapnic and Metabolic Acidosis and the Effect on Perfusion of Gut and Renal System

S. F. X. Kreyer*, T. Muders, J. Zinserling, C. Kricklies, D. Varelmann, H. Wrigge, C. Putensen

Dept. of Anaesthesiology and ICM, University Bonn, Bonn, Germany

Introduction

To reduce the ventilator induced lung injury (VILI) in acute lung injury (ALI) protective ventilation with low tidal ventilation is used. During this low-tidal-ventilation in critical ill patients hypercapnic acidosis (HA)can accur. Hypercapnia has direct effects on the regulation of the vascular bed, which may change organ perfusion. The influence of hypercapnia and acidosis on the regional perfusion (RP) is not exactly known.

Methods

16 anesthetised pigs were intubated, mechanically ventilated and recieved an experimental ALI by acid aspiration. To achieve HA pigs were randomized into two groups: group 1) Hypoventilation leading to HA and 2) acid infusion for metabolic acidosis (MA). The regional perfusion was measured with the coloured microspheres technique (Dye Trak®, Triton Technology, San Diego, CA), while systemic paramaters were measured with the PICCO®-system. Measurements were taken under baseline (BL), HA and MA. Statistics: repeated measures ANOVA and Tukey HSD.

Results

Cardiac output (CO) tended to be higher during HA. While the perfusion of the gut and the spleen was increased, the stomach- and renalperfusion remained unchanged. In contrast MA did not affect global or regional perfusion.

Statistic declaration in Table 1: time + p < 0.05 ++ p < 0.01 +++ p < 0.001; interaction * p < 0.05 ** p < 0.01 ***p < 0.001; post hoc $ p < 0.05 $$$ p < 0.001 vs bl, p < 0.001 HA vs MA.

mean (SD) RBF (ml/g/min)

hypercapnia baseline

hypercapnia hypercapnia

metabolic acid baseline

metabolic acid metabolic acid.

ANOVA

stomach

0,4(0,2)

0,5(0,2)

0,4(0,2)

0,3(0,1)

 

duodenum

0,4(0,2)

0,8(0,3)

0,5(0,2)

0,5(0,2)

+*

jejunum

0,6(0,2)

1,1(0,6) $

0,5(0,3)

0,6(0,3)

++ *

colon

0,4(0,4)

0,6(0,4) $

0,3(0,2)

0,3(0,1)

+*

spleen

2,3(1,0)

4,4(1,9) $ §

2,6(1,0)

2,1(0,8)

**

kidney

3,7(1,4)

4,3(1,7)

3,2(0,9)

2,9(0,6)

 

PaCO2 (mmHg)

41(4)

82(3) $$$§§§

43(3)

46(7)

+++ ***

pH

7,44(0,02)

7,19(0,04) $$$§§§

7,4(0,08)

7,19(0,04) $$$§§§

***

CO (l/min)

4,4(0,7)

5,6(2,3)

4,9(1,2)

4,3(1,1)

p0,07

Conclusion

Not acidosis but hypercapnia leads to increased RP in several intestinal organs. Maybe this is a protective effect of hypercapnia.

Grant Acknowledgement

Departmental funding and NOVALUNG.

Poster Sessions Sepsis therapies: 0592–0605

0592

Newly Developed Cytokine Adsorption Column Ctr-001 in Sepsis

Y. Suzuki*, G. Takahashi, S. Shibata, S. Endo

Critical Care Medicine, Iwate Medical University, Morioka, Japan

Introduction

Sepsis is a common cause of morbidity and mortality in intensive care unit, and delayed diagnosis and therapy is associated with increased mortality. Continuous hemodiafiltration (CHDF) and polymyxin-B immobilized fiber (PMX) have recently been widely performed mainly in critical care. To study effects of newly developed cytokine absorption colmu CTR-001 (Kanaka Co. Osaka, Japan), RCT was performed in patients with sepsis.

Methods

A total of 18 patients with early septic shock or septic organ dysfunction was enrolled.. Nine of 18 were randomized to direct hemoperfusion (DHP). All patients received supportive intensive care, and those randomized to DHP received direct hemoperfusion for 4 hours more than two times up to 14 times during 14 days. MEASUREMENTS(1): We measured the plasma concentration of interleukin 6, 8, 1-beta, tumor necrosis factor alpha. APACHE II score was evaluated for each patient 1st, 7th, and 14th day after starting treatment before the treatment in the morning.

Results

The decrease of APACHE II score from the pretreatment level at 7th day was significantly larger in the treatment group than in the control group (p=0.0189; Mann-Whitney test). Adsorption column related serious adverse events were not observed in DHP group. The concentration of interleukin 6 and interleukin 8 in the plasma decreased from the pretreatment level in the DHP group significantly (p=0.0464, 0.0464 respectively; Wilcoxon test).

Conclusion

Newly developed direct hemoperfusion column improved the septic shock better than the ordinary supportive intensive care.

0593

Effects of Endotoxin Adsorption Therapy on Hemodynamics, Procalcitonin and C-Reactive Protein in Patients with Septic Shock

C. Mitaka*, N. Tsuchiya, K. Kawada, Y. Nakajima, T. Imai

Critical Care Medicine, Tokyo Medical and Dental University Graduate School, Tokyo, Japan

Introduction

Despite of antibiotic therapy, fluid therapy and vasopressive/inotropic therapy, the mortality rate of septic shock remains high. In the pathogenesis of septic shock, endotoxin plays an important role. Recently, endotoxin adsorption therapy, polymyxin B-immobilized fiber column (PMX) hemoperfusion, has been widely used in patients with septic shock in Japan. PMX tretment has been based on the binding property of polymxin B to lipid A of endotoxin. Endotoxin induces excessive generation of cytokines which increase systemic inflammatory response and cause tissue damage. Although decrease in concentration of endotoxin has been reported, the detailed mechanism of PMX treatment is not known. Therefore, we investigated the effects of PMX on hemodynamics and plasma concentrations of procalcitonin and C-reactive protein in patients with septic shock.

Methods

Patients with septic shock (ACCP/SCCM-criteria) were included. Demographic data, routine biochemistry, microbiological data, infection focus, the acute physiology and chronic health evaluation (APACHE) II score, the Sequential Organ Failure Assessment (SOFA) score, and 28-day mortality were recorded. PMX treatment was performed 2 ∼ 27 hr (mean 12 hr) for each application once or twice. Blood concentrations of endotoxin, procalcitonin, and C-reactive protein (CRP) were measured before and after PMX treatment. Blood concentrations of endotoxin were measured using the high-sensitivity endotoxin assay based on the kinetic turbidometric Limulus assay (Toxinometer). Changes in hemodynamic parameters and PaO2/FIO2 ratio were also evaluated.

Results

Sixteen patients (11 men and 5 women, mean age 65 years old) with septic shock were studied. The 28-day mortality rate was 50%. PMX treatment significantly (P<0.05) increased mean arterial pressure and significantly (P<0.05) decreased body temperature, and tended to improve PaO2/FIO2 ratio. Blood concentrations of endotoxin, procalcitonin and CRP were markedly high in all patients before PMX treatment. In the survivors, blood concentrations of endotoxin significantly (P<0.05) decreased, and those of procarcitonin and CRP tended to decrease after PMX treatment.

Conclusion

These findings showed that PMX treatment improved hemodynamics and tended to suppress inflammatory response, suggesting that PMX treatment might be a useful strategy in septic shock by reducing systemic inflammatory responses.

Grant Acknowledgement

Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

0594

Therapeutic Effect of Xigris on Sepsis-Associated Encephalopathy

H. D. Spapen*, D. Nguyen, B. Scott, L. Huyghens, J. De Regt, M. Diltoer

Intensive Care, University Hospital VUB, Brussels, Belgium

Introduction

XigrisR has been shown to improve cardiovascular and respiratory failure in severe sepsis. Sepsis-associated encephalopathy (SAE) is a diffuse cerebral dysfunction induced by the inflammatory response to infection. SAE has an unknown incidence but its occurrence is related to an increased mortality rate. We have previously described high levels of the brain-specific S-100⌆ta protein in 42% of patients with severe sepsis and septic shock(1. Distinct cerebral lesions were found with S-100⌆ta > 4 µg/L. S-100⌆ta levels between 0.06 and 2 µg/L, however, were typically associated with white matter lesions which are thought to represent the pathological substrate of SAE. We studied whether XigrisR influenced S-100⌆ta in patients with acute septic shock who presented with increased baseline levels of this biomarker.

Methods

Patients with septic shock who required mechanical ventilation were recruited. All received standard and goal-directed resuscitation. Glasgow coma scale (GCS) was calculated before start of sedation. Contrast computed tomography (CCT) of the brain was performed in all patients with a GCS <14 to exclude significant pre-existing or acute neurological disease. XigrisR was given as a continuous infusion of 24 µg/kg/h for 96 h. Patients who did not receive XigrisR had a known contra-indication for the drug. S-100⌆ta was measured before sedation and start of XigrisR (d1) and then daily during XigrisR infusion. The lower limit of normal was < 0.5 µg/L.

Results

Sixty-six patients were consecutively enrolled. Twelve patients were excluded from follow-up due to death within 4 days after admission (n=5) baseline S-100⌆ta values> 4 µg/L (n=2), and a pathological brain CTT (n=5).24 patients had S-100⌆ta levelsg<0.5 µg/ L upfront that did not increase thereafter. Of the remaining 30 patients, 16 received XigrisR. In this group, S-100⌆ta values — though higher on d1 than in untreated subjects (1.3 +/- 0.8 µg/L vs. 0.9 +/- 0.4 µg/L; p = 0.16) — progressively and significantly decreased (1.0 +/- 0.8 µg/ L at d2, 0.9 +/- 0.6 µg/L at d3, and 0.8 +/- 0.7 µg/L at d4; all p<0.05 vs. d1). S-100⌆ta tended to increase in the untreated group. In both treatment groups, no correlation was found between GCS and S-100⌆ta at baseline.

Conclusion

S-100⌆ta is increased in almost half of the patients with acute sepsis-induced cardiovascular and respiratory failure. S-100⌆ta levels that are likely to correspond with the presence of SAE decrease during treatment with XigrisR.

0595

Effects of Hemoperfusion with Immobilized Polymyxin-B Fiber Column on Cytokine Plasma Levels in Patients with Intestinal Sepsis: Results of a Single Centre Experience

A. Peris, G. Zagli, S. Damiani, A. Di Filippo, S. Matano*, M. Bonizzoli, S. Batacchi

Intensive Care Unit of Emergency Department, Careggi Teaching Hospital, Florence, Italy

Introduction

Endotoxin (ET) is considered one of the most important pathogenetic factor involved in sepsis/septic shock. A potential beneficial role of hemofiltration with an immobilized polymyxin-B fiber (PMX) column (Toraymyxin®, Toray Industries, Tokyo, Japan) has been proposed, especially in Gram-negative sepsis. However, the effect of PMX on cytokines plasma levels (IL-6, IL-10, TNF-alfa) remain to be clarified.

Methods

Eight patients (mean age: 43.6) with diagnosis of severe sepsis/septic shock from intestinal sources admitted in our Intensive Care Unit from Octr 2006 to Dec 2007 have been treated with PMX cartridge hemofiltration. Data were collected 24h before and 24h after the PMX-treatment, as determinations of plasma levels of IL-6, IL-10 (Bender MedSystem, CA), and TNF-alfa (Biosource Europe, BE). Patients were followed up for 28 days.

Results
Data of all patients and subgroups are represented in Tables Tab1–2. No adverse events have been reported.
Table 1

Clinical and Laboratory Data of All Patients

 

Pre-treatment

Post-treatment

SOFA

17.1±1.1

13.1±1.6*

IL-6 (pg/ml)

272.3±66.2

120.1±20.4*

IL-10 (pg/ml)

148.0±38.7

37.57±13.2*

TNF-alfa (pg/ml)

73.3±14.6

19.9±6.2*

Procalcitonin (ng/ml)

41.7±19.5

20.5±14.5

Norepinephrine (µg/Kg/min)

0.6±0.2

0.2±0.08*

PaO2/FiO2 ratio

115±16.7

227±26.2*

Statistical analysis: Student's t-test (*P<0.05). Values are reported as mean±SEM

Table 2

Clinical and Laboratory Data of Surviving (4) and Deceased Patients (4).

 

Surviving patients Pre-treatment

Surviving patients Post-treatment

Deceased patients Pre-treatment

Deceased patients Post-treatment

SOFA

17.5±0.9

12.5±1*

16.7±2.4

14±4

IL-6 (pg/ml)

304.5±84.1

82.8±16.3*

229.3±64.4

170.0±53.9

IL-10 (pg/ml)

88.5±18.02

8.3±1.4*

227.3±58.9

76.7±19.0*

TNF-alfa (pg/ml)

41.5±11.1

11.0±3.3

115.7±57.2

31.7±11.2

Procalcitonin (ng/ml)

39.8±24.8

27.0±22

44.4±37.9

11.8±7.5

Norepinephrine (µg/Kg/min)

0.85±0.34

0.14±0.09*

0.32±0.04

0.28±0.16

PaO2/FiO2 ratio

133±23.2

277.5±13.2*

91.67±19.65

159.3±21.3*

Statistical analysis: Student's t-test (*P < 0.05). Values are reported as mean±SEM

Conclusion

PMX-hemofiltration significantly reduces plasma levels of IL-6, IL-10, and TNF-alfa (Table 1), especially in patients with a better outcome at 28 days (Table 2). These data suggest a possible correlation between the post-treatment IL-6 levels and a more favourable outcome, whereas pre-treatment IL-6 levels do not seem to be predictive. In contrast, high pretreatment levels of IL-10 and TNF-alfa appears to be related with a poor outcome (Table 2). Finally, significant reduction in cytokine levels seems to be linked with an improvement of hemodynamic and respiratory parameters (Tables 1–2), according with literature.

0596

Polymyxin B Treatment in Septic Patients with Peritonitis Developed after Abdominal Surgery

G. Sbrana*1, F. Franchi2, E. Bonora2, M. Ferraro2, P. Giomarelli2, R. Madonna1, S. Scolletta2

1Anestesia e Rianimazione, Ospedale Misericordia, Grosseto, 2Anestesia e Rianimazione, University of Siena, Siena, Italy

Introduction

Polymyxin B (PMX) bound and immobilized to polystyrene fibers is a medical device that aims to remove circulating endotoxines by adsorption, theoretically preventing the progression of the biological cascade of sepsis. The aim of this study was to evaluate the effects of the haemoperfusion with PMX (HP-PMX) on organ function and mortality in septic patients undergoing surgical abdominal toilette for peritonitis developed after abdominal surgery.

Methods

12 septic patients (age 63±12, male/female 6/6)were studied. Inclusion criteria were: peritonitis due to gut perforation and/or anastomotic leakage after abdominal surgery, and diagnosed organ failure (two or more organs). Exclusion criteria were: age<18, thrombocytopaenia, hemophilia, PMX allergy. After the intensive care unit (ICU) admission, wound, blood, and peritoneal liquid cultures were performed. At the same time, a single 2- hours-cycle of HP-PMX with the Aquarius haemofiltration machine (Edwards, Edwards Lifesciences) was started at 80ml/min. Haemodynamic parameters were derived with a minimally invasive pulse contour monitoring system (Vigileo, Edwards Lifesciences). Several haemodynamic, respiratory, renal and liver data were collected before starting the HP-PMX (T0), at the end of the HP-PMX (T1), at 24 (T2), and at 48 (T3) hours. T-test and analysis of variance for repeated measures were applied.

Results

Haemodynamic data: a significant increase in systolic arterial pressure from T0 to T3 (121±15 vs 145±10) was observed. From T0 to T2, a significant reduction in norepinephrine infusion occured (0,26±0,07, 0,14±0,09, 0,09±0,04 mcg/Kg/min, T0, T1,T2, respectively). A T3, only two patients received norepinephrine infusion (0,1 and 0,14 y/Kg/ min). Central venous pressure (CVP, mean at T0 =12,7±1,5) and stroke volume variation (SVV%, mean at T0 = 4,5%) did not show significant difference over the time of the study. Respiratory data. a significant increase in PaO2/FiO2 ratio from T0 to T1 (180±42,1 vs 295±43, p<0.05), and from T2 to T3 (304±65 and 298±60, p0.05) was found. The average time of mechanical ventilation was 10,2±8 days. Renal data: serum creatinine values decreased from T0 to T1 (3,1±0,6 vs 2,3±0,7, p<0.05) and from T1 to T2 (2,3±0,7 vs 1,9±0,6 p < 0,05). Continuous renal replacement therapy (CRRT) was applied for 3 patients at T3. Only one of them underwent CRRT up to the fifth day because he had chronic renal failure requiring preoperative dialysis. No statistical significance was found for liver function indexes. The ICU stay resulted 15,8±8 days. The 28 days mortality rate was 25%. The cultures confirmed the presence of Gram- bacteria in each patient.

Conclusion

In our institution the mortality rate of this kind of patients was about 65%. An optimal timing about the use of the HP-PMX depends on the pathology. In septic patients, with peritonitis due to perforation and/or anastomotic leakage after abdominal surgery, the presence of Gram- bacteria is actually a common condition. Our findings demonstrated that a single 2-hours-cycle of treatment with HP-PMX seemed to reduce the incidence of organ dysfunction and mortality. An earlier surgical toilette of the infected site, combined with a treatment with HP-PMX, may significanlty improve the outcome in these critically ill subjects.

0597

Pmx Treatment of Septic Patients Improved Viability of Pulmonary Epithelial Cells

V. Puntorieri, E. L. Martin*, D. Pasero, L. Del Sorbo, V. M. Ranieri

Anesthesia and Intensive Care, University of Turin, Turin, Italy<