In-line filtration reduces severe complications and length of stay on pediatric intensive care unit: a prospective, randomized, controlled trial
- First Online:
- 2.9k Downloads
Particulate contamination due to infusion therapy carries a potential health risk for intensive care patients.
This single-centre, prospective, randomized controlled trial assessed the effects of filtration of intravenous fluids on the reduction of complications in critically ill children admitted to a pediatric intensive care unit (PICU). A total of 807 subjects were randomly assigned to either a control (n = 406) or filter group (n = 401), with the latter receiving in-line filtration. The primary endpoint was reduction in the rate of overall complications, which included the occurrence of systemic inflammatory response syndrome (SIRS), sepsis, organ failure (circulation, lung, liver, kidney) and thrombosis. Secondary objectives were a reduction in the length of stay on the PICU and overall hospital stay. Duration of mechanical ventilation and mortality were also analyzed.
Analysis demonstrated a significant reduction in the overall complication rate (n = 166 [40.9 %] vs. n = 124 [30.9 %]; P = 0.003) for the filter group. In particular, the incidence of SIRS was significantly lower (n = 123 [30.3 %] vs. n = 90 [22.4 %]; P = 0.01). Moreover the length of stay on PICU (3.89 [95 % confidence interval 2.97−4.82] vs. 2.98 [2.33−3.64]; P = 0.025) and duration of mechanical ventilation (14.0 [5.6−22.4] vs. 11.0 [7.1−14.9] h; P = 0.028) were significantly reduced.
In-line filtration is able to avert severe complications in critically ill patients. The overall complication rate during the PICU stay among the filter group was significantly reduced. In-line filtration was effective in reducing the occurrence of SIRS. We therefore conclude that in-line filtration improves the safety of intensive care therapy and represents a preventive strategy that results in a significant reduction of the length of stay in the PICU and duration of mechanical ventilation (ClinicalTrials.gov number: NCT00209768).
KeywordsIn-line filtration SIRS Intensive care Particle Inflammation Children Complication
Acute respiratory distress syndrome
Centralized intravenous additive service
Pediatric intensive care unit
- PIM II
Pediatric Index of Mortality II
Systemic inflammatory response syndrome
The intravenous administration of fluids and drugs is essential in the management of critically ill patients. The contamination of infusion solutions by particles is a widely unknown and underestimated side effect of intravenous therapy [1, 2]. Particulate contamination is due to drug incompatibility reactions or their incomplete reconstitution during the preparation process . Various studies have demonstrated the contamination of infusion solutions with glass particles from opening glass ampoules, particles from rubber stoppers or conglomerates of the parenteral nutrition components [4, 5]. Particles have also been shown to be inherent to generic drug formulation . In an intensive care setting the particle burden may rise up to one million infused particles per day, increasing with the complexity and quantity of the administered infusions [6, 7]. Acknowledging the risks associated with such contamination, it is important to optimize infusion therapy in order to minimize medication errors and particle load. Therefore, standard operational procedures for the infusion set-up, databases for the prevention of drug incompatibilities and training of medical staff can be helpful tools to improve patient safety [3, 8, 9]. In addition, in-line filtration has been shown to almost completely prevent particulate infusion . Two intravenous fluid filters are currently in widespread use: 0.2-μm filters for crystalline solutions and 1.2-μm filters for lipid-containing admixtures. Positively charged 0.2-μm filters are able to retain particles, air and micro-organisms and endotoxins .
Post-mortem examinations of adults suffering from acute respiratory distress syndrome (ARDS)  and children with underlying disease  have revealed that infusion therapy can lead to particle-induced mechanical blockage of vessels and the generation of pulmonary foreign body granulomata. Particles harm the pulmonary endothelium either directly or through activation of complement, platelets and/or neutrophils, leading to the formation of occlusive microthrombi [7, 12]. In vitro studies with endothelial cells and macrophages have also demonstrated particle-induced modulation of immune response . The effect of particle infusions is aggravated in situations of disturbed microcirculation, such as ischemia and reperfusion injury: after cardioplegia, particles in cardioplegia solutions lead to impaired coronary artery flow, alterations in endothelium and increased leukocyte adhesion in humans as well as in animals . In an experimental rat liver transplantation model, particles in the preservation solution were found to disturb the microcirculation and aggravate post-ischemic inflammation . In a hamster skinfold chamber model, following ischemia, intravenous infusion of particle-contaminated solutions reduced functional capillary density by almost 50 %, compared to post-ischemic values before particle injection [14, 15]. However, filtration of solutions through 0.2-μm in-line filters prevented the loss of capillaries completely  and attenuated the inflammatory process [1, 13].
Previous clinical trials of in-line filtration primarily focused on the retention of micro-organisms, demonstrating a preventive effect on thrombophlebitis  without any impact on central venous catheter-related sepsis . To date, no clinical trial has taken account of the results of experimental studies on particle contamination. However, a clinical effect of particle retentive in-line filtration would be expected to be most evident in situations of altered microcirculatory homoeostasis or immune response. Intensive care patients who experienced trauma, major surgery or systemic inflammation suffer from conditions that affect the microcirculation of tissues and vital organs .
This single-centre, prospective, randomized, controlled trial in critically ill children was conducted to evaluate the impact of in-line filtration on severe complications, such as systemic inflammatory response syndrome (SIRS), sepsis, thrombosis and different organ failure.
Materials and methods
This was a single-centre, prospective, randomized, controlled trial conducted between February 2005 and September 2008 in an interdisciplinary pediatric intensive care unit (PICU) of a tertiary German hospital. Approval of the study protocol was obtained from the local ethics committee. Funding was primarily provided by a research grant of Hannover Medical School and partially by an unrestricted grant from Pall, Dreieich, Germany and B. Braun Melsungen, Germany. Both companies supplied in-line filters. An open label study design characterized by visible in-line filters was necessary to ensure maximum safety in drug administration and to allow nurses to monitor the in-line filters for imminent blockage. Any adverse events relating to in-line filtration were recorded.
Patient enrolment and randomization
Infusion management and in-line filtration
Prior to this study, the infusion regimen was standardized for patients in both groups. A computer database containing information on compatibilities of drugs (KiK 3.0; ODATA, Rastede, Germany) was used to prevent incompatibilities . Medications were prepared according to the manufacturer’s instructions. Parenteral nutrition and certain drugs (special antibiotics, antiviral drugs, antimycotics, chemotherapy) were supplied by a centralized intravenous additive service (CIVA) to guarantee chemical stability and aseptic standards. The filter group received in-line filtration throughout the period of infusion therapy, and all fluids except blood, plasma proteins or fresh frozen plasma were administered via in-line filters. The appropriate filters (1.2-μm pore size [Intrapur Lipid/Intrapur Neonat Lipid; B. Braun Corporation, Melsungen, Germany] for infusion of lipid-containing admixtures; 0.2-μm pore size positively charged filters [ELD96LLCE/NEO96E; Pall, Dreieich, Germany] for aqueous solutions) were arranged in the lumen of each central and peripheral venous catheter (Fig. 1 of the Electronic Supplementary Material [ESM]). Nurses checked the in-line filters at least every hour for imminent blockage. Filters were replaced after 24 h (Intrapur Lipid/Intrapur Neonat Lipid) or 72 h (ELD96LLCE/NEO96E) of regular use or in case of blockage. In both groups, the administration sets for lipid-containing admixtures were changed every 24 h, others every 72 h as recommended by the Robert Koch Institute . Nurses and physicians were extensively trained prior to the study to ensure correct and safe in-line filter handling and a competent infusion management.
Given the low mortality in PICU, we chose reduction in the overall complication rate of major events as the primary endpoint. Major events included the incidence of SIRS, sepsis, thrombosis, acute liver failure, ARDS and acute renal and circulatory failure. The occurrence of at least one of these major events during the PICU stay accounted for one complication per patient in the calculation of the overall complication rate. Complications prior to PICU stay and on admission were not taken into account. Sepsis and SIRS were defined according to International Pediatric Sepsis Consensus Conference (IPSCC) 2005 [20, 21]. All other endpoints were defined according to accepted pediatric consensus recommendations [22, 23, 24, 25]. Further information is provided in the ESM. Secondary endpoints were reduction of length of stay on PICU and overall hospital stay.
Data collection and statistical analysis
Detailed information on data collection is provided in the ESM. Statistical analysis was performed on an intention-to-treat basis. In a preceding survey a complication rate of approximately 40 % was determined. The study was designed using a chi-square test for equal proportions and with 80 % power to detect a reduction from 40 to 30 % in the complication rate for the filter group. At a significance level of 0.05, based on the previously determined complication rate, this design required 376 subjects to be enrolled in each group. The study design included an interim analysis after recruitment of 50 % of the targeted enrolment to evaluate any possible adverse effects of in-line filtration. Statistical analysis was performed with the use of Predictive Analysis Software for Windows, ver. 18 (SPSS, Chicago, IL). Further information on the statistics is provided in the ESM.
Baseline characteristics of patients
Control group (n = 406)
Filter group (n = 401)
5.58 ± 5.59
6.07 ± 6.01
21.8 ± 20.1
23.0 ± 20.7
Pediatric Index of Mortality II (PIM II)
4.15 ± 8.76
3.42 ± 9.14
Disease category on admission (n)
For the reporting of adverse events, see ESM.
Control group (n = 406)
Filter group (n = 401)
95 % Confidence interval
Primary objectives (n)
Adjusted to PIM II
Adjusted to PIM II
Acute renal failure
Acute liver failure
Length of stay (days)b
Duration of mechanical ventilation (h)b
The length of stay of the filter group in the PICU was significantly reduced (23 %) compared to that of the control group (2.98 [95 % confidence interval 2.33−3.64] vs. 3.89 [2.97−4.82] days, respectively; P = 0.025) (Table 2). Although statistically not significant, overall hospital stay was reduced by 1 day in the filter group verus the control group (15.0 [13.4−16.6] vs. 16.0 [14.2−17.8] days, respectively; P = 0.19). The duration of mechanical ventilation was significantly reduced in the filter group versus that in the control group (11.0 [7.1−14.9] vs. 14.0 [5.6−22.4] h, respectively; P = 0.028). There was a statistical trend towards lower mortality in the filter group (6.7 [n = 27] vs. 4.0 % [n = 16], control vs. filter group; P = 0.09).
In this prospective clinical trial, in-line filtration of infusion solutions led to a significant decrease in major complications in critically ill pediatric patients. A significant reduction from 40.9 to 30.9 % in the overall complication rate of severe events (SIRS, sepsis, circulatory failure, ARDS, thrombosis, acute renal failure and acute liver failure) was demonstrated for the filter group. Although sample size and power were not calculated to detect reduction in single complications, the incidence of SIRS was significantly reduced. Additionally, all other primary objectives had a lower incidence in the filter group, but without statistical significance. In-line filtration significantly decreased both the duration of mechanical ventilation and the length of stay in the PICU. Considering the relatively low incidence of mortality in the PICU compared to an adult cohort, the statistical trend towards a reduction in mortality rate in the filter group is especially noteworthy.
This is the first clinical trial involving more than 800 PICU patients to reveal a significant benefit of in-line filtration on serious complications. However, our results seem to be in contrast to a recent Cochrane analysis including 704 neonates and preterm infants of four different trials, which demonstrated no favorable effect of in-line filtration on morbidity or mortality . Two of the four studies included in the Cochrane analysis focused on thrombophlebitis and cannula patency, and only one study reported on endpoints comparable to those chosen for our study . In this latter study, the authors demonstrated a significant reduction in the incidence of typical neonatal complications for in-line filtration, which is consistent with our data. In the fourth study of van Hoogen et al. , the primary objective was a reduction of sepsis, which, similar to our results, could not be significantly reduced by the use of in-line filters.
In our trial, the majority of severe complications occurred within the first 3 days following admission to the PICU, with 40.9 % of patients in the control group experiencing at least one major complication during this period. Implementation of in-line filters as early as the time of admission prevented severe complications and generated a persistent reduction in the complication rate (Fig. 3). These protective effects of in-line filtration may be explained by the preservation of microcirculation in all organs. The maintenance of microcirculation is particularly crucial in critically ill patients to prevent organ failure, whereas several frequently coexisting pathologies, such as inflammation and trauma, compromise the microcirculation . The early restoration of microcirculation shown for inflammatory syndromes such as sepsis is associated with reduced morbidity and lower organ failure score . In the pathological state of an already reduced microcirculatory perfusion, infused particles may cause additional impairment, leading to a loss of capillary density, as demonstrated in recent preclinical studies [14, 15]. Thus, the threshold of organ recovery is exceeded, and clinical signs of organ dysfunction or failure arise.
Based on our data, in-line filtration is effective in preventing SIRS in intensive care patients. SIRS was first used in adults to describe a nonspecific systemic inflammatory process in the absence of infection [21, 30]. SIRS criteria for adults were incorporated and modified with age-specific norms for children by the IPSCC . SIRS in children is manifested by the presence of at least two of four criteria, one of which must be an abnormal temperature or leukocyte count: (1) hyperthermia or hypothermia, (2) tachycardia, (3) tachy- or bradypnoea or (4) leukocytosis or leukopenia . The occurrence of SIRS predisposes patients to organ failure and frequently determines clinical outcomes . Mortality and morbidity of severe non-infectious SIRS does not differ from those of severe sepsis, as shown in a multicentre trial involving 3,500 patients admitted to adult intensive care units . The incidence of SIRS has been found to be higher than that of sepsis in both adult  and pediatric  intensive care patients. SIRS on admission has been shown to determine mortality and the length of stay of critically ill trauma patients . In a prospective survey of 3,708 hospitalized adults, patients with SIRS had a 26 % chance of developing sepsis . In another study, the more SIRS criteria fulfilled by a patient, the more likely that patient was to develop ARDS, disseminated intravascular coagulation or acute renal failure or even die of SIRS . Pathophysiologically, mechanisms of SIRS and sepsis are similar, but the management of SIRS is typically non-specific . Only supportive intensive care management, including fluid resuscitation, mechanical ventilation, inotropic support combined with the treatment of the initiating insult, might be beneficial . In this setting of limited therapeutic options, in-line filtration represents a potent strategy to prevent SIRS and associated co-morbidities.
Although sample size and power were not calculated to detect a reduction in ARDS, a statistical trend towards a reduction was evident for the in-line filtration group. Consistent with this, a significantly decreased duration of mechanical ventilation was found, supporting the hypothesis that the lung as the physiological filter of intravenous infusion is most vulnerable to particulate damage. The mechanical obstruction of capillaries in the lung and other organs by infused particles has been demonstrated in both clinical and experimental studies [7, 12, 14]; such obstruction may impair coagulation through the consumption of pro- and anticoagulative proteins and platelets . In vitro incubation with endothelial cells and macrophages demonstrated immunomodulating effects of particles . Similar mechanisms have been shown for deposits of inhaled particles in the lung, which initiate a local inflammatory process that expands into a systemic inflammatory response with an increase in circulating inflammatory factors and activation of immune cells .
We adhered to a standardized infusion regimen, prevented incompatibilities and had access to CIVA and we still achieved a further reduction in the incidence of complications by using in-line filtration. The medical staff received extensive training in the handling of in-line filters prior to the initiation of our study, followed by continuous support by the authors during the study period. This practice enabled the implementation of filters without any relevant problems and demonstrated the practicability of the standard infusion arrangement and filter set-up. Thus, during the entire trial period there were only an irrelevant number of blocked filters and no serious adverse events of in-line filtration were noted. To ensure maximum patient safety in drug administration, filter membranes were visible in order to control for imminent blockage or defects. This compulsory open label study design may be a limitation and a potential source of bias.
Consistent with results reported in other trials , we have demonstrated a correlation between SIRS and length of stay in the PICU. Based on our data, one would expect that applying an in-line filter would also reduce diagnostic and therapeutic resources and increase turnover on the PICU. However, economic aspects of in-line filtration were not an endpoint in our study. As already shown by a meta-analysis , additional costs of in-line filters are compensated for by the reduced consumption of intravenous administration sets [27, 28] and decreased time for changing infusion sets . In both of these studies [27, 28], intravenous sets in the control group were changed daily, while in the in-line filter group (interventional group) the changing times were extended up to 96 h. Due to different changing intervals in our study—72 h in both the control and filter group—these calculations are only partially applicable. Costs for disposables in our infusion set-up amounts to an additional expense of approximately 30 Euros for filters plus about 2 min extra work for equipment set-up. By lengthening the changing time from 72 to 96 h, some of the additional cost could be amortized. In summary, diagnostic and therapeutic resource-saving effects are likely to outweigh the additional costs for the filters and, consequently, positive economic effects can be expected.
In conclusion, the results of our trial demonstrate the safety and efficacy of in-line filtration in preventing major complications in patients admitted to the PICU. The overall complication rate and the incidence of SIRS were reduced among those patients with in-line filtration, indicating that filtration is a preventive strategy that can result in decreased morbidity of critically ill patients, reduced duration of mechanical ventilation and reduced length of stay on the PICU. Furthermore, in-line filtration was shown to improve the safety of infusion therapy. Further research is necessary to fully elucidate the pathophysiological mechanisms underlying our clinical findings.
We thank the participants and their guardians who volunteered for the study; the staff of the Department of Pediatric Cardiology and Intensive Care Medicine, Hannover Medical School, Hannover, Germany, for their excellent support—in particular M. Abura, J. Wessels, M. Becker, C. Bormann, V. Dannenberg, F. Dziuba, V. Quartier, D. Yilmaz; A. Vogel, H. Toensfeuerborn and J. Wuttke. We specially thank all members of the nursing staff for their excellent work during the trial. We thank F. Schröder, pharmacist at Klinikum Bremen, and W. Orth for their support and for providing the KIK software.
Conflicts of interest
This study was funded by a research grant from Hannover Medical School and partially by an unrestricted grant from Pall Corporation, Dreieich, Germany and B. Braun Corporation, Melsungen, Germany. Both companies supplied in-line filters during the study period. Dr. Jack, Dr. Sasse and Dr. Boehne report having been paid lecture and travel fees from the Pall and B. Braun Corporation. Dr. Brent received travel fees from the Pall Corporation. No other potential conflict of interest relevant to this article was reported. The sponsors had no influence on the study design, patient enrolment, data collection, analyses, data interpretation or preparation of the manuscript. The trial was independently conducted by the investigators. Data were fully accessible to and interpreted by all authors, who attest to its accuracy and completeness. The authors had final responsibility for the decision to submit for publication.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.