Head-up tilt and hyperventilation produce similar changes in cerebral oxygenation and blood volume: an observational comparison study using frequency-domain near-infrared spectroscopy
- First Online:
- 812 Downloads
During anesthesia, maneuvers which cause the least disturbance of cerebral oxygenation with the greatest decrease in intracranial pressure would be most beneficial to patients with intracranial hypertension. Both head-up tilt (HUT) and hyperventilation are used to decrease brain bulk, and both may be associated with decreases in cerebral oxygenation. In this observational study, our null hypothesis was that the impact of HUT and hyperventilation on cerebral tissue oxygen saturation (SctO2) and cerebral blood volume (CBV) are comparable.
Surgical patients without neurological disease were anesthetized with propofol-remifentanil. Before the start of surgery, frequency-domain near-infrared spectroscopy was used to measure SctO2 and CBV at the supine position, at the 30° head-up and head-down positions, as well as during hypoventilation and hyperventilation.
Thirty-three patients were studied. Both HUT and hyperventilation induced small decreases in SctO2 [3.5 (2.6)%; P < 0.001 and 3.0 (1.8)%; P < 0.001, respectively] and in CBV [0.05 (0.07) mL·100 g−1; P < 0.001 and 0.06 (0.05) mL·100 g−1; P < 0.001, respectively]. There were no differences between HUT to 30° and hyperventilation to an end-tidal carbon dioxide (ETCO2) of 25 mmHg (from 45 mmHg) in both SctO2 (P = 0.3) and CBV (P = 0.4).
The small but statistically significant decreases in both SctO2 and CBV caused by HUT and hyperventilation are comparable. There was no correlation between the decreases in SctO2 and CBV and the decreases in blood pressure and cardiac output during head-up and head-down tilts. However, the decreases in both SctO2 and CBV correlate with the decreases in ETCO2 during ventilation adjustment.
Un basculement de la tête vers le haut et l’hyperventilation entraînent des modifications semblabes de l’oxygénation cérébrale et du volume sanguin: étude comparative observationnelle utilisant la spectroscopie infrarouge proche
Au cours de l’anesthésie, des manœuvres provoquant le minimum de perturbation de l’oxygénation cérébrale et la plus grande diminution possible de la pression intracrânienne auraient de grands avantages pour les patients souffrant d’hypertension intracrânienne. Le basculement de la tête vers le haut (HUT) et l’hyperventilation sont toutes deux utilisées pour diminuer le volume cérébral et les deux peuvent être associées à des baisses de l’oxygénation cérébrale. Dans cette étude observationnelle, notre hypothèse nulle était que l’impact du HUT sur la saturation en oxygène du tissu cérébral (SctO2) et sur le volume sanguin cérébral (CBV) serait comparable à celui de l’hyperventilation.
Des patients de chirurgie sans maladie neurologique ont été anesthésiés au propofol et au rémifentanil. Avant le début de l’intervention chirurgicale, une spectroscopie infrarouge proche a permis de mesurer la SctO2 et le CBV en decubitus dorsal, avec la tête basculée à 30° vers le haut et vers le bas, ainsi qu’en hypoventilation et en hyperventilation.
Trente-trois patients ont été étudiés. La HUT et l’hyperventilation ont induit de petites diminutions de la SctO2 (respectivement: 3,5 [2,6] %; P < 0,001 et 3,0 [1,8] %; P < 0,001) et du CBV (respectivement: 0,05 [0,07] mL·100 g−1; P < 0,001 et 0,06 [0,05] mL·100 g−1; P < 0,001). Il n’y a pas eu de différences entre la HUT à 30° et l’hyperventilation avec une concentration de dioxyde de carbone en fin de respiration (ETCO2) de 25 mmHg (depuis 45 mmHg) pour ce qui concernait à la fois la SctO2 (P = 0,3) et le CBV (P = 0,4).
Les diminutions limitées, mais statistiquement significatives, de la SctO2 et du CBV provoquées par la HUT et l’hyperventilation sont comparables. Aucune corrélation n’a été trouvée entre les baisses de la SctO2 et du CBV avec les diminutions de tension artérielle et de débit cardiaque au cours du basculement de la tête vers le haut ou vers le bas. Cependant, la baisse de la SctO2 et celle du CBV sont corrélées à une diminution de l’ETCO2 au cours du réglage de la ventilation.
In patients with increased brain bulk, head-up tilt (HUT) and hyperventilation are often instituted to decrease intracranial pressure (ICP) or to improve operating conditions. However, these widely applied maneuvers can also have a negative impact on cerebral perfusion and oxygenation, i.e., HUT can severely compromise cerebral perfusion pressure, and hyperventilation can cause profound cerebral vasoconstriction.1 There has been debate about the relative effects of HUT in maintaining cerebral blood flow (CBF) and decreasing ICP.2-5 In contrast, the current point of view is that hyperventilation in head-injured patients can produce more harm than benefit, and it should be strictly limited to the emergent management of life-threatening intracranial hypertension pending definitive measures or to facilitate intraoperative surgery.6 Recent advances in near-infrared spectroscopy (NIRS), such as frequency-domain (FD) and time-domain approaches, allow for absolute quantification of cerebral tissue oxy- and deoxyhemoglobin.7,8 These newer quantitative NIRS technologies can assess not only cerebral tissue oxygen saturation (SctO2) but also cerebral blood volume (CBV) based on total hemoglobin concentration (THC), the sum of oxy- and deoxyhemoglobin.8,9 In this observational study, our null hypothesis was that HUT and hyperventilation cause similar changes in SctO2 (an estimate of cerebral perfusion and oxygenation) and CBV (a contributor of intracranial mass and ICP). Our specific aim was to use FD-NIRS to compare the changes in SctO2 and CBV caused by HUT and hyperventilation in propofol-remifentanil anesthetized non-neurosurgical patients.
After Institutional Research Board approval (HS#: 2010-7521; approved on: May 21st, 2010; Contact: Research Administration, 5171 California, Suite 150, Irvine, CA 92697) and informed verbal and written consent, patients scheduled for elective non-neurosurgical procedures at University of California Irvine Medical Center were recruited for this study. Exclusion criteria were: age ≤ 18 yr old, cerebrovascular disease, symptomatic cardiovascular disease, poorly controlled hypertension (systolic blood pressure ≥ 160 mmHg), and poorly controlled diabetes mellitus (blood glucose ≥ 200 mg·dL−1). The data presented here from our FD-NIRS study were acquired from the same patients we recruited to study the effects of vasopressor treatment. The result regarding the impact of vasopressor administration on SctO2 and the result regarding the comparison of cardiac output (CO) measured by esophageal Doppler and Vigileo FloTrac have been previously published.10,11 As each study has a unique hypothesis and paradigm, they have been reported separately. We took care to ensure vasopressor-induced hemodynamic changes returned to baseline values for at least five minutes before whole body tilt and ventilation adjustment.
Following the patient’s arrival in the operating room, a radial intra-arterial catheter, a bispectral index (BIS) monitor, and two FD-NIRS probes (left and right forehead) were placed in addition to the other routine monitors. Following anesthesia induction with fentanyl 1.5-2 μg·kg−1 and propofol 2-3 mg·kg−1, all patients’ tracheas were intubated and maintained with total intravenous anesthesia using propofol 75-150 μg·kg−1·min−1 and remifentanil 0.3-0.5 μg·kg−1·min−1 to target a BIS of 30. Volume-controlled ventilation was used with a tidal volume of 8-10 mL·kg−1 and a respiratory rate of 8-10 breaths·min−1 to target an end-tidal carbon dioxide (ETCO2) of 35 mmHg. The inspired oxygen was 50%. Muscle relaxation was maintained with cisatracurium. A 30° HUT (reverse-Trendelenburg position) and a 30° head-down tilt (Trendelenburg position) were performed and compared with the supine position (0°). The order of head-up and head-down tilts was randomized. Study measurements were recorded when mean arterial pressure (MAP) decreased to the lowest value with HUT and when MAP increased to the highest value with head-down tilt. After completion of the body tilt component and once tilt-induced hemodynamic changes receded, ventilation adjustment was conducted with the end point of hyperventilation at ETCO2 of 25 mmHg and the end point of mild hypoventilation at ETCO2 of 45 mmHg. The order of hyperventilation and mild hypoventilation was also randomized. Study measurements were recorded once the end points of hyperventilation and mild hypoventilation were reached. All measurements were obtained before the start of surgery.
Cerebral blood volume is in mL·100 g−1; THC is in μMol; MWHb is the molecular weight of hemoglobin (64,458 g·Mol−1); HGB is systemic blood hemoglobin concentration (g·dL−1); Dbt is brain tissue density (1.0335 g·mL−1); and CLVHR = 0.69 is the cerebral to large vessel hematocrit ratio.
Mean arterial pressure was monitored at the external ear canal level via a radial intra-arterial catheter system, Vigileo FloTrac (Edwards Lifesciences, Irvine, CA, USA). Cardiac output was monitored by both esophageal Doppler (CardioQ, Deltex Medical, UK) (COED) and the third-generation Vigileo FloTrac (COFT). End-tidal carbon dioxide was determined by the gas analyzer built into the anesthesia machine (Aisys, GE Healthcare, Madison, WI, USA). Oxygen saturation by pulse oximetry was determined by pulse oximeter (LNOP Adt, Masimo Corp, Irvine, CA, USA). The depth of anesthesia was monitored via the BIS monitor (S/5TM M-BIS, GE Healthcare, Madison, WI, USA).
Sample-size determination for evaluating the impact of vasopressor treatment on SctO2 was reported previously.10 Since the observation we report in this article is a secondary outcome of the experiment, we did not carry out a separate sample-size determination. Data are expressed as mean standard deviation (SD). Ninety-five percent confidence intervals are reported. The P values reported for comparisons between head-up vs supine, head-down vs supine, and hyperventilation vs hypoventilation were compared by paired Student’s t test. The P values reported for comparisons between HUT and hyperventilation were also compared by paired Student’s t test. The P values reported for Pearson’s correlations were calculated by Student’s t test using linear regression analysis. The P values < 0.001 (0.05/45 = 0.001) were regarded as significant, corresponding to the Bonferroni correction to control the familywise error rate at 0.05 for the 45 tests (comparisons) performed.
Thirty-three patients [22 males, 11 females, aged 59 (13) yr, height 173 (9) cm, and weight 77 (13) kg] were recruited for this study. Among the 33 patients, three patients were categorized as American Society of Anesthesiologists’ (ASA) physical status I, 22 patients were categorized as ASA II, and eight were categorized as ASA III. Due to their concern about low blood pressure, the attending anesthesiologists withdrew five patients from the tilt component of the study, and three from the ventilation component. The data from 28 patients were entered into the tilt analysis database, and data from 30 patients were entered into the ventilation analysis.
Physiological measurements at supine, head-up, and head-down positions (n = 28)
−18.1 to −10.6
24.2 to 34.0
−1.0 to −0.3
−0.05 to 0.7
−1.0 to −0.2
0.9 to 1.6
−0.2 to 7.1
−4.8 to 0.3
−1.1 to 0.2
−0.5 to 0.4
−3.2 to −1.2
−0.9 to 1.0
−0.4 to 2.7
−4.4 to 2.2
−4.4 to −2.5
0.7 to 2.4
−1.2 to −0.4
1.1 to 2.2
CBV (mL·100 g−1)
−0.08 to −0.02
0.07 to 0.1
Physiological measurements during hypoventilation and hyperventilation (n = 30)
−3.6 to 1.5
−0.6 to 0.2
−0.6 to 0.08
1.1 to 5.3
0.2 to 0.7
−20.1 to −16.9
−4.4 to 0.3
−3.6 to −2.3
−1.2 to −0.6
CBV (mL·100 g−1)
−0.08 to −0.04
In the 27 patients who received both interventions, the decreases in both SctO2 and CBV showed no differences between HUT to 30° and hyperventilation to an ETCO2 of 25 mmHg (from 45 mmHg) (P = 0.3 and P = 0.4, respectively).
The unique aspect of this study is that it provides a direct comparison of changes in cerebral oxygenation and cerebral blood volume in the same patients due to hyperventilation and HUT. The major findings from this study using FD-NIRS in healthy surgical patients were that both HUT to 30° and hyperventilation to an ETCO2 of 25 mmHg (from 45 mmHg) caused small but significant decreases in SctO2 and CBV, and the decreases in both SctO2 and CBV were not significantly different between the two conditions. We also found that changes in SctO2 and changes in CBV had no correlation with changes in MAP and CO during head-up and head-down tilts. However, changes in both SctO2 and CBV correlated well with changes in ETCO2 during hyperventilation.
Our study based on 30° HUT in propofol-remifentanil anesthetized patients showed a small decrease in both SctO2 [3.5 (2.6)%] and CBV [0.05 (0.07) mL·100 g−1]. In awake healthy subjects, Hunt et al. found a decrease in SctO2 of 2.6 (3.2)% and no change in CBV with 60° HUT using a NIRO 300 spectrophotometer,14 and Suzuki et al. found a decrease in SctO2 of 1.1 (1.0)% with 70° HUT using a PSA-III NIRS instrument.15 In propofol-anesthetized healthy surgical patients, Lovell et al. found a decrease in THC (the measurement used to calculate CBV) of 0.70 (0.99) μMol with 18° HUT using a NIRO 500 spectrophotometer; however, changes in SctO2 were not mentioned.16 It is noteworthy that none of the above studies was based on FD-NIRS technology. We also observed a small decrease in both SctO2 [3.0 (1.8)%] and CBV [0.06 (0.05) mL·100 g−1] induced by hyperventilation. In propofol-anesthetized rabbits, Cenic et al. reported no change in both CBF and CBV based on contrast-enhanced computed tomography measurements when arterial blood carbon dioxide partial pressure (PaCO2) was reduced from 41 to 27 mmHg via hyperventilation.17 In a subsequent study, the same group confirmed that CBF and CBV reactivity to hyperventilation is absent in propofol but present in rabbits anesthetized with isoflurane.18 Thus, although our findings are consistent with others in the literature, they may pertain only to patients anesthetized with propofol and remifentanil, and the results should not be extrapolated to patients receiving inhaled anesthetics.
Both HUT and hyperventilation are common interventions in neurosurgical patients to decrease brain bulk.1,19 The intervention which causes the least decrease in SctO2 and the greatest decrease in CBV would likely be more beneficial to the patient. In this FD-NIRS comparison study, we demonstrated that the decreases in both SctO2 and CBV caused by 30° HUT did not differ significantly with those caused by hyperventilation to an ETCO2 of 25 mmHg (from 45 mmHg). However, before extrapolating these findings to patients with intracranial pathology, a number of important factors should be considered. Our study was conducted in healthy patients without intracranial disorders. Although the impact of HUT and hyperventilation on CBV is similar, their effects on ICP and brain bulk were not directly compared. In addition to CBV, cerebrospinal fluid (CSF) is another key contributor to intracranial volume and ICP. In addition to the decrease caused by CBV reduction, hydrostatic displacement of CSF (from the cranial cavity to the spinal subarachnoid space) induced by HUT—a property hyperventilation does not possess—might also help to decrease ICP.20
There are several methodological limitations to be considered. First, estimation of PaCO2 by ETCO2 has its limitations even though there are data that ETCO2 is a reliable estimate of PaCO2 and the change in ETCO2 strongly approximates the change in PaCO2.24 For example, the use of the HUT position may significantly change the pulmonary blood flow, increasing the amount of zone 1 ventilation and causing significant dead space ventilation and an enlargement in the gradient between ETCO2 and PaCO2. Without knowing the PaCO2, the extent of hyperventilation is not known. Second, this study was conducted before surgical incision in order to avoid the impact of surgical stimulation on cerebral measurements. It was also done in non-neurosurgical patients with normal ICP in order to acquire “baseline” or normal brain information. Therefore, the clinical significance of the decreases in SctO2 and CBV caused by HUT and hyperventilation was not directly addressed by this study.
The specific FD-NIRS technology used for this study provides a better estimate of absolute measurements than the commonly employed NIRS instruments based on continuous wave (CW) technology which provide only relative measurements.25 Near infrared photons (~650-1000 nm) penetrate deeply (several centimetres) into tissues. At the wavelengths used in this study, the dominant absorbers (or chromophores) in tissue are oxy- and deoxyhemoglobin. However, tissues also strongly scatter NIR light, and it is this dominant scattering of the signal that makes it difficult to measure tissue’s hemoglobin absorbance accurately.26 Frequency-domain near-infrared spectroscopy separates the contributions of absorption and scattering to the detected NIR signals, while CW-NIRS does not.25 This may be important because studies have shown large intersubject variations in brain scattering based on measurements of optical path length in neonates, children, and adults.27,28 For this reason, changes in hemoglobin saturation reported by CW-NIRS may be affected by factors such as the intersubject variation in tissue scattering. In contrast, FD-NIRS used in this study avoids this confounding factor by direct and continuous measurement of light scattering in the tissue.7 Therefore, FD-NIRS is regarded as a quantitative method while CW-NIRS is considered a trend monitor only. An additional confounding factor is the contamination of the NIRS signal by extra-cerebral layers. Multiple studies show that the optode spacing distance determines the ability of NIRS to “see through” scalp and skull.12,13,29,30 Studies using FD-NIRS in humans and with phantoms have found that extra-cerebral contamination is negligible when the source-detector spacing is larger than 2 cm.12,13 Evidence also shows that NIRS measurements are consistent with those made by functional magnetic resonance imaging.31 In summary, FD-NIRS is not only a quantitative technology by its ability to separate absorption and scattering, but it is also a technology which is likely less affected by the extra-cerebral tissue layers.
In conclusion, the significant but small decreases in both SctO2 and CBV caused by HUT to 30° and hyperventilation to an ETCO2 of 25 mmHg (from 45 mmHg) in normal healthy individuals are comparable. Changes in SctO2 and CBV do not correlate with changes in MAP and CO during HUT; however, changes in both SctO2 and CBV correlate with changes in ETCO2 during hyperventilation.
We acknowledge the generous loan of the Oxiplex TX oximeter from ISS, Inc. We also thank Christine Lee BS for her help in data analysis and Nam P. Tran BS for her help in data acquisition.
This study is supported by the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), through the following programs: the Institute for Clinical & Translational Science (ICTS) grant UL1 RR031985 (BJT); the Laser Microbeam and Medical Program (LAMMP), a NIH BRTP resource (P41-RR01192) (BJT); and the Laboratory for Fluorescence Dynamics (LFD) grant P41 RR03155 (WWM). It is also supported by the Department of Anesthesiology & Perioperative Care, University of California Irvine Medical Center.
Conflicts of interest
The authors (A.E.C., B.J.T., and W.W.M.) consult for ISS™, Inc.