Cerebral Autoregulation Is Impaired During Deep Hypothermia—A Porcine Multimodal Neuromonitoring Study

Abstract

Effects of brain temperature modulation on cerebral hemodynamics are unclear. We aimed at investigating changes of dynamic cerebral autoregulation (AR) indices during induction of deep hypothermia (HT) in a porcine model mimicking the clinical scenario of accidental HT. Thirteen pigs were surface-cooled to a core temperature of 28°C. High-frequency monitoring included brain temperature, mean arterial blood pressure (MAP), intracranial pressure (ICP), brain tissue oxygen tension (P_btO₂), and regional oxygen saturation (rSO₂) assessed by near-infrared spectroscopy to calculate AR-indices (pressure reactivity index [PRx], oxygen reactivity index [ORx], and cerebral oximetry index [COx]). Brain temperature decreased from 39.3°C ± 0.8°C to 28.8°C ± 1.0°C within a median 160 minutes (interquartile range 146–191 minutes), reflecting a rapid induction of deep HT (−4°C/h). MAP and cerebral perfusion pressure (CPP) remained stable until a brain temperature of 35°C (69 ± 8 mmHg, 53 ± 7 mmHg) and decreased to 58 ± 17 mmHg and 40 ± 17 mmHg at 28°C (p = 0.031 and p = 0.015). Despite the decrease in MAP and CPP, brain oxygenation increased (P_btO_2: +5 mmHg, p = 0.037; rSO₂: +7.3%, p = 0.029). There was no change in ICP during HT induction. Baseline AR-indices reflected normal cerebral AR and did not change until a brain temperature of 34°C (ORx), 33°C (PRx), and 30°C (COx). At lower temperature, AR-indices increased (PRx: p < 0.001, ORx: p = 0.02, COx: p = 0.03), reflecting impaired cerebral AR. Cerebrovascular reactivity is impaired at lower brain temperature levels. Although these temperatures are usually not targeted in clinical routine, this should be kept in mind when treating patients with accidental deep HT.

Introduction

Deep hypothermia (HT) (<28°C) occurs accidentally in avalanche victims and occasionally in out-of-hospital comatose patients (Paal et al., 2016). In the intra-hospital setting, induced deep HT is still used in cardiovascular surgery (Keeling et al., 2018). Hemodynamic management of these patients, especially in the pre-hospital setting, can be challenging, aiming at ensuring adequate organ perfusion. There is little known about the optimal blood pressure that should be targeted to ensure adequate substrate delivery to the brain. As cerebral autoregulation (AR) may be impaired in hypothermic patients (Joshi et al., 2010; Koizumi et al., 2016), energy delivery to the brain may be directly dependent on changes in mean arterial blood pressure (MAP). At the same time, brain metabolism and therefore energy consumption may be decreased during deep HT (Frietsch et al., 2000; Erecinska et al., 2003).

Monitoring cerebral AR is an integral part in the critical care management of hospitalized brain injured patients (Johnson et al., 2016; Rivera-Lara et al., 2017; Gaasch et al., 2018), and a loss of cerebral AR has been linked to secondary brain injury and poor outcome (Donnelly et al., 2016; Gaasch et al., 2018). Most cerebral AR indices are still assessed invasively, including the pressure reactivity index (PRx; moving correlation coefficient of MAP and intracranial pressure [ICP]), the oxygen reactivity index (ORx; moving correlation coefficient between cerebral perfusion pressure [CPP] and brain tissue oxygen tension [P_btO₂]), and the cerebral oximetry index (COx; regional oxygen saturation [rSO₂] and MAP) (Czosnyka et al., 1997; Jaeger et al., 2006; Rivera-Lara et al., 2017). Based on the lack of feasibility to assess cerebral AR invasively in hypothermic patients in the prehospital setting or perioperatively, we studied the complex interaction of pressure-dependent energy delivery to the brain in the setting of induced deep HT in a porcine model.

Materials and Methods

Detailed information on animal preparation and the protocol of the main experiment are reported elsewhere (Putzer et al., 2016). A summary of important information is given here. This study was approved by the Institutional Animal Care and Use Committee of the University of Innsbruck and the Austrian Ministry of Science, Research and Economy (Protocol number: BMWF-66.011/0027-II/3b/2013). It was conducted at our experimental research unit in compliance with EU regulations for animal experimentation (Directive 2010/63/EU of the European Parliament and the European Council) and reporting is in accordance with current ARRIVE guidelines.

Animal preparation and anesthesia

This study was performed in 13 domestic pigs weighing between 35 and 50 kg. The pigs were fasted overnight with free access to water. Premedication was done with azaperone (4 mg/kg IM; Jansen, Vienna, Austria) and atropine (0.01 mg/kg IM). Anesthesia was induced with ketamine (25 mg/kg IM) and propofol (1 mg/kg IV). After endotracheal intubation during spontaneous ventilation, the pigs were ventilated volume-controlled (Evita 2; Draeger, Lübeck, Germany) with 21% inspiratory oxygen and a tidal volume of 6–8 mL/kg body weight to maintain normocapnia (35–45 mmHg). During HT, we used the α-Stat regime to adopt ventilator settings and report blood gas values. No exogenous CO₂ was administered nor did we use profound hypoventilation. Anesthesia was maintained by using propofol [6–8 mg/(kg·h) IV] and repetitive injections of piritramide (15 mg IV; Janssen, Vienna, Austria); neuromuscular blockade was achieved with a continuous infusion of pancuronium [0.3 mg/(kg·h)]. Normovolemia was maintained by administering lactated Ringer's solution [10 mL/(kg·h) IV]. A standard lead II electrocardiogram (ECG) was used to monitor cardiac rhythm, and a pulsoximeter was placed on the tail. A 6.0 Fr saline-filled arterial catheter (Arrow, Reading, PA) was placed in the right femoral artery.

Neuromonitoring

A near-infrared spectroscopy (NIRS) optode (INVOS™ System; Somanetics, Inc., Troy, MI) was placed on the right forehead for rSO₂ measurements. A brain tissue oxygen catheter (LICOX; Integra NeuroSciences, Ratingen, Germany) and a combined ICP/Temperature probe (Neurovent-P Raumedic, Münchberg, Germany) were placed in the white matter of the corresponding region of the left hemisphere. Arterial catheters were attached to pressure transducers (1290A; Hewlett Packard, Böblingen, Germany) and calibrated at the level of the meatus acusticus for CPP calculations. Hemodynamic and respiratory variables were measured and analyzed by using an AS/3 Monitor (Datex-Ohmeda AS/3; GE Health-care, Buckinghamshire, UK).

Study protocol and cooling procedure

The study protocol mimicked the scenario of a severely hypothermic patient. After calibrating the LICOX probe according to the manufacturers' specifications, the experimental protocol was commenced with baseline measurements of hemodynamic, respiratory and cerebral oxygenation parameters and blood gases. The pigs were then surface-cooled with crushed ice until a core temperature of 28°C was reached. At this point, recordings for this analysis were stopped and the protocol continued as previously described (Putzer et al., 2016). After completion of the experimental protocol, the animals were euthanized.

Measurements

High-frequency measurements included MAP, ICP, end-tidal CO₂ (etCO₂), oxygen saturation (SpO₂) and regional tissue (rSO₂), and P_btO₂. CPP was calculated as the difference between MAP and ICP.

AR indices (PRx, ORx, and COx) were calculated as originally described (Czosnyka et al., 1997; Jaeger et al., 2006; Rivera-Lara et al., 2017): A moving Pearson correlation of 30 ten-second intervals between ICP and MAP for PRx and between rSO₂ and MAP for COx was calculated. For ORx, 120 one-minute intervals were used, to correlate P_btO₂ und CPP. For every moving window, a minimum of 50% of data points had to be available to calculate a value. Blood gases were taken at baseline and stepwise during the cooling period at every 2°C temperature drop (Table 1). Overall, brain temperature was used for analysis, if not mentioned otherwise.

Table 1.

Arterial Blood Gases at Different Temperature Levels

	Core temperature°C
	38	36	34	32	30	28
pH	7.54 ± 0.01	7.47 ± 0.04	7.53 ± 0.03	7.5 ± 0.06	7.57 ± 0.06	7.48 ± 0.14
pCO₂	36.69 ± 0.66	43.83 ± 5.17	37.23 ± 2.38	41.33 ± 7.16	35.3 ± 6.47	41.67 ± 12.14
pO₂	77.59 ± 3.84	83.13 ± 13.59	81.83 ± 32.10	98.27 ± 12.19	109.38 ± 9.41^a	118.77 ± 20.88^a
HCO₃-	31.51 ± 0.46	32.43 ± 1.46	31.113 ± 1.21	31.5 ± 1.98	31.92 ± 1.22	31.13 ± 0.99
sBE	8.61 ± 0.65	8.27 ± 1.27	8.09 ± 1.51	7.72 ± 1.31	9.30 ± 0.63	7.50 ± 1.39
Hb	8.15 ± 0.23	8.00 ± 0.52	8.08 ± 0.33	8.02 ± 0.38	8.12 ± 0.50	8.23 ± 0.35
Hct	24.18 ± 0.64	23.33 ± 1.15	24.13 ± 0.99	23.67 ± 0.82	24.20 ± 1.30	24.00 ± 1.00

Note that brain temperature highly correlated with core temperature. (r = 0.995, p < 0.001).

Significant compared with 38°C.

Statistical analysis

We used SPSS (v24; IBM) for statistical analysis. To correct for piglet weight and different duration of cooling time, we calculated the percentage of cooling time (down to 28°C) and used this variable as longitudinal reference. Statistical analysis for repeated measurements was done, using univariate and/or multivariate generalized estimating equation (GEE) models, with the correlation matrix best fitting the data and fitting selection criteria as described elsewhere (Gaasch et al., 2018). Linear mixed-effects analysis was performed in R (R Core Team, 2012) and lme4 (Bates et al., 2012), with predicted values obtained by using the ggpredict() function (Lüdecke, 2018). Important co-variates are specified in the Results section. For group differences, Student's t-test or Mann–Whitney U tests were used as appropriate. Multiple comparisons in GEE models were corrected by using the Least Significant Difference method; otherwise, no correction for multiple comparisons was used. A p-value <0.05 was considered statistically significant.

Results

Data from 13 pigs were obtained and included in the analysis. The pigs had a median weight of 38 kg (interquartile range [IQR]: 36–43) and were 110.5 cm long (IQR: 108.5–115.0). All pigs were male except for one.

Temperature

Mean baseline core temperature was 39.3°C ± 0.8°C and significantly decreased by 10.5°C (95% confidence interval [CI]: 9.7–11.5, p > 0.001) to 28.8°C ± 1.0°C after 160 minutes (IQR: 146–191 minutes). Brain and core temperature were highly correlated (r = 0.995, p < 0.001).

Neuromonitoring

ICP did not change over time (16 ± 0.9 mmHg at baseline and 17 ± 1.7 mmHg at 28°C). Baseline MAP and CPP were 69 ± 8 mmHg and 53 ± 7 mmHg, respectively, and remained constant until a temperature of 35°C was reached. Thereafter, MAP and CPP significantly decreased to 58 ± 17 mmHg (p = 0.031) and 40 ± 17 mmHg (p = 0.015) at 28°C brain temperature (Fig. 1).

FIG. 1.

Time course of CPP during the induction of hypothermia (mean ±95% CI). CPP, cerebral perfusion pressure; MAP, mean arterial blood pressure; P_btO₂, brain tissue oxygen tension.

P_btO₂ significantly increased from 9 ± 3 mmHg to 14 ± 5 mmHg during the cooling period (p = 0.037). This change remained significant even after correcting for end-tidal CO₂ (p = 0.012).

Baseline rSO₂ increased during the cooling period from 54% ± 3% to 62% ± 11% at 29°C (+7.3; 95% CI: 0.5–9.2; p = 0.029).

Cerebral AR indices

The time course of all measured cerebral AR indices is shown in Figure 2. Baseline AR-indices reflected normal cerebral AR in all pigs and did not change significantly until a temperature of 34°C (ORx), 33°C (PRx), and 30°C (COx) was reached. At lower temperatures, AR-indices significantly worsened: from −0.10 ± 0.25 to +0.34 ± 0.35 at 28°C (p < 0.001) for PRx, from 0.12 (±0.18) to 0.16 at 28°C (+0.28, p = 0.018) for ORx, and COx increased by 0.15 to a maximum of 0.17 ± 0.14 (p = 0.03), reflecting impaired cerebral AR. PRx and ORx correlated comparably well (r = 0.15 p < 0.01). There was no correlation between COx and PRx or ORx (p = 0.23, p = 0.28).

FIG. 2.

Time course of cerebral autoregulation indices over the cooling period (mean ±95% CI). *Difference significant p < 0.05. CI, confidence interval; COx, cerebral oximetry index; ORx, oxygen reactivity index; PRx, pressure reactivity index.

In a mixed-effects model, brain temperature remained significant (χ² = 319.9, p ≤ 0.001) and PRx was further associated with heart rate (HR) (χ² = 7.1329, p < 0.01), MAP (χ² = 17.937, p < 0.001), and EtCO₂ (χ² = 4.44, p = 0.03). ORx was associated with brain temperature (χ² = 66.39, p < 0.001) and was also influenced by MAP (χ² = 57.125, p < 0.001) and EtCO₂ (χ² = 655.86, p < 0.001). No significant association was seen for ORx and HR (p = 0.053). In this mixed model, COx showed an insignificant association to brain temperature (p = 0.12). It was, however, associated with MAP (χ² = 114.87, p < 0.001), HR (χ² = 12.28, p < 0.001), and EtCO₂ (χ² = 7.38, p < 0.01).

End-tidal CO₂ significantly decreased from 38.25 mmHg (±1.95) at 39°C to 22.5 mmHg (±5.17) at 28°C (p < 0.01).

Fraction of inspired oxygen (FiO₂) was kept constant during the whole cooling period at 21% ± 0.9%. Hematocrit and hemoglobin levels were comparable in all piglets with a median of 24% (IQR: 23–25%) and 8.2 mg/dL (IQR: 8.2–8.2 mg/dL), respectively.

Discussion

The main finding of this study is that mild HT (36–34°C) did not change systemic hemodynamic variables and cerebral AR indices, whereas moderate-to-deep HT (34–28°C) was associated with a significant decrease in MAP, CPP and a derangement of cerebral AR.

Cerebrovascular AR refers to the ability of the brain vessels to regulate cerebral blood flow (CBF) over a wide range of CPP levels (Donnelly et al., 2016). When CPP decreases below the lower or above the upper inflection point of the cerebral AR curve, oxygen and glucose delivery solely depend on the perfusion pressure. This evidentially results in cerebral hypoperfusion and ischemic brain injury when CPP is too low, or cerebral edema and raised ICP when CPP is raised above the upper inflection point of the cerebral AR curve (Skinhoj and Strandgaard, 1973). Besides other factors (Donnelly et al., 2016), HT itself may influence cerebral AR leading to pressure-dependent changes in CBF.

Controlled HT is routinely used as a neuroprotectant perioperatively during cardiovascular surgery and during intensive care of postcardiac arrest patients (Nielsen et al., 2013; Smith et al., 2017). Further, HT is still an integral part in the neurocritical care management of traumatic brain injured patients through its effect on decreasing raised ICP (Flynn et al., 2015). However, hemodynamic side effects, including a decrease in blood pressure necessitating vasopressors, are common (Andrews et al., 2016). The optimal targeted blood pressure to maintain sufficient energy delivery to the brain has yet to be defined in the setting of moderate and deep HT. It is important to note that most studies investigating the effect of HT on cerebrovascular AR used controlled HT. In contrast to controlled HT used intraoperatively or during targeted temperature management, the drop in body temperature during accidental HT is unintentional and uncontrolled and may even be associated with hemodynamic side effects, including hypotension. Our animal model best represents this clinical scenario. Based on this experimental design, we did not augment blood pressure during episodes of hypotension (e.g., by the use of vasopressors).

Cerebral AR indices, which were introduced in the management of the critically ill neurological patient in the 1990s (Czosnyka et al., 1997), may allow for monitoring of AR integrity and facilitate individual adjustments of MAP to improve cerebral perfusion in hypothermic patients. Today, many indices of cerebrovascular AR have been described. The PRx relates to the correlation between MAP and ICP (Czosnyka et al., 1997). It shows negative values while AR is preserved, whereas with impaired AR an increase in MAP will cause passive dilation of cerebral vessels and, consequently, an increase in CBF and ICP, resulting in positive PRx values. Analogous, the ORx describes the relationship between CPP and P_btO₂ (Jaeger et al., 2006), and the COx between COx between MAP and rSO₂ (Rivera-Lara L et al., 2017). When cerebral vessels are pressure passive because of impaired AR, an increase in CPP will lead to an increase in CBF and oxygen delivery and therefore P_btO₂ and rSO₂. Thus, both indices will be positive. Cerebral oxygenation depends on many aspects alongside cerebral AR. The physiological cascade assessed by oxygen derived indices such as ORx and, to an even greater extend, COx, may be influenced by physiological and environmental circumstances which may further be determined by temperature. This limitation should be kept in mind when comparing ORx and COx to other indices such as the pressure derived PRx.

There is no consensus on which index is best used (Donnelly et al., 2016), as they represent very different physiological aspects (Zeiler et al., 2017). We, therefore, assessed all three AR indices and observed the same trend when analyzing relative changes over time. Cerebral AR was relatively stable during mild HT (36–34°C), which is consistent with previous studies (Koizumi et al., 2016). The PRx, the ORx, and the COx indicated loss of AR at temperatures below 33°C, 34°C, and 30°C, respectively (Fig. 2). Although the COx was shown to be equivalent to invasive AR indices in its capacity to indicate AR (Lee et al., 2012), it reacted slower and less prominent than the other two indices in our study. In contrast to the PRx and ORx, which are derived from invasively measured parameters, COx is calculated from noninvasively measured cerebral rSO₂ obtained by NIRS and MAP, which obviates the need for invasive measurements. The noninvasive character of NIRS may seem advantageous as a potential monitoring tool in the prehospital or perioperative setting. However, integration of NIRS values in the critical care management of acutely brain injured patients is challenging as absolute values are influenced by many factors, including perfusion pressure, CBF, arterial carbon dioxide and oxygen levels, oxygen consumption, temperature, and even ambient light (Kurth and Uher, 1997; Boris Kovalenko and Michael, 2015; Putzer et al., 2016), but most importantly the extracranial venous influence (Davie and Grocott, 2012). Based on our results in addition to technical limitations, COx measurements should be interpreted carefully when assessing AR in hypothermic patients.

Our data further demonstrate that HT below 34°C was associated with a decrease in blood pressure (Fig. 1). It is important to mention that CPP values were always above the lower reflection point of cerebral AR (∼30 mmHg) as reported for healthy pigs (Brady et al., 2008), and AR indices indicated loss of AR during deep HT (Brady et al., 2008). Moreover, we observed an increase in P_btO₂ at lower temperature levels, which indicates that our pigs were still in the range of autoregulatory capacity. Therefore, the decrease in temperature most likely affected AR status and not the decrease in MAP. This is in contrast to a recent experimental study, where AR remained intact until body temperatures around 20°C (Smith et al., 2017). The statistical approach used in this referenced study is limited as the authors simply compared mean values at normothermia and HT and did not analyze AR indices as continuous physiologic variables changing over time. Recently, we demonstrated the importance of such an approach as the true difference observed within a physiologic variable may be washed out if a too simplistic analytical method is used (Rivera-Lara et al., 2017; Gaasch et al., 2018). Moreover, we did not administer vasoactive drugs that may have influenced their findings (Ristagno et al., 2007, 2009; Putzer et al., 2016).

Despite impaired AR status and a decrease in perfusion pressure in our experiment, we observed an increase in brain oxygenation. In a state of impaired cerebral AR, the observed increase in P_btO₂ may be due to hyperemia; however, given the low levels of MAP alongside the lack of ICP increase usually accompanying hyperemia (Kelly et al., 1996), the changes in P_btO₂ may more likely be interpreted as a decrease in energy consumption caused by metabolic suppression during HT (Frietsch et al., 2000; Erecinska et al., 2003). Indeed, HT decreases brain oxygen consumption by ∼6% per 1°C fall in temperature (Kurth and Uher, 1997), improving the tolerance for hypoxia during low-flow states. In fact, survival with excellent neurological outcome may be possible even in patients with deep accidental HT (<24°C) and hypotension (Pasquier et al., 2014).

The findings of this experimental study may not directly be transferable to humans. Piglets have higher body temperature at rest (Hannon et al., 1990), thus, the temperature levels at which impaired AR was documented may not be the same in humans. Also, ICP values observed in our study seem high in comparison to humans. However, our findings are compatible with previous studies reporting ICP values for sedated pigs in supine position (Yannopoulos et al., 2006; Heuer et al., 2012; Putzer et al., 2018). Further, accidental HT differs from controlled HT and our study results may therefore not be applicable for controlled HT settings.

During HT, we used the α-Stat regime to adopt ventilator settings and report blood gas values. Using the α-Stat regime actual partial pressure of CO₂ during HT is lower than measured and may influence cerebral AR (Murkin et al., 1987; Donnelly et al., 2016). As in clinical practice, the α-Stat regime is the common standard of care in cardiovascular surgery as well as the treatment of patients with severe accidental HT. Whether results of our study also apply for pH Stat management is questionable and needs further evaluation.

Conclusion

In conclusion, we found evidence of impaired cerebral AR during induced deep HT. Despite the decrease in MAP and CPP, brain oxygenation increased most likely due to a decrease in brain metabolism during HT. Further studies are needed to define the optimal blood pressure target during deep HT.

Footnotes

Acknowledgments

This work was funded by the Department of Anesthesiology and Critical Care Medicine, Medical University of Innsbruck, Innsbruck, Austria; and the Institute of Mountain Emergency Medicine at the European Academy, Bolzano, Italy.

Author Disclosure Statement

No competing financial interests exist.

References

Andrews

, Harris

, Murray

. Hypothermia for intracranial hypertension after traumatic brain injury. N Engl J Med, 2016; 374:1385.

Bates

, Maechler

, Bolker

. lme4: Linear Mixed-Effects Models Using S4 Classes. R Package Version 0.999999-0. 2012.

Boris Kovalenko

MRBAF

, Michael

. Effect of ambient light on near infrared spectroscopy. Trauma Treat, 2015; 4:258.

Brady

, Lee

, Kibler

, et al. Continuous measurement of autoregulation by spontaneous fluctuations in cerebral perfusion pressure: comparison of 3 methods. Stroke, 2008; 39:2531–2537.

Czosnyka

, Smielewski

, Kirkpatrick

, et al. Continuous assessment of the cerebral vasomotor reactivity in head injury. Neurosurgery, 1997; 41:11–17; discussion 17–19.

Davie

, Grocott

. Impact of extracranial contamination on regional cerebral oxygen saturation: a comparison of three cerebral oximetry technologies. Anesthesiology, 2012; 116:834–840.

Donnelly

, Budohoski

, Smielewski

, et al. Regulation of the cerebral circulation: bedside assessment and clinical implications. Crit Care, 2016; 20:129.

Erecinska

, Thoresen

, Silver

. Effects of hypothermia on energy metabolism in Mammalian central nervous system. J Cereb Blood Flow Metab, 2003; 23:513–530.

Flynn

, Rhodes

, Andrews

. Therapeutic hypothermia reduces intracranial pressure and partial brain oxygen tension in patients with severe traumatic brain injury: preliminary data from the Eurotherm3235 trial. Ther Hypothermia Temp Manag, 2015; 5:143–151.

10.

Frietsch

, Krafft

, Piepgras

, et al. Relationship between local cerebral blood flow and metabolism during mild and moderate hypothermia in rats. Anesthesiology, 2000; 92:754–763.

11.

Gaasch

, Schiefecker

, Kofler

, et al. Cerebral autoregulation in the prediction of delayed cerebral ischemia and clinical outcome in poor-grade aneurysmal subarachnoid hemorrhage patients. Crit Care Med, 2018; 46:774–780.

12.

Hannon

, Bossone

, Wade

. Normal physiological values for conscious pigs used in biomedical research. Lab Anim Sci, 1990; 40:293–298.

13.

Heuer

, Sauter

, Barwing

, et al. Effects of high-frequency oscillatory ventilation on systemic and cerebral hemodynamics and tissue oxygenation: an experimental study in pigs. Neurocrit Care, 2012; 17:281–292.

14.

Jaeger

, Schuhmann

, Soehle

, et al. Continuous assessment of cerebrovascular autoregulation after traumatic brain injury using brain tissue oxygen pressure reactivity. Crit Care Med, 2006; 34:1783–1788.

15.

Johnson

, Engquist

, Howells

, et al. Bedside xenon-CT shows lower CBF in SAH patients with impaired CBF pressure autoregulation as defined by pressure reactivity index (PRx). Neurocrit Care, 2016; 25:47–55.

16.

Joshi

, Brady

, Lee

, et al. Impaired autoregulation of cerebral blood flow during rewarming from hypothermic cardiopulmonary bypass and its potential association with stroke. Anesth Analg, 2010; 110:321–328.

17.

Keeling

, Tian

, Leshnower

, et al. Safety of moderate hypothermia with antegrade cerebral perfusion in total aortic arch replacement. Ann Thorac Surg, 2018; 105:54–61.

18.

Kelly

, Kordestani

, Martin

, et al. Hyperemia following traumatic brain injury: relationship to intracranial hypertension and outcome. J Neurosurg, 1996; 85:762–771.

19.

Koizumi

, Suehiro

, Fujiyama

, et al. Effects of brain temperature on cerebrovascular autoregulation during the acute stage of severe traumatic brain injury. Acta Neurochir Suppl, 2016; 122:193–197.

20.

Kurth

, Uher

. Cerebral hemoglobin and optical pathlength influence near-infrared spectroscopy measurement of cerebral oxygen saturation. Anesth Analg, 1997; 84:1297–1305.

21.

Lee

, Yang

, Wang

, et al. Noninvasive autoregulation monitoring in a swine model of pediatric cardiac arrest. Anesth Analg, 2012; 114:825–836.

22.

Lüdecke

ggeffects: Tidy data frames of marginal effects from regression models. J Open Source Softw, 2018; 3:772–777.

23.

Murkin

, Farrar

, Tweed

, et al. Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass: the influence of PaCO₂. Anesth Analg, 1987; 66:825–832.

24.

Nielsen

, Wetterslev

, Cronberg

, et al. Targeted temperature management at 33 degrees C versus 36 degrees C after cardiac arrest. N Engl J Med, 2013; 369:2197–2206.

25.

Paal

, Gordon

, Strapazzon

, et al. Accidental hypothermia-an update: the content of this review is endorsed by the International Commission for Mountain Emergency Medicine (ICAR MEDCOM). Scand J Trauma Resusc Emerg Med, 2016; 24:111.

26.

Pasquier

, Zurron

, Weith

, et al. Deep accidental hypothermia with core temperature below 24 degrees c presenting with vital signs. High Alt Med Biol, 2014; 15:58–63.

27.

Putzer

, Braun

, Martini

, et al. Effects of head-up vs. supine CPR on cerebral oxygenation and cerebral metabolism—a prospective, randomized porcine study. Resuscitation, 2018; 128:51–55.

28.

Putzer

, Braun

, Strapazzon

, et al. Monitoring of brain oxygenation during hypothermic CPR—a prospective porcine study. Resuscitation, 2016; 104:1–5.

29.

Ristagno

, Sun

, Tang

, et al. Effects of epinephrine and vasopressin on cerebral microcirculatory flows during and after cardiopulmonary resuscitation. Crit Care Med, 2007; 35:2145–2149.

30.

Ristagno

, Tang

, Huang

, et al. Epinephrine reduces cerebral perfusion during cardiopulmonary resuscitation. Crit Care Med, 2009; 37:1408–1415.

31.

Rivera-Lara

, Zorrilla-Vaca

, Geocadin

, et al. Predictors of outcome with cerebral autoregulation monitoring: a systematic review and meta-analysis. Crit Care Med, 2017; 45:695–704.

32.

Rivera-Lara

, Geocadin

, Zorrilla-Vaca

, et al. Validation of near-infrared spectroscopy for monitoring cerebral autoregulation in comatose patients. Neurocrit Care, 2017; 27:362–369.

33.

Skinhoj

, Strandgaard

. Pathogenesis of hypertensive encephalopathy. Lancet, 1973; 1:461–462.

34.

Smith

, Vu

, Kibler

, et al. Does hypothermia impair cerebrovascular autoregulation in neonates during cardiopulmonary bypass?. Paediatr Anaesth, 2017; 27:905–910.

35.

Yannopoulos

, McKnite

, Metzger

, et al. Intrathoracic pressure regulation for intracranial pressure management in normovolemic and hypovolemic pigs. Crit Care Med, 2006; 34:S495–S500.

36.

Zeiler

, Donnelly

, Menon

, et al. Continuous autoregulatory indices derived from multi-modal monitoring: each one is not like the other. J Neurotrauma, 2017; 34:3070–3080.