|Home | About | Journals | Submit | Contact Us | Français|
To date, considerable attention has been focused upon the use of hypothermia as a therapeutic strategy for attenuating many of the damaging consequences of traumatic brain injury (TBI). Despite the promise of hypothermic intervention following TBI, many questions remain regarding the optimal use of hypothermic intervention, including, but not limited to, the rewarming rates needed to assure optimal brain protection. In this review, we revisit the relatively limited literature examining the issue of hypothermia and differing rewarming rates following TBI. Considering both experimental and clinical literature, evidence is presented that the rate of posthypothermic rewarming is an important variable for influencing the protective effects of hypothermic intervention following TBI. In the experimental setting, posttraumatic hypothermia followed by slow rewarming appears to provide maximal protection in terms of traumatically induced axonal damage, microvascular damage and dysfunction, and contusional expansion. In contrast, hypothermia followed by rapid rewarming not only reverses the protective effects associated with hypothermic intervention, but in many cases, exacerbates the traumatically induced pathology and its functional consequences. While similar evaluations have not been conducted in the clinical setting, multiple lines of clinical evidence suggest the benefits of posttraumatic hypothermia are optimized through the use of slow rewarming, with the suggestion that such a strategy reduces the potential for rebound vasodilation, elevated intracranial pressure (ICP), and impaired neurocognitive recovery. Collectively, this review highlights not only the benefits of hypothermic intervention, but also the rate of posthypothermic rewarming as an important variable in assuring maximal efficacy following the use of hypothermic intervention.
As is fully developed in other papers in this compendium, there is ample evidence, both from the laboratory and clinical settings, to support the use of hypothermia as an important strategy for providing neuroprotection following traumatic brain injury (TBI). As noted in many reviews, the mechanisms through which hypothermia exerts its neuroprotective effects are not fully appreciated, yet evidence has emerged that hypothermic intervention can beneficially alter brain metabolism (Vink et al., 1987; Kaibara et al., 1999), reduce brain tissue loss (Bramlett et al., 1997; Matsushita et al., 2001), attenuate axonal damage (Koizumi and Povlishock, 1998) and microvascular dysfunction (Suehiro et al., 2001), reduce blood-brain barrier disruption (Jiang et al., 2006), and attenuate radical-mediated and neuroexcitatory-mediated pathobiology (Globus et al., 1995). Importantly, in animal models of TBI, hypothermic induction has also been shown to have beneficial behavioral effects (Bramlett et al., 1995). Additionally, using Glasgow Outcome Scores, improved outcome has also been demonstrated in traumatically brain injured patients who were subjected to elongated, mild hypothermic intervention (Jiang et al., 2006). While the above passage attests to the demonstrated efficacy of hypothermic intervention, it would be inappropriate to argue that the use of hypothermia should constitute a standard of care in the treatment of traumatically brain-injured patients. As fully detailed in other reviews in this compendium, many questions remain regarding the optimal use of hypothermia and the avoidance of potential complications typically associated with some aspects of this intervention. Specifically, multiple questions remain regarding the therapeutic window over which hypothermia exerts protection, the target temperature needed to achieve optimal protection, and the duration of hypothermic intervention needed to assure its optimal effects. Another important parameter typically unappreciated in the process of hypothermic intervention is the rate of posthypothermic rewarming needed to assure optimal protection. In this review, we will focus on the issue of posthypothermic rewarming following TBI to develop how rewarming rates significantly influence multiple parameters related to the use of hypothermic intervention. In this context, we will examine data gleaned from both the laboratory and clinical settings, demonstrating that the use of slow rewarming optimizes hypothermia's protective effects, whereas rapid posthypothermic rewarming in both injured and normal brains can have adverse consequences, even exacerbating traumatically induced pathology and dysfunction. Lastly, this review will close with a brief consideration of the potential mechanisms involved in the adverse consequences associated with rapid rewarming. For the purpose of clarity, this review will first present evidence derived from the laboratory followed by limited clinical data with a subsequent discussion of the potential mechanisms linked to the adverse consequences of rewarming.
As is alluded to above, the experimental literature supporting the damaging consequences of posthypothermic rewarming following TBI is limited and derived primarily from our laboratory. In this context, we have had a longstanding interest in the protective effects of hypothermic intervention, showing that the use of mild to moderate hypothermia (32–33°C), when administered in the early posttraumatic period, can significantly reduce the number of damaged axons seen throughout the brain and brain stem (Koizumi and Povlishock, 1998) (Fig. 1). In these studies, axonal damage was assessed through the use of antibodies targeting amyloid precursor protein (APP), which detects large intraaxonal accumulations of this protein, resulting from the impaired axonal transport triggered by the traumatic episode (Stone et al., 2000). These intraaxonal accumulations can be seen as axonal swellings distributed throughout multiple fiber pathways coursing throughout the brain and brain stem. Thus, these APP-positive axonal swellings can then be quantified per unit area, and this approach has been used to critically evaluate the protective effects of hypothermia. Importantly, these same APP axonal markers are also routinely used in human postmortem analyses, wherein the pathological axonal response to injury was originally detected (Sherriff et al., 1994), supporting the premise that these animal studies have fidelity to the human condition. While the protection provided by the use of 32–33°C hypothermia was not complete, it was statistically significant (Koizumi and Povlishock, 1998). Further, with the use of hypothermia, the numbers of damaged axons per unit area were differentially affected by both the time of onset and the duration of the posttraumatic hypothermic intervention (Fig. 1), issues further discussed in multiple chapters in this communication. When initially reported, these hypothermic interventions routinely used a rewarming period over which the animal's body temperature was moved from 32°C to 37°C over a span of 90min. In these studies, this was considered a slow rewarming rate, although by human standards, this rewarming strategy would not be considered slow. Despite this caveat, this strategy consistently resulted in axonal protection in multiple brain foci and in multiple fiber systems. Further, when these studies were replicated using other non-APP markers of traumatic axonal injury that revealed a previously underappreciated axonal pathology involving intraaxonal cytoskeletal compaction and nonswelling, similar axonal protection was afforded by the use of hypothermia followed by the same slow rewarming strategy (Büki et al., 1999). However, when these same axonal counting strategies were employed with hypothermia now followed by rapid rewarming wherein normothermia was restored over 20min, a complete reversal of the previously reported axonal protection was seen (Suehiro and Povlishock 2001; Suehiro et al., 2001) (Figs. 2a,b and and3).3). In fact, not only was the previously documented axonal protection eliminated via rapid rewarming, but also, and perhaps more importantly, the overall burden of axonal damage was dramatically increased, in some cases involving a virtual doubling of the numbers of damaged axons seen per unit area (Fig. 3). This observation was consistent across studies and, moreover, was consistent when different markers of axonal injury were employed. Specifically, using antibodies to the APP as well as neurofilament compaction to differentiate different populations of traumatically injured axons, we consistently observed the rapid rewarming reversed this protection and exacerbated the number of APP-containing and neurofilament-compacted axonal segments per unit of the brain (Suehiro et al., 2001). The mechanisms responsible for this exacerbated response to rewarming are not fully appreciated at present and will be discussed further in latter section of this review. It is of note, however, that in companion with these studies of hypothermic intervention and different rewarming rates, we also employed various axonal protection strategies that previously had been demonstrated to attenuate the damaging consequences of traumatic axonal injury without use of hypothermic intervention. Specifically, to date, we have shown that the use of the immunophilin ligands cyclosporin A (CsA) and FK506 exert statistically significant axonal protection (Okonkwo et al., 1999; Okonkwo and Povlishock, 1999; Buki, Okonkwo, and Povlishock, 1999; Singleton et al., 2001; Marmarou and Povlishock, 2006), perhaps through two diverse, but potentially interlinked, pathways involving the blunting of mitochondrial permeability transition and calcineurin inhibition. Using either systemically applied FK506 or intrathecally applied CsA given either preinjury or early postinjury following traumatically induced brain injury alone, both drugs significantly reduced the numbers of damaged axons per unit area. CsA treatment reduced the number of APP-containing and neurofilament-compacted axons, with the parallel demonstration of reduced cysteine protease activation (Okonkwo et al., 1999; Okonkwo and Povlishock, 1999; Buki et al., 1999). This protection was linked indirectly via multiple lines of evidence to CsA's ability to provide intraaxonal mitochondrial protection via the blocking of the mitochondrial permeability transition pore (Okonkwo and Povlishock, 1999). FK506 similarly reduced axonal damage, although its use only reduced the postinjury numbers of APP-containing axons, with no impact on the neurofilament compacted axonal profiles (Singleton et al., 2001; Marmarou and Povlishock, 2006). With FK506, multiple lines of evidence suggested that its attenuation of calcineurin activation was at work; however, the potential for mitochondrial protection could not be excluded (Singleton et al., 2001; Marmarou and Povlishock, 2006). In addition to the primary protective effects of these immunophilin ligands, other studies assessing TBI followed by hypothermia and rapid rewarming also confirmed their usefulness in this situation. Using CsA prior to rapid posthypothermic rewarming after TBI, we significantly attenuated the number of APP-positive axons per unit area within the corticospinal tract (Suehiro and Povlishock, 2001) (Fig. 3). Also, through the use of FK506, we observed similar protection involving a reduced number of APP-positive axons within the corticospinal system after hypothermia and rapid rewarming (Suehiro et al., 2001). Although the use of the immunophilin ligand FK506 did not provide protection to those axons demonstrating focal neurofilament compaction, it is conceivable that this fiber population, which undergoes cysteine protease-mediated degradation, may simply be too severely damaged to be attenuated via any therapeutic strategy (Suehiro et al., 2001). The benefits of both CsA and FK506 therapeutic intervention suggest that the period of rapid posthypothermic rewarming exacerbates the pathology initially attenuated by hypothermic intervention. In the case of axonal injury, this involves both local mitochondrial damage and cytoskeletal perturbation, both of which are intertwined in a complex cascade of subsequent cysteine protease activation, cytoskeletal degradation, and axonal collapse and disconnection.
In addition to the information provided above focusing on traumatically induced axonal injury and its modification via hypothermia and differing posthypothermic rewarming rates, there is now evidence that the cerebral microcirculation is also structurally and functionally perturbed following TBI and that such perturbation can be attenuated by hypothermia followed with slow rewarming. In fact, similar to the situation with axonal injury this hypothermic protection can be reversed or even exacerbated by the use of rapid posthypothermic rewarming. A rich literature spanning over 30 years, demonstrates that TBI is capable of triggering sustained microvascular damage and dysfunction, which has been assessed primarily in the cortical/pial microcirculation (Wei et al., 1980; Kontos and Povlishock, 1981; Kontos et al., 1980, 1981; Wei et al., 1981). The seminal work of Kontos and colleagues, which has been replicated in multiple species, demonstrates that the traumatic episode triggers discreet endothelial and smooth muscle damage and that such structural vascular abnormalities are associated with altered vascular responsiveness to known vasodilator challenges (Wei et al., 1980; Kontos et al., 1980). Through direct visualization and physiological assessment of the pial microcirculation, it has been shown that injured arterioles lose their vasodilator responses to known vasodilators such as acetylcholine, hypercapnia, pinacidil, and sodium nitroprusside. Based on our findings targeting both endothelial-dependent and independent responses, it appears that both the endothelium and smooth muscle are rendered dysfunctional by the TBI. While the overall functional implications of these vascular changes for traumatically brain-injured animals and potentially brain-injured humans is not fully appreciated, it is most likely that the vascular abnormalities and their associated vascular dysfunction to vasodilatory challenges render the brain more susceptible to secondary insults following the initial TBI. Recent work in our laboratory has revealed that this vascular dysfunction persists for multiple weeks following a traumatic episode, even in animals that appear behaviorally normal, suggesting that these animals are at increased risk for further brain injury should they sustain a secondary insult during this same time frame (unpublished observations). Similar to the information now known regarding traumatically induced axonal damage, there is also a rich literature existing on those factors responsible for these vascular abnormalities. Specifically, there is compelling evidence that traumatically induced generation of oxygen radicals is a major contributor to this pathobiology, exerting damaging consequences upon both the endothelium and smooth muscle cells (Kontos et al., 1983, 1984, 1986; Wei et al., 1985). In this context, direct assessments of radical production via the use of nitroblue tetrazolium reveal that oxygen radicals are produced for, at least, 30–60min in animal models of TBI (Kontos and Wei, 1986). The involvement of these oxygen radicals is also substantiated by the fact that the use of radical scavengers, particularly superoxide dismutase, when applied early postinjury, can protect the cerebral microcirculation and prevent vascular dysfunction, in terms of the above-described abnormal structural and vasodilator responses (Kontos and Wei, 1986; Kontos and Povlishock, 1986). Similarly, when comparably injured animals were treated with posttraumatic hypothermic intervention followed by slow rewarming, significant vascular protection was also provided, again in terms of the above-described vascular abnormalities (Suehiro et al., 2003; Ueda et al., 2003). As with previous descriptions derived from our laboratory, the protective effects of hypothermia were once again a function of time of initiation, target temperature, and overall duration (Ueda et al., 2003). Importantly, the use of rapid rewarming over the span of 20 minutes in rodents again reversed the protective effects of hypothermic intervention and in fact, contributed to enhanced vascular dysfunction following traumatic injury (Suehiro et al., 2003) (Fig. 4). The use of posttraumatic hypothermia followed by rapid rewarming exacerbated both endothelial and smooth muscle damage, with abnormal vasodilatory responses seen in terms of the use of endothelial-dependent and independent challenges. The use of acetylcholine, hypercapnia, pinacidil, and sodium nitroprusside consistently resulted in abnormal vascular responsivity, indicating a reversal of any vascular protective effect. From a mechanistic perspective, it appears that similar to TBI alone, the enhanced damaging consequences of posttraumatic hypothermia followed by rapid rewarming are also associated with oxygen radical-mediated processes. While studies of this aspect are ongoing, it appears that the use of oxygen radical scavengers can attenuate some of the damaging consequences of posttraumatic hypothermia and rapid rewarming; however, this situation may be more complex than that seen with trauma alone. It is of interest that in the absence of TBI, the use of hypothermia followed by fast rewarming induced cerebrovascular abnormalities similar to those seen in brain-injured animals receiving hypothermia followed by fast rewarming (Ueda et al., 2004).
While studies on these abnormal vascular responses triggered by uncomplicated hypothermia and rapid rewarming are ongoing, the damaging cascades resulting from rapid rewarming appear to be closely associated with oxygen radical-mediated processes. From a mechanistic perspective, preliminary data from our laboratory on this issue suggests that altered cerebral vascular reactivity to various dilators operates through different radical-mediated mechanisms. Specifically, the impaired vasodilation in response to ACh is mediated primarily through the production/release of superoxide anion since it can be restored by administration of SOD prior to onset of fast rewarming. On the other hand, CO2-induced dilation is impaired primarily by hydrogen peroxide and/or its derivatives, because this vasodilator response can be significantly preserved by catalase treatment prior to onset of fast rewarming. These observations are intriguing from several perspectives because they indicate that rapid posthypothermic rewarming, in itself, is capable of unleashing multiple damaging cascades that can adversely impact upon microvasculature and most likely other CNS components.
Paralleling the above descriptions of hypothermia's protective effects in terms of axonal injury and microvascular dysfunction, similar data also exists in relation to traumatically induced contusional change, with again evidence that the protective effects of posttraumatic hypothermia are lost with rapid rewarming. Dietrich et al. (1995, 1997, 2001) have provided multiple streams of evidence that hypothermia attenuates many of the damaging consequences of TBI including contusional volume, neuronal cell loss, ventricular dilation, and multiple behavioral sequelae. Importantly, however, all of these studies were conducted using hypothermic interventions in which the rewarming rates were not evaluated. Matsushita et al. (2001) reevaluated the issue of posttraumatic hypothermic protection, now considering the potential role of differential rewarming rates. Using fluid percussion brain injury capable of producing focal contusion, they utilized a secondary hypoxic insult which further exacerbated the contusional load. Following this combinational insult, they then employed hypothermic intervention followed by either slow or rapid rewarming. Through such an approach, the authors confirmed that hypothermia followed by slow rewarming provided protection, reducing the contusional volume initially triggered by the traumatic episode and further complicated by the induction of secondary hypoxia. However, when the traumatic and secondary insults were followed by hypothermia with rapid rewarming, this protection was lost, consistent with those observations described above for traumatically induced axonal damage and microvascular dysfunction. While it is beyond the scope of the current review to venture in the area of experimental ischemia and the effects thereupon of hypothermia followed by various rewarming rates, it is of note that important parallels exist between the above-described study of Matsushita et al. and assessment of the cerebral infarction following acute stroke, as modified by postischemic hypothermia followed by different rewarming rates. As the contusional change described above involves strong hemorrhagic and ischemic components, important parallels do exist with that damage ongoing in focal ischemia and subsequent infarction. In this vein, recent studies by Berger et al. (2007) utilizing hypothermia and differing rewarming rates following middle cerebral artery occlusion demonstrated that hypothermia followed by slow rewarming results in a significant reduction in infarct volume compared to the normothermic group. Further, relevant to the current discussion, the use of postischemic hypothermia followed by rapid rewarming did not provide any comparable reduction in infarct volume size. Thus, in the case of both traumatically induced contusion and focal ischemia and infarction, the potential benefits of hypothermic protection appear optimal with the use of hypothermia followed by slow rewarming, while the use of rapid rewarming obviates any potential protective effects.
In the previous passages, we have presented multiple lines of experimental evidence that posttraumatic hypothermic intervention followed by rapid rewarming reverses and perhaps exacerbates many of the protective effects associated with posttraumatic hypothermia followed by slow rewarming. We have shown, in the context of traumatically induced axonal change, microvascular dysfunction, and contusional volume, that any potential beneficial effect of hypothermic intervention is consistently reversed and in some cases, exacerbated, by the use of posttraumatic hypothermia followed by rapid rewarming. While in the preceding passages we have alluded to the potential mechanisms, it appears that in the case of axonal injury, microvascular dysfunction and contusional damage, both direct and indirect mitochondrial perturbation, together with radical-mediated processes, may be major players in the ensuing pathobiologies and dysfunction described. It is of note that the issue of posthypothermic rewarming rates and the potential adverse consequences of rapid rewarming have been considered in the field of transplantation wherein hypothermic organ maintenance and rewarming are integral to successful organ viability and subsequent transplantation. It has been shown with liver transplantation that the use of rapid rewarming is highly damaging, most likely due to the rapid ATP depletion, energy failure, and oxygen radical production associated with hepatic mitochondrial damage. Using cultured hepatocytes subjected to deep hypothermia followed by rapid rewarming, Leducq et al. (1998) demonstrated that this strategy resulted in ATP depletion with energy failure and mitochondrial dysfunction. Importantly, Leducq et al. (1998) posited that mitochondrial damage was most likely caused by an opening of the mitochondrial permeability transition pore, testing this hypothesis via the concomitant administration of CsA which, as noted previously, blocks this pore via binding to an immunophilin site on the pore complex. This strategy not only resulted in mitochondrial protection but importantly, translated into the maintenance of mitochondrial functional integrity. In aggregate, these studies and interpretations made in hepatocytes have remarkable similarities to the axonal, vascular, and contusional pathologies described above wherein hypothermia followed by rapid rewarming was associated with a host of potentially interrelated mitochondrial disorders that translated into axonal failure, microvascular dysfunction, and increased infarct volume. The demonstration that CsA can attenuate many of the damaging axonal consequences associated with posttraumatic hypothermia followed by rapid rewarming (Suehiro and Povlishock, 2001) suggests that, similar to the findings of Leducq et al. (1998), traumatically induced mitochondrial damage is exacerbated by rapid posthypothermic rewarming and that this mitochondrial damage is attenuated through the use of CsA via its binding to the immunophilin site on the pore complex. Although the use of CsA in microvessels sustaining TBI followed by hypothermia and rapid rewarming has not been critically assessed, the microvascular anomalies described also bear a strong linkage to potential mitochondrial abnormalities. Lastly, in that traumatically induced contusion has also been long associated with inflammatory and radical-mediated processes, there is the likelihood that the use of hypothermia followed by rapid rewarming could exacerbate these radical-mediated processes and their damaging cascades. While we believe that evidence derived from multiple streams of investigation support the fact that rapid posthypothermic rewarming may exacerbate mitochondrial pathology and exacerbate/accelerate various radical-mediated processes, it is important to note that this premise is simplistic and most likely will be shown to be more complex. This is underpinned by our current limited understanding of these processes triggered by hypothermic intervention. Currently, our beliefs are moving from a once singular view of hypothermia's ability to exert pan-inhibition of all biological destructive processes associated with trauma and/or ischemia to a much more complex understanding of hypothermia's effects involving both up and down regulation of multiple biological processes that can exert either protective or damaging consequences. While the actual benefits of hypothermia following TBI are incomplete, there is relatively significant literature in ischemia that comprehensively considers to the benefits and limitations of hypothermic intervention. In a recent review by Zhao et al. (2007), the authors summarized the benefits and limitations of hypothermic intervention highlighting some of the inhibitory effects on various biological/pathological processes relevant to the current discussion. However, in addition to addressing the benefits of hypothermia and its potential inhibition of damaging cascades, they also reviewed the less appreciated literature that hypothermic intervention may also inhibit various protective mechanisms and/or genes such as those related to Heat Shock protein and nerve growth factors. In reviewing the protective effects of hypothermia in cerebral ischemia, Sapolsky et al. observed that the majority of experimental hypothermic studies had focused upon interischemic hypothermia rather than postischemic hypothermia, which is more relevant to the clinical situation and for that matter, the situation addressed with TBI. Importantly, they observed that with ischemia, hypothermia most likely accelerated ATP recovery during reperfusion, a fact that also may be relevant to TBI, although the traumatic episode is not complicated by a reperfusion phase. Admittedly the above discussion provides little direct evidence relevant to the benefits or liabilities associated with different posthypothermic rewarming rates yet, it does illustrate the complexity of the questions at hand, not only in the context of posthypothermic rewarming rates but also in the context of the precise mechanisms of hypothermia's purported protective effects. In light of the literature emerging in the use of hypothermia in ischemia, it is most likely the protection afforded by hypothermia with slow rewarming versus the damage precipitated by hypothermia and fast rewarming involves multi-factorial processes that have not been rigorously evaluated. For example, based upon our knowledge of hypothermic intervention with ischemia, it is conceivable that posttraumatic hypothermia with rapid rewarming may exacerbate a host of damaging processes while also potentially downregulating various protective mechanisms. Obviously, these issues all require further detailed investigation.
While the use of hypothermia in the clinical situation continues to generate both interest and controversy, there is also emerging clinical interest in the area of rewarming following the use of posttraumatic hypothermic intervention. The literature in this area remains thin, with few direct assessments and/or comparisons of specific rewarming rates postinjury; however, there appears to be an evolving belief that the use of rapid posthypothermic rewarming does not lead to improved outcome. Several studies have alluded to the fact that rapid posthypothermic rewarming may not be beneficial and in fact, may reverse many of the potential protective effects of hypothermic intervention. Unfortunately, however, detailed assessments of the benefits of differing rewarming rates after hypothermic intervention following TBI have not been reported. Rather, discussions of differing rewarming rates are typically embedded in multi-faceted studies or reviews of hypothermic intervention and its potential benefits and limitations (Alzaga et al., 2006; Bernard and Buist, 2003; Bloch, 1967; Vigue et al., 2006). To date, the majority of the considerations regarding different rewarming rates have been embedded in studies focusing on hypothermic management of severe TBI, where it is well recognized that lack of attention to rewarming rates can lead to sudden vasodilation and rebound increases in intracranial pressure (Alzaga et al., 2006; Jiang et al., 2000; Jiang and Yang, 2007). These abnormal vasodilatory and ICP responses are most significant in those patients in whom the posttraumatic hypothermic period was relatively short, with those patients who were cooled for several days and then slowly rewarmed showing the most benefit (Jiang et al., 2006). This illustrates some of the complexities associated with the clinical interpretation of different rewarming rates in that it appears that several of the purported protective effects of slow rewarming are also influenced by the more extended use of hypothermic intervention. Overall, this issue is also complicated by the fact that with rewarming an overshoot of the baseline temperature of 37°C results in significant temperature-dependent impairment in cerebral vasoreactivity and possibly autoregulation (Lavino et al., 2007). Collectively, these studies suggest that in the clinical setting, more attention must be paid to rewarming rates and the attainment of target temperature to assure hypothermia's optimal effects. In the clinic, the beneficial effects of slow rewarming are also attested to by other literatures that have examined different CNS abnormalities, such as the relatively extensive literature on the use of hypothermia in cardiac surgery, with its potential implications for the patient's neurocognitive outcome. Comparing neurocognitive function following cardiac surgery, Grigore et al. (2002) demonstrated that the rewarming rate was an important variable, with slow rewarming during cardiopulmonary bypass resulting in better cognitive performance at 6 weeks postsurgery than that seen with more rapid rewarming. Collectively, these studies, together with those emerging in the context of TBI, provide evidence that slower rewarming rates should be routinely employed.
Funding for these studies was provided by the National Institutes of Health (grants R01NS045824, R01HD055813, and R21NS057175-01).