Bench-to-bedside review: Hypothermia in traumatic brain injury

Traumatic brain injury remains a major cause of death and severe disability throughout the world. Traumatic brain injury leads to 1,000,000 hospital admissions per annum throughout the European Union. It causes the majority of the 50,000 deaths from road traffic accidents and leaves 10,000 patients severely handicapped: three quarters of these victims are young people. Therapeutic hypothermia has been shown to improve outcome after cardiac arrest, and consequently the European Resuscitation Council and American Heart Association guidelines recommend the use of hypothermia in these patients. Hypothermia is also thought to improve neurological outcome after neonatal birth asphyxia. Cardiac arrest and neonatal asphyxia patient populations present to health care services rapidly and without posing a diagnostic dilemma; therefore, therapeutic systemic hypothermia may be implemented relatively quickly. As a result, hypothermia in these two populations is similar to the laboratory models wherein systemic therapeutic hypothermia is commenced very soon after the injury and has shown so much promise. The need for resuscitation and computerised tomography imaging to confirm the diagnosis in patients with traumatic brain injury is a factor that delays intervention with temperature reduction strategies. Treatments in traumatic brain injury have traditionally focussed on restoring and maintaining adequate brain perfusion, surgically evacuating large haematomas where necessary, and preventing or promptly treating oedema. Brain swelling can be monitored by measuring intracranial pressure (ICP), and in most centres ICP is used to guide treatments and to monitor their success. There is an absence of evidence for the five commonly used treatments for raised ICP and all are potential 'double-edged swords' with significant disadvantages. The use of hypothermia in patients with traumatic brain injury may have beneficial effects in both ICP reduction and possible neuro-protection. This review will focus on the bench-to-bedside evidence that has supported the development of the Eurotherm3235Trial protocol.


Introduction
It is important to remember that traumatic brain injury (TBI) is a major cause of death and severe disability throughout the world. TBI leads to 1,000,000 hospital admissions per annum throughout the European Union. It causes the majority of the 50,000 deaths from road traffi c accidents and leaves 10,000 patients severely handi capped: three quarters of these victims are young people [1]. Additionally, TBI causes 290,000 hospital admissions and 51,000 deaths and leaves 80,000 patients with permanent neurological disabilities in the US annually [2]. Th e consequence of this is both a devastating emotional and physical impact and an enormous fi nancial burden [3].
Th erapeutic hypothermia has been shown to improve outcome after cardiac arrest [3], and consequently the European Resuscitation Council and American Heart Association guidelines [4,5] recommend the use of hypothermia in these patients. Hypothermia is also thought to improve neurological outcome after neonatal birth asphyxia [6]. Cardiac arrest and neonatal asphyxia patient populations present to health care services rapidly and without posing a diagnostic dilemma; there fore, therapeutic systemic hypothermia may be imple men ted relatively quickly. As a result, hypothermia in these two populations is similar to the laboratory models wherein systemic therapeutic hypothermia is com men ced very soon after the injury and has shown so much promise [7].
Th e need for resuscitation and computerised tomography (CT) imaging to confi rm the diagnosis in patients with TBI is a factor that delays intervention with tempera ture reduction strategies. Treatments in TBI have traditionally focussed on restoring and maintaining adequate brain perfusion, surgically evacuating large Abstract Traumatic brain injury remains a major cause of death and severe disability throughout the world. Traumatic brain injury leads to 1,000,000 hospital admissions per annum throughout the European Union. It causes the majority of the 50,000 deaths from road traffi c accidents and leaves 10,000 patients severely handicapped: three quarters of these victims are young people. Therapeutic hypothermia has been shown to improve outcome after cardiac arrest, and consequently the European Resuscitation Council and American Heart Association guidelines recommend the use of hypothermia in these patients. Hypothermia is also thought to improve neurological outcome after neonatal birth asphyxia. Cardiac arrest and neonatal asphyxia patient populations present to health care services rapidly and without posing a diagnostic dilemma; therefore, therapeutic systemic hypothermia may be implemented relatively quickly. As a result, hypothermia in these two populations is similar to the laboratory models wherein systemic therapeutic hypothermia is commenced very soon after the injury and has shown so much promise. The need for resuscitation and computerised tomography imaging to confi rm the diagnosis in patients with traumatic brain injury is a factor that delays intervention with temperature reduction strategies. Treatments in traumatic brain injury have traditionally focussed on restoring and maintaining adequate brain perfusion, surgically evacuating large haematomas where necessary, and preventing or promptly treating oedema. Brain swelling can be monitored by measuring intracranial pressure (ICP), and in most centres ICP is used to guide treatments and to monitor their success. There is an absence of evidence for the fi ve commonly used treatments for raised ICP and all are potential 'double-edged swords' with signifi cant disadvantages. The use of hypothermia in patients with traumatic brain injury may have benefi cial eff ects in both ICP reduction and possible neuro-protection. This review will focus on the bench-to-bedside evidence that has supported the development of the Eurotherm3235Trial protocol. haema tomas where necessary, and preventing or promptly treating oedema [3]. Brain swelling can be monitored by measuring intracranial pressure (ICP), and in most centres ICP is used to guide treatments and to monitor their success. Th ere is an absence of evidence for the fi ve commonly used treatments for raised ICP and all are potential 'double-edged swords' with signifi cant disadvantages. Th e use of hypothermia in patients with TBI may have benefi cial eff ects in both ICP reduction and possible neuro-protection.

Pathophysiology
Ischaemia has a key role in all forms of brain injury and preventing ischaemic (or secondary) injury is at the core of all neuro-protective strategies [3]. A complex cascade of processes ensues at the cellular level after a period of ischaemia (Table 1), beginning from minutes to hours after injury and continuing for up to 72 hours or longer. Th us, there may be a window of opportunity of several hours, or even days, during which injury can be mitigated by treatments such as hypothermia [3].
Early studies that used profound hypothermia in models of brain trauma gave inconsistent results. More recent studies have shown that moderate/mild hypothermia appears to be neuro-protective in well-characterised rodent models of TBI. Th e eff ects of systemic hypothermia (30°C to 36°C) following fl uid percussion brain injury in rats were fi rst investigated by Clifton and colleagues [8], who showed that hypothermia of 33°C reduced mortality rates and attenuated defi cits in motor function and weight loss compared with normothermia. Dietrich and colleagues [9] showed that post-traumatic hypothermia (30°C) initiated 5 minutes after fl uid percussion brain injury reduced overall contusion volume Reduced mitochondrial Mitochondrial dysfunction is a frequent occurrence in the hours to days after a period of Hours to days dysfunction and improved ischaemia and may be linked to apoptosis. Hypothermia reduces metabolic demands and energy homeostasis b may improve mitochondrial function.
Reduction in free radical Production of free radicals (for example, superoxide, peroxynitrate, hydrogen peroxide, and Hours to days production b hydroxyl radicals) is typical in ischaemia. Mild-moderate (30°C to 35°C) hypothermia is able to reduce this event.
Mitigation of reperfusion Cascade of reactions following reperfusion, partly mediated by free radicals but with Hours to days injury b distinctive and various features. These are suppressed by hypothermia.
Reduced permeability of Blood-brain barrier disruptions induced by trauma or ischaemia are moderated by Hours to days the blood-brain barrier and hypothermia. The same eff ect occurs with vascular permeability and capillary leakage. the vascular wall and reduced oedema formation a Reduced permeability of Decreased leakage of cellular membranes with associated improvements in cell function Hours to days cellular membranes and cellular homeostasis, including decrease of intracellular acidosis and mitigation of (including membranes of DNA injury the cell nucleus) b

Improved ion homeostasis
Ischaemia induces accumulation of excitatory neurotransmitters such as glutamate and First minutes to 72 hours prolonged excessive infl ux of Ca2 + into the cell. This activates numerous enzyme systems (kinases) and induces a state of hyperexcitability (exitotoxic cascade) that can be moderated by hypothermia.
Reduced metabolism a Cellular oxygen and glucose requirements are reduced by 5% to 8% per degree Celsius Hours to days decrease in temperature.

Depression of the immune
Sustained destructive infl ammatory reactions and secretion of pro-infl ammatory cytokines First hour to 5 days response and various after ischaemia can be blocked or mitigated by hypothermia. potentially harmful pro-infl ammatory reactions a Reduction in cerebral Some areas in the brain have signifi cantly higher temperatures than others. These diff erences Minutes to days thermopooling a can increase dramatically after injury with temperatures that are up to 2°C to 3°C higher in injured areas. Hyperthermia can increase the damage to the injured brain cells; this is mitigated by hypothermia.
Anticoagulant eff ects a Microthrombus formation may add to brain injury after CPR. Anticoagulant eff ects of Minutes to days hypothermia may prevent thrombus formation.

Suppression of epileptic
Seizures after ischaemic injury or trauma are common and may add to injury. Hypothermia Hours to days activity and seizures a has been shown to mitigate epileptic activity.
and preserved survival of the overlying cortical neurons.
Th erefore, these studies demonstrated that cooling after a TBI provided histological/cellular protection, improved motor and cognitive function, and reduced mortality. Moderate hypothermia (30°C) initiated 5 minutes after TBI improved hippocampal-dependent learning and memory using the Morris water maze [10]. An important predictor of outcome in TBI patients, traumatic axonal pathology, is reduced with moderate post-injury hypothermia therapy [11]. Th erefore, post-traumatic hypother mia modulates the major pathologies in TBI such as contusions, neuronal vulnerability, and traumatic axonal injury. Mild hypothermia is therefore potentially attractive as it modulates multiple mechanisms or pathways and has advantages over unipolar pharmacological attempts to provide neurological protection.

Blood-brain barrier
Alterations in blood-brain barrier (BBB) permeability after acute injury result in the crossing of water, electrolytes, blood-borne substances, and potential neurotoxic agents across the vascular system and into the brain parenchyma. Many studies have demonstrated the importance of brain and body temperature on the microvascular consequences of cerebral ischaemia and trauma. One study that assessed the eff ects of intra-ischaemic brain temperature (mild hypothermia) on BBB was shown to reduce extravasation of the protein tracer horseradish peroxidase [12]. Brain water content is signifi cantly reduced with hypothermia after focal cerebral ischaemia [13,14]. Recent studies have assessed this with magnetic resonance imaging and found that reductions in the apparent diff usion coeffi cient of water (cellular oedema) are also attenuated by hypothermia [15]. In models of post-traumatic injury, hypothermia has also been shown to reduce BBB permeability. Hypothermia may be attenuating BBB permeability by altering matrix metalloproteinases, which are critical extracellular enzymes that can disrupt the BBB [16]. Th ese modulating eff ects of hypothermia on BBB permeability are important benefi cial mechanisms because of the association between BBB permeability, formation of vasogenic oedema, the extravasation of circulating infl ammatory cells and adverse post-injury outcomes.

Infl ammation and oedema
Th e infl ammatory response after TBI is signifi cantly modulated by hypothermia in laboratory and clinical studies. As well as attenuating the increase in BBB permeability and leucocyte margination, the endogenous infl ammatory response of the central nervous system (CNS) is reduced by hypothermia. Astrocytes and micro glia respond to CNS injury by proliferating around the injury areas and releasing pro-infl ammatory communi cation molecules as an endogenous repair mechanism. Hypothermia signifi cantly attenuates the activation of both astrocytes and microglia [17][18][19][20]. Combination therapy including the anti-infl ammatory cytokine inter leukin-10 and hypothermia therapy was attempted in both TBI and focal cerebral ischaemia [21]. Synergistic eff ects were seen in focal cerebral ischaemia but not in TBI, suggesting that the cellular biology of infl ammation in these two major CNS injuries has an infl uence on the eff ect of subsequent hypothermia. Another major aspect of the infl ammatory response to CNS injury is the release of reactive oxygen species by astrocytes and microglia. Hypothermia reduces increases in tissue levels of superoxide, nitric oxide, and the hydroxyl radical [22,23]. Superoxide dismutase, the enzyme responsible for scavenging superoxide, is increased by hypothermia [24,25], and the enzyme responsible for synthesising nitric oxide, nitric oxide synthase, is attenuated by hypo thermia [26].

Metabolism
Exploiting local measures of glucose metabolism using 2-deoxyglucose techniques, researchers showed that moderate hypothermia (30°C) reduced glucose utilisation compared with normothermia [27]. Metabolic eff ects of mild hypothermia have also been shown using nuclear magnetic resonance spectroscopy, in which the metabolic eff ects of diff erent levels of hypothermia were reported [28]. Th erefore, hypothermia lowers metabolic and energy demands, having potentially benefi cial eff ects on cytoplasmic ATP and the maintenance of normal transmembrane ion and neurotransmitter gradients. Th e magnitude of preservation of ATP levels depends on both the temperature reduction and the severity of the injury. Th erefore, an important mechanism for the neuroprotective eff ects of hypothermia is a reduction or delay in metabolic demand during and after an acute CNS injury.

Excitotoxicity
Th e eff ects of moderate hypothermia on glutamate excito toxicity were reported using microdialysis to assay extracellular/extravascular concentrations of neurotransmitters after global ischaemia. Th e middle cerebral artery occlusion model is considered a reasonable model for traumatic hemorrhagic contusion (John Povlishock, personal communi cation). Busto and colleagues [22] showed that intra-ischaemic hypothermia (33°C and 30°C) attenuated the rise in extracellular levels of glutamate and dopamine after global cerebral ischaemia. Th ese studies have been replicated in a variety of models of ischaemia, indicating that one of the major mechanisms by which temperature aff ects neuronal vulnerability is through reducing excito toxicity following cerebral ischaemia [22,29,30]. Delayed pharmacological treatments that reduce excitotoxicity further improve outcome in combination with hypo thermia [31] and may be a promising strategy for further studies. Th e glutamatergic receptors, AMPA (alpha-amino-3-hydroxyl-5methyl-4-isoxazole-propionate) and NMDA (N-methyl d-aspartate), are also modulated by hypothermia. Expression of hippocampal glutamate receptors is decreased after transient global ischaemia and this is completely blocked by intra-ischaemic hypothermia [32].
Other neurotransmitters are also modulated by hypother mia. Lyeth and colleagues [33] demonstrated that hypothermia (30°C) reduced elevations in cerebrospinal levels of acetylcholine after TBI. Conversely, hypothermia delayed decreases in dopamine, norepinephrine, and serotonin after global cerebral ischaemia [34]. But other studies have demonstrated that hypothermia (32°C) can improve outcome after CNS injury without attenuating extracellular levels of glutamate and aspartate [11,28,35]. Th e neurotransmitter response, in various injury models, may be temperature-dependent, but attenuating other injury cascades may be more important in delivering possible benefi cial eff ects of hypothermia.

Cerebrovascular eff ects
Th e eff ects of hypothermia on cerebral blood fl ow are controversial. In 1954, Rosomoff and Holaday [36] demon strated that systemic hypothermia to 25°C signi ficantly lowered cerebral blood fl ow. However, in a model of selective brain cooling (30°C), cortical blood fl ow measured by laser Doppler fl owmetry was shown to increase above control levels [37]. Cerebrovascular changes secondary to cooling the brain are important because reductions in blood fl ow to critical levels could have adverse eff ects on tissue survival and functional outcome.

Intracellular calcium-dependent signalling
Th ere are pronounced changes in calcium-dependent intra cellular signalling pathways after CNS injury. Normal neuronal activity is mediated by signalling through protein kinases and several of these have been documented to be disrupted by TBI and cerebral ischaemia. Temporary cerebral ischaemia inhibits the activity of calcium/calmodulin-dependent protein kinase II (CaMKII), a key protein kinase that mediates synaptic strength and this is attenuated by hypothermia [38]. Protein kinase C (PKC) translocates to the membrane after cerebral ischaemia and undergoes inhibition; hypothermia rescues the inhibition of PKC activity and its translocation to the membrane [39]. Recently, various transcription factors that participate in normal neuronal functioning have been shown to be sensitive to temperature. Th e immediate early gene c-Fos, which regulates key genetic responses of neurons, is activated by hypothermia after transient global ischaemia [17,40,41]. Th ese studies underline that tem pera ture may have profound eff ects on events associated with neuronal injury as well as the normal processing of neuronal signals throughout brain circuits.

Neuronal cell death
Although neuronal necrosis is commonly seen in most CNS injury models, evidence for apoptotic cell death has also been documented using various histochemical and molecular techniques. As with necrosis, apoptotic cell death appears to be sensitive to post-injury hypothermic treatment strategies. Recent studies indicate that apoptotic cell death is another important target by which temperature may aff ect long-term outcome in various models of CNS injury. Various gene families (genes with a similar sequence of DNA nucleotides) have been shown to be sensitive to post-injury temperature manipulations in models of ischaemia and trauma [42]. Th e ability of post-injury temperature to aff ect the acute and more delayed genetic response to injury is important in that these genes may be important in determining the cellular response or responses that result in secondary injury. Genomic studies using high-throughput screening and bioinformatics are ongoing in many laboratories and these contemporary technologies will help to determine how hypothermia may protect and potentially repair CNS tissues after injury.
Systematic review and meta-analysis have not been rigorously applied to TBI models. Such an approach is beyond the scope of this paper but in the future would enhance translational research and (hopefully) reduce futile studies. Th e CAMARADES (Collaborative Approach to Meta Analysis and Review of Animal Data from Experimental Stroke) group have successfully applied such methodology to stroke models [43].
Th erefore (with the proviso above), numerous animal experiments, in diff erent species, have shown that induced hypothermia improves outcome after experimental TBI. Th is has led to the undertaking of a large number of clinical trials [3]. Interpretation of these results is complicated by the fact that these studies have enrolled diff erent categories of patients, with diff erent types of injuries, and have used widely diverging treatment protocols [44]. Most have used elevated ICP as an inclusion criterion although some have used a CT scan criterion. Th e duration of cooling has varied from 24 hours to more than 5 days, and re-warming rates have also varied. Some studies have used ICP to guide depth and duration of treatment, although responses to rebound intracranial hypertension have diff ered [3]. Cointerventions such as osmotic therapy, sedation, analgesia, paralysis, and targets for mean arterial pressure and cerebral perfusion pressure have also varied considerably [3]. All of these factors can aff ect outcome after TBI in general and the potential effi cacy of cooling in particular. Th us, interpreting, comparing, and aggregating the results of these studies present a number of complex challenges.

Review of clinical evidence
Th ere have been eight meta-analyses carried out on this subject from the years 2002 to 2009. Th is section is a literature review of the available evidence. In total, 29 clinical studies have been performed to assess the eff ects of hypothermia in TBI. Twenty-seven of these were performed in adult patients, and 18 of these 27 included control groups. Data from one pilot study were subsequently included in a larger study, therefore leaving 17 studies. As outlined above, study protocols have diff ered considerably, and not all studies were (properly) randomised [3]. A total of 131 patients were enrolled into two studies undertaken in patients with normal ICP. Only one of these studies reported outcome data (at 3 months) and the results showed no signifi cant diff erence between groups (good outcome in 21/45 [hypothermia] versus 27/46 patients [controls], P = 0.251) [45].
In contrast, one of the two largest multi-centre randomised controlled trials (RCTs) failed to show that induced hypothermia improved outcome at 6 months after TBI (relative risk [RR] of a poor outcome 1; 95% confi dence interval [CI] 0.8 to 1.2; P = 0.99) [46]. Significantly more of the patients admitted to hospital with hypothermia who were randomly assigned to normothermia, and consequently re-warmed, had a poor outcome (78%, n = 31) compared with patients admitted with hypothermia and treated with hypothermia (61%, n = 38) (P = 0.09).
On subsequent analysis, it became clear that although this study was methodologically well designed, there was marked inter-centre variance in the treatment eff ect of hypothermia, age of participants, severity of illness scoring between groups, management of intracranial hypertension, and haemodynamic and fl uid management [64]. Induced hypothermia in the hypothermia group was started relatively late with a slow speed of cooling (average time to target temperature of more than 8 hours) in all centres.
Hypotension (lasting more than 2 hours) and hypovolaemia occurred three times more frequently in the hypothermia group. Bradycardia associated with hypotension occurred four times more frequently in this group, and electrolyte disorders and hyperglycaemia were also found more frequently in the hypothermia group [46]. All of these complications are known side eff ects of hypothermia. Most are easily preventable with good intensive care and should not be regarded as inevitable consequences of hypothermia treatment. Since even very brief episodes of hypotension or hypovolaemia can adversely aff ect outcome in TBI, these and other issues may have signifi cantly aff ected the results of this trial [65][66][67]. One possible problem was that some of the participating centres had little or no previous experience in using hypothermia. Large centres familiar with cooling showed apparently favourable neurological outcomes whereas smaller centres showed poor outcomes.

Induction of hypothermia
Th e most widely accepted use of hypothermia is after cardiac arrest. Two RCTs in this patient group have shown signifi cant neurological improvements in patients who were treated with hypothermia many hours after injury and whose initial cardiac rhythm was ventricular fi brillation or ventricular tachycardia [68,69]. Subsequent data from a large study of patients after myocardial infarc tion suggest that infarct size was reduced in patients who were cooled to less than 35°C before coronary intervention [70], thus suggesting that faster cooling rates may be benefi cial to patient outcome.
Methods of cooling can be broadly divided into surface and core cooling techniques [71]. Th e above study used surface cooling devices alone and found that large numbers of patients did not reach the target temperature quickly enough before the start of the coronary intervention [70]. Despite advancing technology in surface cooling devices and the introduction of endovascular catheters for core cooling, average periods of 2 to 3 hours are still required to reach temperatures of 32°C to 34°C [71]. Th e currently available surface cooling devices are also relatively large and cumbersome. Th is coupled with the need for staff with specialist knowledge of the management of induced hypothermia may prevent its use outside of the intensive care unit [71].
A recent study examined the feasibility, speed, and complication rates of infusing refrigerated fl uids intravenously to quickly induce hypothermia in patients with various neurological injuries [71]. Results showed that a 1,500 mL infusion of 0.9% saline, administered over the course of 30 minutes in patients without cardiogenic shock, reduced core temperature from 36.9°C ± 1.9°C to 34.6°C ± 1.5°C at 30 minutes and to 32.9°C ± 0.9°C at 60 minutes. Continuous monitoring of arterial blood pressure, heart rhythm, central venous pressure, arterial blood gasses, and serum levels of electrolytes, platelets, and white blood cells showed no signifi cant adverse events [71].
When hypothermia develops, the body will imme diately try to counteract the temperature drop to maintain homeostasis [72]. One of the key mechanisms of heat production is shivering, which can lead to an increased oxygen consumption of 40% to 100%, which may be detrimental in this patient population. Sedation drugs are known to increase peripheral blood fl ow and, in turn, increase the transfer of heat from the core to the peripheries, thus reducing core temperature [72]. Th erefore, shivering may be counteracted by the administration of sedatives, anaesthetic agents, opiates, and/or paralysing agents [72].
It should be noted, however, that the capacity and eff ectiveness of the mechanisms of controlling body temperature decrease with age. Younger patients will there fore react earlier and with greater intensity than older patients. For this reason, induction of hypothermia in younger patients often requires high doses of sedation drugs to neutralise the counter-regulatory mechanisms [72].

Meta-analyses
Eight meta-analyses have been published between the years 2000 and 2009 [73][74][75][76][77][78][79][80]. Th ese include various numbers of trials, with varying quality of randomisation and blinding procedures. All have found a trend to positive eff ects of hypothermia on neurological outcome, although statistical signifi cance was reached in only two reviews: RR of improved neurological outcome of 0.78 (95% CI 0.63 to 0.98) [73] and RR of 0.68 (95% CI 0.52 to 0.89) [74]. Since submission of this manuscript, two Cochrane systematic reviews (issue 1 [79] and updated issue 2 [80] from 2009) have been published (Figure 1). Th e authors found that hypothermia may be eff ective in reducing death and unfavourable outcomes, but signifi cant benefi t was found only in low-quality trials. Th e high-quality trials found some non-signifi cant benefi t of hypothermia.
Th e meta-analysis of Peterson and colleagues [78] included eight trials that studied comparable patient groups at baseline. Hypothermia was shown to reduce mortality by 20%, although this was not statistically signifi cant (RR 0.80, 95% CI 0.59 to 1.09). Subgroup analysis showed that this eff ect was signifi cantly greatest when hypothermia was maintained for more than 48 hours (RR 0.51, 95% CI 0.33 to 0.79). Hypothermia was also associated with a non-signifi cant increase of 25% in neurological outcome when measured by the Glasgow Outcome Scale at 6 months (RR 1.25, 95% CI 0.96 to 1.62). Despite not reaching statistical signifi cance, results showed an increased likelihood of improved neurological outcome when cooling was maintained for more than 48 hours (RR 1.91, 95% CI 1.28 to 2.85). Another key fi nding of this meta-analysis is that hypothermia was of signifi cant benefi t only to those patients who had not received barbiturate therapy (RR 0.58, 95% CI 0.40 to 0.85).
A criticism of these analyses is that most failed to take account of important diff erences in patient groups (such as those with or without intracranial hypertension) and of diff erences in treatment protocols, except the use of hypothermia. Only one diff erentiated between studies that enrolled patients with normal ICP and those that enrolled patients with intracranial hypertension and that analysis found no neurological improve ment associated with hypothermia [77]. Only two assessed eff ects of treatment duration and speed of re-warming [73,74], concluding that cooling for more than 48 hours and rewarming rates of 24 hours, or 1°C/4 hours, were key factors in reducing mortality (RR 0.70, 95% CI 0.56 to 0.87) and improving neurological outcome (RR 0.79, 95% CI 0.63 to 0.98), respectively.
In summary, the evidence from previous research shows that induced hypothermia may be eff ective in patients with severe TBI and intracranial hypertension provided that the treatment is continued for long enough (between 48 hours and 5 days) and that patients are rewarmed slowly (1°C/4 hours). Experience with cooling also appears to be important if complications that may outweigh the benefi ts of hypothermia are to be avoided.

Rationale for Eurotherm3235Trial
Th e Eurotherm3235Trial will enrol TBI patients who have an ICP of more than 20 mm Hg for at least 5 minutes after fi rst-line treatments with no obvious reversible cause (for example, patient position, coughing, or inadequate sedation). Th ree stages of TBI management have been developed to support the trial and are based upon the best evidence available ( Figure 2) [73,81]. Th e Brain Trauma Foundation's recommended treatment threshold for treatment of ICP is 20 mm Hg [73]. Although early cooling after injury is considered to be benefi cial, this is off set by the failure to show benefi t from hypothermia in the absence of raised ICP. Enrolment to the Eurotherm3235Trial will therefore be allowed for up to 72 hours following injury. Th is potential delay in cooling will be compensated for, to an extent, by inducing hypothermia with 20 to 30 mL/kg of refrigerated 0.9% saline given intravenously over the course of 30 minutes. No maximum duration of cooling is specifi ed, and  Stages of therapeutic management after traumatic brain injury. These 'stages' have been developed for use in the Eurotherm3235Trial using evidence synthesis from the Brain Trauma Foundation [73] and the European Brain Injury Consortium [81]. CSF, cerebrospinal fl uid; PaCO 2 , arterial partial pressure of carbon dioxide; PaO 2 , arterial partial pressure of oxygen.

STAGE 3
Barbiturate therapy Decompressive Craniectomy hypothermia will continue until ICP is no longer dependent on temperature reduction to remain below 20 mm Hg. Patients will then be slowly re-warmed at a rate of 0.25°C per hour (1°C/4 hours). Th e experience from previous hypothermia trials underscores the potential diffi culties in using therapeutic hypothermia treatment for TBI. For this reason, and to reduce inter-centre variance, only centres experienced with the care of TBI patients and the use of hypothermia (after either cardiac arrest or TBI) will be initiated.

Conclusions
Many potential neuro-protective pharmacological interventions have been tested and have failed to show benefi t in TBI. Common reasons that have been cited include inadequate or low methodological quality preclinical studies and poor (and often underpowered) clinical study design. Hypothermia has extensive preclinical data supporting clinical testing and generally meets the STAIR (Stroke Th erapy Academic Industry Roundtable) recommendations [80]. Th e Eurotherm3235Trial will recruit 1,800 patients in 41 months and will be one of the largest TBI studies to date.

Competing interests
HLS is a trial manager and PJDA is the chief investigator of the Eurotherm3235Trial (http://www.eurotherm3235trial.eu).