At cellular level, decreased cerebral oxygen supply manifests as reduced neuronal aerobic metabolism, leading to decreased cellular production of the high-energy substrate adenosine triphosphate (ATP). ATP depletion results in the dysfunction of the ATP-dependent Na+/K+ ion exchange pump, leading to massive entry of sodium and water into the cell, and intracellular cytotoxic edema. Potassium efflux and membrane depolarization lead to opening of voltage-gated Ca++ channels and intracellular calcium influx. Additionally, endothelial dysfunction occurs in the microvasculature, this leading to increased blood-brain barrier permeability, vasogenic cerebral edema formation, the formation of microthrombi, and limited cerebral blood flow exacerbating cellular ischemia.5,6

Reperfusion of the ischemic cerebrovascular bed triggers a series of mechanisms leading to secondary brain injury. Increased intracellular Ca++ caused by the primary injury leads to the release of glutamate, an excitatory neurotransmitter that binds to cell membrane receptors causing further intracellular Ca++ influx. The accumulation of intracellular Ca++ activates lytic enzymes such as caspase, proteases, and phospholipases, and causes mitochondrial dysfunction, and the release of pro-apoptotic proteins and reactive oxygen species. This results in additional neuronal damage and energy failure, leading to apoptosis and cell death.

Another consequence of reperfusion injury is activation of the innate immune system, leading to an inflammatory response. Resident macrophages, called microglia, and circulating monocytes adhere to cerebral microvasculature endothelial cells and infiltrate neuronal tissue, secreting proinflammatory cytokines. Complement cascade activation also occurs, further propagating inflammatory injury.

In 2002, 2 randomized clinical trials (RCTs)18,19 were published, showing improved survival with favorable neurological outcomes after having induced TH between 32 °C and 34 °C for 12–24 h in patients with witnessed out-of-hospital cardiac arrest (OHCA) and initial shockable rhythm. However, these promising results have not been confirmed in subsequent studies. In 2013, the first large-sample RCT—the TTM1 (20)—was published, including 950 unconscious patients after OHCA with shockable and non-shockable initial rhythms, without noticeable differences in mortality or functional outcomes with the application of TH. Unlike the studies from 2002,18,19 the researchers of the TTM1 used a protocolized approach to establish neurological prognosis, and temperature in both arms was actively controlled. However, the main limitations of the TTM1 study were the time to reach the target temperature (approximately 9 h on average to reach 33 °C) and the high rate of withdrawal of life-sustaining therapy based on neurological prognosis.

Then, in 2019, the HYPERION trial21 included 584 comatose patients after in- and out-of-hospital cardiac arrest with non-shockable rhythm and compared TH (33 °C) with thermal control (37 °C) for 24 h. No mortality differences were reported (81.3% and 83.2%, respectively), but in survivors, the rate of patients with favorable neurological outcomes was higher in the hypothermia group (10.2% vs 5.7%, P = .04). We should mention that the HYPERION trial had a fragility index of 1, suggesting that if only 1 patient had a different outcome (changing from a good neurological outcome to a bad one), the findings would not have reached statistical significance.

The most rigorous and largest-sample RCT (TTM2) was published 2 years later.22 In this study, 1850 adults in a coma after out-of-hospital cardiac arrest of presumed cardiac cause or unknown cause were randomized to receive TH (33 °C for 28 h) or normothermia. No differences in mortality rates (50% vs 48%, P = .37) or functional outcomes were reported. No differences were reported wither in clinical outcomes based on time to return of spontaneous circulation (ROSC) (> or <25 min) and the initial rhythm. Among the main limitations of the TTM2 study, we should mention that although the protocol required reaching the target temperature in 90 min or less, half of the study population needed 7 h from ROSC to TH, perhaps due to the low intravascular cooling rate used in the trial (29%). Additionally, nearly 50% of the patients from both groups showed inadequate fever control after 72 h, and, once again, a high rate of withdrawal of life-sustaining therapy based on neurological prognosis established according to the trial protocol was reported. Another argument used to justify the lack of clinical benefit of hypothermia in the TTM2 trial is the high percentage of patients sedated with propofol, a potentially neurotoxic drug associated with mitochondrial dysfunction, one of the causes of hypoxic-ischemic brain injury.23

More recently, the CAPITAL CHILL trial24 addressed the question of whether an even lower temperature would be more beneficial for neurological recovery. The RCT included 367 comatose patients after out-of-hospital cardiac arrest regardless of rhythm and compared moderate (31 °C) to mild hypothermia (34 °C) for 24 h. No differences were reported in the composite outcome of death or unfavorable functional outcomes at 180 days. We should mention that this was a single-center study and probably did not have enough statistical power for the detection of any clinically significant differences. Moreover, this study24 compared 2 hypothermia strategies (moderate and mild) and did not include a group not subjected to hypothermia, which means that no conclusions can be drawn from the study on the efficacy of hypothermia therapy (vs normothermia).

Most patients included in studies conducted so far, except for the HYPERION study,21 had experienced out-of-hospital cardiac arrest with an initial shockable heart rhythm and a known or presumed primary cardiac cause,20,22,25 which suggests that in other categories of patients, temperature control with hypothermia might be more effective. TTM researchers conducted a meta-analysis of individual patient data from the TTM-1 and TTM-2 studies to assess whether the effects of hypothermia differ depending on the circumstances of the cardiac arrest or the characteristics of the patients.26 Patients from the 36 °C group in TTM-1 were combined with those from the normothermia group in TTM-2 and compared to patients from the 33 °C groups in both TTM studies. The primary endpoint, 6-month all-cause mortality rate (47.9% vs 49.4%, P = .41), and results in predefined subgroups including age, sex, initial heart rhythm (shockable or non-shockable), time to ROSC, and circulatory shock at admission showed no significant differences between the 2 groups. On the other hand, a recent study27 conducted among 249 resuscitated patients after in-hospital cardiac arrest in Germany, confirmed that inducing hypothermia (32–34 °C) produced no benefits on mortality. However, other observational studies have identified a benefit in neurological recovery associated with 33 °C TH vs 36 °C in patients with greater severity of post-arrest neurological injury.28–30

In 2021, a systematic review and meta-analysis31 analyzed the safety and efficacy profile of maintaining different temperature targets after out-of-hospital cardiac arrest. A total of 10 RCTs with 4218 patients were included, and their results confirmed that TH at 31−32 °C, 33−34 °C, and 35−36 °C did not improve survival with good functional outcomes vs normothermia at 37–37.8 °C, while a higher incidence of arrhythmias was observed among patients treated with hypothermia at 31−32 °C and 33−34 °C. That same year, another meta-analysis conducted by the International Liaison Committee on Resuscitation (ILCOR)32 identified 9 RCTs, 6 of which were included in the meta-analysis, and concluded that treatment aimed at achieving a target temperature of 32°–34 °C did not improve survival or produced favorable neurological outcomes (low level of evidence). Three trials evaluated different hypothermic temperature targets and found no differences in outcomes between 33 °C and 36 °C or between 32 °C, 33 °C, and 34 °C (low level of evidence). Based on these results, the ILCOR ALS Task Force published its recommendations,33 summarized in Table 1, suggesting actively preventing fever (defined as a temperature > 37.7 °C) by targeting temperatures ≤ 37.5 °C (weak recommendation, low level of evidence), rather than recommending hypothermia to treat patients remaining comatose after post-cardiac arrest ROSC.

In 2022, the European Resuscitation Council (ERC) and the European Society of Intensive Care Medicine (ESICM),34,35 in line with the ILCOR consensus document, recommended continuous monitoring of core temperature and actively preventing fever (>37.7 °C) for, at least, 72 h in patients remaining comatose after cardiac arrest (Table 1). One possible reason for the lack of clinical benefit of hypothermia after cardiac arrest is that in the studies conducted so far, the target temperature was reached several hours after ROSC and potentially outside the therapeutic window. In a RCT,36 the pre-hospital infusion of up to 2 l of physiological saline at 4 °C after ROSC reduced the time needed to reach a temperature of 34 °C by more than an hour. However, this strategy did not improve survival or neurological status in patients resuscitated after pre-hospital ventricular fibrillation, or in patients without ventricular fibrillation, and was associated with significantly higher rates of recurrent arrest and pulmonary edema on the first thoracic x-ray.

The optimal duration of TH remains a topic of discussion. The only trial included in the ILCOR review that addressed the issue of temperature control duration did not show any differences in outcomes between temperature control at 32°–34 °C for 24 h vs 48 h after cardiac arrest.25 The duration of active fever prevention does not seem to have an impact on the clinical outcomes either. In a much more recent trial37 with 789 patients in a coma after out-of-hospital cardiac arrest presumably of cardiac causes, temperature control was compared with a target of 36 °C for 24 h, followed by control at 37 °C for an additional 12 h or 48 h, or until the patient regained consciousness. The occurrence of death or severe brain disability or coma within 90 days did not significantly differ after 36 h or 72 h of active fever prevention.

Finally, the most recent Cochrane review of 12 studies with 3956 patients, concludes that TH with a target temperature of 32–34 °C may improve neurological outcomes after cardiac arrest, although with a low level of evidence.38

Currently, the ICECAP trial (NCT04217551) is in the recruitment phase.39 This is a multicenter, randomized, adaptively allocated clinical trial to determine whether increasing the duration of TH is associated with a higher rate of good neurological outcomes in survivors in a coma after cardiac arrest.

Taking everything into account, the current evidence (Table 2) leaves unanswered the question of what the optimal temperature management is for patients in cardiac arrest with initial non-shockable rhythms, for patients in whom cardiac arrest is due to non-cardiac causes, and for those with more severe neurological injuries. This raises the hypothesis that the clinical benefit of TH after cardiac arrest may depend on factors that we still do not fully understand.

The authors of this review propose prioritizing active fever prevention (≤37.5 °C), optimization of hemodynamics, oxygenation, ventilation, and, perhaps most importantly, avoiding early prognostic assessment and premature withdrawal of life-sustaining therapy, general measures that have consistently been associated with better clinical outcomes.40–42

In 2009, a Cochrane review43 including 5 RCTs found no statistically significant effects of pharmacological or physical temperature reduction therapy on reducing the risk of death or dependence. Both interventions were associated with a non-significant increase in infection rates.

Another subsequent meta-analysis,44 which included 3 studies in 131 patients with large hemispheric infarctions, suggested that TH was not associated with a decrease in mortality; however, it was associated with improved neurological outcomes in survivors, albeit with a higher risk of adverse events during treatment.

A recent meta-analysis evaluated current literature on the efficacy of TH to treat AIS.45 Twelve studies with a total of 778 patients were included. Functional independence did not differ between groups, although 5 studies showed a trend towards better functional outcomes with hypothermia (OR, 1.57, 95%CI, 1.01−2.44; P = .05). General complications were higher with hypothermia (RR, 1.18, 95%CI, 1.06−1.32; P < .01).

The European Stroke Organization (ESO) guidelines46 do not recommend induction of hypothermia as a way improve functional outcome and/or survival in patients with acute ischemic stroke, with a weak grade of recommendation and a very low level of evidence. Similarly, the American Heart Association/American Stroke Association guidelines47 state that, in patients with acute ischemic stroke, the benefit of TH is still under discussion, with a moderate grade of recommendation and a moderate level of evidence. These guidelines recommend identifying the causes of hyperthermia (temperature > 38 °C) and administering antipyretics to lower the temperature in AIS patients, with a strong grade of recommendation and a low level of evidence (Table 1).

The evidence on temperature control in AIS patients is summarized in Table 3.43–45,48–51

Although some studies have shown a trend towards better functional outcomes in AIS patients treated with TH, no overall beneficial effect has been observed, and its use is associated with more complications. Therefore, based on the analysis of the evidence presented here, the application TH in this clinical context is ill-advised, and once again we believe that the priority should be the prevention and active treatment of fever.

Clinical trials on TH in the management of TBI have yielded disparate results, and meta-analyses have drawn contradictory conclusions. Inconsistent results may be due to differences in the study design. RCTs of TH in the management of severe TBI can be categorized into those where TH was used to treat elevated intracranial pressure (ICP) and those where TH was used as neuroprotection to halt the biochemical cascade following injury.

In 2015, the Eurotherm3235 trial52 evaluated 387 patients with severe TBI and intracranial hypertension (ICH) with failed first-line ICP control measures and compared TH (32–35 °C for ≥48 h) to standard care. Increased mortality (31% vs 22%; HR, 1.45, 95%CI 1.01–2.10, P = .047) and more unfavorable functional outcomes 6 months after the TBI (OR, 1.53; 95%CI, 1.02–2.30; P = .04) were observed in the TH group. Another more recent study53 compared prolonged mild TH (34–35 °C for 5 days) to normothermia in 302 patients with severe TBI (Glasgow Coma Scale 4–8) and initial ICH (ICP ≥ 25 mmHg). No inter-group differences were detected in neurological outcomes or mortality. In patients with initial ICP ≥ 30 mm Hg, TH significantly increased favorable outcomes (60.8% vs 42.7%; OR, 1.86; 95%CI, 1.03–3.36; P = .039).

The largest RCT on TH in TBI patients to date, the POLAR trial,54 found no neurological benefit with the application of prophylactic TH (33–35 °C) for 72 h in 511 patients with Glasgow Coma Scale score < 9. This study was included in a systematic review1 along with 6 other RCTs with a total of 1843 TBI patients addressing the relationship between TH (33–35 °C) and functional outcomes and mortality, with again discordant results.

The most recent Brain Trauma Foundation guidelines on the management of severe TBI do not recommend early (within 2.5 h) or short-term (48 h post-TBI) prophylactic TH to improve the outcomes of patients with diffuse injuries, with a II B level of evidence.55

On the other hand, the International Consensus Conference on Brain Injury in Seattle (SIBICC) recommends mild TH (35−36 °C) as a level 3 intervention to reduce ICP in patients with ongoing ICH, once other level 1 and 2 interventions have been exhausted, with a IV level evidence (expert opinion)56,57 (Table 1).

The evidence available on temperature control in TBI patients is summarized in Table 4.52–54,58–61

TH in patients with severe TBI, used as neuroprotection or for ICP treatment, does not improve functional outcomes or mortality, and may even be harmful. Therefore its use is ill-advised.