Cooling for cerebral protection during brain surgery

Patients undergoing neurosurgery are at risk of cerebral ischaemia with resultant cerebral hypoxia and neuronal cell death. This can increase both the risk of mortality and long term neurological disability. Induced hypothermia has been shown to reduce the risk of cerebral ischaemic damage in both animal studies and in humans who have been resuscitated following cardiac arrest. This had lead to an increasing interest in its neuroprotective potential in neurosurgical patients. This review was originally published in 2011 and did not find any evidence of either effectiveness or harm in these patients. This updated review was designed to capture current evidence to readdress these issues.

To evaluate the effectiveness and safety profile of induced hypothermia versus normothermia for neuroprotection in patients undergoing brain surgery. Effectiveness was to be measured in terms of short and long term mortality and functional neurological outcomes. Safety was to be assessed in terms of the rate of the adverse events infection, myocardial infarction, ischaemic stroke, congestive cardiac failure and any other adverse events reported by the authors of the included studies.

For the original review, the authors searched the databases Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE (OvidSP), EMBASE (OvidSP) and LILACS to November 2010. For the updated review all these databases were re-searched from November 2010 to May 2014.For both the original and updated versions, grey literature was sought by searching reference lists of identified studies and relevant review articles, and conference proceedings. No language restrictions were applied.

As in the original review, we included randomized controlled trials (RCTs) of induced hypothermia versus normothermia for neuroprotection in patients of any age and gender undergoing brain surgery, which addressed mortality, neurological morbidity or adverse event outcomes.

Three review authors independently extracted data and two independently assessed the risk of bias of the included studies. Any discrepancies were resolved by discussion between authors.

In this updated review, one new ongoing study was found but no new eligible completed studies were identified. This update was therefore conducted using the same four studies included in the original review. These studies included a total of 1219 participants, mean age 40 to 54 years. All included studies were reported as RCTs. Two were multicentred, together including a total of 1114 patients who underwent cerebral aneurysm clipping, and were judged to have an overall low risk of bias. The other two studies were single centred. One included 80 patients who had a craniotomy following severe traumatic brain injury and was judged to have an unclear or low risk of bias. The other study included 25 patients who underwent hemicranicectomy to relieve oedema following cerebral infarction and was judged to have an unclear or high risk of bias. All studies assessed hypothermia versus normothermia. Overall 608 participants received hypothermia with target temperatures ranging from 32.5 °C to 35 °C, and 611 were assigned to normothermia with the actual temperatures recorded in this group ranging form 36.5 °C to 38 °C. For those who were cooled, 556 had cooling commenced immediately after induction of anaesthesia that was continued until the surgical objective of aneurysm clipping was achieved, and 52 had cooling commenced immediately after surgery and continued for 48 to 96 hours.Pooled estimates of effect were calculated for the outcomes mortality during treatment or follow-up (ranging from in-hospital to one year); neurological outcome measured in terms of the Glasgow Outcome Score (GOS) of 3 or less; and adverse events of infections, myocardial infarction, ischaemic stroke and congestive cardiac failure. With regards to mortality, the risk of dying if allocated to hypothermia compared to normothermia was not statistically significantly different (risk ratio (RR) 0.87, 95% confidence interval (CI) 0.59 to 1.27, P = 0.47). There was no indication that the time at which cooling was started affected the risk of dying (RR with intraoperative cooling 0.95, 95% CI 0.60 to 1.51, P = 0.83; RR for cooling postoperatively 0.67, 95% CI 0.34 to 1.35, P = 0.26). For the neurological outcome, the risk of having a poor outcome with a GOS of 3 or less was not statistically different in those who received hypothermia versus normothermia (RR 0.80, 95% CI 0.61 to 1.04, P = 0.09). Again there was no indication that the time at which cooling was started affected this result. Regarding adverse events, there was no statistically significant difference in the incidence in those allocated to hypothermia versus normothermia for risk of surgical infection (RR 1.20, 95% CI 0.73 to 1.97, P = 0.48), myocardial infarction (RR 1.86, 95% CI 0.69 to 4.98, P = 0.22), ischaemic stroke (RR 0.93, 95% CI 0.82 to 1.05, P = 0.24) or congestive heart failure (RR 0.85, 95% CI 0.60 to 1.21, P = 0.38). In contrast to other outcomes, where time of application of cooling did not change the statistical significance of the effect estimates, there was a weak statistically significant increased risk of infection in those who were cooled postoperatively versus those who were not cooled (RR 1.77, 95% CI 1.05 to 2.98, P = 0.03). Overall, as in the original review, no evidence was found that the use of induced hypothermia was either beneficial or harmful in patients undergoing neurosurgery.

We found no evidence that the use of induced hypothermia was associated with a significant reduction in mortality or severe neurological disability, or an increase in harm in patients undergoing neurosurgery.