ABSTRACT
OBJECTIVE: Therapeutic hypothermia (TH) is defined as the designed reduction of the human body's core temperature to 32°C-35°C for a period of 24-48 hours. TH has been studied extensively in many diseases related to critical care illness. This meta-analysis assesses the effect of TH on mortality across different indications in medical, neurological and cardiothoracic care. DATA SOURCES: The online databases Embase, Ovid MEDLINE, TRIP and CINAHL were searched for eligible studies published between 1940 and October 2018. STUDY SELECTION: Randomised clinical trials of induced TH in adults for any indication. DATA EXTRACTION: Information about baseline characteristics of patients, mortality outcomes, cooling strategy and target temperature achieved in hypothermia and normothermia groups was collected. DATA SYNTHESIS: Eighty studies, with a total of 13 418 patients, were included in this meta-analysis: 22 studies for traumatic brain injury, six studies for stroke, five studies for out-of-hospital cardiac arrest (OHCA), 34 studies for intraoperative cardiopulmonary bypass, and 13 studies for other diseases. A total of 6901 patients (51.4%) were randomly allocated to the TH group and 6517 patients (48.6%) were randomised to the normothermia control group. The unadjusted analysis showed no significant difference in mortality across different critical care illnesses. However, after adjusting for population, gender, age and temperature, only the OHCA group showed a small statistically significant difference favouring TH, but this had a questionable clinical significance. CONCLUSIONS: This meta-analysis suggests that after decades of extensive research, TH has yet to show a beneficial effect on mortality across different critical care diseases.
