Niven DJ, Stelfox HT, Laupland KB.; J Crit Care. 2013 Jun;28(3):303-10.
PURPOSE: To determine whether fever control with antipyretic therapy effects the mortality of febrile critically ill adults.
METHODS: Systematic review using MEDLINE, EMBASE, Cochrane Central Register for Controlled Trials, CINAHL, Google Scholar, and 2 clinical trial registries from inception to April 2012. Randomized clinical trials comparing treatment of fever with no treatment or comparing different thresholds for fever control in adults without acute neurological injury admitted to intensive care units (ICUs) were selected for review. The effect of fever control on all-cause ICU-mortality was determined using a random effects meta-analysis.
RESULTS: Five randomized clinical trials in 399 patients were included. The temperature threshold for treatment in the intervention group was commonly 38.3°C to 38.5°C, whereas it was typically 40.0°C for controls. Four studies used physical measures and 3 used pharmacologic measures for temperature control. There was no significant heterogeneity among the included studies (I(2) = 12.5%, P = .3). Fever control did not significantly effect ICU mortality with a pooled risk ratio of 0.98 (95% confidence interval 0.58-1.63, P = .9).
CONCLUSIONS: This meta-analysis found no evidence that fever treatment influences mortality in critically ill adults without acute neurological injury. However, studies were underpowered to detect clinically important differences.