Fever Prevention in Patients With Acute Vascular Brain Injury

Importance Fever is associated with worse outcomes in patients with stroke, but whether preventing fever improves outcomes is unclear. Objective To determine whether fever prevention after acute vascular brain injury is achievable and impacts functional outcome. Design, Setting, and Participants Open-label randomized clinical trial with blinded outcome assessment that enrolled 686 of 1176 planned critically ill patients with stroke at 43 intensive care units in 7 countries from March 2017 to April 2021 (last date of follow-up was May 12, 2022). Intervention Patients randomized to fever prevention (n = 339) were targeted to 37.0 °C for 14 days or intensive care unit discharge using an automated surface temperature management device. Standard care patients (n = 338) received standardized tiered fever treatment on occurrence of temperature of 38 °C or greater. Main Outcomes and Measures Primary outcome was daily mean fever burden: the area under the temperature curve above 37.9 °C (total fever burden) divided by the total number of hours in the acute phase, multiplied by 24 hours (°C-hour). The principal secondary outcome was 3-month functional recovery by shift analysis of the 6-category modified Rankin Scale, which is scored from 0 (no symptoms) to 6 (death). Major adverse events included death, pneumonia, sepsis, and malignant cerebral edema. Results Enrollment was stopped after a planned interim analysis demonstrated futility of the principal secondary end point. In total, 686 patients were enrolled, and 9 were consented but not randomized, leaving a primary analysis population of 677 patients (254 ischemic stroke, 223 intracerebral hemorrhage, 200 subarachnoid hemorrhage; 345 were female [51%]; median age, 62 years) with 433 (64%) completing the study through 12 months. Daily mean (SD) fever burden was significantly lower in the fever prevention group (0.37 [1.0] °C-hour; range, 0.0-8.0 °C-hour) compared with the standard care group (0.73 [1.1] °C-hour; range, 0.0-10.3 °C-hour) (difference, −0.35 [95% CI, −0.51 to −0.20]; P < .001). Between-group differences for the primary outcome by stroke subtype were −0.10 (95% CI, −0.35 to 0.15) for ischemic stroke, −0.50 (95% CI, −0.78 to −0.22) for intracerebral hemorrhage, and −0.52 (95% CI, −0.81 to −0.23) for subarachnoid hemorrhage (all P < .001 by Wilcoxon rank-sum test). There was no significant difference in functional recovery at 3 months (median modified Rankin Scale score, 4.0 vs 4.0, respectively; odds ratio for a favorable shift in functional outcome, 1.09 [95% CI, 0.81 to 1.46]; P = .54). Major adverse events occurred in 82.2% of participants in the fever prevention group vs 75.9% in the standard care group, including 33.8% vs 34.5% for infections, 14.5% vs 14.0% for cardiac disorders, and 24.5% vs 20.5% for respiratory disorders. Conclusions and Relevance In patients with acute vascular brain injury, preventive normothermia using an automated surface temperature management device effectively reduced fever burden but did not improve functional outcomes. Trial Registration ClinicalTrials.gov Identifier: NCT02996266