Perioperative fever and outcome in surgical patients with aneurysmal subarachnoid hemorrhage.

Objective: We examined the incidence of perioperative fever and its relationship to outcome among patients enrolled in the Intraoperative Hypothermia for Aneurysm Surgery Trial.

Methods: One thousand patients with initial World Federation of Neurological Surgeons grades of I to III undergoing clipping of intracranial aneurysms after subarachnoid hemorrhage were randomized to intraoperative normothermia (36 degrees C-37 degrees C) or hypothermia (32.5 degrees C-33.5 degrees C). Fever (> or =38.5 degrees C) and other complications (including infections) occurring between admission and discharge (or death) were recorded. Functional and neuropsychologic outcomes were assessed 3 months postoperatively. The primary outcome variable for the trial was dichotomized Glasgow Outcome Scale (good outcome versus all others).

Results: Fever was reported in 41% of patients. In 97% of these, fever occurred in the postoperative period. The median time from surgery to first fever was 3 days. All measures of outcome were worse in patients who developed fever, even in those without infections or who were World Federation of Neurological Surgeons grade I. Logistic regression analyses were performed to adjust for differences in preoperative factors (e.g., age, Fisher grade, initial neurological status). This demonstrated that fever continued to be significantly associated with most outcome measures, even when infection was added to the model. An alternative stepwise model selection process including all fever-related measures from the preoperative and intraoperative period (e.g., hydrocephalus, duration of surgery, intraoperative blood loss) resulted in the loss of significance for dichotomized Glasgow Outcome Scale, but significant associations between fever and several other outcome measures remained. After adding postoperative delayed ischemic neurological deficits to the model, only worsened National Institutes of Health Stroke Scale score, Barthel Activities of Daily Living index, and discharge destination (home versus other) remained independently associated with fever.

Conclusion: These findings suggest that fever is associated with worsened outcome in surgical subarachnoid hemorrhage patients, although, because the association between fever and the primary outcome measure for the trial is dependent on the covariates used in the analysis (particularly operative events and delayed ischemic neurological deficits), we cannot rule out the possibility that fever is a marker for other events. Only a formal trial of fever treatment or prevention can address this issue.