Abstract 37: Refining of the Association of Fever with Functional Outcome in SAH

Abstract

Background: Fever has been associated with worse clinical outcomes in aneurysmal subarachnoid hemorrhage (aSAH). However, the impact of the cause, severity, and duration of fever is not clear. We conducted this study to evaluate the impact of fever and subfebrile load and fever characteristics on functional outcome. Methods: We collected detailed information on fever onset, cause, severity, and duration during the ICU stay in a cohort of 586 consecutive patients with aSAH. Fever was defined as core body temperature ≥ 38.3°C. Subfebrile measurements were those between 37 and 38.2°C. Febrile and subfrebile loads were defined as number of hours with fever or subfebrile measurements. Poor outcome was defined as modified Rankin score (mRS) > 2. Univariate and multivariate logistic regression models were developed to define predictors of outcome using various categorizations of fever cause, severity, and duration. Results: 532/586 patients (90.9%) had fever for a mean of 2.1±3.0 days. Fever started within 24 hours in 69 (11.8%) and within 72 hours in 110 (18.8%). Poor outcome occurred in 175 patients (29.9%). On univariate analysis, days of fever, febrile load, fever onset within 24 hours, and fever onset within 72 hours were associated with poor outcome (all p<0.001), but subfebrile load was not (p=0.58). On multivariate model constructed with all variables associated with outcome on univariate analyses (including age, WFNS grade, modified Fisher grade) days of fever remained independently associated with poor outcome (OR 1.14 of poor outcome per day of fever, 95% CI 1.06-1.22; p=0.0006) displacing all other fever measures from the final model. Conclusions: The great majority of patients with aSAH are febrile during their ICU stay. Early onset of fever, number of hours with fever, and especially days of fever are associated with poor functional outcome. Conversely, the number of hours with elevated but subfebrile temperature does not influence clinical outcome. These data suggest that prolonged fever should be avoided, but subfebrile temperatures do not justify intervention.