UNLABELLED: Statins are used to prevent atherosclerosis because of their hypolipemic effects. In vitro testing and murine models suggest that statins may affect outcome in sepsis. Our meta-analysis of epidemiological studies in humans confirms that previous statin use appears to have a protective effect on infection rates and outcomes. The studies considered, however, were very different and heterogeneity was high, especially for mortality criteria. OBJECTIVE: To compare outcome for current statin users and nonusers hospitalized with fever, under pragmatic circumstances. DESIGN: Prospective cohort study. SETTING: University Hospital of Brest (France). PARTICIPANTS: Febrile (>38 degrees C) patients older than 40 years, admitted to the hospital on an emergency basis in 2005. Patients' outcome was compared according to their exposure to statins (current user or nonuser). MAIN OUTCOME MEASURES: Mortality, length of hospitalization, admission to ICU and to convalescent homes RESULTS: Of 40 343 patients referred for admission by the emergency department in 2005, 964 patients older than 40 years had a fever higher than 38 degrees C and were included in the study. Statin-user status, however, was available for only 921. The ICU admission rate (relative risk: 4.69; 95% CI: 2.42-9.08) was significantly higher (p=0.009) among statin users (n=139) than nonusers (n=782). The groups had similar lengths of hospitalization and similar rates of mortality and of convalescent home admissions. CONCLUSION: Our study does not confirm that statin use affects infection outcome and suggests that previous findings may be associated instead with more cautious care or nonspecific cardiovascular prevention rather than with a specific infection-related effect. Thus, these data support the continuation of statin therapy during physiological aggression, including infection.