Antibiotic escalations are frequently guided by fever persistence. Unnecessary antibiotic escalation is associated with resistance induction. We examined whether fever persistence is associated with adverse outcomes among medical inpatients with sepsis. In a single-center prospective cohort study, we included consecutive medical inpatients with suspected or documented bacterial infections. Data were collected on days 0, 2, 4, and 30 days from episode onset. We examined the association between fever persistence at 4 days and 30-day mortality on univariate and multivariate analysis. Inappropriate empirical antibiotic treatment (IAET) was defined for patients with microbiologically documented infections (MDIs). Odds ratios (ORs) are presented with 95% confidence intervals (CIs). A total of 1,621 patients were included. Among patients with MDIs, 38/206 (18.4%) given appropriate empiric therapy had continued fever on day 4, compared to 64/231 (27.7%) of patients receiving IAET, OR 0.59, 95% CI 0.37-0.93. Fever persistence was not associated with mortality after adjustment for other risk factors. Among patients with presumed sepsis who did not have MDIs, persistent fever was significantly associated with 30-day mortality on a multivariate analysis, adjusted OR 2.77 (95% CI 1.78-4.31). Other risk factors for mortality included older age, nosocomial infections, malignancy, dyspnea, shock, decreased albumin, and elevated creatinine. For patients with MDIs, fever persistence for up to 4 days is a marker of IAET, but is not associated with mortality, and should not, in itself, trigger antibiotic escalation. For patients without MDIs, fever persistence should trigger careful re-evaluation, as it is associated with mortality.