Recent studies suggest that macrolides may have beneficial effects for patients at risk for certain infections. The current authors examined the effect of macrolide therapy on 30- and 90-day mortality for patients with severe sepsis caused by pneumonia. A retrospective cohort study was conducted at two tertiary teaching hospitals. Eligible subjects were admitted with a diagnosis of, had chest radiography consistent with, and had a discharge diagnosis of pneumonia and clinical criteria of severe sepsis. Subjects were considered to be on macrolides if they received at least one dose within 48 h of admission. Severe sepsis was present in 237 (30.1%) subjects, out of whom 104 (43.9%) received macrolides. Mortality was 20.3% at 30 days and 24.5% at 90 days. In the multivariable analysis, the use of macrolide was associated with decreased mortality at 30 days (hazard ratio (HR) 0.3, 95% confidence interval (CI) 0.2-0.7) and at 90 days (HR 0.3, 95% CI 0.2-0.6) in patients with severe sepsis and in patients with macrolide-resistant pathogens (HR 0.1, 95% CI 0.02-0.5). Macrolide use was associated with decreased mortality in patients with severe sepsis due to pneumonia and macrolide-resistant pathogens. Confirmatory studies are needed to determine whether macrolide therapy may be protective for patients with sepsis.