Objective: To determine the prevalence and mortality of ICU patients with severe sepsis in Germany, with consideration of hospital size. Design: Prospective, observational, cross-sectional 1-day point-prevalence study. Setting: 454 ICUs from a representative nationwide sample of 310 hospitals stratified by size. Data were collected via 1-day on-site audits by trained external study physicians. Visits were randomly distributed over 1 year (2003). Patients: Inflammatory response of all ICU patients was assessed using the ACCP/SCCM consensus conference criteria. Patients with severe sepsis were followed up after 3 months for hospital mortality and length of ICU stay. Measurements and results: Main outcome measures were prevalence and mortality. A total of 3,877 patients were screened. Prevalence was 12.4% (95% CI, 10.9-13.8%) for sepsis and 11.0% (95% CI, 9.7-12.2%) for severe sepsis including septic shock. The ICU and hospital mortality of patients with severe sepsis was 48.4 and 55.2%, respectively, without significant differences between hospital size. Prevalence and mean length of ICU stay of patients with severe sepsis were significantly higher in larger hospitals and universities (≤ 200 beds: 6% and 11.5 days, universities: 19% and 19.2 days, respectively). Conclusions: The expected number of newly diagnosed cases with severe sepsis in Germany amounts to 76-110 per 100,000 adult inhabitants. To allow better comparison between countries, future epidemiological studies should use standardized study methodologies with respect to sepsis definitions, hospital size, and daily and monthly variability.