Background: In 1992, a consensus conference defined the terms systemic inflammatory response syndrome (SIRS), sepsis, severe sepsis and septic shock. Since then, numerous reports have validated the prognostic usefulness of these operative definitions. Aim: To evaluate if sepsis severity criteria, as defined by the Consensus Conference, can be applied to noninfectious SIRS. Patients and methods: Five hundred eighteen patients admitted to 5 intensive care units (ICU) from 4 hospitals were prospectively evaluated during a 3 months period. Patients that met at least one severity criteria were included. SIRS etiology, organ dysfunction and evolution were recorded in each patient. Results: One hundred two patients were included: 79 with sepsis (group I) and 23 with noninfectious SIRS (group II). ICU and hospital mortality were comparable (43 and 48% in sepsis compared to 43 and 51% in non infectious SIRS). The most common sources of sepsis were pneumonia and peritonitis. Group II patients had a wide variety of diseases. ICU stay, APACHE score and number of organs with dysfunction were not different among groups. Only the incidence of renal dysfunction was higher in the septic group. Conclusions: The Consensus sepsis severity criteria can be applied to noninfectious SIRS, defining a population subset with similar high mortality and organ dysfunction incidence, although with greatly heterogeneous etiologies.