To compare six disease severity scoring systems as predictors of mortality in septic shock when used in the first 24 hrs of diagnosis. The six scoring systems tested were: Multiorgan Failure; the Acute Organ System Failure; the Acute Physiology and Chronic Health Evaluation (APACHE II); the Multisystem Organ Failure scoring system; the Mortality Prediction Model; and the grading of sepsis. Retrospective, case series, consecutive sample. Adult ICUs of three teaching hospitals. Seventy-one patients from 12 to 84 yrs, fulfilling specific criteria for the diagnosis of septic shock, who were admitted to the ICU during 15 consecutive months. The Multiorgan Failure scoring system, Acute Physiology and Chronic Health Evaluation (APACHE II), and Acute Organ System Failure scoring system were found, with our modifications, to be statistically significant predictors of mortality. Predictive data for these three scoring systems were as follows: Multiorgan Failure scoring system p = .008, mean number of points of survivors 5.2 +/- 1.5 (SD), mean number of points of nonsurvivors 6.3 +/- 1.5; APACHE II p = .013, mean number of points of survivors 21.1 +/- 5.9, mean number of points of nonsurvivors 24.6 +/- 6.0; and Acute Organ System Failure scoring system p = .011. None of the other three scoring systems showed significant predictive ability: Multisystem Organ Failure scoring system p = .072, Mortality Prediction Model p = 0.091, and the grading of sepsis p = .27. There was a significant (p = .004) difference in the survival rate of the three hospitals. The Multiorgan Failure scoring system, APACHE II, and the Acute Organ System Failure scoring system, with minor modifications, were found to be useful prognostic tools for patients with septic shock and allowed us to compare the performance and treatment programs of different ICUs.