- Objective. Prognosis studies have indicated that survival of patients with systemic lupus erythematosus (SLE) has improved significantly. We investigate whether the apparent improvement in the survival of patients with SLE is associated with a reduction in the risk of death compared with the general population, or with changes over time in the distribution of various prognostic factors. Methods. The University of Toronto cohort of 720 patients with SLE followed between 1970 and 1994 was divided into 3 groups based on year of entry into study: Group A 1970-77, Group B 1978-85, Group C 1986-1994. Standardized mortality ratios (SMR) were calculated for each cohort. Prognostic factors for death occurring in the first 8 yr period after entry into the study were examined in each of the 3 cohorts. Analysis involved chi-squared tests for categorical values and unpaired t tests for continuous variables. Results. Group A comprised 183 patients, Group B 332 patients, and Group C 205 patients. An examination of the first 8 years of evaluation for each group revealed that the SMR decrease over time ranged from 10.1-fold greater than the general population in Group A, to 4.8-fold in Group B, to 3.3-fold in Group C. Prognostic factors for death varied over time, with vasculitis decreasing and hyperlipidemia increasing. Conclusion. Survival in SLE has improved over 24 yrs in the University of Toronto cohort more than the health of the general population has improved. This improved survival was not related to changing demographics, severity of lupus at presentation, major change in disease patterns, or new modalities of treatment.