All adult patients with systemic lupus erythematosus (SLE) (greater than or equal to 15 years old, n = 86) from a defined population (approximately 160,000 population at risk) were followed prospectively over 6 years. The study area comprised 2 health care districts served by only 1 hospital. Retrieval was based on clinical case finding and computerized diagnosis and laboratory registers. The incidence of the disease was 4.0 cases/100,000 adults/year and was stable during the 6 years, suggesting that completeness of retrieval was high. Point prevalence by the end of 1986 was 42 cases/100,000 population at risk and 5-year survival in the prospective group 97%. Immunologically the group was characterized by a high frequency of positive anti-dsDNA (73%), and a low frequency of rheumatoid factor positivity (10%). The frequency of ARA criteria (median, 6) was comparable with previous larger series of selected patients. Sixteen percent of the patients were males, and they had more serositis and renal manifestations than females. In view of the low mortality we studied the more sensitive outcome measures: disease activity, irreversible organ damage, and functional impairment. After the diagnosis year, disease flares occurred with a constant frequency of 0.2 flares/year/patient, even after long duration of disease. Patients with neuropsychiatric disease, history of drug reactions, and immunological abnormalities such as persistent hypocomplementemia, antibodies to dsDNA, and cardiolipin had a high frequency of relapse. In contrast, elderly patients with serositis during their first flare seldom relapsed. The number of gainfully employed individuals was normal. Neuropsychiatric disease and joint involvement were principal causes of long-standing functional impairment. Notably, patients with renal disease usually fared well, as reflected by preserved renal function and little functional by preserved renal function and little functional impairment. Disease duration and glucocorticoid treatment were major denominators for morbidity due to infections and vascular disease with the incidence of myocardial infarctions being 9 times more common than that in a Swedish control population. Prolonged glucocorticoid treatment was also related to mortality, which was predominantly due to cardiovascular or central nervous system disease. The present prospective and epidemiologically based study of outcome in SLE was made possible by a uniquely coordinated health care system, enabling complete identification of the patients within the study area.(ABSTRACT TRUNCATED AT 400 WORDS)