The outcome of patients admitted to intensive care units is known to be influenced by such factors as age, previous health status, severity of disease, and diagnosis. To estimate the outcome of such patients with systemic rheumatic diseases and to determine if the severity of these diseases unfavourably influences the prognosis at the time of admission to a medical intensive care unit, the clinical courses of all patients with systemic rheumatic disease admitted to two medical intensive care units between January 1978 and December 1988 were studied retrospectively. Sixty nine patients with systemic lupus erythematosus (n = 16), necrotising vasculitis (n = 19), rheumatoid arthritis (n = 19), and other systemic rheumatic diseases (n = 15) were included. The mean (SD) age on admission into the medical intensive care unit was 53 (17) years and the mean simplified acute physiological score was 12 (5.5). The principal diagnoses on admission were infectious complications (29/69 patients) and acute exacerbation of the systemic rheumatic disease (19/69 patients). The death rate in the medical intensive care unit was 33% (23/69 patients) and was similar to that of a non-selected population with comparable simplified acute physiological score. The death rate in hospital was 42% (29/69 patients). Infection was the main cause of death in the medical intensive care unit (19/23 patients) and the infection was mainly acquired in the unit. Only the simplified acute physiological score on admission was a statistically significant prognostic factor: the simplified acute physiological score in patients who died was 15 (5.2) v 9.9 (4.7) for survivors. Long term outcome analysis showed that 83% (33/40 patients) of patients were still alive after admission to the medical intensive care unit with a follow up time between two months and nine years (mean 38 months). The death rate was relatively high and was mainly due to nosocomial infections. It was not different, however, from that of nonselected patients and the long term prognosis was highly favourable. This shows that the complications are often reversible, particularly infectious applications, and justifies admission to the medical intensive care unit of this group of patients.