The frequency of the hemolytic uremic syndrome, characterized by microangiopathic hemolytic anemia, thrombocytopenia, and renal failure, is increasing. Although Escherichia coli serotype 0157:H7 has been implicated as a causative agent, more information is needed about the basic epidemiology and clinical aspects of this syndrome. We conducted a retrospective population-based study of hemolytic uremic syndrome in Minnesota residents less than 18 years of age from 1979 through 1988 to assess trends in disease occurrence, describe the clinical illness, and identify predictors of disease severity and outcome. We also conducted a case–control study of patients with onsets of illness from 1986 through 1988 to examine risk factors. One hundred seventeen patients were identified. The mean annual incidence increased from 0.5 case per 100,000 child-years among children less than 18 in 1979 (6 cases) to 2.0 cases per 100,000 in 1988 (26 cases) (P = 0.000004). E. coli 0157:H7 was isolated from 13 of 28 patients (46 percent) who had stool specimens submitted for testing. For those who presented with typical hemolytic uremic syndrome, an elevated polymorphonuclear-leukocyte count on hospital admission, a shorter duration of prodrome, and the presence of bloody diarrhea were predictive of severe disease. In the case–control study, the patients were more likely to attend large day-care centers (more than 50 children) than were the controls (odds ratio, 10.2; P = 0.03), suggesting that day-care attendance may be a risk factor. On the basis of the population-attributable risk, however, this factor could account for no more than 16 percent of the cases. This study provides evidence for an increase in the incidence of hemolytic uremic syndrome, which is probably related to an increased incidence of E. coli 0157:H7 infections. Hemolytic uremic syndrome has become an important pediatric and public health problem. (N Engl J Med 1990; 323:1161–7).