The influence of epidemic influenza on hospitalizations because of influenza, pneumonia and diabetic acidosis in patients with diabetes mellitus was investigated. Data on the weekly incidence of influenza-like illness were obtained from the Continuous Morbidity Registration and the cumulative data on hospitalizations in short-stay hospitals were obtained from the National Medical Registration. Patients with duodenal ulcer were used as a control population. Epidemic elevations of influenza infections were observed in 1976 and 1978. The estimated relative risk for hospitalization because of influenza infection was 1.1 and 1.0 for the two non-epidemic years 1977 and 1979, respectively. For the epidemic years 1976 and 1978 this risk was calculated to be 5.7 and 6.2, respectively. An increased relative risk was also noted for pneumonia; being 25.6 for both epidemic years. The estimated relative risk of dying during hospitalization rose from 30.9 in 1977 to 91.8 in 1978. The number of hospitalizations for ketoacidosis was 50% higher in 1978 than in the other three years. During the epidemic years, 25.7% of patients hospitalized for pneumonia died, while this percentage was 14.6% in the non-epidemic years (P less than 0.05). Differences in mortality due to diabetic acidosis were similar: 25.4% in epidemic and 14.7% in non-epidemic years (P less than 0.01). During the 1978 epidemic, one out of every 1300 patients with diabetes mellitus was hospitalized because of pneumonia. It is estimated that 1 of every 260 patients with IDDM was hospitalized for diabetic acidosis. It is concluded that patients with diabetes mellitus have indeed a very high influenza-associated morbidity.