In a nationwide incident case referent study we have evaluated vaccinations, early and recent infections and the use of medicines as possible risk determinants for Type 1 (insulin-dependent) diabetes mellitus in childhood. A total of 339 recently onset diabetic and 528 referent children, age 0-14 years, were included. Information about infections was collected from a mailed questionnaire and about vaccinations from childhood health care centres and schools. When vaccinations were considered as possible risk factors for diabetes, a significant decrease in relative risk estimated as odds ratio (OR) was noted for measles vaccination (OR = 0.69; 95% confidence limits 0.48-0.98). For vaccination against tuberculosis, smallpox, tetanus, whooping cough, rubella and mumps no significant effect on OR for diabetes was found. The odds ratios for Type 1 diabetes for children exposed to 0.1-2 or over 2 infections during the last year before diagnosis of diabetes revealed a linear increase (OR = 1.0, 1.96 and 2.55 for 0, 1-2 and over 2 infections, respectively). The trend was still significant when standardized for possible confounders such as age and sex of the children, maternal age and education and intake of antibiotics and analgetics. In conclusion, a protective effect of measles vaccination for Type 1 diabetes in childhood is indicated as well as a possible causal relationship between the onset of the disease and the total load of recent infections.