The aim of this study was to determine the minimum prevalence of coeliac disease in a group of 459 diabetic children and adolescents. Six patients were already known to have coeliac disease. A total of 436 patients with type 1 diabetes mellitus aged 2-21 years and with age at onset at 2 months to 17 years at three paediatric departments agreed to participate in the study. All patients were tested for gliadin IgA antibodies with a commercial kit (Pharmacia Gluten IgA EIA). Later, serum was tested for reticulin IgA/IgG antibodies. Nineteen patients had elevated gliadin IgA levels (> 25 AU). Eighteen underwent jejunal biopsy. Ten had total or subtotal villous atrophy. These 10 patients were reticulin IgA-positive. Of 417 gliadin IgA-negative patients, 408 were reticulin IgA/IgG-negative. Of 6 reticulin IgA-positive patients, 3 had total or subtotal villous atrophy. All 3 had become gliadin IgA-positive at the time of biopsy. Among 3 reticulin IgG-positive patients with IgA deficiency, 2 had total villous atrophy: 1 was not willing to be biopsied. Patients with total or subtotal villous atrophy were judged as having coeliac disease and were recommended a gluten-free diet. Within 2 months, gliadin IgA levels were normal in patients adhering to the diet. Five patients have gone through a second jejunal biopsy to date with normal histology in all 5. The 15 newly diagnosed patients with coeliac disease plus 6 already known patients with coeliac disease and type 1 diabetes mellitus gave a minimum prevalence of coeliac disease in diabetic children and adolescents of 21/459 = 4.6%.