A total of 167 children and adolescents with insulin-dependent (Type 1) diabetes mellitus (97 males; age range 1.9-22.4 yrs) in a UK paediatric diabetic clinic were screened for coeliac disease using the IgA endomysial (EMA) test, or, in IgA deficient subjects, the IgG antigliadin (AGA) test. Antibody positive subjects were selected for small bowel biopsy, and confirmed coeliac cases started on a gluten free diet. Clinical features, height (Ht) standard deviation score (SDS), body mass index (BMI) SDS, HbA1c, insulin requirements' haemoglobin (Hb), mean red cell volume (MCV), serum folate and ferritin levels were evaluated at diagnosis and thereafter at 3-6 month intervals. A total of 156 subjects (93.4%) were antibody negative. Eleven (6.6%) were antibody positive (10 EMA/1 AGA; 6 males), of whom 9 had biopsies: 1 normal: 8 coeliac (4.8%; 5 males; 1 'classical'; 1 anaemia; 3 'atypical'; 3 asymptomatic). Seven coeliac subjects were followed during 12-24 months of dietary therapy. Pretreatment mean (range) Ht SDS = 0.08 (-1.66 to 1.88); BMI SDS = 0.32 (-0.82 to 1.29); HbA1c = 8.9 (6.2 to 11.3%); insulin dose = 0.98 (0.51 to 1.29) U kg(-1) day(-1). During treatment antibody status reverted to and remained negative, and symptoms resolved. By 24 months, there was a trend towards increased BMI SDS (mean (range) 1.31 (0.47 to 2.29), p = 0.248) and to reductions in HbA1c (8.1 (6.4-10.8), p = 0.697). Repeat small bowel biopsies were normal in 6 subjects (1 refused). No statistically significant changes occurred in any other parameters. In conclusion, serological screening is effective, although the therapeutic benefit of dietary therapy in asymptomatic cases remains uncertain.