Gianfranco Meloni, Angelo Dore, Giuseppe Fanciulli, Francesco Tanda, Gian Franco Bottazzo

1607 schoolchildren (845 boys and 753 girls; aged 6–14 years) were recruited in seven towns in the province of Sassari (Sardinia)(during school year 1993–94), as part of a study to assess the prevalence of endemic goitre. Tests were initially done for IgA and IgG anti-gliadin antibodies (AGA), followed by IgA anti-endomisium antibodies (AEA) in samples positive for at least one AGA test. AEA were also assayed in 53 random serum samples negative for both IgA and IgG AGA. To exclude a selective IgA deficiency (S-IgA-D), total serum IgA were measured in 132 serum samples positive only for IgG AGA, but negative for AEA (figure). Overall, 23 (1· 43%) children had abnormal values of at least two of three coeliac disease markers and/or S-IgA-D. Jejunal biopsy was done on 18 children. Apart from the child with S-IgA-D, 17 children had histologically proven coeliac disease (table). Two children had no symptoms, but the parents of the others reported that children had intermittent diarrhoea, mild abdominal pain. or other mild symptoms. Five of 17 children had low ferritin values.

The heterodimer DQA1* 0501/DQB1* 0201 which predisposes to coeliac disease, was found in 13 of 16 children, while the other three had DRB* 0405 or DRB1* 07, which are associated with coeliac disease. By contrast with a previous study in Sardinia, 1 a double dosage of genetic susceptibility to coeliac disease was found in 13 of 16 affected children, including the three who were homozygous. The overall prevalence of silent coeliac disease we found was 10· 6 per 1000 children ([95% CI 0· 56–1· 56] it could be higher if the four children positive for IgG AGA+ AEA, also underwent biopsy [figure]). Similar studies in Italy have shown a lower prevalence of silent coeliac