Search

Summary Background The relative importance of genetic and environmental factors in causing insulin-dependent diabetes mellitus (IDDM) is unknown. We studied this question by assessing the incidence of the disease in children, born in a region with a low incidence of IDDM (Lazio), but whose parents came from a region with high incidence (Sardinia). Methods We identified all IDDM cases that occurred between 1989 and 1994. We used as the denominator the number of children aged 0–14 born in Lazio of Sardinian parents to calculate incidence. We compared this rate with the incidences of IDDM in the populations of Lazio and Sardinia. Findings The age-adjusted incidence of IDDM in Sardinian-heritage children born and living in Lazio was 33·8 per 100 000 per year (95% CI 7·0–99·0) for those with two Sardinian parents, and 15·9 (8·7–26·6) for those with only one parent from Sardinia. The former incidence was not different from that recorded in Sardinia (34·4, 31·3–37·9), but was fourfold that of Lazio-heritage children (7·9, 7·1–8·8). Interpretation Our results show that two different ethnic groups living in the same region have a fourfold difference in incidence of IDDM. Children of Sardinian-heritage born in Lazio have the same incidence as the population of origin, which is genetically prone to the disease. Moreover, children with one Sardinian parent had a rate half that of Sardinians and double that of the indigenous population. We conclude that in a given population genetic susceptibility determines the frequency of IDDM in response to the environmental challenge.

1626 Vol 349 • May 31, 1997 residents. Thus, even assuming that children’s diet is greatly influenced by the diet of their parents, Sardinianheritage children living in Lazio would not have a different milk consumption from Lazio-heritage children, nor would children with one Sardinian parent have intermediate milk intake, as Odawara and Yamashita suggest, to account for the intermediate IDDM incidence (15·9) of these children. Moreover, among 17 regions with data for both cow’s milk consumption and IDDM incidence Sardinia was the highest outlier from the linear regression model in which R was 0·44, thus representing the most considerable exception to the association between the two variables. Finally, we want to point out that every possible influence of the microenvironment, such as persistence of original habits, was carefully explored, even on an individual basis, and could be ruled out. Yudkin proposes that non-genetic intergenerational effects, such as transplacental transmission of cytomegalovirus, could have a role in the high incidence of IDDM in Sardinia and in the Sardinian progeny in Lazio, thus accounting also for the 50% drop in incidence among children of mixed parenthood. However, he does not consider that we reported the same number of IDDM cases for children of mixed couples according to parental gender: seven with a Sardinian mother and seven with a Sardinian father. If the high incidence of IDDM among Sardinians were a consequence of transplacental transmission of a virus, Sardinian mothers should have outnumbered Sardinian fathers among mixed parents of IDDM children; and no different frequency between children of either Sardinian or mixed couples would have been found. Although we do not deny the contribution of environmental factors to the aetiology of IDDM, we reaffirm that the difference in the frequency of IDDM between two populations sharing the same environment is determined by the respective genetic susceptibilities in response to the environmental challenge.