There is little information in the literature about the risk of diabetes in children of type 1 diabetic parents (mother and father affected), and evidence is based on small numbers (1). Analysis of a larger number of such trios could importantly contribute to the clarification of type 1 diabetes inheritance. In the Karlsburg Clinic for Diabetes and Metabolic Diseases, 77 offspring of 61 pairs with type 1 diabetes have registered since 1955. Until 1989, the parents and their children (diabetic and nondiabetic) were repeatedly treated or checked for diabetes in our clinic; since 1990 some of the patients have visited the clinic or have been contacted as outpatients. The diagnosis of type 1 diabetes in the affected individuals is proofed by clinical and in many cases laboratory data (C-peptide, antibodies, and HLA), as is obvious from our previous study (2).
Here we present data of 58 offspring from 46 pairs with type 1 diabetes that were last contacted in December of 2001. There are 35 daughters and 23 sons born between 1955 and 1993. Until now, 25 of 58 (43%) descendants developed diabetes at ages 1–42 years (mean ± SD 11 ± 10). The diabetes incidence did not increase during the observation time in 10-year intervals [54 (7 of 13), 42 (5 of 12), and 50 (11 of 22), respectively, and so far 18% (2 of 11) in those born 1986–1995]. The difference in the diabetes affection of sons (12 of 23) and daughters (13 of 35) is not significant. A strong preponderance of female descendants (18 female and 7 male) was seen in children born before 1975, whereas the sex ratio is nearly one (18 female and 15 male) for those born after 1975.
We observed a dramatic decrease in the age of diabetes onset. Children born during 1955–1975 developed the disease at 1–42 years (mean 15.1 ± 12.1), and children born during 1975–1995 developed the disease at 2–13 years (7.3 ± 5.2) (P = 0.0449). The mean age of mothers when they gave birth to their children was 24 years in both periods.
In children born before 1975 the mean age of diabetes manifestation (15.1 ± 12.1 years) is not different from that of their parents (fathers 16 ± 4.9 years and mothers 16.2 ± 7.4 years), but in children born after 1976 (7.3 ± 5.2) it is significantly (P = 0.014) younger than that of their fathers (13.6 ± 5.9 years) but not of their mothers (11 ± 6.8 years).
The constant diabetes incidence of >40% over >40 years of observation time implies the definite role of a genetic background for type 1 diabetes development. The high diabetes frequency of 43% (near half) in a representative number of children of two type 1 diabetic parents points, according to Mendelian laws, to a small number of responsible genes. On the other hand, we believe that these data demonstrate the influence of the altered environment over time. We previously described (3) a disturbed proportion of male to female (1:2) in children of diabetic mothers, but not of fathers. We now show that this shifted sex ratio does not exist in offspring of diabetic mothers (and fathers) born in the last 20 years. As most children were born in the Karlsburg hospital, we know that the therapeutical principles of diabetes treatment during pregnancy changed to normoglycemia in the late 1970s (4,5). As described in the Swedish study (6), we also confirm the dramatic shift to younger age at diagnosis for this subgroup of German diabetic children.
