Type 1 Diabetes and Multiple Sclerosis

Together at last

Last year, Marrosu et al. (1) reported an increased prevalence of type 1 diabetes among Sardinian individuals with multiple sclerosis and their first-degree relatives. The study was accompanied by a commentary (2) indicating that these autoimmune disorders were “an unlikely alliance” because the HLA haplotype that increases risk for multiple sclerosis (DRB1*15-DQA1*0102-DQB1*0602) protects against type 1 diabetes. To the authors’ knowledge, there have been no published studies of an increased risk of multiple sclerosis for individuals with type 1 diabetes or their families.

Could the association between type 1 diabetes and multiple sclerosis be unique to Sardinia, where the rates of these two disorders are among the highest in the world? Indeed, multiple sclerosis is associated with DRB1*0405-DQA1*0501-DQB1*0301 and DRB1*0301-DQA1*0501-DQB1*0201 in Sardinia (3) and not DRB1*15-DQA1*0102-DQB1*0602, as it is elsewhere in the world. The DRB1*0301-DQA1*0501-DQB1*0201 haplotype is also associated with type 1 diabetes and other autoimmune diseases, such as celiac disease and autoimmune thyroid disease, which are also highly prevalent in Sardinia. Thus, perhaps the association between type 1 diabetes and multiple sclerosis observed in Sardinia is due to its unique HLA haplotype distribution (4).

Alternatively, the “unlikely alliance” hypothesis, which represents the current thinking of many researchers, may have biased us to the point that we failed to look for a possible association between type 1 diabetes and multiple sclerosis. Indeed, we plead guilty to this charge. However, examination of data collected for our Familial Autoimmune and Diabetes (FAD) Study revealed, for the first time, a highly significantly increased prevalence of multiple sclerosis in adults with type 1 diabetes and their first-degree relatives.

The FAD Study focused on the clustering of type 1 diabetes, autoimmune thyroid disease, and rheumatoid arthritis in a cohort of adult type 1 diabetic subjects diagnosed before age 17 years and seen at Childrens’ Hospital of Pittsburgh between 1950 and 1964. Self-report data on the existence of other autoimmune disorders in type 1 diabetic subjects and their siblings and parents were also collected. Seventy-six percent of the eligible probands and 83% of their siblings and parents participated. We recruited 94 nondiabetic control families for comparison.

Two percent of the females with type 1 diabetes and 0.5% of their sisters reported multiple sclerosis. No cases of multiple sclerosis were observed among male probands or male siblings, and multiple sclerosis was totally absent in the control families. Published prevalence rates for multiple sclerosis in the U.S. ranged from 0.06 to 0.17% (∼0.1% on average) for female adults. Thus, we observed a 20-fold increase in the prevalence of multiple sclerosis in our type 1 diabetic women (P = 0.003). Nondiabetic sisters had a fivefold higher risk of multiple sclerosis compared with the general population, which was not statistically significant.

We therefore conclude that adult women with type 1 diabetes are at an enormously increased risk of multiple sclerosis, and that the answer to questions about the clustering of these disorders is that they are “together at last.” Further epidemiologic studies are needed to confirm our findings. Genetic studies are also required to evaluate the DRB1*0301, DRB1*0405, and DRB1*15 extended haplotypes for common alleles at other loci that may contribute to the familial clustering of multiple sclerosis and type 1 diabetes.