Observations Multiple sclerosis and Type I diabetes [4], currently includes over 900 cases from the Lazio region, and none of these patients has multiple sclerosis. We, there- fore, analysed the HLA class II typing data for 450 of these Type I diabetes patients, 315 patients with multiple scle- rosis (unpublished data), and 719 control subjects, all residents of continental Italy. As shown in Table 1, the Type I diabetes susceptibility haplotypes DR3 and DR4 are not associated with multiple sclerosis in our cohort. Susceptibility to the latter dis- ease seems to depend more on having the DR15 haplotype, which is, as reported [5], infrequent among Type I diabetic pa- tients. This finding can explain why the two diseases are not as- sociated in continental Italy, which is highly similar to Sardinia in terms of environment and lifestyle. The common predisposition to both multiple sclerosis and Type I diabetes suggested by the data of Marrosu et al. seems to have a pri- marily genetic basis that could be unique to the Sardinian population. R. Buzzetti, P. Pozzilli, U. Di Mario, C. Ballerini, L. Massacesi Diabetologia (2002) 45:1735–1736 DOI 10.1007/s00125-002-0967-6 Received: 19 August 2002 Published online: 16 November 2002 © Springer-Verlag 2002 To the Editor: Multiple sclerosis (MS) and Type I diabetes mellitus are both believed to result from interaction between genetic and environmental factors. The incidence of these dis- eases varies from country to country, and marked differences can also be observed between ethnically distinct groups within a country. For instance, the incidence of Type I diabetes on the island of Sardinia is four to five times higher than that of continental Italy [1], and similar figures have been reported for multiple sclerosis. Sardinia and continental Central Italy are 200 km apart with residing populations exposed to similar environmen- tal factors. This suggests that the differences in the incidence of Type I diabetes and multiple sclerosis are primarily a result of genetic rather than environmental factors. This conclusion is supported (at least for T1DM) by the observation that, among the children of Sardinian parents who have migrated into the Lazio region of continental Italy, the high incidence of this dis- ease that characterizes their ancestral population has been re- tained [2]. Marrosu et al. [3] have reported an increased pre- valence of Type I diabetes in a cohort of Sardinians with multi- ple sclerosis. They suggested that susceptibility to both of these autoimmune diseases might be conferred by the same HLA haplotype. As they noted, previous studies have shown that multiple sclerosis in Sardinia is most strongly associated with two HLA DR-DQ haplotypes, DRB1*0301-DQA1*0501- DQB1*0201 and DRB1*0405-DQA1*0501-DQB1*0301, the first of which is known to bestow high susceptibility to Type I diabetes. In contrast, in northern European populations, multi- ple sclerosis shows a stronger relation with the DR15 haplo- type, which confers dominant protection against Type I diabe- tes. In the Lazio region of Italy, there is no multiple sclerosis patient registry and no published data on multiple sclerosis pa- tients of Sardinian origin in this area. However, the Eurodiab Registry for Type I diabetes, which was established in 1989 Table 1. Phenotypic distribution of HLA DRB1 in Continental Italians with Type I diabetes or multiple sclerosis and control subjects HLA Type I diabetes Multiple Control mellitus sclerosis subjects n=490 (%) n=315 (%) n=718 (%) DR3 277 (57) a 62 (20) 126 (18) DR4 269 (55) b 59 (19) 113 (16) DR15 17 (3.5) a 98 (31) d 83 (12) a χ 2 test Type I diabetes vs control subjects, p<10–4, OR 6.1 (CI 5–8) b χ 2 test Type I diabetes vs control subjects, p<10–4, OR 6.5 (CI 5–8.5) c χ 2 test Type I diabetes vs controls subjects, p<10–4, OR 0.3 (CI 0.16–0.5) d χ 2 test multiple sclerosis vs control subjects, p<10–4, OR 3.5 (CI 2.5–5)