The high incidence of insulin-dependent diabetes mellitus (IDDM) in Finland contrasts strikingly with the low rates in the neighbouring populations of countries in the Eastern Baltic region: Estonia, Latvia and Russia. To evaluate the possible contribution of genetic factors to these differences, the frequencies of HLA-DQB1 alleles and relevant DQB1-DQA1 or DQB1-DRB1 haplotypes associated with IDDM risk or protection were analysed among IDDM patients and control subjects from these four populations. An increased frequency of HLA-DQB1*0302, DQB1*02-DQA1*05 and DQB1*0302-DRB1*0401 was observed in subjects with IDDM in all studied populations, whereas the prevalence of DQB1*0301 and DQB1*0602 and/or *0603 was decreased among patients. The degree of IDDM risk associated with HLA alleles analysed here did not differ significantly between the populations. Comparisons of the distribution of IDDM-related HLA alleles and haplotypes in the background populations revealed its consonance with IDDM incidence. The combined frequency of high risk genotypes was significantly higher among Finns than in other populations studied. Our data support the hypothesis that variance in the dispersion of HLA alleles is the genetic basis of variation of IDDM incidence observed in the Eastern Baltic region.

We have carried out a comparison of the incidence of childhood onset insulin-dependent diabetes mellitus (IDDM) between five populations around the Baltic Sea. These were Finland, Estonia, Latvia, Lithuania and Poland. The risk of IDDM is highest in the world in Finland and also very high in Sweden, on the western and northern side of the Baltic Sea. The risk of IDDM in children on the eastern side of the Baltic Sea has not been known before. The data collection period covered the years 1983-1988. A marked variation in incidence was seen within this relatively small geographical area. Among these five populations, the incidence increased with the latitude. Our present results confirmed the very high incidence of IDDM in Finland. The average age-standardized yearly incidence of IDDM/100,000 was in males under 15 years of age 36.9 in Finland, 10.7 in Estonia, 6.4 in Latvia, 6.5 in Lithuania and 6.0 in Poland. In females the incidence was 31.6, 10.0, 6.9, 7.0 and 6.4 in these five populations, respectively. The differential in incidence in Estonia as compared with Latvia, Lithuania and Poland was statistically significant (P = 0.0002). A slight male excess in incidence was found in countries with higher incidence--Finland and Estonia, but in lower incidence countries the sex ratio was opposite (P = 0.019 for the interaction sex-population). During 1983-1988 the incidence increased significantly in Finland but not in other populations although a large year-to-year variation in incidence was observed in each country, particularly in males. We recorded a peak in IDDM incidence in most of these populations around 1986.

The genes encoding the HLA-DQ heterodimer molecules, DQB1 and DQA1, have been found to have the strongest association with IDDM risk, although there is cumulative evidence for the effect of other gene loci within the major histocompatibility complex gene region. After the HLA-DQ locus, the HLA-DR locus has been suggested most often as contributing to the disease susceptibility. In this study we analyzed at the population level the effect of DR4 subtypes and class I, HLA-B alleles, on IDDM risk when the influence of the DQ locus was stratified. In all three populations studied (Estonian, Latvian, and Russian), DQB1*0302 haplotypes most frequently carried DRB1*0401 or DRB1*0404. DRB1*0401 was the most prevalent subtype in IDDM patients, whereas DRB1*0404 was decreased in frequency. DRB1*0402 was also prevalent among Russian haplotypes, but was not associated with IDDM risk. When HLA-B alleles were analyzed, strong associations between the presence of specific B alleles and DRB1*04 subtypes were detected. The HLA-B39 allele was found significantly more often in DRB1*0404-DQB1*0302-positive patients than in healthy control subjects positive for this haplotype: 27 of 54 (50%) vs. 4 of 49 (8.2%) (P < 0.0001). The results demonstrate that DQ and DR genes cannot explain all of the HLA-linked susceptibility to IDDM, and that the existence of a susceptibility locus telomeric to DR is probable.

The seasonal pattern with two cycles among older children and one cycle only among younger children may indicate different triggers of Type 1 diabetes mellitus for different age groups.