Abstract

The aim of this study was to estimate annual incidence rate of type 1 diabetes according to the levels of genetic susceptibility provided by HLA-DQA1 and HLA-DQB1 genotypes. Two information sources were used: (1) a population-based incidence study in which 61 incident cases were ascertained during 1 year in Santiago, Chile (incidence rate: 4.11 cases per 100,000 children per year) and (2) a case-control comparison of 57 cases (recruited from the incidence study) and 125 controls. Susceptibility alleles were defined as DQA1*0301 and DQA1*0501 for DQA1 gene and alleles DQB1*0201 and DQB1*0302 for DQB1 gene. In DQA1 gene, the highest point estimate of the incidence rate was calculated for the genotype DQA1*0501/DQA1*501 (33.04 cases per 100,000 children aged less than 15 years old and per year; 95% CI: 9.22-118.33). In the DQB1 gene, the highest risk was estimated for the genotype DQB1*0201/DQB1*0201 (20.35 cases per 100,000 children aged less than 15 years old and per year; 95% CI: 5.26-78.67). This study shows an application on how a transformation of the logistic equation based on Bayes' theorem can be used to estimate incidence rates from case-control studies and population-based incidence rates.