Content area
A population-based juvenile-onset-diabetes registry was established in the United States Virgin Islands. Primary case ascertainment was by retrospective review of hospital and clinic records, and IDDM was defined by Diabetes Epidemiology Research International (DERI) criteria. Validity and completeness of the registry were assessed by use of a physician survey and a mass media campaign. Completeness of the primary sources was estimated to be 92.3%. A total of 39 individuals age 0-19 were found to have developed diabetes in the U.S. Virgin Islands between January 1, 1979, and December 31, 1988. During the 10-year period, 28 children age $<$15 were diagnosed with IDDM resulting in an average annual IDDM incidence rate (IR) of 7.5/100,000 (94% confidence interval = $\pm$2.8). A significant increase in IDDM incidence ($\rho <$.001) was observed when the IR rose to 28.4/100,000 in 1984. A significant ($\rho <$.001) seasonal pattern in IDDM incidence was found which differed from other registries. White children had the highest IR (28.9/100,000). The IR for Hispanics (7.2/100,000) was slightly higher than that for blacks (5.9/100,000). Among black children a slight male excess in incidence was observed (m/f = 1.5). The IR for black boys peaked in the 5-9 age group (8.5/100,000) while the IR for black girls was highest in the 10-14 age group (9.3/100,000). When black cases were compared to age-matched controls with respect to grandparental race, cases were found to have a greater percentage of white ancestry ($\rho <$.02). Similar results were found for Hispanic cases ($\rho <$.05). The incidence of IDDM in Caribbean blacks (West Indians) in the U.S. Virgin Islands was similar to blacks in the United States. Differences in IDDM incidence by race were consistent with other reports. The epidemic of IDDM in 1984 supports an infectious etiology of IDDM.