Identifying diabetes genotype risk scores predicted adult type 2 diabetes in white and black adolescents, according to data presented at the 2012 Pediatric Academic Societies Annual Meeting.
Using data from the Bogalusa Heart Study, Jason L. Vassy, MD, MPH, fellow in training at Harvard Medical School, and colleagues identified 1,058 children (55% girls, 32% black) seen during adolescence and followed into adulthood who had genotype information available. The researchers hypothesized that genotype risk scores for diabetes could predict incident adult type 2 diabetes and improve prediction models based solely on clinical risk factors.
They defined incident type 2 diabetes as fasting glucose of at least 126 mg/dL or diabetes treatment, and used Cox regression to develop nested type 2 diabetes prediction models based on clinical factors examined during adolescence, including parental history of disease, BMI z score, mean arterial pressure, fasting glucose, HDL cholesterol and triglycerides, with and without a 38-variant genotype risk score.
To determine whether genotype risk scores improved prediction in each model, the researchers used likelihood ratio tests and continuous net reclassification improvement indices.
Median follow-up was 26.3 years, during which time 8.5% of participants developed type 2 diabetes. In all models, genotype risk scores predicted disease, with HRs of 1.09 per risk allele (95% CI, 1.03-1.15) in the basic demographic model and 1.06 (95% CI, 1-1.13) in the full model, the researchers wrote.
Adding the genotype risk score improved model fit in all models (all likelihood ratio tests, P<.05) and improved reclassification in the full model (continuous net reclassification improvement, 26.3%; 95% CI, 0.7-52.3). There was no association between race and genotype risk score in any model (P>.05 for all).
According to the researchers, the clinical implications of their findings are uncertain.