Bivariate genetic analysis of fasting insulin and glucose levels

The main aim of this study was to estimate the relative influence of genes and environment on fasting insulin levels, which were considered a proxy of insulin resistance. Possible sex differences in genetic and environmental influences, and the origin of the covariance between fasting insulin and glucose were investigated. Subjects were 209 pairs of middle-aged twins, divided into 5 sex-by-zygosity groups. A general bivariate model and a reciprocal causation model including fasting insulin and glucose were used in the analyses. For both quantitative genetic models, a model specifying additive genetic and unique environmental factors, which were the same in males and females, showed the best fit to the data. Heritability estimates were modest and highly similar in both models: 20-25% of the variance in fasting insulin, and around 50% of the variance in fasting glucose levels could be attributed to genetic factors. The two models could not be discriminated on the basis of their fit to the data. A submodel of the general bivariate model suggested that the covariance between glucose and insulin has a unique environmental basis, whereas for the reciprocal causation model both causal paths were needed to explain the phenotypic correlation between insulin and glucose and estimates of the reciprocal paths were of opposite sign, an indication for the expected negative feedback loop.