Recent data suggest that proinsulin is associated with cardiovascular risk factors in nondiabetic and diabetic subjects. Since most conventional insulin assays cross-react with proinsulin, it has been suggested that the associations of insulin concentrations with dyslipidemia and hypertension could actually reflect associations with proinsulin. We examined these associations by using both a conventional immunoreactive insulin assay and a specific Linco insulin assay that does not cross-react with proinsulin in 623 nondiabetic and in 180 non-insulin-dependent diabetic subjects who participated in the San Antonio Heart Study, a population-based study of diabetes and cardiovascular disease. Both the immunoreactive insulin assay and the specific Linco insulin assay were equally correlated with cardiovascular risk factors in nondiabetic subjects. Insulin concentrations were moderately correlated with high triglyceride and low high-density lipoprotein cholesterol levels and were weakly correlated with increased blood pressure. In diabetic subjects there were only weak associations between insulin and cardiovascular risk factors using either assay. We conclude that the association of insulin concentrations with cardiovascular risk factors is not a function of using insulin assays that cross-react with proinsulin and that for epidemiological studies of cardiovascular risk factors, conventional immunoreactive insulin assays are as good as the newer specific insulin assays.