A systematic evaluation of the insulin resistance syndrome as an independent risk factor for cardiovascular disease mortality and derivation of a clinical index.

Insulin resistance-related risk factor clustering (the insulin resistance syndrome or IRS) may be a cardiovascular disease (CVD) risk factor, but a convincing demonstration of this requires a rigorously derived reference measure and a systematic evaluation of measures and indices that derive from that measure. Using established IRS characteristics, factor analysis in 832 white men, generally healthy at baseline, generated a priori an IRS reference measure. An IRS diagnostic was chosen by evaluating CVD mortality risk in percentiles of the reference measure. An IRS clinical index was derived by (1) identification of readily measured, independent predictors of the IRS reference measure by multiple linear regression; (2) assignment to each predictor of a cut point optimal for discrimination of participants diagnosed with IRS; and (3) selection of a combination of the dichotomized predictors that further optimized IRS discrimination. The reference IRS diagnostic was defined by the top 16.7% of the IRS reference measure and predicted CVD mortality in Cox proportional hazards modeling (hazard ratio, 2.7; 95% confidence interval, 1.5-5.2; P = .002). An optimized IRS index was defined by triglycerides of at least 1.6 mmol/L and uric acid of at least 400 μmol/L plus any one of fasting plasma glucose of at least 5.4 mmol/L, diastolic blood pressure of at least 90 mm Hg, or body mass index of at least 27.0 kg/m(2) and predicted CVD mortality (hazard ratio, 2.14 [1.08-4.24]; P = .02). Prediction was independent of hypertension, hypercholesterolemia, and smoking. Conventionally derived glucoregulatory insulin resistance and metabolic syndrome were not predictive. The IRS is an independent risk factor for CVD mortality; and an effective, clinically usable index can be derived from readily measured variables.