Fasting glucose, fasting insulin, and insulin resistance in the prediction of myocardial infarction and mortality at long-term follow-up

Objective: To assess the additional prognostic value of fasting blood glucose (FBG), fasting plasma insulin (FPI), and homeostasis model assessment derived insulin resistance (HOMA-IR) for predicting incident myocardial infarction (MI) and all-cause mortality, independently of traditional cardiovascular (CV) risk factors Methods: As part of a population-based cohort study aiming to screen for CV risk factors, 6024 men and 1013 women without known diabetes mellitus (DM) or overt CV disease (CVD defined as previous MI, stroke, or transient ischemic attack), recruited 1974-1992, had both FBG and FPI measured at baseline. Subsequently, HOMA-IR was derived using the computerized HOMA calculator and ranked into quartiles due to the non-normal distribution and presumably non-linear biological effect of insulin resistance. Prognostic values of FBG, FPI, HOMA-IR, and traditional risk factors were tested using Cox proportional hazards regression analysis and likelihood-ratio testing. Follow-up time from inclusion until event (first MI or death) or censoring (emigration or last follow-up date (<20 years)) comprised the underlying time scale Results: Median [IQR] age at inclusion was 47.7 [47.1-48.3] years, whereas median [IQR] HOMA-IR was 0.9 [0.4-1.4]. Over a median follow-up time of 20 years, 1448 events occurred (11.3 per 1000 person-years). The simple prediction model, i.e. the model with traditional CV risk factors only, included age, gender, body mass index, systolic blood pressure, total cholesterol, and smoking status. Addition of FBG, but not FPI or HOMA-IR resulted in significant model improvement (FBG: chi2 = 5.82, p = 0.02; FPI: chi2 = 0.29, p = 0.59; HOMA-IR: chi2 = 0.08, p = 0.78). Moreover, we detected significant interactions between both FBG and FPI (chi2 = 12.52, p <0.001), FBG and HOMA-IR (chi2 = 6.07, p = 0.01), and FPI and HOMA-IR (chi2 = 13.15, p <0.001) Conclusion: Addition of FBG, but not FPI or HOMA-IR provided additional significant prognostic value on top of traditional CV risk factors in the prediction of MI or death at long-term follow-up in subjects without CVD and/or DM at baseline. Furthermore, interaction analyses revealed that the prognostic values of both FBG and FPI were significantly greater among subjects in the highest HOMA-IR quartile.