Sex differences in the association between insulin resistance and non-fatal myocardial infarction across glycaemic states

Abstract

Abstract Background Females are generally less prone to cardiovascular (CV) events than males, but this protection is trumped by diabetes. The mechanism behind the increased relative risk in females with diabetes is not fully understood. Insulin resistance (IR) is suggested to be a more important contributor to CV morbidity in females than in males. We aim to investigate differences in the association between IR indexes (Homeostatic Model Assessment of IR - HOMA-IR, visceral adiposity index – VAI, and triglycerides/high-density lipoprotein-cholesterol - TG/HDL-C index), and a first non-fatal myocardial infarction (MI) across different glycaemic states. Methods IR indexes were calculated in a population with (n = 696) and without (n = 707) a first non-fatal MI, free from known diabetes. MI cases were investigated at least six weeks after the event. All participants were categorized by an oral glucose tolerance test as having normal glucose tolerance, impaired fasting glucose, impaired glucose tolerance, or newly diagnosed diabetes. Comparison of proportion of glycaemic states by sex was tested by chi-square test. The associations between sex, a first non-fatal MI, IR indexes, and traditional CV risk factors were analysed by multivariate logistic regression models. Continuous variables were logarithmically transformed. Results Of the total population 19% were females and 81% males, out of whom 47% and 50% had a first non-fatal MI, respectively. Compared with males, females were older, less often smokers, with lower body mass index and higher total cholesterol and high-density lipoprotein cholesterol levels. The proportion of glycaemic states did not differ between the sexes (p = 0.06). Females were less insulin resistant than males, especially among cases and with normal glucose tolerance. In logistic regression models adjusted for major CV risk factors including sex, the associations between VAI and TG/HDL-C index and a first non-fatal MI remained significant only in females (odds ratios and 95% confidence intervals: 1.7, 1.0-2.9, and 1.9, 1.1–3.4 respectively). Conclusions These results support the assumption that IR indexes based on anthropometrics and lipid panel, i.e., VAI and TG/HDL-C, could be a better measure of IR and CV-predictor for non-fatal MI in females, even without glycaemic perturbations.