Non-linear contribution of glucose measures to cardiovascular diseases and mortality: Reclassifying the Framingham's risk categories: A decade follow-up from the Tehran lipid and glucose study

Abstract

BackgroundWe investigated non-linear contribution of fasting plasma glucose (FPG) and 2-hour post-challenge plasma glucose (2h-PCPG) to the risk of CVD and mortality. We hypothesized that glucose measures improve risk-stratification made by the Framingham's general CVD risk algorithm. MethodsAmong participants aged ≥30 (n=8071), not taking glucose-lowering agents, 6169 (3477 women) remained eligible. Non-linear contribution of FPG and 2h-PCPG to incident CVD and mortality were assessed using Cox models incorporating restricted cubic splines functions. Risk reclassification improvement conferred by FPG and 2h-PCPG was examined using an extended version of net reclassification index (NRI) that takes into account the censoring nature of survival data. ResultsWe documented 465 incident CVD events (402 CHD), 212 deaths from any cause (94 CVD deaths). Excluding the contribution of the 2h-PCGP to mortality (that was linear) dose–response relationships between glucose measures and CVD and mortality were curvilinear with nadirs below which decreasing levels of glucose were unlikely to offer any benefit. These nadirs were assigned to FPG of 4.9–5.3 and 2h-PCPG of 6.0mmol.l−1. Glucose measures added to the predictive ability of the Framingham's general CVD risk algorithm with cutpoint-free NRIs ranging from 19 to 54%. ConclusionGlucose measures contributed to the risk of CVD and mortality in a curvilinear fashion, we observed increased risk below glucose thresholds currently used to define diabetes, supporting criteria for the diagnosis of impaired fasting glycemia and impaired glucose tolerance. Glucose measures were observed to add to predictive ability of the predictive model which included established cardiovascular risk factors.