Comparison between indexes of insulin resistance for risk prediction of cardiovascular diseases or development of diabetes

Aim: The predictive effect of various insulin resistance indexes for risk of cardiovascular diseases (CVD) or type 2 diabetes (T2DM) is still unclear. Methods: One thousand and forty-nine 71-years-old male subjects from the Swedish ULSAM study, mean follow-up 9 years. All subjects performed the euglycemic insulin clamp for M/I [glucose disposal/mean insulin], and 75-g oral glucose tolerance test for Ceder-IR: 1/glucose uptake rate/[mean glucose x log mean insulin]; Matsuda-IR: 1/10,000/square root [glucose0 x insulin0 x glucose120 x insulin120]; Belfiore-IR: 1/([glucose0 + glucose120]/normal mean glucose x [insulin0 + insulin120]/normal mean insulin)+1); and HOMA-IR: [glucose0 x insulin0]/22.5. Results: Bland-Altman plots showed best agreement between M/I versus Belfiore-IR and Ceder-IR with mean difference near zero, -0.21 to -0.46, while -0.68 to -0.77 for the other indexes. ISI-Ceder was the strongest predictor for incident nonfatal/fatal ischemic heart disease (CHD) or CVD at Cox regression in all subjects, and for incident T2DM at logistic regression in 1024 subjects with no baseline T2DM, with significantly higher hazard ratios or odds ratios than with all other indexes, also with best model fit, after adjusting for clinical characteristics and the traditional cardiovascular risk factors, including metabolic syndrome for CVD risk. Conclusion: Ceder-IR performed strongest as independent predictor for incidences of CHD/CVD and T2DM. (C) 2015 Elsevier Ireland Ltd. All rights reserved.