Association of insulin resistance and β-cell dysfunction with incident diabetes among adults in China: a nationwide, population-based, prospective cohort study

Background National investigations on the interaction of insulin resistance, beta-cell dysfunction, and obesity with the development of diabetes are scarce in China. We aimed to investigate the individual and joint associations of insulin resistance and beta-cell dysfunction with incident diabetes, and to examine the modifying effect of BMI and waist circumference on these associations among adults with normal glucose tolerance and with prediabetes. Methods In this nationwide, population-based, prospective cohort study, we analysed data from the China Cardiometabolic Disease and Cancer Cohort Study, which recruited adults aged 40 years or older during 2011-12 (baseline) and invited participants to attend follow-up visits in 2014-16. Patients with diabetes at baseline, missing data for baseline measures of glucose tolerance status, missing data for baseline homoeostasis model assessment (HOMA) indexes, missing data for baseline covariates, and missing data for measures of glucose tolerance status at follow-up visits were excluded. At baseline and follow-up visits, a comprehensive set of questionnaires, dinical measurements, oral glucose tolerance tests, and laboratory examinations were carried out following standardised protocols. Glucose tolerance status and prediabetes were defined according to the American Diabetes Association 2010 criteria. In the main analysis, we examined the contributions of insulin resistance (HOMA of insulin resistance [HOMA-IR]) and beta-cell dysfunction (HOMA of beta-cell function [HOMA-B]) to diabetes risk, and evaluated the impact of obesity on these associations. Findings 94952 participants (31517 men and 63 435 women) were included in the analysis. High HOMA-IR was associated with a greater hazard of diabetes (quartile 4 vs 1: hazard ratio [HR] 6.70, 95% CI 6.08-7.39; per unit increase in Z score: HR 2.17, 95% CI 2.10-2.24) than low HOMA-B (quartile 1 vs 4: 4.08, 3.72-4.48; per unit decrease in Z score: 1.92, 1.85-2.00). Approximately 24.4% (95% CI 23.6-25.2) of the incident diabetes could be attributed to insulin resistance and 12.4% (11.2-13.7) could be attributed to beta-cell dysfunction. The HRs for diabetes were 1.83 (95% CI 1.72-1.95) per unit increase in Z score of HOMA-IR and 2.03 (1.86-2.21) per unit decrease in Z score of HOMA-B among participants with normal weight; the corresponding H Rs for diabetes were 2.02 (1.93-2.11) and 1.88 (1.79-1.98) among participants with obesity (p(interaction) =0.0091). These associations and interactions were similar for participants with normal glucose tolerance or prediabetes. Interpretation Insulin resistance shows a stronger association with incident diabetes than does beta-cell dysfunction in Chinese adults, and this association pattern was more prominent among adults with obesity. Given the limitations of HOMA indexes as surrogate measures of insulin resistance and beta-cell dysfunction, these findings should be interpreted with caution. Copyright (C) 2019 Published by Elsevier Ltd. All rights reserved.