Dose-response analyses of uric acid and risk of type 2 diabetes in a Chinese occupational population: a cohort study and meta-analysis

Abstract

BackgroundThe link between hyperuricaemia and risk of type 2 diabetes has long been hypothesised, although the debate about the causality of this is unresolved. To better understand the causal role of high concentrations of serum uric acid in the development of diabetes, we comprehensively assessed the association between serum uric acid concentration and the risk of incident diabetes in an occupational population of Chinese men and women. MethodsWe used Restricted Cubic Splines functions and Cox proportional hazards models to estimate the association between serum uric acid measured at baseline and the risk of incident diabetes in the Jinchang cohort study, an ongoing occupational-based prospective study of 42 122 metal-exposed workers aged 20 years or older at baseline (2011–13). Only those who participated in the baseline survey and first follow-up with repeat medical examinations were eligible. Serum uric acid was measured with the uricase-peroxidase enzymatic method. We also performed a meta-analysis of ongoing and completed studies in both the occupational and general populations to summarise this association. FindingsAmong 30 917 participants without diagnosed diabetes at baseline, there were 934 incident diabetes cases during a median of 2·2 years (IQR 1·8–2·5) of follow-up. We observed a positive linear dose–response association between serum uric acid concentration and incident diabetes by spline analyses (poverall association=0·001; pnon-linearity=0·90). After multiple adjustments such as age, BMI, and occupation, the hazard ratio (HR) for incident diabetes was 1·62 (95% CI 1·26–2·08) for the fourth and highest quartile concentration of serum uric acid (ptrend<0·0001), compared with the first quartile. Furthermore, when modelled as a continuous variable, serum uric acid concentration was associated with incident diabetes: the risk of diabetes was increased by 13% per 1 mg/dL increase in serum uric acid concentration (HR 1·13, 95% CI 1·07–1·19). Analysing men and women separately, the positive trends still remained in men (ptrend=0·01), but was not associated in women (ptrend=0·06). However, when modelled as a continuous variable, serum uric acid concentration was significantly associated with the risk of diabetes both in men and women. In a meta-analysis of 14 studies comprising a total of 137 485 participants and 9351 cases of type 2 diabetes, comparing the highest versus lowest quartile of serum uric acid concentration, the HR for incident diabetes was 1·44 (95% CI 1·34–1·55). InterpretationOur findings from the Jinchang cohort show a positive linear dose–response association between serum uric acid concentration and risk of incident type 2 diabetes in both Chinese men and women. These findings were substantiated in an updated meta-analysis, suggesting that high serum uric acid concentration might identify individuals at risk of diabetes. More studies, such as Mendelian randomisation studies and conventional epidemiological studies, are needed to establish causality. FundingProject of Science and Technology of the Jinchang Nonferrous Metals Corporation (Grant JKB20120013) and US National Institutes of Health (D43TW 008323, D43Tw 007864–01, and DK66401).