Association of acrylamide and glycidamide haemoglobin adduct levels with diabetes mellitus in the general population.

The frequency and duration of exposure to acrylamide (AA) from the environment and diet are associated with a range of adverse health effects. However, whether long-term AA exposure is related to diabetes mellitus (DM) remains unknown. Data from 3577 adults in the National Health and Nutrition Examination Survey (NHANES) 2005-2006 and 2013-2016 aged ≥ 20 years was analysed. The main analyses applied multivariate logistic regression and restricted cubic spline models to investigate the associations between DM and AA haemoglobin biomarkers, including haemoglobin adducts of acrylamide and glycidamide (HbAA and HbGA), the sum of HbAA and HbGA (HbAA + HbGA), and the ratio of HbGA to HbAA (HbGA/HbAA) levels. After multivariable adjustment, the odds ratios (95% confidence intervals) for DM comparing the highest with the lowest AA haemoglobin biomarker quartiles were 0.71 (0.55, 0.93), 0.92 (0.71, 1.18), 0.80 (0.62, 1.03) and 1.95 (1.51, 2.51) for HbAA, HbGA, HbAA + HbGA and HbGA/HbAA, respectively. The restricted cubic spline model demonstrated that HbAA was linearly and inversely associated with risk of DM (P for trend = 0.013), while HbGA/HbAA was nonlinearly and positively associated with the prevalence of DM (P for trend <0.001). These results support for epidemiological evidence that the HbAA and HbGA/HbAA are significantly associated with DM. Further studies are warranted to infer the causal role of AA exposure in the prevalence of DM.