Self-reported smoking, urine cotinine, and risk of type 2 diabetes: findings from the PREVEND prospective cohort study

Abstract

BackgroundSmoking is a major risk factor for type 2 diabetes (T2D), but the evidence has mostly relied on self-reports. We aimed to compare the associations of smoking exposure as assessed by self-reports and urine cotinine with T2D. MethodsUsing the PREVEND prospective study, smoking status was assessed at baseline by self-reports and urine cotinine in 4708 participants (mean age, 53 years) without a history of diabetes. Participants were classified as never, former, light current and heavy current smokers according to self-reports and analogous cut-offs for urine cotinine. Hazard ratios (HRs) with 95% CIs were estimated for T2D. ResultsDuring a median follow-up of 7.3 years, 259 participants developed T2D. Compared with self-reported never smokers, the multivariable adjusted HRs (95% CI) of T2D for former, light current, and heavy current smokers were 1.02 (0.75-1.4), 1.41 (0.89-2.22), and 1.30 (0.88-1.93), respectively. The corresponding adjusted HRs (95% CI) were 0.84 (0.43-1.67), 1.61 (1.12-2.31), and 1.58 (1.08-2.32), respectively, as assessed by urine cotinine. Urine cotinine-assessed but not self-reported smoking status improved T2D risk prediction beyond established risk factors. ConclusionUrine cotinine assessed smoking status may be a stronger risk indicator and predictor of T2D compared to self-reported smoking status.