No study has assessed association between cigarette smoking and new-onset diabetes mellitus (NODM) incidence using two different smoking classification systems: self-reported questionnaire and urine cotinine. The objective of this longitudinal study was to evaluate NODM risk using the above two systems in Korean adults.

Among individuals enrolled in Kangbuk Samsung Health Study and Cohort Study who visited between 2011 and 2012 at baseline and 2014 at follow-up, 78,212 participants without baseline diabetes mellitus were followed up for a median of 27 months. Assessment of NODM incidence was made at the end of follow-up period. Cotinine-verified current smoking was having urinary cotinine ≥50 ng/mL.

Percentages of self-reported and cotinine-verified current smokers were 25.9% and 23.5%, respectively. Overall incidence of NODM was 1.5%. According to multivariate regression analyses, baseline self-reported current smoking (relative risk [RR], 1.33; 95% confidence interval [CI], 1.07 to 1.65) and cotinine-verified current smoking (RR, 1.27; 95% CI, 1.08 to 1.49) increased NODM risk compared to baseline self-reported never smoking and cotinine-verified current non-smoking. Higher daily amount and longer duration of smoking were also associated with increased NODM risk (P for trends <0.05). In particular, self-reported current smokers who smoked ≥20 cigarettes/day (RR, 1.62; 95% CI, 1.25 to 2.15) and ≥10 years (RR, 1.34; 95% CI, 1.08 to 1.67) had the highest RRs for NODM. These results remained significant in males, although there was no gender interaction.

This longitudinal study showed that baseline self-reported and cotinine-verified current smoking were associated with increased risks of NODM, especially in males.