Association of sleep and circadian patterns and genetic risk with incident type 2 diabetes: a large prospective population-based cohort study

Abstract

Objective To examine the association of incident type 2 diabetes (T2D) risk with sleep factors, genetic risk, and their combination effects. Design Large prospective population-based cohort study. Methods This population-based prospective cohort study included 360 403 (mean (s.d.) age: 56.6 (8.0) years) participants without T2D at baseline from the UK Biobank. Genetic risk was categorised as high (highest quintile), intermediate (quintiles: 2–4), and low (lowest quintile) based on a polygenic risk score for T2D. Sleep scores, including long or short sleep duration, insomnia, snoring, late chronotype, and excessive daytime sleepiness, were categorized as an unfavourable, intermediate, or favourable sleep and circadian pattern. Results During a median follow-up of 9.0 years, 13 120 incident T2D cases were recorded. Among the participants with an unfavourable sleep and circadian pattern, 6.96% (95% CI: 6.68–7.24%) developed T2D vs 2.37% (95% CI: 2.28–2.46%) of participants with a favourable sleep and circadian pattern (adjusted hazard ratio (HR): 1.53, 95% CI: 1.45–1.62). Of participants with a high genetic risk, 5.53% (95% CI: 5.36–5.69%) developed T2D vs 2.01% (95% CI: 1.91–2.11%) of participants with a low genetic risk (adjusted HR: 2.89, 95% CI: 2.72–3.07). The association with sleep and circadian patterns was independent of genetic risk strata. Participants in the lowest quintile with an unfavourable sleep and circadian pattern were 3.97-fold more likely to develop T2D than those in the lowest quintile with a favourable sleep and circadian pattern. Conclusions Sleep and circadian patterns and genetic risk were independently associated with incident T2D. These results indicate the benefits of adhering to a healthy sleep and circadian pattern in entire populations, independent of genetic risk.