Introduction Circadian misalignment and sleep quality are intertwined processes that are both associated with diabetes. The association between sleep quality and incident diabetes has not been previously investigated in populations living at polar latitudes who experience extreme seasonal daylight variation and may be at greater risk of circadian misalignment. Using data from adult residents of Tromsø, Norway, this study investigates the association of poor sleep quality, as indicated by self-reported sleeplessness, and incident diabetes above the Arctic circle.

Research design and methods Secondary analysis of cohort data from the Tromsø Study. The study cohort consists of adults who attended both the fourth (Tromsø4) and seventh (Tromsø7) surveys conducted in 1995 and 2016, respectively. Only individuals with complete data were included. Multivariable logistic regression was used to examine the association between sleeplessness measured in Tromsø4 and incident diabetes measured in participants followed up to Tromsø7, adjusted for other diabetes risk factors.

Results Among 10 875 individuals (mean 41 years of age at baseline, 53.6% women), 21.2% (n=2302) reported experiencing sleeplessness at baseline. Diabetes incidence risk over follow-up (20 years) was 7.2% (n=784); incidence risk among individuals reporting sleeplessness was 8.8%, compared with 6.8% among unexposed individuals. After adjustment, sleeplessness-exposed individuals in the study cohort were found to have 23% greater odds (ORadj 1.23, 95% CI 1.03 to 1.47, p=0.022) of incident diabetes.

Conclusions Sleep quality is associated with incident diabetes in a population living above the Arctic circle. The direction and strength of association is consistent with findings from other geographical regions.