Context: The biological mechanism for the association between sleep duration and incident diabetes mellitus (DM) is unclear. Sleep duration and caspase-8, a marker of apoptotic activity, have both been implicated in β-cell function.
Objective: To investigate the associations between sleep duration and plasma caspase-8 and incident DM, respectively.
Design: Prospective cohort study.
Setting: The Malmö Diet and Cancer (MDC) Study is a population-based, prospective study run in the city of Malmö, Sweden.
Participants: A total of 4023 individuals from the MDC Study aged 45 to 68 years at baseline without a history of prevalent DM and with information on habitual sleep duration.
Main outcomes: Incident DM.
Results: Mean follow-up time was 17.8 years. Sleep duration was the only behavioral variable significantly associated with plasma caspase-8. Plasma caspase-8 was significantly associated with incident DM per standard deviation of its transformed continuous form [hazard ratio (HR) = 1.24; 95% confidence interval (CI), 1.13 to 1.36] and when dichotomized into high (quartile 4) (HR = 1.44; 95% CI, 1.19 to 1.74) compared with low (quartiles 1 to 3) concentrations. Caspase-8 interacted with sleep duration; compared with individuals who had 7 to 8 hours of sleep and low plasma caspase-8, those with high plasma caspase-8 and sleep duration <6 hours (HR = 3.54; 95% CI, 2.12 to 5.90), 6 to 7 hours (HR = 1.81; 95% CI, 1.24 to 2.65), and 8 to 9 hours (HR = 1.54; 95% CI, 1.09 to 2.18) were at significantly increased risks of incident DM.
Conclusions: Sleep duration is associated with plasma caspase-8. Caspase-8 independently predicts DM years before disease onset and modifies the effect of sleep duration on incident DM. Future studies should investigate if change of sleep duration modifies plasma concentrations of caspase-8.