Quantity and quality of sleep and incidence of type 2 diabetes: a systematic review and meta-analysis

Abstract

Objective: To assess the relationship between habitual sleep disturbances and the incidence of type 2 diabetes and to obtain an estimate of the risk.

Research design and methods: We conducted a systematic search of publications using MEDLINE (1955-April 2009), EMBASE, and the Cochrane Library and manual searches without language restrictions. We included studies if they were prospective with follow-up >3 years and had an assessment of sleep disturbances at baseline and incidence of type 2 diabetes. We recorded several characteristics for each study. We extracted quantity and quality of sleep, how they were assessed, and incident cases defined with different validated methods. We extracted relative risks (RRs) and 95% CI and pooled them using random-effects models. We performed sensitivity analysis and assessed heterogeneity and publication bias.

Results: We included 10 studies (13 independent cohort samples; 107,756 male and female participants, follow-up range 4.2-32 years, and 3,586 incident cases of type 2 diabetes). In pooled analyses, quantity and quality of sleep predicted the risk of development of type 2 diabetes. For short duration of sleep (< or =5-6 h/night), the RR was 1.28 (95% CI 1.03-1.60, P = 0.024, heterogeneity P = 0.015); for long duration of sleep (>8-9 h/night), the RR was 1.48 (1.13-1.96, P = 0.005); for difficulty in initiating sleep, the RR was 1.57 (1.25-1.97, P < 0.0001); and for difficulty in maintaining sleep, the RR was 1.84 (1.39-2.43, P < 0.0001).

Conclusions: Quantity and quality of sleep consistently and significantly predict the risk of the development of type 2 diabetes. The mechanisms underlying this relation may differ between short and long sleepers.

Figure 1

Quantity of sleep and the risk of developing type 2 diabetes. Results are expressed as RR (95% CI). The size of squares is proportional to the weight of the study. A: Forest plot of the risk of type 2 diabetes associated with short duration of sleep compared with the reference group in nine population cohorts from seven published prospective studies. B: Forest plot of the risk of type 2 diabetes associated with long duration of sleep compared with the reference group in seven population cohorts from six published prospective studies. n/a, not available.

Figure 2

Quality of sleep and the risk of developing type 2 diabetes. Results are expressed as RR (95% CI). The size of squares is proportional to the weight of the study. A: Forest plot of the risk of type 2 diabetes associated with difficulty in initiating sleep (D.I.S.) compared with none in six population cohorts from five published prospective studies. B: Forest plot of the risk of type 2 diabetes associated with difficulty in maintaining sleep (D.M.S.) compared with none in six population cohorts from four published prospective studies. n/a, not available.

Figure 3

Meta-regression of the risk of developing type 2 diabetes by duration of follow-up according to type of sleep disturbance. The size of circles is proportional to the weight of the study. DIS, difficulty in initiating sleep; DMS, difficulty in maintaining sleep.