Short, long sleep duration associated with increased likelihood of type 2 diabetes in middle-aged women

Among middle-aged women, those with a self-reported sleep duration of =8 hours are more likely to have type 2 diabetes relative to those with a sleep duration of 7 hours, according to results of the Finnish type 2 diabetes (FIN-D2D) survey. However, the survey findings did not reveal a similar association in middle-aged men.

The FIN-D2D survey was conducted in 2004 and 2005 to evaluate the FIN-D2D Project, a community-based diabetes prevention program. Researchers selected 4,500...

Among middle-aged women, those with a self-reported sleep duration of <=6 hours or >=8 hours are more likely to have type 2 diabetes relative to those with a sleep duration of 7 hours, according to results of the Finnish type 2 diabetes (FIN-D2D) survey. However, the survey findings did not reveal a similar association in middle-aged men.

The FIN-D2D survey was conducted in 2004 and 2005 to evaluate the FIN-D2D Project, a community-based diabetes prevention program. Researchers selected 4,500 individuals aged 45 to 74 years from the National Population Register to participate; 1,336 men and 1,434 women were included in the current analysis.

All of the participants completed a health examination, an oral glucose tolerance test (OGTT), and a sleep questionnaire.

The mean self-reported sleep duration was 7.3 hours for both women and men. Among the women, the researchers observed a significant U-shaped relationship between sleep duration and diabetes prevalence, which remained significant in adjusted analyses. Although the researchers observed a tendency toward the same U-shaped association in men, this association was not significant in adjusted analyses.

The adjusted odds ratio for having a diabetes diagnosis prior to the survey was 2.55 (95% CI, 1.21-5.35) among women who slept <=6 hours per night and 1.76 (95% CI, 1.12-2.61) among those who slept >=8 hours per night, relative to those who slept 7 hours per night. Similarly, the adjusted odd ratios for having either previously diagnosed diabetes or diabetes detected at the time of the survey health exam were 1.56 (95% CI, 0.95-2.58) and 1.71 (95% CI, 1.13-2.61), respectively.

The authors acknowledged that their study was limited by the overall participation rate (62%), the cross-sectional design, the lack of multiple glucose measurements and objective sleep measures, and the fact that they did not consider menopause status or hormone replacement therapy use in their analyses.  (Tuomilehto H, et al. Sleep Med 2008;9:221-227.)

In an accompanying editorial, Peter Nilsson of Lund University in Sweden noted that these findings, which show a stronger association between sleep duration and diabetes in women than in men, contradict most of the other published literature on the subject.

Nilsson added that intervention studies would need to be undertaken to truly evaluate the relationship between sleep quality/duration and glucose tolerance/type 2 diabetes, although such studies are unlikely to be conducted.

As a result, he advised, "[B]oth men and women should be studied in order to find ways to normalize their sleep habits to avoid consequences of disease and even increased mortality risk, taking into account the interindividual range of what is called normal sleep length." (Sleep Med 2008;9:219-220.)