Table 3—

Cox regression analysis of associations between incidence of diabetes during 32 years (dependent variable) and composite variable sleep problems, as well as sleep complaints, sleep medication, and hours slept per 24 h when controlling for age as well as including different variables indicating obesity in the model

Relative risk (95% CI)Likelihood ratio χ2df
Composite variable sleep problems
    Plus age1.22 (0.82–1.08)72
    Plus age and BMI1.4 (0.95–2.1)1133
    Plus age and WHR1.1 (0.73–1.64)903
    Plus age and subscapular skinfold1.24 (0.84–1.84)1183
Sleep complaints
    Plus age1.03 (0.91–1.17)62
    Plus age and BMI1.06 (0.94–1.2)1113
    Plus age and WHR0.99 (0.88–1.13)903
    Plus age and subscapular skinfold1.04 (0.91–1.18)1183
Sleep medication
    Plus age1.16 (0.89-1.53)82
    Plus age and BMI1.25 (0.96–1.62)1133
    Plus age and WHR1.21 (0.93–1.57)913
    Plus age and subscapular skinfold1.31 (0.99–1.73)1203
Hours slept per 24 h
    Plus age0.95 (0.81-1.12)62
    Plus age and BMI0.97 (0.83–1.14)1113
    Plus age and WHR1.01 (0.86–1.18)903
    Plus age and subscapular skinfold0.98 (0.83–1.15)1183
Multivariate extended
    Composite variable sleep problems plus age plus BMI plus WHR plus triglycerides plus B-glucose plus systolic blood pressure plus education plus socioeconomic group plus subscapular skinfold plus exercise1.35 (0.89–2.1)16010
  • The variables are presented first once at a time controlled for age then in the extended multivariate analysis with all significant variables included. df, degrees of freedom.