Table 2. Association of sleep duration with incident cardiovascular disease (CVD) in the 45 and Up Study.

	Model 1a	Model 2b		Model 3c
Sleep duration	HRd	Healthy HR	Less Healthy HR	HR	Events/pye
<6 h	1·16 (1·01-1·34)*	0·92 (0·75-1·14)	1·38 (1·12-1·70)**	1·03 (0·88-1·21)	165/11,332
6 h	1·09 (1·00-1·19)	1·06 (0·96-1·18)	1·10 (0·93-1·31)	1·06 (0·96-1·17)	629/44,275
7 h	1	1	1	1	1,228/93,789
8 h	1·01 (0·94-1·07)	0·94 (0·87-1·02)	1·18 (1·03-1·35)*	0·98 (0·91-1·05)	1,956/150,222
9 h	1·01 (0·92-1·11)	0·95 (0·85-1·06)	1·14 (0·97-1·34)	0·98 (0·89-1·09)	555/40,405
≥10 h	1·10 (0·98-1·22)	0·97 (0·83-1·14)	1·17 (1·00-1·39)	1·00 (0·88-1·14)	319/20,647

Notes: Individuals with missing data for sleep duration, age, sex, physical functioning or baseline illness were excluded from all analyses. * P<0·05. ** P<0·01.

^a Model 1 (N=181,544) included all eligible participants except those reporting CVD at baseline, and was adjusted for age and sex. ^b Model 2 (N=181,544) included all eligible participants except those reporting CVD at baseline, and was adjusted for age, sex, education, marital status, residential remoteness, alcohol consumption, smoking status, health insurance status, income, body mass index, physical activity and baseline health status. Results are presented separately for healthy and less healthy participants, as significant interaction between sleep duration and health status was observed. ^c Model 3 (N=156,902) excluded all participants reporting any serious illness at baseline (current/past cancer, or diagnosed/ treated heart disease, stroke or type 2 diabetes), and was adjusted for the same set of covariates as Model 2. ^d Hazard ratio with 95% confidence interval, representing the estimated risk of CVD for the specified category of sleep duration, compared with the reference category of 7 h. ^e Number of cardiovascular disease (coronary heart disease, ischaemic stroke or peripheral vascular disease) hospitalisations/total person years at risk during follow-up.