Table 3.
Multivariable‐Adjusted Linear and Logistic Models for Associations of Sleep Characteristics with Cardiovascular Health Defined by the American Heart Association Life's Simple 7 Score (n=1920)*
| Linear regression models | Logistic regression models† | |||
|---|---|---|---|---|
| Sleep characteristics | B(SE) | P value | Sleep characteristics | OR (95% CI) |
| Sleep duration, h‡ | 0.01 (0.04) | 0.023 | Sleep duration (<6 vs ≥6 h) | 1.25 (1.01–1.55) |
| Sleep duration (<6 h vs ≥6 h) | −0.27 (0.12) | 0.025 | ||
| Sleep efficiency (%)c | 0.05 (0.01) | 0.002 | Sleep efficiency (<85% vs ≥ 85%) | 1.19 (0.86–1.65) |
| Sleep efficiency (<85% vs ≥85%) | −0.26 (0.18) | 0.144 | ||
| SD of sleep duration (h)‡ | −0.25 (0.07) | <0.001 | Increased night‐to‐night variability in sleep duration (SD of sleep duration ≥90 vs <90 min) | 1.24 (1.02–1.51) |
| Increased night‐to‐night variability in sleep duration (SD of sleep duration ≥90 vs <90 min) | −0.45 (0.11) | <0.001 | ||
| SD of sleep onset timing (h)‡ | −0.03 (0.03) | 0.382 | Increased night‐to‐night variability in sleep onset timing (SD of sleep onset timing ≥90 vs <90 min) | 1.31 (1.04–1.64) |
| Increased night‐to‐night variability in sleep onset timing (SD of sleep onset timing ≥90 vs <90 min) | −0.50 (0.12) | <0.001 | ||
| WHIIRS‡ | −0.04 (0.11) | 0.746 | Insomnia (WHIIRS score ≥9 vs <9) | 1.03 (0.84–1.25) |
| ESS‡ | −0.40 (0.16) | 0.010 | Excessive daytime sleepiness (ESS >10 vs ≤10) | 1.17 (0.89–1.54) |
| OSA (moderate to severe vs mild or none)§ | −0.95 (0.11) | <0.001 | OSA (moderate to severe vs mild or none)§ | 2.21 (1.78–2.73) |
AHA LS7 indicates American Heart Association Life's Simple 7; ESS, Epworth Sleepiness Scale; OR, odds ratio; OSA, obstructive sleep apnea; and WHIIRS, Women's Health Initiative Insomnia Rating Scale.
Models are adjusted for age, sex, race/ethnicity, education, health insurance, and alcohol use.
In linear models, the sleep characteristics were evaluated in relation to the AHA LS7 score. In logistic regression models, the sleep characteristics were evaluated in relation to odds of having poor cardiovascular health, defined as an AHA LS7 score of 0–7.
Variables were used on the continuous scale. A higher standard deviation of sleep duration and sleep onset timing indicates greater night‐to‐night variability in sleep patterns. A higher ESS score indicates greater daytime sleepiness. A higher Women's Health Initiative Insomnia Rating Scale score indicates greater insomnia severity.
Additional adjustment of models on OSA for BMI attenuated associations, but the results remained significant. In linear models: B(SE)=−0.30(0.10), P‐value=0.005, and in logistic regression models: OR (95%CI), 1.35 (1.07–1.72).