The authors analyzed a nationally representative sample drawn from the 2004 National Health Interview Survey (NHIS) then linked to the National Death Index (NDI) for mortality follow-up through December 31, 2019. The analytic cohort included 3,287 adults aged 18 years or older with self-reported heart disease and complete data on sleep, mortality, and covariates. Follow-up extended up to 15 years; by the end of the observation window nearly half of participants (49.4%) had died. The study used the NCHS probabilistic matching scheme to ascertain vital status from the NDI.
Sleep was measured with a single NHIS question: “On average, how many hours of sleep do you get in a 24-hour period?” Investigators categorized responses as ≤6 hours (short), 7–8 hours (reference), and ≥9 hours (long), consistent with adaptations of existing recommendations cited by the authors. The team also modeled sleep duration as a continuous exposure to avoid loss of information from categorization and to evaluate nonlinear effects.
The study evaluated three primary endpoints: all-cause mortality, heart disease–related mortality (defined using ICD-10 codes I00–I09, I11, I13, I20–I51), and non–heart disease-related mortality. During follow-up, 16.7% of the cohort died of heart-related causes and 32.6% of non-heart-related causes. Using Cox proportional hazards models (with stratification for covariates violating proportional-hazards assumptions), the authors reported that long sleep duration (≥9 hours) was associated with higher risk across all endpoints compared with 7–8 hours. Reported hazard ratios were: all-cause mortality HR 1.51 (95% CI 1.32–1.73); heart-related mortality HR 1.53 (95% CI 1.22–1.92); and non-heart-related mortality HR 1.50 (95% CI 1.27–1.77). These estimates come directly from the published analyses.
To capture nonlinear associations, the investigators applied restricted cubic spline models to continuous sleep duration. The spline analyses produced a nonlinear curve with the lowest predicted mortality risk centered near 6–7 hours of sleep rather than at the categorical 7–8 hour reference. The authors therefore reported a U-shaped or nonlinear pattern, whereby mortality rose at longer sleep durations and did not continue to decline at the lower end of the sleep scale.
The study team conducted multiple sensitivity analyses and stratified checks. They reported that the primary associations for long sleep and higher mortality remained robust across a range of sensitivity models and within age and sex subgroups. The Cox models were stratified when variables violated the proportional-hazards assumption; cause-specific analyses censored non-target deaths at time of death. The article notes these steps without introducing additional new measurements beyond the NHIS self-report and NDI cause-of-death coding.
The authors discussed the role of sleep duration as a potentially simple clinical marker of residual risk among patients living with heart disease, but the observational design precludes causal inference. Sleep was self-reported with a single-question assessment; the source did not report objective sleep measures such as actigraphy or polysomnography. Residual confounding remains possible despite model adjustment, and the study does not provide mechanistic data explaining why longer sleep correlated with higher mortality in this cohort. Finally, the analysis used data from a single NHIS cycle (2004) linked to NDI through 2019, and the report did not describe intervention studies testing whether modifying sleep duration alters risk in this population.
These results emphasize an association: long habitual sleep (≥9 hours) marked higher all-cause, heart-related, and non–heart-related mortality among US adults with established heart disease in this nationally representative cohort. The next clinical questions documented by the authors include whether sleep duration is a modifiable causal risk factor or primarily a marker of underlying illness, and whether objective sleep measures or interventional trials can clarify causality and inform secondary-prevention strategies.