Importance: Short and long sleep durations are adversely associated with cardiovascular disease (CVD), type 2 diabetes, and mortality. It remains unclear how sleep duration trajectories over time are associated with mortality and whether these associations vary by well-documented sex, race, and socioeconomic sleep disparities.
Objective: To investigate the association of 5-year sleep duration trajectories with all-cause and cause-specific mortality among US adults, predominantly those in low-income groups.
Design, setting, and participants: The Southern Community Cohort Study included participants aged 40 to 79 years recruited and enrolled (from March 2002 to September 2009) from community health centers by using random sampling methods across 12 states in the Southeastern US. Participants completed a follow-up survey between 2008 and 2013. Data analysis was performed from August 10 to November 30, 2023.
Exposures: Sleep duration was self-reported at study enrollment and at 5-year follow-up. At each time point, sleep was categorized as short (<7 hours), healthy (7-9 hours), or long (>9 hours). Nine sleep trajectories were defined based on 5-year change or consistency in sleep duration category between enrollment and follow-up.
Main outcome and measures: Cause of death was ascertained via linkage to the National Death Index through December 31, 2022. Multivariable-adjusted Cox proportional hazards regression analysis was performed to estimate hazard ratios (HRs) and 95% CIs for mortality outcomes (all-cause, CVD, cancer, and neurodegenerative disease) associated with sleep duration trajectory.
Results: Participants included 46 928 adults (mean [SD] age, 53.0 [8.8] years; 65.4% women; 63.3% self-identified as Black and 36.7% as White; and 47.5% with a household income <$15 000 per year). Overall, 66.4% of participants had suboptimal 5-year sleep trajectories. Race varied across sleep trajectories; 53.0% of participants in the optimal trajectory were Black, compared with 84.5% in the long-short trajectory. During a median 12.6 (IQR, 11.3-13.1) years of follow-up, 13 579 deaths occurred (4135 from CVD, 3067 from cancer, and 544 from neurodegenerative diseases). Compared with the optimal sleep duration trajectory, suboptimal trajectories were associated with as much as 29% greater risk of all-cause mortality in fully-adjusted models. For all-cause and CVD-specific mortality, the long-long (HRs, 1.27 [95% CI, 1.14-1.41] and 1.22 [95% CI, 1.01-1.48], respectively) short-long (HRs, 1.29 [95% CI, 1.17-1.42] and 1.22 [95% CI, 1.03-1.45], respectively), and long-short (HRs, 1.19 [95% CI, 1.05-1.35] and 1.32 [95% CI, 1.07-1.63], respectively) trajectories were associated with the greatest risk. After adjustment for comorbid conditions, no associations were observed for mortality due to cancer or neurodegenerative disease. Observed associations varied by race and household income, with the greatest risk observed among White adults with greater household incomes.
Conclusions and relevance: In this cohort study of 46 928 US residents, nearly two-thirds of participants had suboptimal 5-year sleep duration trajectories. Suboptimal sleep duration trajectories were associated with as much as a 29% increase in risk of all-cause mortality. These findings highlight the importance of maintaining healthy sleep duration over time to reduce mortality risk.