Combined sleep, domain-specific physical activity, and nutrition in relation to all-cause mortality among US adults

Technical Summary: Combined Sleep, Domain-Specific Physical Activity, and Nutrition in Relation to All-Cause Mortality Among US Adults

Problem Statement
While sleep duration, physical activity (PA), and nutrition (SPAN) are individually established as determinants of mortality risk, existing literature predominantly examines these behaviors in isolation or in pairs. There is a scarcity of research investigating the synergistic associations of all three behaviors simultaneously. Furthermore, previous studies often treat physical activity as a domain-agnostic total volume (e.g., total moderate-to-vigorous physical activity [MVPA]), potentially overlooking distinct health associations based on how activity is accumulated across daily life domains: occupational (OPA), leisure-time (LTPA), and transportational (TPA). Additionally, much of the high-quality evidence on combined SPAN behaviors originates from the UK Biobank, a cohort that may be less socioeconomically heterogeneous and generally healthier than the general population. This study addresses these gaps by examining combined SPAN behaviors and their association with all-cause mortality across different physical activity domains within a representative sample of US adults.

Methodology

• Study Population: The study utilized data from the National Health and Nutrition Examination Survey (NHANES) cycles from 2007 to 2018. The analytical sample included 31,875 adults aged 18 years or older with complete data on SPAN behaviors. Participants with mortality recorded within the first year of follow-up were excluded to mitigate reverse causation.
• Data Sources: Mortality data were linked to the National Death Index (NDI) through December 31, 2019.
• Exposure Assessment:
  • Sleep: Self-reported duration categorized as short (<7 h/day), optimal (7–8 h/day), and long (>8 h/day).
  • Physical Activity: Assessed via the Global Physical Activity Questionnaire (GPAQ). MVPA was calculated by summing moderate activity minutes and twice the vigorous activity minutes. Analyses were conducted for total MVPA and separately for domain-specific MVPA (OPA, LTPA, TPA).
  • Nutrition: Diet quality was estimated using the Healthy Eating Index-2020 (HEI-2020) derived from 24-hour dietary recalls.
• Statistical Analysis:
  • Joint Categorical Analysis: Participants were stratified into 27 mutually exclusive groups based on tertiles of PA/HEI-2020 and guideline-based sleep categories. Cox proportional hazards models estimated Hazard Ratios (HRs) for all-cause mortality, with the lowest combination (short sleep, low PA, low diet quality) as the reference.
  • Minimum Variation Analysis: Restricted cubic spline models were used to identify the minimum joint increments in SPAN behaviors (relative to a 5th percentile baseline) associated with a clinically meaningful 10% reduction in mortality risk.
  • Adjustments: Models were adjusted for age, sex, ethnicity, education, income-to-poverty ratio, alcohol, smoking, sedentary time, total energy intake, and comorbidities (CVD, cancer, T2D). When analyzing specific PA domains, the other two domains were mutually adjusted.
  • Sensitivity: Extensive sensitivity analyses were performed, including adjustments for BMI, exclusion of current smokers and those with baseline disease, alternative diet indices (AHEI), and multiple imputation strategies.

Key Results

• Combined SPAN and Mortality: The combination of optimal sleep (7–8 h/day), high MVPA (>72.9 min/day), and moderate-to-high diet quality (HEI-2020 >45.0) was associated with the lowest all-cause mortality risk (HR = 0.50, 95% CI: 0.37–0.67) compared to the reference group.
• Domain-Specific Findings:
  • Leisure-Time PA (LTPA): The combination of optimal sleep, high LTPA (>35.4 min/day), and high diet quality yielded the lowest mortality risk (HR = 0.44, 95% CI: 0.30–0.66).
  • Occupational PA (OPA): The lowest risk was observed with short sleep (3.0–6.5 h/day), moderate OPA (1.4–111.5 min/day), and high diet quality (HR = 0.53, 95% CI: 0.37–0.75).
  • Transportational PA (TPA): The optimal combination involved optimal sleep, high TPA (>25.7 min/day), and high diet quality (HR = 0.55, 95% CI: 0.38–0.79).
• Minimum Variations: A modest combined increment of 15 minutes/day of sleep, 10.8 minutes/day of total MVPA, and a 5-point increase in HEI-2020 was associated with a 10% lower mortality risk.
  • For LTPA, the 10% risk reduction required only an additional 5.2 minutes/day of activity (plus 15 min sleep and 5 HEI points).
  • For OPA, the requirement was 24.0 minutes/day of activity.
  • For TPA, the requirement was 5.5 minutes/day of activity, but necessitated a larger sleep increment (30 minutes/day) to achieve the same risk reduction.
• Robustness: Results remained consistent across sensitivity analyses, including adjustments for BMI, exclusion of smokers, and the use of alternative diet indices. Subgroup analyses indicated homogeneity across sex, race, education, and income strata.

Significance and Claims
The authors claim this is the first study to examine physical activity accumulation patterns by specific domains within a combined SPAN framework in relation to all-cause mortality. The study highlights that while modest combined improvements in sleep, activity, and nutrition are beneficial across all domains, leisure-time physical activity demonstrates the most favorable mortality profile, requiring the smallest incremental increase in activity time to achieve significant risk reduction when paired with sleep and diet improvements.

The paper suggests that the "physical activity paradox"—where occupational activity shows weaker or different associations with health compared to leisure activity—persists even when combined with optimal sleep and nutrition. The authors posit that LTPA, often self-regulated and allowing for recovery, confers distinct benefits compared to OPA, which may involve prolonged strain and limited recovery. Furthermore, the study provides evidence that even small, achievable increments in SPAN behaviors (e.g., adding ~5 minutes of leisure activity and 15 minutes of sleep daily) can yield clinically meaningful reductions in mortality risk, supporting a public health approach that encourages modest, simultaneous improvements in these three behavioral pillars.

The authors note limitations, including the self-reported nature of the data (potentially introducing recall bias), the inability of NHANES TPA questions to capture vigorous intensity, and the lack of data on sleep disorders. They conclude that while the associations are robust, the specific mechanisms regarding domain differences warrant further investigation.