Minimum and optimal combined variations in sleep, physical activity, and nutrition in relation to all-cause mortality risk

Abstract

Background: Sleep, physical activity, and nutrition (SPAN) are critical behaviours for health, although they have traditionally been studied separately. We examined the combined associations of SPAN and the minimum between-individual variations associated with meaningfully lower all-cause mortality risk.

Methods: This prospective cohort analysis included 59,078 participants from the UK Biobank (median age: 64.0 years; 45.4% male) who wore trackers for 7 days and self-reported dietary data. Wearable-measured sleep (hours/day) and moderate to vigorous physical activity (MVPA; mins/day) were calculated using a machine learning based schema. A 10-item diet quality score (DQS) assessed the intake of vegetables, fruits, fish, dairy, whole grains, vegetable oils, refined grains, processed and unprocessed meats, and sugary beverages (0-100 for all components with higher values indicating higher quality). Cox proportional hazards models were used to estimate hazard ratios (HR) for all-cause mortality risk across 27 separate joint tertile combinations of SPAN behaviours with the lowest tertile for all three as the referent group. For more granular clinical interpretations, we examined combined incremental dose-response changes of the SPAN behaviours using the 5th percentile of each behaviour as the referent point.

Results: Over the 8.1-year median follow-up time, 2,458 mortality events occurred. Compared to the referent group of combined SPAN exposure (lowest tertiles for all three), the optimal SPAN combination involving moderate sleep duration (7.2-8.0 h/day), high MVPA (42-103 min/day), and a DQS between 57.5 and 72.5 was associated with an HR of 0.36 (95% CI: 0.26-0.50). Relative to the 5th percentile of sleep (5.5 h/day), physical activity (7.3 min/day), and nutrition (36.9 DQS), a theoretical minimum combined increase of 15 min/day of sleep, 1.6 min/day MVPA, and 5 DQS points (corresponding to e.g., extra 1/2 serving of vegetables per day or 1 less serving of processed meat per week) was associated with 10% lower all-cause mortality risk (0.90; 0.88-0.93). Combined increases of 75 min/day of sleep, 12.5 min/day MVPA, and 25 DQS points were associated with 50% lower all-cause mortality risk (0.50; 0.44-0.58).

Conclusions: This study highlights the potential health value of subtle combined SPAN modification in relation to mortality risk and expands opportunities for more holistic recommendations.

Keywords: Cohort studies; Diet; Exercise; Mortality; Nutrition; Physical activity; Sleep.

Fig. 1

Multivariable-adjusted associations of combined Sleep, Physical Activity, and Nutrition with all-cause mortality risk. Legend: Model is adjusted for age, sex, ethnicity, smoking, education, Townsend deprivation index, alcohol, discretionary screen time (time spent watching TV or using the computer outside of work), light intensity physical activity, medication (blood pressure, insulin, and cholesterol), previous diagnosis of major CVD (defined as disease of the circulatory system, arteries, and lymph, excluding hypertension), previous diagnosis of cancer, and familial history of CVD and cancer (n = 59,078; events = 2,458). Sleep (hours/day), physical activity (moderate to vigorous intensity (MVPA) minutes/day), and nutrition (Dietary Quality Score (DQS)) were included in the model as a joint term. The specific ranges for each exposure included sleep duration as 4.8-7.2 hours/day (low), 7·2-8.0 hours/day (medium), and 8.0-9.4 hours/day (high); MVPA measurements as 5-23 minutes/day (low), 23-42 minutes/day (medium), and 42-103 minutes/day (high); and diet quality using the DQS as 32.5-50.0 (low), 50.0-57.5 (medium), and 57.5-72.5 (high). The lowest tertiles for all three exposures (sleep, MVPA and DQS) was considered the reference group. Dashed blue lines separate tertiles MVPA and dashed black lines separate tertiles of sleep. Sleep (Slp); Low Diet Quality (LD); Medium Diet Quality (MD); High Diet Quality (HD)

Fig. 2

Multivariable adjusted all-cause mortality risk associated with concurrent variations in Sleep, Physical Activity, and Nutrition. Legend: The correlogram displays changes in sleep (hours/day), physical activity (moderate to vigorous intensity (MVPA) minutes/day), and nutrition (Dietary Quality Score (DQS)) and corresponding mortality risk with the reference being the 5^th percentile of sleep (5.5 hours/day), physical activity (7.3 minutes/day), and nutrition (36.9 DQS). Sleep, physical activity, and nutrition are included as independent terms in the model to allow for more granular predictions. Each square on the grid represents the hazard ratio for all-cause mortality associated with a combination of behaviours, as defined by the x-axis (physical activity), y-axis (sleep), and z-axis (nutrition). The colour corresponds to the hazard ratio where red indicates a higher risk of all-cause mortality and green indicates a lower risk of all-cause mortality. Model is adjusted for age, sex, ethnicity, smoking, education, Townsend deprivation index, alcohol, discretionary screen time (time spent watching TV or using the computer outside of work), light intensity physical activity, medication (blood pressure, insulin, and cholesterol), previous diagnosis of major CVD (defined as disease of the circulatory system, arteries, and lymph, excluding hypertension), previous diagnosis of cancer, and familial history of CVD and cancer (n = 59,078; events = 2,458)