Longitudinal Association Study On Sleep And Physical Activity In Middle Aged And Older People

Aim: Nowadays,the aging of population has become a major problem that lots of countries and regions need to face.Sleep and physical activity are key factors in the prevention and management of chronic health conditions in the middle-aged and older population,including cardiovascular disease,overweight and obesity,depression and anxiety,etc.It is of great public health significance to study and intervene the longitudinal relationship between them.However,most of the previous studies on the relationship between sleep and physical activity had short follow-up time and small sample size,and failed to use the longitudinal data to explore the longitudinal association between sleep and physical activity from between-person and within-person levels.Accordingly,we carried out a large sample prospective study to evaluate the longitudinal causal association between sleep duration,sleep quality and physical activity,and to reveal the dominant factor of the association between sleep and physical activity,which can provide a scientific basis for formulating health policy,reducing the prevalence of chronic diseases and achieving healthy aging.Methods: The data derived from the English Longitudinal Study of Aging(ELSA).ELSA began in 2002,with a total of 12099 participants.Follow-up visits were conducted every 2years and nurse interviews were conducted every 4 years.Up to now,ELSA has completed9 surveys(2002-2018).We used the data from the 4th(T1,2008-2009),6th(T2,2012-2013)and 8th(T3,2016-2017)surveys containing sleep data.Finally,a total of 6450 participants aged 50 years and above were included.Sleep duration,sleep quality and physical activity were assessed by self-reported questionnaire,in which sleep duration was classified as normal sleep duration(6-8 hours)and abnormal sleep duration(< 6 hours or > 8 hours).Sleep quality was assessed by the calculated sleep disorder score.The higher the sleep disorder score,the worse the sleep quality.Physical activity was categorized as very low,low,moderate and high according to the type and frequency of self-reported physical activity.Longitudinal causal associations of sleep duration and sleep disorder score with physical activity were assessed at both between-person and within-person levels by fitting cross-lagged path model(CLPM)and random-intercept cross-lagged path model(RICLPM).We compare the standardized cross-hysteresis path coefficients to determine the direction of the association and the dominant factors.For the robustness of the results and the simplicity of the model,we tested the time-invariance of the cross-lagged path,autoregressive path and autocorrelation path of CLPM and RI-CLPM.We performed three sensitivity analyses:(1)repeating the above analysis after classifying sleep duration as 5-9hours;(2)fitting the CLPM using data from T1 and T3;(3)excluding individuals with missing values for the primary study variables and repeating the above analysis.All models were adjusted for age,sex,marital status,education level,smoking status,drinking status,and body mass index.Results: There were 2,866(44.4%)males and 1276(20.2%)participants with abnormal sleep duration among the 6,450 participants in T1 period.The median age of these participants was 62 years old,and the median sleep disorder score was 2.00.There were405(6.3%),1212(18.8%),3336(51.7%),and 1495(23.2%)participants with very low,low,moderate,and high physical activity levels,respectively.(1)The CLPM results of abnormal sleep duration and physical activity showed that the cross-lagged coefficients between abnormal sleep duration in the previous period and physical activity in the next period were-0.076(P<0.01)and-0.080(P<0.01)respectively,and the cross-lagged coefficients between physical activity in the previous period and abnormal sleep duration in the next period were-0.064(P<0.01)and-0.069(P<0.01),respectively.It showed that at the between-person level,participants with abnormal sleep duration had lower levels of physical activity in the next period than those with normal sleep duration,and vice versa.According to the comparison of standardized cross-lagged coefficients,abnormal sleep duration was the dominant factor.(2)The RI-CLPM results of abnormal sleep duration and physical activity showed that the cross-lagged coefficients between abnormal sleep duration in the previous period and physical activity in the next period were-0.029(P=0.65)and-0.026(P=0.66)respectively,and the cross-lagged coefficients between physical activity in the previous period and abnormal sleep duration in the next period were 0.054(P=0.55)and 0.063(P=0.55),respectively.It showed that at the within-person level,sleep duration of participants tending to be abnormal compared with their own average subsequently reported no association with physical activity,and vice versa.(3)The CLPM results of sleep disorder score and physical activity showed that the cross-lagged coefficients between sleep disorder score in the previous period and physical activity in the next period were-0.074(P<0.05)and-0.065(P<0.05)respectively,and the cross-lagged coefficients between physical activity in the previous period and sleep disorder score in the next period were-0.019(P<0.01)and-0.023(P<0.01),respectively.It showed that at the between-person level,participants with higher sleep disorder score had lower levels of physical activity in the next period than those with lower sleep disorder score,and vice versa.According to the comparison of standardized cross-lagged coefficients,sleep disorder score was the dominant factor.(4)The RI-CLPM results of sleep disorder score and physical activity showed that the cross-lagged coefficients between sleep disorder score in the previous period and physical activity in the next period were-0.031(P<0.05)and-0.028(P<0.05)respectively,and the cross-lagged coefficients between physical activity in the previous period and sleep disorder score in the next period were-0.012(P=0.62)and-0.014(P=0.62),respectively.It showed that at the within-person level,sleep disorder score of participants tending to be higher compared with their own average subsequently reported decrease in the levels of physical activity,while the reduction in participants’ level of physical activity compared to their previous average had no effect on sleep disorder score in the next period.The results of the sensitivity analysis were basically consistent with the main analysis.Conclusions: Abnormal sleep duration and physical activity were negatively associated in both directions at the between-person level,and abnormal sleep duration was the dominant factor.At the within-person level,there is no association between abnormal sleep duration and physical activity.Sleep disorder score and physical activity were negatively associated in both directions at the between-person level,and abnormal sleep duration was the dominant factor.At the within-person level,there was a one-way negative association between the sleep disorder score and physical activity.Taken together,sleep was the dominant factor in the longitudinal causal association between sleep and physical activity in middle aged and older people.Early interventions targeting one healthy behavior in sleep or physical activity may help improve the status of another.