Abstract
Study Objectives:
Cohort studies about the sleep duration on the risk of death among Chinese older adults are still lacking. The aim of this study was to examine whether extremely long or short sleep duration was associated with mortality in Chinese adults aged 65 years or older.
Methods:
We included participants aged 65 years or older in 2011 at baseline in 23 provinces from the Chinese Longitudinal Healthy Longevity Survey who were followed up in 2014/2018 in China. Sleep duration was categorized as short sleep duration (< 7 hours) and long sleep duration (> 8 hours). We used the Cox proportional hazards model and restricted cubic spline analysis to explore the association between sleep duration and mortality.
Results:
Among 9578 participants, short sleep duration was associated with an 11% higher risk of death (adjusted hazard ratio [aHR]: 1.11; 95% confidence interval [CI]: 1.02–1.20) and long sleep duration was associated with a 24% higher risk of death (aHR: 1.24; 95% CI: 1.15–1.34), after adjustment for all covariates. There was a U-shaped association between sleep duration and all-cause mortality (nonlinear, P < .0001). Stratified analyses showed that the risk was higher among older people who smoked and with a higher level of education both for short and long sleepers than for those who never smoked and were illiterate (P value for interaction < .05).
Conclusions:
There was a U-shaped association between sleep duration and all-cause mortality in Chinese older adults, especially in more educated individuals and smokers.
Citation:
Du M, Liu M, Liu J. The association between sleep duration and the risk of mortality in the Chinese older adults: a national cohort study. J Clin Sleep Med. 2021;17(9):1821–1829.
Keywords: sleep duration, mortality, Chinese, older adults, cohort study
BRIEF SUMMARY
Current Knowledge/Study Rationale: The association between sleep duration and all-cause death among older adults is still inconsistent. In addition, nationwide prospective cohort studies of this subject among Chinese older people are lacking.
Study Impact: There was a U-shaped association between sleep duration and all-cause mortality in Chinese older adults, especially for more-educated individuals and smokers. Health care management could incorporate sleep education in plans to promote longevity and decrease costs of health care for the older population.
INTRODUCTION
Average human life expectancy has doubled during the last 200 years. The median age of adults has increased over time globally,1 which shows that the world is facing up to the great challenges of aging.2 China is encountering formidable aging challenges. According to the 2020 China Statistical Yearbook, there were 176 million people aged 65 or above, which accounted for 12.6% of the whole population in China.3 It is estimated that there will be 400 million Chinese citizens aged ≥ 65 years, 150 million of whom will be ≥ 80 years by 2050.4 The issue of aging implies broad global health implications. Increased life expectancy means a global burden of late-life disease, including chronic diseases such as diabetes, hypertension, etc, and mental health conditions such as dementia, which all are linked to a higher risk of death in older adults.5 Geriatric health has become a critical health concern in China, as well as in other countries.6 Therefore, determining factors that influence the health of older adults is becoming increasingly critical.
Sleep problems have been considered as one of the preventable risk factors. It was reported that older adults were more likely to have sleep problems, such as trouble falling asleep, daytime napping, and multiple nocturnal awakenings, etc.7,8 Among them, extremely short or long sleep duration was a vital risk factor for health outcomes among the older adult population, including memory deficits and limited attention span,8 depression,9 dementia,10 and even mortality. Until now, the results of studies investigating the relationship between sleep duration and mortality in elderly individuals have been controversial. A U-shaped association between hours of sleep and overall mortality has been reported in some epidemiological studies,11–13 while some studies reported that only long sleepers were at higher risk of death, not shorter-duration sleepers.14–18 Some longitudinal studies have shown that sleep duration was associated with mortality among older adults in America,16,19 Japan,10,20 Spain,12 and Israel.17 However, national longitudinal studies about the relationship between sleep duration and mortality among elderly individuals in China are still lacking. Hou et al11 reported that, compared with a sleep duration of 7–9 hours, the risk of total deaths was higher for short duration (< 7 hours; adjusted hazard ratio [aHR]: 1.08; 95% confidence interval [CI]: 1.03–1.13) and for long duration (> 9 hours; aHR: 1.12; 95% CI: 1.07–1.17) among the oldest old people aged ≥ 80 years in mainland China. Previous cohort studies that reported the association of sleep duration with mortality among Chinese older adults were mainly conducted in Taiwan.13–15,18 Furthermore, limited studies controlled for basic demographic characteristics (age, education, etc),21,22 lifestyle (physical activity, diet, etc),23,24 health status (body mass index [BMI], depression, disability, comorbidities, etc),21,22 and sociodemographic status (disability, childhood socioeconomic status [SES], and adult SES, etc),21 which were reported to be associated with sleep or mortality previously.
The relationship between sleep duration and mortality might be influenced by other factors, such as basic demographic characteristics, lifestyle, health, and sociodemographic status. Although several longitudinal studies have shown that sleep duration is associated with all-cause mortality among older adults, the results are still inconsistent and need to be clarified. Furthermore, nationwide prospective cohort studies that explored the association between sleep duration and all-cause mortality among Chinese individuals aged 65 years or older are lacking. Thus, we hypothesized that reduced and/or prolonged sleep duration predisposes individuals to an increased risk of all-cause mortality among individuals aged 65 or older in China after controlling for related covariates. We aimed to explore the association between sleep duration and the risk of death among adults aged 65 or older in a nationwide prospective cohort study in mainland China, after adjustment for potential confounders.
METHODS
Data source and study population
Our study analyzed data obtained from the Chinese Longitudinal Healthy Longevity Survey (CLHLS), an ongoing, prospective cohort study that covers 23 out of 31 provinces in China. This study was established in 1998, with subsequent follow-up and recruitment of new participants in 2000, 2002, 2005, 2008, 2011, 2014, and 2018. The CLHLS provides information regarding health status and quality of life of older adults aged 65 and older; more details on the study have been described elsewhere.25 The present analysis includes data from the seventh wave in 2011 (at baseline), which included the question, “How many hours do you sleep normally?” The follow-up survey was conducted in 2014 and 2018. The study was approved by the Research Ethics Committee of Peking University (IRB00001052-13074). All of the participants or their legal representatives signed written consent forms to participate in the baseline and follow-up surveys.
The 2011 survey wave included 9,765 Chinese older adults. We excluded 86 participants who were younger than 65 years old and 101 participants who were missing data on sleep duration. Hence, a total of 9,578 participants were included in the final analysis. Figure 1 shows the study flowchart of inclusion and exclusion criteria of the research participants.
Figure 1. Flowchart of the inclusion of participants.
CLHLS = Chinese Longitudinal Healthy Longevity Survey.
Assessment of sleep duration
Participants’ sleep duration was assessed by a questionnaire that included the question, “How many hours do you sleep normally?” Participants answered at baseline. According to the National Sleep Foundation’s sleep time duration recommendations for older adults (7–8 hours), sleep duration was classified into the short sleep duration (< 7 hours) group, normal sleep duration group (7–8 hours), and long sleep duration group (> 8 hours).26
Covariates
We attempted to examine as many factors as possible that have been found to be associated with hours of sleep and mortality in our analyses.21,27 Trained investigators collected information, including basic demographic and lifestyle characteristics and health and socioeconomic status, using a standardized questionnaire. All of the surveys were face-to-face interviews conducted at the participant’s home. If participants were illiterate, investigators helped them to complete the questionnaire.
Basic demographic characteristics included age (< 90/≥ 90 years), sex (male/female), education (no school/≥ 1 year), residence (urban/rural), marital status (unmarried/married/divorced or widowed), and living arrangement (living with family members/living alone or in an institution). Lifestyle included smoking status (nonsmoker/smoker), drinking status (nondrinker/drinker), regular exercise (yes/no), and dietary diversity score (good/poor), which was categorized according to the Food and Agriculture Organization of the United Nations recommendations and previous research.28 Health and socioeconomic status were measured by BMI (weight/height squared; kg/m2) (underweight/normal/overweight/obese), depression (yes/no/unknown), history of chronic diseases (hypertension, diabetes, heart diseases, and stroke; yes/no/unknown), activities of daily living (ADL in disability; yes/no), childhood SES (yes/no/unknown), and adult SES (good/poor).
Participants’ weights and standing heights were measured directly by trained investigators. The remaining variables were directly collected using the standardized questionnaire at baseline. According to the World Health Organization cutoff points, BMI was categorized as underweight (< 18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), or obese (≥ 30 kg/m2). Depression was assessed by the question, “Have you felt sad, blue, or depressed for two weeks or more in last 12 months?” ADL refers to basic personal-care tasks of everyday life. In this study, ADL in disability were defined as self-reported difficulty with any of the following ADL items29: dressing, eating, bathing, continence, toileting and cleaning, and indoor movement. We determined childhood SES (yes/no/unknown) and adult SES (good/poor) by asking how “often went to bed hungry as a child” and “how do you rate your economic status compared with other local people? A good adult SES was defined as an answer of “very rich” and “rich” and poor adult SES was defined as an answer of “so-so,” “poor,” or “very poor.”
Data analysis
Baseline characteristics of the study population are described as means ± standard deviations for continuous variables or percentages for categorical variables. We compared the characteristics of older adults among different sleep durations by using the chi-square test or Kruskal–Wallis test. Time to death (event = 1) was defined as the period between the baseline survey and death. Censoring (event = 0) was performed for surviving participants or those lost to follow-up in 2014; the censoring time was calculated from baseline to the survey time of 2014/2018, which was updated. Cox proportional hazards models were used to assess the association of sleep duration with all-cause mortality. We performed sensitivity analysis by fitting different models to examine the robustness of the estimation. Model 1 was a univariate model without adjustment for any confounders. Basic demographic characteristics were added in model 2, including age, sex, education, residence, marital status, and living arrangement. All of the covariates in model 3 were adjusted by adding smoking status; drinking status; regular exercise; dietary diversity; BMI; depression; ADL in disability; self-reported chronic diseases including hypertension, diabetes, heart diseases, and stroke; childhood SES; and adult SES. In addition, in order to test the robustness of the results, we conducted (1) an additional full model 3 by replacing categorical variables into continuous variables including age and BMI (Table S1 (130.7KB, pdf) in the supplemental material) and (2) we tested 6 hours as the threshold value of the short sleep duration (Table S2 (130.7KB, pdf) ). Our results are presented as pooled HRs with 95% CIs. We also tested for effect modification of the association of sleep duration with all-cause mortality by performing analyses stratified by age; sex; education; residence; marital status; living arrangement; smoking status; drinking status; regular exercise; dietary diversity; BMI; depression; ADL in disability; self-reported chronic diseases including hypertension, diabetes, heart diseases, and stroke; childhood SES; and adult SES. We used a heterogeneity test to examine the difference between groups; P values < .05 indicated significance. To examine the shape of the association between sleep duration and all-cause mortality, we conducted a restricted cubic spline analysis to the fully adjusted model. All of the analyses were performed with SPSS 26.0 (IBM Corporation, Armonk, NY), Stata 16.0 (StataCorp, College Station, TX), and R 3.4.0 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
Basic characteristics of the participants
Characteristics of the 9578 study participants at baseline are shown in Table 1. The mean ± standard deviation age was approximately 85.94 ± 11.21 years old, and 44.82% of participants were men. A total of 64.32% of participants had an extremely short or long sleep duration (< 7 hours: 2938, 30.67%; > 8 hours: 3223, 33.65%).
Table 1.
Characteristics of the study participants according to sleep duration at baseline.
| Median (IQR) or Total n | Sleep Duration (Hours) | K-W/χ2 | P | |||
|---|---|---|---|---|---|---|
| 7–8; n = 3,417 (35.68%) | < 7; n = 2,938 (30.67%) | > 8; n = 3,223 (33.65%) | ||||
| Demographic characteristics | ||||||
| Age (years), median (IQR) | 86.00 (76.00, 94.00) | 82.00 (74.00, 91.25) | 84.00 (76.00, 93.00) | 90.00 (80.00, 97.00) | 230.88 | < .000a |
| Age (years) | 359.436 | < .0001 | ||||
| <90 | 5,750 | 2,351 (40.89) | 1,885 (32.78) | 1,514 (26.33) | ||
| ≥90 | 3,828 | 1,066 (27.85) | 1,053 (27.51) | 1,709 (44.64) | ||
| Sex | 61.639 | < .0001 | ||||
| Male | 4,293 | 1,695 (39.48) | 1,170 (27.25) | 1,428 (33.26) | ||
| Female | 5,285 | 1,722 (32.58) | 1,768 (33.45) | 1,795 (33.96) | ||
| Education | 55.308 | < .0001 | ||||
| No school | 5,569 | 1,820 (32.68) | 1,757 (31.55) | 1,992 (35.77) | ||
| ≥1 year | 3,968 | 1,585 (39.94) | 1,162 (29.28) | 1,221 (30.77) | ||
| Residence | 0.822 | .663 | ||||
| Urban | 4,561 | 1,634 (35.83) | 1,379 (30.23) | 1,548 (33.94) | ||
| Rural | 5,017 | 1,783 (35.54) | 1,559 (31.07) | 1,675 (33.39) | ||
| Marital status | 177.234 | < .0001 | ||||
| Unmarried | 97 | 37 (38.14) | 28 (28.87) | 32 (32.99) | ||
| Married | 3,596 | 1,553 (43.19) | 1,077 (29.95) | 966 (26.86) | ||
| Divorced or widowed | 5,808 | 1,791 (30.84) | 1,815 (31.25) | 2,202 (37.91) | ||
| Living pattern | 52.553 | < .0001 | ||||
| Living with family members | 7,665 | 2,741 (35.76) | 2,237 (29.18) | 2,687 (35.06) | ||
| Living alone/or in an institution | 1,809 | 624 (34.49) | 675 (37.31) | 510 (28.19) | ||
| Lifestyle characteristics | ||||||
| Smoking status | 17.296 | < .0001 | ||||
| Nonsmoker | 6,261 | 2,148 (34.31) | 1,990 (31.78) | 2,123 (33.91) | ||
| Smoker | 3,225 | 1,238 (38.39) | 924 (28.65) | 1,063 (32.96) | ||
| Drinking status | 27.107 | < .0001 | ||||
| Nondrinker | 6,477 | 2,257 (34.85) | 2,097 (32.38) | 2,123 (32.78) | ||
| Drinker | 2,961 | 1,108 (37.42) | 801 (27.05) | 1,052 (35.53) | ||
| Regular exercise | 6.096 | .047 | ||||
| Yes | 4,298 | 1,575 (36.64) | 1,274 (29.64) | 1,449 (33.71) | ||
| No | 5,117 | 1,768 (34.55) | 1,622 (31.70) | 1,727 (33.75) | ||
| Dietary diversity score | 42.193 | < .0001 | ||||
| Poor | 4,873 | 1,622 (33.29) | 1,636 (33.57) | 1,615 (33.14) | ||
| Good | 4,658 | 1,775 (38.11) | 1,294 (27.78) | 1,589 (34.11) | ||
| Health and socioeconomic status | ||||||
| Body mass index (kg/m2), median (IQR) | 20.76 (18.37, 23.44) | 21.11 (18.73, 23.92) | 20.76 (18.37, 23.26) | 20.40 (18.05, 23.15) | 7.88 | .005a |
| Body mass index (kg/m2) | 37.078 | < .0001 | ||||
| Underweight (< 18.5) | 2,391 | 753 (31.49) | 751 (31.41) | 887 (37.10) | ||
| Normal (18.5–24.9) | 5,441 | 2,019 (37.11) | 1,678 (30.84) | 1,744 (32.05) | ||
| Overweight (25–29.9) | 1,077 | 427 (39.65) | 326 (30.27) | 324 (30.08) | ||
| Obese (≥ 30) | 290 | 113 (38.97) | 86 (29.66) | 91 (31.38) | ||
| Depression | 149.967 | < .0001 | ||||
| Yes | 1,198 | 354 (29.55) | 485 (40.48) | 359 (29.97) | ||
| No | 6,763 | 2,596 (38.39) | 2,007 (29.68) | 2,160 (31.94) | ||
| Unknown | 1,617 | 467 (28.88) | 446 (27.58) | 704 (43.54) | ||
| Hypertension | 30.353 | < .0001 | ||||
| Yes | 2,720 | 1,011 (37.17) | 901 (33.13) | 808 (29.71) | ||
| No | 6,455 | 2,277 (35.27) | 1,920 (29.74) | 2,258 (34.98) | ||
| Unknown | 403 | 129 (32.01) | 117 (29.03) | 157 (38.96) | ||
| Diabetes | 2.84 | .585 | ||||
| Yes | 395 | 140 (35.44) | 126 (31.90) | 129 (32.66) | ||
| No | 8,693 | 3,113 (35.81) | 2,667 (30.68) | 2,913 (33.51) | ||
| Unknown | 490 | 164 (33.47) | 145 (29.59) | 181 (36.94) | ||
| Heart diseases | 20.971 | < .0001 | ||||
| Yes | 1,165 | 421 (36.14) | 409 (35.11) | 335 (28.76) | ||
| No | 7,945 | 2,845 (35.81) | 2,387 (30.04) | 2,713 (34.15) | ||
| Unknown | 468 | 151 (32.26) | 142 (30.34) | 175 (37.39) | ||
| Stroke | 7.365 | .118 | ||||
| Yes | 793 | 266 (33.54) | 248 (31.27) | 279 (35.18) | ||
| No | 8,355 | 3,004 (35.95) | 2,573 (30.80) | 2,778 (33.25) | ||
| Unknown | 430 | 147 (34.19) | 117 (27.21) | 166 (38.60) | ||
| ADL in disability | 236.577 | < .0001 | ||||
| Yes | 2,471 | 658 (26.63) | 685 (27.72) | 1,128 (45.65) | ||
| No | 6,836 | 2,666 (39.00) | 2,181 (31.90) | 1,989 (29.10) | ||
| Childhood SES | 45.52 | < .0001 | ||||
| Yes | 6,322 | 2,175 (34.40) | 1,927 (30.48) | 2,220 (35.12) | ||
| No | 2,022 | 752 (37.19) | 584 (28.88) | 686 (33.93) | ||
| Unknown | 1,234 | 490 (39.71) | 427 (34.60) | 317 (25.69) | ||
| Adult SES | 26.253 | < .0001 | ||||
| Good | 1,620 | 608 (37.53) | 413 (25.49) | 599 (36.98) | ||
| Poor | 7,813 | 2,753 (35.24) | 2,489 (31.86) | 2,571 (32.91) | ||
Continuous variables are presented as medians (IQRs). Categorical variables are presented as n (%). aKruskal–Wallis test. Missing data: Demographic characteristics: education, n = 41 (0.43%); marital status, n = 77 (0.80%); living arrangement, n = 104 (1.09%); Lifestyle: smoking status, n = 92 (0.96%); drinking status, n = 140 (1.46%); regular exercise, n = 163 (1.70%); dietary diversity score, n = 47 (0.49%); Health status: body mass index, n = 379 (3.96%); ADL in disability, n = 271 (2.83%); adult SES 145 (1.5%). ADL = activities of daily living, IQR = interquartile range, K-W = Kruskal-Wallis test, SES = socioeconomic status.
Participants who were female, aged < 90 years, illiterate, divorced or widowed, living alone or in an institution, never smoked, never consumed alcohol, were underweight, had depression, had nonregular exercise, had poor dietary diversity, had a history chronic disease including hypertension and heart disease, without ADL disability, an unknown childhood SES, and poor adult SES were more likely to have short sleep duration (P < .05). Older participants who were female, aged ≥ 90 years, were illiterate, divorced or widowed, living with family members, nonsmokers, consumed alcohol, were underweight, had nonregular exercise, had good dietary diversity, had unknown depression status, a history of chronic disease including hypertension and heart disease, had ADL disability, and had poor childhood SES and good adult SES were more likely to have long sleep duration (P < .05). Importantly, extremely short or long sleep duration was more likely in those who were female, illiterate, divorced or widowed, underweight, never smoked, and had nonregular exercise. More details are shown in Table 1.
Association of sleep duration with all-cause mortality
During 7 years of follow-up, a total of 4,657 deaths were observed. In the unadjusted analysis, both short sleep duration and long sleep duration were positively associated with an increased risk of all-cause mortality (< 7 hours: hazard ratio [HR]: 1.16; 95% CI: 1.08–1.25; > 8 hours: HR: 1.72; 95% CI: 1.60–1.84). Upon adjusting for basic demographic characteristics including age, sex, education, residence, marital status, and living arrangement, the HR and 95% CI remained significant (< 7 hours: aHR: 1.12; 95% CI: 1.04–1.21; > 8 hours: aHR: 1.30, 95% CI: 1.21–1.40). In multivariable models, after adjustment for all of the covariates, the short sleep duration was associated with an 11% higher risk of mortality (aHR: 1.11; 95% CI: 1.02–1.20). Compared with the participants with normal sleep, the risk of all-cause mortality was still higher among long sleepers (aHR: 1.24; 95% CI: 1.15–1.34) (Table 2). Similar results were shown in the additional full model 3 (Table S1 (130.7KB, pdf) ) and the threshold value of the short sleep duration (< 6 hours) tested model (Table S2 (130.7KB, pdf) ). According to the restricted cubic spline analysis, the sleep duration from 7 to 8 hours had the lowest risk of death and the U-shaped association between sleep duration and all-cause mortality was significant (nonlinear, P < .0001), as shown in Figure 2.
Table 2.
Association of sleep duration with mortality in the univariate and multivariable models.
| Sleep Duration (Hours) | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| 7–8 | 1 (reference) | 1 (reference) | 1 (reference) | |||
| < 7 | 1.16 (1.08, 1.25) | < .0001 | 1.12 (1.04, 1.21) | .004 | 1.11 (1.02, 1.20) | .016 |
| > 8 | 1.72 (1.60, 1.84) | < .0001 | 1.30 (1.21, 1.40) | < .0001 | 1.24 (1.15, 1.34) | < .0001 |
Model 1 was a univariate model. In model 2, we adjusted for basic demographic characteristics, including age, sex, education, residence, marital status, and living arrangement. In model 3, we adjusted for all the covariates, including age, sex, education, residence, marital status, living pattern smoking status, drinking status, regular exercise, BMI, depression, ADL in disability, history of chronic disease (hypertension, diabetes, heart diseases, and stroke), childhood SES, and adult SES. ADL = activities of daily living, BMI = body mass index, CI = confidence interval, HR = hazard ratio, SES = socioeconomic status.
Figure 2. Associations of sleep duration with mortality in the study population of Chinese older individuals using Cox models with penalized splines after adjustment for confounders.
Lines represent estimated hazard ratios of sleep duration for risk of mortality; dotted lines represent for 95% CIs. CI = confidence interval.
Subgroup analysis
We stratified the analysis by age; sex; education; residence; marital status; living arrangement; smoking status; drinking status; regular exercise; dietary diversity; BMI; depression; ADL in disability; history of chronic diseases including hypertension, diabetes, heart diseases, and stroke; childhood SES; and adult SES. Although group differences in education and smoking status were significant for the association of extremely short or long sleep duration with all-cause mortality in the multivariable model (all P values for interaction < .05), the risk was higher among the older individuals with a higher level of education (< 7 hours: aHR: 1.31; 95% CI: 1.14–1.51; > 8 hours: aHR: 2.78; 95% CI: 2.49–3.09) and who smoked (< 7 hours: aHR: 1.28; 95% CI: 1.12–1.47; > 8 hours: aHR: 1.42; 95% CI: 1.25–1.61). Furthermore, group differences in marital status and dietary diversity were significant for the association of short sleep duration with mortality in the multivariable model (all P values for interaction < .05) (Figure 3 and Table S3 (130.7KB, pdf) ).
Figure 3. Subgroup analysis stratified by age, sex, education, residence, marital status, living arrangements, smoking status, drinking status, regular exercise, dietary diversity, body mass index, depression, ADL in disability, history of chronic disease (hypertension, diabetes, heart diseases, and stroke), childhood SES, and adult SES.
ADL = activities of daily living, CI = confidence interval, HR = hazard ratio, SES = socioeconomic status.
DISCUSSION
To our knowledge, this is the first nationwide cohort study that examined the association between sleep duration and all-cause mortality among Chinese older individuals aged ≥ 65 years old. This prospective cohort study found that older people who reported extremely long or short sleep duration had a significantly increased risk of all-cause mortality. There was a U-shaped association between sleep duration and all-cause mortality. In addition, the results suggested that education and smoking status modified the association of extremely short or long sleep duration with all-cause mortality in Chinese older individuals.
Although some studies have found that long sleep duration was associated with mortality among older people,15–18,20 our results are still in line with other studies. Jung et al19 found that mortality risk was U-shaped in both men and women aged 60–96 years in the United States, and the risk was lowest among those sleeping 7–7.9 hours per night. Samples from 12 southeastern states of the United States found that there was a U-shaped relationship between weekday and weekend sleep duration and all-cause mortality.30 In Spain, mortality was higher in those who slept < 5 hours, 8 hours, 9 hours, 10 hours, and 11 hours or more than in those who slept 7 hours (P = .001) among a cohort study of 3,820 persons aged 60 years and older.12 The similar associations still remained unchanged in community-dwelling Japanese individuals aged 60 years and older.10 However, a related, large-sample, nationwide cohort study of Chinese is lacking, and the relationship between sleep duration and all-cause mortality among Chinese older individuals is still unclear. One study reported a U-shaped association between sleep duration and all-cause mortality among the oldest people (> 80 years old) in China.11 Our study was conducted in a large area involving 23 research locations in 23 provinces in mainland China and controlled for basic demographic characteristics, lifestyle, health, and SES, which related to sleep duration or mortality. We found, after adjustment for all the covariates, compared with those who reported normal sleep duration, reported short sleep duration was associated with an 11% higher risk of death and long sleep duration was associated with a 24% higher risk of death. There was a U-shaped association between sleep duration and all-cause mortality. Therefore, we should pay more attention to the sleep duration of older people to reduce the risk of death.
Our study found group differences in education and smoking status were significant for the association of extremely long or short sleep duration. Higher education may be the important influencing factor on the association between sleep duration and death. Our study found that the risk of death was higher among the short and long sleepers with a higher level of education. This finding required more research to confirm and analyze the deeper reasons. Our study also found that the risk of death was higher among older individuals who smoked. Cigarette smoking is associated with the development of various lung diseases,31 which may cause death in elderly individuals. In addition, both smoking and sleep loss cause abnormal oxidative stress, which is associated with death.31,32 Although there were no group differences in other factors, the risk of mortality related to long sleep duration was higher among older men, whereas the risk of mortality related to long sleep duration was lower for older women. With regard to the impact of sex on mortality, many studies have found that women have a longer life expectancy.33 Our research also indirectly suggested that sex may have an impact on the relationship between sleep duration and mortality, which requires more research in the future. Furthermore, the effect of long sleep duration on mortality was higher among older people aged less than 90 years old compared with those aged 90 years or above. The phenomenon of “the younger participants were, higher the effect was” as was also seen in other studies.12,34,35 Depression may increase the risk of mortality among older people with long sleep duration. The association between long sleep duration (10 vs 7 hours) and mortality was higher among older adults with depression.12 Older people with long sleep duration without a history of stroke, heart diseases, and diabetes had a lower risk for mortality than all other participants. Zawisza et al35 reported that chronic disease had an influence on the association between long sleep duration and mortality. Based on previous research, our study suggests that sleep duration plays a key role in the health of older people and provides a reference for health care of older adults. In our study, approximately 64% of participants had an extremely short or long sleep duration, which revealed that the sleep quality of older people was poor. In addition, we found that older people who were female, without education, divorced or widowed, never smoked, did not have regular exercise, and were underweight were more likely to have an extremely short or long sleep duration. These results suggest sleep medicine health professionals should pay more attention to older people who are female, illiterate, never regularly exercise, are underweight, and divorced or widowed.
The specific potential mechanism of extremely long or short sleep duration–related death is still unclear. Inflammation may be an important biological explanation. Research shows that shorter and longer sleep duration are associated with higher levels of C-reactive protein, which may predict excessive weight gain among older adults; furthermore, excessive inflammation may cause some chronic diseases including type 2 diabetes, cardiovascular events, and hypertension, which all contribute to mortality.36 Reactive oxygen species (ROS) may be another potential mechanism. Animal experiments suggest that sleep deprivation can cause death through ROS accumulation in the gut, which triggers oxidative stress in this organ.32 Furthermore, the association between extremely short or long sleep duration and health status may be a bidirectional relationship. It is not clear whether this association is due to sleep duration itself or is a consequence of health status, which influences both sleep duration and mortality.12 In our study, we took health status into consideration, including chronic conditions, functional limitations, and depression. But the nonsignificant group difference in health status may indicate that extremely short or long sleep duration may predict mortality independently.
There were several limitations to our study. First, we included participants aged 65 years or older in mainland China, so the results might not be generalizable to older people in other countries because of different habits and customs in different countries. Second, the sleep duration was self-reported. Objective methods such as polysomnography to monitor sleep were not used in this study. Further studies including a more objective assessment are needed to confirm our findings. Furthermore, distinguishing daytime sleep duration and nighttime sleep duration may be more conducive to detailed analysis, so a validated questionnaire is needed in the future. However, the CLHLS collected sleep durations using the question, “How many hours do you sleep normally?” Third, some covariates (eg, adult SES and depression) were not measured with a validated questionnaire/scale, which should be improved in the future. Finally, the CLHLS did not collect the specific cause of deaths, so we could not explore the association of sleep duration with specific cause of death.
In this nationwide prospective cohort study, there was a U-shaped association between sleep duration and all-cause mortality in Chinese older adults, especially for educated individuals and smokers. Our findings suggest that shorter and longer sleepers are both high-risk populations and extremely short or long sleep duration should be a modifiable lifestyle risk factor for mortality. Health care management could incorporate sleep education in plans to promote longevity and decrease costs of health care for the older population.
DISCLOSURE STATEMENT
All authors gave approval for the final version of the manuscript. Work for this study was performed at the Department of Epidemiology and Biostatistics, School of Public Health, Peking University, Beijing, China. This study was funded by the National Key Research and Development Project of China (2019YFC1710301, 2020YFC0846300). The authors report no conflicts of interest.
ACKNOWLEDGMENTS
The authors thank the members of the CLHLS study group for data collection and all of the participants. Author contributions: J.L. conceptualized and designed the study; J.L. and M.D. performed data acquisition; M.D. performed data curation and formal analysis and visualization and wrote the draft manuscript. M.D., M.L., and J.L. revised the manuscript. The data are available on the CLHLS study website (https://opendata.pku.edu.cn).
ABBREVIATIONS
- ADL,
activities of daily living
- aHR,
adjusted hazard ratio
- BMI,
body mass index
- CI,
confidence interval
- CLHLS,
Chinese Longitudinal Healthy Longevity Survey
- HR,
hazard ratio
- SES,
socioeconomic status
REFERENCES
- 1. Chang AY , Skirbekk VF , Tyrovolas S , Kassebaum NJ , Dieleman JL . Measuring population ageing: an analysis of the Global Burden of Disease Study 2017 . Lancet Public Health. 2019. ; 4 ( 3 ): e159 – e167 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Partridge L , Deelen J , Slagboom PE . Facing up to the global challenges of ageing . Nature. 2018. ; 561 ( 7721 ): 45 – 56 . [DOI] [PubMed] [Google Scholar]
- 3.Liu A, Ye Z. Population age structure and dependency ratio. In: National Bureau of Statistics, ed. China Statistical Yearbook. China Statistics Press; 2020.
- 4. Zeng Y , George LK . Population ageing and old-age insurance in China . In: The SAGE Handbook of Social Gerontology . London: : SAGE Publications Ltd; ; 2010. : 420 – 429. [Google Scholar]
- 5. GBD 2016 DALYs and HALE Collaborators . Global, regional, and national disability-adjusted life-years (DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016 . Lancet. 2017. ; 390 ( 10100 ): 1260 – 1344 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Fang EF , Scheibye-Knudsen M , Jahn HJ , et al . A research agenda for aging in China in the 21st century. Ageing Res Rev. 2015. ;24(Pt B):197-205. [DOI] [PMC free article] [PubMed]
- 7. Cochen V , Arbus C , Soto ME , et al . Sleep disorders and their impacts on healthy, dependent, and frail older adults . J Nutr Health Aging. 2009. ; 13 ( 4 ): 322 – 329 . [DOI] [PubMed] [Google Scholar]
- 8. Crowley K . Sleep and sleep disorders in older adults . Neuropsychol Rev. 2011. ; 21 ( 1 ): 41 – 53 . [DOI] [PubMed] [Google Scholar]
- 9. Motivala SJ , Levin MJ , Oxman MN , Irwin MR . Impairments in health functioning and sleep quality in older adults with a history of depression . J Am Geriatr Soc. 2006. ; 54 ( 8 ): 1184 – 1191 . [DOI] [PubMed] [Google Scholar]
- 10. Ohara T , Honda T , Hata J , et al . Association between daily sleep duration and risk of dementia and mortality in a Japanese community . J Am Geriatr Soc. 2018. ; 66 ( 10 ): 1911 – 1918 . [DOI] [PubMed] [Google Scholar]
- 11. Hou C , Lin Y , Zimmer Z , Tse LA , Fang X . Association of sleep duration with risk of all-cause mortality and poor quality of dying in oldest-old people: a community-based longitudinal study . BMC Geriatr. 2020. ; 20 ( 1 ): 357 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Mesas AE , López-García E , León-Muñoz LM , Guallar-Castillón P , Rodríguez-Artalejo F . Sleep duration and mortality according to health status in older adults . J Am Geriatr Soc. 2010. ; 58 ( 10 ): 1870 – 1877 . [DOI] [PubMed] [Google Scholar]
- 13. Chien KL , Chen PC , Hsu HC , et al . Habitual sleep duration and insomnia and the risk of cardiovascular events and all-cause death: report from a community-based cohort . Sleep. 2010. ; 33 ( 2 ): 177 – 184 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lan TY , Lan TH , Wen CP , Lin YH , Chuang YL . Nighttime sleep, Chinese afternoon nap, and mortality in the elderly . Sleep. 2007. ; 30 ( 9 ): 1105 – 1110 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lee WJ , Peng LN , Liang CK , Chiou ST , Chen LK . Long sleep duration, independent of frailty and chronic Inflammation, was associated with higher mortality: a national population-based study . Geriatr Gerontol Int. 2017. ; 17 ( 10 ): 1481 – 1487 . [DOI] [PubMed] [Google Scholar]
- 16. Stone KL , Ewing SK , Ancoli-Israel S , et al . Self-reported sleep and nap habits and risk of mortality in a large cohort of older women . J Am Geriatr Soc. 2009. ; 57 ( 4 ): 604 – 611 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Cohen-Mansfield J , Perach R . Sleep duration, nap habits, and mortality in older persons . Sleep. 2012. ; 35 ( 7 ): 1003 – 1009 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chen HC , Su TP , Chou P . A nine-year follow-up study of sleep patterns and mortality in community-dwelling older adults in Taiwan . Sleep. 2013. ; 36 ( 8 ): 1187 – 1198 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Jung KI , Song CH , Ancoli-Israel S , Barrett-Connor E . Gender differences in nighttime sleep and daytime napping as predictors of mortality in older adults: the Rancho Bernardo Study . Sleep Med. 2013. ; 14 ( 1 ): 12 – 19 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Kawachi T , Wada K , Nakamura K , et al . Sleep duration and the risk of mortality from stroke in Japan: the Takayama Cohort Study . J Epidemiol. 2016. ; 26 ( 3 ): 123 – 130 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Gu D , Sautter J , Pipkin R , Zeng Y . Sociodemographic and health correlates of sleep quality and duration among very old Chinese . Sleep. 2010. ; 33 ( 5 ): 601 – 610 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Reis C , Dias S , Rodrigues AM , et al . Sleep duration, lifestyles and chronic diseases: a cross-sectional population-based study . Sleep Sci. 2018. ; 11 ( 4 ): 217 – 230. 10.5935/1984-0063.20180036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Driver HS , Taylor SR . Exercise and sleep . Sleep Med Rev. 2000. ; 4 ( 4 ): 387 – 402 . [DOI] [PubMed] [Google Scholar]
- 24. Duke CH , Williamson JA , Snook KR , Finch KC , Sullivan KL . Association between fruit and vegetable consumption and sleep quantity in pregnant women . Matern Child Health J. 2017. ; 21 ( 5 ): 966 – 973 . [DOI] [PubMed] [Google Scholar]
- 25. Zeng Y , Feng Q , Hesketh T , Christensen K , Vaupel JW . Survival, disabilities in activities of daily living, and physical and cognitive functioning among the oldest-old in China: a cohort study . Lancet. 2017. ; 389 ( 10079 ): 1619 – 1629 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Hirshkowitz M , Whiton K , Albert SM , et al . National Sleep Foundation’s sleep time duration recommendations: methodology and results summary . Sleep Health. 2015. ; 1 ( 1 ): 40 – 43 . [DOI] [PubMed] [Google Scholar]
- 27. Lv YB , Gao X , Yin ZX , et al . Revisiting the association of blood pressure with mortality in oldest old people in China: community based, longitudinal prospective study . BMJ. 2018. ; 361 : k2158 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Yin Z , Fei Z , Qiu C , et al . Dietary diversity and cognitive function among elderly people: a population-based study . J Nutr Health Aging. 2017. ; 21 ( 10 ): 1089 – 1094 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Katz S , Ford AB , Moskowitz RW , Jackson BA , Jaffe MW . Studies of illness in the aged: the index of ADL: a standardized measure of biological and psychosocial function . JAMA. 1963. ; 185 ( 12 ): 914 – 919 . [DOI] [PubMed] [Google Scholar]
- 30. Xiao Q , Blot WJ , Matthews CE . Weekday and weekend sleep duration and mortality among middle-to-older aged White and Black adults in a low-income southern US cohort . Sleep Health. 2019. ; 5 ( 5 ): 521 – 527 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Aoshiba K , Nagai A . Oxidative stress, cell death, and other damage to alveolar epithelial cells induced by cigarette smoke. Tob Induc Dis. 2003. ;1(3):219-226. https://doi.org/10.1186/1617-9625-1-3-219 [DOI] [PMC free article] [PubMed]
- 32. Vaccaro A , Kaplan Dor Y , Nambara K , et al . Sleep loss can cause death through accumulation of reactive oxygen species in the gut . Cell. 2020. ; 181 ( 6 ): 1307 – 1328, e15 . [DOI] [PubMed] [Google Scholar]
- 33. World Health Organization . World health statistics 2019: monitoring health for the SDGs, sustainable development goals. Geneva, Switzerland: World Health Organization, 2019. . https://apps.who.int/iris/handle/10665/324835 . Accessed June 2, 2021.
- 34. Åkerstedt T , Ghilotti F , Grotta A , Bellavia A , Lagerros YT , Bellocco R . Sleep duration, mortality and the influence of age . Eur J Epidemiol. 2017. ; 32 ( 10 ): 881 – 891 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Zawisza K , Tobiasz-Adamczyk B , Galas A , Brzyska M . Sleep duration and mortality among older adults in a 22-year follow-up study: an analysis of possible effect modifiers . Eur J Ageing. 2015. ; 12 ( 2 ): 119 – 129 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Irwin MR , Olmstead R , Carroll JE . Sleep disturbance, sleep duration, and inflammation: a systematic review and meta-analysis of cohort studies and experimental sleep deprivation . Biol Psychiatry. 2016. ; 80 ( 1 ): 40 – 52 . [DOI] [PMC free article] [PubMed] [Google Scholar]