Characteristics and correlates of sleep duration, daytime napping, snoring and insomnia symptoms among 0.5 million Chinese men and women

Background Inadequate sleep duration and insomnia can affect both physical and mental health. There is limited evidence, however, on characteristics and correlates of sleep patterns and insomnia in urban and rural China. Methods This cross-sectional study, involving 512,891 adults aged 30–79 years from ten (five urban and five rural) diverse areas in China, recorded detailed information, using interviewer-administered laptop-based questionnaires, on sleep patterns (duration, daytime napping and snoring) and insomnia symptoms. Logistic regression was used to examine the associations of sleep patterns and insomnia symptoms with a range of socio-economic, lifestyle, behaviour and health-related factors. Results Overall, the mean (SD) sleep duration was 7.38 (1.37) h, with 23% reporting short (≤6 h) and 16% reporting long (≥9 h) sleep duration, 21% taking daytime naps and 22% having frequent snoring. Overall, 17% reported having insomnia symptoms, with a higher proportion in women than in men (19% vs 13%), in rural than in urban residents (19% vs 15%), and in individuals who were living alone (23%). The adjusted odds ratios (ORs) of having insomnia symptoms were significantly higher among people with major depressive episodes (6.10, 95% CI: 5.69–6.55), generalised anxiety disorders (7.46, 6.65–8.37) and any chronic diseases (1.46; 1.44–1.49). In contrast, the ORs of insomnia symptoms were significantly lower among those reporting napping (0.77, 0.75–0.78) and frequent snoring (0.86, 0.84–0.87). Conclusions Among Chinese adults, sleep patterns varied greatly by socio-economic, lifestyle and health-related factors. The risk of insomnia symptoms was associated with both poor mental and physical health status.


Introduction
Insomnia is a common problem in middle-aged and older people and has been associated with poor health status. Recent systematic reviews of prospective studies conducted mainly in Western populations have demonstrated that individuals who reported either short sleep (ie <6 h) or long sleep duration (ie >9 h) had a higher prevalence of cardiovascular risk factors (hypertension, diabetes and obesity) and higher risks of both cardiovascular disease and all-cause mortality [1,2]. However, relatively little is known about the correlates of sleep duration or sleep disturbances or about the characteristics and determinants of sleep duration and insomnia in low-and middle-income countries such as China.
In recent decades in China, there has been a rapid economic transition accompanied by major changes in lifestyle, including changes in sleep duration and sleep habits. The increased use of night-time TV, internet, mobile phones and social media has resulted in reduced sleep duration and increased prevalence of insomnia. Such changes varied substantially by demographic, socioeconomic, lifestyle and health factors. Previous studies of the correlates and health consequences of sleep patterns and insomnia in China have been constrained by small sample size [3], collection of only limited data on other lifestyle, behavioural and health-related correlates, and restriction to local occupational or urban cohorts [4]. We report on findings from a nationwide study of 0.5 million Chinese adults in the China Kadoorie Biobank (CKB). The aims of the present study were to examine the following: the patterns of sleep duration, napping, snoring and prevalence of insomnia symptoms, overall and in certain population subgroups (eg by age, gender and region); the socio-economic and lifestyle correlates of sleep duration (short and long) and insomnia symptoms and the association of sleep duration and insomnia symptoms with physical and mental disorders (eg depression and anxiety).

Study participants
Details of the design and methods of the CKB study have been described elsewhere [5,6]. Briefly, 512,891 women and men, aged 30e79 years (mean 52 years), were recruited between 2004 and 2008 from 1175 local communities across ten (five rural and five urban) geographically defined areas in China. The regions were selected according to local disease patterns, exposure to certain risk factors, population stability, qualities of local death and disease registries, local commitment and capacity. In each chosen area, permanent residents aged 30e79 years old were identified through local residential records and invited to participate in the study by letter and information leaflet. Overall, approximately 30% (33% rural, 27% urban areas) of potentially eligible residents participated in the study. Nonparticipants were mainly those who were absent from home or reluctant to spend time to visit the local assessment centre according to anecdotal reports by field staff. Periodic resurveys of~5% of randomly selected surviving participants were conducted using identical approaches. Ethics approval was obtained from all relevant local, national and international authorities prior to commencement of the study. All participants provided written informed consent.
The baseline survey and subsequent re-surveys were conducted at local assessment centres set up specifically for the study. Trained health workers administered laptop-based questionnaire on demographic and socio-economic status, prior medical history, health status, smoking, alcohol consumption, diet, physical activity and other aspects of lifestyle (http://www.ckbiobank.org). The questionnaire had built-in checks to identify and minimise missing items, data entry errors and inconsistencies. Physical measurements including height, weight, waist and hip circumference, heart rate and blood pressure were also recorded. A 10-ml non-fasting blood sample (with time since last meal recorded) was collected for longterm storage. Standard operational procedures, training manuals and onsite training were provided to the local study team in each of ten areas. All devices were regularly maintained and calibrated to ensure consistency of the measurements across the ten areas and over time.

Assessment of insomnia symptoms, napping, snoring and sleep duration
Participants were first asked whether they had any of the following four symptoms in at least three days or more in a week during the last month: (i) taking >30 min to fall asleep after going to bed or waking up in the middle of the night, (ii) waking up early and not being able to go back to sleep, (iii) needing to take medicine (including herbal or sleeping pills) at least once a week to help sleep and (iv) having difficulty staying alert while at work, eating or meeting people during daytime. Participants who answered 'Yes' to any of the above i, ii and iii symptoms were classified as having insomnia symptoms (Section 10.3 Baseline questionnaire: http:// www.ckbiobank.org/site/binaries/content/assets/resources/pdf/ qs_baseline-final-from10june2004.pdf).
Subsequently, participants were asked the following multiple choice questions: Do you usually take a daytime nap? ('Yes, usually'; 'Yes, but only in summer' and 'No'); Do you snore during sleep? ('Yes, Frequently'; 'Yes, Sometimes' or 'No/Don't know'). Finally, participants were asked 'How many hours do you typically sleep per day including naps?'

Assessment of physical and mental health
Participants were asked if they had ever been diagnosed by a doctor as having any major diseases, including ischaemic heart disease, stroke/TIA, emphysema/chronic bronchitis, diabetes mellitus and cancer (Section 7 Baseline questionnaire: http://www. ckbiobank.org).
Participants were assessed using the CIDI-SF (A) questionnaire for major depressive episode (MDE) and CIDI-SF (B) for generalised anxiety disorder (GAD) [7]. Those who met the CIDI-SF criteria (A) were classified as having had a MDE, and those who met CIDI-SF (B) [7] were classified as having GAD. Questions on ten major stressful life events related to death of spouse, family conflict, financial difficulty, violence, etc. were also included, along with level of life satisfaction and level of self-rated health status (http://www. ckbiobank.org/site/binaries/content/assets/resources/pdf/qs_ baseline-final-from10june2004.pdf).

Statistical methods
The adjusted mean sleep duration (in hours) and proportions of individuals reporting insomnia symptoms, napping and frequent snoring were calculated by direct standardisation to the age (in ten year groups), gender and study area (ten groups) structure of the CKB cohort. Logistic regression was used to estimate adjusted odds ratios (OR) of short (6 h) and long (9 h) sleep duration (reference group: 7e8 h) and OR of insomnia symptoms by selected baseline characteristics adjusted for age, gender, region, body mass index (BMI) (five groups: <18.5, 18.5e24.99, 25e29.99, 30e34.99 and 35 kg/m 2 ), systolic blood pressure (SBP mmHg) (<110, 110e129, 130e139, >140), smoking (three groups: never regular, former regular and current regular smoker), alcohol (four groups: never regular, former regular, occasional and weekly drinkers) and metabolic equivalent of task (MET) (four groups: <10, 10e19.99, 20e29.99 and 30 MET-h per day). For variables with three or more groups, 'floating' standard errors [8] were used to facilitate comparisons between any two groups. These analyses were repeated after excluding individuals with prior disease (any of ten diseases reported at study entry) and MDE or GAD as sensitivity analyses. All analyses were conducted in R version 3.3.1 (R Core Team, 2016).

Results
The overall mean (SD) sleep duration was 7.38 (1.37) h and was higher (7.46 h) among those without prior diseases, MDE or GAD (Table 1 and Web Table 1). The duration of sleep decreased with age, whereas the converse was true for the proportion of individuals reporting insomnia symptoms (Fig. 1). Compared with  Table 1) and a higher proportion of insomnia symptoms, especially after age 50 years. Participants living in urban areas had shorter sleep duration (7.21 vs 7.51 h; Table 1), but a lower proportion reported insomnia symptoms across all age groups (Fig. 1). The patterns were unaltered after excluding individuals with prior diseases, MDE or GAD (Web Table 1). The proportion of individuals reporting daytime napping was higher in men than in women, in urban than in rural areas (36.4% vs 8.5%) and increased with age (Table 1). However, the proportion reporting only seasonal napping (ie napping only in the summer) was much higher in rural than in urban areas (58.3% vs 16.3%; data not shown). The proportion reporting frequent snoring was higher in men, in urban residents and those with annual household income more than 30,000 RMB (Table 1).

Patterns and correlates of sleep duration, daytime napping and snoring
Overall, about 23% reported short sleep duration (ie 6 h), and 16% reported long sleep duration (ie 9 h; Table 1). Exclusion of people with prior diseases and 12-month MDE or GAD significantly decreased the proportion of individuals with short but not long sleep duration (Web Table 1). Among people without prior diseases and MDE or GAD, the proportion of individuals reporting short sleep duration was more common at age >70 years, in women, in those with low socio-economic status (including no formal education, annual household income <10,000 RMB and unemployed) and in those with high physical activity (MET >30 h per day) (Web Table 1). The proportion reporting short sleep duration in participants without prior diseases was higher in those who were widowed (28%), living alone (30%), not employed (21%) and ex-regular smokers (28%) (Web Table 1). The proportion of individuals reporting long sleep duration in participants without prior diseases was higher in those with higher BMI (30 kg/m 2 ) or higher SBP (140 mmHg). Although low BMI was associated with short sleep duration (Table 1), the association became U-shaped after excluding prior diseases including 12-month MDE and GAD (Web Table 1).
The results for short and long sleep duration after excluding individuals with any insomnia symptoms are shown in Fig. 2A (Fig. 2A). Compared to short sleep duration, those reporting long   (Fig. 2B).
Overall, one-fifth of participants reported daytime napping, with a higher proportion in individuals with higher socio-economic status (eg university education, household income 35,000 RMB and office worker), married, not living alone, non-smokers and regular drinkers. Moreover, daytime napping was also associated with higher levels of BMI and lower levels of physical activity (Table 1). In the study, almost a quarter of the participants reported frequent snoring, and it was associated with various socioeconomic and lifestyle factors similar to those associated with daytime napping. Individuals reporting daytime napping or frequent snoring were less likely to report insomnia symptoms. Snoring was associated with slightly longer sleep duration (5 min; P < 0.0001, data not shown).

Characteristics and correlates of insomnia symptoms
Overall, 16.8% of the participants reported having insomnia symptoms (Table 1). Of those surveyed, 69% had difficulty falling sleep and 64% waking up early. About 7% of individuals with insomnia symptoms used medication, and 14% participants had difficulty staying alert during daytime (Table 2). Among those that answered insomnia questions, 58% had one insomnia symptom, 31% had two insomnia symptoms, and very few had three or more insomnia symptoms (Table 2).
Individuals reporting insomnia symptoms were more likely to have lower socio-economic status, with a particularly high rate among those with poor education, lower household income, widowed, living alone, unemployed, ex-regular smokers or ex-regular drinkers (Table 1). Additionally, in all individuals, after adjusting for age, sex, region, smoking, alcohol, MET, snoring, SBP and BMI, Each solid square represents an odds ratio (OR), and the size of the squares is inversely proportional to the variance of the log OR in that group after considering the variance of the log risk in the reference group where appropriate. The horizontal lines represent the 95% CI. All ORs were stratified by age, gender and region and adjusted for BMI, smoking, alcohol, MET-h/day and systolic blood pressure (SBP). Prior disease was defined as having been diagnosed by a doctor with at least one of the following diseases: diabetes, coronary heart disease, stroke or TIA, hypertension, rheumatic heart disease, tuberculosis, emphysema/bronchitis, asthma, cirrhosis/hepatitis, peptic ulcer, gallbladder disease, kidney disease, rheumatoid arthritis, head injury and cancer. hypertension was associated with slightly but statistically significantly longer sleep duration (2 min; P < 0.0001; data not shown), BMI (kg/m 2 ) was also significantly associated with longer sleep duration (about 8e9 min in individuals with BMI 18.5 compared to underweight individuals; P < 0.0001; data not shown). With the same adjustments, hypertension was not associated with insomnia symptoms, and BMI (kg/m 2 ) had an inverse dose-response association with insomnia symptoms (ORs 0.74, 0.62 and 0.56 for BMI 18.5e24.99, 25e29.99 and 30, respectively, compared to <18.5; data not shown). These associations remained largely unaltered after excluding prior diseases and 12-month MDE and GAD (Web Table 1). Participants who rated themselves unsatisfied or very unsatisfied with their life were much more likely to report insomnia symptoms than those who rated themselves very satisfied (OR: 2.53; 95% CI: 2.46e2.61; Fig. 3). Similarly, participants who rated their health status as poor were also more likely to have insomnia symptoms than those with excellent self-rated health status (OR 3.78; 3.71e3.85; Fig. 3). As with life satisfaction and level of selfrated health, there was a positive dose-response relationship between the number of stressful life events and insomnia symptoms (Fig. 3). Similarly, individuals who had MDE or GAD during the previous 12 months were 6e7 times more likely to have insomnia symptoms, and those with prior history of mental disorders were six times more likely to have insomnia symptoms (Fig. 3). Prior diseases were also found to be associated with moderately increased odds of insomnia symptoms (Fig. 3). Because major life events could affect sleep, in sensitivity analyses, we excluded individuals who experienced major life events, and the results remained unchanged (data not shown). We also examined ORs of insomnia symptoms by physical illness and mental disorders in those with short (6 h) and long (9 h) sleep duration, but the patterns remained unchanged (data not shown).

Discussion
This is the largest study to date of sleep patterns and correlates in middle-aged Chinese men and women. The study showed that sleep duration shortened as age increased, and women had shorter sleep duration and more insomnia symptoms at least after 50 years old, whereas rural residents had longer sleep duration but higher proportion of insomnia symptoms. Moreover, short sleep duration and insomnia symptoms were strongly associated with poor mental health status, particularly depression, and, albeit to a lesser extent, with poor physical health status.
Several population-based studies have previously reported associations of either short or long sleep duration among adults with mental and physical diseases. In the present study, the mean sleep duration was 7.46 h among those without major health problems, consistent with those previously reported in much smaller studies in China (7.76 h) conducted during 2009 [3], but appeared to be longer than those reported in other studies in high-income countries such as Korea (7 h) [9] or North America (6.99 h) [10]. In the present study, the total sleep duration also included daytime napping, which is much more common in China than in the US and Korea, which may explain the slightly longer duration observed in the present study.
A few large cross-sectional studies have examined the correlates of sleep duration in high-income countries such as US and Korea [9,10]. In general, these studies demonstrated that both short (6 h) and long (9 h) sleep duration were each associated with older age, women and lower SES. The present study in Chinese adults demonstrated similar associations with these factors. The association of short sleep duration with obesity has been reported in both children and adults [11]. A meta-analysis of 17 cross-sectional studies, involving 600,000 individuals aged 15e102 years, showed that short sleep duration was associated with 55% and 89% higher risks of obesity in adults and children, respectively [11]. However, a small cross-sectional study of about 6000 adults aged >15 years in Beijing, China [3], reported that individuals with short sleep duration had lower BMI, similar to what we found in the present study without excluding prior diseases. After excluding those with prior disease and 12-month MDE and GAD, both high BMI and low BMI were associated with short sleep duration. The relationship of BMI with sleep duration in adults warrants further study in prospective studies with longitudinal measurements adjusting for regression dilution of both BMI and sleep duration [12].
Insomnia symptoms were more often associated, although not exclusively, with short sleep duration. We examined the associations of mental and physical illness with short and long sleep duration while excluding participants who reported any insomnia symptoms. Both short and long sleep duration were associated with higher risks of psychiatric disorders including depressive and anxiety disorders in studies conducted in the US [13], Korea [9] and China [3], but most such studies included only a modest number of participants. However, in a Korean study of 6510 adults [9], only short sleep duration was associated with increased risk of depression and anxiety. The present study showed that people with several mental health-related traits (eg poor life satisfaction, poor self-rated health status, stressful life events, and 12-month MDE and GAD) were much more likely to have short and, to a lesser extent, long sleep duration. Short or long sleep duration may represent a marker of susceptibility to psychiatric disorders, including depression (as either may indeed indicate the development of depression) [13]. A meta-analysis of prospective studies, involving 474,684 adults with 16,000 cardiovascular events, showed that both short and long sleep duration were associated with higher risks of cardiovascular diseases [14]. Another US study of over 55,000 adults also showed that in addition to CVD, the risks of hypertension and diabetes were also elevated among people with either short or long sleep duration [2,15]. The findings of the present study were broadly consistent with previous crosssectional studies conducted in Chinese adults [4] [16] and with prospective studies [14].
In the present study, insomnia symptoms were comparable to those used in Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) [17] or International Classification of Sleep Disorders (ICSD) [18] and Research Diagnostic Criteria (RDC) for insomnia [19]. Among individuals without apparent physical and mental disorders, the proportion reporting at least one insomnia symptom in the present study (17%) was less than half of that reported in a US population using the same criteria [20] and lower than the estimates in two small cross-sectional studies in China [3,21], in which sleep disturbance was measured by Pittsburgh Sleep Quality Index (PSQI). The reasons for the apparent discrepancy are not clear but could be due at least in part to the use of a non-representative sample and low overall response rate in the present study, which could lead to underestimation of the prevalence of insomnia symptoms. Insomnia symptoms can be either a risk factor or symptom of mental disorders. A few prospective studies, including the US-based HUNT study of 24,715 people, reported that insomnia predicts long-term risk of several major physical and mental diseases [22,23]. The present cross-sectional study demonstrated strong positive associations of insomnia symptoms with a range of mental conditions including MDE, GAD and prior mental disorders. Consistent with previous reports [20] [24] the current study found that women, those with older age, lower levels of education and household income, and unemployment were more likely to have insomnia symptoms. Those who were divorced/separated, widowed and/or living alone were also more likely to report insomnia symptoms. The higher proportion of insomnia symptoms in rural residents could be largely attributed to the lower socioeconomic status. As a result of rapid urbanisation and introduction of one child policy in China, living alone became particularly common among the elderly rural population, highlighting the need for provision of adequate social support in rural communities.
Unlike sleep duration, few previous studies have examined the relationship between BMI and insomnia symptoms. The present study showed that individuals with low BMI were more likely to have insomnia symptoms, both in those with or without prior physical and mental disorders. These results are consistent with a report from CKB demonstrating that the prevalence of MDE was inversely associated with BMI [25]. These results should be interpreted with caution as only 4% of the present study population was obese. The present study also showed an inverse association of snoring with insomnia symptoms. A possible explanation is that frequent snoring was associated with both higher BMI and longer sleep duration, which were inversely associated with reporting insomnia symptoms ( Table 1). The longer sleep duration reported by snorers may be due to sleep apnoea associated with snoring, although it was not explicitly examined in the present study. Hence, snorers, who are at increased risk of apnoea, may exhibit less insomnia symptoms. The association between snoring and insomnia symptoms needs to be investigated further in individuals with and without apnoea in future studies.
The chief strength of the present study included the large sample size and diverse areas covered; the use of internationally recognised criteria for insomnia symptoms, major depression and generalised anxiety disorder; and directly measured physical measurements. Moreover, the information collected covered a range of socio-economic factors, health-related behaviour characteristics, and physical and mental health status, which enabled comprehensive assessment of their associations with sleep duration and insomnia symptoms in a single study.
The present study also has several limitations. First, the main objective of the prospective CKB was to investigate genetic and non-genetic determinants of major chronic diseases rather than population prevalence or incidence of such diseases. Hence, the selection of areas was based on the diversity of risk factor profile and disease patterns rather than representativeness. Because of the low response rate, which is typical of large cohort studies such as CKB, the present study may have underestimated the prevalence of insomnia symptoms as individuals with severe insomnia symptoms may have been less likely to participate. Nonetheless, the patterns of socio-economic and health-related correlates with sleep duration and insomnia symptoms and the association of extreme sleep duration and insomnia symptoms with mental and physical health conditions are likely to be generalisable to the general population. Second, the information collected was selfreported, using questionnaire rather than objective measures, so the data may be subject to recall bias, but the effects would likely be non-differential on the observed associations. Third, because of the cross-sectional nature of the study, the direction of the associations reported in the present study cannot be reliably established. Finally, information about sleep hygiene, shift work or other common sleep disorders such as RLS and sleep apnoea was not recorded.
Among Chinese adults, sleep patterns and insomnia symptoms varied greatly by socio-economic, lifestyle and health-related factors. People with more extreme sleep duration (6 h or 9 h), with or without insomnia symptoms, were more likely to have concurrent mental and physical health problems. Similarly, insomnia symptoms, regardless of sleep duration, were also associated with higher odds of both mental and physical diseases.

Contributors
YC, LL, ZC had full access to the data. All authors were involved in study design, conduct, long-term follow-up, analysis of data, interpretation or writing the report.