Fan-Fan Zhao,Wei-Min Li,Qing-Qing Liu,Xiao-Jie Feng,Yuan-Jie Li,Jun Lyu＊

1Department of Clinical Research, The First Affiliated Hospital of Jinan University, Guangzhou 510630, China.2School of Public Health, Xi’an Jiaotong University Health Science Center, Xi’an 710061, China.3Department of Nutrition, The First Affiliated Hospital of Xi'an Jiaotong University,Xi’an 710061, China.4Department of Human Anatomy, Histology and Embryology,School of Basic Medical Sciences,Xi’an Jiaotong University Health Science Center,Xi’an 710061,China.

Abstract Background:To examine the association between alcohol drinking and short sleep duration among Chinese adults.Methods: Based on inclusion and exclusion criteria, 28,167 records of 15,074 adults longitudinal data were obtained from the China Health and Nutrition Survey for the period from 2004 to 2011.A multilevel logistic regression model was employed to explore the association between alcohol drinking and sleep duration. Results:Compared with nondrinking participants, a high drinking frequency was positively correlated with a short sleep duration in both the males (odds ratio (OR) = 1.33, 95% confidence interval (CI) = 1.20-1.48, P ＜0.001) and females (OR = 1.60, 95% CI = 1.18-2.18, P = 0.003) before the covariates were adjusted.After adjusting for age,residence area, education level, marital status, smoking, coffee consumption, tea consumption, and activity level,this association remained significant in the males (OR = 1.31, 95% CI = 1.17-1.46, P ＜0.001) but not in the females(OR=1.16,95%CI=0.85-1.59,P=0.340).Conclusion:The present results suggest that a high drinking frequency was positively correlated with a short sleep duration in male population.Comprehensive measures therefore need to be implemented to improve sleep duration, especially to increase awareness of the dangers of drinking alcohol.

Keywords:Sleep duration,Alcohol drinking,Longitudinal study,China Health and Nutrition Survey

Background

Alcohol is one of the top-five causes of the global burden of disease [1], and there is now considerable epidemiological evidence that drinking is related to a range of health conditions and social outcomes [2].Chronic noncommunicable diseases caused by alcohol account for nearly 60% of the net global disease burden due to conditions such as cancer,neuropsychiatric diseases, diabetes, cardiovascular disease, and cirrhosis [3].As an endocrine disruptor,alcohol interacts in a complex manner with the human circadian rhythm and has become of widespread concern to health scientists [4].The relationship between alcohol consumption and sleep is complex.

For more than a century, the use of certain doses of alcohol has constituted part of the clinical and research studies on sleep and other health conditions [5].On one hand, as an effective somnogen, alcohol has a profound effect on sleep.Acute alcohol intake reduces the time to fall asleep (sleeping latency) and improves the quality of non-rapid eye movement sleep.However,on the other hand, the sleep-promoting effects of alcohol are transient, and can be severely disturbed during the second half of the night [6].At present,studies on sleep quality and alcohol are mainly conducted from two aspects: objective sleep time and subjective sleep time.Objective sleep quality was assessed using the Compumedics Somte polysomnography to standardize patient sleep,including rapid eye movement latency, sleep latency,wake time,total sleep time,and sleep efficiency.Ian M.Colrain et al.reported that alcoholics suffer from severe and persistent sleep disruptions, which manifest as severe insomnia and extreme daytime sleepiness[7].Subjective sleep quality assessment was based on the Pittsburgh Sleep Quality Index (PSQI) scale survey,which included subjective feelings such as sleep latency, total sleep time, and total bed rest time.Cindy L.Ehlers et al.quantified sleep disturbance as indexed by the PSQI in a sample of young adults, and reported that adolescent alcohol abuse was associated with detrimental effect on sleep quality in young adulthood[8].Another study using PSQI reported that concurrent use of alcohol, khat as well as alcohol, khat, and tobacco smoking was associated with poor sleep[9].

As we stated earlier, alcohol-related sleep problems have significant socioeconomic impacts on both individuals and society as a whole.Yet, there have been few studies evaluating the relationship between sleep duration and drinking behavior in Chinese teens,and no study has used a nationally representative sample of Chinese adults, with the samples either not being adults or only being partial, such as subjects from the Ji Dong cohort [10].These limitations mean that the obtained results are not necessarily representative of the general Chinese population.In this study, we want to examine the relationship between alcohol drinking behavior and insufficient sleep in a large, nationally representative sample in China.Analyzing a nationally representative sample would be an important advance in this area.

Methods

Study sample

Participants in the study were 28,167 person year adults from the publicly accessible database of the China Health and Nutrition Survey (CHNS).CHNS was jointly conducted by the Center for Nutrition and Food Safety of the Chinese Center for Disease Control and Prevention and the University of North Carolina Population Center.The study was approved by the relevant institutional review boards and full details about the research design can be found elsewhere[11].The present study analyzed CHNS data for the period from 2004 to 2011 related to adults older than 18 years, with complete data obtained on all of the analyzed independent and dependent variables.The study sample comprised 13,050 person-years (46.3%)for males and 15,117 person-years (53.7%) for females.

Data definitions

Sleep duration.Slee p duration was assessed with a single question.“How many hours each day do you usually sleep, including daytime and nighttime?”Individuals who answered that their sleep duration was≤8 h were included in the study.Their answers were coded as integers and classified into short (＜8 h per day)and normal(8 h per day)sleep durations.

Drinking.Subject defined as drinkers had consumed beer, liquor, or other alcoholic beverages within the previous year.Individuals were asked, “how often did you drink beer or any alcoholic beverage?” Drinking frequency was divided into four categories: never(nondrinkers), low (no more than twice a month),moderate (no more than twice a week), and high(almost every day).In the questionnaire, we consulted how much subjects drank each week.Alcohol consumption was quantified as the average weekly intake of pure alcohol, calculated as the sum of the weekly number of various types of alcohol beverages consumed multiplied by their average alcohol contents.Weekly alcohol intake = (number of beer (bottles) ×600 (ml) × 0.043 (%, v/v) + number of wine (ml) ×0.129(%,v/v)+number of liquor(ml)×0.45(%,v/v))×0.8[12].

Activity levels.Participants were asked about their activity level in their normal working environment.Based on the job and the average time of exercise(sitting, standing, walking, and lifting) during a workday,interviewers were categorized into five levels.(1)very light(working in a sitting position,e.g., office worker, watch repairer, etc.), (2) light (working in standing position, e.g., salesperson, laboratory technician, teacher,etc.), (3) moderate (student, driver,electrician, metal worker, etc.), (4) heavy (farmer,dancer, steel worker, athlete, etc.), (5) very heavy(loader, logger, miner, stonecutter, etc.).These categories were combined into three groups: low (very light or light), moderate (moderate), and high (heavy or very heavy).

Other sociodemographic and health covariates.Certain potential confounders were assessed, including age, sex, residence area (rural and urban), education level (primary school and below, junior high school,senior high school, and college and above), marital status (unmarried and married), smoking, coffee consumption and tea consumption.

Statistical analysis

The longitudinal tracking data used in this study presented a nested structure, with the first level repeating the observations for each subject and the second level being the individual study subjects.Multilevel analysis methods were used to value parameters accurately.Taking the sleep duration as the dichotomous variable(short and normal),a generalized linear mixed model was used to analyze the effect of drinking behavior on the risk of a short sleep duration.

First, IBM SPSS Statistics (version 20.0) was used for basic statistical analysis, we examined the differences in demographic variables between sleep-duration groups using thet-test for continuous variables and Pearson’s chi-square test for categorical variables.Next, R software(version 3.5.1)was used to establish a multilevel logistic regression model to explore the relationship between drinking behavior and sleep duration.Two models were evaluated: (1)unadjusted associations were first examined, and (2)then these associations were adjusted for age,residence area,education level,marital status,smoking,coffee consumption, tea consumption, and activity level.Odds ratios(ORs)with 95%confidence intervals(CIs)were presented to show the strength and direction of the association.P＜ 0.05 (two-tailed) was considered significant.

Results

Demographic characteristics

The demographic characteristics of the sample are presented in Table 1.The mean age of the 5,751 subjects at the baseline analysis in 2004 was 46.68 years, and they included 2,720 males (47.3%).The survey population was followed for 28,167 person-years: 5,751, 6,150, 6,915, and 9,351 in 2004,2006, 2009, and 2011, respectively.Table 1 indicates that there was no large fluctuation in the variables over the 8-year study period.The proportion of the sample with a short sleep duration increased during the follow-up period,this is shown in Figure 1.

The results of the univariate analysis are presented in Table 2, which reports the subject characteristics according to gender and sleep duration category.The sociodemographic and health variables differed between the sleep duration categories, demonstrating the rationality of including them as covariates.

In the male population of the study,this study found that the mean age was higher in the short-sleep-duration group than in the group with a normal sleep duration (P＜0.001).Urban adults were more likely to be short sleepers than were rural adults(P＜0.001).The lowest (primary school and below)and highest (college and above) education levels were associated with a short sleep duration (P＜0.001).People with a low activity level tended to have a short sleep duration (P＜0.001).A short sleep duration was more common in people who were married(P＜0.001),consumed coffee (P= 0.026) or tea (P＜0.001).People who had a high drinking frequency tended to be more likely to exhibit a short sleep duration (P＜0.001).The mean weekly alcohol intake was 21.6 g higher in short sleepers than in normal sleepers (P＜0.001).

The same results can be obtained in the female population.A short sleep duration was more common in people who were older (P＜0.001), urban adults (P＜0.001), people with the lowest education levels (P＜0.001), people with a low activity level (P＜0.001),married (P＜ 0.001), smokers (P＜ 0.001), and consumed tea (P= 0.018).People who had a high drinking frequency tended to be more likely to exhibit a short sleep duration (P= 0.010).The mean weekly alcohol intake was 1.6 g higher in short sleepers than in normal sleepers(P=0.019).

Longitudinal analysis of drinking behavior and sleep duration

Table 3 presents the results from the multilevel logistic regression.Compared with the nondrinking participants, a high drinking frequency was positively correlated with a short sleep duration in both the males(OR = 1.33, 95% CI = 1.20-1.48,P＜0.001) and females (OR = 1.60, 95% CI = 1.18-2.18,P= 0.003)before the covariates were adjusted.After adjusting for age, residence area, education level, marital status,smoking, coffee consumption, tea consumption, and activity level, this association remained significant in the males(OR=1.31,95%CI=1.17-1.46,P＜0.001)but not in the females(OR=1.16,95%CI=0.85-1.59,P=0.340).

Discussion

In this cross-sectional study, we found that (1) the proportion of the sample with a short sleep duration increased during the follow-up period; (2) in the male population of the study, a short sleep duration was more common in people who were older, married,urban adults, with the lowest and highest education levels,a low activity level.The adults who smoke,and consume coffee or tea, or drink frequency undergo a short sleep duration; (3) the same results can be obtained in the female population; (4) compared with the non-drinking participants, a high drinking frequency was positively correlated with a short sleep duration in both the males and females before the covariates were adjusted; (5) after adjusting for age,residence area,education level,marital status,smoking,coffee consumption, tea consumption, and activity level, this association remained significant in the males.

Table 1 Demographic characteristics of study participants,2004-2011(N=28,167).

Table 2 Subject characteristics by sleep duration category

Figure 1 Changes in the proportion of sleep duration from 2004 to 2011

Table 3 Multilevel logistic regression model analysis of drinking behavior and short sleep duration (Control group=8h)

This study was designed to explore the association between drinking behavior and sleep duration in China.Sleep is a significant determinant of health as it is closely related to both mortality and morbidity [13].It has been widely believed that modern industrialized societies have become sleep-deprived—a lack of sleep has become a public health crisis that should provoke widespread concern.

Some studies have showed that the prevalence of a short sleep duration has gradually increased over the past few decades [14, 15].Our findings support previous findings, in that the proportion of the sample in the short-sleep-duration group increased overall during the follow-up period in the analyzed nationally representative data.Consistent with existing epidemiological data, poor sleep quality generally increased with age [16].We found that the mean age was higher among the subjects who experienced insufficient sleep.

Several demographic variables such as age,residence area,and marital status have previously been linked to insomnia [17].Our study indicated that being older, married, and residing in an urban area are associated with a greater risk of short sleep duration.In males, the proportion of short sleepers vs.normal sleepers was higher for urban residents (37.9% vs.34.1%) and those who were married (94.3% vs.91.6%), and shorter sleepers was older (51.6 vs.47.6 years old) than normal sleepers.A correlation between cigarette smoking and insomnia is well-established in the general population [18].Another study also found that smoking has a dose-response relationship with sleep, with greater smoking being associated with a shorter sleep duration [16].In our study, subjects who consumed coffee or tea were more likely to have a short sleep duration, which is consistent with a previous study that considered the consumption of tea or coffee to be a risk factor for insomnia[19].

We also found that people with a high frequency and consumption of alcohol tended to be more likely to exhibit a short sleep duration.It is consistent with previous findings for sleep deprivation[20].Decreased sleep time,increased sleep latency,and increased wake time after sleep onset seem to be common symptoms associated with the chronic consumption of alcohol[21].Similarly, the amount and frequency of drinking have been shown to be related to insomnia in some studies [22].The present longitudinal analysis suggested that compared with nondrinkers, a high drinking frequency is positively correlated with a short sleep duration both in males and in females, although the association was not significant in the females after adjusting for covariates.The gender difference in the relationship between alcohol drinking and sleep duration might be ascribed to following explanation.It is well known that gender difference in alcohol consumption is one of the universal gender differences in human social behavior.Cultural factors should be considered when interpreting the results of this study,in China, females have a low rate of alcohol consumption, while alcohol drinking was widely accepted in males [23].Compared to males, more females abstain from alcohol throughout their lives,drink less, and are less likely to develop alcohol-related illnesses.Overall, the association was observed between high drinking frequency and short sleep duration, while female who drink high was extremely rare.Our study suggested that sociodemographic and health-related variables were associated with sleep duration, and so comprehensive measures are necessary to improve sleep duration.It is especially important for populations to be aware of the dangers of drinking—this can be achieved by educating people about the importance of moderate drinking,which is both reasonable and feasible.

It is particularly interesting that some research findings suggest that the association between alcohol use and sleep problems is bidirectional [24],with insomnia symptoms increasing the risk of subsequent heavy alcohol use[25],and heavy drinking leading to acute or chronic sleep problems [26].Nevertheless, many alcohol-dependent people drink alcohol to help them get to sleep, but as their consumption of alcohol becomes chronic, the hypnotic effect of alcohol is diminished [27].Based on the limited literature and research results, we proposed several possible mechanisms of the association between alcohol consumption and insomnia: alcohol can alter sleep architecture, including the initial promotion of slow wave sleep, inhibition of rapid eye movement sleep, and ultimately a reduction in slow wave sleep later at night.For alcoholics, alcohol consumption may modulate sleep homeostasis through adenosine and arousal inducing cholinergic neurons in the basal forebrain.For healthy non-alcoholics, while alcohol consumed at night has a profound effect on non-rapid eye movement sleep promotion,sleep can be disrupted especially during the second half of the sleep time [28].Chronic consumption of alcohol seems to cause common symptoms such as decreased sleep time,increased sleep latency, and increased wake time after sleep onset [21].In addition, there was evidence that alcohol had a stimulating effect on the nervous system,which may stimulate the brain to excite and damage the neurons, making it difficult for the drinkers to initiate sleep [29].It is warranted to further examine the mechanisms between drinking behaviors and sleep problems.

We acknowledge that this study was subject to several limitations.First, although our multivariate model adjusted for the effects of important confounders, causal relationships could not be established because both the exposure (drinking behavior)and outcome(sleep duration)indicators were collected concurrently.Second, all the data were obtained and verified using simple self-report questions.The obtained data might therefore be inaccurate since some individuals might underreport their drinking characteristics.Moreover, due to the high expenditure of polysomnography [30], our cohort study lacked objective measures for insomnia.Moreover, limitations of the CHNS questionnaire meant that the study did not consider other relevant characteristics related to sleep, such as sleep quality.Third, this study did not assess mental health, while several researchers have found that certain aspects of the mental health status, such as depression and anxiety, are important factors affecting sleep quality[31].

The greatest strength of this study is that it analyzed a nationally representative sample of Chinese adults for which adequate data were available from 2004 to 2011.The largeness of the sample allowed secular trends in sleep duration and drinking behavior to be determined,and guaranteed the stability of the research results.Nonetheless, the present findings might not be generalizable to other countries, including even other Asian ones.Despite these limitations, the findings reported findings that link alcohol-drinking behaviors to short sleep duration in adults.Understanding this complex relationship will be an important first step to raise vigilance for the consumption of alcohol among adults with sleep problems.