The Link of Self-Reported Insomnia Symptoms and Sleep Duration with Metabolic Syndrome: A Chinese Population-Based Study

Abstract

Study Objectives:

The aims of this study are to investigate the relationships of metabolic syndrome (MetS) with insomnia symptoms and sleep duration in a Chinese adult population.

Methods:

Data from a nationwide epidemiological survey conducted on residents from randomly selected districts in Taiwan in 2007 were used for this cross-sectional population-based study. A total of 4,197 participants were included in this study. Insomnia symptoms, including difficulty initiating sleep (DIS), difficulty maintaining sleep (DMS), early morning awakening (EMA), were assessed using the Insomnia Self-Assessment Inventory questionnaire. Subjects were divided into 3 groups based upon their reported sleep duration (< 7, 7–8, and ≥ 9 h per night). Odds ratios (ORs) and 95% confidence intervals (CIs) derived from multivariable logistic regression were used to evaluate the study aims.

Results:

The endorsement of DIS and DMS were cross-sectionally associated with the MetS after adjustment for sleep duration (OR [95% CI] was 1.24 [1.01–1.51] and 1.28 [1.02–1.61], respectively). In addition, short sleep duration was significantly associated with the prevalence of MetS independent of insomnia symptoms (OR [95% CI] was 1.54 [1.05–2.47]). However, there was no significant combined effect of insomnia symptoms and sleep duration on the prevalence of MetS.

Conclusions:

The current investigation shows that short sleep duration and insomnia symptoms, specifically DIS and DMS, were significant correlates of MetS. These findings should be replicated in prospective studies using both sleep duration and sleep quality measures.

Citation:

Lin SC, Sun CA, You SL, Hwang LC, Liang CY, Yang T, Bai CH, Chen CH, Wei CY, Chou YC. The link of self-reported insomnia symptoms and sleep duration with metabolic syndrome: a Chinese population-based study. SLEEP 2016;39(6):1261–1266.

Significance.

In previous studies, assessment of sleep duration has been mainly used to evaluate the relationship of status of sleep with metabolic syndrome (MetS). However, measuring sleep duration only is insufficient for understanding global sleep status, which consists of not only sleep duration but also sleep quality. Taking both aspects into consideration, our study results indicated that the endorsement of Insomnia symptoms, including difficulty initiating sleep and difficulty maintaining sleep were cross-sectionally associated with the MetS after adjustment for sleep duration. In addition, short sleep duration was significantly related to high prevalence of MetS independent of insomnia symptoms. However, there was no significant joint effect of insomnia symptoms and sleep duration on the prevalence of MetS.

INTRODUCTION

The metabolic syndrome (MetS) is characterized by elevated blood pressure, impaired fasting glycemia, dyslipidemia, and abdominal obesity.¹ The condition is closely associated with the development of diabetes and cardiovascular disease.^2,3 MetS is also a major and escalating public health challenge worldwide; therefore, identifying modifiable risk factors associated with the development of this medical disorder is crucial.

Adequate sleep quantity and quality are essential for the normal functioning of daily metabolic and hormonal processes.⁴ Chronic sleep deprivation, which is increasingly prevalent in modern society, is associated with metabolic and endocrine alterations that may have pathological consequences in the long term.^5–7 A considerable number of epidemiological studies have investigated the possible associations between both quantitative and qualitative sleep disturbances and risk of cardiometabolic diseases. Many studies have indicated that sleep duration is associated with obesity, diabetes, hypertension, and cardiovascular disease.^8–12 In turn, sleep duration may also be a significant risk factor for MetS. Indeed, findings from several epidemiological studies show a U-shaped association between sleep duration and MetS with both short and long sleep being risk factors of MetS.^13–16 Additionally, a handful of cross-sectional studies have documented that self-reported symptoms of insomnia (i.e., difficulty initiating sleep [DIS], difficulty maintaining sleep [DMS], and unrefreshing sleep) and poor sleep quality were associated with MetS.^17–24 In the literature, there is also evidence of a combined effect of symptoms of insomnia and short sleep duration on the risk of cardiovascular diseases.^25–28 Yet epidemiological studies exploring the relationships of MetS with insomnia symptoms in conjunction with sleep duration are scarce. In the present study, we examined both the individual and combined effects of self-reported sleep duration and insomnia symptoms on the prevalence of MetS in a nationally representative sample of Chinese adult populations.

METHODS

Study Participants

Data for this study came from the Taiwanese survey on prevalence of hypertension, hyperglycemia, and hyperlipidemia conducted in 2002 (TwSHHH) and a subsequent follow-up program carried out in 2007 (TwSHHH-II). The detailed study design and subject recruitment have been described elsewhere.²⁹ The initial TwSHHH was based on a multistage stratified systematic sampling scheme with the method of probability proportional to size. In total, 10,292 individuals were randomly selected for the TwSHHH. Of these 10,292 subjects, 7,578 (73.6%) completed a questionnaire and 6,600 (64.1%) permitted additional blood pressure measurement and blood examination for biomarkers. Subsequently, these 6,600 individuals who completed all examinations in the TwSHHH were eligible for recruitment in the TwSHHH-II. Among them, 242 subjects had died and 581 persons could not be contacted. The remaining 5,777 individuals were invited to participate in the TwSHHH-II. A total of 1,095 subjects refused to take part in the study, resulting in 4,682 persons enrolled in the TwSHHH-II (a response rate of 70.9%). Differences in sex, age, and educational level were not statistically significant between those participating in TwSHHH-II and non-participants. The TwSHHH and TwSHHH-II studies were approved by the Institutional Review Board Ethics Committee at the Bureau of Health Promotion, Department of Health, Executive Yuan in Taiwan. Written informed consent was obtained from all participants in the TwSHHH and TwSHHH-II. Given the fact that data on self-reported sleep duration and insomnia symptoms were collected only in TwSHHH-II, the current study used data from the TwSHHH-II to investigate the associations between sleep disturbances and MetS.

Baseline Measurements

At study entry in the TwSHHH and TwSHHH-II, participants underwent questionnaire interviews and anthropometric measurements by well-trained nurses under a standardized protocol. Body mass index (BMI) was calculated as weight divided by height squared (kg/m²). Waist circumference (WC) was measured at the narrowest point with a tape measure placed parallel to the floor at the end of a relaxed expiration with the participants standing akimbo. During the visit, sitting blood pressure (BP) was taken from each participant using an electric sphygmomanometer (BP3AC1-1, Microlife Cooperation, Berneck, Switzerland). Two BP readings were taken 30 seconds apart in the right arm after the participant had sat and rested for 5–10 minutes. A third BP measurement was performed if the first 2 BP readings differed by > 10 mm Hg. The average of the 2 closest readings was calculated to determine the reported BP of each participant.

Participants were instructed to fast ≥ 12 h before the blood was drawn. Sodium fluoride (NaF) plasma was collected for fasting plasma glucose (FPG) analysis, and serum was collected for measurement of lipids. The blood samples were centrifuged immediately and were transported at −10°C in dry ice to the Taiwan Public Health Association-certified research laboratory at the Chiu Clinic in Taipei City, stored at −20°C, and analyzed within 2 weeks. The serum levels of total cholesterol (TC) and triglycerides (TG) were measured using the enzymatic colorimetric test (Denka Seiken, Tokyo, Japan). In addition, electrophoresis was performed using commercial reagents (Beckman Coulter, Inc., Brea, CA) to measure size, density, and cholesterol subclasses, including high- and low-density lipoprotein cholesterol (HDL-C and LDL-C, respectively).³⁰ The FPG level was measured by the hexokinase glucose-6-phosphate dehydrogenase procedure with the use of commercial reagents (Denka Seiken, Tokyo, Japan). Blind duplicate coefficients of variation for TC, TG, HDL-C, LDL-C, and FPG were 1.7%, 1.6%, 4.8%, 3.9%, and 2.2%, respectively. These reliability estimates were generally compatible with published values.³¹ All biochemical tests were performed using automatic analyzers (TBA-200FR, Toshiba Corporation, Tokyo, Japan).

Assessment of Sleep Disturbances

Measures of sleep disturbances included self-reported insomnia symptoms and sleep duration. Information about perceived symptoms of insomnia was obtained by self-report using the Insomnia Self-assessment Inventory (ISAI) questionnaire, which was designed by the World Health Organization worldwide project on sleep and health³² and was used in the Taiwanese 2001 social trend survey.³³ The ISAI consists of 13 questions and measures insomnia symptoms, including DIS, DMS, early morning awakening (EMA), nonrestorative sleep complaints, and daytime fatigue as a result of poor sleep during the past month. The internal consistency reliability of the ISAI measured by Cronbach α was 0.91.³² Further, the ISAI ascertains symptoms of insomnia and effectively identifies insomnia caseness based on the Diagnostic and Statistical Manual IV criteria.³⁴ The present study used 3 of the 13 questions to assess insomnia symptoms due to feasibility considerations of large-scale population-based studies. The DIS question asks the subject to indicate the occurrence of taking more than one hour to fall asleep at night. The DMS question asks the subject to indicate the occurrence of waking up > 3 times during the night. The EMA question asks the subject to indicate the occurrence of waking up too early in the morning as a sleep problem. Each question was scored with a 5-point (1–5) Likert-type scale with 1 indicating “never,” 2 indicating “seldom,” 3 indicating “sometimes,” 4 indicating “usually,” and 5 indicating “all the time.” Scores of 4 and 5 for each of these questions were treated as the occurrence of the respective problem. In addition, participants were asked the number of hours per night they had slept over the past 7 days and were asked separately about sleep duration on weeknights (Monday to Friday) and weekend nights (Saturday and Sunday). Weekly average reported sleep duration was calculated as the weighted average of weeknights and weekend nights: [(5 × weekday sleep duration) + (2 × weekend sleep duration) / 7].³⁵ This definition makes no assumption regarding work, leisure, or other constraints on weekday and weekend sleep hours. Participants were divided into the following groups based upon their reported sleep duration: < 7 h per night = short sleepers, 7–8 h per night = normal sleepers (reference group), and ≥ 9 h per night = long sleepers.

Definition of MetS

In this study, MetS was defined according to the criteria established by a joint statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity.¹ The participants who fulfilled ≥ 3 of the following 5 attributes were defined as having MetS: (1) elevated blood pressure: systolic blood pressure (SBP) ≥ 130 mm Hg or diastolic blood pressure (DBP) ≥ 85 mm Hg or the use of antihypertensive medications; (2) elevated TG: serum TG levels ≥ 150 mg/dL or specific treatment for its abnormality; (3) reduced HDL-C: HDL-C levels < 40 mg/dL in men and < 50 mg/dL in women or specific treatment for its abnormality; (4) elevated FPG: a FPG level ≥ 100 mg/dL or the use of hypoglycemic agents; and (5) abdominal obesity: a waist circumference ≥ 90 cm in men and ≥ 80 cm in women. Therefore, participants receiving medication treatments for hypertension (n = 396), elevated TG and/or reduced HDL-C (n = 184), and elevated FPG (n = 97) were included in the MetS definition. The components of MetS were measured at study entry in the TwSHHH-II.

Participant Eligibility

Of the eligible 4,682 participants, 214 were excluded because of the lack of data on blood pressure (n = 151), BMI (n = 11), or WC (n = 52) measurements. Two hundred seventy-one participants for whom laboratory information was missing (68 for HDL-C, 95 for TG, and 108 for FPG) were also excluded. As a result, 4,197 subjects were included in the analysis.

Statistical Analysis

Descriptive statistics were used to characterize study participants based on presence or absence of MetS. Differences were evaluated by χ² tests for categorical variables or t-tests for continuous variables. Multivariable logistic regression analyses were used to examine the associations between the prevalence of MetS and sleep-related items (insomnia symptoms and sleep duration). Covariates considered were sex, age, cigarette smoking status (dichotomized as never or ever), and betel-nut chewing status (dichotomized as never or ever). The results from the multivariable data analysis were presented as adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs). The combined effects of individual insomnia symptoms and sleep duration on the prevalence of MetS were evaluated by entering respective interaction terms (products of respective individual insomnia symptoms and sleep duration) in the multivariable logistic regression models. Subsequently, the likelihood ratio test was used to compare models with and without the interaction term. All of the statistical analyses were performed using the IBM SPSS statistics version 22 (IBM SPSS statistics 22), and all of the statistical tests were 2-tailed with an α level of 0.05.

RESULTS

Of the total 4,197 subjects with complete data, 880 (20.9%) had metabolic syndrome. As shown in Table 1, participants with MetS were significantly older and had a significant higher prevalence of cigarette smoking and betel-nut chewing than those without MetS. As expected, subjects with MetS had significant higher BMI (27.9 ± 4.0 vs. 23.2 ± 3.5 kg/m²), TC (188.5 ± 40.3 vs. 178.1 ± 37.2 mg/dL), LDL-C (116.6 ± 35.5 vs. 109.5 ± 32.1 mg/dL), and TG (206.2 ± 64.2 vs. 104.1 ± 41.2 mg/ dL) than those without MetS (data not shown).

Table 1.

Baseline characteristics of study participants according to the presence or absence of the metabolic syndrome.

Table 2 shows the associations of prevalent MetS with self-reported individual insomnia symptoms and sleep duration. Of the individual symptoms of insomnia, subjects with MetS had a significant higher proportion of DIS and DMS than those without MetS. The OR (95% CI) of MetS associated with DIS and DMS was 1.24 (1.01–1.51) and 1.28 (1.02–1.61), respectively, after adjustment for sleep duration and potential confounders. However, the EMA was not related significantly to the prevalence of MetS (OR [95% CI] was 1.07 [0.86–1.32]). With respect to sleep duration, with subjects reporting normal sleep duration serving as the reference group, short sleep duration was significantly positively associated with MetS (OR, 1.54; 95% CI, 1.05–2.47) after adjustment for individual insomnia symptoms and all relevant confounders. However, there was no significant association between long sleep duration and MetS after adjustment for individual symptoms of insomnia and potential confounders.

Table 2.

Associations of metabolic syndrome with sleep characteristics.

The logistic regression results presented in Table 3 examined the combined effects of individual symptoms of insomnia and short sleep duration on the prevalence of MetS. Compared to participants who reported normal sleep duration without respective insomnia symptoms, subjects who experienced short sleep duration had a significant higher prevalence of MetS regardless of the presence or absence of respective insomnia symptoms. Meanwhile, the presence of respective insomnia symptoms was significantly related to a high prevalence of MetS among short sleepers. There was a positive but nonsignificant association between respective insomnia symptoms and prevalent MetS among normal sleepers. Collectively, there was no evidence to suggest an interaction between short sleep duration and individual symptoms of insomnia on the prevalence of MetS on the basis of likelihood ratio tests. On the other hand, results from logistic regression analyses documented that co-presence of long sleep duration and individual symptoms of insomnia and the presence of long sleep duration or respective insomnia symptoms on its own were not significantly associated with the prevalence of MetS (data not shown).

Table 3.

Associations of metabolic syndrome with the joint effects of short sleep duration and individual symptoms of insomnia.

We further conducted a sensitivity analysis to assess the robustness of our findings. It has been noted that people with comorbidities could be taking medications that might adversely affect sleep.⁵ We excluded participants who had taken antihypertensive medications, specific treatments for dyslipidemia, or hypoglycemia agents. As a result, we found very similar results to the main analysis even though the risk estimates were attenuated slightly (data not shown).

DISCUSSION

The present study examined associations between sleep disturbances and MetS using data from a nationwide epidemiological survey conducted in Taiwan. The results showed that the endorsement of DIS and DMS were cross-sectionally associated with the MetS after adjustment for sleep duration. In addition, short sleep duration was related significantly to the prevalence of MetS independent of insomnia symptoms. The current report complements previous reports on sleep and MetS by the simultaneous use of information on insomnia symptoms and sleep duration.

The present study showed significant associations between the prevalence of MetS and specific symptoms of insomnia, DIS, and DMS. Our results are partly comparable with those from previous related studies. Troxel et al. conducted a longitudinal study of 812 Americans aged 45–74 years over a 3-year period. They found that difficulty falling asleep and having unrefreshing sleep, both of which are symptoms of insomnia, significantly predicted the development of MetS.²⁴ However, in their study no significant associations were observed between MetS and the syndromal definition of insomnia. Contrary to the results of Troxel et al., Ikeda et al. demonstrated significant cross-sectional associations between MetS and insomnia, defined by exhibiting several insomnia symptoms, but not between MetS and individual insomnia symptoms among people in Japan.²³ The discrepancy among these studies may be attributable to different methods used for questioning individuals about sleep-related items and defining insomnia and its symptoms, as well as the different race and age distributions of participants and study designs used. Further epidemiological investigation is required to clarify these associations.

In recent years, studies focused on sleep duration and cardiometabolic risk have proliferated. A number of studies in adults have demonstrated an association between short sleep duration and MetS.^13–16,36 Conversely, some of these studies have also shown that long sleep duration is related to MetS.^{13–16,37,38} Indeed, pooled analyses of cross-sectional studies have indicated that those who sleep less (≤ 5–6 h/night) are at a greater risk of MetS than those who sleep 7–8 hours per night. Those who sleep longer (≥ 8–10 h/night) also showed an increased risk for MetS.³⁹ In the current study, a short sleep duration was significantly associated with the prevalence of MetS after adjustment for insomnia symptoms. However, no significant association was found between MetS and a long sleep duration. It is noteworthy that in the present study more than half of the participants reported a short sleep duration but a limited fraction of subjects mentioned a long sleep duration (Table 2). Additionally, DIS and DMS were more prevalent in those who slept less than in those who slept more (21.6% vs. 17.2% and 15.9% vs. 10.3%, respectively) in this study. A cross-sectional study of Taiwanese adults revealed that short sleep duration but not long sleep duration was associated with poor sleep quality.¹⁹ However, only sleep duration was used as a covariate in the aforementioned studies, whereas insomnia and sleep complaints were not. Accordingly, we took sleep quality into account when investigating the relationship between sleep duration and MetS.

This work had several strengths and limitations. Among its strengths are that it was representative of the adult population of an entire country. In addition, the current study simultaneously used information on sleep quality and duration. Among its limitations are that this study was based on cross-sectional data, so causal inferences could not be made. In addition, our information on sleep disturbances relied on self-reported questions on insomnia symptoms and sleep duration, and was not as precise as if it were objectively measured. However, it is often not feasible to obtain objective measures of sleep in large population samples, and self-reported assessments of sleep have previously been shown to be a reasonably valid estimate of sleep characteristics assessed by actigraph or polysomnography data.^40,41 Further, our measures on self-reported symptoms of insomnia were measures of sleep quality, but did not distinguish clinically diagnosed insomnia, which is likely to have resulted in some misclassification. Additionally, because the number of participants in some exposure categories was relatively modest, we had limited statistical power in the assessment of interactions between individual insomnia symptoms and sleep duration in the prevalence of MetS. Other limitations of note included that, like most large population-based studies, we lacked information on sleep apnea, a major and prevalent sleep disorder. Sleep apnea often results in disturbed sleep and also constitutes a risk factor for MetS,⁴² making the absence of data related to sleep apnea a concern.

In conclusion, the current investigation shows that short sleep duration and symptoms of insomnia, specifically DIS and DMS, were significant correlates of MetS. These findings should be replicated in prospective studies using both sleep duration and sleep quality measures.

DISCLOSURE STATEMENT

This was not an industry supported study. This study was supported by a grant from the Health Promotion Administration, Ministry of Health and Welfare (DOH95-HP-2103), Executive Yuan, Taiwan, Republic of China. The authors have indicated no financial conflicts of interest.