- Email: [email protected]
U-shaped relationships between sleep duration and metabolic syndrome and metabolic syndrome components in males: a prospective cohort study
Sleep Medicine 16 (2015) 949–954
Contents lists available at ScienceDirect
Sleep Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s l e e p
Original Article
U-shaped relationships between sleep duration and metabolic syndrome and metabolic syndrome components in males: a prospective cohort study Xue Li, Liqun Lin, Lin Lv, Xiuyu Pang, Shanshan Du, Wei Zhang, Guanqiong Na, Hao Ma, Qiao Zhang, Shuo Jiang, Haoyuan Deng, Tianshu Han, Changhao Sun, Ying Li * Department of Nutrition and Food Hygiene, College of Public Health, Harbin Medical University, Harbin, China
A R T I C L E
I N F O
Article history: Received 16 November 2014 Received in revised form 22 January 2015 Accepted 4 March 2015 Available online 18 May 2015 Keywords: Sleep duration Metabolic syndrome Metabolic syndrome components Cohort study
A B S T R A C T
Objective: Based on cross-sectional studies, sleep duration has been shown to have a relationship with the prevalence of metabolic syndrome (MS); however, no prospective studies have verified a correlation between the incidence of MS and the gender difference. Herein we prospectively determined the association between MS and gender using a large sample. Methods: A total of 4774 subjects without MS, 30–65 years of age, participated in this study. One-way ANOVA and Chi-square test were used to analyze the baseline variables. Cox regression models were performed separately in a mixed-gender population, males and females, while controlling for lifestyle and sleep-related factors. Results: During an average of 4.4-year follow-up, 1506 subjects developed MS. Both short (<6 h) and long sleep durations (8–9 and ≥9 h) increased the incidence of MS and elevated the fasting blood glucose (FBG) level in the mixed-gender population (MS: HR = 1.43, 1.25, and 1.45, respectively; elevated FBG: HR = 1.61, 1.65, and 1.98, respectively) and males (MS: HR = 1.87, 1.73, and 1.96, respectively; elevated FBG: HR = 2.27, 2.28, and 3.16, respectively). The HR8–9 and ≥9 h for hypertriglyceridemia in males was 1.48 and 19.4, and the HR<6, 6–7, and ≥9 h for hypertension in females was 1.25, 1.46, and 1.72, respectively. Conclusion: Both short and long sleep durations were associated with a greater incidence of MS and elevated FBG in a mixed-gender population and in males, and hypertension in females. Males who sleep longer were also at a higher risk for hypertriglyceridemia. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Metabolic syndrome (MS) is closely linked to lifestyle and clustering of risk factors, including abdominal adiposity, hypertension, dyslipidemia, and hyperglycemia. Patients with MS are at twice the risk for cardiovascular disease compared with patients without MS [1]. Currently, the global prevalence of MS is increasing rapidly. In recent years, the relationship between metabolic disorders and sleep duration has been of high concern; however, the results from epidemiologic studies are controversial, especially with respect to long sleep duration and gender differences. The findings presented herein were drawn from cross-sectional studies which could not evaluate the true association between sleep duration and the incidence of MS. Therefore, we have asked the following question: What is the exact cause-and-effect relationship between sleep
* Corresponding author. Department of Nutrition and Food Hygiene, College of Public Health, Harbin Medical University, 157 Baojian Road, 150081 Harbin, China. Tel: +86 451 87508731; fax: +86 451 87502885. E-mail address: [email protected] (Y. Li). http://dx.doi.org/10.1016/j.sleep.2015.03.024 1389-9457/© 2015 Elsevier B.V. All rights reserved.
duration and MS? Based on cross-sectional studies, short sleep duration has a relationship with a higher BMI [2]. In contrast, obese people have been reported to more likely develop sleep disorders that lead to disturbed and insufficient sleep [3]. Thus, it is necessary to assess the temporal relationship between sleep duration and MS, as well as the components of MS using a prospective study. Because males and females have different social roles, daily life stressors, and sex hormones, all of which can affect sleep quality and quantity [4], it was necessary to perform a gender-stratified analysis. Such an analysis could provide clues for the pathologic mechanism and the effect of gender on the relationship between sleep duration and MS. In addition, there is also a need to verify whether or not short and long sleep durations have different risks on the development of MS. With the increasing pace of life in developed and developing countries, increased suffering from chronic sleep deprivation is common [5], and has been observed to be linked with MS-related diseases, such as systemic inflammation, circulating adipokine levels, increased food intake, and energy expenditure [6–8]. Moreover, most of the evidence suggests that short sleep duration is likely to influence the prevalence of MS [9]. In contrast, long sleep duration is also an important lifestyle risk factor for
950
X. Li et al./Sleep Medicine 16 (2015) 949–954
diabetes, impaired fasting glucose, and cardiovascular disease [10–13]; however, it is still unclear whether or not long sleep duration adds to the risk for developing MS. Although a greater carotid intima–media thickness, and high triglycerides (TG) and low highdensity lipoprotein-cholesterol (HDL-C) levels may play important roles in the development of MS, as observed among those with long sleep duration [12], there are still no significant increased risks for the prevalence of MS based on cross-sectional studies. These outcomes may be influenced by single-gender surveys and a series of important confounding factors, such as sleep quality, psychological pressure, insomnia, use of hypnotics, daytime sleep, mental illness, menopausal status, and prevalence of stroke and cardiovascular disease. Therefore, a prospective cohort study in which additional confounding factors are controlled should be designed to assess the relationship between long sleep duration and incident MS to minimize bias. The aim of the current longitudinal study was to determine whether or not the relative risks of the MS incidence and MS components correlate with short and long sleep durations after controlling for a series of important confounding factors, and to make comparisons between males and females using a large-samplesize prospective cohort study in Chinese adults. 2. Methods 2.1. Subjects The study population was based on a cohort study of chronic disease in Harbin (ChiCTR-ECH-12002938; www.chictr.org). The 7696 participants (3514 males and 4182 females; 30–65 years of age) for this 4.4-year cohort study were recruited between March 2008 and June 2013. The exclusion criteria were as follows: diagnosed with MS at baseline (2742 [35.6%]); and missing questionnaire data on sleep and lost to follow-up (180 [3.6%]). The final sample size for analysis was 4774 (2496 males and 2278 females). This study protocol was approved by the Harbin Medical University Ethics Committee. At baseline, all subjects underwent health examinations and completed questionnaires, which included demographic data, the history of the present illness, medical and medication histories, sleep habits, education level, and current alcohol and cigarette consumption status. The end point of this study was the development of MS at the follow-up visit; the presence of any three of the following components was required to establish a clinical diagnosis of MS [14]: (1) abdominal obesity, defined as a waist circumference (WC) ≥ 85 cm for men and ≥80 cm for women; (2) hypertriglyceridemia, defined as a serum TG concentration ≥ 1.7 mmol/L; (3) reduced HDL-C (drug treatment for reduced HDL-C is an alternate indicator), defined as <1.0 mmol/L in males and <1.3 mmol/L in females; (4) systolic blood pressure (SBP) ≥ 130 mmHg, and/or a diastolic blood pressure (DBP) ≥ 85 mmHg; and (5) a fasting glucose ≥ 5.6 mmol/L (drug treatment for elevated glucose is an alternate indicator). 2.2. Sleep assessment and lifestyle A self-reported questionnaire was used to assess the sleep duration at night. The total hours of sleep were categorized as follows: <6 h; 6–7 h; 7–8 h; 8–9 h; and ≥9 h. The subjective sleep quality was assessed based on self-report. The insomnia status, use of hypnotics, snoring, daytime naps, bad mood, and psychological pressure were also included in the assessment. Alcohol consumption intake per day was calculated based on the following information: frequency per week; amount of consumption daily; and the varieties of alcoholic beverage. Current smoking status was categorized as follows: never or <1 cigarette per day; <10 cigarettes per day; 10–20 cigarettes per day; and >20 cigarettes per day.
Educational levels were divided into six parts (no formal education, primary school drop-out, elementary education, secondary education, upper secondary education, college education or above). For physical activity, we adopted the International Physical Activity Questionnaire (IPAQ, short version) [15]. Physical activity was expressed as metabolic equivalent task (MET)-minutes/week (=MET level × minutes of activity/day × days/week), then the participants were classified into three levels of physical activity as follows: level 1, ‘low’ physical activity level; level 2, ‘moderate’ physical activity level; and level 3, ‘vigorous’ physical activity level. 2.3. Measurements Anthropometric indices were measured by well-trained examiners with participants wearing light and thin clothing and without shoes. Body weight and height were measured to the nearest 0.1 kg and 0.1 cm, respectively; body mass index (BMI) was calculated as weight divided by the square of the height (kg/m2). We measured blood pressure three times after a 10-min rest in a sitting position, and calculated the mean of three measured results. Blood samples were drawn from study participants after an overnight fast, and immediately centrifuged at 2500 × g for 15 min to obtain serum, then stored at −80 °C. Fasting blood glucose (FBG), total cholesterol (TC), HDL-C, and TG were determined using a Roche Modular P800 Automatic Biochemical Analyzer (Roche Diagnostics, Mannheim, Germany). 2.4. Statistical analysis SPSS (version 20.0) was used to analyze the data. The baseline continuous variables were compared by ANOVA. The differences between proportions were assessed by the Pearson chi-squared test according to the sleep duration. Person-years of follow-up were calculated from the date of enrollment to the date of the initial development of MS or the date of follow-up, whichever occurred first. Cox proportional hazard regression models were used to calculate the adjusted hazard ratios (HRs) for the incidence of MS in the general population for males and females. In model 3, we adjusted age, gender, SBP, WC, smoking status, alcohol use, physical activity level, education level, psychological pressure, bad mood, use of hypnotics, sleep quality, sleep duration in daytime, and the prevalence of stroke, cardiovascular disease, mental disease, insomnia, and snoring, as well as serum FBG, TG, and HDL-C levels. In the analysis of females, menopausal status was also adapted to COX models. A two-tailed p value for trend <0.05 indicated statistical significance. 3. Results During an average of 4.4 years of follow-up, 1506 new cases of MS (incidence rate of 31.6%) were developed in this cohort study, including 814 males (32.6%) and 692 females (30.4%). The characteristics of subjects at baseline were reported by groups of sleep duration (Table 1). Short sleepers were, on average, older than long sleepers. There were significant differences in current smoking status, physical activity levels, and prevalence of stroke across the categories of sleep duration. Participants sleeping ≤6 h had significantly higher mean levels of SBP, DBP, FBG, TC, and the prevalence of insomnia. In addition, participants sleeping ≤6 h had a significantly lower prevalence of mental disease, snoring, and psychological pressure than other participants. In contrast, participants sleeping ≥9 h tended to have an increased likelihood of mental disease, cardiovascular disease, psychological pressure, and bad mood. In addition, participants sleeping ≥9 h were more likely to be under medication for hypnotics, and reported higher mean levels of TG and nap time, and lower means levels of SBP, DBP, FBG, and HDL-C. Among participants sleeping 7–8 h, a higher mean level of alcohol
X. Li et al./Sleep Medicine 16 (2015) 949–954
951
Table 1 Baseline characteristics of participants according to sleep duration. Sleep duration (h)
Number Age (years) Female gender (%) BMI (kg/m2) WC (cm) SBP (mmHg) DBP (mmHg) Current smoking (%) Never or <1 cigarette per day <10 cigarette per day 10–20 cigarette per day >20 cigarette per day Alcohol consumption (g/day) Physical activity level Level 1 (%) Level 2 (%) Level 3 (%) Lower education (%) Mental disease (%) Insomnia (%) Use of hypnotics (%) Sleep in daytime (h) Sleep quality (%) Snoring (%) Psychological pressure (%) Bad mood (%) Stroke (%) Cardiovascular disease (%) FBG (mmol/L) TC (mmol/L) TG (mmol/L) HDL-c (mmol/L)
P value for trend
<6
6 to < 7
7 to <8
8 to <9
≥9
1099 51.5 (9.0) 48.8 24.5 (3.3) 82.4 (9.8) 130.8 (20.3) 78.3 (10.5)
1067 50.2 (9.9) 59.0 24.4 (3.3) 82.7 (11.3) 127.1 (20.6) 77.2 (11.4)
1466 50.6 (9.5) 36.4 24.3 (3.1) 82.4 (9.9) 129.2 (19.7) 77.4 (11.8)
936 48.4 (10.4) 50.6 24.3 (3.3) 82.1 (9.5) 127.8 (19.6) 77.7 (11.5)
206 49.0 (11.3) 51.0 23.8 (2.9) 82.7 (9.1) 121.7 (17.6) 75.7 (11.0)
<0.001 <0.001 0.095 0.731 <0.001 0.021
80.3 9.6 6.1 4.0 7.76 (15.78)
87.6 6.0 4.7 1.7 4.61 (13.55)
79.5 9.8 7.2 3.4 7.78 (20.31)
82.2 9.6 4.8 3.4 5.29 (14.30)
74.8 15.5 9.7 0.0 7.44 (17.95)
<0.001
79.1 19.4 1.5 3.5 8.7 57 0.2 0.20 (0.42) 11.10 43.0 0.9 0.2 3.3 26.0 4.76 (1.47) 4.89 (0.93) 1.37 (0.78) 1.33 (0.31)
85.5 13.7 0.8 3.6 10.9 46.8 0.6 0.15 (0.34) 19.70 54.3 1.5 0.0 5.8 14.4 4.62 (0.73) 4.68 (0.91) 1.30 (0.68) 1.34 (0.29)
84.0 13.5 2.5 0.3 12.8 45.6 0.1 0.17 (0.38) 10.50 55.8 1.8 0.0 3.4 17.1 4.73 (0.98) 4.81 (0.90) 1.46 (0.95) 1.36 (0.32)
77.4 20.1 2.6 2.4 10.6 55.4 0.1 0.35 (0.59) 23.80 45.2 1.3 0.2 2.8 11.6 4.60 (0.68) 4.81 (0.90) 1.40 (0.77) 1.31 (0.32)
72.8 22.8 4.4 1.9 13.6 53.9 1.0 0.47 (0.59) 20.90 46.1 4.9 4.9 3.9 26.7 4.55 (0.50) 4.84 (1.10) 1.47 (0.91) 1.31 (0.34)
<0.001
<0.001
<0.001 0.019 <0.001 0.007 <0.001 <0.001 <0.001 <0.001 <0.001 0.004 <0.001 <0.001 <0.001 <0.001 0.007
Data are presented as means (SD) for continuous variables and percentages for categorical variables.
consumption, a lower proportion of females, low levels of education, and high sleep quality were observed compared to the other groups. BMI and WC had non-significant associations with sleep duration (p value for trends = 0.095 and 0.731, respectively). Over the 20,984 person-years of follow-up, the incidence per 1000 person-years of observation was 71.8. Table 2 displays the HRs of being diagnosed with MS over the follow-up period, as computed with a Cox proportional hazards model. Age, gender, SBP, WC, smoking status, alcohol use, physical activity level, education level, psychological pressure, bad mood, use of hypnotics, sleep quality, daytime sleep duration, prevalence of stroke, cardiovascular disease, mental disease, insomnia, and snoring, as well as serum FBG, TG, and HDL-C levels were designated as confounding factors in model 3. After adjusting for the confounding factors, participants who slept <6 h and >8 h per night had an increased risk of developing MS compared to participants who slept 7–8 h (all p values for trends < 0.05). A U-shaped relationship between sleep duration and the incidence of MS was observed. Of all the MS components, higher relative risks of an increased incidence of FBG existed in shorter and longer sleepers (HR <6 h = 1.61, 95% CI = 1.24–2.08; HR 8–9 h = 1.65, 95% CI = 1.26–2.17; HR≥9 h = 1.98, 95% CI = 1.28–3.05). To elucidate the gender differences affecting the relationship between sleep duration and MS, Cox regression model was conducted separately in males and females. As with the outcomes in both genders, a U-shaped relationship between sleep duration and the incidence of MS was observed in males. Sleep less than 6 h, 8–9 h, and ≥9 h were risk factors for developing MS in males (HR = 1.87, 1.73, and 1.96; 95% CI = 1.51–2.30, 1.37–2.19, and 1.35–2.85, respectively; Table 3). Sleeping less than 6 h, 8–9 h, or ≥9 h also increased the incidence of an elevated FBG level. The HRs were 2.27 (95% CI = 1.63–3.17), 2.28 (95% CI = 1.62–3.21), and 3.16 (95% CI = 1.83–5.46), respectively. Moreover,
people who slept >8 h were more likely to have elevated serum TG (HR8–9 h = 1.48, 95% CI = 1.20–1.82; HR≥9 h = 1.94, 95% CI = 1.36–2.76). Although the HRs were not significantly different, subjects sleeping ≤6 h tended to have higher TG levels (HRs = 1.23, p for trend = 0.051). In the analysis of females, we also adjusted for menopausal status in addition to the aforementioned multiple potential confounding factors; however, the HRs for incident MS among females who slept shorter or longer were not significantly different (all p trends > 0.05) compared with those who slept 7–8 h (Table 4). Of all the components of MS, sleep < 6 h, 6–7 h, and > 9 h were associated with a significantly higher risk of hypertension (HR6 h = 1.25, 95% CI = 1.01–1.56; HR6–7 h = 1.46, 95% CI = 1.17–1.84; HR≥9 h = 1.72, 95% CI = 1.20–2.47). 4. Discussion This is the first and largest sample-size cohort study focusing on the relationship between sleep duration and incidence of MS. The results demonstrated that sleep <6 h or >8 h added to the risk of MS in the general population and in males; however, no such association was observed in females. For the components of MS, both short and long sleep durations tended to increase the incidence of impaired fasting glucose in the general population and in males, while long sleep duration increased the risk of hypertriglyceridemia in males. Both short and long sleep durations increased the risk of hypertension in females. Several studies have suggested that sleep deprivation alone adds to the risk of MS [6,8,16]; in contrast, this cohort study showed that a longer sleep duration may also be a risk factor. Various factors may contribute to this difference, such as single-gender surveys, geographic areas and ethnic variations, differences in the definition of
952
X. Li et al./Sleep Medicine 16 (2015) 949–954
Table 2 Risks for development of metabolic syndrome and its components according to sleep duration among 4774 participants during follow-up. HRs (95% CI)
MS Cases (%) Person-years Incidence density (per 1000 person-years) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Elevated FBG Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Hypertension Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Central obesity Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Reduced HDL-c Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Hypertriglyceridemia Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI)
<6
6 to < 7
7 to <8
8 to <9 h
≥9 h
435 (39.6) 5030 86.4 1.28* (1.12–1.47) 1.32* (1.14–1.54) 1.43* (1.23–1.67)
324 (30.4) 4637 69.9 1.09 (0.94–1.26) 1.08 (0.92–1.28) 1.09 (0.92–1.29)
404 (27.6) 6271 64.4 1 (Ref.) 1 (Ref.) 1 (Ref.)
274 (29.3) 4112 66.6 1.03 (0.88–1.20) 1.06 (0.89–1.26) 1.25* (1.05–1.50)
69 (33.5) 934 73.9 1.12 (0.87–1.45) 1.32 (1.00–1.75) 1.45* (1.09–1.93)
188 (17.1) 1.38* (1.12–1.70) 1.44* (1.13–1.84) 1.61* (1.24–2.08)
119 (11.2) 1.03 (0.82–1.31) 0.95 (0.72–1.25) 1.02 (0.76–1.36)
164 (11.2) 1 (Ref.) 1 (Ref.) 1 (Ref.)
120 (12.8) 1.17 (0.92–1.48) 1.43* (1.10–1.85) 1.65* (1.26–2.17)
34 (16.5) 1.42 (0.98–2.05) 1.76* (1.16–2.69) 1.98* (1.28–3.05)
612 (55.7) 1.05 (0.94–1.17) 1.07 (0.95–1.20) 1.10 (0.98–1.24)
476 (44.6) 0.96 (0.85–1.08) 1.04 (0.92–1.18) 1.01 (0.89–1.15)
680 (46.4) 1 (Ref.) 1 (Ref.) 1 (Ref.)
406 (43.4) 0.93 (0.82–1.05) 0.94 (0.83–1.08) 1.00 (0.87–1.14)
99 (48.1) 0.96 (0.78–1.19) 1.03 (0.81–1.31) 1.07 (0.84–1.37)
725 (66.0) 1.01 (0.91–1.11) 1.02 (0.92–1.14) 1.06 (0.95–1.18)
660 (61.9) 1.07 (0.96–1.18) 1.03 (0.92–1.16) 1.00 (0.89–1.13)
834 (56.9) 1 (Ref.) 1 (Ref.) 1 (Ref.)
583 (62.3) 1.08 (0.97–1.20) 1.03 (0.92–1.16) 1.08 (0.96–1.22)
148 (71.8) 1.16 (0.98–1.38) 1.12 (0.91–1.37) 1.15 (0.94–1.42)
351 (31.9) 0.80* (0.70–0.92) 0.85* (0.73–0.99) 0.89 (0.77–1.04)
396 (37.1) 0.98 (0.86–1.13) 0.98 (0.84–1.14) 0.96 (0.82–1.12)
462 (31.5) 1 (Ref.) 1 (Ref.) 1 (Ref.)
323 (34.5) 0.92 (0.80–1.06) 0.92 (0.78–1.08) 0.97 (0.82–1.14)
54 (26.2) 0.67* (0.51–0.89) 0.83 (0.62–1.11) 0.87 (0.65–1.16)
444 (40.4) 1.06 (0.93–1.20) 1.06 (0.92–1.22) 1.11 (0.96–1.29)
310 (29.1) 0.85* (0.73–0.98) 0.82* (0.70–0.96) 0.88 (0.75–1.03)
494 (33.7) 1 (Ref.) 1 (Ref.) 1 (Ref.)
319 (34.1) 0.97 (0.84–1.12) 0.95 (0.82–1.12) 1.10 (0.93–1.29)
84 (40.8) 1.11 (0.88–1.40) 1.08 (0.83–1.41) 1.21 (0.93–1.59)
Data are HRs (95% CI, confidence interval); Model 1 adjusted by age, sex; Model 2 adjusted by model 1 plus SBP, smoking, alcohol use, physical activity level, education, psychological pressure, bad mood, stroke and cardiovascular disease; Model 3 adjusted by model 2 plus mental illness, insomnia, use of hypnotics, sleep quality, sleep in daytime, snoring, WC, FBG and TG. *P < 0.05.
MS, and limited confounding factors. Thus, we analyzed a mixedgender population, as well as males and females separately. In addition, a greater quantity of important influencing factors than previous research has been analyzed. First, napping is a common
practice in Chinese adults. Frequent napping and long nap duration were shown to increase the risk of type 2 diabetes mellitus [17,18]. In in vitro experiments, sleeping during the daytime resulted in higher levels of insulin-like growth factor binding protein
Table 3 Relative risks for development of MS according to sleep duration among males. HRs (95% CI)
MS Elevated FBG Hypertension Central obesity Reduced HDL-c Hypertriglyceridemia
<6
6 to < 7
7 to < 8
8 to <9 h
≥9 h
1.87* (1.51–2.30) 2.27* (1.63–3.17) 1.06 (0.91–1.24) 1.13 (0.98–1.31) 0.80 (0.61–1.05) 1.23 (1.00–1.50)
0.78 (0.59–1.03) 0.74 (0.48–1.15) 0.85 (0.71–1.01) 0.96 (0.81–1.12) 0.83 (0.62–1.12) 0.79 (0.62–1.02)
1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.)
1.73* (1.37–2.19) 2.28* (1.62–3.21) 0.94 (0.79–1.12) 1.07 (0.90–1.25) 1.21 (0.93–1.58) 1.48* (1.20–1.82)
1.96* (1.35–2.85) 3.16* (1.83–5.46) 0.78 (0.55–1.12) 1.19 (0.90–1.58) 1.46 (0.93–2.30) 1.94* (1.36–2.76)
* P < 0.05.
Table 4 Relative risks for development of MS according to sleep duration among females. HRs (95% CI)
MS Elevated FBG Hypertension Central obesity Reduced HDL-c Hypertriglyceridemia * P < 0.05.
<6
6 to < 7
7 to < 8
8 to <9 h
≥9 h
0.93 (0.73–1.19) 0.68 (0.44–1.05) 1.25* (1.01–1.56) 0.97 (0.81–1.16) 0.93 (0.76–1.14) 0.91 (0.73–1.14)
1.27 (0.99–1.64) 0.84 (0.53–1.29) 1.46* (1.17–1.84) 1.07 (0.89–1.29) 1.08 (0.88–1.33) 1.04 (0.81–1.32)
1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.)
0.87 (0.65–1.17) 0.84 (0.52–1.34) 1.19 (0.93–1.52) 1.13 (0.94–1.38) 0.90 (0.72–1.13) 0.80 (0.61–1.05)
0.93 (0.60–1.53) 0.53 (0.19–1.52) 1.72* (1.20–2.47) 1.11 (0.81–1.53) 0.76 (0.51–1.12) 0.80 (0.51–1.25)
X. Li et al./Sleep Medicine 16 (2015) 949–954
1 (IGFBP1) than nocturnal sleep [19]. In addition, cell and molecular links exist between disorders of circadian rhythms and sleep with MS [20]. Thus, independence from daytime sleeping should be considered with respect to the relationship between nocturnal sleep duration and incidence of MS. Second, subjects with MS have poorer sleep quality [21]. Poor sleep efficiency, such as sleepdisordered breathing, poor sleep continuity, and “lighter” sleep, has a significant correlation with overactivity of the sympathetic nervous system, which could result in insulin resistance [22,23]. In view of this, a series of sleep parameters were considered as confounding factors in the assessment of the relative risks for MS, which included sleep in the daytime, insomnia, use of hypnotics, selfreported sleep quality, and snoring. Bad mood was also adjusted in Cox regression models because a significant interaction between sleepiness and depression was also identified [24]. In this longitudinal research, U-shaped relationships between sleep duration and MS were observed in a mixed gender population and in males. This conclusion was supported by several previous cross-sectional studies without gender-stratification or just among special ages [25,26]. Regarding the MS components, shorter and longer sleepers had increased risks of elevated FBG and males who slept long were at high risk for hypertriglyceridemia, as reported in several crosssectional studies [12,27]; however, cross-sectional analysis could not identify a potential cause and effect relationship. The results derived from a prospective cohort study further indicated the causality may exist between sleep duration and some of the aforementioned MS components. The mechanisms underlying the relationship between sleep duration and MS remain to be elucidated. Indeed, sleeplessrelated hypothalamic–pituitary–adrenal hyperactivity, elevated cortisol secretion levels, altered growth hormone metabolism, and low-grade activated inflammation have been shown to be associated with the pathogenesis of MS [28–30]. In contrast, short sleep duration has adverse effects on levels of leptin, visfatin, and ghrelin, increases appetite and caloric intake, then reduces energy secretion and impairs glycemic control [10,31,32]. Theories to explain the relationship between long sleep duration and metabolic abnormalities have been advanced [33,34]. For example, the waking time to undertake physical activity was shorter for long sleepers, which may lead to an increased risk of MS [33]. Both short and long sleep durations increased the risk for hypertension in females, but not in males. One of the potential mechanisms to explain the increased risk for hypertension in females may be associated with menopause and psychosocial stress [35], which may be interrelated with the increased risk for MS and risk factors [36]. In the current study, both menopausal status and psychological pressure were analyzed as confounding factors. Except for short sleep duration, a correlation between long sleep duration and hypertension was also noted in females, while a U-shaped relationship was not demonstrated in males. Similarly, 24 h of total sleep deprivation reduced muscle sympathetic nerve activity in males, but not in females. Gender differences in muscle sympathetic nerve activity were associated with gender differences in sympathetic baroreflex function and testosterone [37]. There were several limitations in this study. First, this study was limited to a Chinese population, and the results may not be extended to other races and ethnic groups. Second, we did not examine hormone levels, which are increasingly indicated to regulate energy balance. A further limitation was that self-reported sleep duration may not be as precise as when measured objectively. We only analyzed subjective sleep quality, and failed to adopt the poor global sleep-quality scores on the Pittsburgh Sleep Quality Index [38]; however, we have adjusted sleeping habits as well as sleep quality from self-reported questionnaire information as confounding factors to minimize the bias.
953
In conclusion, both short and long sleep durations increased the incidence of MS and hyperglycemia in the general population and males. In addition, males who slept >8 h were shown to be at higher risk for hypertriglyceridemia. Females who slept ≤7 h and ≥9 h were associated with incident hypertension.
Conflicts of interest There are no conflicts of interest. The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2015.03.024.
Acknowledgements This work was financially supported by the Chang Jiang Scholar Candidates Program of the Provincial Universities in Heilongjiang, China (grant No. 2012CJHB006).
References [1] Zhao Y, Yan H, Yang R, et al. Prevalence and determinants of metabolic syndrome among adults in a rural area of Northwest China. PLoS ONE 2014;9:e91578. [2] Bayon V, Leger D, Gomez-Merino D, et al. Sleep debt and obesity. Ann Med 2014;46:264–72. [3] Xiao Q, Keadle SK, Hollenbeck AR, et al. Sleep duration and total and causespecific mortality in a large us cohort: interrelationships with physical activity, sedentary behavior, and body mass index. Am J Epidemiol 2014;doi:10.1093/ aje/kwu222. [4] Antonijevic IA, Murck H, Frieboes R, et al. On the gender differences in sleependocrine regulation in young normal humans. Neuroendocrinology 1999;70:280–7. [5] Bonnet MHAD. We are chronically sleep deprived. Sleep 1995;18:908–11. [6] Choi JK, Kim MY, Kim JK, et al. Association between short sleep duration and high incidence of metabolic syndrome in midlife women. Tohoku J Exp Med 2011;225:187–93. [7] Rechtschaffen A, Bergmann BM, Everson CA, et al. Sleep deprivation in the rat: X. Integration and discussion of the findings. Sleep 1989;25:68–87. [8] Otsuka T, Kawada T, Yanai M, et al. Incidence of metabolic syndrome and associated lifestyle factors in a worksite male population. Sangyo Eiseigaku Zasshi 2011;53:78–86. [9] Wu MC, Yang YC, Wu JS, et al. Short sleep duration associated with a higher prevalence of metabolic syndrome in an apparently healthy population. Prev Med 2012;55:305–9. [10] Hayes AL, Xu F, Babineau D, et al. Sleep duration and circulating adipokine levels. Sleep 2011;34:147–52. [11] Sharma S, Kavuru M. Sleep and metabolism: an overview. Int J Endocrinol 2010;2010. [12] Kaneita Y, Uchiyama M, Yoshiike N, et al. Associations of usual sleep duration with serum lipid and lipoprotein levels. Sleep 2008;31:645–52. [13] Buxton OM, Marcelli E. Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States. Soc Sci Med 2010;71:1027–36. [14] Alberti KG, Eckel RH, Grundy SM, et al. Harmonizing the metabolic syndrome: a joint interim 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. Circulation 2009;120:1640–5. [15] Tierney M, Fraser A, Kennedy N. Criterion validity of the International Physical Activity Questionnaire Short Form (IPAQ-SF) for use in patients with rheumatoid arthritis: comparison with the SenseWear Armband. Physiotherapy 2014;doi:10.1016/j.physio.2014.07.005. [16] Chaput JP, McNeil J, Despres JP, et al. Short sleep duration as a risk factor for the development of the metabolic syndrome in adults. Prev Med 2013;57. [17] Picarsic JL, Glynn NW, Taylor CA, et al. Self-reported napping and duration and quality of sleep in the lifestyle interventions and independence for elders pilot study. J Am Geriatr Soc 2008;56:1674–80. [18] Lam KB, Jiang CQ, Thomas GN, et al. Napping is associated with increased risk of type 2 diabetes: the Guangzhou Biobank Cohort Study. Sleep 2010;33:402–7. [19] Rehman JU, Brismar K, Holmback U, et al. Sleeping during the day: effects on the 24-h patterns of IGF-binding protein 1, insulin, glucose, cortisol, and growth hormone. Eur J Endocrinol 2010;163:383–90. [20] Maury E, Ramsey KM, Bass J. Circadian rhythms and metabolic syndrome: from experimental genetics to human disease. Circ Res 2010;106:447–62. [21] Hung HC, Yang YC, Ou HY, et al. The association between self-reported sleep quality and metabolic syndrome. PLoS ONE 2013;8:e54304.
954
X. Li et al./Sleep Medicine 16 (2015) 949–954
[22] Hall MH, Okun ML, Sowers M, et al. Sleep is associated with the metabolic syndrome in a multi-ethnic cohort of midlife women: the SWAN Sleep Study. Sleep 2012;35:783–90. [23] Tasali E, Leproult R, Ehrmann DA, et al. Slow-wave sleep and the risk of type 2 diabetes in humans. Proc Natl Acad Sci U S A 2008;105:1044–9. [24] Ronksley PE, Hemmelgarn BR, Heitman SJ, et al. Excessive daytime sleepiness is associated with increased health care utilization among patients referred for assessment of OSA. Sleep 2011;34:363–70. [25] Hall MH, Muldoon MF, Jennings JR, et al. Self-reported sleep duration is associated with the metabolic syndrome in midlife adults. Sleep 2008;31:635– 43. [26] Choi KM, Lee JS, Park HS, et al. Relationship between sleep duration and the metabolic syndrome: Korean National Health and Nutrition Survey 2001. Int J Obes (Lond) 2008;32:1091–7. [27] Yaggi HK, Araujo AB, McKinlay JB. Sleep duration as a risk factor for the development of type 2 diabetes. Diabetes Care 2006;29:657–61. [28] Miller MA, Cappuccio FP. Inflammation, sleep, obesity and cardiovascular disease. Curr Vasc Pharmacol 2007;5:93–102. [29] Straznicky NE, Eikelis N, Lambert EA, et al. Mediators of sympathetic activation in metabolic syndrome obesity. Curr Hypertens Rep 2008;10:440–7. [30] Vgontzas AN, Chrousos GP. Sleep, the hypothalamic-pituitary-adrenal axis, and cytokines: multiple interactions and disturbances in sleep disorders. Endocrinol Metab Clin North Am 2002;31:15–36.
[31] St-Onge MP, Roberts AL, Chen J, et al. Short sleep duration increases energy intakes but does not change energy expenditure in normal-weight individuals. Am J Clin Nutr 2011;94:410–16. [32] Nedeltcheva AV, Kilkus JM, Imperial J, et al. Sleep curtailment is accompanied by increased intake of calories from snacks. Am J Clin Nutr 2009;89:126–33. [33] Arora T, Jiang CQ, Thomas GN, et al. Self-reported long total sleep duration is associated with metabolic syndrome: the Guangzhou Biobank Cohort Study. Diabetes Care 2011;34:2317–19. [34] Ayas NT, White DP, Al-Delaimy WK, et al. A prospective study of self-reported sleep duration and incident diabetes in women. Diabetes Care 2003;26:380–4. [35] Cappuccio FP, Stranges S, Kandala NB, et al. Gender-specific associations of short sleep duration with prevalent and incident hypertension: the Whitehall II Study. Hypertension 2007;50:693–700. [36] Janssen I, Powell LH, Crawford S, et al. Menopause and the metabolic syndrome: the Study of Women’s Health Across the Nation. Arch Intern Med 2008;168:1568–75. [37] Carter JR, Durocher JJ, Larson RA, et al. Sympathetic neural responses to 24-hour sleep deprivation in humans: sex differences. Am J Physiol Heart Circ Physiol 2012;302:H1991–7. [38] Jennings JR, Muldoon MF, Hall M, et al. Self-reported sleep quality is associated with the metabolic syndrome. Sleep 2007;30:219–23.
Contents lists available at ScienceDirect
Sleep Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / s l e e p
Original Article
U-shaped relationships between sleep duration and metabolic syndrome and metabolic syndrome components in males: a prospective cohort study Xue Li, Liqun Lin, Lin Lv, Xiuyu Pang, Shanshan Du, Wei Zhang, Guanqiong Na, Hao Ma, Qiao Zhang, Shuo Jiang, Haoyuan Deng, Tianshu Han, Changhao Sun, Ying Li * Department of Nutrition and Food Hygiene, College of Public Health, Harbin Medical University, Harbin, China
A R T I C L E
I N F O
Article history: Received 16 November 2014 Received in revised form 22 January 2015 Accepted 4 March 2015 Available online 18 May 2015 Keywords: Sleep duration Metabolic syndrome Metabolic syndrome components Cohort study
A B S T R A C T
Objective: Based on cross-sectional studies, sleep duration has been shown to have a relationship with the prevalence of metabolic syndrome (MS); however, no prospective studies have verified a correlation between the incidence of MS and the gender difference. Herein we prospectively determined the association between MS and gender using a large sample. Methods: A total of 4774 subjects without MS, 30–65 years of age, participated in this study. One-way ANOVA and Chi-square test were used to analyze the baseline variables. Cox regression models were performed separately in a mixed-gender population, males and females, while controlling for lifestyle and sleep-related factors. Results: During an average of 4.4-year follow-up, 1506 subjects developed MS. Both short (<6 h) and long sleep durations (8–9 and ≥9 h) increased the incidence of MS and elevated the fasting blood glucose (FBG) level in the mixed-gender population (MS: HR = 1.43, 1.25, and 1.45, respectively; elevated FBG: HR = 1.61, 1.65, and 1.98, respectively) and males (MS: HR = 1.87, 1.73, and 1.96, respectively; elevated FBG: HR = 2.27, 2.28, and 3.16, respectively). The HR8–9 and ≥9 h for hypertriglyceridemia in males was 1.48 and 19.4, and the HR<6, 6–7, and ≥9 h for hypertension in females was 1.25, 1.46, and 1.72, respectively. Conclusion: Both short and long sleep durations were associated with a greater incidence of MS and elevated FBG in a mixed-gender population and in males, and hypertension in females. Males who sleep longer were also at a higher risk for hypertriglyceridemia. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Metabolic syndrome (MS) is closely linked to lifestyle and clustering of risk factors, including abdominal adiposity, hypertension, dyslipidemia, and hyperglycemia. Patients with MS are at twice the risk for cardiovascular disease compared with patients without MS [1]. Currently, the global prevalence of MS is increasing rapidly. In recent years, the relationship between metabolic disorders and sleep duration has been of high concern; however, the results from epidemiologic studies are controversial, especially with respect to long sleep duration and gender differences. The findings presented herein were drawn from cross-sectional studies which could not evaluate the true association between sleep duration and the incidence of MS. Therefore, we have asked the following question: What is the exact cause-and-effect relationship between sleep
* Corresponding author. Department of Nutrition and Food Hygiene, College of Public Health, Harbin Medical University, 157 Baojian Road, 150081 Harbin, China. Tel: +86 451 87508731; fax: +86 451 87502885. E-mail address: [email protected] (Y. Li). http://dx.doi.org/10.1016/j.sleep.2015.03.024 1389-9457/© 2015 Elsevier B.V. All rights reserved.
duration and MS? Based on cross-sectional studies, short sleep duration has a relationship with a higher BMI [2]. In contrast, obese people have been reported to more likely develop sleep disorders that lead to disturbed and insufficient sleep [3]. Thus, it is necessary to assess the temporal relationship between sleep duration and MS, as well as the components of MS using a prospective study. Because males and females have different social roles, daily life stressors, and sex hormones, all of which can affect sleep quality and quantity [4], it was necessary to perform a gender-stratified analysis. Such an analysis could provide clues for the pathologic mechanism and the effect of gender on the relationship between sleep duration and MS. In addition, there is also a need to verify whether or not short and long sleep durations have different risks on the development of MS. With the increasing pace of life in developed and developing countries, increased suffering from chronic sleep deprivation is common [5], and has been observed to be linked with MS-related diseases, such as systemic inflammation, circulating adipokine levels, increased food intake, and energy expenditure [6–8]. Moreover, most of the evidence suggests that short sleep duration is likely to influence the prevalence of MS [9]. In contrast, long sleep duration is also an important lifestyle risk factor for
950
X. Li et al./Sleep Medicine 16 (2015) 949–954
diabetes, impaired fasting glucose, and cardiovascular disease [10–13]; however, it is still unclear whether or not long sleep duration adds to the risk for developing MS. Although a greater carotid intima–media thickness, and high triglycerides (TG) and low highdensity lipoprotein-cholesterol (HDL-C) levels may play important roles in the development of MS, as observed among those with long sleep duration [12], there are still no significant increased risks for the prevalence of MS based on cross-sectional studies. These outcomes may be influenced by single-gender surveys and a series of important confounding factors, such as sleep quality, psychological pressure, insomnia, use of hypnotics, daytime sleep, mental illness, menopausal status, and prevalence of stroke and cardiovascular disease. Therefore, a prospective cohort study in which additional confounding factors are controlled should be designed to assess the relationship between long sleep duration and incident MS to minimize bias. The aim of the current longitudinal study was to determine whether or not the relative risks of the MS incidence and MS components correlate with short and long sleep durations after controlling for a series of important confounding factors, and to make comparisons between males and females using a large-samplesize prospective cohort study in Chinese adults. 2. Methods 2.1. Subjects The study population was based on a cohort study of chronic disease in Harbin (ChiCTR-ECH-12002938; www.chictr.org). The 7696 participants (3514 males and 4182 females; 30–65 years of age) for this 4.4-year cohort study were recruited between March 2008 and June 2013. The exclusion criteria were as follows: diagnosed with MS at baseline (2742 [35.6%]); and missing questionnaire data on sleep and lost to follow-up (180 [3.6%]). The final sample size for analysis was 4774 (2496 males and 2278 females). This study protocol was approved by the Harbin Medical University Ethics Committee. At baseline, all subjects underwent health examinations and completed questionnaires, which included demographic data, the history of the present illness, medical and medication histories, sleep habits, education level, and current alcohol and cigarette consumption status. The end point of this study was the development of MS at the follow-up visit; the presence of any three of the following components was required to establish a clinical diagnosis of MS [14]: (1) abdominal obesity, defined as a waist circumference (WC) ≥ 85 cm for men and ≥80 cm for women; (2) hypertriglyceridemia, defined as a serum TG concentration ≥ 1.7 mmol/L; (3) reduced HDL-C (drug treatment for reduced HDL-C is an alternate indicator), defined as <1.0 mmol/L in males and <1.3 mmol/L in females; (4) systolic blood pressure (SBP) ≥ 130 mmHg, and/or a diastolic blood pressure (DBP) ≥ 85 mmHg; and (5) a fasting glucose ≥ 5.6 mmol/L (drug treatment for elevated glucose is an alternate indicator). 2.2. Sleep assessment and lifestyle A self-reported questionnaire was used to assess the sleep duration at night. The total hours of sleep were categorized as follows: <6 h; 6–7 h; 7–8 h; 8–9 h; and ≥9 h. The subjective sleep quality was assessed based on self-report. The insomnia status, use of hypnotics, snoring, daytime naps, bad mood, and psychological pressure were also included in the assessment. Alcohol consumption intake per day was calculated based on the following information: frequency per week; amount of consumption daily; and the varieties of alcoholic beverage. Current smoking status was categorized as follows: never or <1 cigarette per day; <10 cigarettes per day; 10–20 cigarettes per day; and >20 cigarettes per day.
Educational levels were divided into six parts (no formal education, primary school drop-out, elementary education, secondary education, upper secondary education, college education or above). For physical activity, we adopted the International Physical Activity Questionnaire (IPAQ, short version) [15]. Physical activity was expressed as metabolic equivalent task (MET)-minutes/week (=MET level × minutes of activity/day × days/week), then the participants were classified into three levels of physical activity as follows: level 1, ‘low’ physical activity level; level 2, ‘moderate’ physical activity level; and level 3, ‘vigorous’ physical activity level. 2.3. Measurements Anthropometric indices were measured by well-trained examiners with participants wearing light and thin clothing and without shoes. Body weight and height were measured to the nearest 0.1 kg and 0.1 cm, respectively; body mass index (BMI) was calculated as weight divided by the square of the height (kg/m2). We measured blood pressure three times after a 10-min rest in a sitting position, and calculated the mean of three measured results. Blood samples were drawn from study participants after an overnight fast, and immediately centrifuged at 2500 × g for 15 min to obtain serum, then stored at −80 °C. Fasting blood glucose (FBG), total cholesterol (TC), HDL-C, and TG were determined using a Roche Modular P800 Automatic Biochemical Analyzer (Roche Diagnostics, Mannheim, Germany). 2.4. Statistical analysis SPSS (version 20.0) was used to analyze the data. The baseline continuous variables were compared by ANOVA. The differences between proportions were assessed by the Pearson chi-squared test according to the sleep duration. Person-years of follow-up were calculated from the date of enrollment to the date of the initial development of MS or the date of follow-up, whichever occurred first. Cox proportional hazard regression models were used to calculate the adjusted hazard ratios (HRs) for the incidence of MS in the general population for males and females. In model 3, we adjusted age, gender, SBP, WC, smoking status, alcohol use, physical activity level, education level, psychological pressure, bad mood, use of hypnotics, sleep quality, sleep duration in daytime, and the prevalence of stroke, cardiovascular disease, mental disease, insomnia, and snoring, as well as serum FBG, TG, and HDL-C levels. In the analysis of females, menopausal status was also adapted to COX models. A two-tailed p value for trend <0.05 indicated statistical significance. 3. Results During an average of 4.4 years of follow-up, 1506 new cases of MS (incidence rate of 31.6%) were developed in this cohort study, including 814 males (32.6%) and 692 females (30.4%). The characteristics of subjects at baseline were reported by groups of sleep duration (Table 1). Short sleepers were, on average, older than long sleepers. There were significant differences in current smoking status, physical activity levels, and prevalence of stroke across the categories of sleep duration. Participants sleeping ≤6 h had significantly higher mean levels of SBP, DBP, FBG, TC, and the prevalence of insomnia. In addition, participants sleeping ≤6 h had a significantly lower prevalence of mental disease, snoring, and psychological pressure than other participants. In contrast, participants sleeping ≥9 h tended to have an increased likelihood of mental disease, cardiovascular disease, psychological pressure, and bad mood. In addition, participants sleeping ≥9 h were more likely to be under medication for hypnotics, and reported higher mean levels of TG and nap time, and lower means levels of SBP, DBP, FBG, and HDL-C. Among participants sleeping 7–8 h, a higher mean level of alcohol
X. Li et al./Sleep Medicine 16 (2015) 949–954
951
Table 1 Baseline characteristics of participants according to sleep duration. Sleep duration (h)
Number Age (years) Female gender (%) BMI (kg/m2) WC (cm) SBP (mmHg) DBP (mmHg) Current smoking (%) Never or <1 cigarette per day <10 cigarette per day 10–20 cigarette per day >20 cigarette per day Alcohol consumption (g/day) Physical activity level Level 1 (%) Level 2 (%) Level 3 (%) Lower education (%) Mental disease (%) Insomnia (%) Use of hypnotics (%) Sleep in daytime (h) Sleep quality (%) Snoring (%) Psychological pressure (%) Bad mood (%) Stroke (%) Cardiovascular disease (%) FBG (mmol/L) TC (mmol/L) TG (mmol/L) HDL-c (mmol/L)
P value for trend
<6
6 to < 7
7 to <8
8 to <9
≥9
1099 51.5 (9.0) 48.8 24.5 (3.3) 82.4 (9.8) 130.8 (20.3) 78.3 (10.5)
1067 50.2 (9.9) 59.0 24.4 (3.3) 82.7 (11.3) 127.1 (20.6) 77.2 (11.4)
1466 50.6 (9.5) 36.4 24.3 (3.1) 82.4 (9.9) 129.2 (19.7) 77.4 (11.8)
936 48.4 (10.4) 50.6 24.3 (3.3) 82.1 (9.5) 127.8 (19.6) 77.7 (11.5)
206 49.0 (11.3) 51.0 23.8 (2.9) 82.7 (9.1) 121.7 (17.6) 75.7 (11.0)
<0.001 <0.001 0.095 0.731 <0.001 0.021
80.3 9.6 6.1 4.0 7.76 (15.78)
87.6 6.0 4.7 1.7 4.61 (13.55)
79.5 9.8 7.2 3.4 7.78 (20.31)
82.2 9.6 4.8 3.4 5.29 (14.30)
74.8 15.5 9.7 0.0 7.44 (17.95)
<0.001
79.1 19.4 1.5 3.5 8.7 57 0.2 0.20 (0.42) 11.10 43.0 0.9 0.2 3.3 26.0 4.76 (1.47) 4.89 (0.93) 1.37 (0.78) 1.33 (0.31)
85.5 13.7 0.8 3.6 10.9 46.8 0.6 0.15 (0.34) 19.70 54.3 1.5 0.0 5.8 14.4 4.62 (0.73) 4.68 (0.91) 1.30 (0.68) 1.34 (0.29)
84.0 13.5 2.5 0.3 12.8 45.6 0.1 0.17 (0.38) 10.50 55.8 1.8 0.0 3.4 17.1 4.73 (0.98) 4.81 (0.90) 1.46 (0.95) 1.36 (0.32)
77.4 20.1 2.6 2.4 10.6 55.4 0.1 0.35 (0.59) 23.80 45.2 1.3 0.2 2.8 11.6 4.60 (0.68) 4.81 (0.90) 1.40 (0.77) 1.31 (0.32)
72.8 22.8 4.4 1.9 13.6 53.9 1.0 0.47 (0.59) 20.90 46.1 4.9 4.9 3.9 26.7 4.55 (0.50) 4.84 (1.10) 1.47 (0.91) 1.31 (0.34)
<0.001
<0.001
<0.001 0.019 <0.001 0.007 <0.001 <0.001 <0.001 <0.001 <0.001 0.004 <0.001 <0.001 <0.001 <0.001 0.007
Data are presented as means (SD) for continuous variables and percentages for categorical variables.
consumption, a lower proportion of females, low levels of education, and high sleep quality were observed compared to the other groups. BMI and WC had non-significant associations with sleep duration (p value for trends = 0.095 and 0.731, respectively). Over the 20,984 person-years of follow-up, the incidence per 1000 person-years of observation was 71.8. Table 2 displays the HRs of being diagnosed with MS over the follow-up period, as computed with a Cox proportional hazards model. Age, gender, SBP, WC, smoking status, alcohol use, physical activity level, education level, psychological pressure, bad mood, use of hypnotics, sleep quality, daytime sleep duration, prevalence of stroke, cardiovascular disease, mental disease, insomnia, and snoring, as well as serum FBG, TG, and HDL-C levels were designated as confounding factors in model 3. After adjusting for the confounding factors, participants who slept <6 h and >8 h per night had an increased risk of developing MS compared to participants who slept 7–8 h (all p values for trends < 0.05). A U-shaped relationship between sleep duration and the incidence of MS was observed. Of all the MS components, higher relative risks of an increased incidence of FBG existed in shorter and longer sleepers (HR <6 h = 1.61, 95% CI = 1.24–2.08; HR 8–9 h = 1.65, 95% CI = 1.26–2.17; HR≥9 h = 1.98, 95% CI = 1.28–3.05). To elucidate the gender differences affecting the relationship between sleep duration and MS, Cox regression model was conducted separately in males and females. As with the outcomes in both genders, a U-shaped relationship between sleep duration and the incidence of MS was observed in males. Sleep less than 6 h, 8–9 h, and ≥9 h were risk factors for developing MS in males (HR = 1.87, 1.73, and 1.96; 95% CI = 1.51–2.30, 1.37–2.19, and 1.35–2.85, respectively; Table 3). Sleeping less than 6 h, 8–9 h, or ≥9 h also increased the incidence of an elevated FBG level. The HRs were 2.27 (95% CI = 1.63–3.17), 2.28 (95% CI = 1.62–3.21), and 3.16 (95% CI = 1.83–5.46), respectively. Moreover,
people who slept >8 h were more likely to have elevated serum TG (HR8–9 h = 1.48, 95% CI = 1.20–1.82; HR≥9 h = 1.94, 95% CI = 1.36–2.76). Although the HRs were not significantly different, subjects sleeping ≤6 h tended to have higher TG levels (HRs = 1.23, p for trend = 0.051). In the analysis of females, we also adjusted for menopausal status in addition to the aforementioned multiple potential confounding factors; however, the HRs for incident MS among females who slept shorter or longer were not significantly different (all p trends > 0.05) compared with those who slept 7–8 h (Table 4). Of all the components of MS, sleep < 6 h, 6–7 h, and > 9 h were associated with a significantly higher risk of hypertension (HR6 h = 1.25, 95% CI = 1.01–1.56; HR6–7 h = 1.46, 95% CI = 1.17–1.84; HR≥9 h = 1.72, 95% CI = 1.20–2.47). 4. Discussion This is the first and largest sample-size cohort study focusing on the relationship between sleep duration and incidence of MS. The results demonstrated that sleep <6 h or >8 h added to the risk of MS in the general population and in males; however, no such association was observed in females. For the components of MS, both short and long sleep durations tended to increase the incidence of impaired fasting glucose in the general population and in males, while long sleep duration increased the risk of hypertriglyceridemia in males. Both short and long sleep durations increased the risk of hypertension in females. Several studies have suggested that sleep deprivation alone adds to the risk of MS [6,8,16]; in contrast, this cohort study showed that a longer sleep duration may also be a risk factor. Various factors may contribute to this difference, such as single-gender surveys, geographic areas and ethnic variations, differences in the definition of
952
X. Li et al./Sleep Medicine 16 (2015) 949–954
Table 2 Risks for development of metabolic syndrome and its components according to sleep duration among 4774 participants during follow-up. HRs (95% CI)
MS Cases (%) Person-years Incidence density (per 1000 person-years) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Elevated FBG Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Hypertension Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Central obesity Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Reduced HDL-c Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI) Hypertriglyceridemia Cases (%) Model 1 (95% CI) Model 2 (95% CI) Model 3 (95% CI)
<6
6 to < 7
7 to <8
8 to <9 h
≥9 h
435 (39.6) 5030 86.4 1.28* (1.12–1.47) 1.32* (1.14–1.54) 1.43* (1.23–1.67)
324 (30.4) 4637 69.9 1.09 (0.94–1.26) 1.08 (0.92–1.28) 1.09 (0.92–1.29)
404 (27.6) 6271 64.4 1 (Ref.) 1 (Ref.) 1 (Ref.)
274 (29.3) 4112 66.6 1.03 (0.88–1.20) 1.06 (0.89–1.26) 1.25* (1.05–1.50)
69 (33.5) 934 73.9 1.12 (0.87–1.45) 1.32 (1.00–1.75) 1.45* (1.09–1.93)
188 (17.1) 1.38* (1.12–1.70) 1.44* (1.13–1.84) 1.61* (1.24–2.08)
119 (11.2) 1.03 (0.82–1.31) 0.95 (0.72–1.25) 1.02 (0.76–1.36)
164 (11.2) 1 (Ref.) 1 (Ref.) 1 (Ref.)
120 (12.8) 1.17 (0.92–1.48) 1.43* (1.10–1.85) 1.65* (1.26–2.17)
34 (16.5) 1.42 (0.98–2.05) 1.76* (1.16–2.69) 1.98* (1.28–3.05)
612 (55.7) 1.05 (0.94–1.17) 1.07 (0.95–1.20) 1.10 (0.98–1.24)
476 (44.6) 0.96 (0.85–1.08) 1.04 (0.92–1.18) 1.01 (0.89–1.15)
680 (46.4) 1 (Ref.) 1 (Ref.) 1 (Ref.)
406 (43.4) 0.93 (0.82–1.05) 0.94 (0.83–1.08) 1.00 (0.87–1.14)
99 (48.1) 0.96 (0.78–1.19) 1.03 (0.81–1.31) 1.07 (0.84–1.37)
725 (66.0) 1.01 (0.91–1.11) 1.02 (0.92–1.14) 1.06 (0.95–1.18)
660 (61.9) 1.07 (0.96–1.18) 1.03 (0.92–1.16) 1.00 (0.89–1.13)
834 (56.9) 1 (Ref.) 1 (Ref.) 1 (Ref.)
583 (62.3) 1.08 (0.97–1.20) 1.03 (0.92–1.16) 1.08 (0.96–1.22)
148 (71.8) 1.16 (0.98–1.38) 1.12 (0.91–1.37) 1.15 (0.94–1.42)
351 (31.9) 0.80* (0.70–0.92) 0.85* (0.73–0.99) 0.89 (0.77–1.04)
396 (37.1) 0.98 (0.86–1.13) 0.98 (0.84–1.14) 0.96 (0.82–1.12)
462 (31.5) 1 (Ref.) 1 (Ref.) 1 (Ref.)
323 (34.5) 0.92 (0.80–1.06) 0.92 (0.78–1.08) 0.97 (0.82–1.14)
54 (26.2) 0.67* (0.51–0.89) 0.83 (0.62–1.11) 0.87 (0.65–1.16)
444 (40.4) 1.06 (0.93–1.20) 1.06 (0.92–1.22) 1.11 (0.96–1.29)
310 (29.1) 0.85* (0.73–0.98) 0.82* (0.70–0.96) 0.88 (0.75–1.03)
494 (33.7) 1 (Ref.) 1 (Ref.) 1 (Ref.)
319 (34.1) 0.97 (0.84–1.12) 0.95 (0.82–1.12) 1.10 (0.93–1.29)
84 (40.8) 1.11 (0.88–1.40) 1.08 (0.83–1.41) 1.21 (0.93–1.59)
Data are HRs (95% CI, confidence interval); Model 1 adjusted by age, sex; Model 2 adjusted by model 1 plus SBP, smoking, alcohol use, physical activity level, education, psychological pressure, bad mood, stroke and cardiovascular disease; Model 3 adjusted by model 2 plus mental illness, insomnia, use of hypnotics, sleep quality, sleep in daytime, snoring, WC, FBG and TG. *P < 0.05.
MS, and limited confounding factors. Thus, we analyzed a mixedgender population, as well as males and females separately. In addition, a greater quantity of important influencing factors than previous research has been analyzed. First, napping is a common
practice in Chinese adults. Frequent napping and long nap duration were shown to increase the risk of type 2 diabetes mellitus [17,18]. In in vitro experiments, sleeping during the daytime resulted in higher levels of insulin-like growth factor binding protein
Table 3 Relative risks for development of MS according to sleep duration among males. HRs (95% CI)
MS Elevated FBG Hypertension Central obesity Reduced HDL-c Hypertriglyceridemia
<6
6 to < 7
7 to < 8
8 to <9 h
≥9 h
1.87* (1.51–2.30) 2.27* (1.63–3.17) 1.06 (0.91–1.24) 1.13 (0.98–1.31) 0.80 (0.61–1.05) 1.23 (1.00–1.50)
0.78 (0.59–1.03) 0.74 (0.48–1.15) 0.85 (0.71–1.01) 0.96 (0.81–1.12) 0.83 (0.62–1.12) 0.79 (0.62–1.02)
1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.)
1.73* (1.37–2.19) 2.28* (1.62–3.21) 0.94 (0.79–1.12) 1.07 (0.90–1.25) 1.21 (0.93–1.58) 1.48* (1.20–1.82)
1.96* (1.35–2.85) 3.16* (1.83–5.46) 0.78 (0.55–1.12) 1.19 (0.90–1.58) 1.46 (0.93–2.30) 1.94* (1.36–2.76)
* P < 0.05.
Table 4 Relative risks for development of MS according to sleep duration among females. HRs (95% CI)
MS Elevated FBG Hypertension Central obesity Reduced HDL-c Hypertriglyceridemia * P < 0.05.
<6
6 to < 7
7 to < 8
8 to <9 h
≥9 h
0.93 (0.73–1.19) 0.68 (0.44–1.05) 1.25* (1.01–1.56) 0.97 (0.81–1.16) 0.93 (0.76–1.14) 0.91 (0.73–1.14)
1.27 (0.99–1.64) 0.84 (0.53–1.29) 1.46* (1.17–1.84) 1.07 (0.89–1.29) 1.08 (0.88–1.33) 1.04 (0.81–1.32)
1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.) 1 (ref.)
0.87 (0.65–1.17) 0.84 (0.52–1.34) 1.19 (0.93–1.52) 1.13 (0.94–1.38) 0.90 (0.72–1.13) 0.80 (0.61–1.05)
0.93 (0.60–1.53) 0.53 (0.19–1.52) 1.72* (1.20–2.47) 1.11 (0.81–1.53) 0.76 (0.51–1.12) 0.80 (0.51–1.25)
X. Li et al./Sleep Medicine 16 (2015) 949–954
1 (IGFBP1) than nocturnal sleep [19]. In addition, cell and molecular links exist between disorders of circadian rhythms and sleep with MS [20]. Thus, independence from daytime sleeping should be considered with respect to the relationship between nocturnal sleep duration and incidence of MS. Second, subjects with MS have poorer sleep quality [21]. Poor sleep efficiency, such as sleepdisordered breathing, poor sleep continuity, and “lighter” sleep, has a significant correlation with overactivity of the sympathetic nervous system, which could result in insulin resistance [22,23]. In view of this, a series of sleep parameters were considered as confounding factors in the assessment of the relative risks for MS, which included sleep in the daytime, insomnia, use of hypnotics, selfreported sleep quality, and snoring. Bad mood was also adjusted in Cox regression models because a significant interaction between sleepiness and depression was also identified [24]. In this longitudinal research, U-shaped relationships between sleep duration and MS were observed in a mixed gender population and in males. This conclusion was supported by several previous cross-sectional studies without gender-stratification or just among special ages [25,26]. Regarding the MS components, shorter and longer sleepers had increased risks of elevated FBG and males who slept long were at high risk for hypertriglyceridemia, as reported in several crosssectional studies [12,27]; however, cross-sectional analysis could not identify a potential cause and effect relationship. The results derived from a prospective cohort study further indicated the causality may exist between sleep duration and some of the aforementioned MS components. The mechanisms underlying the relationship between sleep duration and MS remain to be elucidated. Indeed, sleeplessrelated hypothalamic–pituitary–adrenal hyperactivity, elevated cortisol secretion levels, altered growth hormone metabolism, and low-grade activated inflammation have been shown to be associated with the pathogenesis of MS [28–30]. In contrast, short sleep duration has adverse effects on levels of leptin, visfatin, and ghrelin, increases appetite and caloric intake, then reduces energy secretion and impairs glycemic control [10,31,32]. Theories to explain the relationship between long sleep duration and metabolic abnormalities have been advanced [33,34]. For example, the waking time to undertake physical activity was shorter for long sleepers, which may lead to an increased risk of MS [33]. Both short and long sleep durations increased the risk for hypertension in females, but not in males. One of the potential mechanisms to explain the increased risk for hypertension in females may be associated with menopause and psychosocial stress [35], which may be interrelated with the increased risk for MS and risk factors [36]. In the current study, both menopausal status and psychological pressure were analyzed as confounding factors. Except for short sleep duration, a correlation between long sleep duration and hypertension was also noted in females, while a U-shaped relationship was not demonstrated in males. Similarly, 24 h of total sleep deprivation reduced muscle sympathetic nerve activity in males, but not in females. Gender differences in muscle sympathetic nerve activity were associated with gender differences in sympathetic baroreflex function and testosterone [37]. There were several limitations in this study. First, this study was limited to a Chinese population, and the results may not be extended to other races and ethnic groups. Second, we did not examine hormone levels, which are increasingly indicated to regulate energy balance. A further limitation was that self-reported sleep duration may not be as precise as when measured objectively. We only analyzed subjective sleep quality, and failed to adopt the poor global sleep-quality scores on the Pittsburgh Sleep Quality Index [38]; however, we have adjusted sleeping habits as well as sleep quality from self-reported questionnaire information as confounding factors to minimize the bias.
953
In conclusion, both short and long sleep durations increased the incidence of MS and hyperglycemia in the general population and males. In addition, males who slept >8 h were shown to be at higher risk for hypertriglyceridemia. Females who slept ≤7 h and ≥9 h were associated with incident hypertension.
Conflicts of interest There are no conflicts of interest. The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2015.03.024.
Acknowledgements This work was financially supported by the Chang Jiang Scholar Candidates Program of the Provincial Universities in Heilongjiang, China (grant No. 2012CJHB006).
References [1] Zhao Y, Yan H, Yang R, et al. Prevalence and determinants of metabolic syndrome among adults in a rural area of Northwest China. PLoS ONE 2014;9:e91578. [2] Bayon V, Leger D, Gomez-Merino D, et al. Sleep debt and obesity. Ann Med 2014;46:264–72. [3] Xiao Q, Keadle SK, Hollenbeck AR, et al. Sleep duration and total and causespecific mortality in a large us cohort: interrelationships with physical activity, sedentary behavior, and body mass index. Am J Epidemiol 2014;doi:10.1093/ aje/kwu222. [4] Antonijevic IA, Murck H, Frieboes R, et al. On the gender differences in sleependocrine regulation in young normal humans. Neuroendocrinology 1999;70:280–7. [5] Bonnet MHAD. We are chronically sleep deprived. Sleep 1995;18:908–11. [6] Choi JK, Kim MY, Kim JK, et al. Association between short sleep duration and high incidence of metabolic syndrome in midlife women. Tohoku J Exp Med 2011;225:187–93. [7] Rechtschaffen A, Bergmann BM, Everson CA, et al. Sleep deprivation in the rat: X. Integration and discussion of the findings. Sleep 1989;25:68–87. [8] Otsuka T, Kawada T, Yanai M, et al. Incidence of metabolic syndrome and associated lifestyle factors in a worksite male population. Sangyo Eiseigaku Zasshi 2011;53:78–86. [9] Wu MC, Yang YC, Wu JS, et al. Short sleep duration associated with a higher prevalence of metabolic syndrome in an apparently healthy population. Prev Med 2012;55:305–9. [10] Hayes AL, Xu F, Babineau D, et al. Sleep duration and circulating adipokine levels. Sleep 2011;34:147–52. [11] Sharma S, Kavuru M. Sleep and metabolism: an overview. Int J Endocrinol 2010;2010. [12] Kaneita Y, Uchiyama M, Yoshiike N, et al. Associations of usual sleep duration with serum lipid and lipoprotein levels. Sleep 2008;31:645–52. [13] Buxton OM, Marcelli E. Short and long sleep are positively associated with obesity, diabetes, hypertension, and cardiovascular disease among adults in the United States. Soc Sci Med 2010;71:1027–36. [14] Alberti KG, Eckel RH, Grundy SM, et al. Harmonizing the metabolic syndrome: a joint interim 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. Circulation 2009;120:1640–5. [15] Tierney M, Fraser A, Kennedy N. Criterion validity of the International Physical Activity Questionnaire Short Form (IPAQ-SF) for use in patients with rheumatoid arthritis: comparison with the SenseWear Armband. Physiotherapy 2014;doi:10.1016/j.physio.2014.07.005. [16] Chaput JP, McNeil J, Despres JP, et al. Short sleep duration as a risk factor for the development of the metabolic syndrome in adults. Prev Med 2013;57. [17] Picarsic JL, Glynn NW, Taylor CA, et al. Self-reported napping and duration and quality of sleep in the lifestyle interventions and independence for elders pilot study. J Am Geriatr Soc 2008;56:1674–80. [18] Lam KB, Jiang CQ, Thomas GN, et al. Napping is associated with increased risk of type 2 diabetes: the Guangzhou Biobank Cohort Study. Sleep 2010;33:402–7. [19] Rehman JU, Brismar K, Holmback U, et al. Sleeping during the day: effects on the 24-h patterns of IGF-binding protein 1, insulin, glucose, cortisol, and growth hormone. Eur J Endocrinol 2010;163:383–90. [20] Maury E, Ramsey KM, Bass J. Circadian rhythms and metabolic syndrome: from experimental genetics to human disease. Circ Res 2010;106:447–62. [21] Hung HC, Yang YC, Ou HY, et al. The association between self-reported sleep quality and metabolic syndrome. PLoS ONE 2013;8:e54304.
954
X. Li et al./Sleep Medicine 16 (2015) 949–954
[22] Hall MH, Okun ML, Sowers M, et al. Sleep is associated with the metabolic syndrome in a multi-ethnic cohort of midlife women: the SWAN Sleep Study. Sleep 2012;35:783–90. [23] Tasali E, Leproult R, Ehrmann DA, et al. Slow-wave sleep and the risk of type 2 diabetes in humans. Proc Natl Acad Sci U S A 2008;105:1044–9. [24] Ronksley PE, Hemmelgarn BR, Heitman SJ, et al. Excessive daytime sleepiness is associated with increased health care utilization among patients referred for assessment of OSA. Sleep 2011;34:363–70. [25] Hall MH, Muldoon MF, Jennings JR, et al. Self-reported sleep duration is associated with the metabolic syndrome in midlife adults. Sleep 2008;31:635– 43. [26] Choi KM, Lee JS, Park HS, et al. Relationship between sleep duration and the metabolic syndrome: Korean National Health and Nutrition Survey 2001. Int J Obes (Lond) 2008;32:1091–7. [27] Yaggi HK, Araujo AB, McKinlay JB. Sleep duration as a risk factor for the development of type 2 diabetes. Diabetes Care 2006;29:657–61. [28] Miller MA, Cappuccio FP. Inflammation, sleep, obesity and cardiovascular disease. Curr Vasc Pharmacol 2007;5:93–102. [29] Straznicky NE, Eikelis N, Lambert EA, et al. Mediators of sympathetic activation in metabolic syndrome obesity. Curr Hypertens Rep 2008;10:440–7. [30] Vgontzas AN, Chrousos GP. Sleep, the hypothalamic-pituitary-adrenal axis, and cytokines: multiple interactions and disturbances in sleep disorders. Endocrinol Metab Clin North Am 2002;31:15–36.
[31] St-Onge MP, Roberts AL, Chen J, et al. Short sleep duration increases energy intakes but does not change energy expenditure in normal-weight individuals. Am J Clin Nutr 2011;94:410–16. [32] Nedeltcheva AV, Kilkus JM, Imperial J, et al. Sleep curtailment is accompanied by increased intake of calories from snacks. Am J Clin Nutr 2009;89:126–33. [33] Arora T, Jiang CQ, Thomas GN, et al. Self-reported long total sleep duration is associated with metabolic syndrome: the Guangzhou Biobank Cohort Study. Diabetes Care 2011;34:2317–19. [34] Ayas NT, White DP, Al-Delaimy WK, et al. A prospective study of self-reported sleep duration and incident diabetes in women. Diabetes Care 2003;26:380–4. [35] Cappuccio FP, Stranges S, Kandala NB, et al. Gender-specific associations of short sleep duration with prevalent and incident hypertension: the Whitehall II Study. Hypertension 2007;50:693–700. [36] Janssen I, Powell LH, Crawford S, et al. Menopause and the metabolic syndrome: the Study of Women’s Health Across the Nation. Arch Intern Med 2008;168:1568–75. [37] Carter JR, Durocher JJ, Larson RA, et al. Sympathetic neural responses to 24-hour sleep deprivation in humans: sex differences. Am J Physiol Heart Circ Physiol 2012;302:H1991–7. [38] Jennings JR, Muldoon MF, Hall M, et al. Self-reported sleep quality is associated with the metabolic syndrome. Sleep 2007;30:219–23.