Relationship between sleep duration and the metabolic syndrome ...

International Journal of Obesity (2008) 32, 1091–1097 & 2008 Macmillan Publishers Limited All rights reserved 0307-0565/08 $30.00 www.nature.com/ijo

ORIGINAL ARTICLE Relationship between sleep duration and the metabolic syndrome: Korean National Health and Nutrition Survey 2001 KM Choi1, JS Lee2, HS Park3, SH Baik1, DS Choi1 and SM Kim4 1 Department of Internal Medicine, Division of Endocrinology and Metabolism, Department, College of Medicine, Korea University, Seoul, South Korea; 2Division of Biostatistics, Graduate School of Public Health Department, College of Medicine, Korea University, Seoul, South Korea; 3Department of Family Medicine, University of Ulsan College of Medicine, Seoul, South Korea and 4Department of Family Medicine, College of Medicine, Korea University, Seoul, South Korea

Objectives: Previous studies have revealed that both short and long sleep durations are linked to obesity, hyperglycemia and hypertension. We evaluate the relationship between sleep duration and the metabolic syndrome using representative national survey data from the Korean population. Methods: We analyzed data from the 2001 Korean National Health and Nutrition Survey. The average amount of sleep per night was categorized as: p5, 6, 7, 8 and X9 h. Those reporting 7 h per night served as a reference group. In this cross-sectional study, the data of 4222 participants were finally analyzed. Results: A majority of the components of the metabolic syndrome demonstrated U-shaped patterns based on sleep duration. Although the prevalences of abdominal obesity and hypertension were highest in subjects who slept p5 h per night, those of hyperglycemia and high triglyceridemia were highest in subjects who slept X9 h per night. Prevalence of the metabolic syndrome also exhibited U-shape pattern based on sleep duration. More components of the metabolic syndrome were highly associated with sleep duration in subjects under the age of 60 compared to those over the age of 60. Subjects who slept p5 h per night demonstrated the highest risk for the metabolic syndrome (OR 1.74, 95% CI 1.33–2.26, Po0.001). Subjects who slept X9 h per night exhibited increased risk for the metabolic syndrome even after adjustment of other risk variables (OR 1.69, 95% CI 1.17–2.45, P ¼ 0.006). Conclusions: Both short and long sleep durations are related to increased risk of the metabolic syndrome and its components in the Korean population, although adjustment for risk factors attenuates their relationship. Subjects reporting sleep duration of 7 h demonstrated the lowest prevalence of the metabolic syndrome. International Journal of Obesity (2008) 32, 1091–1097; doi:10.1038/ijo.2008.62; published online 13 May 2008 Keywords: metabolic syndrome; sleep duration; hypertension; Korean National Health and Nutrition Survey

Introduction Sleep loss due to self-imposed bedtime restriction has become a common phenomenon in modern society.1 Previous studies have indicated that short sleep duration is associated with elevated body mass index (BMI).2,3 In a recent prospective study, Hasler et al.4 determined a longitudinal relationship between short sleep duration and future weight gain. Additionally, evidences from laboratory and

Correspondence: Dr SM Kim, Department of Family Medicine, Graduate School of Public Health Department, College of Medicine, Korea University Guro Hospital, 80 Guro-Dong, Guro-Gu, Seoul 152-050, Korea. E-mail: [email protected] Received 18 September 2007; revised 23 March 2008; accepted 24 March 2008; published online 13 May 2008

epidemiological studies have indicated that insufficient sleep may result in decreased glucose tolerance and increased risk for type 2 diabetes.5,6 Decreased insulin sensitivity has been observed with different lengths of sleep deprivation.7 Furthermore, a recent epidemiological study demonstrated an increased risk of hypertension incidence with short sleep duration, even after controlling for obesity and diabetes.8 Sleep deprivation studies have shown significant increases in blood pressure and sympathetic nervous system activity in both normotensive and hypertensive subjects.9,10 Altogether, the data suggest that sleep duration may be an important risk factor for the metabolic syndrome. Although scanty previous reports showed an association between sleep duration and several components of the metabolic syndrome,11 to our knowledge, there are no previous reports on a direct relationship between sleep duration and the

Sleep duration and the metabolic syndrome KM Choi et al

1092 metabolic syndrome. Furthermore, most previous studies on sleep duration and components of the metabolic syndrome have been based on a Caucasian population. In the present study, we have examined the relationship between sleep duration and the metabolic syndrome using representative data from the Korean population.

Subjects and methods Study subjects The Korean National Health and Nutrition Survey (KNHNS) was performed by the Korean Ministry of Health and Welfare. The 2001 KNHNS was a cross-sectional and nationally representative survey, the details of which have been previously published.12,13 A stratified multistage probability sampling design was used, with selections made from sampling units based on geographical area, sex, and age based on household registries. There were 246 097 primary sampling units, each of which contained approximately 60 households. Two hundred sampling frames (12 180 households) from primary sampling units were randomly selected throughout South Korea. The survey was completed by 9770 of 12 642 individuals who participated in the Health Examination Study, resulting in a participation rate of 77.3%. After exclusion of participants under 20 years of age (n ¼ 3169), we utilized data collected from 6601 subjects over 20 years of age for this study. The diagnostic criteria of the metabolic syndrome in children were different from those used for adults, and previous studies showed extreme heterogeneity for the sets of variables and cutoff values. Therefore, we excluded subjects under 20 years of age in the present study. Among the 6601 participants, those without fasting time data (n ¼ 236), those who did not fast for a long enough period of time (n ¼ 1253), and those without sleep duration data (n ¼ 342) were excluded. Then, we further excluded subjects who were taking medications to treat the metabolic syndrome (n ¼ 548), resulting in a total final sample size of 4222 subjects (1822 men, 2400 women) for analysis. Excluded subjects were more likely to have higher age (48.1±0.5 for excluded vs 44.1±0.4 for included), have higher waist circumference (82.5±0.3 cm for excluded vs 80.3±0.2 cm for included), have higher BMI (23.8±0.1 for excluded vs 23.3±0.1 for included), have higher systolic (126.2±0.7 mm Hg for excluded vs 120.9±0.5 mm Hg for included) and diastolic blood pressure (78.6±0.4 mm Hg for excluded vs 76.6±0.3 mm Hg for included). However, gender distribution was similar (men 44.3% for excluded vs 43.2% for included). Each respondent was assigned a weight, based on geographic and demographic characteristics, to allow findings to be extrapolated for the entire Korean population. Nurses were trained to carry out anthropometric measurements, serum collection, blood pressure measurements and questionnaires management. The questionnaires included items concerning the demographic, socioeconomic, dietary International Journal of Obesity

and medical history details of each respondent. Self-reported alcohol intake, smoking and physical exercise were estimated from the questionnaire. Alcohol consumption was categorized based on the usual frequency and intake of alcoholic beverages. The categories are as follows: none, X1 cup once a week, X5 cups in men or X4 cups in women 1–2 times a week and X5 cups in men or X4 cups in women three times a week or more.14 This quantity-frequency method is based on the guideline of WHO (World Health Organization) and modification of the Korean Ministry of Health and Welfare.15 In the survey, individuals were classified as nonsmokers, ex-smokers or current smokers. Physical exercise was defined according to estimated energy consumption per week. Categories included: none, o7.5, 7.5–15 and 415 kcal kg1 per week .14 Using the definition by the Korean government, small towns and villages with a population of p50 000 were designated as rural areas.16 A family history of hypertension or diabetes was considered positive if at least one of the parents or one sibling had diabetes or hypertension.

Anthrophometric and laboratory measurement Body weight and height were measured with subjects wearing light clothing and no shoes. BMI was calculated as weight (in kg) divided by height (in m2). Waist circumference was measured from the narrowest point between the lower borders of the rib cage and the iliac crest. Blood pressure was measured in a sitting position after a 10-min rest period. Two systolic and diastolic blood pressure readings were recorded with a 5-min interval and averaged for analysis. Fasting blood sample was taken in the morning following at least an 8-h fast. Blood samples were centrifuged, refrigerated at the examination site and transferred in ice boxes to a central laboratory in Seoul on the same day samples were taken. Plasma glucose, total cholesterol, triglyceride and high-density lipoprotein (HDL)-cholesterol were measured using an auto-analyzer (Hitachi 747 autoanalyzer, Tokyo, Japan). Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation for subjects with serum triglyceride levels p400 mg per 100 ml.17

Definition of the metabolic syndrome We defined subjects as having the metabolic syndrome who had three or more of the following criteria as defined by the National Cholesterol Education Program:18,19 (1) central obesity; (2) hypertriglyceridemia with fasting plasma triglycerides X1.69 mmol l1 (150 mg per 100 ml); (3) low HDL cholesterol with fasting HDL cholesterol o1.04 mmol l1 (40 mg per 100 ml) in men and o1.29 mmol l1 (50 mg per 100 ml) in women; (4) hypertension with systolic or diastolic blood pressureX130/85 mm Hg and (5) hyperglycemia with fasting plasma glucose X5.6 mmol l1 (100 mg per 100 ml). To define central obesity, we used the Asia-Pacific criteria for


1093 obesity based on waist circumference as defined by WHO.20 Specifically, the waist circumference threshold for central obesity is X90 cm in men and X80 cm in women.

Statistical analysis All statistical analyses were performed using SUDAAN release 8.0 (Research Triangle Institute, Research Triangle Park, NC, USA) to reflect the characteristics of the study’s multistage sampling design. The age-adjusted prevalence of each component or cluster of the metabolic syndrome components were calculated using a direct adjustment method, presented as % and 95% confidence interval (CI). The 2001 National Census data from the Korea National Statistical Office was used to define the standard population. Survey weights were used to calculate standard errors in prevalence rates. Crude and adjusted ORs for the metabolic syndrome were obtained using multiple logistic regression analysis and adjusting for age, gender, family history of hypertension or diabetes, residential area, education level, monthly income, alcohol, smoking and exercise according to each model.

Results Table 1 summarizes the characteristics of study population according to self-reported sleep duration. Cardiovascular risk variables such as waist circumference, blood pressure and fasting lipid profiles exhibited a trend of U-shape pattern. The post-hoc analysis showed that in most of the cardiovascular disease (CVD) risk factors the main significant difference is between people who slept o5 h and those who slept 7 h. Subjects reporting 7 h of sleep per night demonstrated the least propensity for these risk variables, Table 1

although they had the highest family history of hypertension or diabetes. In Table 2, prevalences and 95% CIs of the metabolic syndrome components classified by sleep duration are shown. Subjects reporting sleep duration of 7 h exhibited the lowest prevalence of any components of the metabolic syndrome except low HDL cholesterol. Although the prevalences of abdominal obesity and hypertension were highest in subjects who slept p5 h per night, those of hyperglycemia and high triglyceridemia were highest in subjects who slept X9 h per night. The prevalence of the metabolic syndrome also exhibited U-shape pattern based on sleep duration. Subjects reporting sleep duration of 7 h demonstrated the lowest prevalence of the metabolic syndrome. Table 3 shows the prevalence of the metabolic syndrome components classified by sleep duration in younger and elderly subjects. The parameters that significantly differ by sleep duration for younger subjects did not appreciably differ for older subjects. The parameters that approached being significantly different by sleep duration for older subjects did not appreciably differ for younger subjects. More components of the metabolic syndrome were highly associated with sleep duration in subjects under the age of 60 compared to those over the age of 60. We calculated ORs and 95% CIs using multiple logistic regression analyses for the metabolic syndrome classified according to sleep duration (Table 4). With those reporting sleep duration of 7 h serving as the reference group, both short and long sleep durations were linked to an increased risk for the metabolic syndrome in an unadjusted model. Subjects who slept p5 h per night were at the greatest risk for the metabolic syndrome (OR 1.74, 95% CI 1.33–2.26, Po0.001). Even after adjustment for age, gender, BMI, family

Clinical and biochemical characteristics classified by sleep duration Pz

Sleep duration (hour per day)

Age (years) Waist circumference (cm) Body mass index Systolic BP (mm Hg) Diastolic BP (mm Hg) FBG (mmol l1) Total cholesterol (mmol l1) HDL cholesterol (mmol l1) LDL cholesterol (mmol l1) Triglyceride (mmol l1) Monthly incomea, n (%) p100 101–200 201–300 X301

p5

6

7

8

X9

51.3±0.9abcd 81.4±0.4a 23.6±0.2 124.6±1.1abc 77.6±0.7a 5.35±0.05 4.96±0.04 1.19±0.02 3.05±0.04abc 1.52±0.04

45.0±0.6aef 80.5±0.3 23.4±0.1 121.5±0.7ad 77.2±0.4b 5.35±0.03 4.87±0.03 1.20±0.01 2.95±0.03a 1.52±0.03

41.8±0.5be 76.7±0.3a 23.3±0.1 118.6±0.8bdef 75.4±0.4abc 5.27±0.03 4.80±0.03 1.20±0.01 2.91±0.03b 1.46±0.03

42.5±0.6cf 80.4±0.3 23.3±0.1 120.8±0.7ce 77.0±0.5c 5.37±0.04 4.88±0.03 1.22±0.01 2.93±0.03c 1.55±0.03

43.5±1.2d 80.0±0.7 22.9±0.2 122.0±1.2f 76.3±0.9 5.34±0.05 4.84±0.06 1.17±0.02 2.95±0.05 1.51±0.06

271 (23.9) 454 (26.7) 170 (24.4) 86 (21.6)

96 (8.0) 126 (7.5) 45 (5.4) 17 (5.6)

o0.001 0.015 0.098 o0.001 0.003 0.079 0.057 0.190 0.024 0.244 o0.001

186 (16.1) 175 (10.4) 83 (11.2) 37 (10.1)

316 446 203 119

(28.9) (25.0) (25.0) (29.0)

260 536 248 131

(23.0) (30.4) (33.1) (33.8)

Abbreviations: BP, blood pressure; FBG, fasting blood glucose; HDL, high-density lipoprotein; LDL, low-density lipoprotein. Demographic characteristics are presented as the mean±s.d. or as numbers and percentages, as appropriate. zP-values were calculated using ANOVA test or w2-test. a,b,c,d,e,fSignificant difference based on linear contrasts multiple comparisons. Units are 10 000 Korean won per month.

International Journal of Obesity


1094 Table 2

Prevalence (95% confidence interval) of the metabolic syndrome components classified by sleep duration p5 h

Elevated blood glucose Abdominal obesity Elevated blood pressure High triglyceride Low HDL cholesterol

32.4 41.4 42.7 29.4 54.2

6h

7h

8h

X9 h

Pa

(27.6–37.6) (35.9–47.2) (37.3–48.4) (25.1–34.2) (49.0–59.4)

34.1 31.5 33.2 31.9 50.4

(30.6–37.9) (28.4–34.8) (29.9–36.6) (28.8–35.2) (46.6–54.1)

30.9 29.2 25.9 27.8 50.2

(27.5–34.4) (26.5–32.2) (22.8–29.2) (25.0–30.8) (46.3–54.0)

35.7 32.1 32.4 32.5 46.8

(31.7–40.0) (28.7–35.7) (28.7–36.3) (29.3–35.8) (43.1–50.5)

35.9 34.5 33.3 34.3 53.9

(30.3–41.8) (28.0–41.7) (27.3–40.0) (27.9–41.3) (47.1–60.6)

0.168 0.008 o0.001 0.129 0.198

Number of components of the metabolic syndrome 0 15.0 (11.8–18.7) X1 85.0 (81.3–88.2) X2 62.1 (57.4–66.5) X3 34.3 (29.2–39.9) X4 12.3 (9.9–15.3) 5 1.4 (0.6–3.0)

18.5 81.5 52.9 27.7 11.3 2.7

(15.9–21.5) (78.5–84.1) (49.3–56.5) (24.6–31.0) (9.5–13.4) (1.9–3.8)

20.5 79.5 48.6 23.2 8.5 1.2

(17.9–23.3) (76.7–82.1) (45.1–52.1) (20.6–26.0) (6.6–10.9) (0.6–2.1)

19.1 80.9 53.3 27.7 9.8 2.5

(16.4–22.2) (77.8–83.6) (49.7–53.8) (24.2–31.5) (7.8–12.2) (1.5–4.3)

17.4 82.6 57.5 31.8 9.9 3.1

(12.8–23.1) (76.9–87.2) (50.8–63.8) (26.0–38.1) (6.3–15.1) (1.6–5.9)

0.157 0.157 o0.001 o0.001 0.138 0.026

Abbreviation: HDL, high-density lipoprotein. aP-values were calculated using w2-test. Table 3

Prevalence (95% confidence interval) of the metabolic syndrome components classified by sleep duration in younger and elderly subjects p5 h

Age o60 Elevated blood glucose Abdominal obesity Elevated blood pressure High triglyceride Low HDL cholesterol

32.7 37.9 34.5 26.4 49.9

6h

7h

8h

X9 h

Pa

(27.3–38.6) (31.3–45.0) (28.8–40.8) (21.8–31.7) (43.8–55.9)

33.1 27.4 27.1 29.6 48.5

(29.1–37.4) (24.0–31.0) (23.6–30.9) (25.9–33.5) (44.5–52.6)

31.5 27.2 21.8 25.0 49.5

(28.0–35.2) (24.5–30.1) (19.1–24.8) (22.3–27.8) (45.5–53.5)

35.1 30.4 27.6 31.2 46.2

(30.9–39.5) (26.8–34.2) (23.8–31.7) (27.7–34.8) (42.2–50.3)

33.6 32.3 23.5 33.4 49.5

(27.3–40.5) (25.0–40.6) (18.0–30.1) (26.7–40.8) (42.0–57.1)

0.659 0.070 0.009 0.035 0.730

Number of components of the metabolic syndrome 0 18.0 (13.9–23.0) X1 82.0 (77.0–86.1) X2 55.7 (50.1–61.2) X3 28.4 (23.1–34.4) X4 10.6 (7.9–14.2)

21.0 79.0 48.1 23.4 9.0

(17.9–24.6) (75.4–82.1) (44.1–52.2) (19.9–27.3) (7.1–11.5)

22.1 77.9 46.2 20.9 6.6

(19.3–25.2) (74.8–80.7) (42.6–49.9) (18.4–23.5) (5.2–8.2)

20.7 79.3 51.0 24.8 8.2

(17.5–24.3) (75.7–82.5) (47.3–54.8) (20.8–29.2) (6.3–10.7)

20.5 79.5 52.8 25.7 6.8

(15.0–27.5) (72.5–85.0) (45.0–60.4) (20.2–32.1) (3.8–12.0)

0.666 0.666 0.024 0.074 0.111

Age X60 Elevated blood glucose Abdominal obesity Elevated blood pressure High triglyceride Low HDL cholesterol

(23.8–40.5) (38.9–59.0) (49.9–69.2) (28.4–43.9) (54.0–72.2)

38.8 50.2 61.0 41.9 59.0

(31.0–47.3) (42.6–57.8) (53.0–68.5) (35.5–48.7) (50.6–66.8)

26.0 44.6 56.9 49.4 55.3

(18.1–35.9) (32.7–57.1) (44.4–68.6) (37.1–61.8) (43.5–66.5)

39.8 43.0 62.5 40.5 50.4

(29.9–50.5) (34.5–51.9) (52.3–71.8) (32.7–48.7) (42.0–58.8)

45.1 43.2 71.5 37.7 70.6

(33.1–57.7) (30.3–57.2) (55.0–83.7) (25.4–51.9) (56.8–81.5)

0.068 0.716 0.623 0.458 0.076

Number of components of the metabolic syndrome 0 8.5 (5.0–14.3) X1 91.5 (85.7–95.0) X2 75.5 (66.3–82.9) X3 46.9 (37.2–56.9) X4 15.9 (11.0–22.5)

7.4 92.6 74.4 46.7 21.3

(4.0–13.3) (86.7–96.0) (66.9–80.7) (39.8–53.7) (15.9–28.0)

7.9 92.1 66.5 40.4 22.8

(4.1–15.0) (85.0–95.9) (55.1–76.2) (28.9–53.1) (12.6–37.7)

9.3 90.7 67.2 46.0 19.7

(4.8–17.3) (82.7–95.2) (57.7–75.7) (38.0–54.3) (12.8–29.1)

4.9 95.1 75.7 55.6 21.7

(1.8–12.3) (87.7–98.2) (63.0–85.1) (41.3–68.9) (12.6–34.7)

0.752 0.752 0.483 0.626 0.591

31.6 48.9 60.0 35.7 63.6

Abbreviation: HDL, high-density lipoprotein. aP-values were calculated using w2-test.

history of hypertension or diabetes, alcohol, smoking, exercise and other risk factors, the relationship between sleep duration and the metabolic syndrome persisted particularly in group reporting 49 h of sleep a night (OR 1.69, 95% CI 1.17–2.45, P ¼ 0.006).

Discussion In the present study, we found a U-shaped association between sleep duration and the metabolic syndrome in the Korean population. International Journal of Obesity

In recent years, self-reported sleep duration of Americans has decreased by 1.5–2 h over the past 40 years, with many Americans sleeping only 5–6 h per night.21 According to the 2001 National Sleep Foundation Survey, approximately 31% of Americans sleep 6 or fewer hours per day.22 In concert with declining sleep duration, there has been an increase in the prevalence of obesity. Short sleep duration has recently drawn attention as a possible cause of obesity. Several epidemiological studies have observed an association between sleep duration and obesity.2,3,23,24 Taheri et al.25 and Patel et al.26 observed a U-shaped relationship between sleep duration and BMI. Furthermore, in a recent longitudinal study, short sleep duration was associated with a


1095 Table 4 Odds ratio (95% confidence interval) of the metabolic syndrome based on sleep duration Unadjusted p5 h 6h 7h 8h X9 h

Model 1a

1.74 (1.33–2.26) 1.23 1.27 (1.04–1.56) 1.13 1 1.27 (1.01–1.63) 1.28 1.55 (1.15–2.07) 1.47

(0.92–1.64) (0.91–1.40) 1 (1.00–1.64) (1.08–2.00)

Model 2b

Model 3c

1.15 (0.85–1.55) 1.17 1.11 (1.01–1.57) 1.07 1 1.26 (1.01–1.57) 1.32 1.46 (1.07–1.98) 1.69

(0.87–1.59) (0.85–1.34) 1 (1.01–1.73) (1.17–2.45)

Subjects were considered to have the metabolic syndrome if they had more than three components according to the definition of National Cholesterol Education Program. aModel 1, adjusted for age, gender and family history of hypertension or diabetes. bModel 2, adjusted for the variables in model 1 plus residential area, education level, monthly income, alcohol, smoking and exercise. cModel 3, adjusted for the variables in model 2 plus BMI.

modest increase in future weight gain and incident obesity.27 In the present study, we confirmed this U-shaped relationship using central obesity criteria based on waist circumference. Participants who slept p5 h per night had the highest prevalence of central obesity. As a mechanism, Taheri et al.25 demonstrated that short sleep duration was associated with reduced leptin and elevated ghrelin levels, which are known to control appetite and body weight. In a randomized crossover clinical trial, sleep deprivation was linked to reductions in the anorexic hormone leptin, elevations in the orexigenic factor ghrelin and increased hunger and appetite.28 In long sleepers, reduced energy expenditure due to increased time in bed may affect their obesity, as a study has shown that long sleeper exercise less.29 Ayas et al.5 reported that both short and long self-reported sleep durations were associated with an increased risk for developing diabetes. Interestingly, the relative risk of diabetes was attenuated in short sleepers, and disappeared after controlling for BMI. However, in long sleepers, a significant positive association between sleep duration and diabetes persisted even after controlling for BMI. Consistent with previous reports,5,6 we found U-shaped pattern for hyperglycemia based on sleep duration in the present study. Previous studies have demonstrated that subjects in the sleep-deprived state have impaired glucose tolerance, higher evening cortisol levels, increased sympathetic nervous system activity and a reduction in leptin secretion compared to those in the recovery state.30 Although mechanisms underlying the association between long sleep duration and diabetes are unclear, chronic subclinical inflammation associated with visceral obesity may trigger long sleep duration and hyperglycemia due to the sleep-inducing31 and metabolic32 effects of pro-inflammatory cytokines, including interleukin-1 and tumor necrosis factor-a. Blood pressure gradually falls with the onset of sleep, and promptly rises after awakening.33 After nights where sleep was restricted to 3.6–4.5 h, significant increases in blood pressure and sympathetic nervous activity were reported.10 Habitually short sleep durations may lead to the development and persistence of hypertension through longer

exposure to daytime stress and elevated sympathetic system activity.8 Recently, Gangwisch et al.8 found an association between short sleep duration and the incidence of hypertension in middle-aged subjects and found no relationship in elderly subjects. Another study confirmed this result in the elderly.34 In the present study, we observed that subjects who slept p5 h per night exhibited the highest prevalence of hypertension, with U-shaped pattern for hypertension based on sleep duration. This association was found only in younger subjects, which is consistent with previous studies. Metabolic syndrome represents a cluster of several metabolic risk factors that are linked to an increase in CVD. the metabolic syndrome results in an approximate twofold increase in the relative risk for cardiovascular events compared to subjects without the syndrome.35 Recently, short and long self-reported sleep duration was reported to be associated with a modestly increased risk of coronary events in women.29 In a prospective study, sleeping o6 h or 47 h a night was associated with increased mortality following adjustment for other risk factors.26 Our results may provide a mechanism linking sleep duration and mortality via mediation of the metabolic syndrome. In the present study, subjects reporting 7 h of sleep per night showed the lowest prevalence for the metabolic syndrome, which is compatible with the previous report’s finding that mortality risk was lowest among women reporting sleep duration of 7 h.26 The results of our study showed U-shaped pattern of prevalence for the metabolic syndrome based on sleep duration in an unadjusted model. We found the attenuation in the relationship between short sleep duration and the metabolic syndrome after adjustment of other variables. The demographic variable indicative of low socio-economic status (SES) may account for this. The relationship between poor health outcomes and low SES has been hypothesized to be due in part to short sleep duration among this group.36 The present study showed that the association of the metabolic syndrome components and sleep duration were different between younger and elderly subjects. It is compatible with previous studies that reported no relationship between sleep duration and obesity in elderly adults.24,27 We performed both weighted and unweighted analyses in the present study and found no appreciable differences in the results between the two. We chose to show the weighted results, because those reflect the characteristics of the study’s multistage sampling design. There are some limitations in our study. First, self-reported sleep duration was used in our analyses rather than actual measurement of sleep duration. However, good correlations have been reported in previous studies between self-reported sleep duration and values obtained through actigraphic monitoring.37 Due to the nature of the present study, which used national epidemiological survey data, it would be impractical to measure sleep duration. Second, the crosssectional design precludes defining causal relationships. International Journal of Obesity


1096 Third, different components of the metabolic syndrome are risk factors for each other. Nevertheless, this study had its own advantages using KNHNS data, as it ensured reliable countrywide sampling and utilized a survey that was both large-scale and nationally representative. Furthermore, the present study analyzed the relationship between sleep duration and the metabolic syndrome after adjustment of various potential risk factors for the metabolic syndrome.

Conclusion These national survey data indicate that both short and long sleep durations are associated with an increased risk for the metabolic syndrome and its components in the Korean population, although adjustment for risk factors attenuates their relationship. Sleep duration might be a potential risk factor for the metabolic syndrome.

9

10

11

12

13

14

15 16

Acknowledgements 17

We thank the members of the Korea Institute for Health and Social Affairs who conducted the national survey. This study was supported by a grant of the Seoul R&BD Program, Republic of Korea (10526).

18

Conflict of interest None. 19

References 1 Spiegel K, Knutson K, Leproult R, Tasali E, Van Cauter E. Sleep loss: a novel risk factor for insulin resistance and type 2 diabetes. J Appl Physiol 2005; 99: 2008–2019. 2 Vorona RD, Winn MP, Babineau TW, Eng BP, Feldman HR, Ware JC. Overweight and obese patients in a primary care population report less sleep than patients with a normal body mass index. Arch Intern Med 2005; 165: 25–30. 3 Kohatsu ND, Tsai R, Young T, Vangilder R, Burmeister LF, Stromquist AM et al. Sleep duration and body mass index in a rural population. Arch Intern Med 2006; 166: 1701–1705. 4 Hasler G, Buysse DJ, Klaghofer R, Gamma A, Ajdacic V, Eich D et al. The association between short sleep duration and obesity in young adults: a 13-year prospective study. Sleep 2004; 27: 661–666. 5 Ayas NT, White DP, Al-Delaimy WK, Manson JE, Stampfer MJ, Speizer FE et al. A prospective study of self-reported sleep duration and incident diabetes in women. Diabetes Care 2003; 26: 380–384. 6 Gottlieb DJ, Punjabi NM, Newman AB, Resnick HE, Redline S, Baldwin CM et al. Association of sleep time with diabetes mellitus and impaired glucose tolerance. Arch Intern Med 2005; 165: 863–867. 7 Gonzalez-Ortiz M, Martinez-Abundis E, Balcazar-Munoz BR, Pascoe-Gonzalez S. Effect of sleep deprivation on insulin sensitivity and cortisol concentration in healthy subjects. Diabetes Nutr Metab 2000; 13: 80–83. 8 Gangwisch JE, Heymsfield SB, Boden-Albala B, Buijs RM, Kreier F, Pickering TG et al. Short sleep duration as a risk factor for


20

21 22 23

24

25

26

27

28

hypertension: analyses of the first National Health and Nutrition Examination Survey. Hypertension 2006; 47: 833–839. Tochikubo O, Ikeda A, Miyajima E, Ishii M. Effects of insufficient sleep on blood pressure monitored by a new multibiomedical recorder. Hypertension 1996; 27: 1318–1324. Lusardi P, Zoppi A, Preti P, Pesce RM, Piazza E, Fogari R. Effects of insufficient sleep on blood pressure in hypertensive patients: a 24-h study. Am J Hypertens 1999; 12: 63–68. Bjorvatn B, Sagen IM, Oyane N, Waage S, Fetveit A, Pallesen S et al. The association between sleep duration, body mass index and metabolic measures in the Hordaland Health Study. J Sleep Res 2007; 16: 66–76. Kim SM, Lee JS, Lee J, Na JK, Han JH, Yoon DK et al. Prevalence of diabetes and impaired fasting glucose in Korea: Korean National Health and Nutrition Survey 2001. Diabetes Care 2006; 29: 226–231. Choi KM, Park HS, Han JH, Lee JS, Lee J, Ryu OH et al. Prevalence of prehypertension and hypertension in a Korean population: Korean National Health and Nutrition Survey 2001. J Hypertens 2006; 24: 1515–1521. Ministry of Health and Social Welfare. Analysis of Major Disease and Health Behavior in Koreans. Ministry of Health and Social Welfare: Gyeonggi-do, South Korea, 2003. WHO. International Guide for Monitoring Alcohol Consumption and Related Harm. WHO: Geneva, 2000. Ministry of Health and Social Welfare. Report on 1998 National Health and Nutrition Survey. Ministry of Health and Social Welfare: Gyeonggi-do, South Korea, 1999. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 2001; 285: 2486–2497. Grundy SM, Brewer Jr HB, Cleeman JI, Smith Jr SC, Lenfant C. Definition of metabolic syndrome: report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation 2004; 109: 433–438. Health Communications Australia Pty Limit. The Asia-Pacific Perspective: Redefining Obesity and its Treatment. Health Communications Australia Pty Limit: Sydney, Australia,, 2000. Jean-Louis G, Kripke DF, Ancoli-Israel S. Sleep and quality of wellbeing. Sleep 2000; 23: 1115–1121. National Sleep Foundation. National Sleep Foundation Sleep Survey. National Sleep Foundation: Washington, D.C., 2001. Reilly JJ, Armstrong J, Dorosty AR, Emmett PM, Ness A, Rogers I et al. Early life risk factors for obesity in childhood: cohort study. BMJ 2005; 330: 1357. Gangwisch JE, Malaspina D, Boden-Albala B, Heymsfield SB. Inadequate sleep as a risk factor for obesity: analyses of the NHANES I. Sleep 2005; 28: 1289–1296. Taheri S, Lin L, Austin D, Young T, Mignot E. Short sleep duration is associated with reduced leptin, elevated ghrelin, and increased body mass index. PLoS Med 2004; 1: e62. Patel SR, Ayas NT, Malhotra MR, White DP, Schernhammer ES, Speizer FE et al. A prospective study of sleep duration and mortality risk in women. Sleep 2004; 27: 440–444. Patel SR, Malhotra A, White DP, Gottlieb DJ, Hu FB. Association between reduced sleep and weight gain in women. Am J Epidemiol 2006; 164: 947–954. Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med 2004; 141: 846–850.


1097 29 Ayas NT, White DP, Manson JE, Stampfer MJ, Speizer FE, Malhotra A et al. A prospective study of sleep duration and coronary heart disease in women. Arch Intern Med 2003; 163: 205–209. 30 Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet 1999; 354: 1435–1439. 31 Kapas L, Hong L, Cady AB, Opp MR, Postlethwaite AE, Seyer JM et al. Somnogenic, pyrogenic, and anorectic activities of tumor necrosis factor-alpha and TNF-alpha fragments. Am J Physiol 1992; 263: R708–R715. 32 Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 1993; 259: 87–91. 33 Kario K, Schwartz JE, Pickering TG. Changes of nocturnal blood pressure dipping status in hypertensives by nighttime dosing of

34

35

36

37

alpha-adrenergic blocker, doxazosin : results from the HALT study. Hypertension 2000; 35: 787–794. van den Berg JF, Tulen JH, Neven AK, Hofman A, Miedema HM, Witteman JC et al. Sleep duration and hypertension are not associated in the elderly. Hypertension 2007; 50: 585–589. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA et al. Diagnosis and management of the metabolic syndrome: an American heart association/national heart, lung, and blood institute scientific statement. Curr Opin Cardiol 2006; 21: 1–6. Van Cauter E, Spiegel K. Sleep as a mediator of the relationship between socioeconomic status and health: a hypothesis. Ann N Y Acad Sci 1999; 896: 254–261. Lockley SW, Skene DJ, Arendt J. Comparison between subjective and actigraphic measurement of sleep and sleep rhythms. J Sleep Res 1999; 8: 175–183.
