Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

Contents lists available at ScienceDirect

Progress in Neuro-Psychopharmacology & Biological Psychiatry

Subjective depressive symptoms and metabolic syndrome among the general population Sang Jin Rhee a, Eun Young Kim a, Se Hyun Kim b, Hyun Jeong Lee a, Bora Kim a, Kyooseob Ha c,d, Dae Hyun Yoon e,⁎, Yong Min Ahn a,c,⁎⁎ a

Department of Psychiatry, Seoul National University Hospital, Yongon-Dong, Chongno-Gu, Seoul 110-744, Republic of Korea Institute of Human Behavioral Medicine, Medical Research Institute, Seoul National University, Yongon-Dong, Chongno-Gu, Seoul 110-744, Republic of Korea Institute of Human Behavioral Medicine, Seoul National University College of Medicine, Yongon-Dong, Chongno-Gu, Seoul 110-744, Republic of Korea d Mood Disorders Clinic and Affective Neuroscience Laboratory, Department of Neuropsychiatry, Seoul National University Bundang Hospital, Seongnam, Gyeonggi 463-707, Republic of Korea e Department of Psychiatry, Seoul National University Hospital, Healthcare System Gangnam Center, Yeoksam-Dong, Gangnam-Gu, Seoul 135-984, Republic of Korea b c

a r t i c l e

i n f o

Article history: Received 3 February 2014 Received in revised form 17 June 2014 Accepted 18 June 2014 Available online 27 June 2014 Keywords: Depression Depressive symptoms Gender Lipid Metabolic syndrome

a b s t r a c t Objective: The evidence of the association between depression and metabolic syndrome is increasing, but the existence of sex differences in this association remains controversial. The aim of this study was to investigate the association between subjective depressive symptoms and metabolic syndrome and each of its components by sex in the Korean population. Methods: The study sample comprised 15,073 men and 15,034 women who underwent routine health examinations. They completed the Beck Depression Inventory for depressive symptoms, and medical examinations provided data regarding metabolic syndrome. Adjustments for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, stroke, angina, and thyroid disease were performed. The association between depressive symptoms and metabolic syndrome and each of its components was analyzed by multiple logistic regression. Results: In women, depressive symptoms were associated with metabolic syndrome (OR = 1.35, 95% CI = 1.11–1.64, p = 0.002) and the high-density lipoprotein cholesterol component (OR = 1.26, 95% CI = 1.09–1.46, p = 0.002) of metabolic syndrome. There was also an association between the severity of depressive symptoms and metabolic syndrome in women (OR = 1.046, 95% CI = 1.002–1.091, p = 0.039). In men, depressive symptoms were inversely associated with the hypertension component of metabolic syndrome (OR = 0.73, 95% CI = 0.58–0.91, p = 0.005). Conclusions: Subjective depressive symptoms were associated with metabolic syndrome only in women. Further research should consider sex differences and dyslipidemia. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Major depressive disorder is one of the most prevalent psychiatric diseases, with a 12-month point prevalence rate of 2.5% (1.7% in men, 3.2% in women) and a lifetime prevalence rate of 5.6% (3.6% in men, 7.6% in women) in the Korean population (Cho et al., 2010). Evidence supports increasing prevalence (Cho et al., 2010) and disease burden

Abbreviations: TG, Triglyceride; HDL, High-density lipoprotein cholesterol; BDI, Beck Depression Inventory-1; OR, Odds ratio; CI, Confidence interval; df, Degrees of freedom; HPA, Hypothalamic–pituitary–adrenal. ⁎ Correspondence to: D.H. Yoon, Department of Psychiatry, Gangnam Health Promotion Center, Seoul National University Hospital, 737 Yeoksam-Dong, Gangnam-Gu, Seoul 135984, Republic of Korea. Tel.: +82 2 2112 5572; fax: +82 2 2112 5794. ⁎⁎ Correspondence to: Y.M. Ahn, Department of Psychiatry, Seoul National University Hospital, Institute of Human Behavioral Medicine, Seoul National University College of Medicine, 28 Yongon-Dong, Chongno-Gu, Seoul 110-744, Republic of Korea. Tel.: +82 2 2072 0710; fax: +82 2 744 7241. E-mail addresses: [email protected] (D.H. Yoon), [email protected] (Y.M. Ahn).

http://dx.doi.org/10.1016/j.pnpbp.2014.06.006 0278-5846/© 2014 Elsevier Inc. All rights reserved.

(Lepine and Briley, 2011) of depression over time. However the increasing prevalence may be explained due to methodological problems (Simon and VonKorff, 1992). As the importance of depression increases, evidence is accumulating with respect to other important outcomes associated with depression, such as cardiovascular disease (Vaccarino et al., 2008), diabetes (Windle and Windle, 2013), and mortality (Mykletun et al., 2007). Metabolic syndrome, which includes a cluster of cardiovascular disease risk factors such as abdominal obesity, high triglyceride (TG) levels, low high-density lipoprotein cholesterol (HDL) levels, high blood pressure, and high glucose levels (Grundy et al., 2004), helps to explain the association between depression and cardiovascular disease. Obesity and insulin resistance are proposed core mechanisms for the development of metabolic syndrome, and lipid profiles are known to be associated with both obesity and insulin resistance (Grundy et al., 2004). Several studies have investigated the association between depression and metabolic syndrome, and a recent meta-analysis (Pan et al., 2012) has shown that depression and metabolic syndrome are cross-

224

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

sectionally as well as bidirectionally associated, increasing the evidence for the association between the two. However there are studies conducted with large samples which have found no association (Herva et al., 2006; Hildrum et al., 2009) between the two conditions. Data showing sex differences in the association between depression and metabolic syndrome have been especially controversial. Some studies have reported that this association is confined to women (Gil et al., 2006; Kinder, 2004; Laudisio et al., 2009; Pulkki-Raback et al., 2009; Toker et al., 2008) or to men (Nishina et al., 2011; Viinamaki et al., 2009), and other studies have reported no sex interactions in the association (Hildrum et al., 2009; Kahl et al., 2012; Muhtz et al., 2009). It is possible that these controversies are attributable, in part, to the existence of different subtypes of depression, given the evidence that the somatic–affective symptom cluster (Marijnissen et al., 2013) and chronic atypical depression (Lamers et al., 2013) are associated with metabolic abnormalities. The present study examined the association between subjective depressive symptoms and metabolic syndrome and its individual components in a large sample of the Korean general population who participated in a health screening program. Because there have been reports of sex differences in this association, we analyzed men and women separately. We hypothesized that depressive symptoms and metabolic abnormalities are associated and that there are sex differences in this association. 2. Methods 2.1. Patient population Patients who underwent routine health examinations at the Seoul National University Hospital Healthcare System Gangnam Center from October 2004 to July 2012 were screened for this retrospective crosssectional study. These health examinations are performed to screen for and enable early diagnosis of certain diseases and are available to all individuals for a fee. 15,073 men and 15,034 women aged 12 to 90 years who provided complete data from the Beck Depression Inventory-1 (BDI) and who had data available on metabolic syndrome were eligible for the study. The Institutional Review Board of Seoul National University Hospital approved the study protocol. 2.2. Assessment of subjective depressive symptoms The severity of subjective depressive symptoms was assessed using the BDI, a 21-item self-report questionnaire that scores each question from 0 to 3 points (total score = 0–63 points) (Beck et al., 1961). A threshold of 19 points has been proposed for moderate depression (Beck et al., 1988), but previous studies have used different cutoff scores to indicate depression (Marijnissen et al., 2013; Miettola et al., 2008) when studying its association with metabolic syndrome. A cutoff score of 21 points had been initially proposed for the Korean population (Hahn et al., 1986). However, the majority of subsequent studies proposed 16 points as the cutoff for clinical depression (Jo et al., 2007; Lee and Song, 1991; Shin et al., 1993). The sensitivity and specificity of the cutoff in those studies were 67.9–71.6% and 65.7%– 77.3%, respectively (Jo et al., 2007; Lee and Song, 1991; Shin et al., 1993). In our study, patients with a total BDI score ≥16 were classified as having clinical depressive symptoms. Assessments of associations involving symptom severity relied on the square-root transformation of the total BDI score (because the distribution was skewed) as a continuous predictor variable. 2.3. Assessment of metabolic syndrome Based on the guidelines of the National Cholesterol Education Program Adult Treatment Panel III and the modified waist circumference criteria of the Korean Society for the Study of Obesity, metabolic

syndrome was defined as the presence of three or more of the following components: (a) abdominal obesity as measured by a waist circumference of N90 cm for men and N85 cm for women, (b) TG level ≥ 150 mg/dL or use of antilipidemic medication, (c) HDL cholesterol level b40 mg/dL for men and b50 mg/dL for women, (d) systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥85 mm Hg or use of antihypertensive medication, and (e) fasting glucose level ≥100 mg/dL or use of antidiabetic medication (Grundy et al., 2004). Waist circumference was measured at the midpoint between the lower border of the rib cage and iliac crest. TG, HDL, and glucose levels were measured after 12 h of fasting. Blood pressure was measured in the right upper arm with an automated sphygmomanometer after at least 10 min of rest. Data on medication use were provided by self-reports.

2.4. Other assessments Structured questionnaires were used to collect information on demographic variables. These included age, marriage, cigarette smoking, alcohol use, exercise, education, previously diagnosed diseases, current medication use including psychotropic medications and oral contraceptives/hormone treatment particularly for women, and menopause for women. Marriage (currently married or currently not married) was treated as a dichotomous variable. Cigarette smoking was divided into three groups (current smoker, ex-smoker, or nonsmoker). Alcohol use (≤ 1/month, 2/month to 2/week, 3–4/week, or ≥5/week) and exercise (b1/week, 1–2/week, 3–4/week, or ≥5/week) were categorized by average number of days per week or month. Education was classified into four groups (less than high school graduate, less than college graduate, college graduate, or post-graduate). Cancer, stroke, angina, and thyroid disease were treated as dichotomous variables.

2.5. Statistical analyses Differences in demographic and metabolic characteristics between patients with and those without depressive symptoms were assessed based on the cutoff value of a BDI score of 16 points. Continuous variables were analyzed with a two-tailed Student's t-test for normally distributed variables and with the Mann–Whitney U-test for variables not normally distributed. Categorical variables were analyzed with the chi-square test. Logistic regression models were used to evaluate the association between depressive symptoms and metabolic syndrome and each of its components. First, single regression analysis was performed with depressive symptoms as a dichotomous independent variable and metabolic syndrome and each of its components as the individual outcome variables. In the multiple regression analyses, age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, stroke, angina, and thyroid disease were added as covariates. Finally, menopause and oral contraceptive/hormone treatment were entered into the models in women to determine whether they influenced the associations. To determine if certain cutoff scores defining metabolic syndrome influenced the association we also conducted linear regression between depressive symptoms with the total number of metabolic syndrome components and the individual continuous metabolic syndrome components. Logistic transformations were used for skewed components. We also used the square-root transformation of total BDI scores as a continuous variable and performed multiple logistic regression to determine whether there was a graded relationship between depressive symptom severity and metabolic syndrome. Statistical significance was defined as p b 0.05. Statistical analyses were performed with SPSS 21.0 (SPSS, Inc., Chicago, IL).

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

3. Results 3.1. Demographic and metabolic characteristics Depressive symptoms were present in 1900 (6.3%) patients, including 456 (3.0%) men and 1444 (9.4%) women. Compared with their nondepressed counterparts, more depressed men and women did not exercise more than 1 day per week, were currently smoking, and had a history of stroke or angina. Additionally, a larger percentage of depressed subjects had completed less than a college education, whereas more non-depressed subjects had at least completed college. Depressed men were significantly more likely to be unmarried at the time of the survey compared with non-depressed men; a similar but non-significant tendency was observed in depressed women. Depressed men and women were more likely than their non-depressed counterparts to consume alcohol at least 5 days per week. Those who consumed alcohol 1 day per month or less were more prevalent among depressed than among non-depressed men but less prevalent among depressed than among non-depressed women. Depressed women were significantly older than their non-depressed counterparts, whereas depressed men were younger than their non-depressed counterparts, although the latter difference was not statistically significant. No significant differences between depressed and not depressed groups of either sex were observed regarding history of cancer or thyroid disease. The detailed baseline demographic characteristics of all subjects are shown in Table 1. Table 2 shows the baseline metabolic syndrome characteristics. In total, 16.1% of patients with clinical depressive symptoms and 14.2% without clinical depressive symptoms had metabolic syndrome, a statistically significant difference (p = 0.030). More patients with subjective clinical depressive symptoms met the HDL criteria for metabolic syndrome, but fewer patients with depressive symptoms met the TG, hypertension, and glucose criteria.

225

Further analysis by sex was performed. In women depressive symptoms were significantly associated with metabolic syndrome (16.1% vs. 11.2%, p b 0.001), each of the components of metabolic syndrome, and the total number of components. However, there was no significant relationship between depressive symptoms and metabolic syndrome (15.8% vs. 17.1%, p = 0.46) or any individual component of metabolic syndrome in men, with the exception of the hypertension, which was more prevalent in the non-depressed group. In terms of the continuous components of metabolic syndrome, all factors, with the exception of systolic blood pressure, were associated with depressive symptoms in women. However, the only differences noted in men were that depressed men had a smaller waist circumference, higher TG levels, and lower blood pressure compared with nondepressed men. 3.2. Association between depressive symptoms and metabolic syndrome and each of its components In the total study population, subjective depressive symptoms were positively associated with metabolic syndrome and the HDL component, and they were inversely associated with the TG, hypertension, and glucose components. After adjusting for confounding factors including sex, only the relationship with HDL was significant. Stratification by sex showed different patterns of associations. In women, depressive symptoms were associated with metabolic syndrome and all of its individual components, but the only association in men was an inverse association between depressive symptoms and hypertension. After adjusting for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease, depressive symptoms were still associated with metabolic syndrome (OR = 1.35, 95% CI = 1.11–1.64, p = 0.002) and with the presence of the HDL component (OR = 1.26, 95% CI = 1.09–1.46, p = 0.002) in women, and they were inversely associated with the hypertension

Table 1 Demographic characteristics of the patient population by depressive symptoms and sex. p-Valuea

Total

Age, mean ± SD, years Marriage Currently not married, n (%) Cigarette smoking Current smoker, n (%) Ex-smoker, n (%) Non-smoker, n (%) Alcohol use ≤1/month, n (%) 2/month to 2/week, n (%) 3–4/week, n (%) ≥5/week, n (%) Exercise ≤1/week, n (%) 1–2/week, n (%) 3–4/week, n (%) ≥5/week, n (%) Education bHigh school graduate, n (%) bCollege graduate, n (%) College graduate, n (%) Post-graduate, n (%) Cancer Stroke Angina Thyroid disease Oral contraceptive/hormone treatment Menopause

p-Valuea

Men

p-Valuea

Women

Depressed

Not depressed

Depressed

Not depressed

Depressed

Not depressed

n = 1900

n = 28,207

n = 456

n = 14,617

n = 1444

n = 13,590

47.5 ± 10.7

47.1 ± 10.4

47.5 ± 11.9

48.5 ± 10.2

47.5 ± 10.4

45.7 ± 10.4

312 (16.6)

2991 (10.7)

101 (22.3)

1125 (7.8)

211 (14.8)

1866 (14.0)

89 (6.9) 71 (5.5) 1136 (87.7)

399 (3.2) 534 (4.3) 11,630 (92.6)

898 (67.7) 369 (27.8) 46 (3.5) 14 (1.1)

8779 (70.3) 3295 (26.4) 327 (2.6) 83 (0.7)

684 (48.9) 285 (20.4) 305 (21.8) 124 (8.9)

5674 (43.0) 2835 (21.5) 3358 (25.4) 1333 (10.1)

137 (9.6) 353 (24.9) 753 (53.0) 177 (12.5) 47 (3.4) 6 (0.4) 25 (1.7) 97 (7.2) 112 (8.4) 548 (40.4)

650 (4.9) 2732 (20.5) 7594 (56.9) 2367 (17.7) 380 (2.9) 16 (0.1) 137 (1.0) 884 (6.9) 730 (5.9) 4257 (33.7)

0.17 b0.001 b0.001

279 (16.0) 251 (14.4) 1215 (69.6)

4947 (18.3) 7144 (26.5) 14,913 (55.2)

1024 (58.2) 570 (32.4) 120 (6.8) 46 (2.6)

11,618 (43.9) 11,477 (43.4) 2665 (10.1) 686 (2.6)

858 (46.5) 431 (23.3) 389 (21.1) 168 (9.1)

9216 (33.3) 8305 (30.0) 7003 (25.3) 3136 (11.3)

165 (8.8) 432 (23.1) 983 (52.6) 289 (15.5) 66 (3.6) 14 (0.8) 49 (2.6) 121 (6.7)

890 (3.2) 4020 (14.4) 15,066 (54.1) 7853 (28.2) 846 (3.1) 97 (0.4) 465 (1.7) 1733 (6.3)

b0.001 190 (42.3) 180 (40.1) 79 (17.6)

4548 (31.5) 6610 (45.8) 3283 (22.7)

126 (29.1) 201 (46.4) 74 (17.1) 32 (7.4)

2839 (20.3) 8182 (58.6) 2338 (16.7) 603 (4.3)

174 (38.8) 146 (32.6) 84 (18.8) 44 (9.8)

3542 (24.5) 5470 (37.8) 3645 (25.2) 1803 (12.5)

28 (6.2) 79 (17.6) 230 (51.2) 112 (24.9) 19 (4.2) 8 (1.8) 24 (5.3) 24 (5.3)

240 (1.7) 1288 (8.9) 7472 (51.6) 5486 (37.9) 466 (3.2) 81 (0.6) 328 (2.3) 849 (5.9)

b0.001

0.048

b0.001

b0.001

b0.001

b0.001

0.24 0.001 b0.001 0.63

b0.001 0.37 b0.001

0.001

b0.001

0.21 0.005 0.003 0.53

0.09 b0.001

b0.001

Abbreviations: SD = standard deviation. a Two-tailed Student's t-test for normally distributed variables and chi-square test for categorical variables. Boldface p-values are significant at α = 0.05.

0.30 0.005 0.012 0.69 b0.001 b0.001

226

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

Table 2 Metabolic syndrome characteristics of the patient population according to sex and depressive symptoms. p-Valuea

Total Depressed

Metabolic syndromeb, n (%) Waist circumference componentc, n (%) TG componentd, n (%) HDL componente, n (%) Hypertension componentf, n (%) Glucose componentg, n (%) Number of components, mean ± SD Waist circumference, mean ± SD, cm TG, median, IQR, mg/dL HDL, mean ± SD, mg/dL Systolic blood pressure, mean ± SD, mm Hg Diastolic blood pressure, mean ± SD, mm Hg Fasting glucose, median, IQR, mg/dL Medication Antilipidemic medication Antidiabetic medication Antihypertensive medication

Not depresse

n = 1900

n = 28,207

305 (16.1) 538 (28.3) 336 (17.7) 436 (22.9) 501 (26.4) 476 (25.1) 1.19 ± 1.22 82.1 ± 8.3 86.0, 57 57.1 ± 13.8 113.2 ± 15.8 72.2 ± 11.4 93.0, 13

4018 (14.2) 7515 (26.6) 5787 (20.5) 4565 (16.2) 8301 (29.4) 7659 (27.2) 1.18 ± 1.17 83.4 ± 8.0 89.0, 62 55.8 ± 13.2 114.9 ± 14.6 74.1 ± 11.3 93.0, 13

95 (5.0) 57 (3.0) 224 (11.8)

1767 (6.3) 742 (2.6) 3382 (12.0)

0.030 0.11 0.003 b0.001 0.005 0.046 0.61 b0.001 0.011 b0.001 b0.001 b0.001 0.008 0.027 0.33 0.79

p-Valuea

Men

p-Valuea

Women

Depressed

Not depressed

Depressed

Not depressed

n = 456

n = 14,617

n = 1444

n = 13,590

72 (15.8) 120 (26.3) 148 (32.5) 56 (12.3) 141 (30.9) 149 (32.7) 1.33 ± 1.12 85.8 ± 7.9 110.5, 79 51.5 ± 12.2 116.4 ± 13.9 76.3 ± 10.9 95.5, 12

2502 (17.1) 4290 (29.3) 4340 (29.7) 1694 (11.6) 5577 (38.2) 5113 (35.0) 1.41 ± 1.14 86.7 ± 6.7 107.0, 70 51.3 ± 11.2 118.2 ± 12.9 77.8 ± 10.4 96.0, 13

0.46 0.16 0.20 0.65 0.002 0.31 0.13 0.018 0.024 0.75 0.003 0.002 0.28

233 (16.1) 418 (28.9) 188 (13.0) 380 (26.3) 360 (24.9) 327 (22.6) 1.15 ± 1.24 81.0 ± 8.1 80.0, 50 58.9 ± 13.8 112.2 ± 16.2 71.0 ± 11.3 92.0, 13

1516 (11.2) 3225 (23.7) 1447 (10.6) 2871 (21.1) 2724 (20.0) 2546 (18.7) 0.93 ± 1.14 79.8 ± 7.6 74.0, 44 60.6 ± 13.5 111.3 ± 15.4 70.2 ± 11.1 91.0, 11

b0.001 b0.001 0.006 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.05 0.019 b0.001

41 (9.0) 26 (5.7) 61 (13.4)

1238 (8.5) 555 (3.8) 2280 (15.6)

0.69 0.037 0.197

54 (3.7) 31 (2.1) 163 (11.3)

529 (3.9) 187 (1.4) 1102 (8.1)

0.78 0.020 b0.001

Abbreviations: SD = standard deviation, TG = triglyceride, IQR = interquartile range, HDL = high-density lipoprotein cholesterol. a Two-tailed Student's t-test for normally distributed variables, Mann–Whitney U-test for skewed variables, and chi-square test for categorical variables. Boldface p-values are significant at α = 0.05. b National Cholesterol Education Program Adult Treatment Panel III (ATP III) criteria for metabolic syndrome were used with the modified waist circumference criteria from the Korean Society for the Study of Obesity. c Waist circumference N90 cm for men and N85 cm for women. d Serum TG ≥150 mg/dL or antilipidemic medication. e Serum HDL b40 mg/dL for men and b50 mg/dL for women. f Systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥85 mm Hg or antihypertensive medication. g Fasting plasma glucose ≥100 mg/dL or antidiabetic medication.

component (OR = 0.73, 95% CI = 0.58–0.91, p = 0.005) in men. Among women, the association between depression and metabolic syndrome and its HDL component was significant after additional adjustment for menopause and oral contraceptive/hormone treatment (OR = 1.41, 95% CI = 1.14–1.72, p = 0.001 and OR = 1.26, 95% CI = 1.07–1.47, p = 0.005, respectively) (Table 3).

The associations between depressive symptoms and the continuous components of metabolic syndrome were also individually analyzed. Adjustment for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease were performed for each sex. Depressive symptoms were negatively associated with waist circumference (β = −0.021, p = 0.015), systolic blood pressure

Table 3 Logistic regression analysis for the association between depressive symptoms and metabolic syndrome and each of its components. Metabolic syndrome OR (95% CI) Total Model 1 Model 2

Men Model 1 Model 2′

Women Model 1 Model 2′ Model 3

1.15 (1.01–1.31) 1.15 (0.99–1.34)

0.91 (0.70–1.17) 0.84 (0.63–1.11)

1.53 (1.32–1.78) 1.35 (1.11–1.64) 1.41 (1.14–1.72)

p-Valuea

0.030 0.08

0.46 0.21

b0.001 0.002 0.001

Waist circumference component OR (95% CI) 1.09 (0.98–1.21) 1.02 (0.91–1.16)

0.86 (0.70–1.06) 0.83 (0.66–1.04)

1.31 (1.16–1.48) 1.09 (0.93–1.27) 1.12 (0.95–1.31)

p-Valuea

0.11 0.71

0.16 0.10

b0.001 0.29 0.19

TG component

HDL component

Hypertension component

OR (95% CI)

p-Valuea

OR (95% CI)

p-Valuea

OR (95% CI)

0.83 (0.74–0.94) 1.15 (1.00–1.03)

0.003

1.54 (1.38–1.72) 1.19 (1.05–1.36)

b0.001

0.86 (0.77–0.95) 1.00 (0.87–1.13)

1.07 (0.80–1.42) 0.96 (0.71–1.30)

0.65

1.33 (1.18–1.51) 1.26 (1.09–1.46) 1.26 (1.07–1.47)

b0.001

1.14 (0.93–1.39) 1.11 (0.90–1.37)

1.26 (1.07–1.48) 1.16 (0.95–1.42) 1.15 (0.93–1.43)

0.06

0.20 0.35

0.006 0.14 0.19

0.008

0.79

0.002 0.005

Glucose component p-Valuea

0.005 0.94

0.73 (0.59–0.89) 0.73 (0.58–0.91)

0.002

1.33 (1.17–1.50) 1.16 (0.98–1.36) 1.14 (0.96–1.37)

b0.001

0.005

0.08 0.13

OR (95% CI) 0.90 (0.81–1.00) 1.06 (0.93–1.21)

p-Valuea

0.046 0.37

0.90 (0.74–1.10) 0.95 (0.76–1.18)

0.31

1.27 (1.11–1.48) 1.11 (0.95–1.31) 1.17 (0.98–1.39)

b0.001

0.63

0.19 0.08

Model 1: No adjustment. Model 2: Adjusted for sex, age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease. Model 2′: Adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease. Model 3: Adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, thyroid disease, menopause, and oral contraceptive/hormone treatment. Abbreviations: OR = odds ratio, CI = confidence interval, TG = triglyceride, HDL = high-density lipoprotein cholesterol. a Boldface p-values are significant at α = 0.05.

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

(β = −0.027, p = 0.001) and diastolic blood pressure (β = − 0.021, p = 0.014) in men, and they were positively associated with waist circumference (β = 0.019, p = 0.037) and log-transformed TG levels (β = 0.029, p = 0.003) and negatively associated with HDL levels (β = − 0.025, p = 0.011) in women (Table 4). Depressive symptoms was significantly associated with the number of metabolic syndrome components according to the linear regression analysis adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease in women (β = 0.029, p b 0.001), but no association was found in men (β = −0.014, p = 0.10). 3.3. Relationship between depressive symptom severity and metabolic syndrome In a separate analysis, we inserted the square-root-transformed BDI as a continuous independent variable and adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease. Depressive symptom severity was associated with metabolic syndrome in women (OR = 1.046, 95% CI = 1.002–1.091, p = 0.039). Even after additional adjustment for menopause and oral contraceptive/hormone treatment in women, the association remained statistically significant (OR = 1.049, 95% CI = 1.003–1.098, p = 0.038). However, no association was identified in men (OR = 1.01, 95% CI = 0.97–1.05, p = 0.72). 3.4. Sensitivity analysis We performed an analysis that excluded data from 3075 men and 1376 women who were receiving antidepressants, anxiolytics, or hypnotics or who had missing data for these variables to determine whether current psychotropic medication use influenced the results. This analysis yielded similar results to those reported above except that the association between depressive symptoms and metabolic syndrome in the total population was statistically significant after adjusting for sex, age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease (OR = 1.21, 95% CI = 1.02–1.43, p = 0.027). 4. Discussion After controlling for confounding factors, subjective depressive symptoms were associated with metabolic syndrome and with the number of metabolic syndrome components only in women. Women

227

with clinical depressive symptoms had an approximately 1.35-fold higher risk of metabolic syndrome than did women without clinical depressive symptoms, and we found a relationship between subjective depressive symptom severity and the risk of metabolic syndrome. Focusing on the individual components, the HDL component was associated with depressive symptoms after adjusting for confounding factors in women, and the hypertension component was inversely associated with depressive symptoms in men. Linear regression yielded the same results and also showed that TG levels were associated with depressive symptoms in women and that waist circumference was positively associated with depressive symptoms in women and negatively associated with depressive symptoms in men. Importantly, the finding that the association between depressive symptoms and metabolic syndrome was confined to women replicates some previous studies (Kinder, 2004; Laudisio et al., 2009; Pulkki-Raback et al., 2009; Toker et al., 2008). However there are studies which have shown that the association is confined to men (Nishina et al., 2011; Viinamaki et al., 2009) or studies that were only conducted in men (Almeida et al., 2009; Takeuchi et al., 2009) which reported the association. The use of different study designs and study populations may explain these discrepancies. Studies reporting that the association was confined to women (Kinder, 2004; Laudisio et al., 2009; Pulkki-Raback et al., 2009; Toker et al., 2008) have included a variety of age ranges, whereas most studies (Almeida et al., 2009; Nishina et al., 2011; Viinamaki et al., 2009) reporting the association in men included older participants and a narrower age range than did our study. Takeuchi et al. (2009), who reported the association only in men, used a sample with an average age of 42.7 years, which is similar to our population. However, that study sample was smaller than ours and was selected from among the full-time employees of a company, which may have resulted in selection biases. It is also possible that the discrepant findings are related to differences in ethnicity. Vogelzangs et al. (2007b) found that depression was associated with metabolic syndrome only in Whites and not in Blacks, although this conclusion remains debatable, as there is also evidence that this association does not differ by ethnicity (Kinder, 2004). In terms of Asian populations, only a few studies have investigated the association between depression and metabolic syndrome in Japanese samples, and these have reported that the association is confined to men (Nishina et al., 2011; Takeuchi et al., 2009), which is inconsistent with our results. As noted above, the difference between these studies and ours may be attributable to the use of different research designs. Additional research with large samples drawn from Asian populations is needed before drawing conclusions about these discrepant findings. It is also possible that these

Table 4 Linear regression analysis for the association between depressive symptoms and metabolic syndrome and each of its components. Waist circumference

Log TG

HDL

Systolic blood pressure

Diastolic blood pressure

Log glucose

β

p-Valuea

β

p-Valuea

β

p-Valuea

β

p-Valuea

β

p-Valuea

β

p-Valuea

Total Model 1 Model 2

−0.038 0.005

b0.001 0.33

−0.014 0.021

0.015 b0.001

0.025 −0.011

b0.001 0.058

−.0.028 −0.006

b0.001 0.29

−0.041 −0.004

b0.001 0.44

−0.013 0.008

0.029 0.20

Men Model 1 Model 2′

−0.022 −0.021

0.006 0.015

0.019 0.015

0.017 0.08

0.97 0.75

Women Model 1 Model 2′ Model 3

0.045 0.016 0.019

b0.001 0.049 0.037

0.048 0.029 0.029

b0.001 0.001 0.003

0.003 0.008 −0.036 −0.023 −0.025

0.75 0.34

−0.024 −0.027

0.003 0.001

−0.025 −0.021

0.002 0.014

b0.001 0.003

b0.001 0.011 0.011

0.017 b0.001 −0.006

0.042 0.97 0.53

0.019 0.004 −0.001

0.019 0.68 0.93

0.028 0.014 0.018

0.001 0.12 0.05

Model 1: No adjustment. Model 2: Adjusted for sex, age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease. Model 2′: Adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, and thyroid disease. Model 3: Adjusted for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, angina, stroke, thyroid disease, menopause, and oral contraceptive/hormone treatment. Abbreviations: OR = odds ratio, CI = confidence interval, TG = triglyceride, HDL = high-density lipoprotein cholesterol. a Boldface p-values are significant at α = 0.05.

228

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

differences are attributable to our use of a cross-sectional design. Viinamaki et al. (2009) found that the association between metabolic syndrome and depression was confined to men. However, theirs was a prospective study of patients with chronic depressive symptoms, which would have a different influence on metabolic parameters. Finally, depressive subtypes should be considered. Individuals with nonmelancholic depressive symptoms may be more likely than those with melancholic depressive symptoms to suffer from metabolic syndrome (Seppala et al., 2012), but most studies have used cutoff scores to classify depression and have not provided data on subtypes. In women, the HDL component was associated with depressive symptoms in both the linear and logistic regressions, whereas TG levels and waist circumference were associated with depressive symptoms only in the linear regression. This suggests that different definitions of each component and the different cutoff points for each component may explain the discrepancy between previous studies. Although both linear regression and logistic regression are reasonable approaches, the interpretation of the results of our linear regression analysis is limited, as the values of some components may have been directly affected by medication. Nevertheless, the HDL component had substantial evidence with the association between depression, followed by the TG component and the least with waist circumference, which are the three most common components that have shown association with depression in previous studies (van Reedt Dortland et al., 2010). The inverse relationship of depression with hypertension and waist circumference in men was also interesting. A few studies with large and broad-based samples such as ours have shown a negative association between hypertension and depression (Herva et al., 2006; Hildrum et al., 2009). As positive associations between depression and hypertension have also been reported, further investigation is needed. The association between waist circumference and depression in both men and women should be interpreted with caution, because logistic regression revealed no statistically significant associations in this regard. Nonetheless, the negative association with waist circumference is similar to that with hypertension. There are several reasons that these associations may be sex specific. At first, there are physiological hormone differences between men and women. Particularly in women menopause and female hormone medications are known to have effects on depression (Freeman et al., 2014; Kornstein et al., 2010) and on the risk of metabolic syndrome (Goyal et al., 2013; Salpeter et al., 2006). However, the association in women remained significant after controlling for these factors, so it cannot be explained by hormonal factors alone. Lifestyle behaviors should also be considered. Cigarette smoking, alcohol intake, and lack of exercise are known to influence depression (Berk et al., 2013), and the patterns of these behaviors differed between each sex in our study. However, we controlled for these lifestyle factors, which implies that the associations reported herein cannot be completely explained by them. Another explanation may relate to the use of self-reports to classify depression. Depression is reportedly more common in women (Cho et al., 2010), which we also found in our study (3.0% of men, 9.4% of women). However, it is possible that men under-reported the severity of their depressive symptoms, potentially resulting in a classification bias. Other explanations for sex differences may involve the mechanisms that have been proposed as underlying both depression and metabolic syndrome. Thus, depression is associated with dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis (Min et al., 2012), which is also implicated in the development of metabolic syndrome (Kazakou et al., 2012). Vogelzangs et al. (2007a) reported that urinary cortisol levels were associated with TG and HDL levels in patients with depression, implying that the HPA axis may be a link between metabolic syndrome and depression. It is known that women show greater activation of the HPA axis than do men when depressed (Young and Korszun, 2010) and that cortisol mediates the association of metabolic syndrome with depression in women (Muhtz et al., 2009), which may explain the sex differences in the association between the conditions.

Insulin resistance, one of the core pathologies of metabolic syndrome (Grundy et al., 2004), may constitute a link between depression and metabolic syndrome. Insulin resistance is known to be associated with dyslipidemia (Raghavan, 2012), which was significantly related to subjective depressive symptoms in our study. Akbaraly et al. (2013) recently reported that low insulin secretion appears to be a risk factor for depression in middle-aged women, and they proposed that insulin resistance may be among the reasons for sex differences of the association between depression and metabolic syndrome. Another potential common mechanism is inflammation, as previous studies have shown that it is associated with depression (Leonard, 2014; Lopresti et al., 2014). However, research examining the relationship between inflammation and depression by each sex has yielded inconsistent results and has been confined to White women (Morris et al., 2011) or to men (Vetter et al., 2013). Genetic factors (Kloiber et al., 2010), autonomous neurotransmission imbalance (Hamer and Malan, 2012), and leptin resistance (Chirinos et al., 2013) have also been proposed as mechanisms linking depression and metabolic syndrome. However, all the aforementioned mechanisms require further investigation and additional replications to clarify their role in sex differences in metabolic syndrome. It should also be noted that recent studies have shown that different depressive subtypes have different associations with these mechanisms. IL-6 was significantly elevated in melancholic but not in atypical depression (Dunjic-Kostic et al., 2013), implying that inflammatory markers differ among depression subtypes. However, another study found that melancholic depression was associated with HPA axis hyperactivity and that atypical depression was associated with inflammatory and metabolic dysregulation (Lamers et al., 2013). As different subtypes of depression exert different effects on these proposed mechanisms and the metabolic parameters under examination, differences in subtype may explain discrepancies in the results of studies in this domain. The association between the square-root transformation of the total BDI score and metabolic syndrome observed in women also suggests that depressive symptom severity may be important in evaluating the risk of metabolic syndrome. A few studies examining this trend have shown similar results (Skilton et al., 2007; Vaccarino et al., 2008). This highlights the importance of assessing not only the prevalence of depression but also its severity as the risk of metabolic syndrome increases. There are several limitations to the present study. First, it was a cross-sectional study; thus, causality cannot be determined. Depressive symptoms may be a risk factor for metabolic syndrome, or metabolic syndrome may be a risk factor for depressive symptoms (Akbaraly et al., 2011). However, a meta-analysis has increased the evidence that the association is bidirectional (Pan et al., 2012). Second, this study used self-reported scales and did not use a comprehensive psychiatric evaluation. The diagnosis of depression and comorbid psychiatric diseases could not be established. However, we used a cutoff score validated for the Korean population, and previous studies have shown that this approach has high sensitivity and specificity (Jo et al., 2007; Lee and Song, 1991; Shin et al., 1993) for diagnosing depression. Third, the patients in our study were drawn entirely from the Korean population, so ethnic differences could not be considered. Fourth, certain selection biases should be considered; for instance, obstacles may prevent people with low income from receiving these examinations. Despite these limitations, our study has major strengths. First, it was conducted in a large sample of the general population including a wide age range. It provided sufficient statistical power, and generalization to similar populations is reasonable. Second, the association was separately investigated in each sex, a variable that, when not considered, may have biased the results of previous studies. Our study showed no association between metabolic syndrome and depression in the total population after adjusting for confounders including sex; however, a significant association between depressive symptoms and metabolic syndrome was observed in women. Third, we analyzed not only metabolic syndrome but

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

also each of its components to determine whether any specific components were associated with depressive symptoms. The lipid components may play a critical role in the development of metabolic syndrome, particularly in women; this requires further investigation. 5. Conclusion In conclusion, our large-sample study showed that subjective depressive symptoms were associated with metabolic syndrome in women after controlling for confounding factors. This emphasizes, especially in women, the importance of screening for and diagnosing depression, along with performing laboratory evaluations, including lipid profiles, because metabolic syndrome can be comorbid. Further studies including longitudinal studies with multiethnic samples and evaluations of specific mechanisms (especially dyslipidemia) should consider sex differences. Potential conflict of interest Dr. Yong Min Ahn receives research grants and served as a lecturer for Janssen, Lilly, Lundbeck and Otsuka. All other authors report no financial relationships with commercial interests. Submission declaration and previous presentation We certify that the submission is an original work and is not under review at any other journal. Contributors All authors contributed to the analysis, interpretation of the results and to drafting the critical review of manuscript. All authors reviewed and approved the final version of the manuscript. Acknowledgments This work was supported by grant A121987 from the Korea Healthcare Technology R & D Project, Ministry of Health and Welfare, Republic of Korea. References Akbaraly TN, Ancelin ML, Jaussent I, Ritchie C, Barberger-Gateau P, Dufouil C, et al. Metabolic syndrome and onset of depressive symptoms in the elderly: findings from the three-city study. Diabetes Care 2011;34:904–9. Akbaraly TN, Kumari M, Head J, Ritchie K, Ancelin ML, Tabak AG, et al. Glycemia, insulin resistance, insulin secretion, and risk of depressive symptoms in middle age. Diabetes Care 2013;36:928–34. Almeida OP, Calver J, Jamrozik K, Hankey GJ, Flicker L. Obesity and metabolic syndrome increase the risk of incident depression in older men: the health in men study. Am J Geriatr Psychiatry 2009;17:889–98. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71. Beck AT, Steer RA, Garbin MG. Psychometric properties of the Beck Depression Inventory — 25 years of evaluation. Clin Psychol Rev 1988;8:77–100. Berk M, Sarris J, Coulson CE, Jacka FN. Lifestyle management of unipolar depression. Acta Psychiatr Scand Suppl 2013;443:38–54. Chirinos DA, Goldberg R, Gellman M, Mendez AJ, Gutt M, McCalla JR, et al. Leptin and its association with somatic depressive symptoms in patients with the metabolic syndrome. Ann Behav Med 2013;46:31–9. Cho MJ, Chang SM, Lee YM, Bae A, Ahn JH, Son J, et al. Prevalence of DSM-IV major mental disorders among Korean adults: a 2006 National Epidemiologic Survey (KECA-R). Asian J Psychiatry 2010;3:26–30. Dunjic-Kostic B, Ivkovic M, Radonjic NV, Petronijevic ND, Pantovic M, Damjanovic A, et al. Melancholic and atypical major depression—connection between cytokines, psychopathology and treatment. Prog Neuropsychopharmacol Biol Psychiatry 2013;43:1–6. Freeman EW, Sammel MD, Boorman DW, Zhang R. Longitudinal pattern of depressive symptoms around natural menopause. JAMA Psychiatry 2014;71:36–43. Gil K, Radziwittowicz P, Zdrojewski T, Pakalska-Korcala A, Chwojnicki K, Piwonski K, et al. Relationship between the prevalence of depressive symptoms and metabolic syndrome. Results of the SOPKARD Project. Kardiol Pol 2006;64:464–9. Goyal S, Baruah M, Devi R, Jain K. Study on relation of metabolic syndrome with menopause. Indian J Clin Biochem 2013;28:55–60.

229

Grundy SM, Brewer Jr HB, Cleeman JI, Smith Jr SC, Lenfant C, National Heart L, et al. Definition of metabolic syndrome: report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Arterioscler Thromb Vasc Biol 2004;24:e13–8. Hahn HM, Yun TH, Shin YW, Kim KH, Yoon DJ, Chung KJ. A standardization study of Beck Depression Inventory in Korea. J Korean Neuropsychiatr Assoc 1986;25:487–500. Hamer M, Malan L. Sympathetic nervous activity, depressive symptoms, and metabolic syndrome in black Africans: the sympathetic activity and ambulatory blood pressure in Africans study. Stress 2012;15:562–8. Herva A, Rasanen P, Miettunen J, Timonen M, Laksy K, Veijola J, et al. Co-occurrence of metabolic syndrome with depression and anxiety in young adults: the Northern Finland 1966 Birth Cohort Study. Psychosom Med 2006;68:213–6. Hildrum B, Mykletun A, Midthjell K, Ismail K, Dahl AA. No association of depression and anxiety with the metabolic syndrome: the Norwegian HUNT study. Acta Psychiatr Scand 2009;120:14–22. Jo SA, Park MH, Jo I, Ryu SH, Han C. Usefulness of Beck Depression Inventory (BDI) in the Korean elderly population. Int J Geriatr Psychiatry 2007;22:218–23. Kahl KG, Greggersen W, Schweiger U, Cordes J, Balijepalli C, Losch C, et al. Prevalence of the metabolic syndrome in unipolar major depression. Eur Arch Psychiatry Clin Neurosci 2012;262:313–20. Kazakou P, Kyriazopoulou V, Michalaki M, Ierodiakonou V, Psyrogiannis A, Habeos I. Activated hypothalamic pituitary adrenal axis in patients with metabolic syndrome. Horm Metab Res 2012;44:839–44. Kinder LS. Depression and the metabolic syndrome in young adults: findings from the Third National Health and Nutrition Examination Survey. Psychosom Med 2004;66: 316–22. Kloiber S, Kohli MA, Brueckl T, Ripke S, Ising M, Uhr M, et al. Variations in tryptophan hydroxylase 2 linked to decreased serotonergic activity are associated with elevated risk for metabolic syndrome in depression. Mol Psychiatry 2010;15:736–47. Kornstein SG, Young EA, Harvey AT, Wisniewski SR, Barkin JL, Thase ME, et al. The influence of menopause status and postmenopausal use of hormone therapy on presentation of major depression in women. Menopause 2010;17:828–39. Lamers F, Vogelzangs N, Merikangas KR, de Jonge P, Beekman AT, Penninx BW. Evidence for a differential role of HPA-axis function, inflammation and metabolic syndrome in melancholic versus atypical depression. Mol Psychiatry 2013;18:692–9. Laudisio A, Marzetti E, Pagano F, Pozzi G, Bernabei R, Zuccala G. Depressive symptoms and metabolic syndrome: selective association in older women. J Geriatr Psychiatry Neurol 2009;22:215–22. Lee YH, Song JY. A study of the reliability and the validity of the BDI, SDS and MMPI-D scales. Korean J Clin Psychol 1991;10:98–113. Leonard BE. Impact of inflammation on neurotransmitter changes in major depression: an insight into the action of antidepressants. Prog Neuropsychopharmacol Biol Psychiatry 2014;48:261–7. Lepine JP, Briley M. The increasing burden of depression. Neuropsychiatr Dis Treat 2011; 7:3–7. Lopresti AL, Maker GL, Hood SD, Drummond PD. A review of peripheral biomarkers in major depression: the potential of inflammatory and oxidative stress biomarkers. Prog Neuropsychopharmacol Biol Psychiatry 2014;48:102–11. Marijnissen RM, Smits JE, Schoevers RA, van den Brink RH, Holewijn S, Franke B, et al. Association between metabolic syndrome and depressive symptom profiles—sexspecific? J Affect Disord 2013;151:1138–42. Miettola J, Niskanen LK, Viinamaki H, Kumpusalo E. Metabolic syndrome is associated with self-perceived depression. Scand J Prim Health Care 2008;26:203–10. Min W, Liu C, Yang Y, Sun X, Zhang B, Xu L, et al. Alterations in hypothalamic–pituitary– adrenal/thyroid (HPA/HPT) axes correlated with the clinical manifestations of depression. Prog Neuropsychopharmacol Biol Psychiatry 2012;39:206–11. Morris AA, Zhao L, Ahmed Y, Stoyanova N, De Staercke C, Hooper WC, et al. Association between depression and inflammation—differences by race and sex: the METAHealth study. Psychosom Med 2011;73:462–8. Muhtz C, Zyriax BC, Klahn T, Windler E, Otte C. Depressive symptoms and metabolic risk: effects of cortisol and gender. Psychoneuroendocrinology 2009;34:1004–11. Mykletun A, Bjerkeset O, Dewey M, Prince M, Overland S, Stewart R. Anxiety, depression, and cause-specific mortality: the HUNT study. Psychosom Med 2007;69:323–31. Nishina M, Nishina K, Ohira T, Makino K, Iso H. Associations of psychological distress with metabolic syndrome among Japanese urban residents. J Atheroscler Thromb 2011; 18:396–402. Pan A, Keum N, Okereke OI, Sun Q, Kivimaki M, Rubin RR, et al. Bidirectional association between depression and metabolic syndrome: a systematic review and metaanalysis of epidemiological studies. Diabetes Care 2012;35:1171–80. Pulkki-Raback L, Elovainio M, Kivimaki M, Mattsson N, Raitakari OT, Puttonen S, et al. Depressive symptoms and the metabolic syndrome in childhood and adulthood: a prospective cohort study. Health Psychol 2009;28:108–16. Raghavan VA. Insulin resistance and atherosclerosis. Heart Fail Clin 2012;8:575–87. Salpeter SR, Walsh JM, Ormiston TM, Greyber E, Buckley NS, Salpeter EE. Meta-analysis: effect of hormone-replacement therapy on components of the metabolic syndrome in postmenopausal women. Diabetes Obes Metab 2006;8:538–54. Seppala J, Vanhala M, Kautiainen H, Eriksson J, Kampman O, Mantyselka P, et al. Prevalence of metabolic syndrome in subjects with melancholic and non-melancholic depressive symptoms. A Finnish population-based study. J Affect Disord 2012;136: 543–9. Shin MS, Kim ZS, Park KB. The cut-off score for the Korean version of Beck Depression Inventory. Korean J Clin Psychol 1993;12:618–23. Simon GE, VonKorff M. Reevaluation of secular trends in depression rates. Am J Epidemiol 1992;135:1411–22. Skilton MR, Moulin P, Terra JL, Bonnet F. Associations between anxiety, depression, and the metabolic syndrome. Biol Psychiatry 2007;62:1251–7.

230

S.J. Rhee et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 54 (2014) 223–230

Takeuchi T, Nakao M, Nomura K, Inoue M, Tsurugano S, Shinozaki Y, et al. Association of the metabolic syndrome with depression and anxiety in Japanese men: a 1-year cohort study. Diabetes Metab Res Rev 2009;25:762–7. Toker S, Shirom A, Melamed S. Depression and the metabolic syndrome: genderdependent associations. Depress Anxiety 2008;25:661–9. Vaccarino V, McClure C, Johnson BD, Sheps DS, Bittner V, Rutledge T, et al. Depression, the metabolic syndrome and cardiovascular risk. Psychosom Med 2008;70:40–8. van Reedt Dortland AK, Glitay EJ, van Veen T, Zitman FG, Penninx BW. Metabolic syndrome abnormalities are associated with severity of anxiety and depression and with tricyclic antidepressant use. Acta Psychiatr Scand 2010;122: 30–9. Vetter ML, Wadden TA, Vinnard C, Moore RH, Khan Z, Volger S, et al. Gender differences in the relationship between symptoms of depression and high-sensitivity CRP. Int J Obes (Lond) 2013;37(Suppl. 1):S38–43.

Viinamaki H, Heiskanen T, Lehto SM, Niskanen L, Koivumaa-Honkanen H, Tolmunen T, et al. Association of depressive symptoms and metabolic syndrome in men. Acta Psychiatr Scand 2009;120:23–9. Vogelzangs N, Suthers K, Ferrucci L, Simonsick EM, Ble A, Schrager M, et al. Hypercortisolemic depression is associated with the metabolic syndrome in late-life. Psychoneuroendocrinology 2007a;32:151–9. Vogelzangs N, Beekman AT, Kritchevsky SB, Newman AB, Pahor M, Yaffe K, et al. Psychosocial risk factors and the metabolic syndrome in elderly persons: findings from the Health, Aging and Body Composition study. J Gerontol A Biol Sci Med Sci 2007b;62: 563–9. Windle M, Windle RC. Recurrent depression, cardiovascular disease, and diabetes among middle-aged and older adult women. J Affect Disord 2013;150:895–902. Young E, Korszun A. Sex, trauma, stress hormones and depression. Mol Psychiatry 2010; 15:23–8.