The relation between obesity and depressed mood in a multi-ethnic population. The HELIUS study

  • Deborah Gibson-Smith
  • Mariska Bot
  • Marieke Snijder
  • Mary Nicolaou
  • Eske M. Derks
  • Karien Stronks
  • Ingeborg A. Brouwer
  • Marjolein Visser
  • Brenda W. J. H. Penninx
Open Access
Original Paper

Abstract

Purpose

To examine the association between obesity and depressed mood in a large multi-ethnic population and check for consistency in this association across six ethnic groups.

Methods

Data of 21,030 persons (18–70 years) were sourced from the HELIUS study. Cross-sectional relationships between obesity measures [body mass index (kg/m2) and waist circumference (cm)] and depressed mood (PHQ-9 score ≥ 10) were analysed. Consistency of associations was investigated across ethnic groups by interaction terms (ethnicity*obesity measures) in basic (age, sex, education) and fully (health behaviours and somatic health) adjusted models.

Results

Obesity was prevalent in all ethnic groups, but varied substantially. After sociodemographic adjustment, obesity measures were associated with increased odds of depressed mood but this was inconsistent across ethnic groups. Obesity (BMI ≥ 30 or highest waist circumference quartile) was strongly and significantly associated with depressed mood in the Dutch [Odds Ratio (OR) = 1.72; 95% Confidence intervals (CI) 1.24–2.40, and OR = 1.86; 95% CI 1.38–2.50], respectively, and African Surinamese (OR = 1.60; 95% CI 1.29–1.98 and OR = 1.59; 95% CI 1.27–2.00, respectively) but had a weaker, non-significant association in other ethnic groups (South-Asian Surinamese, Ghanaian, Moroccan, Turkish groups). Adjustment for health behaviours and somatic health had limited effect on this pattern.

Conclusion

Obesity was associated with a higher risk of depressed mood. However, ethnic differences were found: the obesity-depressed mood association was strong in the Dutch and African Surinamese populations, but not in other ethnic groups. Future studies should explore whether differential normative values or pathophysiology across ethnic groups explain why the obesity-depression association is inconsistent across ethnic groups.

Keywords

Depressed mood Obesity Overweight Ethnicity HELIUS study 

Introduction

Obesity is a growing public health problem [1] that has been associated with many short-term effects such as social stigmatisation [2] and joint pain [3], and longterm negative health outcomes such as diabetes [4], hypertension [5] and some cancers [6]. Cross-sectional reviews and meta-analyses have shown that obesity and depression are associated [7]. Furthermore, longitudinal studies analysing the temporal trends between obesity and the development of depression have demonstrated that a higher BMI, and in particular obesity (BMI ≥ 30 kg/m2), increases the risk of developing depression and vice versa [8]. Depression is an important public health concern [9] as depression is the leading cause of years lived with disability [10].

Several studies from the United States suggest that the association between obesity and depression may differ between ethnic groups [11, 12, 13, 14, 15, 16, 17]. Most studies observed a stronger association among White and Hispanic Americans as compared to Blacks [11, 12, 13, 14, 15, 16]. One longitudinal study, however, found that obesity had a stronger association with depression for African Americans, as opposed to White Americans [17]. In addition, the literature is not entirely unanimous with several studies finding no differences across ethnic groups [18, 19]. Moreover, existing literature is limited to the study of White Americans, Afro-Caribbean Americans and Hispanics, thereby omitting minority groups more commonly found in Europe. For Example in the Netherlands, apart from the ~ 90% Dutch and other Western origins, the largest ethnic groups comprise 2.3% Turkish, 2.3% Moroccan, 2.1% Surinamese, 0.9% Antilleans and 4.5% other [20].

Diverse mechanisms explaining ethnic differences in the causal link between obesity and depression have been proposed. First, differences may be due to the fact that obesity is more prevalent and, therefore, more normative in certain cultural groups, thus having a smaller impact on mental health. Contributing attributes here could be psychosocial factors such as greater body dissatisfaction or increased weight discrimination. Second, health behaviours, such as alcohol consumption, smoking behaviour and the level of physical activity, are directly related to both obesity [21] and depression [22] and could, therefore, underlie the relationship between the two. As alcohol consumption, smoking behaviour and levels of physical activity are different between ethnic groups [23, 24, 25] they could be ethnicity-differential contributing factors explaining the obesity-depression relationship. Moreover, certain diseases associated with obesity, such as diabetes and heart disease, which have been shown to influence depressive symptoms [26], are also more prevalent among certain ethnic groups [27].

We investigated the relationship between obesity and depressed mood and whether this association was different among six ethnic groups living in Amsterdam, a European city. Ethnic differences will be explored in different models with adjustment for sociodemographic factors, health behaviours and somatic health to explore the effect these variables have on the obesity/depression association among ethnic groups.

Methods

Study population

The HEalthy LIfe in an Urban Setting (HELIUS) study is a multi-ethnic cohort study conducted in Amsterdam, the Netherlands which has been described in detail elsewhere [28]. In brief, baseline data collection took place in 2011–2015 and included people aged 18–70 years from different ethnic origins. Participants of Dutch, Surinamese, Turkish, Moroccan and Ghanaian ethnic origin were randomly selected, by ethnic group, from the municipal register ensuring roughly equal numbers from each ethnic group. Socio-historical information on the ethnic minority groups included in this study can be found elsewhere [29]. Data were collected by questionnaire and a physical examination in which biological samples were also obtained. Participants unable to fill out questionnaires in Dutch were offered questionnaires in English and Turkish, or assistance from an ethnically matched, trained interviewer (all ethnic minority groups).

For the current study, cross-sectional baseline data were used, including 22,165 participants for whom questionnaire data as well as data from the physical examination were available. We excluded those of Javanese Surinamese (n = 233) or unknown Surinamese (n = 267) origin due to small numbers. We also excluded those with another/unknown ethnic origin (n = 48). Additionally, due to low numbers, those who were underweight (BMI < 18.5 kg/m2, n = 354), as well as those with missing BMI data (n = 22) or missing depression scores (n = 211) were also excluded, which resulted in an analytic sample of 21,030 participants (4477 Dutch, 2938 South-Asian Surinamese, 4035 African Surinamese, 2265 Ghanaian, 3520 Turkish, 3795 Moroccan origin).

Ethnicity

Ethnicity was defined according to the country of birth of the participant as well as that of his/her parents, which is currently the most widely accepted assessment of ethnicity in the Netherlands [30]. Specifically, a participant is considered as of non-Dutch ethnic origin if he/she fulfils either of the following criteria: (1) he or she was born abroad and has at least one parent born abroad (first generation); or (2) he or she was born in the Netherlands but both his/her parents were born abroad. There were no third generation ethnic minority individuals in the cohort. Of the Surinamese immigrants in the Netherlands, approximately 80% are either African or South-Asian origin. Surinamese subgroups were classified according to self-reported ethnic origin. Participants were considered as of Dutch origin if the both parents were born in the Netherlands.

Depressed mood

Depressed mood was assessed using the Patient Health Questionnaire-9 (PHQ-9), an instrument consisting of 9 items with 4 response options for each item [never (0), several days (1), more than half the days (2) and nearly every day (3)], giving a sum score range of 0–27, with higher scores indicating more depressive symptoms [31]. If one of the items was missing (n = 406, 1.7%), the mean score of the other eight items was used to replace the missing item. If more than one item was missing, the variable was considered missing. This questionnaire assesses depressive symptoms during the previous 2 weeks. Depressed mood was considered present when a person had a PHQ-9 score equal or greater than 10. This cut-off has a sensitivity and specificity of 88% of predicting major depressive disorder [31], is a commonly used cut-off [32], and has been shown to be consistent across American ethnic groups [33]. The PHQ-9 has been shown to measure the same concepts across all six ethnic groups included in this study, and there are no systematic differences in reporting depressive symptoms between the groups [34].

Anthropometric measurements

Obesity measures used were BMI and waist circumference, a measure of abdominal visceral fat, which is generally seen as a more pathogenic metabolic risk factor [35]. Body weight and body height were measured in duplicate by a trained research assistant in barefoot participants wearing light clothes only. Waist circumference was measured in duplicate using a tape measure at the level midway between the lowest rib margin and the iliac crest. BMI was calculated as weight in kilograms divided by height squared in meters (kg/m2). In addition to a continuous indicator, BMI categories were made according to the World Health Organization classification, 18.5 to < 25 kg/m2 (normal), 25 to < 30 kg/m2 (overweight), ≥ 30 kg/m2 (obese). Waist circumference (cm), continuous and as sex-specific quartiles, were used as a measure of abdominal obesity. As the original units were small and to make the results of the two continuous obesity measures more comparable, both values were standardised.

Covariates

Based on previous literature, adjustments were made for three different groups of covariates [22, 26, 36, 37]. Socio-demographic covariates included age, gender and level of education (the highest level of education completed with a diploma or certificate of proficiency in the Netherlands or in the country of origin). The second group of covariates were health behaviours which included physical activity (achieving the Dutch guideline for physical activity) measured with the SQUASH questionnaire [38], smoking (current, former, never) and alcohol use. Alcohol intake was reported as weekly or monthly frequency and typical number of consumptions per drinking day. This was subsequently converted into 3 categories non-drinker, low-moderate (0–14 drinks/week women, 0–21 drinks men), heavy (≥ 14 drinks/week women, ≥ 21 drinks/week men) assuming mode values for each frequency category. The third group of covariates was somatic health. Somatic health was derived from the number of self-reported diseases: presence of hypertension, diabetes, cardiovascular diseases (cardiovascular disease, cardiovascular accident, myocardial infarction), chronic lung disorders (asthma, chronic bronchitis, lung emphysema), osteoarthritis (arthrosis, rheumatoid arthritis) and cancer. Participants were asked if they had been diagnosed by a doctor with any of these diseases. Data were complete for all covariates with the exception of physical activity (26 missing) and waist circumference (16 missing) thus missing data was ignored.

Statistical analysis

Baseline socio-demographic characteristics, health behaviours, somatic and anthropometric data are presented according to ethnic origin. Initially the overall association between obesity measures (both continuous and categorical) and depressed mood was analysed adjusted for socio-demographic variables (age, gender, educational level). Normal BMI and waist circumference in the lowest quartile were used as reference categories. The analysis estimating the association of waist circumference with depressed mood also included adjustment for height (cm) to adjust for body size.

To examine ethnic differences in the association between obesity measures and depressed mood, interaction terms for the product of the dichotomised obesity measures [obese/waist circumference highest quartile (yes/no)] with ethnicity were added to the regression models. Three models of increasing complexity were made: the first model was adjusted for socio-demographics and ethnicity (model 1), the second added health behaviours (model 2) and the third added somatic health (model 3). The Dutch group was taken as a reference group, although we also ran models varying the reference groups to test for differences between all ethnic groups. Results of significant modifications of the obesity-depressed mood relationship by ethnicity, as identified by the overall interaction term, were stratified and the odds ratios for the individual ethnic groups were simultaneously displayed in a figure along with the overall odds ratio. Three separate models were made to illustrate whether the ethnic differences were present after adjustment for socio-demographic variables (model 1), after additional adjustment for health behaviours (model 2), and after additional adjustment for number of chronic diseases (model 3). We also performed a couple of post hoc analyses. Given that some studies have found that ethnic differences in the obesity-mood relationship are gender specific, we also tested for gender*ethnicity*obesity measures interaction. Additionally, we also examined whether length of residence modified the obesity-depression relationship of the non-Dutch residence by adding a obesity*length_of_stay variable to the models. Statistical significance was set at P < 0.05 (P < 0.1 for the testing of interaction terms) and all analyses were performed in SPSS version 22 (Inc., Chicago, Illinois, USA).

Results

Table 1 shows that the Dutch and African Surinamese population were moderate drinkers whilst the other ethnic groups were mostly non-drinkers. Smoking was particularly low among Ghanaians. Another notable difference was that depressed mood was more prevalent in all ethnic minorities compared to those of Dutch origin, with particularly high prevalence in people of Turkish, Moroccan and South-Asian Surinamese origin. The prevalence of overweight and particularly obesity was higher among all ethnic minority groups as compared with Dutch, with the highest obesity rates among those of Ghanaian and Turkish origin.

Table 1

Characteristics of the HELIUS study participants by ethnicity

Variables

Dutch (n = 4477)

South Asian Surinamese (n = 2938)

African Surinamese (n = 4035)

Ghanaian (n = 2265)

Turkish (n = 3520)

Moroccan (n = 3795)

Sex [n (%)], male

2074 (46)

1335 (45)

1567 (39)

880 (39)

1593 (45)

1482 (39)

Age (years), mean (SD)

46.3 (13.9)

45.8 (13.3)

48.1 (12.4)

44.7 (11.1)

40.5 (12.1)

40.6 (12.9)

1st generation (%)

4477 (100%)

2270 (77)

8380 (88)

2164 (96)

2481 (71)

2624 (69)

Residence duration (years)

N/A

33.1 (8.7)

32.1 (10.5)

18.3 (8.1)

28.8 (8.2)

29.0 (8.9)

Educational level

 No or elementary [n (%)]

146 (3)

419 (14)

223 (6)

643 (28)

1183 (31)

1103 (31)

 Lower vocational or lower secondary [n (%)]

636 (14)

990 (34)

1437 (36)

923 (40)

894(25)

693 (19)

 Intermediate vocational /higher/secondary [n (%)]

997 (23)

857 (29)

1449 (35)

561 (26)

997 (29)

1262 (33)

 Higher vocational or university [n (%)]

2698 (60)

672 (23)

926 (23)

138 (6)

526 (15)

675 (17)

Smoking status

 Never [n (%)]

1640 (37)

1699 (58)

1966 (49)

1974 (87)

1655 (47)

2802 (74)

 Former [n (%)]

1731 (38)

417 (14)

808 (19)

193 (8)

660 (18)

489 (13)

 Current [n (%)]

1106 (25)

822 (28)

1261 (32)

98 (4)

1205 (34)

504 (13)

Alcohol

 Non-drinker [n (%)]

397 (9)

1248 (43)

1248 (31)

1179 (52)

2708 (77)

3508 (92)

 Low-moderate [n (%)]

3637 (81)

1577(54)

2675(66)

1065(47)

775(22)

266(7)

 Heavy [n (%)]

443 (10)

81 (3)

112 (3)

21 (1)

37 (1)

21 (1)

Physical activity [n (% achieving guideline)]a,b

3383 (76)

1572 (53)

2484 (61)

1212 (53)

1483 (42)

1786 (47)

Chronic diseases [n (% with n > 1)]

291 (6.5)

599 (20.4)

691 (17.1)

277 (12.2)

468 (13.3)

374 (9.9)

Chronic diseases, median (IQR)

0(0–0)

0 (0–1)

0 (0–1)

0 (0–1)

0 (0–1)

0 (0–1)

Height (cm), mean (SD)

175.5 (9.5)

164.5 (9.3)

168.6 (4.6)

165.5 (7.8)

165.3 (9.4)

166.4 (9.2)

BMI (kg/m2), mean (SD)

24.8 (4.2)

26.6 (4.6)

28.0 (5.4)

28.6 (5.0)

28.7 (5.6)

27.7 (5.1)

BMI categories

 Normal (18.5 to < 25 kg/m2) [n (%)]

2657 (59)

1231 (42)

1308 (33)

557 (25)

949 (27)

1242 (33)

 Overweight (25 to < 30 kg/m2) [n (%)]

1359 (31)

1135 (39)

1506 (37)

930 (41)

1327 (38)

1435 (38)

 Obese (≥ 30 kg/m2) [n (%)]

461 (10)

572 (19)

1221 (30)

778 (34)

1244 (35)

1118 (29)

Waist circumference (cm)

 Mean (SD)

89.2 (11.9)

91.2 (11.6)

92.4 (12.7)

91.9 (11.6)

93.9 (13.4)

92.8 (13.2)

 Quartile 1c

1461 (32)

642 (22)

921 (23)

444 (20)

705 (20)

818 (22)

 Quartile 2

1274 (29)

805 (27)

1010 (25)

565 (25)

797 (23)

881 (23)

 Quartile 3

935 (21)

838 (29)

1003 (25)

657 (29)

916 (26)

1008 (26)

 Quartile 4

805 (18)

651 (22)

1095 (27)

597 (26)

1099 (31)

1087 (29)

Depressed mood score (PHQ-9) median (IQR)

3.0 (1.0–5.0)

4.0 (1.0–8.0)

3.0 (0.0–5.0)

2.0 (0.0–5.0)

5.0 (2.0–9.0)

4.0 (2.0–8.0)

Depressed mood, yes [n (%)]2

316 (7)

548 (19)

429 (11)

209 (9)

820 (23)

793 (21)

SD standard deviation, BMI body mass index, IQR interquartile range

Low = less than 1 drink a week, moderate = men: 1–21 drinks/week; women: 1–14/week, heavy = men: >21/week; women: >14/week

aAchieved the Dutch national guidelines of ≥ 5 days a week 30 mins moderate-intensive activity

PHQ-9 sumscore ≥10

cQuartile 1 (<86.0cm)m (<80.1cm)f, Quartile 2 (86.0–<93.4cm)m(80.1–<89.65cm)f, Quartile 3 (93.4–<101.5cm)m(89.7–<100.3cm)f, Quartile 4 (>101.5cm)m(>100.3cm)f; m males, f females

Logistic regression analysis showed that both higher BMI and higher waist circumference were significantly associated with depressed mood among the total sample, after adjustment for age, sex and educational level [Odds Ratio (OR) for 1 SD higher BMI = 1.16; 95% confidence interval (CI) 1.12–1.21 and for 1 SD higher waist circumference: OR = 1.20; 95% CI 1.15–1.25] (supplementary Table 1). Dividing the obesity measures into categories showed that it was participants in the highest categories that had the greatest odds of depressed mood. Thus, participants with obesity or having a waist circumference in the highest quartile had greater odds of having depressed mood than those not in the extreme weight categories (OR = 1.43; 95% CI 1.29–1.54, OR = 1.56; 95% CI 1.38–1.77, respectively).

The relationship between obesity measures and depressed mood was significantly different across ethnic groups, as evidenced by significant overall p-values for the interaction terms ethnicity *obesity measures (supplementary Table 2), suggesting that the association between obesity measures and depressed mood is not consistent across ethnic groups. These interactions were significant for both obesity and waist circumference in the model adjusted for socio-demographic factors where the p values for the overall interaction terms were 0.03 and 0.05 for obesity and a high waist circumference, respectively. The effect modification of ethnicity on obesity measures remained significant with the addition of health behaviours [p value = 0.07 (obesity) 0.04 (waist circumference)] and persisted after the addition of somatic health for having a high waist circumference (p value = 0.06). However, the significance was diminished for the interaction between obesity*ethnicity after adjustment for somatic health (p value = 0.21).

To illustrate the ethnicity interaction, Fig. 1 shows the odds ratios for the relationship between obesity measures and depressed mood, stratified by ethnicity. Model 1 showed that Dutch and African Surinamese had a significant, positive association between obesity and depressed mood (OR = 1.72; 95% CI 1.24–2.40, OR = 1.60; 95% CI 1.29–1.98, respectively), whilst obesity in South-Asian Surinamese, Ghanaian, Moroccans and Turks showed a weaker non-significant association with depressed mood. Waist circumference quartiles showed a similar pattern with slightly larger differences between the ethnic groups. For those of Dutch and African Surinamese origins a high waist circumference had a positive association with depressed mood (OR = 1.86; 95% CI 1.38–2.50, OR = 1.59; 95% CI 1.27–2.00, respectively) whilst in South-Asian Surinamese, Ghanaian, Moroccans and Turks having a high waist circumference had a weaker non-significant association with depressed mood. This pattern of associations between ethnic groups remained the same after adjustment for health behaviour and somatic health and if anything, were slightly more pronounce for having a waist circumference in the highest quartile after adjustment for both health behaviours and somatic health.

Fig. 1

The association between obesity (diamonds) and waist circumference (triangles) with depressive symptoms (PHQ-9 score ≥ 10) by ethnicity, adjusted for covariates in different models. aSignificantly different (P < 0.05) from Dutch, bSouth-Asian Surinamese, cAfrican Surinamese, dTurkish, eMoroccan. ASignificantly different (P < 0.01) from Dutch, BSouth-Asian Surinamese, EMoroccan. Models 1:Socio-demographic: age, sex and education, 2:Socio-demographic plus lifestyle: smoking, alcohol use and physical activity, 3:Socio-demographic

Three way interaction terms showed that no gender*ethnic differences were present in the obesity–depression relationship [for both obese vs non-obese and waist circumference in quartile 4 (y/n)], both before and after adjustment for health behaviours and somatic health. Additionally, the length of residence in Amsterdam did not modify the relationship between obesity and depression.

Discussion

This study aimed to investigate whether the cross-sectional relationship between BMI or waist circumference (continuous and in categories) and the presence of depressed mood is consistent across six ethnic groups, i.e. Dutch, South-Asian Surinamese, African Surinamese, Turkish, Moroccan and Ghanaian, living in the same European city (Amsterdam, The Netherlands). Our results indicate that the association between being obese (BMI ≥ 30 kg/m2) and depressed mood tend to differ across ethnic groups: this association was stronger and statistically significant among Dutch and African Surinamese but weaker among those of Ghanaian, South-Asian Surinamese, Turkish or Moroccan origin. This pattern remained after additional adjustment for health behaviours and somatic health, illustrating that these factors are not largely involved in explaining the ethnic differences in the association between obesity and depressed mood.

Our study is one of the first to compare the relationship of BMI/waist circumference with depressed mood among different ethnic groups in Europe. The finding that obesity measures do not have a unanimous association with depressed mood across ethnic groups has been found in some other studies in the US [11, 12, 13, 14, 15, 16, 17]. Our results are partially in accordance with the previously performed American studies in that these studies also found that obesity increased the risk of depressed mood in the (non-Hispanic) White population but not the other ethnic groups [Mexican, Hispanic or Black (African American)] [11, 13, 15]. However, in our study, we found that African Surinamese with obesity were also relatively more likely to suffer from depressed mood.

Not all earlier studies found that ethnicity modifies the relationship between BMI/obesity and depressed mood. Remigio-Baker et al. [19] examined the cross-sectional relationship between visceral fat measured by computed tomography and elevated depressed symptoms in 1017 men and women aged 47–84. They found no significant interaction between elevated depressive symptoms and race/ethnicity (White, Black, Chinese and Hispanic) for visceral fat; however, they attributed this to lack of statistical power. Another study by Dong et al. [18] found that extreme obesity was consistently associated with an increased risk for depression across two racial groups (African American vs. European American). The population of this study differed considerably from ours due to the large proportion of extremely obese (BMI > 40 kg/m2) individuals (44% across racial groups), whilst only 2.2% of our participants had a BMI > 40 kg/m2.

Considering that the observed ethnic differences in the association between obesity and depression remained after adjustment for health behaviours and, in the case of central adiposity, somatic health, it is unlikely that these are strong mechanisms. Additional explanations for the differences across ethnic groups could be based on sociocultural differences. Higher prevalence of obesity among certain ethnic groups may lead to protective social norms. Among cultures where a “larger body size” is the norm, obesity is socially more acceptable [39] resulting in less body dissatisfaction, less mental stress [40, 41] and a buffering from weight discrimination. All of these can lead to a higher self-esteem which itself is a protective factor for depressed mood [42, 43, 44]. In America, perceived weight discrimination has been found to be greater in White Americans, where the relationship between obesity and depression is stronger than in African Americans [45]. Among our population sample, obesity was most prevalent among the Turkish, Moroccan and Ghanaian populations in whom, along with the South-Asian Surinamese, the association between obesity and depression is weakest. However, a study of Turkish and Moroccan residence of Amsterdam has shown that this population does not express a preference for a larger body size [46]. Hence, body satisfaction is unlikely to explain the weak association between obesity and depression for the Turkish and Moroccan population, although social acceptance of a larger body size may be a contributing factor, especially considering that a large proportion of the overweight Turkish and Moroccan men perceived their weight to be average. A study by Hoeninka et al. in the HELIUS population found that African Surinamese and Ghanaians experienced body dissatisfaction at a higher BMI than did Dutch origin populations [47]. Hence, for the Dutch population, this is consistent with our hypothesis that body dissatisfaction and, therefore, also depression, is greater in ethnic groups where the prevalence of obesity is lower. However, this does not explain the strong obesity-depression relationship among African Surinamese, thus this theory does not hold for this group. Another possibility is that how strongly being obese is related to depression may be dependent on how long the non-Dutch participants have resided in the Netherlands. The assumption being that the longer the duration of residence, the more assimilated a person is to the Dutch culture and, therefore, the more similar they would be in terms of their relationship between their obesity status and depression. However, given that the length of residence is similar between all ethnic groups (except for the Ghanaians), and in particular the fact that both Surinamese ethnic groups migrated at the same time, this could not explain the differences between groups. Furthermore, there was no interaction between length of stay and obesity. Alternatively, in addition to ethnic normative differences, another explanation for ethnic differences in the obesity-depression association could be an ethnic-differential pathophysiology. Several biological mechanisms, ranging from systemic inflammation, leptin resistance, metabolic syndrome disturbances, dysregulated hypothalamus–pituitary–adrenal axis (HPA-axis) have been proposed to be linking mechanisms between depression and obesity as these all occur in both conditions [48]. There are indications that occurrence of these biological mechanisms are ethnicity dependent [49]. We should also take into account the fact that the relationship between obesity and depression is bidirectional [8] and, therefore, it could be that when suffering from a depressed mood some ethnic groups gain weight whilst other remain the same or perhaps even loose weight.

This study is the first to examine the association of obesity with depressed mood among a multi-ethnic European sample. The strength of this study is its large sample size comprising of roughly equal numbers of different ethnic groups, including a variety of migrant groups found in Europe along with a large number of covariates. Additionally, we had multiple measures of obesity. The inclusion of waist circumference is important as waist circumference is generally considered a more pathogenic metabolic risk factor.

There are a few limitations. First, this study was cross-sectional and as such no causal pathways could be investigated. However, longitudinal studies have indicated that having obesity is a risk factor for developing depression but it is generally accepted that the relationship is bidirectional [8]. Second, the PHQ-9 is not equivalent to a diagnosis of clinical depression and only captures depressed mood, thus these results are not generalizable to those with clinical depression. Finally, adjustment for health behaviours excluded nutritional intake, and the measurement of physical activity was subjective, therefore, leaving potential for residual confounding.

We found that obesity measures were positively associated with depressed mood but only in certain ethnic groups. The association was stronger and statistically significant in the Dutch and African Surinamese whereas the association was weaker and not statistically significant in South-Asian Surinamese, Ghanaian, Moroccan and Turkish origin groups. The pattern among ethnic groups remained the same even after adjusting for differences in health behaviours and somatic health and became even more pronounced for high waist circumference compared to having obesity. Knowing that the relationship between obesity and depressed mood is not universal among ethnic groups, may help target prevention strategies with the knowledge that for some ethnic groups, programmes aimed at targeting obesity may result in an improvement in both somatic and mental health and whilst in other groups the improvement in physical health would be the main focus. Future studies should explore whether differential social-cultural based normative values or underlying pathophysiology across ethnic groups explain why obesity and depression are strongly related in some but not all ethnic groups.

Notes

Acknowledgements

The HELIUS study is conducted by the Academic Medical Center Amsterdam and the Public Health Service of Amsterdam. Both organisations provided core support for HELIUS. The HELIUS study is also funded by the Dutch Heart Foundation, the Netherlands Organization for Health Research and Development (ZonMw), the European Union (FP-7), and the European Fund for the Integration of non-EU immigrants (EIF). We gratefully acknowledge the AMC Biobank for their support in biobank management and high-quality storage of collected samples. We are most grateful to the participants of the HELIUS study and the management team, research nurses, interviewers, research assistants and other staff who have taken part in gathering the data of this study. DGS, MB, MV, IB and BP were supported by the EU-MoodFood grant. Funding for this research is provided by EU FP7 MooDFOOD Project ‘Multi-country cOllaborative project on the rOle of Diet, FOod-related behaviour, and Obesity in the prevention of Depression’, Grant agreement no. 613598. The study reported here was additionally supported by an additional grant from the Amsterdam Public Health research institute.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Integrity of research and reporting

The HELIUS study is conducted in accordance with the Declaration of Helsinki and has been approved by the Amsterdam Medical Center Ethical Review Board. All participants provided written informed consent.

Supplementary material

127_2018_1512_MOESM1_ESM.docx (20 kb)
Supplementary material 1 (DOCX 20 KB)

References

  1. 1.
    Wyatt SB, Winters KP, Dubbert PM (2006) Overweight and obesity: prevalence, consequences, and causes of a growing public health problem. Am J Med Sci 331:166–174CrossRefPubMedGoogle Scholar
  2. 2.
    Puhl RM, Heuer CA (2009) The stigma of obesity: a review and update. Obesity 17:941–964.  https://doi.org/10.1038/oby.2008.636 CrossRefPubMedGoogle Scholar
  3. 3.
    Hartz AJ, Fischer ME, Bril G et al (1986) The association of obesity with joint pain and osteoarthritis in the HANES data. J Chronic Dis 39:311–319CrossRefPubMedGoogle Scholar
  4. 4.
    Tyrovolas S, Koyanagi A, Garin N et al (2015) Diabetes mellitus and its association with central obesity and disability among older adults: a global perspective. Exp Gerontol 64:70–77.  https://doi.org/10.1016/j.exger.2015.02.010 CrossRefPubMedGoogle Scholar
  5. 5.
    Deng W-W, Wang J, Liu M-M et al (2013) Body mass index compared with abdominal obesity indicators in relation to prehypertension and hypertension in adults: the CHPSNE study. Am J Hypertens 26:58–67.  https://doi.org/10.1093/ajh/hps001 CrossRefPubMedGoogle Scholar
  6. 6.
    Calle EE, Kaaks R (2004) Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer 4:579–591.  https://doi.org/10.1038/nrc1408 CrossRefPubMedGoogle Scholar
  7. 7.
    de Wit L, Luppino F, van Straten A et al (2010) Depression and obesity: a meta-analysis of community-based studies. Psychiatry Res 178:230–235.  https://doi.org/10.1016/j.psychres.2009.04.015 CrossRefPubMedGoogle Scholar
  8. 8.
    Luppino FS, de Wit LM, Bouvy PF et al (2010) Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry 67:220–229.  https://doi.org/10.1001/archgenpsychiatry.2010.2 CrossRefPubMedGoogle Scholar
  9. 9.
    Bloom DE, Cafiero E, Jané-Llopis E et al (2011) The global economic burden of noncommunicable diseases. World Econ Forum 1–46Google Scholar
  10. 10.
    Ferrari AJ, Charlson FJ, Norman RE et al (2013) Burden of depressive disorders by country, sex, age, and year: findings from the Global Burden of Disease Study 2010. PLoS Med 10(11):e1001547.  https://doi.org/10.1371/journal.pmed.1001547 CrossRefPubMedPubMedCentralGoogle Scholar
  11. 11.
    Anderson SE, Murray DM, Johnson CC et al (2011) Obesity and depressed mood associations differ by race/ethnicity in adolescent girls. Int J Pediatr Obes 6:69–78.  https://doi.org/10.3109/17477161003728477 CrossRefPubMedGoogle Scholar
  12. 12.
    BeLue R, Francis LA, Colaco B (2009) Mental health problems and overweight in a nationally representative sample of adolescents: effects of race and ethnicity. Pediatrics 123:697–702.  https://doi.org/10.1542/peds.2008-0687 CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    Gavin AR, Rue T, Takeuchi D (2010) Racial/ethnic differences in the association between obesity and major depressive disorder: findings from the comprehensive psychiatric epidemiology surveys. Public Health Rep 125:698–708CrossRefPubMedPubMedCentralGoogle Scholar
  14. 14.
    Heo M, Pietrobelli A, Fontaine KR et al (2006) Depressive mood and obesity in US adults: comparison and moderation by sex, age, and race. Int J Obes (Lond) 30:513–519.  https://doi.org/10.1038/sj.ijo.0803122 CrossRefGoogle Scholar
  15. 15.
    Hicken MT, Lee H, Mezuk B et al (2013) Racial and ethnic differences in the association between obesity and depression in women. J women’s Heal 22:445–452.  https://doi.org/10.1089/jwh.2012.4111 CrossRefGoogle Scholar
  16. 16.
    Pratt LA, Brody DJ (2014) Depression and obesity in the US Adult household population, 2005–2010. NCHS Data Brief 167:1–8Google Scholar
  17. 17.
    Sachs-Ericsson N, Burns AB, Gordon KH et al (2007) Body mass index and depressive symptoms in older adults: the moderating roles of race, sex, and socioeconomic status. Am J Geriatr Psychiatry 15:815–825.  https://doi.org/10.1097/JGP.0b013e3180a725d6 CrossRefPubMedGoogle Scholar
  18. 18.
    Dong C, Sanchez LE, Price RA (2004) Relationship of obesity to depression: a family-based study. Int J Obes 28:790–795.  https://doi.org/10.1038/sj.ijo.0802626 CrossRefGoogle Scholar
  19. 19.
    Remigio-Baker RA, Allison MA, Schreiner PJ et al (2014) Difference by sex but not by race/ethnicity in the visceral adipose tissue-depressive symptoms association: the Multi-Ethnic Study of Atherosclerosis. Psychoneuroendocrinology 47:78–87.  https://doi.org/10.1016/j.psyneuen.2014.05.004 CrossRefPubMedPubMedCentralGoogle Scholar
  20. 20.
    Centraal Bureau voor Statistiek (2017) Bevolking; kerncijfersGoogle Scholar
  21. 21.
    Galani C, Schneider H (2007) Prevention and treatment of obesity with lifestyle interventions: review and meta-analysis. Int J Public Health 52:348–359CrossRefPubMedGoogle Scholar
  22. 22.
    Cassidy K, Kotynia-English R, Acres J et al (2004) Association between lifestyle factors and mental health measures among community-dwelling older women. Aust N Z J Psychiatry 38:940–947.  https://doi.org/10.1080/j.1440-1614.2004.01485.x CrossRefPubMedGoogle Scholar
  23. 23.
    de Munter JSL, Agyemang C, van Valkengoed IGM et al (2013) Cross national study of leisure-time physical activity in Dutch and English populations with ethnic group comparisons. Eur J Public Health 23:440–446.  https://doi.org/10.1093/eurpub/cks088 CrossRefPubMedGoogle Scholar
  24. 24.
    Brathwaite R, Smeeth L, Addo J et al (2017) Ethnic differences in current smoking and former smoking in the Netherlands and the contribution of socioeconomic factors: a cross-sectional analysis of the HELIUS study. BMJ Open 7:e016041.  https://doi.org/10.1136/bmjopen-2017-016041 CrossRefPubMedPubMedCentralGoogle Scholar
  25. 25.
    Sharma S, Malarcher AM, Giles WH, Myers G (2004) Racial, ethnic and socioeconomic disparities in the clustering of cardiovascular disease risk factors. Ethn Dis 14:43–48PubMedGoogle Scholar
  26. 26.
    Ryu E, Chamberlain AM, Pendegraft RS et al (2016) Quantifying the impact of chronic conditions on a diagnosis of major depressive disorder in adults: a cohort study using linked electronic medical records. BMC Psychiatry 16:114.  https://doi.org/10.1186/s12888-016-0821-x CrossRefPubMedPubMedCentralGoogle Scholar
  27. 27.
    Dijkshoorn H, Uitenbroek DG, Middelkoop BJC (2003) Prevalence of diabetes mellitus and cardiovascular disease among immigrants from Turkey and Morocco and the indigenous Dutch population. Ned Tijdschr Geneeskd 147:1362–1366PubMedGoogle Scholar
  28. 28.
    Stronks K, Snijder MB, Peters RJG et al (2013) Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health 13:402.  https://doi.org/10.1186/1471-2458-13-402 CrossRefPubMedPubMedCentralGoogle Scholar
  29. 29.
    Ikram UZ, Snijder MB, de Wit MAS et al (2016) Perceived ethnic discrimination and depressive symptoms: the buffering effects of ethnic identity, religion and ethnic social network. Soc Psychiatry Psychiatr Epidemiol 51:679–688.  https://doi.org/10.1007/s00127-016-1186-7 CrossRefPubMedPubMedCentralGoogle Scholar
  30. 30.
    Stronks K, Kulu-Glasgow I, Agyemang C (2009) The utility of “country of birth” for the classification of ethnic groups in health research: the Dutch experience. Ethn Health 14:255–269.  https://doi.org/10.1080/13557850802509206 CrossRefPubMedGoogle Scholar
  31. 31.
    Kroenke K, Spitzer RL, Williams JBW, Löwe B (2010) The patient health questionnaire somatic, anxiety, and depressive symptom scales: a systematic review. Gen Hosp Psychiatry 32:345–359.  https://doi.org/10.1016/j.genhosppsych.2010.03.006 CrossRefPubMedGoogle Scholar
  32. 32.
    Moriarty AS, Gilbody S, McMillan D, Manea L (2015) Screening and case finding for major depressive disorder using the patient health questionnaire (PHQ-9): a meta-analysis. Gen Hosp Psychiatry 37:567–576.  https://doi.org/10.1016/j.genhosppsych.2015.06.012 CrossRefPubMedGoogle Scholar
  33. 33.
    Huang FY, Chung H, Kroenke K et al (2006) Using the patient health questionnaire-9 to measure depression among racially and ethnically diverse primary care patients. J Gen Intern Med 21:547–552.  https://doi.org/10.1111/j.1525-1497.2006.00409.x CrossRefPubMedPubMedCentralGoogle Scholar
  34. 34.
    Galenkamp H, Stronks K, Snijder MB, Derks EM (2017) Measurement invariance of the PHQ-9 across six ethnic groups: the HELIUS study. BMC Psychiatry 17:349.  https://doi.org/10.1186/s12888-017-1506-9 CrossRefPubMedPubMedCentralGoogle Scholar
  35. 35.
    Fox CS, Massaro JM, Hoffmann U et al (2007) Abdominal visceral and subcutaneous adipose tissue compartments: Association With Metabolic Risk Factors in the Framingham Heart Study. Circulation 116:39–48.  https://doi.org/10.1161/CIRCULATIONAHA.106.675355 CrossRefPubMedGoogle Scholar
  36. 36.
    Schuch JJJ, Roest AM, Nolen WA et al (2014) Gender differences in major depressive disorder: results from the Netherlands study of depression and anxiety. J Affect Disord 156:156–163.  https://doi.org/10.1016/j.jad.2013.12.011 CrossRefPubMedGoogle Scholar
  37. 37.
    Nunes SOV, Vargas HO, Prado E et al (2013) The shared role of oxidative stress and inflammation in major depressive disorder and nicotine dependence. Neurosci Biobehav Rev 37:1336–1345.  https://doi.org/10.1016/j.neubiorev.2013.04.014 CrossRefPubMedGoogle Scholar
  38. 38.
    Wendel-Vos GCW, Schuit AJ, Saris WHM, Kromhout D (2003) Reproducibility and relative validity of the short questionnaire to assess health-enhancing physical activity. J Clin Epidemiol 56:1163–1169CrossRefPubMedGoogle Scholar
  39. 39.
    Wilson DB, Sargent R, Dias J (1994) Racial differences in selection of ideal body size by adolescent females. Obes Res 2:38–43.  https://doi.org/10.1002/j.1550-8528.1994.tb00042.x CrossRefPubMedGoogle Scholar
  40. 40.
    Grabe S, Hyde JS (2006) Ethnicity and body dissatisfaction among women in the United States: a meta-analysis. Psychol Bull 132:622–640.  https://doi.org/10.1037/0033-2909.132.4.622 CrossRefPubMedGoogle Scholar
  41. 41.
    Ross CE (1994) Overweight and depression. J Health Soc Behav 35:63–79CrossRefPubMedGoogle Scholar
  42. 42.
    Roberts RE, Duong HT (2013) Perceived weight, not obesity, increases risk for major depression among adolescents. J Psychiatr Res 47:1110–1117.  https://doi.org/10.1016/j.jpsychires.2013.03.019 CrossRefPubMedPubMedCentralGoogle Scholar
  43. 43.
    Robinson E, Sutin A, Daly M (2016) Perceived weight discrimination mediates the prospective relation between obesity and depressive symptoms in U.S. and U.K. adults. Health Psychol.  https://doi.org/10.1037/hea0000426 Google Scholar
  44. 44.
    Jackson SE, Beeken RJ, Wardle J (2015) Obesity, perceived weight discrimination, and psychological well-being in older adults in England. Obesity (Silver Spring) 23:1105–1111.  https://doi.org/10.1002/oby.21052 CrossRefGoogle Scholar
  45. 45.
    Dutton GR, Lewis TT, Durant N et al (2014) Perceived weight discrimination in the CARDIA study: differences by race, sex, and weight status. Obesity (Silver Spring) 22:530–536.  https://doi.org/10.1002/oby.20438 CrossRefGoogle Scholar
  46. 46.
    Nicolaou M, Doak C, Van Dam R et al (2008) Body size preference and body weight perception among two migrant groups of non-Western origin. Public Health Nutr 11:1332–1341.  https://doi.org/10.1017/S1368980008002966 CrossRefPubMedGoogle Scholar
  47. 47.
    Hoeninka JC, Beune EJ, Hartman MA et al The Body Size Ideals and Body Size Satisfaction of Dutch Residents and African-origin Residents Living in Amsterdam. The HELIUS Study. (unpublished) Google Scholar
  48. 48.
    Hryhorczuk C, Sharma S, Fulton SE (2013) Metabolic disturbances connecting obesity and depression. Front Neurosci 7:1–14.  https://doi.org/10.3389/fnins.2013.00177 CrossRefGoogle Scholar
  49. 49.
    Case SM, Stewart JC (2014) Race/ethnicity moderates the relationship between depressive symptom severity and C-reactive protein: 2005–2010 NHANES data. Brain Behav Immun 41:101–108.  https://doi.org/10.1016/j.bbi.2014.04.004 CrossRefPubMedGoogle Scholar

Copyright information

© The Author(s) 2018

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Deborah Gibson-Smith
    • 1
  • Mariska Bot
    • 1
  • Marieke Snijder
    • 2
    • 3
  • Mary Nicolaou
    • 2
  • Eske M. Derks
    • 4
    • 5
  • Karien Stronks
    • 2
  • Ingeborg A. Brouwer
    • 6
  • Marjolein Visser
    • 6
    • 7
  • Brenda W. J. H. Penninx
    • 1
  1. 1.Department of Psychiatry, Amsterdam Public Health Research InstituteVU University Medical CenterAmsterdamThe Netherlands
  2. 2.Department of Public HealthAmsterdam Public Health research institute, Academic Medical CenterAmsterdamThe Netherlands
  3. 3.Department of Clinical Epidemiology, Biostatistics and BioinformaticsAmsterdam Public Health research institute, Academic Medical CenterAmsterdamThe Netherlands
  4. 4.Department of PsychiatryAcademic Medical CenterAmsterdamThe Netherlands
  5. 5.QIMR Berghofer, Translational Neurogenomics groupBrisbaneAustralia
  6. 6.Department of Health Sciences, Faculty of Earth and Life Sciences and Amsterdam Public Health research instituteVrije Universiteit AmsterdamAmsterdamThe Netherlands
  7. 7.Department of Internal Medicine, Nutrition and DieteticsVU University Medical CenterAmsterdamThe Netherlands

Personalised recommendations