ABSTRACT

OBJECTIVES

To explore the association between adiposity, major depressive disorder and generalized anxiety disorder, and to assess the role of inflammation, diet quality and physical activity in this association.

METHODS

We used data from 2,977 individuals from the 1993 Pelotas Cohort (Brazil) who attended the 18- and 22-year follow-ups. We assessed general obesity using body mass index, fat mass index, and abdominal obesity using waist circumference. Major Depressive Disorder and generalized anxiety disorder were assessed using the mini-international neuropsychiatric interview. C-reactive protein and interleukin-6 (IL-6) levels were used as a measure of inflammation; diet quality was estimated using the revised diet quality index, and physical activity was assessed by the International physical activity questionnaire (IPAQ, min/day). The association between adiposity and major depressive disorder and generalized anxiety disorder was assessed using logistic regression, and the natural indirect effect via the mediators was estimated using G-computation.

RESULTS

General obesity assessed by body mass index (OR: 2.3; 95% CI:1.13; 4.85), fat mass index (OR: 2.6; 95%CI: 1.37; 4.83), and abdominal obesity (OR: 2.5; 95%CI: 1.18; 5.39) were associated with higher odds of major depressive disorder, whereas major depressive disorder was only associated with obesity assessed by body mass index (OR=1.9; 95% CI: 1.09; 3.46). Obesity and generalized anxiety disorder were not associated. C-reactive protein, diet quality and physical activity did not mediate the effect of obesity on major depressive disorder, and C-reactive protein mediated about 25% of the effect of major depressive disorder on adiposity.

CONCLUSIONS

Depression, but not generalized anxiety disorder, is associated with adiposity in both directions, with a stronger evidence for the direction obesity-depression. Inflammation explains part of the effect of major depressive disorder on obesity but not the other way around. Further research should explore other mechanisms that could be involved in the association between obesity and depression.

Adiposity; Obesity; Depression; Anxiety Disorders; Measures of Association, Exposure, Risk or Outcome

INTRODUCTION

Emerging adulthood (18 to 29 years old) is a critical life period of high instability, when most people have not yet set up the stable structure of an adult life. Obesity and mental disorders are common during this period¹1. Cheng HL, Medlow S, Steinbeck K. The health consequences of obesity in young adulthood. Curr Obes Rep. 2016;5(1):30-7. https://doi.org/10.1007/s13679-016-0190-2
https://doi.org/10.1007/s13679-016-0190-... , ²2. Suvisaari J, Aalto-Setälä T, Tuulio-Henriksson A, Härkänen T, Saarni SI, Perala J, et al. Mental disorders in young adulthood. Psychol Med. 2009;39(2):287-99. https://doi.org/10.1017/S0033291708003632
https://doi.org/10.1017/S003329170800363... , and are associated with the occurrence of several chronic diseases later in life³3. Clarke DM, Currie KC. Depression, anxiety and their relationship with chronic diseases: a review of the epidemiology, risk and treatment evidence. Med J Aust. 2009;190(7 Suppl):S54-60. https://doi.org/10.5694/j.1326-5377.2009.tb02471.x
https://doi.org/10.5694/j.1326-5377.2009... , ⁴4. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham CL, Anis AH. The incidence of co-morbidities related to obesity and overweight: a systematic review and meta-analysis. BMC Public Health. 2009;9:88. https://doi.org/10.1186/1471-2458-9-88
https://doi.org/10.1186/1471-2458-9-88... .

Several meta-analyses of cohort studies showed a bidirectional association between major depressive disorder (MDD) (symptoms and clinical diagnosis) and obesity in adolescents and adults, and the direction from depression to obesity seems to be more common than the opposite⁵5. Luppino FS, Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220-9. https://doi.org/10.1001/archgenpsychiatry.2010.2
https://doi.org/10.1001/archgenpsychiatr...

6. Mannan M, Mamun A, Doi S, Clavarino A. Prospective associations between depression and obesity for adolescent males and females: a systematic review and meta-analysis of longitudinal studies. PLoS One. 2016;11(6):e0157240. https://doi.org/10.1371/journal.pone.0157240
https://doi.org/10.1371/journal.pone.015... - ⁷7. Mannan M, Mamun A, Doi S, Clavarino A. Is there a bi-directional relationship between depression and obesity among adult men and women? Systematic review and bias-adjusted meta analysis. Asian J Psychiatr. 2016;21:51-66. https://doi.org/10.1016/j.ajp.2015.12.008
https://doi.org/10.1016/j.ajp.2015.12.00... . The effect of abdominal obesity on MDD might be higher than the effect of general obesity⁸8. Xu Q, Anderson D, Lurie-Beck J. The relationship between abdominal obesity and depression in the general population: a systematic review and meta-analysis. Obes Res Clin Pract. 2011;5(4):e267-360. https://doi.org/10.1016/j.orcp.2011.04.007
https://doi.org/10.1016/j.orcp.2011.04.0... , and the association between depressive symptoms with changes in abdominal visceral fat has been shown to be stronger than that with changes in overall obesity⁹9. Vogelzangs N, Kritchevsky SB, Beekman ATF, Newman AB, Satterfield S, Simonsick EM, et al. Depressive symptoms and change in abdominal obesity among older persons. Arch Gen Psychiatry. 2008;65(12):1386-93. https://doi.org/10.1001/archpsyc.65
https://doi.org/10.1001/archpsyc.65... . However, the evidence currently available comes from high-income countries (HIC), which limits the extrapolation of results to low- and middle-income countries (LMIC) since differences in sociocultural characteristics and lifestyle behaviors could affect such association.

The bidirectional association between anxiety and obesity has been less investigated. A prospective study with 3,134 adolescents in the United States showed that anxiety disorders predicted overweight/obesity, but not the opposite; in addition, this effect was observed only in males¹⁰10. Roberts RE, Duong HT. Do anxiety disorders play a role in adolescent obesity? Ann Behav Med. 2016;50(4):613-21. https://doi.org/10.1007/s12160-016-9786-8.
https://doi.org/10.1007/s12160-016-9786-... . Evidence regarding generalized anxiety disorder (GAD) and obesity is also weak, and the only study assessing the bidirectional relationship between body mass index (BMI) and GAD found no association, in any direction¹¹11. Pickering RP, Goldstein RB, Hasin DS, Blanco C, Smith SM, Huang B, et al. Temporal relationships between overweight and obesity and DSM-IV substance use, mood, and anxiety disorders: results from a prospective study, the National Epidemiologic Survey on Alcohol and Related Conditions. J Clin Psychiatry. 2011;72(11):1494-502. https://doi.org/10.4088/JCP.10m06077gry
https://doi.org/10.4088/JCP.10m06077gry... .

Several genetic, biological, social, psychological and behavioral factors could explain the interrelationship between obesity and depression/anxiety¹²12. Milaneschi Y, Simmons WKA-Ohoo, Rossum EFC, Penninx BW. Depression and obesity: evidence of shared biological mechanisms. Mol Psychiatry. 2019;24(1):18-33. https://doi.org/10.1038/s41380-018-0017-5
https://doi.org/10.1038/s41380-018-0017-... , ¹³13. Preiss K, Brennan L, Clarke D. A systematic review of variables associated with the relationship between obesity and depression. Obes Rev. 2013;14(11):906-18. https://doi.org/10.1111/obr.12052
https://doi.org/10.1111/obr.12052... . One of the biological mechanisms is immune-inflammatory activation, as both individuals with obesity and depression have increased inflammatory markers, such as C-reactive protein (CRP), interleukin-6 (IL-6) and tumour necrosis factor-alpha (TNF-α)¹²12. Milaneschi Y, Simmons WKA-Ohoo, Rossum EFC, Penninx BW. Depression and obesity: evidence of shared biological mechanisms. Mol Psychiatry. 2019;24(1):18-33. https://doi.org/10.1038/s41380-018-0017-5
https://doi.org/10.1038/s41380-018-0017-... . Some modifiable lifestyle behaviors, such as decreased levels of physical activity and unhealthy diet, are associated with a higher risk of both obesity and depression/anxiety, and could also be involved in such association¹²12. Milaneschi Y, Simmons WKA-Ohoo, Rossum EFC, Penninx BW. Depression and obesity: evidence of shared biological mechanisms. Mol Psychiatry. 2019;24(1):18-33. https://doi.org/10.1038/s41380-018-0017-5
https://doi.org/10.1038/s41380-018-0017-... , ¹³13. Preiss K, Brennan L, Clarke D. A systematic review of variables associated with the relationship between obesity and depression. Obes Rev. 2013;14(11):906-18. https://doi.org/10.1111/obr.12052
https://doi.org/10.1111/obr.12052... . These mechanisms, however, remain understudied.

Using data from a birth cohort in a middle-income country (Brazil), this study aimed to explore the directionality of the association between adiposity and MDD/GAD, as well as to assess the role of inflammation (CRP and IL-6 level) and lifestyle behaviors (diet quality and physical activity) in young adults.

METHODS

Study design and participants

Data from the 1993 Pelotas Cohort were used. In that year, all maternity hospitals in the city of Pelotas were visited daily and births were identified. Livebirths whose family lived in the urban area of the city were examined and their mothers interviewed (less than 1% of the deliveries were at home or other places). A total of 5,265 births were recorded, and 5,249 children participated in the study. Such subjects have been prospectively followed-up, and subsamples of the participants were attained until they reached the age of 11, when the full cohort was contacted. Follow-ups of the full cohort took place at 11, 15, 18 and 22 years of age. After the 18-year follow-up, interviews took place at the research clinic instead of at home. The response rate at 18 and 22 years old was 81.4% and 76.3%, respectively. Further details on study methodology are available elsewhere¹⁴14. Gonçalves H, Assunção MCF, Wehrmeister FC, Oliveira IO, Barros FC, Victora CG, et al. Cohort profile update: The 1993 Pelotas (Brazil) Birth Cohort follow-up visits in adolescence. Int J Epidemiol. 2014;43(4):1082-8. https://doi.org/10.1093/ije/dyu077
https://doi.org/10.1093/ije/dyu077... , ¹⁵15. Gonçalves H, Wehrmeister FC, Assunção MCF, Tovo-Rodrigues L, Oliveira IO, Murray J, et al. Cohort profile update: The 1993 Pelotas (Brazil) Birth Cohort follow-up at 22 years. Int J Epidemiol. 2018;47(5):1389-90e. https://doi.org/10.1093/ije/dyx249
https://doi.org/10.1093/ije/dyx249... .

Measures

Depression and anxiety

The MDD and GAD were assessed at the 18- and 22-year follow-ups using the Mini International Neuropsychiatric Interview (MINI), version 5.0¹⁶16. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59 Suppl 20:22-33;quiz 4-57. . This instrument was applied by trained psychologists and was validated for the Brazilian population¹⁷17. Azevedo Marques JM, Zuardi AW. Validity and applicability of the Mini International Neuropsychiatric Interview administered by family medicine residents in primary health care in Brazil. Gen Hosp Psychiatry. 2008;30(4):303-10. https://doi.org/10.1016/j.genhosppsych.2008.02.001
https://doi.org/10.1016/j.genhosppsych.2... . Mental disorders definitions followed the DSM-5 criteria, and individuals were diagnosed with current GAD and MDD if they had presented symptoms in the 6 months and 15 days prior to the interview, respectively, in addition to moderate or severe levels of impairment. Impairment was assessed through the following question: “Thinking about your life, school, work, home, family and friends, how much have these problems impaired you? The answer options were: no impairment (0), mild impairment (1), moderate impairment (2), or severe impairment (3). Neuropsychiatric interview was not performed with deaf-mute individuals or with those with cognitive disability (n=3 at 18 years old and n=19 at 22 years old).

Adiposity

At the 18- and 22-year follow-ups, a detailed assessment of body composition was performed, which included anthropometric measures and dual X-ray absorptiometry (DXA; GE Medical Systems Lunar). Body composition evaluation was carried out by trained and standardized assessors. For the examinations, participants wore tight-fitting gym suits (shorts and sleeveless top) and were barefoot. Exams were not performed in women who were pregnant or suspect to be pregnant and individuals with physical problems (wheelchair users and individuals with osteoarticular deformities). For DXA examination, those with weight over 120 kg or height over 192 cm (maximum device capacity), and those who had implanted metal pins, screws, plates and non-removable metallic objects (body piercings and/or chains) were not evaluated.

In this study, BMI and fat mass index (FMI) were used as indicators of total adiposity, and waist circumference (WC) as a measurement of central adiposity. Height was measured twice, using an aluminum stadiometer with accuracy of 1mm, and the mean of the two measurements was used. Weight was measured with an accuracy of 10g, using an electronic scale connected to BodPod®. The BMI was calculated dividing total weight (kg) by height squared (m²2. Suvisaari J, Aalto-Setälä T, Tuulio-Henriksson A, Härkänen T, Saarni SI, Perala J, et al. Mental disorders in young adulthood. Psychol Med. 2009;39(2):287-99. https://doi.org/10.1017/S0033291708003632
https://doi.org/10.1017/S003329170800363... ). WC was measured at the narrowest point of the waist, with a flexible tape, to the nearest 1mm. FMI was calculated by dividing total fat mass (kg), obtained by DXA, by height squared (m²2. Suvisaari J, Aalto-Setälä T, Tuulio-Henriksson A, Härkänen T, Saarni SI, Perala J, et al. Mental disorders in young adulthood. Psychol Med. 2009;39(2):287-99. https://doi.org/10.1017/S0033291708003632
https://doi.org/10.1017/S003329170800363... ).

General obesity was defined according to both BMI (≥ 30 kg/m²2. Suvisaari J, Aalto-Setälä T, Tuulio-Henriksson A, Härkänen T, Saarni SI, Perala J, et al. Mental disorders in young adulthood. Psychol Med. 2009;39(2):287-99. https://doi.org/10.1017/S0033291708003632
https://doi.org/10.1017/S003329170800363... ) and FMI (> 9 for men and > 13 for women)¹⁸18. Kelly TL, Wilson KE, Heymsfield SB. Dual energy X-Ray absorptiometry body composition reference values from NHANES. PLoS One. 2009;4(9):e7038. https://doi.org/10.1371/journal.pone.0007038
https://doi.org/10.1371/journal.pone.000... . Abdominal obesity was defined as WC ≥ 88 cm for women and WC ≥ 102 for men. Continuous measures of BMI, FMI and WC were also used for sensitivity analysis.

Mediators: inflammation, diet quality and physical activity

CRP (mg/L) e IL-6 (pg/ml), diet quality, and physical activity level in leisure time or commuting, all assessed at the 18-y follow-up, were used as possible mediators.

To assess CRP and IL-6 concentration, non-fasting blood samples were collected by venipuncture, using vacutainer tubes. All samples were processed in the laboratory, stored in the same freezers at ultra-low temperature and registered in a central bio-repository. The CRP was measured by immunoturdimetric assay (Labtest Diagnóstica SA, Minas Gerais, Brazil), whereas IL-6 was measured using the Quantikine® HS Human IL-6 immunoassay kit (R&D Systems®, Inc.; Minneapolis, MN55413, USA). As CRP and IL-6 were asymmetrically distributed, they were log-transformed to obtain normalized residuals.

Total caloric intake was assessed using a food frequency questionnaire with 88 items¹⁹19. Schneider BC, Motta JVS, Muniz LC, Bielemann RM, Madruga SW, Orlandi SP, et al. Design of a digital and self-reported food frequency questionnaire to estimate food consumption in adolescents and young adults: birth cohorts at Pelotas, Rio Grande do Sul, Brazil. Rev Bras Epidemiol. 2016;19(2):419-32. https://doi.org/10.1590/1980-5497201600020017
https://doi.org/10.1590/1980-54972016000... . Diet quality was estimated using the revised diet quality index, as previously described²⁰20. Castilhos CB, Schneider BC, Muniz LC, Assunção MCF. Qualidade da dieta de jovens aos 18 anos de idade, pertencentes à coorte de nascimentos de 1993 da cidade de Pelotas (RS), Brasil. Cienc Saude Coletiva. 2015;20(11):3309-3318. https://doi.org/10.1590/1413-812320152011.17822014
https://doi.org/10.1590/1413-81232015201... , and analyzed as a continuous variable.

Physical activity level in leisure time and commuting (min/day) was self-reported, using the long version of the IPAQ²¹21. The IPAQ Group. Guidelines for data processing and analysis of the International Physical Activity Questionnaire (IPAQ): short form; version 2.0; 2004. Available in www.ipaq.ki.se
www.ipaq.ki.se... ; hence, individuals were classified as active (> 150 minutes) or insufficiently active (< 150 minutes).

Base confounders (confounders of the association between mental disorders and adiposity)

At the perinatal visit, the following variables were assessed: sex (male/female); family income (earned by family members in the month before the interview); mother’s education level (complete years), age (years), smoking during pregnancy (yes/no), alcohol intake during pregnancy (yes/no) and pre-pregnancy BMI (calculated using self-reported weight and height); birth weight (measured in grams). Maternal smoking and alcohol intake in pregnancy were defined as any reported use of cigarettes or alcohol, respectively, at any time during pregnancy.

Maternal mental health was measured at the 11-year follow-up using the short version of the Self-Reporting Questionnaire (SRQ-20)²²22. Mari JJ, Williams P. A validity study of a psychiatric screening questionnaire (SRQ-20) in primary care in the city of Sao Paulo. Br J Psychiatry. 1986;148:23-6. https://doi.org/10.1192/bjp.148.1.23
https://doi.org/10.1192/bjp.148.1.23... . At the 15-year follow-up, information on the participant’s self-reported skin color (white, black/brown, yellow/indigenous), alcohol use (yes/no) and tobacco use in the previous month (yes/no) were assessed. Parenthood (yes/no) was assessed at the 18-year follow-up.

Intermediate confounders (confounders of association between each mediator and the outcome)

For mediation analysis by physical activity, diet quality at 18 years of age was considered as intermediate confounder. For the mediation analysis by diet quality, physical activity at 18 years old was considered an intermediate confounder, whereas for the mediation analysis by inflammation markers, diet quality and physical activity were both considered intermediate confounders.

Statistical Analysis

Analyses were conducted using Stata, version 12.1 (Stata Corp., College Station, TX, USA). To assess the association between mental disorders and adiposity, logistic regression models were performed considering each direction, with and without adjustment for the potential base confounders defined above. In all analyses, those who had the outcome of interest at 18 years old were excluded to reduce the probability of reverse causality. Supplementary analyses were performed using continuous measures of adiposity, considering all participants. Standard deviations were used when adiposity was assessed as exposure, and changes in the adiposity measure between 18 and 22 years old were used when assessing it as the outcome. Tests were carried out for interactions between sex and each exposure, and no evidence was found (all p-values > 0.05).

Mediation analyses were performed using G-computation with 10,000 bootstrap replications, to estimate the natural indirect effect (NIE) ‒ mediated effect ‒ of the association between mental disorders and adiposity. Each mediator was individually included on models so that NIE corresponds to the effect by each mediator separately, instead of including it all in the same model. All mediation analyses were adjusted for the base and intermediate confounders described above, considering possible interactions between exposures and mediators. The proportion mediated was calculated as: log(OR NIE)/log(OR Total effect) ×100%.

Ethical Considerations

All follow-ups were approved by the Medical Ethics Committee of the Federal University of Pelotas, and participants or their legal caregivers voluntarily signed an informed consent form prior to participating in the study. Individuals with high suicidal risk in the neuropsychiatric interview were referred to health services.

RESULTS

Sample Characteristics

This study sample size comprised 2,977 participants who had complete information on MDD/GAD and adiposity measures at both the 18-year and 22-year follow-ups. Compared to those excluded from analysis, the included were more likely to be female, to have lower socioeconomic status (lower family income and maternal education level), and to have better health indicators, e.g. lower prevalence of maternal smoking during pregnancy, common maternal mental disorders, and adolescent smoking at the age of 15 (Data not showed in Table).

The prevalence of all obesity indicators and GAD were higher at 22 than at 18 years old, whereas the prevalence of MDD was higher at the age of 18 ( Figure 1 ).

Association of obesity at 18 years old with MDD and GAD at 22 years old

After adjustment for confounders ( Figure 2 ), all obesity measures were associated with higher odds of MDD, with a similar magnitude (OR varying from 2.3 to 2.6). No association between obesity and incidence of GAD was observed in both unadjusted or adjusted analyses. The same pattern of associations was observed using continuous measures of adiposity (data not showed in table).

Mediation of the association between adiposity and MDD

CRP, IL-6, diet quality, and physical activity did not seem to mediate the association between adiposity and MDD ( Table 1 ).

Association of MDD and GAD at 18 years old with adiposity at 22 years old

The MDD was associated with higher odds of general obesity regarding to BMI, in both unadjusted (OR: 2.20; 95% CI: 1.20–3.73) and adjusted analyses (OR: 1.94, 95% CI: 1.09–3.46), but no association was found with other adiposity measures ( Figure 3 ). No association was observed between GAD and any obesity measure, in unadjusted or adjusted analyses. Similar results were observed when adiposity measures were assessed as continuous measures (data not showed in table).

Mediation of the association between MDD and adiposity

The CRP concentration explained about 25% of the association between MDD and obesity measured by BMI and no evidence was found for IL-6. Diet quality and physical activity did not mediate the association between MDD and adiposity measures ( Table 2 ).

DISCUSSION

The major findings of this study were: (1) a bidirectional association was found between MDD and obesity, even after accounting for many important, well-known confounders; (2) the effect of adiposity on MDD was stronger and more consistent than the effect of MDD on obesity; (3) GAD was not associated with obesity, and vice versa; (4) CRP, IL-6, physical activity and caloric intake did not mediate the effect of obesity on MDD, and CRP accounted for about 25% of the association between MDD and obesity (BMI).

Despite the various methods used to assess the bidirectionality of associations (longitudinal studies have used regression models, auto-regressive cross-lagged panel analyses and generalized estimating equations to access reciprocal risks), the bidirectional association between obesity and depression is consistent with the recent literature⁵5. Luppino FS, Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220-9. https://doi.org/10.1001/archgenpsychiatry.2010.2
https://doi.org/10.1001/archgenpsychiatr...

6. Mannan M, Mamun A, Doi S, Clavarino A. Prospective associations between depression and obesity for adolescent males and females: a systematic review and meta-analysis of longitudinal studies. PLoS One. 2016;11(6):e0157240. https://doi.org/10.1371/journal.pone.0157240
https://doi.org/10.1371/journal.pone.015... - ⁷7. Mannan M, Mamun A, Doi S, Clavarino A. Is there a bi-directional relationship between depression and obesity among adult men and women? Systematic review and bias-adjusted meta analysis. Asian J Psychiatr. 2016;21:51-66. https://doi.org/10.1016/j.ajp.2015.12.008
https://doi.org/10.1016/j.ajp.2015.12.00... , ²³23. Rajan TM, Menon V. Psychiatric disorders and obesity: a review of association studies. J Postgrad Med. 2017;63(3):182-90. https://doi.org/10.4103/jpgm.JPGM_712_16
https://doi.org/10.4103/jpgm.JPGM_712_16... . In previous studies, the OR varies from 1.2 to 3.8 when depression leads to obesity, and from 1.1 to 5.8 when obesity leads to depression⁵5. Luppino FS, Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220-9. https://doi.org/10.1001/archgenpsychiatry.2010.2
https://doi.org/10.1001/archgenpsychiatr... , ²³23. Rajan TM, Menon V. Psychiatric disorders and obesity: a review of association studies. J Postgrad Med. 2017;63(3):182-90. https://doi.org/10.4103/jpgm.JPGM_712_16
https://doi.org/10.4103/jpgm.JPGM_712_16... . In our study, the effect of obesity on MDD was more consistent than in the opposite direction. A recent Mendelian randomization study (using genetic variants as instrumental variables) found that higher BMI predicts depressive symptoms but not the other way around, but the authors acknowledge the weakness of genetic instruments for instrumenting depression²⁴24. Broek N, Treur JL, Larsen JK, Verhagen M, Verweij KJH, Vink JM. Causal associations between body mass index and mental health: a Mendelian randomisation study. J Epidemiol Community Health. 2018;72(8):708-10. https://doi.org/10.1136/jech-2017-210000
https://doi.org/10.1136/jech-2017-210000... . No difference by sex was observed in any of the associations, in accordance with previous meta-analyses⁵5. Luppino FS, Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BW, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220-9. https://doi.org/10.1001/archgenpsychiatry.2010.2
https://doi.org/10.1001/archgenpsychiatr... , ⁷7. Mannan M, Mamun A, Doi S, Clavarino A. Is there a bi-directional relationship between depression and obesity among adult men and women? Systematic review and bias-adjusted meta analysis. Asian J Psychiatr. 2016;21:51-66. https://doi.org/10.1016/j.ajp.2015.12.008
https://doi.org/10.1016/j.ajp.2015.12.00... . However, one should note the effect of abdominal obesity on MDD seemed to be stronger in men, even though we had little power to detect such difference.

We found no evidence of mediation by IL-6 and CRP levels on the association of obesity leading to MDD, different from a previous cohort study with English older adults, which showed that CRP levels explained about 20% of the effect of obesity on depressive symptoms over a 4-year period of follow-up²⁵25. Daly M. The relationship of C-reactive protein to obesity-related depressive symptoms: a longitudinal study. Obesity (Silver Sprimg). 2013;21(2):248-50. https://doi.org/10.1002/oby.20051
https://doi.org/10.1002/oby.20051... . We found that about 25% of the MDD effect on obesity (BMI) was mediated by the CRP level, but no evidence of mediation was found with others adiposity measures or IL-6; Atypical MDD ‒ one distinct subtype of MDD that is characterized by reversed neurovegetative symptoms (i.e. increased appetite, leaden paralysis, and hypersomnia) ‒ has been frequently associated with high levels of inflammatory markers²⁶26. Hickman RJ, Khambaty T, Stewart JC. C-reactive protein is elevated in atypical but not nonatypical depression: data from the National Health and Nutrition Examination survey (NHANES) 1999-2004. J Behav Med. 2014;37(4):621-9. https://doi.org/10.1007/s10865-013-9510-0
https://doi.org/10.1007/s10865-013-9510-... as well as obesity²⁷27. Lasserre AM, Glaus J, Vandeleur CL, Marques-Vidal P, Vaucher J, Bastardot F, et al. Depression with atypical features and increase in obesity, body mass index, waist circumference, and fat mass: a prospective, population-based study. JAMA Psychiatry. 2014;71(8):880-8. https://doi.org/10.1001/jamapsychiatry.2014.411
https://doi.org/10.1001/jamapsychiatry.2... , differently than melancholic MDD. Pro-inflammatory cytokines, such as IL-6, are associated with heightened hypothalamic-pituitary-adrenal (HPA) axis activity, which can lead to elevated cortisol levels, which are known to increase appetite and preference for energy-dense food and to promote the accumulation of fat, especially in the abdominal region¹²12. Milaneschi Y, Simmons WKA-Ohoo, Rossum EFC, Penninx BW. Depression and obesity: evidence of shared biological mechanisms. Mol Psychiatry. 2019;24(1):18-33. https://doi.org/10.1038/s41380-018-0017-5
https://doi.org/10.1038/s41380-018-0017-... . As the role of CRP level on MDD and obesity association was no consistent across all adiposity measures in our study, it is important that this hypothesis would be confirmed in others longitudinal studies to rule out that this result was observed by chance.

Our study did not confirm the hypothesis that decreased levels of physical activity and consumption of low-quality diet mediated the association between obesity and MDD. We further considered other diet indicators (i.e. consumption of ultraprocessed foods and caloric intake at 18 years old), and results were similar (data not shown in this paper). A cross-sectional study using data from National Health and Nutrition Examination Surveys in the United States showed that physical activity mediated 79% of the effect of MDD on BMI in young adults, but diet quality was not involved in this association²⁸28. Beydoun MA, Wang Y. Pathways linking socioeconomic status to obesity through depression and lifestyle factors among young US adults. J Affect Disord. 2010;123(1-3):52-63. https://doi.org/10.1016/j.jad.2009.09.021
https://doi.org/10.1016/j.jad.2009.09.02... . A longitudinal study carried out with female Mexican health workers, exploring the bidirectional association between depression and body fat, showed that higher caloric intake and lower physical activity in leisure time mediated the effect of depressive symptoms on body fat, and no effects of obesity on depression were found²⁹29. Quezada AD, Macias-Waldman N, Salmerón J, Swigart T, Gallegos-Carrillo K. Physical activity and calorie intake mediate the relationship from depression to body fat mass among female Mexican health workers. Int J Behav Nutr Phys Act. 2017;14(1):160. https://doi.org/10.1186/s12966-017-0612-x
https://doi.org/10.1186/s12966-017-0612-... . In our data, obesity was associated with higher physical activity, contrary to what would be expected, which could explain the lack of an indirect effect by this variable. Although counterintuitive, this association may be explained by a tendency of over-reporting physical activity among obese people, so we cannot rule out that information bias led to this result. Moreover, the exposures and mediators variables are both collected in the same follow-up and reverse causality may be present, masking the indirect effect these variables may have in the reciprocal association of adiposity and the mental disorders evaluated.

One important mechanism that could explain why depression is related to a higher risk of obesity is the use of antidepressant drugs. We intended to test such association, but the prevalence of antidepressant drug use at 18 years old in our sample was only 0.7%, and antidepressants use was not associated with obesity. Hence, it is unlikely that the effect of MDD on obesity would be completely attributable to antidepressant drugs use in our data. We acknowledge that other mechanisms linking MDD and adiposity, such as inadequate sleep, low self-esteem, body image dissatisfaction, binge eating/emotional eating, and experience of weight-related or MDD-related stigma¹³13. Preiss K, Brennan L, Clarke D. A systematic review of variables associated with the relationship between obesity and depression. Obes Rev. 2013;14(11):906-18. https://doi.org/10.1111/obr.12052
https://doi.org/10.1111/obr.12052... , were not explored in this study and require further investigation.

Although plausible, the association between anxiety and adiposity has been poorly investigated. The mechanisms are very similar to those linking MDD and obesity. Biological mechanisms included the possibility of sharing a common genetic basis and HPA axis dysregulation³⁰30. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes (Lond). 2010;34(3):407-19. https://doi.org/10.1038/ijo.2009.252
https://doi.org/10.1038/ijo.2009.252... . Weight discrimination, poorer social support and social networks could increase the risk of anxiety in obese individuals³⁰30. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes (Lond). 2010;34(3):407-19. https://doi.org/10.1038/ijo.2009.252
https://doi.org/10.1038/ijo.2009.252... . The pressure to gain control over their weight could be distressing for obese people and increase the risk of anxiety, particularly when many attempts to lose weight have failed³⁰30. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes (Lond). 2010;34(3):407-19. https://doi.org/10.1038/ijo.2009.252
https://doi.org/10.1038/ijo.2009.252... . Symptoms of anxiety may increase appetite and the preference for and consume of palatable “comfort foods” as a mean to alleviate their negative feelings and may, as consequence, lead to obesity³¹31. Singh M. Mood, food, and obesity. Front Psychol. 2014;5:925. https://doi.org/10.3389/fpsyg.2014.00925
https://doi.org/10.3389/fpsyg.2014.00925... .

This study did not show any association between obesity and GAD, corroborating a previous bidirectional study on this same association¹¹11. Pickering RP, Goldstein RB, Hasin DS, Blanco C, Smith SM, Huang B, et al. Temporal relationships between overweight and obesity and DSM-IV substance use, mood, and anxiety disorders: results from a prospective study, the National Epidemiologic Survey on Alcohol and Related Conditions. J Clin Psychiatry. 2011;72(11):1494-502. https://doi.org/10.4088/JCP.10m06077gry
https://doi.org/10.4088/JCP.10m06077gry... . Even though some studies have found a positive association between anxiety disorders and obesity (BMI)³⁰30. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a systematic review and meta-analysis. Int J Obes (Lond). 2010;34(3):407-19. https://doi.org/10.1038/ijo.2009.252
https://doi.org/10.1038/ijo.2009.252... , such association might be related to specific anxiety disorder other than GAD. Future prospective studies should test the effect of other types of anxiety disorders and their association with adiposity.

Our study has several strengths. First, we used data from a birth cohort carried out in a middle-income country, where this association has been poorly investigated. We could examine directionality of effects in the same population, using the same instruments. Besides BMI, we also used different and more accurate measures of adiposity and were able to control for several relevant confounders, such as birth weight, maternal BMI and maternal common mental disorders, often not accounted for in previous studies.

Nevertheless, results of this study should also be considered in the context of several limitations. We were able to complete assessments with 56.7% of the original cohort, raising the possibility of non-response bias. Those excluded from the analysis were more likely to have lower socioeconomic status, which is associated with higher prevalence of mental disorders and obesity in our data. Therefore, we expect that associations reported here might be underestimated.

Although we excluded those with MDD and GAD at 18-year follow-up to reduce the probability of reverse causality, we cannot guarantee the cases were re-incident, as no information on previous diagnosis of MDD and GAD was available. Despite having high specificity, MINI is a screening tool and does not assess clinical diagnosis of mental disorders, so we might have misclassification error on the assessment of these disorders. However, the non-differential misclassification error is expected to underestimate associations. Finally, we were not able to distinguish between atypical and melancholic MDD, and some studies indicate that only individuals with atypical MDD have a higher risk of obesity.

This study suggests a bidirectional association between obesity and MDD. CRP level mediated about 25% of the effect of MDD on adiposity, whereas diet quality and physical activity had no indirect effect in this relationship. The GAD-obesity association was not confirmed in this study, but evidence of this association in other contexts should be investigated. Individuals with obesity and MDD could benefit from receiving a multidisciplinary treatment that includes psychological evaluation and weight gain monitoring, respectively. This could help prevent co-occurrence of these diseases and to better manage these conditions in those who already present both obesity and depression. Although we found no association with diet and physical activity in this analysis, a healthy lifestyle should be encouraged in these patients, as it could help obese individuals to lose weight and, consequently, alleviate depressive symptoms and reduce inflammatory markers that apparently can be involved in the association between MDD and obesity.

REFERENCES

Publication Dates

History