INTRODUCTION
Obesity and depression are considered notorious health problems, not only because of their significant prevalence but also because of their high morbidity mortality rates. According to recent data from the World Health Organization (WHO) in 2014 it was estimated that more than 600 million people were affected with obesity whilst at least 2.6 million people die from obesity every year. 44% of global cases of diabetes, 23% of ischemic heart diseases and 7-41% of certain cancers are attributable to overweight and obesity [1]. When it comes to depression the figures are no less alarming, it is estimated that depression affects 350 million people in the world, making it one of the main causes of disability and morbidity worldwide. In addition, depression is also an important cause of premature mortality, primarily due to suicide [2].
The probable association between obesity and depression has been studied repeatedly over time in scientific literature. This is mainly due to the fact that they both carry a high prevalence and an increased risk of cardiovascular disease [3]. While many cross-sectional studies have documented this relationship its significance still remains unclear.
Prospective studies have revealed inconsistent findings regarding the sequence in the onset of depression and obesity [4]. This discrepancy could be due to the methodological variation of the different studies (including variations in the sample selection) the duration of the follow-up and/or the evaluation and diagnosis of depression and obesity. The relationship between obesity and depression has also been studied in childhood and adolescence. A prospective study [5] determined that adolescent women affected with obesity predicted an increased risk of major depression (up to almost 4 times more). However this risk was not significant for men. A meta-analysis that included 8 longitudinal studies [6], concluded that there was a bidirectional relationship between depression and obesity. Thus he determined that people with obesity had a 55% increased risk of developing depression over time, and that depressed people had a 58% increased risk of obesity. In addition, the relationship between obesity and depression was stronger than the relationship between overweight and depression, which reflected a dose-response gradient.
The main objectives of this review will be three; 1) to define the methodological quality of those more recent studies that analysed the association between depression and obesity, 2) to determine whether there is indeed an association between both conditions and 3) to clarify what influence they may have over each other. Additionally it will examine the importance of certain socio-demographic, clinical and therapeutic variables in the relation between obesity and depression.
METHODS
A literature review was conducted of those articles published in the PUBMED database with an inclusion deadline from January 1, 2012 to December 31, 2017, using the following MeSH terms: “obesity” AND “depressive disorder.” The filters applied were those of studies published in the last 6 years and carried out in humans. 179 articles were obtained and the main researcher (BV) made an exhaustive reading of the abstracts, selected the publications based on the inclusion criteria and was responsible for the successive stages. We follow the international recommendations for systematic reviews as Preferred Items for Reporting of Systematic Reviews and Meta-Analyses (PRISMA) [7].
The following inclusion criteria was applied: available abstract, English written articles, already established diagnostic criteria for both conditions (obesity according to body mass index (BMI) and depression according to major depression criteria in the DSM IV or ICD-10, studies of prevalence in both pathologies and finally, the evaluation of the possible prognostic association between these conditions when they affect the same person simultaneously. The exclusion criteria where the following: no available abstract, articles written in a non-English language, articles of very low quality according to the GRADE system (Grading of Recommendations, Assessment, Development and Evaluation) [8,9], and the failure to fulfil the objectives of the study.
The GRADE system initially classifies the evidence into high or low, coming from experimental or observational studies; subsequently and following a series of considerations (risk/benefit balance, values and preferences of the patients and professionals, and the use of resources or costs), the evidence is classified into high, moderate, low or very low [9]. In comparison with other systems, the advantages of GRADE classification includes a clear separation between quality of evidence and strength of recommendations, provide comprehensive criteria and acknowledgment of values and preferences and explicit evaluation of the importance of outcomes of alternative management strategies [8]. Finally studies which main objective was the evaluation of interventions aimed at reducing the frequency of these conditions were also excluded.
By reading the abstracts, a total of 39 articles were selected, which were then thoroughly revised, eventually discarding 21 of them for not complying with the inclusion criteria. Most of these articles were discarded for failing to adapt to the international diagnostic standards rendering it difficult to draw clear conclusions. The third stage corresponded to the critical reading of the 18 selected articles. Figure 1 summarizes each of these stages. Subsequently, the methodological quality of the 18 selected works was evaluated, for which the GRADE classification was used [8,9]. This system is a rigorous and transparent instrument that classifies scientific publications according to their level of evidence and determines the strength of their conclusions. There are 4 levels of quality, so that initially the publications are classified as high or low quality, depending on whether they are experimental or observational studies. In the final phase, a series of characteristics are analysed (limitations, biases and possible confusion biases among others) that allow us to sort the final classification in high, moderate, low or very low evidence level.
RESULTS
Different types of studies and methodological quality
Of the 18 selected studies, 15 were observational and 3 were experimental clinical trials. Within the observational group (15), 6 had a longitudinal design (one retrospective and five prospective) and the other 9 were cross-sectional. The main features as well as the evaluation of the quality of these studies are summarized in Tables 1, 2 and 3. The heterogeneity in the ages of the selected samples should be noted, so that four of the studies were performed in adolescents aged 11 to 17 years [10-13] and the other two [14,15] in patients aged 65 and over.
Depression as a risk factor for obesity
In this section we will focus on those 10 studies that studied depression as a risk factor for obesity. The conclusions here obtained were not uniform. Three of these studies [15-17] concluded that only atypical depression (DSM subtype) was a risk factor for obesity. However, there was no correlation between the other subtypes such as classic or melancholic depression and obesity. Others studies found an association between depressive symptoms and obesity. A higher OR was found for abdominal obesity compared to BMI >30 [18] whilst there was a higher OR in patients with obesity in comparison than those with overweight [19].
Another study focused on age and ethnicity [13], concluding that in the selected sample (African-American adolescent males), depression was a risk factor for the development of obesity. It found no association between obesity and the rest of the sample. Finally, 3 of the studies in this first section [13,20,21] did not obtain significant differences, concluding that depression was not a risk factor for obesity in the selected sample (Figure 2).
Obesity as a risk factor for depression
In this second section, we discovered 5 studies that concluded that obesity was a risk factor for depression. Only one of the articles could not establish a relationship between both disorders, although the conclusions must be adjusted [11]. The group of Nigatu et al. [20] selected a sample with recurrent depressive disorder and another group of participants who had suffered a single depressive episode. They found that that obesity acted as a risk factor for depression only for the group with recurrent depressive episodes. Another group analysed the association between obesity, body image and depression in a sample of more than 4,000 adolescents [12]. Although it was initially established that obesity and depression were indeed associated, this was not maintained when adjusting the results for other factors such as sociodemographic and clinical variables. Regarding sex, there is data that suggests a different obesity-depression relationship between men and women, since it was only in woman that obesity increased the risk of depression [22]. Finally, in a different work on the same sample of 4,000 adolescents [11], the authors concluded that adolescents affected with obesity did not have higher depression rates, although they did detect an increased risk of future obesity in the group of adolescents with depression (Figure 3).
Obesity-depression association. Prognostic implications
An overall of four articles viewed this subject. The first article [10], studied the response to fluoxetine in patients admitted to an acute care unit when diagnosed with a depressive episode. The results concluded that those participants who did not reach symptomatic remission had a higher BMI at the beginning of the treatment. A second article [23] analysed the association between obesity, depression and bipolar disorder in a population sample. They observed that patients affected with obesity that had suffered a mayor depressive episode were more likely to suffer from bipolar disorder in the future than those patients that had depression but did not suffer from obesity. The association between obesity and the response to antidepressant treatment was also studied in one of the clinical trials [24]. Patients with depression where randomly assigned to take different therapeutic options. All patients had their BMI calculated at the beginning and end of the clinical trial. There were no significant differences in the response to treatment when classified by BMI yet there was disparity in the relationship between the BMI and comorbidities. This association was studied leading to the conclusion that BMI in this sample was related to the clinical presentation of depression, with those patients with higher BMI being more prone to suffer from comorbid pathologies (medical and psychiatric, such as social phobia and bulimia). Finally, the results of a multidisciplinary and intensive treatment for depression were not related to the BMI baseline of participants according to data from another study [25].
DISCUSSION
Our chief aim when undertaking this systematic review has been to clarify the association between obesity and depression, two diseases with great impact over global health. In none of the three sections in which the selected articles have been divided (depression>risk factor>obesity, obesity>risk factor> depression and prognosis when depression and obesity are associated) have we observed unanimity in the conclusions.
In the first group (depression as risk factor for obesity) 8 out of the 11 selected articles presented data in favour of this association. We highlight the results of a clinical trial [26] in which patients with depression who took a high-fat diet had higher rates of obesity than those patients with the same diet but without depressive disorder diagnosis. However, the subtype of depression must be taken into account in order to be able to relate it to obesity. Data derived from this clinical trial revealed that it is only the subtype of atypical depression that has been statistically associated. These findings support the current trends that suggest that within the depressive syndrome there would be different entities with different features that would be influenced in different ways by other variables. Regarding this subject, one of the selected articles [13] analysed if there was an association between depression and obesity in adolescents when divided according to ethnic group and sex. They found that it was only in African-American males that depression was a risk factor for obesity; however this did not occur in the rest of the sample.
We also analysed articles that could not demonstrate a correlation between obesity and depression [20]. However, here we must point out that the methodological quality of these studies was not very high, for they did not take into account, for example, the considerable number of losses. They also failed to monitor the influence of certain important variables such as the antidepressant treatment.
In the second section there is a greater consensus when it comes to affirming that obesity increases the risk of depression. Out of the six selected articles, five concluded that there is an association between both factors. Once again, the depressive subtype, the average age of the sample and sex are variables that influenced the results. For example, according to Nigatu et al. [20] obesity would only act as a risk factor for depression in case of recurrent major depressive disorder, however for a sample of adolescents that were followed in time, there was no association between both factors [11].
In the third section we analysed the influence on the prognosis of the obesity-depression combination in the same patient. We found two clinical trials [10,24] that analysed this issue. There was no homogeneity in the samples, since Lin selected hospitalized patients whilst Toups worked with an ambulatory sample with comorbid anxiety disorder. While in the first article the results show that in patients affected with obesity the response to the antidepressant treatment was significantly worse, in the second article no significant associations were obtained.
It has been suggested that the association between depression and obesity is probably due to the action of certain genes involved in both pathologies. For example the genes that encode glucocorticoids, leptin and dopamine receptors. The role of environmental factors should also be noted, especially all those situations that contribute to maintaining a situation of chronic stress. The results of a study conducted on twins by Afari et al. [27] concluded that only 12% of the genetic component of depression was shared by obesity, so that environmental factors are fundamental to understand how both pathologies are related [28]. The current etiopathogenic model states that chronic stress sets in motion a series of mechanisms involving the central nervous system, the hypothalamic-pituitary-adrenal axis and the autonomic nervous system [29]. These mechanisms would then generate a pro-inflammatory state through peripheral resistance to the glucocorticoid action, intestinal bacterial translocation, increase of catecholamines and the secretion of cytokines (TNF-α and IL-6). This pro-inflammatory environment would act over white fat adipocytes leading to an inappropriate local activation, with an increase in the production of leptin and a decrease of adiponectin, causing inflammation and the accumulation of fatty tissue [30]. In addition to these local changes, inflammatory cytokines act over the central nervous system, inducing changes in synaptic plasticity and in neurogenesis that would be similar to those that occur in depression.
As previously explained, the wide methodological variety of the studies included in this review hampered the conclusion of unequivocal results. This is seen in the heterogeneity of the sample size, the design of the studies or the different instruments used to measure obesity and depression, all which could significantly influence the results. However, it can be stated that a high percentage of the analysed bibliography, demonstrates that both pathologies are associated, although in some cases it is difficult to determine the significance of this association.
Another strong point in this study is that it highlights the importance of the subtypes of depression, a feature that is not generally taken into account although there is increasing evidence that depressive disorder will bring different entities together, each with their own clinical and therapeutic characteristics. Furthermore, the influence of age, sex or ethnicity over depressive symptomatology validates the change in diagnostic manuals towards a more dimensional perspective, as well as the inclusion of sub-threshold forms of presentation associated with clinical and functional discomfort. These results highlight that the reality of the great psychiatric syndromes is probably much more complex than what it was initially believed to be.
The current systematic review has several limitations. Firstly, the articles were only reviewed by a single researcher and most of these studies were cross-sectional, rendering it impossible to obtain causality relationships from them. Following the type of design, a high percentage of the studies presented a low to moderate methodological quality according to the GRADE scale [8,9]. Finally, the search was carried out only in Pubmed, and although it is one of the most powerful databases to day, it is possible that additional publications collected in other bibliographic sources were not detected.
CONCLUSIONS
Obesity and depression are disorders with a high prevalence and an extraordinary effect over global morbidity and mortality. The relationship between both disorders has been analysed in scientific literature, obtaining significant associations but also contradictory results. The most current data demonstrates that there is a relationship between both entities, although there is no unanimity when it comes to establishing the meaning of this association. Certain variables such as the subtype of atypical depression, female sex, and African-American ethnicity could influence the relationship between depression and obesity thus it is advised that they be examined in future studies. The limited methodological quality of the articles included in this review, with a large proportion of cross-sectional studies that are very heterogeneous in their design has influence the difficulty to draw clear conclusions. In the future it is recommended to include a larger number follow up of studies that are based on unified criteria.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
