Introduction

Although bariatric surgery is effective regarding substantive weight loss, there is great variability in outcomes, and about 20 % of individuals either fail to achieve substantial weight loss or regain excessive amounts of weight after initial success [1]. There has been considerable effort to identify possible predictors of weight outcomes after bariatric surgery. One area of focus is the relationship between comorbid depressive symptoms and problematic eating behaviors.

Compared to non-surgical obese individuals and to the general population, comorbid psychopathology is quite prevalent in bariatric surgery candidates [2, 3]. While anxiety disorders are the most common psychiatric disorders in general population, affective disorders including major depressive disorder, are the most frequent lifetime form of psychopathology among bariatric surgery candidates [4, 5]. A recent meta-analysis of studies reporting the prevalence of preoperative mental health conditions among patients seeking and undergoing bariatric surgery showed that the most common mental health condition was depression [6]. The presence of premorbid depression may result in compromised weight loss. Kalarchian et al. reported an association between higher rates of lifetime mood and anxiety disorders were associated with less weight loss in the first 6 months after surgery [7]. Similarly, de Zwaan et al. showed that preoperative depression was a predictor of postoperative depression and was associated with less weight loss late 24–36 months follow-up period [8]. However, Dawes et al. have found no association between preoperative depression and weight outcomes in their meta-analysis [6].

Depression appears to be related to obesity in a complex way. Depression can increase the risk for obesity; while obesity can trigger depression perhaps attributable weight-related problems. Obese individuals have a 55 % increased risk of developing depression over time, whereas depressed persons have a 58 % increased risk of becoming obese [9].

Depression is also associated with problematic eating behaviors, which may contribute to the development of obesity. Depressed people at times consume excessive amounts of food, perhaps as a coping mechanism [10–12]. According to one theory of affect regulation, depression, and emotional eating are considered mediators of obesity. Affect regulation theory suggests that there is also a relationship between emotional eating and depression. Heatherton and Baumeister proposed that eating is motivated by a desire to avoid unpleasant affective states. Depression has also been reported to be associated with external eating [13], that is defined as being more sensitive to external food cues, regardless of their internal state of hunger and satiety [14].

Another concept of relevance is restrained eating, first developed by Herman and Polivy [15]. This idea suggests that the desire for food may result in obstinate dietary restrictions [15] as a means to cope with this desire. However, disinhibition theory postulates that strong emotional states such as depression and anxiety can disrupt restrictive eating. Restraint theory also proposes that dietary restriction induces counter-regulatory responses, reduces sensitivity to satiety signals and can result in disinhibited, problematic overeating. Studies have also shown that those who demonstrate a pattern of restrained eating also evidence greater depression, lower self-esteem, and greater emotional responsivity [15, 16].

In considering the possible role of depressive symptoms in the development of problematic eating, there is a need to understand better the relationship between depressive symptoms and problematic eating behavior in bariatric surgery candidates. To our knowledge, depressive symptoms and problematic eating behaviors have never been studied in Turkish bariatric surgery populations.

The purpose of this study was to explore whether depression and body mass index (BMI) have an impact on eating patterns (emotional, restrictive or external eating) among bariatric surgery candidates in a Turkish sample. In the light of previous studies in non-surgical populations, we hypothesized that problematic eating behaviors would increase with the severity of depression in the sample of bariatric surgery candidates.

Materials and methods

Study population

This descriptive study was conducted using 168 consecutive bariatric surgery candidates evaluated between Jun 2012 and Nov 2013. The Ethical Review Board of the Medical Faculty approved the study. Informed consent was obtained from all participants prior to study participation. Participants were asked to complete the Dutch Eating Behavior Questionnaire (DEBQ), the Beck Depression Inventory (BDI), and the data forms requesting information concerning sociodemographic and clinical variables. Seventeen patients were not included in the final analyses, due to missing data.

Measurements

The Beck Depression Inventory (BDI)

BDI is a 21-item self-report inventory for the measurement of emotional, somatic, cognitive, and motivational symptoms seen in depression [17, 18]. The purpose of the scale is to determine the severity of depressive symptoms. Items in the scale are rated between 0 and 3, according to the severity of depression. The highest score obtainable is 63. Participants were subtyped by presence or absence of depressive symptoms according to a BDI cutoff of 14, indicating “probable depression”, and scores were labeled 0–12 for minimal, 13–18 for mild, 19–28 for moderate and 29–63 for severe depression [19]. The BDI has good reliability and validity [20, 21].

Dutch Eating Behavior Questionnaire (DEBQ)

The DEBQ was used for assessment of restrained, emotional and external eating [22]. This self-report measure contains three scales: “restraint eating”-10 items, “external eating”-10 items and “emotional eating”-13 items. All items were rated on a five-point scale from 1 (never) to 5 (very often). The restraint scale measures intentions to restrict food intake and control of food intake. Emotional eating corresponds to the tendency to overeat in response to negative emotions. External eating corresponds to the tendency to overeat in response to food-related stimuli. Each of these subscales has good internal consistency and accurate factorial analyses [23]. The DEBQ was adapted into Turkish by Bozan et al., with a clear factorial validity and the reliability. The internal consistency of Turkish DEBQ subscales has been found 0.9 for external eating, 0.9 for emotional eating, and 0.9 for restrained eating subscales. (Cronbach’s alpha > 0.8) [24].

Sociodemographic and clinical data form

Demographic and clinical data were obtained through the use of a standardized self-report questionnaire, developed by the authors. The questionnaire includes information on age, sex, marital status, education, employment status, socioeconomic status, psychiatric history, medical history, suicide history and family history of obesity.

Body Mass Index (BMI)

Body weight and height were measured (without shoes, in clothes) and BMI (kg/m2) was calculated.

Statistical analysis

All statistical analyses were performed using SPSS 21.0 statistical software program. For evaluation of the data, descriptive statistical methods (frequency, percentage, mean, and standard deviation) were used. Pearson correlation analysis was used to determine the relationship between BDI scores, BMI, DEBQ scores and its subscales. To address the aims of this study, linear regression analyses were performed with each one of the eating styles as dependent variables. The significance level was set at 0.05.

Results

One hundred sixty-eight patients (male; n = 35; 20.8 %, female; n = 133; 79.2 %) were included in this study. On average, the participants’ age was 37.7 years (SD = 11.25; range 18–68), and mean BMI was 46.4 kg/m2 (SD = 6.7; range 31.2–67.1). The mean duration of education was 11.4 years (SD = 4.7). Participants of this study were primarily from the middle socioeconomic class depending on the participants’ claims (n = 98; 59 %) and were married (n = 102; 60.7 %).

DEBQ Total scores and sub-scores (emotional, external and restrictive eating) did not show significant difference between gender group (p > 0.05). DEBQ Total scores and sub-scores (emotional, external and restrictive eating) did not differ significantly regarding their marital status and socio-economic level. BMI and age of the subjects did not show any correlations with the DEBQ Total and sub-scores.

Description of the problematic eating behaviors (restrictive, external and emotional) as measured with DEBQ and level of depression of the entire sample is shown in Table 1. On average, 75.5 % of the participants reported from mild to severe level of depressive symptoms and 29.8 % of participants reported moderate-to-severe depressive symptoms. Analysis of the correlations between DEBQ sub-scores and BDI scores are shown in Table 2. Restrictive eating scores showed a negative correlation (r = −0.17; p = 0.04) with BDI scores. Conversely, emotional eating (r = 0.3; p = 0.002) and external eating (r = 0.2; p = 0.04) were found to be positively correlated with the BDI scores.

Table 1 Demographic variables, clinical variables and description of the eating behaviors according to Dutch Eating Behavior Questionnaire
Full size table
Table 2 Correlation analysis of the Beck Depression Inventory and Dutch Eating Behavior Questionnaire scores
Full size table

Separate regression analyses were performed to analyze the effect of the BMI and BDI scores on all sub-scores of DEBQ (Table 3). The model for evaluating relationship between BDI, BMI and DEBQ Restrictive Eating sub-scores was not statistically significant (F = 2.069; p = 0.13, R 2 = 0.014).

Table 3 Regression analysis results of the variables
Full size table

The model for assessing the relationship between BDI, BMI and DEBQ Emotional eating sub-scores was significant (F = 4.96; p = 0.008), in which the BMI (β = 0.001, p = 0.10) were not found to be associated with DEBQ Emotional eating scores. BDI scores significantly increased with increasing DEBQ Emotional eating scores (ß = 0.03, p < 0.01). These independent predictors together weakly explained the variance (R 2 = 0.1).

The model for assessing the relationship between BDI, BMI and DEBQ External eating sub-scores was significant (F = 3.1; p = 0.02), in which the BMI (β = −0.02, p = 0.07 > 0.1) were not associated with DEBQ External eating scores. BDI scores significantly increased with increasing DEBQ External eating scores (ß = 0.03, p < 0.02). The variance was only weakly explained by these independent predictors (R 2 = 0.04).

The model for assessing the relationship between BDI, BMI and DEBQ Total scores was significant (F = 3.3; p = 0.04 < 0.05), in which the BMI (β = −0.01, p = 0.33 > 0.15) were not associated with DEBQ Total scores. BDI scores significantly increased with increasing DEBQ Total scores (ß = 0.01, p < 0.01). Variance was weakly explained by these independent predictors (R 2 = 0.03).

Discussion

In this study, we found that 62.25 % (n = 94) of 151 eligible patients had probable depression according to the BDI cutoff point of 14. Problematic eating behavior and psychopathology including eating disorders are more prevalent among bariatric surgery candidates than other obese individuals who seek treatment. Depression is a common comorbid condition in patients seeking bariatric surgery [25, 26]. Mitchell et al. found that the most common lifetime psychiatric diagnosis was “any affective disorder” with a rate of 44.2 %. The most common affective disorder was major depressive disorder with a rate of 38.7 % in preoperative bariatric surgery sample [5]. The findings of this study support the high prevalence of depression among bariatric surgery candidates.

It is important to evaluate the patients both before and after the surgery with regard to current depression and eating pathologies. Eating disorder symptoms in response to negative emotions and non-restrained eating behavior can result in weight gain and worsening of the existing depression. Accordingly, treating the depression, educating patients about coping with negative emotions and resisting food cues may contribute to the improvement of the eating pathology [9, 27]. Vinai et al. suggested that a combination of high disinhibition and low restraint has been associated with a higher susceptibility to weight gain and sedentary behavior [28]. There are many other studies available emphasizing that high-level cognitive control, low-level disinhibition and accurate competing ability with negative emotions are critical in long-term weight loss and protection of weight [11, 29].

This is the first study to examine the relationship between depressive symptomatology and eating behavior in a large sample of Turkish bariatric surgery candidates. As hypothesized, patients with higher depression scores endorsed higher eating behavior pathology than those with lower depression scores.

BMI of the subjects did not show any correlation with the DEBQ Total score and sub-scores. The results are somewhat inconsistent with restrained eating theory. Discrepancy could be explained by diverse characteristics of the studied samples. In the present study, the sample was relatively homogeneous with an average BMI of 46.4 kg/m2.

It has also been shown that external eating and emotional eating are related to overeating and weight gain [30]. Mason et al. have found the more emotional and external eating scores, more depressive symptoms and more binge eating among undergraduate students with a higher BMI [31]. Although we couldn’t detect any relation between DEBQ and BMI in bariatric population, studies indicate this relation in other groups such as adolescents.

We found a negative correlation between Beck Depression Inventory score and DEBQ Restrained scores. This may reflect that strong emotions may disinhibit those with restrained eating according to the restraint hypothesis. It has been shown that among clinically depressed patients, those with unrestrained eating reported a significant weight loss and while those with restrained eating demonstrate significant weight gain after the onset of depression [32]. Thus, disinhibition of restrained eating behaviors may play a mediatory role in weight gain.

We found a positive correlation between Beck Depression Inventory score and DEBQ Emotional Eating scores. (r = 0.252; p = 0.002), and it has been suggested that there is a relationship between depression and emotional eating. Studies have shown that emotional eating is associated with obesity and depression. Also, emotional eating mediates the relationship between obesity and depression [29, 33]. Fereidouni et al. found that compared to normal weight individual, obese women had more difficulties in controlling their impulses, engaging in goal directed behaviors and accessing functional strategies while experiencing negative emotions [34]. Emotional eating is quite common, especially in obese individuals who seek treatment. Emotional eating has been found to be predictive of negative weight outcomes after bariatric surgery. Canetti et al. reported that there is a relationship between emotional eating and postoperative snacking and grazing [30, 35]. Thus, the significant correlation between emotional eating and depression underscores the prognostic value of the proper depression management in bariatric surgery patients.

According to our results, Beck Depression Inventory score was related positively with DEBQ_X (r = 0.173; p = 0.035). According to the Schachter’s obese-externality theory, internal sensation, depression, anxiety or fear have effects of food consumption suppression in normal-weighted individuals, but not in obese individuals due to their insensitivity to endogenous stimuli [36]. These individuals are more sensitive to the external stimuli of food, and they eat with response to these external stimuli rather than the feeling of hunger or fasting. According to the Escape theory of Heatherton and Baumeister [13], some individuals limit their attention with environmental current and immediate stimuli to stay away from negative effects such as depression. As in our findings, increased depression and external eating behavior are important factors in weight gain of these individuals. In a study performed by Jansen et al., it was concluded that overweight, obese and more-depressive individuals are prone to external eating pattern with negative mood stimulation [37]. Zijlstra et al. also suggested that external eating was associated with negative affect and with difficulty in identifying feelings in morbid obese women [38], although these correlations were not strong and no longer significant after adjustment for the other eating behaviors.

This study has several limitations. The first limitation is the descriptive design, which limits confidence in making causal inferences between depression and problematic eating patterns. Second, DEBQ scores could be affected by the non-measured eating pathologies such as binge eating or night eating syndrome [39]. Also, the expression power of the regression model is small. These independent predictors weakly explained the variance (R 2 = 0.051). Finally, depression screening instruments such as BDI provide the severity of depressive symptoms at a given time, but they do not diagnose depression.

Conclusion

This study emphasizes that there is a positive correlation between depression level of the bariatric surgery candidates and their emotional/external eating behaviors, and a negative correlation between their depression levels and restrictive eating behavior. Proper management of the depression in bariatric surgery candidates may reduce problematic eating behaviors as well as improve weight outcomes. Future prospective studies investigating the recognition and appropriate treatment of the depression in bariatric surgery candidates are needed.