Introduction

Sleeve gastrectomy (SG) is the most common bariatric surgery (BS) type performed today; however, data on its impact on psychological outcomes and eating behaviors are scarce [1]. Obesity state is related to significant psychological difficulties including depression, anxiety disorder, emotional eating, and maladaptive eating behaviors including binge eating disorder (BED) [1]. Binge eating is highly prevalent among candidates for BS, and has been reported in the range of 6–50% [1, 2]. Presently, most data on binge eating in relation to BS are derived from studies on Roux-en-Y gastric bypass and adjustable gastric banding [3, 4], while only a few studies evaluated binge eating among individuals who underwent SG [1, 2, 5]. Most of the relevant studies published to date reported reductions in binge eating symptoms post-surgery [3, 4]; however, at longer follow-up, binge eating has been shown to revert and even return to pre-surgery levels [1]. Food addiction is characterized by craving and consumption of highly palatable foods (e.g., processed foods high in fat and sugar) [6]. Although not formally recognized by the DSM-5, food addiction (FA) has been well described in the scientific literature [7]. Food addiction shares some clinical features with binge eating and is also common among individuals with obesity [8]. Food addiction has been found to be associated with psychosocial impairment including depression, impulsivity, eating disorder psychopathology, and poorer self-control [6]. The Yale Food Addiction Scale (YFAS) is most often used to conceptualize and measure food addiction [6, 9]. The prevalence of food addiction has been reported as up to 42% in candidates for BS, and up to 57% in adults with obesity and binge eating [9, 10]. A paucity of studies examined food addiction prospectively among individuals who underwent BS, both before and following the surgical procedure [6, 10,11,12]. This prospective study aimed to examine the prevalence of food addiction and binge eating before and 1 year following SG surgery, and to assess possible associations of food addiction and binge eating with weight loss outcomes and eating/behavioral patterns postoperatively.

Materials and Methods

Subjects

This prospective cohort study is a part of a randomized clinical trial of 2-month treatment with preoperative vitamin administration vs. standard pre-surgery care and 12-month follow-up, aimed to assess bone density among 62 women who were candidates for SG. The data presented in this study are for the combined treatment groups, since differences between them were not observed for the measurements discussed here. Inclusion criteria were age 18–65 years and BMI ≥ 40 kg/m2 or BMI ≥ 35 kg/m2 with comorbidities. Major exclusion criteria included untreated mental illness or unstable mental state, pregnancy, lactation, chronic conditions affecting bone metabolism and previous BS. All the procedures performed in this study were approved by the institutional review boards of the participating hospital (Hadassah Ein Kerem Medical Center) and all the participants signed an informed consent. The study was pre-registered in the NIH registration website (identifier: NCT02483026). Participants were evaluated during January 2018 to January 2019 at baseline (2 months preoperatively, M0), 3 (M3), 6 (M6), and 12 (M12) months postoperatively.

Anthropometrics

Weight was measured on a digital medical scale; height was measured by a stadiometer and waist circumference was measured twice according to a uniform protocol. BMI was calculated by weight (in kilograms) divided by the height squared (in square meter). Percentages of excess weight loss (% EWL) were calculated as follows: [(preoperative weight-postoperative weight) / (preoperative weight-ideal weight)] × 100. Ideal body weight (IBW) was considered the weight for BMI 25 kg/m2. Percentages of total body weight loss (TBWL) were calculated as follows: ([preoperative weight-postoperative weight] / [preoperative weight]) × 100.

Dietary Intake, Food Tolerance, and Physical Activity Evaluation

Participants filled a 3-day food diary to assess macronutrient and micronutrient consumption from food. This was computed with the Israeli nutritional software “Zameret.” Participants were asked to complete a physical activity (PA) diary that assessed the frequency, type, and duration of each activity, for a 7-day period [13]. Additionally, patients filled a questionnaire for quick assessment of food tolerance after BS, with a score of 1 to 27 (a high score represented high food tolerance) [14]. Translated versions of this questionnaire have been used in studies in Israel [15].

Binge Eating Scale and Food Addiction Assessment

The Binge Eating Scale (BES) [16, 17] is a 16-item instrument that is widely used as a self-report measure of binge eating severity [18]. In the context of individuals who underwent BS, the instrument showed high internal consistency (Cronbach’s alpha = 0.90) [19], as well as high sensitivity and adequate specificity in classifying persons with and without binge eating [18]. For the purpose of this study, a frequent previously used adaptation was used, in which a score greater than 17 classifies “binge eaters” [17]. Food addiction was assessed by the YFAS, a 25-item self-report measure that assesses symptoms of food addiction [9]. The YFAS is scored both dimensionally (a “symptom” score, based on the number of symptoms endorsed) and dichotomously (a “diagnosis,” by endorsing three or more symptoms and clinically significant impairment or distress) [3, 9]. The YFAS has demonstrated validity and reliability in bariatric candidates and post-surgical patients [20]. Hebrew and Arab translations of the original validated BES questionnaire and the YFAS were attained in steps, according to developed methodologies [21].

Statistical Methods

All the statistical analyses were performed using SPSS version 25.0. If normality was rejected by the Kolmogorov–Smirnov test, nonparametric tests were used. Continuous variables are presented as means ± SD or median and IQR, as appropriate; and dichotomous/categorical variables as proportions. To test differences in continuous variables between 2 groups, the independent samples t test or the Mann-Whitney test were performed. For comparison of dichotomous or categorical variables, the Pearson Chi-Square test was performed. To compare continuous variables between 2 time points, the paired t test was performed or the Wilcoxon test when needed. For dichotomous/categorical variables, McNemar’s test was performed. A mixed-model repeated-measures analysis was conducted to examine associations of anthropometrics and other variables with weight loss and behavioral outcomes.

Results

Characteristics of the Participants

Of the 62 women who enrolled in the study, 4 eventually did not undergo SG surgery and 4 did not complete the 12-month follow-up post-surgery. Therefore, the final sample included 54 women who had anthropometric measurements at the M12 follow-up visit, and the dropout rate was considered 12.9%. Preoperative characteristics according to diagnosis of food addiction and binge eating at baseline are described in Table 1.

Table 1 Characteristics of women who underwent sleeve gastrectomy according to food addiction (FA) and binge eating (BE) status at baseline (n = 54)
Full size table

Prevalence of Food Addiction and Binge Eating Categorization Before and After SG

Before SG surgery, food addiction was identified in 22 (40.7%) and binge eating in 26 (48.1%) participants. All patients with food addiction at baseline (n = 22) have also met the criteria for binge eating; thus, 40.7% of the preoperative sample presented both features. Following SG, food addiction was reduced at M3 (10.2%) and M6 (6.7%); however, at M12, 29.3% of the women met the criteria for food addiction (P > 0.05 compared with baseline). The prevalence of binge eating was significantly lower at all post-SG time points compared with baseline, and was 17.4% at M12 (P < 0.05 compared with baseline) (Fig. 1). At M12, 14.3% of the women met the criteria for both food addiction and binge eating. Moreover, food addiction and binge eating were resolved at M12 among 66.7% and 86.4% of women who met the criteria for food addiction and binge eating at baseline, respectively.

Fig. 1
figure 1

Prevalence of food addiction and binge eating before sleeve gastrectomy, and at 3, 6, and 12 months (M3, M6 and M12) postoperative. Food addiction measured by the Yale Food Addiction Scale (YFAS) (a) and binge eating measured by the BES (b). *Significantly different from baseline, P < 0.05. Food addiction was measured by the Yale Food Addiction Scale (YFAS) and was available for: n = 54 at baseline, n = 49 at M3, n = 45 at M6, and n = 41 at M12. Binge eating was measured by the BES questionnaire and was available for: n = 54 at baseline, n = 49 at M3, n = 46 at M6, and n = 46 at M12

Full size image

Anthropometrics, Physical Activity, and Dietary Intake According to Binge Eating and Food Addiction Categorization at Baseline

Table 2 presents anthropometric and lifestyle parameters at baseline and at M3, M6, and M12, according to the presence of food addiction and binge eating before SG. BMI was significantly lower at M12 than at baseline, for women with and without food addiction at baseline, with no differences between these groups (P time*group = 0.405). However, BMI was stabilized at M6 among women with food addiction at baseline, while it continued to decrease from M6 to M12 among those without food addiction (Table 2). The same trend was noted for BMI during the postoperative year, comparing women with and without binge eating at baseline. The reported hours spent in weekly PA increased significantly from baseline at all time points among women without food addiction and women without binge eating. However, those with food addiction or binge eating at baseline reported significantly less time in PA at M12 (P = 0.041, P = 0.006, respectively). The mean food tolerance score decreased significantly from baseline to M3, M6, and M12 for all the participants. However, scores were significantly lower at M12 among those with than without food addiction, and with than without binge eating at baseline (P = 0.020, P = 0.024, respectively).

Table 2 Anthropometrics, health, and eating behaviors in the year following sleeve gastrectomy, according to baseline diagnosis of food addiction (FA) and binge eating (BE)
Full size table

Characteristics of Women Who Met Criteria for Food Addiction and Binge Eating at 12-Month Postoperative

Characteristics of the cohort are presented in Table 3 and Fig. 2, according to meeting the criteria of food addiction and binge eating at M12. Women who met the criterion of food addiction at M12 reported spending fewer weekly hours in PA and consuming greater daily amounts of carbohydrates than did those who did not meet this criterion at M12 (P < 0.001 and P = 0.013, respectively). Women without food addiction achieved significantly higher %EWL and %TBWL, and lower mean BMI at M12 than did women who met the criterion for food addiction at M12 (P = 0.005, P = 0.005, and P = 0.003, respectively, Table 3). Food addiction score at M12 was negatively correlated with weekly PA (r = − 0.559, P < 0.001), food tolerance scores (r = − 0.428, P = 0.005), and weight loss from M6 to M12 (r = − 0.342, P = 0.041). Additionally, women who met the criterion of binge eating at M12 reported significantly higher rates of vomiting at M12 (50.0 vs. 13.2%, P = 0.017).

Table 3 Anthropometrics, health, and eating behaviors in the year following sleeve gastrectomy, according to food addiction (FA) and binge eating (BE) at 12-month postoperative (M12)
Full size table
Fig. 2
figure 2

Excess weight loss (EWL) and reported physical activity 12 months following sleeve gastrectomy, according to food addiction and binge eating categorization at 12 months post-surgery. Excess weight loss (a) and weekly physical activity (b). EWL, excess weight loss; FA, food addiction. Food addiction was measured by the Yale Food Addiction Scale (YFAS) and was available for n = 41 at M12

Full size image

Discussion

This prospective study of women who underwent SG found different trends between binge eating and food addiction over the postoperative year, and different associations of these classifications with weight loss outcomes and eating and lifestyle behaviors. About half the women met the criteria of binge eating prior to surgery according to the BES questionnaire; this decreased to less than one-fifth at 1-year postoperative. In contrast, food addiction prevalence, as measured by the YFAS, decreased from 41% at baseline to 7% at 6 months postoperative, but rose to 29% during the subsequent 6 months. Women who met the criteria for food addiction or binge eating at baseline reported spending less time in physical activity and their food tolerance scores were significantly lower at 1-year postoperative, compared with women without food addiction or binge eating at baseline. Women with food addition at 1-year postoperative had lower weight loss outcomes, including lower %EWL and %TBWL, and higher mean BMI. Although BS is generally effective in reducing body weight, a substantial proportion of individuals achieves suboptimal weight loss. Thus, identifying modifiable prognostic indicators for postoperative weight outcomes is important [6]. The existing literature on potential predictors of success following BS is far from conclusive [10]. The diverse factors that have been investigated for their effect on postoperative outcomes include various problematic eating behaviors such as binge eating [6]. Our results corroborate previous reports that described a high prevalence of binge eating among candidates for BS, up to 50% [1]. Our findings also concur with studies that showed reductions during the first postoperative year in binge eating after different types of BS, including SG [1, 2, 5, 22]. This is not surprising since BS procedures anatomically restrict the capacity to binge eat or overeat. However, at longer-term follow-up, binge eating symptoms may revert and even return to pre-surgery levels [1]. In the current study, women who were categorized with binge eating at 1-year postoperative also had a higher prevalence of vomiting than did women without this categorization. Vomiting may result from overeating or attempts to binge eat [1]. Vomiting commonly occurs postoperatively, both acutely and involuntarily, or in response to dysphagia [23]. Self-induced vomiting related to concerns of body weight and shape are considered to appear more rarely following BS, though these might reflect compensatory behaviors and reflect the development of new eating disorders post-surgery [23]. In the current study, the categorization of binge eating, either at pre-surgery or post-surgery, was not related to BMI or %EWL outcomes at 1-year postoperative. However, focusing on binge eating findings alone may be misleading as other maladaptive eating patterns can emerge post BS, such as grazing, loss-of-control eating, and vomiting [1, 3]. The current study did not investigate these. Notably, binge eating may evolve over time, as the stomach can enlarge in the long-term following BS [1]. The prevalence of food addiction prior to SG reported in the current study is within the range reported by others, 14–58% [6]. Moreover, our finding of a decrease in food addiction concurs with reports of lower rates of food addiction postoperatively, ranging from 2 to 14 [10, 12]. However, our postoperative rate at 12 months was closer to the rate of one-third demonstrated in a cohort of 44 individuals following BS [12]. We report that baseline food addiction was not related to %EWL or BMI post-surgery. Similarly, a few prospective studies reported that food addiction prior to BS was not associated with weight loss outcomes up to 12 months post-surgery [10, 12, 24]. However, BMI was stabilized at 6-month postoperative among the women of our cohort who were categorized with food addiction at baseline, while continuing to decrease up to 12-month postoperative among women without food addiction at baseline. Moreover, food addiction categorization at 1-year postoperative was related to poorer weight loss outcomes. We report that meeting the criteria of food addiction classification pre- and post-surgery was related to less hours reported for PA and with lower food tolerance scores at 1-year postoperative. This corroborates associations that have been demonstrated of food addiction prior to BS with various post-surgical psychopathologies and specific eating/lifestyle behaviors [6]. However, the current study did not examine changes in other related psychopathologies and behaviors that were previously associated with food addiction, such as night eating syndrome [25], emotional eating [20], depression, and quality of life [24]. It is important to note that food addiction is still considered a controversial concept, thus not formally integrated into the DSM-5 [8]. Generally, there appears to be an excessive conceptual heterogeneity within behavioral addictions, and lack of agreement as to how various repetitive and problematic behaviors should be conceptualized [26]. Moreover, addiction to eating or food is different from other behavioral addictions, and perhaps more akin to substance addictions because of the ingested ‘substance’, i.e., food [26]. Nevertheless, emerging evidence links food addiction to different subthreshold psychopathological domains (e.g., depressive symptoms and eating subthreshold symptoms), suggesting that it might be a dimension underlying different conditions or symptom clusters [6, 27]. Overall, more rigorous research methods with larger and more diverse patients are needed to better understand the clinical utility and validity of food addiction following BS [6]. Importantly, addictive eating type and food addiction in the context of BS have been investigated in greater depth in recent years, though clinical implications, such as the integration of pre-BS psychological evaluations, have not become routine practice [6]. Strengths of the present report include its being one of the few documented studies to prospectively examine food addiction by YFAS at pre- and post-SG. However, our study has several limitations. First, the sample size included only women; thus, generalizability of results may be limited. Nevertheless, women comprise more than 70% of the persons who undergo BS worldwide [28]. Moreover, previous reports have shown associations of female sex with eating disorders and addictive behaviors in various populations, including persons who undergo BS [29]. Furthermore, the relatively small sample size made it difficult to perform further subgroups analysis, such as whether outcomes differed between women whose initial diagnosis of food addiction or binge eating was to those who initial diagnosis was not resolve at M12. The second limitation stems from the study design, which did not assess associations of food addiction and binge eating on weight outcomes beyond 12 month, when weight regain might occur [10]. Third, some of the measures in this study were based on patients’ self-reports (e.g., PA and eating patterns); nevertheless, this study included mostly validated assessment tools (e.g., the YFAS, BES, and food tolerance questioners). Finally, examining changes in other related psychopathologies and behaviors that were previously associated with food addiction, (e.g., depression and decreased quality of life) could contribute to bridge the gaps in existing knowledge on the clinical implications of food addiction in the context of BS.

In conclusion, this study suggests that women may experience sustained reductions in food addiction in the short-term of up to 6 months following BS. However, this reduction was not maintained at M12, and was accompanied by poorer weight and behavior-related outcomes. Additional follow-up studies are required to evaluate longer-term effects of pre- and postoperative food addiction and binge eating following BS. Furthermore, clinical practice should be focused on the psychological aspects associated with obesity, as these may affect postoperative outcomes.