Disordered eating after bariatric surgery: clinical aspects, impact on outcomes, and intervention strategies

Abstract

Purpose of review:

Disordered eating behaviors (DEBs) are associated with poor weight outcomes following bariatric surgery. We describe DEBs most relevant to this population, their associations with weight outcomes, and emerging data on interventions for DEBs.

Recent findings:

Loss of control eating episodes and grazing have been the most well studied DEBs in bariatric samples. Although DEBs often remit after surgery even without targeted intervention, a subgroup of patients have persistent or newly developed DEBs post-operatively. Pre-operative DEBs have little effect on weight outcomes, while pre-operative impulse control-related features commonly associated with DEBs (e.g., inhibitory control) may have stronger predictive value. Post-operatively, DEBs appear to exert robust effects on concurrently measured weight. Post-operative interventions hold promise for optimizing treatment outcomes.

Summary:

We recommend the following to improve clinical care and move research forward: 1) a common language for DEB constructs is needed to improve cross-talk among researchers and care providers; 2) diagnostic schemes and assessment tools may require tailoring for the bariatric population; 3) mechanisms underlying improvements in DEBs following surgery should be clarified; 4) ongoing monitoring of DEBs in the post-operative period is warranted; and 5) a stepped-care approach may improve weight outcomes in a cost-effective manner.

Introduction

Bariatric surgery (BS) is currently the most effective treatment for obesity. Although most patients experience marked weight loss following surgery, a subset present with suboptimal weight loss and/or weight regain in the long-term [1]. Attempts to identify risk factors for poor post-operative outcomes have focused on disordered eating behaviors (DEBs), which are relatively common in pre-surgery samples, often produce imbalances between energy intake and expenditure, and may be associated with post-operative onset of full-syndrome eating disorders [2]. This review will provide an updated description of DEBs common to the bariatric surgery population and their relation to weight outcomes; propose a clinical management approach based on the latest empirical findings; and suggest priorities for future research.

Clinical presentation

Anatomical alterations produced by surgery place several physiological constraints on eating behavior. As a result, constructs, measures, and diagnostic categories originally developed for non-bariatric eating disorder patients may not adequately address the clinical presentation of DEBs among post-surgery patients [3]. Moreover, the eating disorders and obesity fields are plagued by a lack of consistency when describing and labeling eating-related constructs, resulting in considerable overlap among constructs [e.g., food “addiction,” compulsive eating, loss of control eating (LOCE)]. In addition to limiting cross-study comparisons, this lack of consistency impedes recognition of DEBs and appropriate referrals/continuity of care by clinicians.

Research on the development of eating disorders after surgery is scarce, and there are no data on prevalence rates of eating disorders other than binge eating disorder (BED) in post-operative samples. The published reports in this area suggest that a) post-bariatric patients with eating disorders tend to be older than non-bariatric patients with these disorders which may indicate that eating disorder symptoms/risk factors predate surgery; b) what constitutes a “low” BMI in this population is debatable, as several post-surgical patients with anorexia nervosa-like symptoms present with a BMI in the normal weight range despite extreme dietary restraint and substantial weight loss; and c) surgery-specific gastrointestinal symptoms (e.g., vomiting or diarrhea experienced as “dumping” syndrome) may function as compensatory behaviors to avoid weight gain [2].

BED involves recurrent episodes of objective binge eating (OBE; consumption of an unambiguously large amount of food in a discrete time period, accompanied by LOCE) in the absence of inappropriate compensatory behaviors [4]. Approximately 12.7% of pre-surgery patients present with BED [5**], marking it the second most common psychiatric disorder in this population [6**].

In the post-operative period, anatomical alterations induced by the surgical procedure, resulting in limited gastric capacity, make it impossible for most patients to accommodate large amounts of food. Thus, BED is rarely diagnosed after surgery since the primary criterion (ingestion of an unambiguously large amount of food in a discrete time period) is physically unfeasible. As a result, there is evidence for a significant decrease in BED diagnoses after surgery [5**], which suggests a post-operative improvement of eating behavior. Nevertheless, several studies demonstrate that episodes involving LOCE over smaller amounts of food (subjective binge eating – SBE) may continue or emerge shortly after surgery, increasing in frequency with follow-up time [5**,7*,8]. Research suggests that LOCE is a unique indicator of psychopathology, and that the amount of food eaten reflects the degree of LOCE experienced during the episode [9,10**]. Indeed, post-bariatric patients meeting the criteria for BED, except for endorsing SBE rather than OBE, show comparable levels of eating-related and depressive symptomatology relative to non-BS individuals with full-syndrome BED [7*]. Our group [10**] used a continuous rating scale to assess the degree of LOCE in relation to four DEB, including OBE and SBE. We found that psychopathology across the DEBs were more strongly related to the degree of LOCE than amount of food eaten during eating episodes. At the same time, OBE and SBE are associated with similar levels of psychopathology, while the degree of LOCE is highly correlated with the amount of food eaten in laboratory tests [9].

Recent research indicates that LOCE among bariatric surgery patients may be present in the context of eating behaviors outside of OBE or SBE episodes, including grazing [10**]. Grazing involves eating small/modest amounts of food in an unplanned and repetitious manner, and not in response to hunger/satiety sensations. Researchers have described two subtypes of grazing, compulsive (related to one’s inability to resist eating) and non-compulsive (related to the experience of mindless eating). These features, along with the unplanned nature of grazing, suggest that a certain degree of LOCE may be present [11].

Our research suggests that OBE/SBE and compulsive/non-compulsive grazing among bariatric patients could be conceptualized along a continuum of LOCE and eating-related psychopathology, whereby non-compulsive grazing is associated with the lowest levels of LOCE and psychopathology, and OBE with the highest (resulting in larger amounts of food eaten during the episode) [9,10**]. Thus, assessing LOCE on a dimensional scale rather than dichotomously (present/absent) may clarify different behavioral presentations of DEBs in this population.

Stability and course

Although DEBs typically decrease immediately following surgery, they may emerge or re-appear as early as 4–9 months following surgery [7*], with rates continuing to increase over follow-up [5**,12] and peaking at 3 years after surgery [5**]. Recent data from the large, multisite Longitudinal Assessment of Bariatric Surgery (LABS) study showed that the presence of BED and LOCE decreased from 12.7% and 35% pre-operatively, to 2.1% and 24% one year after surgery, respectively [5**], with remittance rates reaching 70% for BED vs. only 27% for LOCE. Thus, LOCE may be more stable than full-syndrome BED. Indeed, 46.6% of those whose BED remitted continued to report sub-syndromal LOCE after surgery. Notably, 4.8% and 25.6% of LABS participants reported post-operative BED and LOCE de novo, respectively, indicating that while risk for post-surgical LOCE is highest for individuals who engaged in DEBs prior to surgery, a small subset of patients evidence new onset of these behaviors after surgery. With respect to other DEBs, we reported that picking/nibbling, an eating behavior resembling grazing, persists from pre- to post-surgery for 43.8% of the sample. Half of those whose LOCE remitted post-surgery engaged in picking/nibbling two years after surgery [13], suggesting that transition among DEBs is not uncommon.

Impact on weight outcomes

Identifying pre-surgical DEBs predictors of weight outcomes could help clinicians screen patients at risk for poor outcomes and inform personalized care. Unfortunately, identification of pre-surgical prognostic indicators has proved elusive [14], and recent research sustains the mixed findings.

Recently, Marek and colleagues [15] found that pre-operative BED was associated with higher post-operative BMI, but not BMI reduction over 5 years of follow-up. Similarly, García-Ruiz-de-Gordejuela and colleagues [16] showed that patients with slower weight loss trajectories in the year following surgery presented with more pre-operative psychopathology, including DEBs. Another group showed that pre-operative emotional eating was associated with lower weight loss one year after surgery, and “food addiction” symptoms with lack of weight loss from 3 to 12 months post-surgery [17]. Unfortunately, none of these studies controlled for the presence of concurrent DEBs after surgery, such that associations with weight outcomes may be accounted for by DEBs that persist among the subset of patients who reported pre-surgery DEBs [5**,13].

The latest reports from LABS [18] found that pre-surgery eating disorder diagnoses were not associated with weight change over 7 years of follow-up. In addition, patients presenting with pre-surgical LOCE had similar 7-year variations in weight loss relative to those with no pre- or post-surgical LOCE [12]. In their investigation of weight trajectories over 2 years following surgery among patients who underwent primary or reoperative surgery [19], Pinto-Bastos and colleagues found that none of the pre-operative disordered eating constructs (eating-related psychopathology, grazing, or OBE/SBE) was a significant predictor of weight loss or regain. Using latent class analysis, Schäfer and colleagues identified different disordered eating subtypes among bariatric surgery candidates, none of which differed in terms of weight loss [20]. Finally, our research showed that patients presenting with pre-operative DEBs did not differ in weight loss trajectories across 30 months of follow-up [13]. A recent literature review found that pre-operative grazing was not associated with post-operative outcomes [21]. Together, these data highlight the limited prognostic value of pre-operative eating disorder diagnoses or subclinical DEBs in relation to weight loss after surgery.

In response to the lack of support for pre-operative DEBs as predictors of post-surgical weight outcomes, several investigators have suggested that personality features [22] such as impulsivity [6**] may account for overlap among DEBs, and may have greater utility as outcome predictors than specific behaviors [22]. Indeed, pre-operative inhibitory control has been shown to predict variability in BMI as early as 6 months after surgery [23], although it is unclear whether effects are mediated by changes in self-regulation of eating [24]. Future research should investigate how impulse control constructs vary across time, and clarify mechanisms of action by which these constructs impact post-operative eating and weight outcomes.

The literature has been far more consistent regarding the impact of post-operative DEBs on weight outcomes. Post-operative eating-related psychopathology has been associated with weight loss and regain, and post-operative grazing with weight regain for primary and reoperative patients [19]. In addition, evidence suggests an association between post-operative grazing and weight loss/regain, gastrointestinal symptoms and compliance with a healthy lifestyle [21].

With respect to other DEBs, we showed that weight loss trajectories were poorer for patients presenting with versus without post-operative DEBs (OBE, SBE, picking/nibbling), with differences observed by 18 months following surgery [13]. Furthermore, data from a subsample of LABS participants suggest a prospective effect of DEBs on weight outcomes, such that LOCE at one year follow-up predicts subsequent weight loss at two years [12]. On the contrary, more recent data on the full LABS sample found no evidence for a prospective (LOCE at previous assessments) or cumulative (total number of LOCE endorsements over time) effect of LOCE on weight outcomes over 7 years of follow-up. However, LOCE was concurrently associated with weight outcomes assessed at the same time point [4]. Thus, past history of DEBs may not be as prognostically meaningful as current DEBs in predicting weight outcomes.

Clinical management

To summarize, there is significant heterogeneity in the timing and impact of DEBs after surgery. A subgroup of patients with pre-operative BED presents with subthreshold LOCE post-operatively, while another subgroup remits. A considerable percentage of patients develop new onset BED, LOCE, and/or grazing post-operatively; these behaviors may appear relatively early after surgery, but tend to impact weight only after a year following surgery. Finally, prospective effects of post-operative DEBs on weight outcomes are inconsistent, but there is strong evidence for a concurrent effect. The latest findings should be considered to inform clinical management of DEBs in the context of surgery.

Screening and assessment.

Pre- and post-operative assessment should focus on early detection of DEBs and psychological/personality characteristics that may persist after surgery and/or increase the risk for poor weight outcomes. A systematic assessment battery incorporating measures that have been validated in the bariatric population should be used throughout the post-operative period to identify high-risk behaviors before they significantly impact weight.

Prevention and treatment.

Given limited evidence for effects of pre-operative DEBs on post-surgical DEB and weight outcomes, it may be most expedient to target interventions towards the post-surgical period. Importantly, pre-operative interventions may not generalize to patient’s post-operative needs, and patients who develop DEBs post-operatively may not retain skills learned by the time their application is necessary. Moreover, bariatric surgery itself may achieve more robust impacts on weight-related and psychological outcomes than psychological/behavioral interventions, although head-to-head comparisons are needed [25**]. Indeed, there appears to be no effect of insurance-mandated pre-operative weight management on rates of readmission/reoperation, follow-up visits, or weight loss one year after surgery [26], and pre-surgical cognitive behavioral therapy seems have no impact on DEBs or BMI at one and at four years after surgery despite showing post-intervention effects [25**]. Of note, in the study reporting an association between pre-operative BED and higher BMI post-surgery [15], patients with pre-operative BED received pre-surgery intervention.

Taken together, there is limited support for the efficacy of pre-operative interventions. However, we believe that undergoing bariatric surgery should follow an informed decision-making process in which patients are fully aware of the need for lifestyle changes, treatment requirements, and risks of suboptimal response. Thus, we propose that pre-operative interventions should be geared towards psychoeducation with the intent of informing patients about their role as active agents in the treatment process, changes they will experience, importance of having a strong support system [27,28], and risk for re-emergence of DEBs and possible impacts on weight.

In contrast, there is growing support for the need to address risk behaviors in the post-operative period to optimize weight outcomes. Clinical approaches should consider the variability of weight loss trajectories across individuals, and the limited public health resources to meet demands of an increasing number of patients. Research should investigate how to optimize the allocation of resources in a cost-effective manner, and identify when and whom needs additional treatment [29*]. A stepped-care approach may hold promise as a strategy for more adaptive and personalized care [29*]. For example, low-intensity strategies such as self-monitoring of weight and food intake may be offered first as a preventive strategy for weight regain, before moving on to more intense interventions, such as behavioral/psychological treatment or pharmacotherapy for weight loss. Such an approach should pivot from systematic and ongoing monitoring of key risk behaviors to screen patients as they develop needs for more intense clinical attention.

Alternative delivery methods can improve access to specialized care while utilizing fewer human resources. New technologies [30] and telephone-based interventions [31] offer novel platforms for contacting the patients without the caveats associated with face-to-face care [32]. These platforms allow for frequent monitoring of risk behaviors and delivery of “real time” intervention content [32], while supporting a greater number of patients at reduced cost [33]. Indeed, pilot data suggest that an Internet-based behavioral intervention may be an acceptable and effective for weight regain and eating-related problems after surgery [30], with high rates of retention (70%) and satisfaction. Other technology-based approaches, such as just-in-time adaptive interventions, should be tested as well [34].

Conclusion

There is strong support for an association between post-operative DEBs and weight outcomes following BS, and growing data suggest that intervening in the post-surgical period may improve patient outcomes. Yet, several gaps in our understanding of DEBs and surgery outcomes need to be addressed through future research. First, researchers from different fields (e.g., eating disorders, obesity, personality) should consider the ways in which different DEBs overlap and develop a common language for DEB constructs to improve cross-study comparisons and screening/intervention. Relatedly, existing diagnostic schemes and assessment protocols may need to be tailored for the post-bariatric population. Second, an improved understanding of the mechanisms by which bariatric surgery improves DEBs should be undertaken to inform both surgical and non-surgical treatment options for these behaviors [35]. Third, considering the marked heterogeneity of the bariatric population [36], optimal modalities, timing, and delivery formats of interventions should be identified [32], with a focus towards cost-effectiveness, given the growing demand for surgical procedures. In this era of personalized medicine, a stepped-care approach based on systematic screening of risk behaviors seems like a reasonable strategy to deliver care for patients in need. Rigorous trials with large samples, state-of-the-art assessment tools, and ability to dismantle critical intervention components and their cost are essential to address the burden of obesity on patients and public health.

Key-points.

1. Although DEBs often remit after surgery even without targeted intervention, a subgroup of patients have persistent or newly developed DEBs post-operatively

2. Pre-operative DEBs have little effect on weight outcomes.

3. A concurrent effect, rather than prospective or cumulative effect, of DEBs on weight outcomes yields the stronger predictive value.

4. Post-operative stepped-care interventions hold promise for optimizing treatment outcomes in a cost-effective manner.

5. Diagnostic schemes and assessment tools may require tailoring for the bariatric population.