Keywords

1 Introduction

As the obesity epidemic continues to grow worldwide, bariatric surgery is increasingly being utilized as a treatment for the severely obese. It is now widely recognized that about two out of three adults in the USA are overweight (BMI = 25–29.9 kg/m2) or obese (BMI > 30 kg/m2), and rates of obesity continue to rise across the age spectrum. Given this increase, and the growing awareness that obesity is associated with a variety of untoward health consequences, including cardiovascular disease, asthma, liver dysfunction, obstructive sleep apnea, various malignancies, and infertility (Bray, 2003; Shah & Ginsburg, 2010), bariatric surgery continues to grow as a treatment for the severely obese. This reflects several additional factors, including an increased awareness that most psychosocial treatments and many of the available medications impact only modestly on body weight and often provide only a temporary improvement. Also, increasingly, bariatric surgery has been found to be a safe and effective approach, with the subsequent significant weight loss frequently resulting in reversal or marked improvement in medical comorbidities (Latifi, Kellum, De Maria, & Sugerman, 2002).

2 Bariatric Surgery Procedures

In considering bariatric surgery, there are three procedures that are commonly used in the USA, and increasingly worldwide. There are other procedures that are used, though rarely, and new procedures are under development. The most commonly used procedure in the USA is the Roux-en-Y gastric bypass (RYGB), which involves the creation of a small gastric pouch in the upper stomach and bypass of the first portion of the small intestine, resulting in both a restriction of food intake and a degree of malabsorption. Also widely used is the laparoscopic adjustable gastric band (LAGB), which involves the placement of a band around the upper stomach creating a small pouch. In this procedure, there is restriction in the amount that can be eaten, but no malabsorption. The most recent procedure to be introduced, which is gaining popularity, is the gastric sleeve (sleeve), which was originally envisioned as the first of a two-stage procedure and is now increasingly being utilized as a single-stage operation. This involves creating a sleeve out of the stomach and removing a large portion of the stomach. In general, the amount of weight loss is superior with the RYGB and the sleeve to that achieved following the LAGB, and the majority of patients eventually lose 50 % or more of their excess body weight, while weight losses following LAGB placement are generally more modest. However, it must be noted that bariatric surgery procedures are not without complications, including complications during the intraoperative period (e.g., the risks of anesthesia, trauma to other internal organs), short-term complications (e.g., intestinal leaks, bowel obstruction, pulmonary embolism), and late complications (e.g., internal hernias, vitamin and mineral deficiencies, and dumping syndrome) (Byrne, 2001; Elliot, 2003; Koch & Finelli, 2010; Latifi et al., 2002). However, for the majority of severely obese patients, the benefits of these procedures clearly outweigh the risks. The presurgical mentality rate is 0.3 % while the adjusted long-term mortality rate at a mean of 7.1 years was reduced 40 % (Adams et al., 2007; Flum et al., 2009).

In addition to the focus on gastric restriction and malabsorption as the mechanisms resulting in weight loss following RYGB, it is increasingly thought that various normal changes, including alterations of various peptides, including glucagon-like peptide 1 (GLP-1), an important regulator of glucose homeostasis, are involved. Also changes in binding to certain receptor sites are now considered possible mechanisms, including the binding to nuclear receptor Farnesoid XR (FXR) in the gut, which is also involved in glucose regulation as well as regulation of lipids and a variety of other metabolic processes. Therefore, some of the changes actually result from physiological and endocrinological changes rather than purely anatomical modifications to the gut.

Although obesity is generally not considered to be primarily an impulse control or addictive disorder, this has become an area of intense interest and debate in recent years (DiLeone, Taylor, & Picciotto, 2012; Volkow, Wang, Tomasi, & Baler, 2013). However, there is considerable overlap in the behaviors associated with obesity and these disorders. There is some literature to support the idea that patients who undergo bariatric surgery may go on to develop a new addiction or new impulse control disorder (ICD) to replace a food “addiction,” depending on how this addiction is defined. Recent data have also shown that patients who are status-post bariatric surgery have an increased prevalence of alcohol disorders. Despite earlier suggestions that support the development of other addictive disorders, there are very few published data in this area. Such disorders are summarized in Table 9.1. We will first review the data on problems with alcohol before turning to other potential problems.

Table 9.1 Bariatric surgery and addictive disorders
Full size table

3 Bariatric Surgery and Alcohol Abuse/Dependence

Ertelt and colleagues (2008) surveyed post-bariatric surgery patients to investigate self-reported alcohol abuse and dependence before and 6–10 years after bariatric surgery. Of the 70 respondents, six individuals were identified who met self-reported criteria for a diagnosis of alcohol dependence and one participant reported alcohol abuse after surgery. In this sample, two patients developed new, postoperative alcohol dependence. Additionally, two patients reported that they did not drink alcohol prior to surgery, but did consume alcohol after surgery. Two other patients reported that alcohol consumption had increased after surgery. However, most of those in the sample reported no change in their alcohol consumption. Reports in this area should be interpreted with the knowledge that active alcohol use disorders (AUDs) are usually considered exclusionary for these surgeries, but patients may not admit to such problems prior to surgery in order to not jeopardize their chances of receiving the surgery

Sogg, Hatoum, and Turbett (2011) surveyed 340 RYGB patients about their alcohol use frequency before and at various durations after surgery and found that 73 participants reported consuming alcohol at a level that was defined as “problem drinking.” Thirty-two participants reported problem drinking after surgery, and importantly, 19 of these participants did not report preoperative problem drinking.

Another published study by Suzuki, Haimovici, and Chang (2012) provided some additional relevant data. These authors surveyed 530 post-bariatric surgery patients about their alcohol consumption. Of these 530 patients, 51 agreed to be interviewed and assessed more thoroughly at a mean of 43.4 months postsurgery. Two key findings from this study were as follows (1) individuals with a lifetime history of an AUD were more likely to have an AUD after surgery compared to those who did not have such a history and (2) patients who underwent RYGB were more likely to have an AUD after surgery than those who received an LAGB.

Welch and colleagues (2011) assessed 75 RYGB patients 2 years after surgery using the CAGE, a screening tool for alcohol problems. Although not the primary focus of the study, the authors provided data on self-reported levels of alcohol abuse in this sample. Of the 75 participants who completed follow-up assessments, only 1.3 % of them responded in a manner suggesting that they abused alcohol.

Östlund (2011) presented data from 12,277 patients who had undergone bariatric surgery. She reported that patients who underwent bariatric surgery were at greater risk for inpatient treatment for alcoholism than a comparison group from the general population in Sweden, who were matched for age and gender. Importantly, she further reported that those who underwent gastric bypass were 2.3 times more likely to develop alcohol problems than those who received LAGB.

Most recently, our group along with other investigators from the NIDDK-funded Longitudinal Assessment of Bariatric Surgery-2 (LABS-2) consortium published a paper on this issue (King et al., 2012). Among 1,945 patients who underwent bariatric surgery and completed baseline (presurgery), and follow-up assessments 1 and 2 years postsurgery, there was a significant increase in AUD noted between baseline and year 2. The percentages of patients with an AUD, as defined using the AUDIT, were 7.6 % at baseline, 7.3 % at 1 year (a nonsignificant change from baseline), and 9.6 % at 2 years (p = 0.01 vs. baseline). The odds of a postoperative AUD were increased among males and younger patients and those who reported presurgery smoking, regular alcohol consumption, AUD, recreational drug use, a lower sense of belonging, and among those who underwent a RYGB as opposed to an LAGB. Another recent paper (Fogger & McGuinness, 2012) reported that 14 % of a series of 173 participants in a state monitoring system for nurses with addiction problems had undergone bariatric surgery, and 10 % had developed an addiction after surgery, most commonly to alcohol or hydrocodone.

Therefore, several lines of evidence, most recently the large data set collected by the LABS-2 consortium, suggest that patients who undergo RYGB have a higher risk of developing an AUD versus those who receive an LAGB. There is currently insufficient evidence regarding the development of AUD in patients who undergo sleeve gastrectomy. In summary, the RYGB is the most common bariatric surgery performed in the USA, and the majority of available data suggest that RYGB leads to an increase in the risk of a postoperative AUD.

A number of other studies have explored pharmacokinetic changes that occur following bariatric surgery. These studies are particularly relevant as pharmacokinetic changes inevitably may act as a causal mechanism for increased alcohol use problems longitudinally. The studies in this area have included a variety of methodologies including pre- and postsurgery designs, weight-based versus fixed dosing of alcohol, and alcohol concentrations measured via both breath (breath alcohol concentrations; BrAC) and blood sampling (blood alcohol concentrations; BAC). Despite the variety of methodologies used, the data have shown an accelerated and higher peak alcohol concentration in postsurgery patients. Some of these studies have also shown differences in alcohol metabolism following surgery, although this finding has been inconsistent.

In 2002, Klockoff, Naslund, and Jones published a study that examined the pharmacokinetics of alcohol in women who had undergone RYGB. They compared 12 healthy control participants to 12 patients who had undergone surgery on BAC after administration of a bolus, weight-based dose of alcohol (0.3 g kg−1). They found that compared to the control group, the RYGB group experienced significantly higher BAC at the first (10 min) and second (20 min) measurement points. Groups were markedly and significantly different at the first assessment point: at the 10-min measurement, the operated group had a BAC of 0.713 g l−1 and the control group had only reached a level of 0.171 g l−1.

In 2007, Hagedorn, Encarnacion, Brat, and Morton reported on the metabolism of alcohol in 17 control and 19 post-gastric bypass patients. They measured BrAC every 5 min from 15 to 140 min following ingestion of a fixed, 5-oz drink of red wine. The main findings from this study were that the RYGB patients had a greater peak BrAC and, unlike the findings of Klockhoff, Naslund, and Jones (2002), a longer time for alcohol levels to reach 0 compared to controls. Notably, at the first assessment point (15 min), RYGB patients experienced higher BrAC than control participants. Limitations of this study included differing BMIs of the surgical and nonsurgical groups and a fixed dose that did not control for differences in body weight.

Woodard, Downey, Hernandez-Boussard, and Morton (2011) studied 19 RYGB patients’ BrAC prior to surgery as well as 3 and 6 months after surgery. Patients drank a fixed dose (5 oz) of red wine and provided BrAC recordings at multiple time points in the study. They found that peak BrAC was higher and it took greater time to return to a BrAC of 0 in the postoperative conditions than the preoperative condition. Additionally, patients reported more dizziness and double vision in the postoperative conditions. Finally, it is important to note that the peak BrAC was highest at the first assessment point (15 min post-drink). Like the study by Hagedorn and colleagues (2007), this study is also limited by the administration of a fixed, non-weight-based dose of alcohol. This is also true for another paper published by this group (Changchien, Woodard, Hernandez-Boussard, & Morton, 2012), which failed to find evidence of changes in kinetics 3 months and 6 months after LAGB or laparoscopic sleeve, using a fixed dose, non-weight-based amount of alcohol.

Maluenda et al. (2010) studied 12 patients who had undergone laparoscopic sleeve gastrectomy. These patients drank a weight-based dose (3.6 ml per liter of water body mass) on two occasions, once prior to surgery and once 30 days after surgery. They found that BrAC was higher and it took longer to return to a BrAC value of 0 in the postsurgery condition. Again, the peak time BrAC values were collected at the first measurement point was 10 min post-alcohol consumption.

The findings from the studies reviewed above are important in that they show that postoperative bariatric surgery patients reach higher levels of intoxication compared to their presurgery state or weight-matched controls. These data, combined with the prevalence of the data cited earlier, suggest that a subgroup of patients, some of whom have a history of AUD but some of whom do not, abuse alcohol after bariatric surgery. A number of variables may presumably contribute to this outcome, including the pharmacokinetic changes in alcohol after surgery.

4 Bariatric Surgery and Eating Disorders

One of the areas of interest in terms of eating pathology after bariatric surgery is what is termed “loss of control” (LOC) eating. Questions remain as to whether LOC eating should include either binge eating (i.e., a quantifiably large amount of food is consumed) or subjective overeating (when a patient feels as though they are unable to stop eating but the amount of food ingested is not necessarily large) or whether both should be considered. Some studies have noted that there is a subgroup of people who indicate that they have difficulty with a sense of LOC at times while eating after bariatric surgery and that this population is more likely to regain weight after RYGB. There is, however, a concern that LOC may be the self-labeling of an epiphenomenon associated with depression. Also, gradations of LOC are likely, as opposed to it being simply a dichotomous variable.

This issue of LOC eating following bariatric surgery will be discussed in detail given that this is the only eating-related variable which has consistently been shown to impact on weight outcomes across a variety of studies. In 1992, Rowston et al. reported follow-up data on 16 patients 2 years after biliopancreatic diversion (BPD) using the Bulimic Inventory Test, Edinburgh (BITE) questionnaire. They found that patients overall lost a significant amount of weight and that eating disordered behavior usually improved within 3 months after surgery. However, those who reported binge eating (BE) and/or LOC eating lost less weight over the longer follow-up period, although preoperative BE was not predictive of postoperative LOC. Pekkarinen, Koskela, Huikuri, and Mustajoki (1994) studied 27 vertical banded gastroplasty (VBG) patients 5 years after surgery. They found that patients who reported BE had a statistically significant greater weight regain, and very few had what they defined as a successful outcome at the end of follow-up. They defined BE as eating large amounts of food, accompanied by a sense of LOC, and subsequent feelings of guilt and shame. They used the Binge Eating Scale (BES) and BITE to identify BE behavior postoperatively and identified it as the “main predictive factor for poor outcome.”

Eating disturbances in patients undergoing VBG (a practice no longer widely used in the USA) were studied by Hsu, Betancourt, and Sullivan (1996). They retrospectively analyzed the eating behavior of 24 female patients 3.5 years after surgery using the Eating Disorder Examination (EDE). A trend was found for those with an eating disturbance before surgery to regain more weight compared to those without such a disturbance. However, this study was retrospective and cross-sectional in design, of short duration, and had low statistical power owing to the sample size.

In 2001, our group did a structured long-term (13–15 year) follow-up of patients after RYGB (Mitchell et al., 2001). Seventy-eight patients were interviewed, and it was found that BE/LOC was associated with more weight regain, if present after surgery, but not before surgery. Limitations of this study included its retrospective design and the long time to follow-up that may have reduced the veracity of the data obtained. Also, some patients were lost during the follow-up interval.

Kalarchian et al. (2002) studied BE among RYGB patients at long-term follow-up. They studied 99 patients between 2 and 7 years after RYGB using the Eating Disorder Examination Questionnaire (EDE-Q) and found that BED or LOC eating was associated with less weight loss, which occurred in 46 % of patients. Another study was completed by Guisado Macias and Vaz Leal (2003). They followed up 140 patients after VBG for 18 months postoperatively. They assessed patients using the BES and did find a correlation between those who reported LOC eating postoperatively (about 18 %) and poor outcomes. However, they did not assess predictive factors.

Larsen et al. (2006) followed 157 patents approximately 34 months after undergoing LAGB. They administered the BES and found that those patients who reported engaging in BE were more likely to have regained weight. They also found that patients who engaged in what was labeled “emotional eating” or “external eating” were also more likely to meet the BE criteria, but they did not assess preoperative characteristics that might be predictive of those outcomes.

Colles, Dixon, and O’Brien (2008) studied “grazing” and LOC eating in 129 patients 12 months after undergoing banding. They used the Questionnaire on Eating and Weight Patterns-Revised (QEWP-R) and found that BE and LOC were associated with less weight loss. They also assessed preoperative characteristics. They found that only 50 % of patients who exhibited preoperative BE went on to evidence LOC eating and weight regain.

In 2010 White, Kalarchian Masheb, Marcus, and Grilo assessed 361 patients preoperatively and 1–2 years after RYGB. They used the EDE-Q, and their data revealed that preoperative LOC eating was associated with less weight loss/more weight regain. They found a high prevalence of LOC (approximately 61 % preoperatively, and about 36–39 % postoperatively). They also found that preoperative LOC was predictive of postoperative LOC. It also appears to be important as to when the LOC eating patterns returns, since reoccurrence shortly after surgery was associated with a better outcome than that reoccurring longer term after surgery in this study. The reason for this difference is not obvious.

de Zwaan and colleagues (2010) studied 59 patients after RYGB 2 years postoperatively. They used a comprehensive interview, the Eating Disorder Examination-Bariatric Surgery Version (EDE-BSV), as well as QEWP-R preoperatively, and found that preoperative LOC was associated with less weight loss. In their study population, 29 % of patients preoperatively met criteria for BE/BED and 25 % did so postoperatively. They also found that preoperative LOC was predictive of postoperative LOC in 47 % of patients. Beck, Mehlsen, and Stoving (2012) also found that BE postsurgery, at a follow-up mean of 2 years, produced less weight loss in a sample of 45 previous RYGB patients using an instrument designed for this study.

A review of these data raises several important points: first, the studies reveal that BE/LOC postoperatively is associated with less weight loss or more weight regain. However, only half of the studies evaluated preoperative eating disordered behavior, with only one study examining what was labeled preoperative LOC per se, although LOC must have been assessed in several other studies as well given that some patients satisfied criteria for BE. Second, the studies that evaluated postoperative BE/BED or LOC found rates varying widely from 3 % in Colles et al. (2008) to 46 % in the Kalarchian et al. (2002) papers. Possible contributors to this wide variance include type of surgery (LAGB vs. RYGB) and type of assessment (e.g., QEWP-R vs. EDE-Q), length of follow-up, and whether LOC or BE/BED were both assessed.

Despite this body of data, important questions remain unanswered. Why do some people experience LOC eating after surgery, while others do not? What are the predictive factors for the emergence or reemergence of this LOC eating? Is the core psychopathology of an eating disorder present in this patient group, despite the fact that these people may not meet criteria for a current eating disorder diagnostic category? Further study is warranted to address these issues.

5 Bariatric Surgery and Other Addictive Disorders

We will now discuss literature regarding the presence of other ICDs and behavioral addictions. These disorders have undergone a shift in the DSM-5, which now includes kleptomania, pyromania, and intermittent explosive disorder as specific ICDs. Gambling disorder is now included as a non-substance-related addictive disorder and trichotillomania in the obsessive-compulsive and related disorders section.

“Healthy” behaviors that are performed excessively and thus may become harmful are sometimes grouped as “behavioral addictions” (see Chap. 18) and include compulsive buying, pathological skin picking, nonparaphilic compulsive sexual behavior, pathological internet use, and excessive exercising (Berczik et al., 2012; Coleman, Raymond, & McBean, 2003; Dell’Osso, Altamura, Allen, Marazziti, & Hollander, 2006; Karim & Chaudhri, 2012; Kuzma & Black, 2008; Meyer, Taranis, Goodwin, & Haycraft, 2011; Odlaug & Grant, 2010).

Corresponding to the relatively sparse research on non-substance addictions and ICDs-NOS in general, only a few studies have investigated their role in bariatric surgery patients. As obesity is associated with increased impulsivity and reduced impulse control (Davis, 2010; Gruss, Mueller, Horbach, Martin, & de Zwaan, 2012; Lent & Swencionis, 2012; Mobbs, Crepin, Thiery, Golay, & Van der Linden, 2010; Müller et al., 2012; van Hout, van Oudheusden, & van Heck, 2004), one might expect elevated prevalence rates of ICDs or behavioral addictions among bariatric surgery patients. ICDs have been shown to be relatively prominent in obese, pre-bariatric surgery candidates. In a sample of 100 presurgery individuals, 19 % met criteria for at least one current ICD and 27 % had a lifetime history of at least one ICD (Schmidt, Korber, de Zwaan, & Müller, 2012). The most prevalent ICDs were excoriation disorder and compulsive buying, with current prevalence rates of 8 % and 6 % and lifetime rates of 9 % and 8 %, respectively. With regard to other ICDs, the following current/lifetime prevalence rates were reported: intermittent explosive disorder 5 % and 10 %, pathological gambling 1 % and 3 %, and pathological internet use 2 % and 5 %. No participant was diagnosed with pyromania, kleptomania, trichotillomania, or hypersexual behavior. Of note, the prevalence rates in this obese sample differed barely from those among consecutive psychiatric inpatients who were assessed with the same structured interview (Müller et al., 2011). The pre-bariatric surgery sample consisted of consecutive morbidly obese individuals who were seen for a routine psychosocial pre-bariatric surgery evaluation but not for psychiatric treatment like the psychiatric inpatients. Considering this, it is surprising that there were no differences found between psychiatric inpatients and obese individuals who were not seeking psychotherapy or psychopharmacological treatments.

There is an increasing literature on excoriation disorder, which is now listed in DSM-5’s section on obsessive-compulsive and related disorders (Grant & Odlaug, 2009; Odlaug & Grant, 2010). Excoriation disorder is characterized by repetitive picking of the skin, commonly on the head and face, although virtually any area of the body can be involved. Compulsive picking results in tissue damage and often leads to scarring, with resultant significant body image concerns. This is a widely recognized and relatively common psychiatric problem, with a lifetime prevalence estimated to be between 1.4 % and 5.4 % (Hayes, Storch, & Berlanga, 2009; Keuthen, Koran, Aboujaoude, Large, & Serpe, 2010). The aforementioned findings of Schmidt et al. (2012) are in line with earlier reports that indicate a much higher prevalence of skin picking in clinical samples than in the general population, including patients with obsessive-compulsive disorders (Grant, Mancebo, Pinto, Eisen, & Rasmussen, 2006), trichotillomania (Odlaug & Grant, 2010), and eating disorders (Favaro, Ferrara, & Santonastaso, 2007). It is important to note that pathological skin picking should be carefully delineated from non-suicidal self-injury.

Compulsive buying (see Chap. 19), also referred to as compulsive shopping or shopping addiction, also seems to be prevalent in obese individuals who are seeking bariatric surgery. This behavioral addiction is characterized by repetitive maladaptive purchasing behavior that is difficult to control and results in distress and significant social, psychological, and financial consequences. Compulsive buying is common in consumer societies with prevalence estimates of about 5–8 % (Koran, Faber, Aboujaoude, Large, & Serpe, 2006) and often is associated with other psychiatric illnesses, particularly mood and anxiety disorders, other ICDs, and BED. Whether or not it is associated with substance use disorders is unclear (Mueller et al., 2010).

The prevalence of other ICDs and behavioral addictions has not been reported in the literature following bariatric surgery. Some have argued that the problematic behaviors may increase or transfer to another problem in order to replace the presurgery “food addiction” through the development of other non-substance addictions (Buffington, 2007; Moorehead & Alexander, 2007; Spencer, 2006). Thus far, only anecdotal reports have examined the development of postoperative behavioral addictions. For example, instead of overeating/binge eating, some patients who have undergone bariatric surgery engage in compulsive buying of clothing, hypersexual behavior, or other extreme activities that appear to be triggered by the massive weight loss. Excessive exercising (see Chaps. 7 and 28) that is often linked to eating disorder pathology and extreme fear of regaining weight may also result (Blum et al., 2011; Larsen et al., 2006). Post-bariatric surgery patients frequently suffer from redundant skin. Beyond appearance-related issues, hanging skin after substantial weight loss can result in skin irritation and provoke or aggravate pathological skin picking. To our knowledge, the “addiction-transfer” hypothesis has yet to be empirically investigated adequately in terms of non-substance addictions. However, we have seen individuals within our clinical practice that suffered from behavioral addictions, and we think that the potential link between behavioral addictions and bariatric surgery should be further addressed.

Conclusions

The hypothesis that bariatric surgery may eventually result in the development of new addictive disorders, or the reemergence of a disorder in remission, has received a great deal of attention in the media and increasing attention in the bariatric surgery literature. A growing literature suggests that AUD may reemerge or develop de novo after bariatric surgery, in particular after RYGB. Additional literature suggests that some patients postsurgery will develop so-called LOC eating. Some of these individuals will have a history of BE or BED prior to surgery. The emergence of LOC eating seems to be associated with less weight loss or greater weight regain, making it a clinically important variable in long-term outcome. Of great interest, it is currently unclear whether or not the patients that develop AUD after surgery are the same ones who develop BE/LOC eating, or how a history of either problem presurgery contributes to the development of the other problems after surgery. Currently data on other substance abuse problems after surgery, including opioids and amphetamine, are limited to a small number of case reports. Other “cross addictions,” such as skin picking and excessive exercising, have been suggested as possible sequelae of bariatric surgery, although data demonstrating this cross addiction potential empirically are currently lacking.