Increased alcohol intake and alcohol use disorder following bariatric surgery: potential mechanisms

Abstract

Sleeve gastrectomy (SG) and Roux-en-Y gastric bypass (RYGB) are the most effective weight loss procedures for severe obesity. However, there is recent evidence of increased alcohol intake and new onset alcohol use disorder (AUD) by 2 yr following both operations. Although the two surgeries differ anatomically, they lead to similar increased drinking. The mechanisms behind increased alcohol intake post-surgery remain unclear. One theory is that with a marked reduction in food intake post-surgery, there is an “addiction” transfer from food to alcohol intake. Another theory implicates the higher rate of absorption of alcohol post-SG and post-RYGB for the increased alcohol intake. Elucidating the mechanisms could help provide new therapeutic targets for preventing increased alcohol intake and AUD as well as identify measures that could be used in a clinical setting to help predict the likelihood of AUD. These findings could also help guide the development of new surgical procedures that do not lead to increased alcohol intake.

1. Introduction

Bariatric surgery is the most effective obesity treatment, producing sustained weight loss and reductions in obesity-related morbidity and mortality [1]. Globally, the most common procedures are sleeve gastrectomy (SG) (61%) and Roux-en-Y gastric bypass (RYGB) (26%) [2]. The operations differ anatomically: In RYGB, the upper portion of the stomach is detached (left in situ with a 30 ml volume) from the lower portion, and anastomosed to the mid-jejunum, bypassing the pylorus. In contrast, SG, entails removal of the outer portion of the stomach along the greater curvature, leaving a narrow gastric sleeve with ~100 ml volume and an intact pylorus. With a markedly reduced gastric capacity post-surgery, solid foods are not easy to ingest and produce marked fullness.

Although lifestyle changes, including diet, can induce medically significant weight loss, up to 50% of the weight lost is typically regained within 1 yr, and 90% within 5 yr [3–5]. Until recently, FDA-approved medications for weight loss have produced only modest effects, and only for as long as the medication is taken [6]. Individuals with Class 3 obesity (BMI ≥ 40) face even greater difficulty in achieving and maintaining weight loss [7]. Recent FDA-approved GLP-1 receptor agonists (GLP-1 RAs), such as tirzepatide, a dual GLP-1/GIP agonist, show considerable promise [8], but have to be injected and may incur moderate to serious GI side effects. Indeed, the discontinuation in real-life use of GLP-1 RAs, including tirzepatide, the most efficacious weight loss drug, was 65% among nondiabetics, partly due to side effects [9]. Discontinuation commonly results in rapid weight regain [10]. The drugs are also expensive, with tirzepatide costing over $1000 per month in the U.S. [10], and insurance coverage is often limited or unavailable.

In 1991, the NIH Consensus panel recommended bariatric surgery for those with BMI≥40, and for moderate obesity, BMI 35–39.9 with comorbid conditions, e.g., Type 2 diabetes [11]. More recently, the American Society for Metabolic and Bariatric Surgery (ASMBS) recommended more relaxed guidelines of BMI ≥35 without comorbidities and 30–34.9 with comorbidities [12]. However, insurance companies in general have not adopted these guidelines [13], and therefore most surgeons are adhering to the older NIH guidelines. Bariatric surgery is more effective (~ 30% weight loss) [14] than non-surgical treatments, including tirzepatide, and leads to a 29% reduction in overall mortality [15].

Although weight loss is similar for the two procedures [16], RYGB has the edge over SG in the longer term [17, 18], which may be related to the smaller gastric capacity in RYGB. Alterations in gut peptides do not help explain the greater weight loss following RYGB: Postprandial release of GLP-1 and PYY [16, 19] (satiety peptides) is increased in both operations, whereas ghrelin (hunger hormone) declines more following SG [20]. However, changes in brain activation may partly contribute to the differences in weight loss. Specifically, there is reduced reward activation, assessed by fMRI, following both RYGB and SG in response to visual and auditory high-energy-dense (HED) palatable food cues relative to low-energy-dense (LED) food cues [19]. Although there is greater activation of the dorsolateral prefrontal cortex (DLPFC), an inhibitory area, in response to HED vs LED in both operations, this activation is larger in RYGB than SG [19, 21]. At 1 and 2 yr post-RYGB, energy intake [22] and energy density are reduced, with lower fat consumption and greater fruit intake [23]. Similarly, post-SG, liking is reduced for high-calorie food [24], i.e., high in fat and carbohydrate [25]. Energy intakes post SG and RYGB are similar [26], and 2 years post-SG, food intake and dietary energy density remain reduced [27]. Although nutritional deficiencies can occur following both operations (20), there is a greater risk of vitamin B₁₂ deficiency following RYGB [28]. Gastric emptying rate is accelerated in both operations but appears to be somewhat faster following RYGB [20, 29]. Indeed, dumping syndrome, the result of rapid emptying, especially of sugary drinks, and characterized by symptoms such as light-headedness, sweating, dizziness, nausea, and abdominal discomfort, is more common following RYGB [30].

Alcohol Intake

There are, however, concerning reports of increased alcohol intake and AUD following both operations and typically occurring by 2 yr post-surgery [31–33]. A large survey found that 83% of patients were consuming alcohol post-surgery, and 28% of these patients had difficulty controlling their alcohol use compared to just 4.5% pre-surgery [34]. Additionally, 84% of those consuming alcohol reported heightened sensitivity to alcohol’s effects [34].

In the largest study [35] to date, AUD rates for the year pre-surgery, 1 yr, and 2 yr post-surgery were 7.6%, 6.3%, and 11.9% respectively, for RYGB, and 10.1%, 9.0%, and 14.4% for SG. The rates increased (p < .001) from pre-surgery to 2 yr post-surgery, without differing between the surgeries. New onset AUD increased modestly (n.s.) from baseline to 1 yr (+0.54% for RYGB and +0.75% for SG), and markedly (p <.0001) from 1 yr to 2 yr (+7.25% for RYGB and +8.5% for SG), without differing from each other [35]. Notably, 61% of AUDs for RYGB and 59% for SG, were new onset in the 2^nd year. The results are consistent with a large NIH-funded study, Longitudinal Assessment of Bariatric Surgery (LABS) [32], comparing RYGB and GB (gastric banding). For RYGB, AUD rates in the years pre-surgery, 1 yr, and 2 yr postsurgery were: 7.0%, 7.9%, and 10.7%, respectively, with p <.001 for the change from 1 yr to 2 yr as compared with GB (9.3%, 5.6%, and 7.0%, respectively, with p =.24, n.s., for the change from 1 yr to 2 yr; i.e., rates increased for RYGB but not for GB. A recent systematic review confirmed the high rate of new onset AUDs following both RYGB and SG [36]. There is also evidence of increased postsurgical alcohol use in adolescents at 8 yr after RYGB or SG [37]. There are only a few predictors of new-onset AUD post-surgery [35]: baseline alcohol consumption for RYGB and SG, and higher education, lower income [38], male sex, young age, and smoking for RYGB [32, 39].

According to the Diagnostic and Statistical Manual-5 (DSM-5), AUD is based on ≥ 2 symptoms of 11, classified as mild category (2–3 symptoms), moderate (4–5), and severe (≥ 6). AUD onset is unrelated to pre-surgery depression or other psychopathology or to post-surgery weight loss [39, 40]. Development of other substance use disorders is uncommon post-surgery [31, 41] or may increase modestly [40]. The development of AUD is of concern as AUD is a leading cause of morbidity (i.e., heart disease, liver cirrhosis [42], various cancers [43]) and mortality[44]. In sum, bariatric surgery has emerged as a new potential risk factor for AUD.

2. Potential Mechanisms

The mechanisms for the increase in alcohol intake post-surgery remain incompletely understood. There are overlapping neural reward pathways underlying substance addiction, including alcohol, smoking, and food intake [45, 46]. One hypothesis is that surgical patients may be susceptible to addiction transfer, whereby alcohol use is substituted for overeating, particularly for HED (high-energy-dense) foods, which is markedly reduced post-surgery [39, 47, 48]. Unlike solid HED foods, alcohol, itself a high-calorie substance in liquid form, can be consumed easily, does not require digestion, and empties rapidly from the stomach [49]. Substitution of alcohol for HED foods may also be related to anatomical changes in the GI tract post-surgery, facilitating faster alcohol absorption. Such anatomical effects could help explain why alcohol intake does not increase following GB, which restricts the stomach without altering GI anatomy [31, 32]. Unlike RYGB and SG, where alcohol absorption is faster [39, 50], there is no pharmacokinetic change following GB [41].

Gastric emptying of liquids is accelerated following both RYGB [51] and SG [52]. After RYGB, peak blood alcohol concentration (BAC) following a glass of wine increased and remained elevated for a longer period of time relative to nonsurgical controls [53]. Similarly, higher BACs were found after SG [54] and remained elevated for a longer period than pre-surgery [54, 55]. Consequently, even small amounts of alcohol can produce intoxication [55], and AUD could develop with relatively small increases in alcohol intake.

In addition to these pharmacokinetic changes, the rewarding effects of alcohol may be enhanced post-surgery. Repeated exposure to small doses of alcohol may produce euphoric effects that were not experienced preoperatively [56, 57]. As this is a learning process, it could help explain why it takes up to 2 yr for AUD to develop. Supporting the reinforcing aspect, one study found that one year after RYGB, alcohol was rated as more subjectively reinforcing compared to pre-surgery when administered at the same dose per body weight [60]. Another explanation of increased alcohol intake, based on the likelihood of more socialization after weight loss, when people feel more comfortable being in public, with more exposure to drinking and participation in drinking, is unlikely, as it should also occur after GB.

In one of the first longitudinal studies to observe an increase in alcohol intake post-surgery [31], our group followed 155 participants (132 f and 23 m) who underwent RYGB (n=100) or GB (n=55). Those who had RYGB reported an increase in alcohol use from pre to 2 yr post-surgery (p =.011) vs. no change in GB. This may be because gastric emptying is not accelerated following GB [59] as compared to RYGB and SG [60]. In another study from our group [33], 27 subjects with RYGB, SG, dietary weight loss, or no treatment (NT) participated. They completed the Yale Food Addiction Scale ( YFAS) and reported alcohol intake pre- and post-intervention at 4 and 24 mo. Between baseline and 24 mo., YFAS scores decreased (p = .006) and alcohol intake increased in the surgery groups (p = .005), but there were no changes in the diet or NT groups [33]. We did not, however, find a significant correlation between the change in YFAS scores and the change in alcohol intake among the surgery participants. This may be due to the surgery participants having relatively few YFAS symptoms at baseline and the small sample size. A systematic review and meta-analysis of studies using the YFAS confirmed the significant decline of the YFAS symptoms postsurgery [61].

It may be worthwhile to use other scales besides the YFAS to assess the more compulsive aspects of eating as opposed to withdrawal and tolerance, such as the Palatable Eating Motives Scale (PEMS) [62] or the Addiction-like Eating Behaviour Scale (AEBS) [63]. The eating behavior may not be a full-fledged addictive disorder but rather a problematic relationship with food, especially HED foods, whose desirability is diminished following surgery, and is then replaced by alcohol, which may activate similar neural circuits. In one study of 147 adults at least 24 months post-RYGB, higher baseline YFAS scores were associated with greater risk of postoperative substance use, including alcohol [64], reflecting behavioral substitution.

In a well-controlled cross-sectional study by Pepino’s group, approximately 2 alcoholic drinks (0.5 g/ kg of fat-free mass [FFM]) were ingested following RYGB, SG, or GB [65]. An intravenous catheter was retrogradely inserted into a superficial dorsal hand vein, with the hand kept heated by a thermostatically controlled box, to obtain “arterialized” venous blood. Blood alcohol samples from a heated dorsal hand vein yield similar concentrations as those obtained from an artery [66]. BAC rose faster and to a higher peak post RYGB (p < .05) than post SG, and both rose higher than GB (p < .05) [65] (Figure 1). They also directly compared BAC by gas chromatography (the gold standard), with alcohol breath testing, which proved to underestimate BAC by 27%.

Figure 1.

Blood alcohol concentrations (BAC) after alcohol ingestion (0.5 g/kg fat free mass; ~2 standard drinks) in women who had undergone sleeve gastrectomy surgery (SG; turquoise symbols), Roux-en-Y gastric bypass surgery (RYGB; black symbols), or laparoscopic adjustable gastric banding surgery (LAGB; white symbols).

*, †, ‡, P < .05, surgery groups differ significantly within a timepoint

§ P < .05 LAGB differs significantly from both RYGB and SG within a timepoint.

Thus, both SG and RYGB tend to accelerate alcohol absorption and elevate peak BAC. Since both surgeries also decrease the size and volume of the stomach, they also reduce the amount of alcohol metabolized in the stomach by alcohol dehydrogenase (ADH), an enzyme which breaks down alcohol, as part of first pass metabolism, leaving more alcohol to enter the small intestine [53] and absorbed into the blood [67]. These effects are not present following GB, which does not alter GI anatomy [65].

Alcohol clearance from the blood is also reduced following bariatric surgery, which could be related to a reduction in fat-free mass (FFM) [68] and may contribute to increased inebriation following alcohol ingestion post-surgery. Reduced clearance, however, is likely a smaller contributor to AUD development than reduced first-pass metabolism [69]. Administration of alcohol to the surgery group (RYGB and SG) and a control group, led to a reduction of GLP-1 by 34% in both groups, but ghrelin was reduced less in the surgery group (13%) than in the control group (27%). Thus, the smaller decrease of ghrelin could also be a contributing factor to increased alcohol ingestion post-surgery [70].

Collectively, the findings show that the two obesity surgeries significantly alter alcohol pharmacokinetics and reward sensitivity to alcohol, which may contribute to increased drinking and AUD.

3. Animal Studies

Animal models have enabled mechanistic investigations of alcohol consumption following bariatric surgery under controlled experimental conditions that are not feasible in human subjects. At 4 months post-RYGB, obese rats on a high-fat diet were placed on an operant progressive ratio schedule of reinforcement for 2, 4, and 8% ethanol. The RYGB rats worked harder to earn access to an alcohol reward and ingested more alcohol when made available than sham-operated controls. When pretreated with a ghrelin receptor antagonist, the operated rats reduced their effort and the amount of alcohol ingested, unlike the sham-operated group. This suggests that increased sensitivity of the reward system to ghrelin post-surgery contributes to increased alcohol ingestion [71]. Increased alcohol intake was even observed when dietary obese rats post-RYGB worked to administer alcohol intravenously. These results show that RYGB can increase the rewarding effects of ethanol independent of GI absorption [72]. The increased alcohol intake is not just due to weight loss as rats that were diet restricted to match the weight loss of the RYGB rats did not consume more alcohol [73].

Unlike humans, increased alcohol ingestion is not observed in rats receiving SG [74]. In rats receiving RYGB, when the remnant stomach was removed, the alcohol intake was reduced compared to the sham-operated group. The reduction in alcohol intake may be related to a possible reduction in ghrelin in the absence of the remnant stomach, although ghrelin levels were not measured in that study [74]. Alterations in the gut microbiome may also play a role. The RYGB rats showed lower taxonomic diversity in the fecal microbiome as compared to SG and sham groups. The reduced diversity in RYGB rats conceivably could contribute to increased alcohol intake through the gut-brain-microbiome axis [75]. Lastly, another explanation could be related to ghrelin levels, which decrease following the obesifying high-fat diet. After RYGB, fasted ghrelin levels returned to their previously higher levels but remained lower following SG and the sham operation. The high ghrelin levels may contribute to increased alcohol intake following RYGB vs SG [76].

RYGB surgery in rats also causes damage to the vagus [77]. To test the effects of vagotomy by itself, the gastric vagal branch was severed in one group of rats vs a sham-operated group. The vagotomized rats consumed more 2% and 4% alcohol solution when offered in a two-bottle preference test vs water [77]. This suggests that another mechanism in humans could be due to vagal damage from the RYGB operation.

In sum, the animal studies confirm that RYGB can increase alcohol intake and suggest a potential role of vagal damage as a contributing factor. However, unlike in humans, the animal studies failed to observe an alcohol effect with SG.

4. Strategies to reduce postsurgical alcohol intake

There has not been much work to evaluate prevention and treatment for AUD post-surgery. Bariatric surgery centers are advised to help educate patients about the risks of post-surgical alcohol use, and some of the programs encourage lifelong abstinence. Nevertheless, most patients do consume alcohol after surgery, suggesting that education may be insufficient. Although some patient characteristics have been shown to be predictive of post-surgical alcohol use, identifying patients at high risk of developing an AUD remains challenging [78]. Patients who were consuming alcohol at least twice monthly were interviewed at 1–3 yr post-bariatric surgery. They were asked about their knowledge of the risk of post-surgical alcohol use and reasons for drinking [78]. Although nearly all knew about the risks of alcohol use, most thought that complete abstinence from alcohol post-surgery was not realistic. Reasons given for drinking alcohol included participating in social gatherings, resuming pre-surgical alcohol use, and citing the concept of addiction transfer. Participants consumed alcohol differently compared to pre-surgery, e.g., drinking more slowly and avoiding carbonated mixed drinks and beer. Patients also reported experiencing more intoxicating effects of alcohol in amounts equivalent to pre-surgery. These patient-reported experiences may be leveraged to tailor more effective interventions.

Traditional alcohol relapse prevention strategies can be modified for bariatric patients to provide tailored pre- and post-surgical interventions to decrease alcohol use. Unfortunately, no formal guidelines exist for AUD prevention or treatment, specifically in the post-bariatric population. Regular monitoring of alcohol use should be integrated into standard post-operative care. The Alcohol Use Disorders Identification Test (AUDIT) is a validated screening tool suitable for this purpose, with a score of ≥8 indicating AUD [81,82].

There is some evidence that adding a GLP-1 agonist drug for obesity treatment may help prevent or reduce alcohol intake [81]. Referral for treatment to an alcohol addiction provider or to various lay groups, e.g., Alcoholics Anonymous, should be considered and discussed with the patient when AUD is identified.

Conclusions

Among the proposed mechanisms for increased alcohol intake following bariatric surgery, the most compelling is the enhanced absorption of alcohol due to anatomical changes following RYGB and SG, which also explains the absence of AUD following GB. The addiction transfer hypothesis remains plausible but inadequate, as it does not account for the lack of increased alcohol use after GB. Likewise, greater post-surgical socialization as a mechanism fails to explain the absence of AUD following GB. Animal research implicating damage to the vagus as a factor needs to be confirmed in humans. Regardless of the potential mechanisms, a formalized program manual for the prevention and treatment of AUD is currently needed for bariatric patients.

Highlights.

• Alcohol intake and alcohol use disorder increase following bariatric surgery

• May be related to a higher rate of alcohol absorption post-surgery

• Surgery candidates should be informed of the elevated risk of alcohol use disorder