Obesity Surgery (2023) 33:1494–1505 https://doi.org/10.1007/s11695-023-06490-w ORIGINAL CONTRIBUTIONS Bariatric Surgery Is Associated with Alcohol‑Related Liver Disease and Psychiatric Disorders Associated with AUD Edilmar Alvarado‑Tapias1,2,3 · David Marti‑Aguado1,4 · Kevin Kennedy5 · Carlos Fernández‑Carrillo1,3,6 · Meritxell Ventura‑Cots1,3,7 · Dalia Morales‑Arraez1,8 · Stephen R. Atkinson1,9 · Ana Clemente‑Sanchez1,3,10 · Josepmaria Argemi1,11 · Ramon Bataller1,12 Received: 1 July 2022 / Revised: 22 January 2023 / Accepted: 30 January 2023 / Published online: 7 March 2023 © The Author(s) 2023 Abstract Background/Aims Bariatric surgery can increase the risk of addictive disorders and nutritional deficiencies. The aim of this study was to evaluate the association between bariatric surgery and alcohol use disorder (AUD), alcohol-related liver disease (ALD), and psychiatric disorders associated with AUD. The impact of vitamin D deficiency in these associations was also investigated. Methods A cross-sectional study was performed using the National Inpatient Sample database and its ICD-9 codes information. Diagnostic and comorbidity data from hospital discharges were obtained from patients with bariatric surgery and other abdominal surgeries between 2005 and 2015. The two groups were then compared for alcohol-related outcomes after propensity-score matching. Results The final study cohort included 537,757 patients with bariatric surgery and 537,757 with other abdominal surger- ies. The bariatric surgery group had an increased risk of AUD [odds ratio (OR): 1.90; 95% CI: 1.85–1.95], ALD [OR: 1.29; 95% CI: 1.22–1.37], cirrhosis [OR, 1.39; 95% CI: 1.37–1.42], and psychiatric disorders associated with AUD [OR, 3.59; 95% CI: 3.37–3.84]. Vitamin D deficiency did not impact in the association between bariatric surgery and AUD, ALD, or psychiatric disorders associated with AUD. Conclusions Bariatric surgery is associated with an increased prevalence of AUD, ALD, and psychiatric disorders associated with AUD. These associations appear to be independent from vitamin D deficiency. Keywords Bariatric surgery · Alcohol use disorder · Alcohol-related liver disease · Alcohol-related mental disorder · Vitamin D Abbreviations ALD Alcohol-related liver disease AUD Alcohol use disorder AMD Alcohol-related mental disorder Edilmar Alvarado-Tapias and David Martí-Aguado have contributed BMI Body mass index equally to this work. CI Confidential interval GLP-1 Glucagon-like peptide-1 Ramon Bataller and Josepmaria Argemi share senior authorship. ICD International Classification of Diseases Key Points LAGB Laparoscopic adjustable gastric banding • Bariatric surgery was associated with an increased prevalence NNIDDK National Institute of Diabetes, Digestive of alcohol use disorder (AUD) and psychiatric disorders and Kidney Diseases associated with AUD. LAB-2 Study NIDDK-funded Longitudinal Assessment • Bariatric surgery was associated with an increased risk of alcohol-related liver disease (ALD) and liver cirrhosis. of Bariatric Surgery-2 • The association between AUD, ALD, and psychiatric disorders NIS National Inpatient Sample associated with AUD was independent of vitamin D deficiency. OR Odds ratio RYGB Roux-en-Y gastric bypass * Ramon Bataller bataller@clinic.cat SD Standard deviation SG Sleeve gastrectomy Extended author information available on the last page of the article 13 Vol:.(1234567890) Obesity Surgery (2023) 33:1494–1505 1495 Introduction sponsored by the Agency for Healthcare Research and Quality. Despite the benefits of bariatric surgery among patients with The NIS is a well-characterized database of hospital clinically severe obesity, this procedure can be associated inpatient stays derived from clinical data submitted by with deleterious health effects, particularly those associ- hospitals to statewide data collection organizations across ated with alcohol use. Data from large cohorts suggest that the USA. It contains a wide range of hospital admission metabolic surgery can increase the risk of alcohol use dis- information, of which it was considered for this study the order (AUD) [1–5]. For this reason, US clinical guidelines following: demographic characteristics, discharge diagno- consider current alcohol or drug use disorders as exclusion sis (outcomes of interest), and comorbidities (past medical criteria for bariatric procedures [6]. However, it is not clear history). Data are collected from a systematic sampling whether a previous metabolic surgery increases the risk of of 20% discharges from all hospitals and stratified by alcohol-related liver disease (ALD) and liver-related events. center, census division, ownership status, urban vs rural Only few data from small samples have investigated the location, teaching status, bed size, and patient diagnosis- association between bariatric surgery and ALD [7–9]. Men- related group, and admission month [23]. This weighting doza et al. reported that more than 50% of postoperative scale provides a sample of hospital discharges from all decompensated liver events were attributable to alcohol use hospitals nationally. The International Classification of [8]. Among the mechanisms involved in this association, the Diseases system (ICD-9-CM codes) was used to identify following have been proposed: (1) changes in alcohol phar- bariatric surgeries, abdominal surgeries, and outcomes of macokinetic, with a rapidly attained maximum blood alcohol interest, as described elsewhere [24]. Table 1S summarizes concentration (BAC) disproportionately high in relation to the the ICD-9 codes used to define the variables of interest in dose consumed [10–13], and (2) changes in hormonal secre- this study. tion, with neuro-hormonal shifts after bariatric surgery deter- Patients admitted during the years 2005–2015, with mining changes in reward processing and dopamine signaling medical history of bariatric surgery and other abdominal [14]. (3) Neurobiological findings in preclinical models with surgeries were included in the study. Cases with abdomi- Roux-en-Y gastric bypass (RYGB) have shown that changes in nal or bariatric surgery were identified using their specific leptin, ghrelin, and GLP-1 secretion could play a major role in ICD-9 codes (Table 1S) among the secondary discharge the regulation of food and alcohol intake [14–18]. diagnosis. The “bariatric surgery group” included sleeve A further potential mechanism linking bariatric surgery gastrectomy, adjustable gastric band, and RYGB. The with liver disease is vitamin D deficiency. RYGB is associated “abdominal surgery group” included appendectomy and with a deficit in the absorption of vitamin D [3, 19], and low cholecystectomy. From eligible population, cases with any levels of vitamin D may predispose patients to certain mental outcome of interest among the pre-surgical comorbidities health disorders such as depression and addictions, including were excluded (any code of AUD, liver diseases, psychiat- AUD [20, 21]. In this line, fewer sunlight hours are associated ric disorders associated with AUD, and/or vitamin D defi- with a higher prevalence of AUD and alcohol-related cirrhosis ciency). To balance the presence of risk factors between [22]. In addition, in patients with cirrhosis and/or ALD, vita- groups, a propensity score match was performed by con- min D deficiency is associated, through different mechanisms, founding variables present prior to surgery. Patient match- with disease severity and mortality [16–19]. ing was performed considering potential factors related to Based on this evidence, the hypothesis of the present liver disease and mental disorder: age, sex, race, obesity, study was that a prior history of bariatric surgery could diabetes, arterial hypertension, renal failure, weight loss, increase the risk of developing ALD and psychiatric dis- anemia, and primary payer. From finally selected popula- orders associated with AUD and that vitamin D deficiency tion, study outcomes were identified from the main pri- could mediate the alcohol-associated outcomes following mary discharge diagnosis. The study design is graphically this procedure. summarized in Fig. 1S. Methods Inclusion and Exclusion Criteria Data Source for the Study Population All hospital discharges between 2005 and 2015 from the NIS database were reviewed for eligibility criteria. A cross-sectional observational study was performed using Cases with bariatric and other abdominal surgeries were the National Inpatient Sample (NIS) database, which is identified with ICD-9 codes and selected for possible part of the Healthcare Cost and Utilization Project inclusion. From the eligible population, cases with the following codes in the pre-surgical comorbidities’ files 13 1496 Obesity Surgery (2023) 33:1494–1505 were excluded: AUD, liver disease, psychiatric disorders To account for the non-random selection of groups, a associated with AUD, vitamin D deficiency, digestive 1:1 propensity score was performed, matching patients by malignant neoplasm, children, and pregnancy. Viral liver potential factors related to liver disease and mental disor- disease was not excluded since the prevalence of this con- der. This matching was performed prior to surgery based dition in both study groups was extremely low (< 0.3%). on the following variables: age, sex, race, obesity, diabetes, hypertension, renal failure, weight loss, anemia, and pri- mary payer. The decision to calculate a propensity score Definition of Study Outcomes and Variables was based on a well-validated method to balance the large number of covariates across the two groups [26, 27]. In this The main study outcomes were AUD, ALD, and psychi- study, the propensity score was defined as the conditional atric disorders associated with AUD. The three of them probability of being in the bariatric surgery group in a mul- are composite outcomes defined by different ICD-9 codes tivariable logistic regression model. A caliper width of 0.2 (Table 1S). Liver disease and psychiatric disorder out- times the SD of the propensity score on the logit scale was comes were composed by specific codes related to their used and determined the balance of matched variables by disease and alcohol. For instance, AUD was defined as using the standardized difference, with values < 10 being the any code related to alcohol abuse, alcohol dependence threshold for a quality match [27]. After matching, the out- syndrome, and/or alcoholism according to the DSM-5 comes of interest were compared between bariatric surgery manual. Recreational drug use was considered an addic- and abdominal surgery using conditional stratified logistic tive disorder with a proper ICD-9 code. ALD was defined regression, accounting for matched pairs as random effects. as the presence of any of the following diagnoses: alco- The results are reported as odds ratios (OR) with 95% con- holic fatty liver, acute alcoholic hepatitis, cirrhosis of fidence intervals (CI). Was used the SAS 9.4 software (SAS the liver with alcoholism, and/or alcoholic liver disease. Institute, Inc.; Cary, NC; USA) for all analyses. Obesity-associated liver disease was defined with the spe- cific code “nonalcoholic fatty liver disease.” Psychiatric disorders associated with AUD were defined with any of Results the following codes: alcohol abuse with psychotic and panic disorders, depression, stress, anxiety, and bipolar Baseline Characteristics syndrome [25]. Depression without AUD was also con- sidered a mental disorder with a proper ICD-9 code. All A total of 82,750,785 hospital discharges were evaluated. Of these conditions being studied were not present prior to these, 2,197,589 patients met the study inclusion criteria; 34% surgery. had undergone bariatric surgery (bariatric surgery group) and The main aim of the study was to compare the preva- 66% other surgeries (abdominal surgery group). Different types lence of ALD and psychiatric disorders associated with of bariatric surgery included a 46.88% of RYGB, 33.76% lapa- AUD in the bariatric surgery group vs abdominal surgery roscopic sleeve gastrectomy (LSG), 10.66% adjustable gas- group. Demographic characteristics and clinical data were tric banding (AGB), and 8.69% others. After propensity-score collected. The influence of vitamin D deficiency was also matching, the final study cohort included 537,757 patients with determined. bariatric surgery and 537,757 patients with other abdominal surgeries. (Fig. 1). All analyses were performed in this final Ethical Considerations The study was performed in accord- matched cohort. The two groups had similar baseline demo- ance with the Declaration of Helsinki and good clinical prac- graphic characteristics and comorbidities (Table 1). The two tice guidelines. The local institutional review board (Univer- groups (bariatric surgery vs abdominal surgery, respectively) sity of Pittsburgh Institutional Revision Board) approved the were comparable with homogeneous proportion of patients study protocol in May 14, 2019 (STUDY19010143). having metabolic comorbidities, including diabetes (18.5% vs 19.2%), hypertension (43.4% vs 46.3%), obesity (16.9% Statistical Analysis vs 19.3%), and renal failure (4.3% vs 5.3%). The prevalence of vitamin D deficiency was significantly higher in the bari- For continuous and categorical variables, data were shown atric surgery group (2.41% vs 0.48%; OR: 5.12 [95% CI: as means ± standard deviation (SD) or n (%). Prevalence 4.91–5.34]; p < 0.001) (Fig. 2S). rates for the main outcome variables were shown as the total prevalence (%) and as overall and yearly fold-changes. Alcohol Use Disorder The fold-change refers to the changes in prevalence in both groups, considering fold-increase changes for each outcome Patients in the bariatric surgery group had a significantly in the bariatric surgery group vs abdominal surgery group. increased risk of AUD (3.32% vs 1.74%; OR: 1.90 [95% 13 Obesity Surgery (2023) 33:1494–1505 1497 Fig. 1  Flow chart showing patient selection from the NIS database. ICD-9-CM: Inter- national Classification of Dis- eases, Ninth Revision—Clinical Modification. *Criteria for selection of patients with “Bari- atric surgery” (ICD-9-CM): Bariatric surgery v45.86, 44.68, 43.82, 44.39. †Criteria for selection of “Abdominal surgery” patients (ICD-9-CM): Appendectomy 47.0, incidental appendectomy 47.1, laparo- scopic appendectomy 47.01, laparoscopic incidental appen- dectomy 47.1, cholecystectomy 51.22, laparoscopic cholecystec- tomy 51.23 CI: 1.85–1.95]; p < 0.001). In the bariatric surgery group, Psychiatric Disorders Associated with AUD the prevalence of AUD increased significantly from 2005 to 2015 (Fig. 2). The magnitude of the association between The prevalence of psychiatric disorders associated with AUD and bariatric surgery was stronger in patients without AUD, recreational drug use, and depression was higher vitamin D deficiency (OR 2.48 in normal vitamin D group in the bariatric surgery group (Fig. 3). Compared to vs OR 1.27 in vitamin D deficiency group). abdominal surgery, bariatric procedures had a higher risk of developing psychiatric disorders associated with Alcohol‑Related Liver Disease AUD (Table 2). Bariatric surgery was more strongly associated with psychiatric disorders in AUD (OR: The prevalence of ALD was significantly higher in the bari- 3.59 [95% CI: 3.37–3.84]) than with alcohol-related atric surgery group (Table 2). Compared to abdominal sur- liver disease (OR: 1.29 [95% CI: 1.22–1.37]) (Fig. 4). gery, bariatric group had a greater risk of ALD (0.71% vs Psychiatric disorders in AUD and bariatric surgery 0.43%, OR: 1.29 [95% CI: 1.22–1.37]; p < 0.001). Among association were similar regardless of vitamin D status patients with AUD, the rate of ALD was higher after bari- (Table 3). atric surgery than other abdominal surgeries (3.4% vs 1.4%; OR: 2.47 [95% CI: 2.45–2.50]; p < 0.001). The prevalence of cirrhosis was also significantly higher in the bariatric sur- Discussion gery group (5.34% vs 3.92%, OR: 1.39 [95% CI: 1.37–1.42]; p < 0.001). The prevalence of both ALD and cirrhosis This study determined whether bariatric surgery was associ- showed a significant increase among bariatric surgery cases ated with the development of liver and mental disorders, and from 2005 to 2015 (Fig. 2). The main etiology of liver dis- the role—if any—of vitamin D deficiency. These findings ease associated to bariatric surgery was nonalcoholic fatty show that patients who underwent bariatric surgery had a liver disease followed by ALD, with a 4.4% and 0.7% preva- significantly increased risk of developing AUD, ALD, cir- lence, respectively (Fig. 3S). The prevalence of liver disease rhosis, and psychiatric disorders associated with AUD. Vita- differed according to vitamin D status (Fig. 4S). Among min D deficiency did not appear to influence the association patients with vitamin D deficiency, ALD prevalence was between bariatric surgery and AUD, ALD, or psychiatric similar between surgical groups (Table 3). disorders associated with AUD. 13 1498 Obesity Surgery (2023) 33:1494–1505 Table 1  Baseline characteristics of hospitals admissions in matched patients with previous bariatric surgery vs abdominal s­ urgery† Bariatric surgery Abdominal surgery Std. diff (%) (n = 537,757) (n = 537,757) Baseline characteristics Gender, M/F (%) 143,365(26.7)/394,392(73.3) 149,171(27.7)/388,586 (72.3) 2.4 Age (yr) 48.78 ± 15 51.51 ± 18 16.3 Ethnic origin (%) Caucasian 389,159 (72.4) 380,259 (70.7) 3.8 Af. American 66,960 (12.5) 71,044 (13.2) Hispanic 56,525 (10.5) 58,908 (11.0) Others 25,113 (4.7) 27,546 (5.1) Primary expected payer (%) Medicare 140,423 (26.1) 153,868 (28.6) 7.6 Medicaid 65,504 (12.2) 67,650 (12.6) Private 284,424 (52.9) 264,819 (49.2) Self-pay 26,260 (4.9) 29,211 (5.4) Median household income national quartile for patient ZIP code (%) 1 124,661 (23.7) 141,065 (26.8) 7.3 2 134,429 (25.6) 132,421 (25.1) 3 138,664 (26.4) 130,878 (24.9) 4 127,975 (24.3) 122,265 (23.2) AHRQ comorbidity measure Deficiency anemias (%) 69,401 (12.9) 73,837 (13.7) 5.2 Rheumatoid (arthritis/collagen vascular) (%) 12,385 (2.3) 10,949 (2.0) 2.4 Chronic blood loss anemia (%) 8,306 (1.5) 5,021 (0.9) 1.8 Congestive heart failure (%) 16,914 (3.1) 25,178 (4.7) 5.5 Chronic pulmonary disease (%) 90,034 (16.7) 75,355 (14.0) 7.9 Coagulopath (%) 13,912 (2.6) 15,224 (2.8) 1.5 Diabetes, uncomplicated (%) 99,618 (18.5) 103,409 (19.2) 1.8 Diabetes (with chronic complications) (%) 12,420 (2.3) 13,995 (2.6) 1.9 Hypertension (%) 233,376 (43.4) 248,876 (46.3) 5.8 Hypothyroidism (%) 66,099 (12.3) 52,016 (9.7) 8.4 Liver disease (%) 31,394 (5.8) 25,603 (4.8) 4.8 Lymphoma (%) 1,500 (0.) 1,884 (0.4) 1.3 Fluid and electrolyte disorders (%) 82,482 (15.3) 104,101 (19.4) 10.6 Metastatic cancer (%) 8,985 (1.7) 10,557 (2.0) 2.2 Other neurological disorders (%) 25,597 (4.8) 20,681 (3.8) 4.5 Obesity (%) 90,672 (16.9) 103,714 (19.3) 6.3 Paralysis (%) 5,226 (1.0) 5,567 (1.0) 0.6 Peripheral vascular disorders (%) 9,584 (1.8) 15,002 (2.8) 6.7 Psychoses (%) 29,146 (5.4) 15,377 (2.9) 12.9 Pulmonary circulation disorders (%) 5,697 (1.1) 6,671 (1.2) 1.7 Renal failure (%) 22,944 (4.3) 28,546 (5.3) 4.9 Solid tumor without metastasis (%) 4,091 (0.8) 6,183 (1.1) 4.0 Peptic ulcer disease (excluding bleeding) (%) 1,113 (0.2) 214 (0.0) 4.8 Valvular disease (%) 9,679 (1.8) 16,017 (3.0) 7.7 Weight loss (%) 24,581 (4.6) 25,983 (4.8) 1.2 Data are presented as mean ± standard deviation or frequencies (%) Abbreviations: AHRQ agency for healthcare research and quality, F female, M male, Yr years, Af African American, Std. diff standard differences † Study groups matched: within the bariatric surgery group, all the patients with previous bariatric surgery. Within the abdominal surgery group, all the patients with other abdominal surgeries according to ICD-9 codes 13 Obesity Surgery (2023) 33:1494–1505 1499 Fig. 2  Impact of bariatric sur- gery on the incidence of alcohol a Alcohol Use Disorder 4 use disorder, alcohol-related p<0.001 liver disease, and liver cirrhosis 5 Prevalence (Fold-increase) from 2005 to 2015. Results are expressed as annual fold 3 Prevalence (%) changes (left side) and overall 4 prevalence rate (right side) during the study period (2005 to 3 2 2015). a Prevalence of alcohol use disorder; b prevalence of 2 alcohol-related liver disease; 1 and c prevalence of liver cirrho- 1 sis in the bariatric surgery and abdominal surgery groups 0 0 2005 2010 2015 Abdominal Bariatric Time (Years) surgery surgery b Alcohol-related Liver Disease 0.8 p<0.001 3 Prevalence (Fold-increase) 0.6 Prevalence (%) 2 0.4 1 0.2 0 0.0 2005 2010 2015 Abdominal Bariatric surgery surgery Time (Years) c Cirrhosis 6 p<0.001 4 Prevalence (Fold-increase) 5 3 Prevalence (%) 4 2 3 2 1 1 0 2005 2010 2015 0 Abdominal Bariatric Time (Years) surgery surgery Bariatric surgery Prevalence Abdominal surgery The motivation to perform the present study was based on bariatric surgery has been shown to have beneficial effects clinical observations in the routine practice, in which some in patients with nonalcoholic steatohepatitis, long-term out- patients who had previously undergone bariatric surgery comes need to be interpreted with caution [28–30]. Moreo- developed ALD, including alcoholic hepatitis. Although ver, given that many clinicians are not aware that bariatric 13 1500 Obesity Surgery (2023) 33:1494–1505 Table 2  Alcohol-related organ Bariatric surgery Abdominal surgery p value damage in both groups matched, (n = 537,757) (n = 537,757) with bariatric surgery vs abdominal ­surgery† Alcohol use disorder, (%) 17,853 (3.32) 9,359 (1.74) < 0.001 Alcohol-related liver disease, (%) 3,818 (0.71) 2,335 (0.43) < 0.001 Liver cirrhosis/complications, (%) 28,716 (5.34) 21,080 (3.92) < 0.001 Psychiatric disorders associated with 4,180 (0.80) 1,169 (0.21) < 0.001 AUD, (%) Recreational drugs use, (%) 13,981 (2.6) 6,453 (1.2) < 0.001 Depression, (%) 108,089 (20.1) 31,189 (5.8) < 0.001 Data are presented as frequencies (%) † Study groups not matched: Within the bariatric surgery group, all the patients with previous bariatric sur- gery. Within the abdominal surgery group, all the patients with other abdominal surgeries according to ICD-9 codes Table 3  Alcohol-related diseases in the matched cohort, stratified by vitamin D status in bariatric surgery group vs abdominal surgery group Vitamin D deficiency Clinical outcome Bariatric surgery Abdominal surgery OR (95% CI) p value (n = 17,677) (n = 3,562) Alcohol use disorder 2.90% 2.30% 1.27 (1.02–1.61) 0.04 Alcohol-related liver disease 0.40% 0.31% 1.25 (0.66–1.37) 0.40 Liver cirrhosis/complications 10.5% 7.0% 1.56 (1.36–1.78) < 0.001 Psychiatric disorders associated with AUD 0.5% 0.1% 3.88 (1.58–9.55) 0.003 Vitamin D normal Clinical outcome Bariatric surgery Abdominal surgery OR (95% CI) p value (n = 729,202) (n = 1,447,148) Alcohol use disorder 3.40% 1.40% 2.48 (2.44–2.53) < 0.001 Alcohol-related liver disease 0.61% 0.10% 3.90 (3.70–4.11) < 0.001 Liver cirrhosis/complications 5.20% 2.50% 2.09 (2.06–2.12) < 0.001 Psychiatric disorders associated with AUD 0.80% 0.21% 4.79 (4.57–5.02) < 0.001 Data are presented as frequencies (%). Odds ratio (OR) (95%, confidential interval) and p value surgery may predispose patients to AUD, it was considered previously assessed in the LABS-2 (Longitudinal Assess- necessary to study the relationship between metabolic sur- ment of Bariatric Surgery-2) prospective cohort studies, gery and alcohol-related outcomes, such as liver diseases which showed a cumulative incidence of AUD symptoms and psychiatric disorders associated with AUD. of 10–21% in the 2–5 years period after surgery [1, 2]. Our The data for this study was obtained from the NIS data- results are similar, although the prevalence of AUD was base. Other studies have also used this database, as to dem- lower. Different prevalences may be due to several factors: onstrate an increasing prevalence of ALD cirrhosis in the first, the underreporting of psychiatric diagnoses in coding- U.S [31–33]. Diagnostic and comorbidity data from all based medical record databases. Our study highlights this hospital discharges between 2005 and 2015 was analyzed. major issue that should alert clinicians to evaluate AUD in The outcomes being studied were excluded prior to surgery. bariatric surgery patients. Second, previous studies deter- Given that patients undergoing bariatric surgery are consid- mined AUD symptoms with the Alcohol Use Disorders ered to have a major risk factor for liver disease, a propen- Identification Test (AUDIT) which is more specific. Third, sity match was also performed prior to surgery to reduce previous studies included preoperative AUD cases with the impact of confounding variables. This design allowed a prevalence of 8–10% [1, 35]. Our study did not include for the comparison between groups and to establish risk preoperative AUD cases, showing a similar fold increase in associations. postoperative AUD prevalence. Nevertheless, beyond coding Our data show that AUD was more prevalent among and AUDIT evaluation, objective alcohol biomarkers such patients who have undergone bariatric surgery, confirm- as phosphatidylethanol in whole blood or ethyl glucuronide ing previous experimental and clinical studies [1, 2, 5, 34]. in hair specimens are needed to establish the true prevalence The association between bariatric surgery and AUD was of AUD in this population. 13 Obesity Surgery (2023) 33:1494–1505 1501 Fig. 3  Impact of bariatric sur- a Recreational drugs gery on the incidence of psychi- 3 p<0.001 atric disorders associated with AUD. Results are expressed as annual fold changes (left 6 Prevalence (Fold-increase) side) and overall prevalence Prevalence (%) 5 2 rate (right side) during the study period (2005 to 2015). a 4 Prevalence of recreational drugs use; b prevalence of psychiatric 3 disorders associated with AUD; 1 and c prevalence of depression 2 in the bariatric surgery and 1 abdominal surgery groups 0 0 Abdominal Bariatric 2005 2010 2015 surgery surgery Time (Years) b Psychiatric disorders associated with AUD 1.0 p<0.001 15 0.8 Prevalence (Fold-increase) Prevalence (%) 0.6 10 0.4 5 0.2 0 0.0 2005 2010 2015 Abdominal Bariatric Time (Years) surgery surgery c Depression 25 p<0.001 6 20 Prevalence (Fold-increase) Prevalence (%) 5 15 4 3 10 2 5 1 0 0 2005 2010 2015 Abdominal Bariatric surgery surgery Time (Years) Bariatric surgery Prevalence Abdominal surgery Several mechanisms have been suggested to explain and higher peak of BAC mainly in RYGB, disproportion- the association between bariatric surgery and AUD: (1) ately high compared to the dose consumed [10–12, 34]; (2) changes in alcohol pharmacokinetic profile, with an earlier changes in the taxonomic profile of the gut microbiota after 13 1502 Obesity Surgery (2023) 33:1494–1505 Fig. 4  Alcohol-related organ Alcohol-related organ damage damage in the bariatric surgery group. Results are expressed Patients with previous Bariatric Surgery as odds ratios (OR) with 95% Odds Ratio (95% CI) confidence intervals (CI). The OR were calculated using con- Alcohol use disorder 1.90 (1.85, 1.95) ditional stratified logistic regres- sion accounting for matched pairs as random effects Alcohol-related liver disease 1.29 (1.22, 1.37) Liver cirrhosis/complications 1.39 (1.37, 1.42) Psychiatric disorders associated with AUD 3.59 (3.37, 3.84) 1 2 4 <<< Less likely More likely >>> bariatric surgery, particularly RYGB, are associated with psychiatric disorders associated with AUD were also more an increased alcohol intake and may contribute to increased prevalent in the bariatric surgery group. Although psychiat- AUD risk through the gut-brain-microbiome axis [36]; ric disorder and alcohol abuse developed after surgery, it is and (3) neuro-hormonal shifts after RYGB that determine not possible to know which of the two disorders led to the changes in reward processing [14, 15, 17, 18, 37, 38]. Con- other one [25]. One of the mechanisms underlying these sidering these mechanisms, related organ damage is likely comorbidities could be modifications in the pharmacokinet- from a high but quickly resolving peak of BAC. Our data ics of alcohol [11, 34], which generates greater susceptibility support the need for more clinical and basic research in this to its effects, thus leading to alcohol-mediated organ dam- area to better elucidate which of these postulated mecha- age. Given the weight loss associated with bariatric surgery, nisms have the greatest impact. Such studies would also be patients who drink will experience an increased tissue dose useful to identify ways to target or prevent AUD in patients postoperatively, even if they do not increase their alcohol undergoing bariatric surgery. intake. In the study group was observed an increase in the In this study, it was also found that ALD and liver cirrho- prevalence of the main outcome variables (AUD, ALD, sis were more prevalent in the bariatric surgery group, with psychiatric disorders associated with AUD over the 10-year a trend towards significant annual increases in the ten-year study period), which could be partially explained by the study period (Fig. 2). These results are consistent with the greater awareness of alcohol-related diseases and nutritional recent findings of a study performed to assess the poten- deficiencies in recent years. Increased awareness of these tial long-term increased risk of alcohol-related cirrhosis in conditions may have led to an increase in the diagnosis and women who underwent bariatric surgery [4, 24]. Although subsequent ICD coding for these disorders. statistically significant, the bariatric surgery group had a low Another mechanism that has been proposed to explain prevalence of ALD (0.7%). Within selected patients with the association between bariatric surgery and alcohol-related AUD, ALD prevalence raised up to 3.4%. Despite of these liver diseases is vitamin D deficiency [39–41]. However, low prevalences, the clinical significance of our findings in this study, the association between bariatric surgery and from a large sample database relies on the 29% increased ALD seemed to be independent from vitamin D deficiency, risk of developing ALD (OR: 1.29) after bariatric surgery indicating that other mechanisms besides nutritional defi- in comparison to other abdominal surgeries. To differenti- ciencies may be involved. ate obesity-associated liver disease from AUD-associated The findings of this study, based on a large national data- liver disease, different ICD codes were used to define these set, have a practical clinical application. This is the need to conditions, and a propensity score analysis was performed increase awareness about the association between bariatric considering obesity as a confounding variable. The overall surgery and alcohol-related issues and thus the need for pre- prevalence of liver disease in bariatric surgery patients was vention and education, as well as the monitoring of alcohol 4.4% and 0.7% for nonalcoholic fatty liver disease and ALD, use and liver injury before and after bariatric procedures. respectively. The current study has several limitations. First, due to the Despite a higher risk of mental disorders among patients cross-sectional design, it is possible to identify associations undergoing bariatric surgery, the current study takes a step but not to test for causality. Second, the data source has a further and focuses on psychiatric disorders associated limited capacity to verify the quality of the diagnostic and with AUD. Beyond recreational drug use and depression, comorbidity coding at hospital discharge. Third, bariatric 13 Obesity Surgery (2023) 33:1494–1505 1503 surgery included procedures with different reported risks for patients who present advanced liver disease and/or alcohol new onset AUD in previously published articles [1, 2, 35]. misuse [45]. Finally, although our data indicate that vita- The most frequent procedures were RYGB and LSG (around min D deficiency does not appear to be associated with the 80%) which have shown a higher risk of AUD compared to risks of AUD, ALD, or psychiatric disorders associated with AGB [1, 2, 35]. Procedures other than RYGB and LSG were AUD, further studies should explore other potential mecha- included because the prevalence of AUD post-AGB is non- nisms underlying the development of these conditions after neglectable (up to 11% in the LABS-2 cohort) [2], and they bariatric surgery. represented a minor proportion of the surgical procedures Supplementary Information The online version contains supplemen- from the NIS database. Still, a risk association between tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 11695-0​ 23-0​ 6490-w. AUD and bariatric surgery was seen, although interpreta- tion should be mainly subject to RYGB and LSG procedures. Acknowledgements We are very grateful to David Bridgwater for his Further studies should evaluate whether these different inter- grammar check and editing of this manuscript. ventions differ in terms of incident ALD and psychiatric Author Contribution EAT, DMA, JA, and RB: study concept and disorders associated with AUD, among others. Fourth, AUD design. EAT, CFC, MVC, DMA(II), SRA, and ACS: data acquisition. and psychiatric disorders associated with AUD can be under- EAT, DMA, JA and RB: analysis and interpretation of data. KK: sta- reported and underrecognized, particularly in patients being tistical analysis. EAT and DMA: drafting of manuscript. CFC, MVC, DMA(II), SRA, ACS, JA, and RB: critical revision of the manuscript. treated for medical conditions unrelated to alcohol use, and All authors approved the final version of the article: EAT, DMA, KK, especially in databases based on ICD codes [42]. Finally, the CFC, MVC, DMA(II), SRA, ACS, JA, and RB. analysis was performed until 2015 because the clinical terms used corresponded to ICD-9 codes. At the end of 2015, codi- Funding Open Access funding provided thanks to the CRUE- CSIC agreement with Springer Nature. Ramón Bataller is a recipi- fication was updated from ICD-9 to ICD-10 codes, resulting ent of NIAAA grants U01AA021908 and U01AA020 and NIDDK in the unavailability of previous databases. P30DK120531821. Dalia Morales-Arraez, Meritxell Ventura-Cots, Notwithstanding the potential limitations described Carlos Fernández-Carrillo, Edilmar Alvarado-Tapias, and Ana Clem- above, our data show that the prevalence of AUD, ALD, ente-Sanchez are recipients of a scholarship grant for study extension abroad, sponsored by the Spanish Association for the Study of the and psychiatric disorders associated with AUD was higher Liver (AEEH). Edilmar Alvarado-Tapias is recipient of a Río Hort- in patients who had undergone bariatric surgery than ega award ISCIII Instituto de Salud Carlos III (Spanish Ministry of other abdominal surgeries. Although prospective studies Science and Innovation) (CM16/00133) and Joan Rodes award ISCIII are needed to better determine alcohol use and whether (JR20/00047). David Marti-Aguado is recipient of a Río Hortega award ISCIII (CM19/00212) and Joan Rodes award ISCIII (JR22/00002). there is a causal relationship between metabolic surgery Josepmaria Argemi is recipient of the PI20/01663 grant from the ISCIII and alcohol-related diseases, our results underscore the and an award from the “Fundación Echebano” (Pamplona, Spain). importance of counseling for alcohol-associated disorders in patients undergoing bariatric surgery. In this regard, all Declarations candidates for bariatric surgery should undergo in-depth Conflict of Interest The authors declare no conflict of interest regard- clinical interviews to detect patterns of alcohol use and ing this manuscript. risk factors for AUD, including a comprehensive assess- ment of specific alcohol biomarkers (such as phosphati- Open Access This article is licensed under a Creative Commons Attri- dylethanol or ethyl glucuronide) with blood, urine, or hair bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long samples analysis before and after bariatric surgery [43, as you give appropriate credit to the original author(s) and the source, 44]. Prospective studies are needed to evaluate long-term provide a link to the Creative Commons licence, and indicate if changes outcomes in this patient population. Ideally, these studies were made. The images or other third party material in this article are would include regular clinical evaluations with biospeci- included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in men collection performed at baseline and annually thereaf- the article's Creative Commons licence and your intended use is not ter. This study design would allow us to establish causality permitted by statutory regulation or exceeds the permitted use, you will between bariatric surgery and the subsequent development need to obtain permission directly from the copyright holder. To view a of alcohol-related conditions. In addition, it would help copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. to better elucidate the molecular mechanisms underlying these associations and to identify high-risk patients. In summary, our data suggest that the evident benefits of bariatric surgery must be weighed against the risks of References developing AUD, which can lead to ALD and/or psychiatric 1. King WC, Chen J-Y, Mitchell JE, Kalarchian MA, Steffen KJ, disorders associated with AUD. It is especially important Engel SG, et al. Prevalence of alcohol use disorders before and to exercise caution before performing bariatric surgery in after bariatric surgery. JAMA. 2012;307:2516–25. 13 1504 Obesity Surgery (2023) 33:1494–1505 2. King WC, Chen JY, Courcoulas AP, Dakin GF, Engel SG, Flum 19. Mangan A, le Roux CW, Miller NG, Docherty NG. Iron and DR, et al. Alcohol and other substance use after bariatric surgery: vitamin D/calcium deficiency after gastric bypass: mechanisms prospective evidence from a U.S. multicenter cohort study. Surg involved and strategies to improve oral supplement disposition. Obes Relat Dis. 2017;13:1392–402. Curr Drug Metab. 2018;20:244–52. 3. Syn NL, Lee PC, Kovalik JP, Tham KW, Ong HS, Chan WH, 20 Tardelli VS, do Lago MPP, da Silveira DX, Fidalgo TM. Vitamin et al. Associations of bariatric interventions with micro- D and alcohol: a review of the current literature. Psychiatry Res. nutrient and endocrine disturbances. JAMA Netw Open. 2017;248:83–6. 2020;3:e205123. 21. Neupane SP, Lien L, Hilberg T, Bramness JG. Vitamin D defi- 4. White GE, Courcoulas AP, Richardson GA, Mair C, King WC. ciency in alcohol-use disorders and its relationship to comorbid Alcohol use thresholds for identifying alcohol-related problems major depression: a cross-sectional study of inpatients in Nepal. before and following Roux-en-Y gastric bypass. Ann Surg. Drug Alcohol Depend. 2013;133:480–5. 2019;269:1001–9. 22. Ventura-Cots M, Watts AE, Cruz-Lemini M, Shah ND, Ndugga N, 5. Östlund MP, Backman O, Marsk R, Stockeld D, Lagergren J, McCann P, et al. Colder weather and fewer sunlight hours increase Rasmussen F, et al. Increased admission for alcohol dependence alcohol consumption and alcoholic cirrhosis worldwide. Hepatol- after gastric bypass surgery compared with restrictive bariatric ogy. 2019;69:1916–30. surgery. JAMA Surg. 2013;148:374. 23. Houchens RL, Ross DN, Elixhauser A, Jiang J. Nationwide inpa- 6. Mechanick JI, Apovian C, Brethauer S, Garvey WT, Joffe AM, tient sample redesign final report. HCUP NIS Related Reports Kim J, et al. Clinical practice guidelines for the periopera- ONLINE. 442014. U.S. Agency for Healthcare Research and tive nutrition, metabolic, and nonsurgical support of patients Quality. 2014. https://​www.​hcup-​us.a​ hrq.​gov/​db/​nation/​nis/​repor​ undergoing bariatric procedures - 2019 update: cosponsored by ts/​NISRe​desig​nFina​lRepo​rt040​914.​pdf American Association of Clinical Endocrinologists/American 24. Mellinger JL, Shedden K, Winder GS, Tapper E, Adams M, College of Endocrinology. Endocr Pract. 2019;25:1346–59. Fontana RJ, et al. The high burden of alcoholic cirrhosis in 7. Mellinger JL, Shedden K, Winder GS, Fernandez AC, Lee BP, privately insured persons in the United States. Hepatology. Waljee J, et al. Bariatric surgery and the risk of alcohol-related 2018;68:872–82. cirrhosis and alcohol misuse. Liver Int. 2021;41:1012–9. 25. Castillo-Carniglia A, Keyes KM, Hasin DS, Cerdá M. Psychi- 8. Mendoza YP, Becchetti C, Wan T, Nett P, Rodrigues SG, Dufour atric comorbidities in alcohol use disorder. Lancet Psychiatry. JF, et al. Malnutrition and alcohol in patients presenting with 2019;6:1068–80. severe complications of cirrhosis after laparoscopic bariatric 26. Austin PC. Optimal caliper widths for propensity-score matching surgery. Obes Surg. 2021;31:2817–22. when estimating differences in means and differences in propor- 9. Mendoza YP, Becchetti C, Watt KD, Berzigotti A. Risks and tions in observational studies. Pharm Stat. 2011;10:150–61. rewards of bariatric surgery in advanced chronic liver diseases. 27. Austin PC, Grootendorst P, Anderson GM. A comparison of the Semin Liver Dis Semin Liver Dis. 2021;41:448–60. ability of different propensity score models to balance measured 10. Acevedo MB, Teran-Garcia M, Bucholz KK, Eagon JC, Bar- variables between treated and untreated subjects: a Monte Carlo tholow BD, Burd NA, et al. Alcohol sensitivity in women after study. Stat Med. 2007;26:734–53. undergoing bariatric surgery: a cross-sectional study. Surg Obes 28. Lassailly G, Caiazzo R, Ntandja-Wandji LC, Gnemmi V, Baud G, Relat Dis. 2020;16:536–44. Verkindt H, et al. Bariatric surgery provides long-term resolution 11. Steffen KJ, Engel SG, Pollert GA, Li C, Mitchell JE. Blood of nonalcoholic steatohepatitis and regression of fibrosis. Gastro- alcohol concentrations rise rapidly and dramatically after Roux- enterology. 2020;159:1290-1301.e5. en-Y gastric bypass. Surg Obes Relat Dis. 2013;9:470–3. 29. Lee Y, Doumouras AG, Yu J, Brar K, Banfield L, Gmora S, et al. 12. Acevedo MB, Eagon JC, Bartholow BD, Klein S, Bucholz KK, Complete resolution of nonalcoholic fatty liver disease after bari- Pepino MY. Sleeve gastrectomy surgery: when 2 alcoholic atric surgery: a systematic review and meta-analysis. Clin Gastro- drinks are converted to 4. Surg Obes Relat Dis. 2018;14:277–83. enterol Hepatol. 2019;17:1040-1060.e11. 13. Woodard GA, Downey J, Hernandez-Boussard T, Morton JM. 30. Miñambres I, Rubio MA, de Hollanda A, Breton I, Vilarrasa N, Impaired alcohol metabolism after gastric bypass surgery: a Pellitero S, et al. Outcomes of bariatric surgery in patients with case-crossover trial. J Am Coll Surg. 2011;212(2):209–14. cirrhosis. Obes Surg. 2019;29:585–92. 14. Orellana ER, Covasa M, Hajnal A. Neuro-hormonal mecha- 31. Barritt AS, Jiang Y, Schmidt M, Hayashi PH, Bataller R. Charges nisms underlying changes in reward related behaviors following for alcoholic cirrhosis exceed all other etiologies of cirrhosis com- weight loss surgery: Potential pharmacological targets. Biochem bined: a national and state inpatient survey analysis. Dig Dis Sci. Pharmacol. 2019;164:106–14. 2019;64:1460–9. 15. Beckman LM, Beckman TR, Sibley SD, Thomas W, Ikramuddin 32. Moon AM, Yang JY, Barritt AS, Bataller R, Peery AF. Rising S, Kellogg TA, et al. Changes in gastrointestinal hormones and mortality from alcohol-associated liver disease in the United leptin after Roux-en-Y gastric bypass surgery. JPEN J Parenter States in the 21st century. Am J Gastroenterol. 2020;115(1):79–87. Enteral Nutr. 2011;35(2):169–80. 33. Hirode G, Saab S, Wong RJ. Trends in the burden of chronic 16. Orellana E, Horvath N, Farokhnia M, Leggio L, Hajnal A. liver disease among hospitalized US adults. JAMA Netw Open. Changes in plasma ghrelin levels following surgical and non- 2020;3(4): e201997. surgical weight-loss in female rats predict alcohol use. Brain 34. Polston JE, Pritchett CE, Tomasko JM, Rogers AM, Leggio L, Res Bull. 2022;188:179–86. Thanos PK, et al. Roux-en-Y gastric bypass increases intrave- 17. Hajnal A, Zharikov A, Polston JE, Fields MR, Tomasko J, nous ethanol self-administration in dietary obese rats. PLoS One. Rogers AM, et al. Alcohol reward is increased after Roux-en-Y 2013;8:283741. gastric bypass in dietary obese rats with differential effects fol- 35. Ibrahim N, Alameddine M, Brennan J, Sessine M, Holliday C, lowing ghrelin antagonism. PLoS One. 2012;7:e49121. Ghaferi AA. New onset alcohol use disorder following bariatric 18. Sirohi S, Richardson BD, Lugo JM, Rossi DJ, Davis JF. Impact surgery. Surg Endosc. 2019;33:2521–30. of Roux-en-Y gastric bypass surgery on appetite, alcohol intake 36. Martin OA, Grant-Beurmann S, Orellana ER, Hajnal A, Fraser behaviors, and midbrain ghrelin signaling in the rat. Obesity. CM. Changes in the gut microbiota following bariatric surgery 2017;25:1228–36. 13 Obesity Surgery (2023) 33:1494–1505 1505 are associated with increased alcohol intake in a female rat model. 42. Grant BF, Goldstein RB, Saha TD, Patricia Chou S, Jung Alcohol Alcohol. 2021;56:605–13. J, Zhang H, et al. Epidemiology of DSM-5 alcohol use 37. Shin AC, Zheng H, Townsend RL, Sigalet DL, Berthoud HR. disorder: results from the national epidemiologic survey Meal-induced hormone responses in a rat model of roux-en-Y on alcohol and related conditions III. JAMA Psychiat. gastric bypass surgery. Endocrinology. 2010;151:1588–97. 2015;72:757–66. 38. Orellana ER, Jamis C, Horvath N, Hajnal A. Effect of vertical 43. Neumann J, Beck O, Helander A, Böttcher M. Performance of sleeve gastrectomy on alcohol consumption and preferences in PEth compared with other alcohol biomarkers in subjects present- dietary obese rats and mice: a plausible role for altered ghrelin ing for occupational and pre-employment medical examination. signaling. Brain Res Bull. 2018;138:26–36. Alcohol Alcohol. 2020;55:401–8. 39. Udomsinprasert W, Jittikoon J. Vitamin D and liver fibrosis: 44. Staufer K, Huber-Schönauer U, Strebinger G, Pimingstorfer P, Molecular mechanisms and clinical studies. Biomed Pharmaco- Suesse S, Scherzer TM, et al. Ethyl glucuronide in hair detects ther. 2019;109:1351–60. a high rate of harmful alcohol consumption in presumed non- 40. Trépo E, Ouziel R, Pradat P, Momozawa Y, Quertinmont E, Gervy alcoholic fatty liver disease. J Hepatol. 2022;77:918–30. C, et al. Marked 25-hydroxyvitamin D deficiency is associated 45. Muthiah MD, Sanyal AJ. Current management of non-alcoholic with poor prognosis in patients with alcoholic liver disease. J steatohepatitis. Liver Int. 2020;40(Suppl 1):89–95. Hepatol. 2013;59:344–50. 41. Cimini FA, Barchetta I, Carotti S, Morini S, Cavallo MG. Over- Ramon Bataller guarantees the article. view of studies of the vitamin D/vitamin D receptor system in the development of non-alcoholic fatty liver disease. World J Gastro- Publisher's Note Springer Nature remains neutral with regard to intest Pathophysiol. 2019;10:11–6. jurisdictional claims in published maps and institutional affiliations. Authors and Affiliations Edilmar Alvarado‑Tapias1,2,3 · David Marti‑Aguado1,4 · Kevin Kennedy5 · Carlos Fernández‑Carrillo1,3,6 · Meritxell Ventura‑Cots1,3,7 · Dalia Morales‑Arraez1,8 · Stephen R. Atkinson1,9 · Ana Clemente‑Sanchez1,3,10 · Josepmaria Argemi1,11 · Ramon Bataller1,12 1 7 Center for Liver Diseases, Division of Gastroenterology, Liver Unit, Hospital Universitari Vall d’Hebron, Vall Hepatology and Nutrition, Department of Medicine, d’Hebron Institute of Research (VHIR), Universitat University of Pittsburgh Medical Center, 201.19 Kaufmann Autònoma de Barcelona, Barcelona, Spain Medical Building, 3471 Fifth Avenue, Pittsburgh, PA 15213, 8 Department of Gastroenterology, Hospital Universitario de USA Canarias, Tenerife, Spain 2 Department of Gastroenterology, Hospital Santa Creu I Sant 9 Department of Metabolism, Digestion and Reproduction, Pau, Institut de Recerca Sant Pau, Universidad Autónoma de Faculty of Medicine, Imperial College London, London, UK Barcelona, Barcelona, Spain 10 3 Liver Unit, Department of Digestive Diseases, Hospital Centro de Investigación Biomédica en Red de Enfermedades General Universitario Gregorio Marañón, Madrid, Spain Hepáticas Y Digestivas (CIBERehd), Instituto de Salud 11 Carlos III, Madrid, Spain Liver Unit, Clínica Universidad de Navarra (CUN), 4 Hepatology Program, Centro de Investigacion Medica Digestive Disease Department, Clinic University Hospital, Aplicada (CIMA), Instituto de Investigación de Navarra Biomedical Research Institute INCLIVA, Valencia, Spain (IdisNA), University of Navarra, Pamplona, Spain 5 Kansas City, MO, USA 12 Department of Medicine, Faculty of Medicine and Health 6 Liver Unit, Department of Gastroenterology and Hepatology, Sciences, University of Barcelona, Barcelona, Spain Hospital Universitario Puerta de Hierro-Majadahonda, IDIPHISA, Madrid, Spain 13