Psychiatric co-morbidity and substance abuse after gastric bypass surgery

Bariatric surgery1,2 may reduce the incidence of depressive disorders in patients with obesity3–5. However, there are several potential adverse postoperative side effects, both physical (for example, malnutrition, gallstone development and bowel obstruction)1 and behavioural (for example, alcohol-related disorders and opioid use)6–8. Moreover, the rate of self-harm and suicide attempts are higher in this patient group compared with the general population6,9 and, possibly, non-operated patients with obesity10,11. Also, a long-term cohort study found more major depression in obese patientes of both sexes after gastric bypass compared to restrictive surgery12. Several studies have investigated the association between gastric bypass and psychiatric disorders3,5,6 but none has compared those risks with non-operated patients with obesity and matched non-obese population controls. In the current study, we used the Swedish Patient Registry to investigate all 71,488 patients (20-65 years) with a principal diagnosis of obesity between 2001–2013, and compared the long-term incidence of psychiatric co-morbidities in gastric bypass patients with non-operated patients with obesity (and matched population controls without obesity). In this nationwide registry-based cohort study, we included all individuals aged 20–65 years with a first recorded principal diagnosis of obesity (ICD-10 codes E65 or E66) in the Patient Registry between 1 January 2001 and 31 December 2013. For each patient with obesity, two control participants, matched by year of birth, sex and area of residence, and without an obesity diagnosis or bariatric surgery code, were randomly selected from Sweden’s Registry of the Total Population. Outcomes compared between gastric bypass patients and non-operated patients with obesity and matched population controls included depressive disorders, anti-depression medications, neurotic, stress-related and somatoform disorders, alcohol-related disorders, other substance use disorders and suicide. Detailed information on methodology, including data acquisition, inclusions and exclusions, and bias adjustments have been published previously13. See Supplementary material for a description of the methods specific to this study, including codes from the NOMESCO classification of surgical procedures (Table S1), codes from the international definition of diseases (Table S2), definition of co-morbidity and outcomes, bias adjustments and statistical analysis, and study limitations. Among all patients with an obesity diagnosis, 30 214 individuals underwent gastric surgery surgery, of which 28 204 remained after exclusions (Fig. S1). There were 41 274 non-operated individuals with an obesity diagnosis, of which 40 827 remained after exclusions (Fig. S1). Gastric bypass patients and non-operated individuals with an obesity diagnosis were matched with 55 903 and 80 800 population controls, respectively. Table 1 presents baseline study population characteristics. Table 2 presents events, median follow-up times and incidence rates for all outcomes and study groups. Characteristics for gastric bypass patients, non-operated patients with obesity and the two matched population control groups without obesity Missing education, n = 2456 (1.2 per cent). †Missing marital status, n = 33 (0.0 per cent). ‡Dispensed prescription from the pharmacotherapeutic group N06A (according to the Anatomical Therapeutic Chemical classification system, ATC), 2005 and onwards. Characteristics for gastric bypass patients, non-operated patients with obesity and the two matched population control groups without obesity Missing education, n = 2456 (1.2 per cent). †Missing marital status, n = 33 (0.0 per cent). ‡Dispensed prescription from the pharmacotherapeutic group N06A (according to the Anatomical Therapeutic Chemical classification system, ATC), 2005 and onwards. Numbers, events, follow-up times and incidence rates for all psychiatric outcomes c.i.: confidence intervals; IQR: Inter. Numbers, events, follow-up times and incidence rates for all psychiatric outcomes c.i.: confidence intervals; IQR: Inter. Gastric bypass patients were younger and included a higher proportion of women compared to non-operated patients with obesity (Table 1). Individuals diagnosed with obesity had a higher baseline prevalence of psychiatric co-morbidities compared with matched population controls (Table 1). Patients with obesity (operated and non-operated) of both sexes had almost twice the risk of being diagnosed with either depressive disorder or neurotic, stress-related and somatoform disorders (DNSS) compared with the general population (Fig. 1a, Table S3). The risk of DNSS was slightly higher for gastric bypass-operated men compared with non-operated men with obesity (Fig. 1b, Table S4). Overall, antidepression medication had a similar pattern to DNSS (Fig. 1a–b). Hazard ratios Hazard ratios for all outcomes by men and women, comparing a gastric bypass patients and non-operated patients with obesity with non-obese population controls and b gastric bypass patients with non-operated patients with obesity The risk of alcohol-related disorders was 4–5 times higher for gastric bypass-operated individuals of both sexes compared with population controls, and 3–4 times higher compared with non-operated patients with obesity (Fig. 1b, Table S4). The risk was generally higher for men. The risk of substance use disorders showed a similar pattern to alcohol-related disorders but with lower overall risks (Fig. 1a–b, Table S3 and S4). The risk of suicide was higher for gastric bypass patients of both sexes compared with population controls and compared with non-operated patients with obesity (Fig. 1a–b, Table S3), but the overall incidence rates were very low (Table 2). In non-operated patients, only women had a higher risk of suicide compared with population controls (Fig. 1a, Table S3). This is the first long-term prospective cohort study from Sweden that compares the incidence of all major psychiatric co-morbidities in adult gastric bypass patients with that in non-operated patients with obesity and matched population controls. We show that gastric bypass surgery is associated with substantially increased risk of alcohol-related disorders and moderately increased risk of other types of substance abuse, particularly in men. In line with previous studies, both depression and DNSS were associated with obesity at baseline in our study groups14,15. However, so was male sex, lower education level and being single, and neither these nor other baseline characteristics differed markedly between operated and non-operated patients with obesity. Hence, none of the co-morbidities analysed here was fuelling the very high incidence of alcohol-related disorders observed among gastric bypass patients. Our study was the first to study DNSS in gastric bypass patients in Sweden. This outcome was more common in gastric bypass patients compared to non-operated patients with obesity, presenting a possible link to the higher incidence of alcohol-related disorders post-surgery. Earlier studies have shown that neurotic disorders are closely associated with alcoholism, such as in younger patients who use alcohol as a way of coping with nervousness and agitation15,16. However, studies on siblings and twins suggest that these disorders are often a consequence rather than a cause of alcoholism17. Like others6,9, we found a higher risk of suicide among gastric bypass patients compared with non-operated patients with obesity. However, the absolute incidence of suicide was very low (0.1–0.3 per cent) and it is difficult to draw general conclusions based on such sparse data. Other large cohort studies have shown that pre-existing depression and anxiety, male sex and gastric bypass operation led to augmented suicide risk after bariatric surgery11,18. Bypassing the greater part of the stomach, duodenum and parts of the jejunum affect the absorption of ingested food and liquids, including alcohol and other drugs19,20. A recent review on the relationship between bariatric surgery and alcohol use discussed a possible mechanism derived from rodent studies, which was that reduced gastric ghrelin secretion may cause malfunction of the growth hormone secretagogin receptor signalling that is involved in appetite control21, which then leads to increased central dopamine discharge that promotes alcohol intake and alcohol rewarding behaviours. Bypassing the stomach also leads to a reduction of first-pass metabolism resulting in higher blood alcohol content19. This is in part explained by the reduced gastric alcohol dehydrogenase (ADH) activity found after gastric bypass, which can also be seen in healthy females, who have similar gastric emptying times to men but lower ADH activity22. These physical aspects can partly explain why gastric bypass patients, and particularly men, have a higher incidence of alcohol-related disorders. Gastric bypass surgery was strongly associated with increased risk of alcohol-related and other substance use disorders, and potentially suicide, particularly in males. Psychiatric co-morbidity did not explain this pattern; rather, these outcomes were directly associated with gastric bypass surgery, possibly because of changed alcohol and drug absorption and metabolism. The increased risks associated with male sex and gastric bypass surgery deserve improved attention and perhaps also different management before and after surgery. Carl Johan Svensson (Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Writing—original draft), Wai Giang Kok (Conceptualization, Formal analysis, Investigation, Methodology, Writing—review & editing), John Wallert (Methodology, Writing—review & editing), Christian Ruck (Validation, Writing—review & editing), and Christina Lundberg (Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Writing—original draft). This work was financially supported by the Department of Anaesthesia, Operation & Intensive Care, Sahlgrenska University Hospital/Östra, Gothenburg, Sweden, and by the Heart-Lung foundation 2019-0494. The authors declare no conflicts of interest with respect to the present study. Data are available from the sources stated in the paper on request to the data providers, fulfilling legal and regulatory requirements and with permission from the Swedish Ethical Review Authority. Supplementary material is available at BJS online.