A 13-year nationwide analysis of outcomes of non-variceal upper gastrointestinal bleeding in post-bariatric surgery patients.

Background: Non-variceal upper gastrointestinal bleeding (NVUGIB) is a common cause of hospitalization and associated morbidity and mortality. Bariatric surgery is a widely performed category of gastrointestinal (GI) surgery that attempts to induce weight loss by reconstructing the upper GI tract. Bleeding is a common complication of bariatric surgery; however, limited research exists on outcomes for these patients when admitted for NVUGIB compared to those without a history of bariatric surgery. Our study aims to evaluate the outcomes of post-bariatric surgery patients (PBSPs) admitted with NVUGIB over a 13-year span.

Methods: The National Inpatient Sample (NIS) from 2008 to 2020 was used to identify patients over 18 years old admitted for NVUGIB using the International Classification of Disease (ICD), 9^th revision (ICD-9) and 10^th revision (ICD-10) codes. Records were weighted using the algorithms provided by the NIS. Primary outcomes of interest were all-cause hospital mortality, shock, acute myocardial infarction (AMI), acute kidney injury (AKI), and a composite of these. Groups were defined based on the history of bariatric surgery, and demographics and incidence of comorbidities were compared. Outcomes were compared between the two groups, and odds ratios (ORs) were calculated using two-stage weighted logistic regression. ORs were adjusted for common co-founders such as age, gender, race, Charlson comorbidity index (CCI), region, hospital size, hospital teaching status, elective vs. emergency admission, and income quartile.

Results: A total of 2,231,826 patients admitted for NVUGIB were included in this study. Of these, 28,167 had a history of bariatric surgery. Overall, bariatric surgery patients were younger, less complicated (CCI: 2.71 vs. 4.76), had a shorter length of stay (LOS), and were less likely to be on Medicare/Medicaid than those without a history of bariatric surgery (P<0.05). Several comorbidities were more common in patients without a history of bariatric surgery, including coronary artery disease, hyperlipidemia (HLD), congestive heart failure (CHF), type 2 diabetes mellitus (T2DM), end-stage renal disease (ESRD), and liver cirrhosis (P<0.05). Other comorbidities were more common in the post-bariatric surgery group, including gastroesophageal reflux disease and anemia (P<0.05). We found that patients admitted for NVUGIB with a history of bariatric surgery had significantly lower odds of all outcomes, including all-cause mortality (OR =0.48; P<0.001), AKI (OR =0.71; P<0.001), AMI (OR =0.62; P<0.05), shock (OR =0.88; P<0.05), and a composite of these four (OR =0.77; P<0.001).

Conclusions: Our study found that patients with NVUGIB and a history of bariatric surgery had substantially decreased odds of mortality, AMI, shock, and AKI compared to patients without a history of bariatric surgery. This suggests that patients with a history of bariatric surgery experienced more minor bleeds than other patients. Further studies are therefore warranted to understand the postoperative risk of bariatric surgery.