Timing and management of bleeding after bariatric surgery.

Abstract

The timing of bleeding after bariatric surgery and subsequent management (characterized as surgical versus non-surgical (i.e., interventions including endoscopic or interventional radiology approaches)) has not been thoroughly studied. As such, we sought to describe the rates of reoperation or non-operative intervention after bleeding following sleeve gastrectomy (SG) or Roux-en-Y gastric bypass (RYGB). The Metabolic and Bariatric Accreditation and Quality Improvement Program (MBSAQIP) database was queried between 2015 and 2018 for any bleeding after SG or RYGB and subsequent reoperation or non-operative intervention. Multivariable Fine-Gray models were used to compare the hazard of reoperation/non-operative intervention. Multivariable generalized linear regression models were used to test the number of subsequent reoperations/non-operative interventions depending on initial management. 6251 patients with bleeding after SG or RYGB were identified, of which 2653 patients underwent subsequent procedures (n = 1375 [51.83%] RYGB index procedure, n = 1278 [48.17%] SG index procedure). 1892 (71.32%) and 761 (28.68%) patients had reoperation and non-operative intervention, respectively. For patients who developed bleeding, SG was associated with significantly higher reoperation risk, while RYGB was associated with significantly higher risk of non-operative intervention. Early bleeding was associated with significantly increased risk of reoperation and decreased risk of non-operative intervention, regardless of initial procedure. The total number of subsequent reoperations/non-operative interventions did not differ significantly depending on whether the patients had non-operative intervention or reoperation first [ratio 1.01, 95% CI (0.75, 1.36), p value 0.9418]. Patients after SG who experience bleeding are more likely to undergo reoperation than RYGB patients. On the other hand, patients with bleeding after RYGB are more likely to undergo non-operative intervention compared to SG patients. Early bleeding is associated with higher risk of reoperation and lower risk of non-operative intervention both after SG and RYGB. The initial approach did not play a role in the total number of subsequent reoperations/non-operative interventions.