Do all roads lead to Rome? A retrospective analysis on surgical technique in sleeve gastrectomy

Background: New York University Langone Health has three accredited bariatric centers, with 10 different bariatric surgeons. This retrospective analysis compares surgeon techniques in laparoscopic or robotic sleeve gastrectomy (SG) to identify associations with perioperative morbidity and mortality.

Methods: All adults who underwent SG between 2017 and 2021 at NYU Langone Health were evaluated via EMR and MBSAQIP 30-day data. We also surveyed all 10 bariatric surgeons and compared their techniques and total adverse outcomes. Bleeding, SSI, mortality, readmission, and reoperation were specifically sub-analyzed via logistic regression.

Results: 86 (2.77%) out of 3,104 patients who underwent SG encountered an adverse event. Lower adverse outcomes were observed with a laparoscopic approach, 40-Fr bougie, buttressing, not oversewing the staple line, using hemostatic agents, stapling 3-cm from pylorus, and no routine UGI series. Lower bleeding rates were observed in a laparoscopic approach, 40-Fr bougie, buttressing, not oversewing the staple line, using hemostatic agents, stapling 3-cm from pylorus, no routine UGI series, and not proceeding with SG if hiatal hernia is present. Lower SSI rates were observed with ViSiGi™ bougie, no hemostatic agents, and routine EGD. Lower readmission rates were observed with 40-Fr bougie, buttressing, not oversewing, and stapling 3-cm from pylorus. Hemostatic agents had higher reoperation rates. It was not feasible to test for mortality given the low incidence.

Conclusion: Certain surgical techniques in SG among our bariatric surgeons had a significant effect on the rates of adverse outcomes, bleeding, readmission, reoperation, and SSI. Our findings warrant further investigation into these techniques via multivariate regression or prospective design.

Limitations: This study was limited by its retrospective and univariate design. We did not account for interaction. The sample size was small, and follow-up of 30 days was relatively short. We did not include patient characteristics in the model or control for surgeon skill.