Repair, Reconstruct, or Divert

Objectives: The aim of this study was to determine differences in esophageal perforation populations undergoing different advanced interventions for perforated esophagus and identify predictors of treatment outcomes. Summary Background Data: Contained esophageal perforation can often be managed expectantly, but uncontained perforation is uniformly fatal without invasive intervention. Treatment options for the latter range from simple endoscopic control through advanced intervention. Clinical presentation varies greatly and directs which intervention is most appropriate. Methods: From 1996 to 2017, 335 patients were treated for esophageal perforation, and 166 for advanced interventions: 74 primary repair with tissue flap (repair), 26 esophagectomy and gastric pull-up (resection), and 66 esophagectomy and immediate diversion with planned delayed reconstruction (resection-diversion). Patient characteristics, clinical presentation, operative outcomes, and survival were abstracted. Pittsburgh Severity Scores (PSS) were retrospectively calculated. Random survival forest analysis was performed for 90-day mortality and competing risks for reconstruction after resection-diversion. Results: Repair and resection patients had lower PSS than resection-diversion patients (3 vs 3 vs 6, respectively). Ninety-day mortality for repair, resection, and resection-diversion was 11% vs 7.7% vs 23%, and 5-year survival was 71% vs 63% vs 47%. Risk of death after resection-diversion was highest within 1 year, but 52% of patients had reconstruction of the upper alimentary tract within 2 years. Conclusions: Several advanced interventions exist for critically ill patients with uncontained esophageal perforation. Repair and organ preservation are always preferred; however, patients at extremes of illness might best be treated with resection-diversion, with the understanding that the competing risk of death may preclude eventual reconstruction.