Stephanie Rodriguez, Spencer R Goble, Gaurav Suryawanshi, Khalid Ahmed, Nabeel Azeem
Most foreign body-associated esophageal perforations are managed conservatively. Among patients selected for intervention, endoscopic management was associated with lower rates of mortality, selected complications, and healthcare utilization than surgery. These findings reflect clinical triage pathways and should be interpreted cautiously, given potential selection bias, confounding by indication, and administrative data limitations.
BACKGROUND/AIMS: Esophageal perforations can be managed conservatively, endoscopically, or surgically. We aimed to characterize the outcomes of these approaches and assess the impact of intervention timing.
METHODS: We retrospectively analyzed hospitalizations in the National Inpatient Sample from 2016 to 2022 for foreign body-associated esophageal perforations. Patients were grouped into endoscopic, surgical, or conservative management. Outcomes included mortality, respiratory failure, sepsis, hemorrhage, parenteral nutrition, and gastrostomy.
RESULTS: Among the 2,045 patients, 145 (7.1%) underwent endoscopic management, 190 (9.3%) underwent surgical management, and 1,710 (83.6%) received conservative treatment. Food impaction accounted for 76% of foreign bodies. The mortality rates were 0%, 5.3%, and 2.9%, respectively. Respiratory failure occurred in 17.2% of endoscopically managed patients and 42.1% of surgically managed patients. Surgical cases had higher use of parenteral nutrition (15.8%) and gastrostomy (31.6%), whereas conservative management had the highest hemorrhage rate (18.1%). Intervention timing was not significantly associated with outcomes.
CONCLUSIONS: Most foreign body-associated esophageal perforations are managed conservatively. Among patients selected for intervention, endoscopic management was associated with lower rates of mortality, selected complications, and healthcare utilization than surgery. These findings reflect clinical triage pathways and should be interpreted cautiously, given potential selection bias, confounding by indication, and administrative data limitations.