Esophagorespiratory Fistulas

Goal: The purpose of this study was to characterize outcomes of esophagorespiratory fistulas (ERF) by etiology and initial treatment strategy. Background: ERF is a morbid condition for which optimal treatment strategies and outcomes are still in evolution. Study: Medical records and images were reviewed for all patients diagnosed with ERF at Mayo Clinic in Rochester, MN, between September 1, 2001 and January 1, 2012. Fistulas were classified as malignant or benign. Treatment strategies were classified as surgical or nonsurgical (typically esophageal stent placement). Technical and clinical success, survival, and survival free of second intervention were assessed. Results: A total of 123 patients with acquired ERF were identified, of whom 65 (53%) were malignant and 58 (47%) benign. Initial treatment strategy was nonsurgical in 88 (72%) patients and surgical in 35 (28%); lower Charlson comorbidity scores were associated with increased likelihood of surgery. Technical and clinical success was seen in a majority of patients treated both surgically and nonsurgically. Patients with malignant ERF treated surgically survived longer than patients undergoing nonsurgical treatment (hazard ratio=5.6, P=0.005). In contrast, those with benign ERF had similar overall survival regardless of whether they received initial surgical or nonsurgical treatment; reintervention was more common in those who underwent nonsurgical treatment (hazard ratio=2.3, P=0.03). Conclusions: We conclude that survival in malignant ERF is better with surgical intervention in selected patients. Surgical and nonsurgical techniques achieve similar survival in benign ERF, but reintervention is more common in those treated endoscopically.