Challenges and limitations in the diagnosis of atrioesophageal fistula

Introduction Atrioesophageal fistula (AEF) is a dire complication of atrial fibrillation ablation. The diagnostic yield of computed tomography (CT) chest, the role and timing of repeat testing, and the value of other investigations in the diagnosis of AEF is uncertain. Methods We systematically reviewed published AEF cases to evaluate radiological, bedside, and biochemical investigations for AEF (registered on PROSPERO [CRD42017077493]). Results Eighty-seven articles with 126 patients (median age, 59 years; male, 71%) were included in the analysis. CT chest was performed in 88% (111/126) and was abnormal in 87%. A clear diagnosis of AEF (fistula/perforation) was only detected in 35% (34/97). Other major findings included free air in mediastinum (26%), left atrium (LA), or LA wall (24%). In 11 patients with normal/nonspecific initial CT chest, major abnormalities were detected in 91% (10/11) of repeat CT chest performed 6 days (median; range, 4-22) after initial scan. Initial CT head was normal in 51%; diffuse air emboli was identified in 79% (22/28). Initial transthoracic echocardiography was normal in 61% of cases. The spectrum of radiological abnormalities included Air (mediastinum/LA), Effusion (pleural/pericardial), Fistula/Perforation, and Thickening (esophagus/LA) - "AEF-Tests." Esophagram demonstrated contrast extravasation in 87% (13/15). Blood culture was consistently positive (100%; 28/28), particularly for streptococcus species (93%; 26/28). Conclusion The diagnosis of AEF remains challenging. Clinicians should be aware of the limitations in the yield of CT chest, the variety of major abnormalities reported, the need for repeat testing, unique brain imaging findings, and the importance of positive blood cultures and raised inflammatory markers.