Emergency Computed Tomography Predicts Caustic Esophageal Stricture Formation

Abstract

Background: Endoscopy is the best predictor of stricture formation after caustic ingestion.

Objective: Our aim was to compare the accuracy of emergency computed tomography (CT) and endoscopy in predicting risks of esophageal stricture.

Methods: We describe a CT classification of esophageal caustic injuries: Grade I show normal esophagus; Grade IIa display internal enhancement of the esophageal mucosa and enhancement of the outer wall conferring a "target" aspect; Grade IIb present as a fine rim of external esophageal wall enhancement. In 152 patients (56 males, median age 45) who underwent esophageal preservation after caustic ingestion we compared the accuracy of the CT and endoscopic (Zargar) classifications in predicting esophageal stricture.

Results: On endoscopy esophageal injuries were classified as grade 1 (n = 50; 33%), grade 2a (n = 11; 7%), grade 2b (n = 19; 13%), grade 3a (n = 14; 9%), and grade 3b (n = 58; 38%). On CT, 47 (31%) patients had grade I, 47 (31%) had grade IIa and 58 (38%) had grade IIb esophageal injuries. Fifty-six (37%) patients developed esophageal strictures. The risk of esophageal stricture formation was 0%, 17%, and 83%, for grade I, IIa, and IIb CT injuries and 0, 0, 28, 50, and 76% for endoscopic grade 1, 2a, 2b, and 3a and 3b injuries, respectively. ROC curve analysis at 120 days after ingestion showed that CT outperformed endoscopy in predicting stricture formation (AUC: 85.1 [95% CI, 74.9-95.3] vs 77.8 [95% CI, 66.5-89.0], P = 0.047) and did just as well as a combined CT-endoscopy algorithm (AUC: 85.8 [95% CI, 76.5-95.0] vs 85.1 [95% CI, 74.9-95.3], P = 0.73).

Conclusion: Emergency CT outperforms endoscopy in predicting esophageal stricture formation after caustic ingestion. Emergency endoscopy evaluation after caustic ingestion is not indispensable.