Capsule Endoscopy Is Not as Accurate as Esophagogastroduodenoscopy in Screening Cirrhotic Patients for Varices

BACKGROUND & AIMS: Esophagogastroduodenoscopy (EGD) is the standard technique for screening cirrhotic patients for high-risk varices and other significant upper gastrointestinal lesions (HRVLs). We investigated whether esophageal capsule endoscopy (ECE) is as convenient and accurate as EGD for the detection of HRVLs. METHODS: We analyzed data from 65 cirrhotic patients without prior upper gastrointestinal bleeding who were examined for varices and HRVLs by ECE and EGD (both procedures were performed on the same day). EGD was performed by 2 physicians (75% of patients were unsedated) who used standard grading for esophageal and gastric varices, portal hypertensive gastropathy, and HRVLs. Coded capsule tracings were read by 2 investigators, blinded to the EGD findings, using standard grading. RESULTS: The median procedure time for EGD (with or without biopsy collection) was 3 minutes, compared with 20 minutes for ECE. The overall accuracy for diagnosis of esophageal varices was 63.2% +/- 5.9%; for detection of esophageal varices red marks was 68.8% +/- 5.4%; and for diagnosis of other HRVLs was 51.5% +/- 4.2%. The interobserver agreement in the diagnosis of esophageal varices was 90.8%; in the detection of esophageal varices red marks was 86.2%; and in the diagnosis of other HRVLs was 7.3%. CONCLUSIONS: ECE is not as accurate as EGD in the diagnosis of esophageal varices and red markings or in grading esophageal varices. Moreover, ECE had poor accuracy in grading portal hypertensive gastropathy and detecting ulcers, gastric varices, and other significant upper gastrointestinal lesions. It took significantly longer to perform ECE and interpret the results than for EGD. These findings do not support ECE as a preferred tool for screening esophageal varices and HRVLs.