Defining the Threshold: New Data on the Ability of Capsule Endoscopy to Discriminate the Size of Esophageal Varices

Abstract

Endoscopy (esophagogastroduodenoscopy, EGD) to screen for esophageal varices (EV) is recommended in patients with portal hypertension. Reports indicate that capsule endoscopy (CE) is capable of identifying large/medium varices (L/MV) when the varix comprises more than 25% of the circumference of the field of view. We evaluated the ability of CE to discriminate the size of EV using this grading scale. Patients underwent CE and EGD on the same day. A blinded investigator interpreted capsule findings. CE labeled EV as L/MV if ≥25% of the lumen circumference was occupied, and small/none for <25%. A total of 37 patients were enrolled in this prospective, observational study at a single tertiary-care academic center. Three CE were excluded due to rapid esophageal transit time or technical malfunction. Using a 25% threshold, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for EC to discriminate L/MV were 23.5%, 88.2%, 66.7%, and 53.6%, respectively (κ=0.12). Reducing the threshold to 12.5% resulted in sensitivity, specificity, PPV, and NPV of 88.2%, 64.7%, 71.4%, and 84.6%, respectively (κ=0.53). A receiver-operator characteristic (ROC) curve showed a 15% threshold to be optimal in discriminating EV size using CE, resulting in sensitivity, specificity, PPV, and NPV of 76.5%, 82.4%, 81.3%, and 77.8%, respectively (κ=0.59). This study indicates that discriminating EV size by the current capsule scale is unreliable. Lowering the grading threshold improved the ability to discriminate EV size by CE. In the proper context, CE is an alternative to EGD to screen for EV.