Diagnostic Value of Capsule Endoscopy (CE) in Obscure Digestive Bleeding (ODB) and Effect of Erythromycin Injection

Abstract

Diagnostic Value of Capsule Endoscopy (CE) in Obscure Digestive Bleeding (ODB) and Effect of Erythromycin Injection Dimitri Coumaros, Laurent Claudel, Patrick Levy, Michel Doffoel Aims : Evaluate the diagnostic yield of CE in ODB and the effect of erythomycin injection. Patients andmethods: In this prospective randomized study, 37 patients (pts) (mean age : 61 years) with an occult (n = 19) or overt (n = 18) ODB, were investigated by the Given M2A capsule, with (n = 19) or without (n = 18) intravenous injection of 3 mg/kg erythomycin, 30 to 60 min before CE. Colonic cleansing with 4 L PEG (n= 34) or 2 bottles of sodium phosphate (n = 3) ending 1 h before CE was undertaken. Lesions were classified according to their bleeding potential : nil (L0), probable (L1) or certain (L2) if an active bleedingwas visible. A single diagnosis per patient was delivered. Diagnostic agreement coefficient between a senior and a junior endoscopist was calculated based on 21 examinations. Gastric and small bowel transit times as well as the possibility of visualizing the colon were compared between the 2 groups. The duration of video analysis was measured. Results : 36 examinations out of 40 were valid (3 technical failures, 1 prolonged retention in the lower esophagus). The examination of 3 pts was repeated. 72 lesions L1 and 2 L2 were detected in 30 pts (83%). These pts had one type of lesion (n = 11), 2 (n = 15) or 3 types of different lesion (n = 4). The resulting diagnosis was angiectasias (n = 14), ulcerations (n = 7), red spots (n = 5), varix (n= 1), tumor (n= 1), nematodes (n= 1), polypoid angiectasia (n= 1). The location of these lesions was jejuno-ileal (n = 23), duodenal (n = 3), gastric (n = 1), colonic (n = 1) or multifocal (n = 2). The interobserver kappa coefficient was 0.64. The presence of red spots was linked to the presence of angiectasias (p<0.001). Erythromycin decreased the gastric time by 55 min (p<0.01) and increased the small bowel time by 70 min (p<0.03) without change of the rate of complete examination of the small bowel (84 vs 80%). Colonic visualization was allways limited to the caecum and ascending colon. Colonic cleansing was insufficient in 89% of pts. The duration of the video analysis decreased significantly after the first 24 examinations (63 vs 80 min; p<0.02). 2 cases of oral time superior to 5 min and 1 temporary blocage of the CE in the hypopharynx were noted. Conclusion : CE detected a bleeding or a potential bleeding lesion in 83% of pts. The resulting diagnosis concerned the stomach, the duodenum or the colon in 14% of the pts. Erythromycin reduced the gastric time without increasing the rate of complete examination of the small bowel. The exploration of the colon was limited to the caecum and ascending colon.