109
VIDEOCAPSULE VERSUS PUSH ENTEROSCOPY IN PATIENTS WITH OBSCURE GI BLEEDING: PLANNED INTERIM ANALYSIS OF AN OUTCOMES RCT
A Barkun1, D Segarajasingam1, E Seidman1, F Costea1, K Waschke1, P Burtin2, V Baffis1, J Parent1, S Mayrand1, C Fallone1, G Jobin3, M Martel1, J Strazzulla1, V Villard-Bellier1, K Durnin1, Y Li1
1Division of Gastroenterology, McGill University Health Center, Montreal, Quebec; 2Service de Gastro-enterologie, CHU, Angers, France; 3Service de Gastro-enterologie, Hopital Maisonneuve-Rosemont, Université de Montréal, Montreal, Quebec
Aims: Videocapsule endoscopy (VCE) is associated with increased detection of mucosal lesions in patients with obscure gastrointestinal bleeding compared to Push Enteroscopy (PE). Its clinical relevance is debated with few high quality outcomes trials.
Methods: We randomly allocated adult patients with occult bleeding (iron deficiency anemia >6 months needing transfusion/ iron with repeated occult blood positive testing), or with overt bleeding (>1 episode of melena/ hematochezia within 6 months with hemoglobin drop) to VCE or PE after a negative initial work-up (gastroscopy and colonoscopy with/without small bowel radiology). The main outcome was recurrence/ persistence of bleeding using standardized a priori criteria over 1 year. Cross-overs were permitted after the main outcome was reached. We present descriptive characteristics, safety considerations, and an inferential analysis for the main outcome, as per the planned interim analysis protocol.
Results: Over 28 months, 20 patients (68±15yrs, 36% female) were allocated to VCE and 27 to PE. Overall, 74% presented with overt bleeding. At baseline, the mean number of co-morbidities were 2.8±1.7, 37% were taking ASA, 17% an anticoagulant, and 5% an NSAID; 30% of patients had had a nuclear scan, and 13% an angiography. No marked between-group imbalances were noted. One or multiple lesions were detected in 77% of patients (thought to be related to bleeding in 26%). The mean duration of follow-up to date has been 304±116 days, with 45% requiring a blood transfusion (mean 2.4±1.2 units), 74% a hospitalization (mean stay: 8.9±10.8 days; 6% had surgery). 34% of patients have so far crossed over (75% from PE to VCE). Overall, the proportion of rebleeding or persistent bleeding in the 36 patients with complete follow-up has been 21% (95% CI: 6.8-48%) in the VCE and 41% (95% CI: 23-61%) in the PE group (p=0.40).
Conclusions: This planned interim analysis suggests no safety concerns; there exists a trend for less rebleeding or persistent bleeding amongst patients randomized to VCE.