S602 Underutilization of Digital Rectal Exam by Trainee Physicians in the Evaluation of Patients with Gastrointestinal Bleed in the Inpatient Setting

analysis was conducted by the random-effects model and heterogeneity was assessed by the I2% statistic and the 95% prediction interval. Results: The analysis included nine studies and 176 patients. The pooled rate of treatment responders was 81.2% (95% CI 58.2-93.1) and the pooled rate of GAVE recurrence was 13.3% (95% CI 2.3-49.6). The pooled mean sessions of treatment was 2.45 (95% CI 2.12-2.77). The pooled mean Hb difference pre and post procedure was 1.65 (95% CI 0.87-2.43), p50.001. The pooled mean difference of pre to post treatment units of pRBCs transfused was 1.37 (95% CI 0.54-2.21), p50.001. The pooled mean difference of pre to post treatment hospitalizations was 0.67 (95% CI 0.19-1.14), p50.006. The pooled adverse events with EBL treatment for GAVE was 12% (95% CI 7.2-19.5). Heterogeneity was moderate based on the 95% prediction interval (Table 1). Conclusion: The high rate of response with as low as 2 to 3 sessions is a key finding. Statistically significant changes in Hb, pRBC transfusions and hospitalizations establish the efficacy of EBL in treatment of GAVE. Although the pooled adverse events was 12%, none of these events warranted admission or prolonged observation. This meta-analysis demonstrates high response to treatment and excellent treatment outcomes of EBL in patients with GAVE.