S3174 Examining Endoscopic Evaluation of Acute Gastrointestinal Bleeding

Abstract

INTRODUCTION: Gastrointestinal bleeding (GIB) is a common medical problem encountered by physicians. GIB is divided into upper or lower GIB based on location in relation to the ligament of Treitz. The incidence of annual upper GIB is estimated to be 50-150 per 100,000. Lower GIB has an annual incidence of approximately 20-35.7 per 100, 000 with a mortality rate. The average cost associated with an upper GIB in the US is ∼$2500 and ∼ $9,700 for lower GIB. The aim of our study was to analyze if the number endoscopies being performed are justified by examining the incidence of acute findings. METHODS: Single institution retrospective observational study conducted from 2014-2016. All patients that presented with GIB were included. Main outcome: incidence of a positive endoscopic finding for an active GIB. Secondary outcome: reason for GIB, association with antiplatelet (AP) or anticoagulants (AC), and if repeat endoscopy was performed. Statistical methods used included: Chi-square/Fisher’s exact test, Mann-Whitney U test, and univariate/multivariate analyses. A P< 0.05 was used to determine statistical significance. RESULTS: 730 observations were collected. 223/730 subjects had a reason for GIB (table 1). EGD, repeat endoscopy and taking AC were statistically significant. The highest reported reason for GIB was diverticulitis/diverticulosis (table 2). When the reason for GIB were listed for colonoscopy and EGD overall and compared between the two groups, the highest findings were diverticular, normal, or adenoma/polyp for colonoscopy overall and normal, gastritis, and hiatal hernia for EGD (table 3). Logistic regression analysis (table 4) was used and using no as a reference group for each variable we found those who used AC were less likely to have a reason found for bleeding than those who did not have use AC (OR = 0.65, 95% CI = 0.42, 1.01). When using yes as the reference group for each variable we found those who did not use AC were more likely to have a reason found for bleeding than those who used AC (OR = 1.53, 95% CI = 0.99, 2.38). Lastly we examined AP and AC use by repeat endoscopy status and found no significant differences in AP or AC use between those who has a repeat endoscopy and those who did not. CONCLUSION: Due to the high cost of GIB we examined if the number of endoscopies being performed were justified by finding the reason for GIB. This is a complex task, we found over a 3-year period, only 30.55% of subjects were found to have a reason for GIB.Table 1.: Characteristics of patientsTable 2: and Table 3.Table 4: and Table 5.