Prospective analysis of work-up of acute lower gastro-intestinal bleeding: Can an optimal algorithm be designed?

The most accurate and cost effective method of investigating the etiology of acute lower gastrointestinal bleeding (LGIB) remains elusive. To develop a clinically useful algorithm, we prospectively collected data on all patients with acute LGIB admitted to our institution. Demographics and clinical parameters were recorded without influencing clinical decision making. Over a two-month period, 38 pts presented with acute LGIB; three pts had prior bowel resections and were excluded. Of the 35 pts in this study (17F/18M, age 68±2.6 yr), 51% had hemodynamic instability at presentation. Nasogastric aspiration performed in 21 pts (60%) excluded upper GI bleeding. Sixty percent were taking aspirin, non-steroidal anti-inflammatory drugs (NSAIDs), or anticoagulants. Using a pocket-sized stool color confirmation card test*, 77% of pts had the two brightest red card colors (color #1 & #2) at presentation. Initial investigation consisted of tagged red blood cell (TRBC) scan in 14 pts (40%) and colonoscopy in 19 (54%). Of the 9 pts who had card colors #1 or #2 at the time of TRBC scan, 67% had a positive scan; in contrast, none of the pts with maroon or brown stool had a positive scan (p = 0.02). All 8 pts with card colors #1 or #2 after bowel purge had a source identified at colonoscopy, in contrast to 10 of 22 pts with maroon or clear stool (p=0.009). The source of bleeding was identified in 18 pts, and included: mass (5 pts), diverticula (5), colitis (3), fissure (2), others (2), aorto-duodenal fistula (1). Localization without etiologic diagnosis was achieved in 3 pts. Endoscopic therapy was successful in 2 pts. Stool color → card colors # 1, 2*other card colors (#3.5)*p at presentation n = 27 n=8 TRBC scan positive † 67% 0% 0.02 final diagnosis ‡ 70% 12% 0.01 left colon sites ‡ 53% 0% 0.01 right colon sites ‡ 33% 12% ns*color confirmation card, Dig Dis Sci 40:1614, 1995 † tin patients with the respective stool colors during scan ‡ includes all methods of investigation Conclusions: The utilization of a simple objective color test appears to optimize the predictability of diagnostic testing in acute LGIB. The passage of blood corresponding to card colors #1 and #2 during TRBC scanning and colonoscopy significantly increases diagnostic yield. These two tests appear complementary in acute LGIB.