A prospective study of endoscopy in HIV-associated diarrhea

OBJECTIVE: Diarrhea commonly occurs in persons with human immunodeficiency virus (HIV) infection. The optimal use of endoscopic procedures remains poorly studied for patients with HIV-related diarrhea. The purpose of this study is to compare the diagnostic yield of a complete endoscopic work-up including an esophagogastroduodenoscopy and colonoscopy to a more limited approach of biopsies obtainable by flexible sigmoidoscopy. METHODS: A prospective study of 79 patients with HIV-related diarrhea. Upper endoscopy and colonoscopy were performed with tissue biopsies labelled according to location within the colon or small intestine. RESULTS: A new infection was diagnosed in 22 of 79 patients (28%). Biopsy of the left colon yielded an enteric pathogen in 17 of 22 patients (sensitivity: 77%) and in 15 of 15 patients with cytomegalovirus colitis (sensitivity: 100%). Combined left and right colonic biopsies had a sensitivity of 82%. Combined colonic and terminal ileum biopsies missed no pathogens. Duodenal biopsies yielded no additional pathogens beyond those identified by colonoscopy and terminal ileal biopsy. Patients with a new pathogen diagnosed had significantly lower CD4 lymphocyte counts as compared to patients without a new pathogen (p = 0.001). CONCLUSIONS: For patients with CD4 counts < 100/mm3 and unexplained AIDS-related diarrhea, flexible sigmoidoscopy with biopsy is a sufficiently thorough endoscopic evaluation.