Background: The potential negative impact of cytomegalovirus (CMV) in ulcerative colitis (UC) and Crohn's disease (CD) warrants efforts to improve the yield of diagnostic techniques.
Methods: We retrospectively determined the optimal biopsy location and number from sixty-eight patients with inflammatory bowel disease (66% UC, 31% CD, and 3% inflammatory bowel disease-unclassified) with CMV disease between 2005 and 2011. Biopsies with endoscopic and histologic inflammation were analyzed by immunohistochemistry and/or in situ hybridization. The proportion of positive biopsies was determined, and using data from the 25th percentile, we assessed the number of biopsies required to achieve an 80% probability of a single positive biopsy.
Results: Of the patients with a diagnosis by immunohistochemistry and/or in situ hybridization, 27 of 61 (44%; 95% confidence interval, 32-57) were positive by hematoxylin and eosin, and 11 of 36 (31%; 95% confidence interval, 16-46) had systemic CMV by polymerase chain reaction. Of the patients with biopsies proximal and distal to the splenic flexure, 1 of 11 with UC and 4 of 8 with CD had a diagnosis limited to the right colon. Twenty percent of biopsies were positive by immunohistochemistry or in situ hybridization (20% in UC and 17% in CD). Eleven biopsies in UC and 16 in CD were required to achieve an 80% probability of a positive biopsy.
Conclusions: Biopsy location and number are important considerations when assessing for CMV. We recommend a flexible sigmoidoscopy with 11 biopsies in UC and a colonoscopy with 16 biopsies in CD.