Interobserver Reproducibility and Feasibility of Polymerase Chain Reaction (PCR)-Based Assay in Distinguishing Ischemic Colitis from Clostridium Difficile Colitis in Endoscopic Mucosal Biopsies
Homer O Wiland, Gary W Procop, Marion Tuohy, John R Goldblum, Deepa T Patil. Cleveland Clinic, Cleveland
Background: Ischemic colitis and C. difficile colitis may be difficult to distinguish on histology alone. More recently, PCR-based assays on stool samples have shown higher sensitivities than the widely used C. difficile toxin enzyme immunoassay (EIA), which has a sensitivity of only 48%. This study examines the utility of C difficile PCR-assay in endoscopic mucosal biopsies. In addition, we also evaluate the interobserver reproducibility of distinguishing ischemic colitis from C. difficile colitis based on previously proposed histologic features. (Dignan CR AJSP 1997).
Design: Biopsies with an ischemic-pattern of injury (N=40) were culled from our database between Jan 2005-Dec 2010 and evaluated for: pseudomembranes, lamina propria hyalinization, withered crypts, thrombi, mucosal hemorrhage, and necrosis. Endoscopic findings, C. difficile EIA result, treatment, and vascular work-up were documented. Biopsies were blindly reviewed by two GI pathologists and categorized as either C. difficile or ischemic colitis and interobserver reproducibility was assessed using kappa statistics. Following microdissection and DNA extraction, PCR assay was performed using the BD Geneohm (Franklin Lakes, NJ) C. difficile assay. Considering clinical work-up, including C. difficile EIA results as the gold standard, the histologic diagnosis and PCR assay results were compared and statistically analyzed.
Results: The cohort comprised of 21 females and 19 males; there was no significant difference in age between patients with ischemic colitis (N= 20) and C difficile colitis (N=18; mean age 56 yrs vs 63 yrs). In 2/40 cases, the etiology was unclear. There was good agreement between histologic and clinical diagnosis (kappa=0.7). The interobserver agreement between the two GI pathologists was excellent (kappa = 0.84). A total of 34 cases (19 ischemia, 15 C. difficile) were analyzed by PCR. None of the ischemic colitis cases were positive by PCR (100% specificity), but only 3/15 C. difficile-confirmed cases were PCR positive (20% sensitivity). One of these 3 cases, however, was C. difficile EIA negative.
Conclusions: Interobserver reproducibility for histologically distinguishing ischemic colitis from C difficile colitis was excellent in this study. The C. difficile PCR test, as applied to endoscopic mucosal biopsies, has excellent specificity but limited sensitivity. Nevertheless, it may be helpful in identifying C. difficile infection in patients with histologic evidence of C. difficile enterocolitis, but who had a negative EIA test.
Wednesday, March 21, 2012 9:30 AM
Poster Session V # 80, Wednesday Morning