Stool and Serum CMV-PCR in Diagnosing CMV Colitis

The gold standard for diagnosing CMV colitis is colonoscopy with tissue biopsy. Histology by hematoxylin and eosin stain (H\&E) has a high specificity of 92-100% but low sensitivity of 10-87%. Immunohistochemistry stain(IHC) increases diagnostic sensitivity to 78-93%.Tissue polymerase chain reaction assay (PCR), the method that amplifies CMV DNA, has the highest sensitivity (92-93%); however, there is no clear cut-off value to differentiate CMV bystanders from CMV colitis. Although colonoscopy is the gold standard, it is not permitted to be performed in some situations due to a high risk of procedural complications, such as in patients with profound neutropenia, thrombocytopenia, or severe illnesses. Serum CMV-PCR is a non-invasive test that detects CMV viremia, which could be associated with CMV-GI disease. However, the results of previous studies are inconsistent.

In recent years, there has been an increasing interest in stool CMV-PCR in diagnosing CMV colitis. Stool CMV-PCR is a non-invasive test that can be either qualitative or quantitative. In the first pilot study, Herfarth et al. reported stool CMV-PCR sensitivity of 83% and specificity of 93% in 19 inflammatory bowel disease patients. Since that, there have been only a few studies reporting the performance of stool CMV-PCR in diagnosing CMV colitis, and their results are inconsistent. The reported sensitivity ranges from 16.7% to 85%, and the specificity ranges from 71 to 96%. Furthermore, the diagnostic performance of combining serum and stool CMV-PCR, which could improve diagnostic performance, has not been studied. Therefore, we aimed to investigate the diagnostic performance of stool CMV-PCR, serum CMV-PCR, and their combination in diagnosing CMV colitis using tissue histopathology as the standard reference in patients with clinical suspicion of CMV colitis. In addition, we also evaluated the correlation of viral load in stool, serum, colonic tissue, and numbers of CMV-infected cells in colonic tissue.

Patients older than 18 years with clinical suspicion of CMV colitis were enrolled. The criteria for clinically suspicious of CMV colitis included: i) presence of at least one of the risk factors, including human immunodeficiency virus infection, solid organ or hematologic stem cell transplantations, hematologic malignancy, inflammatory bowel disease, taking immunosuppressive agents or chemotherapy, and critically ill with multiple comorbidities, and ii) presenting with gastrointestinal tract symptoms, including abdominal pain, gastrointestinal bleeding, diarrhea, and bowel ileus.

All participants underwent colonoscopy with tissue biopsy for both H\&E stain and immunohistochemistry for CMV (IHC-CMV). Patients with the detection of either cytomegalic cells or IHC-CMV cells were diagnosed with CMV colitis. The other two pieces of colonic tissue were sent for quantitative CMV-PCR (CMV R-gene® kit, limit of detection 450 copies/ml, Biomerieux, France). Patients who could not proceed with the colonoscopy and tissue biopsy were excluded. Quantitative stool CMV-PCR (CMV R-gene® kit, limit of detection 450 copies/ml, Biomerieux, France) and quantitative serum CMV-PCR (Cobas® CMV, limit of detection 150 copies/ml, Roche, USA) were collected within seven days of colonoscopy.