Sensitivity and specificity of diagnostic modalities to diagnose CMV in coexisting IBD in patients hospitalized with acute colitis: a systematic review and meta-analysis

BACKGROUND: Cytomegalovirus (CMV), a common infection, can reactivate and cause CMV-associated colitis in immunosuppressed people, like those with known inflammatory bowel disease (IBD). Diagnosis of CMV colitis is limited by the broad spectrum of CMV infection and the invasiveness of biopsy-based investigations. We aimed to generate diagnostic accuracy estimates for serum, stool, and tissue polymerase chain reaction (PCR), pp65 antigenemia, and hematoxylin and eosin (H&E) staining in diagnosing CMV colitis in IBD patients hospitalized with acute colitis. METHODS: We systematically searched the MEDLINE and EMBASE databases via OVID from inception until March 3, 2025, for eligible studies for a meta-analysis. Included studies needed to compare 2 tests and report true positives, false negatives, false positives, and true negatives to enable sensitivity and specificity calculations. RESULTS: In total, 23 studies were included. Tissue PCR had the highest log diagnostic odds ratio at 2.93, with a pooled sensitivity of 95% and pooled specificity of 66%. The remaining pooled sensitivity and specificity, respectively, were stool PCR 70% and 92%, serum PCR 61% and 91%, pp65 antigenemia 33% and 98%, and H&E 44% and 99%. CONCLUSIONS: Tissue PCR had the best performance and was the most sensitive marker of CMV colitis in IBD, albeit with poor specificity. Stool PCR is a promising test, but further diagnostic studies are needed. Serum PCR, pp65 antigenemia, and H&E, when positive, seem to be useful in confirming CMV colitis. There is a need for further studies to develop an optimal PCR cutoff to differentiate low-level CMV reactivation from true CMV colitis on PCR testing, which may enhance the diagnostic yield from an endoscopic biopsy that tests both tissue PCR and IHC given the higher sensitivity of tissue PCR.