Background: It is unclear if traditional histopathology and noninvasive blood-based tests are sufficiently accurate to detect cytomegalovirus (CMV) reactivation in inflammatory bowel disease. Therefore, we assessed the diagnostic accuracy of these tests compared with immunohistochemistry (IHC) and tissue polymerase chain reaction (PCR).
Methods: A systematic search of electronic databases was performed from inception through January 2016 for observational studies comparing diagnostic tests for CMV reactivation in inflammatory bowel disease. IHC and tissue PCR were considered reference standards and were used to evaluate the accuracy of blood-based tests and hematoxylin and eosin histopathology. Weighted summary estimates with 95% confidence intervals (CIs) were calculated using bivariate analysis.
Results: Nine studies examined the accuracy of blood-based tests for predicting colonic CMV reactivation: 5 studies by pp65 antigenemia and 4 studies by blood PCR. The overall sensitivity was 50.8% (95% CI, 19.9-81.6), the specificity was 99.9% (95% CI, 99-100), and the positive predictive value was 83.8% (95% CI, 58.6-95.0). The sensitivities of pp65 and blood PCR were 39.7% (95% CI, 27.4-52.1) and 60.0% (95% CI, 46.5-73.5), respectively. Nine studies examined the sensitivity of histopathology. The overall sensitivity was 12.5% (95% CI, 3.6-21.4), 34.6% by IHC (95% CI, 13.8-55.4), and 4.7% by tissue PCR (95% CI, 1.2-17.1).
Conclusions: Although blood-based tests seem to predict colonic CMV reactivation, they are insensitive tests. Similarly, histopathology has poor sensitivity for detecting colonic CMV. In agreement with current guidelines, these tests should not replace IHC or tissue PCR for detecting CMV reactivation in inflammatory bowel disease.