Cytomegalovirus (CMV) in gastrointestinal mucosal biopsies: should a pathologist perform CMV immunohistochemistry if the clinician requests it?

Cytomegalovirus (CMV) causes clinically significant gastrointestinal (GI) injury. CMV inclusions can be identified on routine hematoxylin and eosin (H&E) stain, but immunohistochemistry (IHC) is also available for identifying CMV in tissue. The advent of accountable care organization models of care bring into question whether it is cost-effective for immunohistochemistry to be performed upfront at the request of clinicians and whether the quality of viral detection is compromised when the diagnosis of CMV is predicated on histologic review. In this study, a retrospective review of GI biopsies with CMV evaluations was performed. There were 449 cases with clinical requests to rule out CMV and 238 CMV analyses initiated by the pathologist without a clinical request. Among the cases that included a clinician's request, 37 had CMV detected. Immunostaining was performed on 26 cases, while a diagnosis based on readily identifiable viral inclusions on H&E-stained slides was made in 11. Among pathologist-initiated work-ups, 15 were CMV+, 3 of which had inclusions identified by H&E only. Among 38 CMV cases for which IHC had been performed, 27 had overt viral inclusions obvious on H&E. Seventy-two cases revealed uninflamed GI mucosa, and although a clinical concern about CMV infection was present, a CMV IHC work-up was not initially performed; all were negative for CMV by IHC and H&E. Clinical suspicion for CMV has a high yield for CMV detection, but "upfront" testing is likely unnecessary. Careful histopathologic review by a pathologist remains critical in the efficient and cost-effective detection of CMV.

Special histologic stains are rarely beneficial for the evaluation of HIV-related gastrointestinal infections.

During a 28-month period, endoscopic mucosal biopsy specimens from all HIV-infected patients were submitted for routine histologic evaluation. Immunoperoxidase staining for cytomegalovirus and herpesvirus antigens (esophagus), mycobacterial and fungal staining, and Gram staining of mucosal biopsy specimens were done. Special fungal and acid-fast stains were selectively performed in patients with absolute CD4 cell counts of less than 200 cells per microliter (200 x 10(6)/L) and/or with diarrhea and or wasting syndrome. Treatment was based on the endoscopic and histologic findings, and long-term follow-up was performed. The 121 symptomatic HIV-infected patients underwent 221 upper and/or lower endoscopies with 285 biopsy sites. The sensitivity and specificity of H&E staining for the diagnosis of gastrointestinal cytomegalovirus were 97% and 100%, respectively. The results of fungal and mycobacterial stains neither altered therapy nor identified previously undiagnosed infections in any patient. Long-term follow-up revealed no patient in whom an infection was missed on routine H&E, which affected outcome. Routine H&E staining is accurate for the diagnosis of gastrointestinal opportunistic infections in HIV-infected patients. Special histologic stains for fungal, mycobacterial, and viral infections did not increase the diagnostic yield or alter medical therapy but doubled the costs.