Histological diagnosis of cytomegalovirus hepatitis in liver allografts.

AIMS--To determine the incidence of histologically documented cytomegalovirus (CMV) hepatitis following orthotopic liver transplantation (OLT) and to assess the effectiveness of immunohistochemistry and in situ hybridisation (ISH) in detecting CMV. To describe the histological pattern most frequently associated with CMV hepatitis in order to select the biopsy group in which these modern techniques are most effective. METHODS--A prospective histological study was carried out on 853 biopsy specimens, obtained from 191 liver allografts (160 patients). Specimens were stained with haematoxylin and eosin and immunohistochemically (avidin-biotin complex) using monoclonal antibodies directed against early and late CMV antigens. A retrospective selection was made of 23 specimens with viral inclusion bodies in cytomegalic cells (group A) to characterise the most frequently associated histological pattern, and of 34 other specimens without viral inclusion bodies (group B) but with the same microscopic features as group A. Re-cuts from both specimen groups were studied using immunohistochemistry and ISH with a CMV specific complementary DNA probe. RESULTS--CMV infection was confirmed in 35 specimens (29 by immunohistochemistry, 23 by presence of inclusion bodies in haematoxylin and eosin stained sections, 16 by ISH) from 27 patients (incidence 16.9%). CMV hepatitis was diagnosed within 46 +/- 19 (range 21-114) days posttransplant. Twenty on (91.3%) of the 23 biopsy specimens with inclusion bodies (group A) displayed heterogeneous inflammatory foci disseminated throughout the hepatic lobule. Nineteen specimens (82.6%) were positive by immunohistochemistry and 14 (60.9%) by ISH. In eight (23.5%) of the 34 group B specimens CMV infection was confirmed by immunohistochemistry (n = 6) or ISH (n = 2). Another 12 (35.3%) of the group B specimens negative on staining with haematoxylin and eosin, immunohistochemistry and ISH came from allografts in which previous or subsequent biopsy specimens were CMV positive. CONCLUSIONS--Demonstration of cytomegalic inclusion bodies in haematoxylin and eosin sections is sufficient for a diagnosis of CMV hepatitis. The routine use of immunohistochemistry in all allograft biopsy specimens in more sensitive than demonstration of inclusion bodies by staining with haematoxylin and eosin but may yield false negative results because of the focal distribution of positive cells. ISH was less sensitive than staining with haematoxylin and eosin and/or immunohistochemistry. A histological picture of "disseminated focal hepatitis" without viral inclusion bodies selects a group of allograft biopsy specimens in which immunohistochemistry and/or ISH may improve detection of CMV.