Liver Pathology

EBV Mononucleosis

Dale Snover
Fairview Southdale Hospital
Edina, MN


Clinical History
64 year old male who presented with an acute febrile illness with abdominal and back pain and headache. Antibiotics were given but he continued to deteriorate. CT and ultrasound showed a thickened gallbladder and gallstones and the possibility of cholecystitis and choledocholithiasis was considered. At that point he had developed abnormal liver function tests. CT of the head was unremarkable. He was admitted on October 27 and underwent cholecystectomy and wedge liver biopsy on November 6. During his course he developed a peripheral blood lymphocytosis and right middle lobe atelectasis. The lymphocytes were initially described as reactive appearing, however over time increased basket cells with atypical lymphocytes were reported. ERCP was reported to be normal and hepatitis A, B and C studies were negative.

Pertinent Laboratory Data:

October 27: ALT 270, AST 225, Alk Phos 217, T. Bili 1.6, Hgb 13.3

October 31: ALT 326, AST 310, Alk Phos 495, T. Bili 4.6, Hgb 12.9, WBC 16.9 X 10^9/L (12% Neutrophils, 88% lymphocytes)

November 4: ALT 209, AST 246, Alk Phos 600, T. Bili 3.0, Hgb 10.9, WBC 26.7 X 10^9/L (13% Neutrophils, 78% lymphocytes, 9% others)

[Normals: ALT 0-70 U/L, AST 0-55 U/L, Alk Phos 40-150 U/L, T. Bili 0.2 – 1.3 mg/dl, Hgb 13.3 – 17.7 X 10^9/L , WBC 4.0 – 11.0 X 10^9/L ]



Slide 1
Digitized whole slide
Click to view with ImageScope
Click to view with a Web-Based Viewer


Figure 1
Tissue stained for CD3
Figure 2
Tissue stained for CD20
Figure 3
Tissue stained for EBER

Pathological/Microscopic Findings and any Immunohistochemical or Other Studies:
This wedge biopsy demonstrates a diffuse marked portal infiltrate that is expanding every portal tract. Despite the expansion, there is no evidence of fibrosis on trichrome stain. The infiltrate consists mainly of lymphocytes with some interspersed eosinophils and plasma cells. The lymphocytes are a mixture of small mature appearing lymphocytes and some larger activated appearing cells. Despite the heavy infiltrate, piecemeal necrosis is minimal. The hepatocellular lobule demonstrates a diffuse sinusoidal infiltrate of small lymphocytes without much in the way of hepatocellular necrosis. Hepatocellular mitoses are present in relative abundance. Granulomas are not identified and the bile ducts appear intact. The peripheral blood smear demonstrated a relative and absolute lymphocytosis with most of the lymphocytes having a reactive appearance although over the course of time the cells appeared somewhat more atypical with large numbers of basket cells. Therefore there was some concern about malignancy, although the overall impression favored a reactive process. Immunohistochemical staining demonstrated that most of the portal infiltrate and essentially all of the sinusoidal cells were CD3 positive T-cells. There were a few scattered CD20 positive B-cells in the portal areas. Serological studies for CMV were negative however EBV VCA IgM and IgG were both present and EBV Nuclear Antigen (EBNA) antibody was negative. In situ EBER staining was positive in scattered cells in the portal infiltrate corresponding to the B-cells. Immunophenotyping was performed on the blood to rule out a malignant lymphoid process and the majority of the circulating lymphocytes were T-cells (86%) with normal cell surface marker expression (CD4:CD8 ratio 0.15:1). There was no evidence of a clonal B-cell population. Other negative studies included antinuclear and antismooth muscle antibodies and a normal ERCP.

Differential Diagnoses:
The differential diagnosis of the liver biopsy in this case included a viral infection, autoimmune hepatitis, medication reactions and a lymphoproliferative process (e.g. lymphoma or lymphoid leukemia). The presentation with an elevated alkaline phosphatase and bilirubin raised clinical concern about biliary tract disease but the biopsy was not compatible with biliary tract disease. Regarding autoimmune hepatitis, the diffuse portal infiltrate with intermixed eosinophils and plasma cells would be very compatible with that diagnosis, however the nature of the lobular inflammation (i.e. the exclusively sinusoidal infiltrate) and the lack of frank necrosis would be atypical as would the limited amount of piecemeal necrosis in the face of such a heavy portal infiltrate. The absence of fibrosis also argues somewhat against autoimmune hepatitis since autoimmune hepatitis will often have considerable fibrosis, even in cases with an apparently acute clinical onset. In the end, although autoimmune hepatitis was not the favored interpretation of this biopsy, it was thought that autoimmune hepatitis could not be ruled out entirely. In regard to the possibility of lymphoma or leukemia, the diffuse portal infiltrate and the sinusoidal infiltrate could be due to malignancy, although the combination is a bit unusual, especially given the immunohistochemical findings. Low grade B-cell lymphomas do demonstrate a diffuse portal infiltrate, but the sinusoidal infiltrate would be unusual and the mix of cells in the portal tracts did not resemble that of a low grade B-cell lymphoma. High grade B-cell lymphomas tend to have focal replacement of the liver without the diffuse infiltrate. On the other hand, hepatosplenic T-cell lymphoma and many acute leukemias typically shows a sinusoidal infiltrate but not such a prominent portal infiltrate. The immunohistochemical results, with the preponderance of T-cells, rule out a B-cell malignancy. Therefore, based on morphology and limited immunohistochemical staining alone a T-cell lymphoma was still possible but unlikely. The fact that the peripheral blood failed to demonstrate an abnormal T-cell phenotype also argued strongly against a malignant diagnosis. In addition, the positive staining for EBER and the fact that the cells marked with CD5 also argues strongly against a T-cell lymphoma. In the end, although lymphoma was considered on the initial interpretation of the biopsy, it was not the favored diagnosis and was essentially ruled out by the ancillary studies. Medication reactions are in the differential diagnosis. In this case, the patient was not on any likely hepatotoxic medications on admission. The only medication likely to produce a sinusoidal pattern of liver inflammation would Dilantin, which the patient was not taking. Therefore, medication reaction was considered unlikely. In regard to viral infection, the histology would be unusual for hepatitis B or C but could represent hepatitis A. Hepatitis C will often show scattered small lymphocytes in the sinusoids, even in chronic disease, however it is never as extensive as in this biopsy and the portal tract infiltrate would be inconsistent with hepatitis C. Acute hepatitis A can produce a somewhat similar pattern although typically plasma cells are more prominent. In this patient serology was negative for these viruses. The histology is compatible with acute EBV or CMV hepatitis as occurs in the immunocompetent host, although the degree of portal infiltrate and the presence of a large number of large lymphocytes is somewhat unusual for either of these infections. Large cells in the portal tracts are seen in EBV infections in immunocompromised hosts as part of an EBV related immunoproliferative disorder, however these are B-cells, not T-cells. In addition, there was no indication that this patient was immunocompromised. In the end, viral infection was considered most the most likely diagnosis on the biopsy alone, and was supported by the ancillary studies.

Final Diagnosis:
The combination of histological, serological and in situ hybridization results in this case indicate that this is acute EBV mononucleosis. The patient was discharged from the hospital with some improvement in transaminase values and with an improving peripheral blood picture (at the time of this writing there is less than a month of follow-up; more follow-up should be available by the time of the meeting).

Case Discussion:
This case points out several issues that can occur with non-hepatotropic viral infections of the liver. Typical cases of EBV or CMV mononucleosis hepatitis occur in younger patients (most commonly adolescents) and are associated with mild elevations of transaminases without significant bilirubin or alkaline phosphatase elevations. The diagnosis is commonly suspected in an adolescent presenting with peripheral lymphocytosis and mild transaminase elevations and liver biopsy is rarely performed. However, in older individuals these infections commonly are more cholestatic, as in this case, and hence viral hepatitis may not be the primary consideration based on the routine "liver function" tests. This issue can be complicated further by the fact that both of these infections can cause granulomatous hepatitis in which case either medication reactions or primary biliary cirrhosis may be considerations on biopsy. In either case, however, the presence of a prominent sinusoidal lymphocytic infiltrate with minimal hepatocellular necrosis and abundant hepatocellular mitoses should suggest the correct diagnosis. The fact that the patient had evidence of an active EBV infection does not rule out lymphoma entirely and there are reported cases of T-cell hepatosplenic lymphoma associated with EBV infection. Therefore, consideration of this rare malignancy should not be immediately discharged when the EBV studies are positive, and the diagnosis may be reconsidered in a patient if the patient does not recover from the infection in a reasonable length of time. The presence of CD5 expression in addition to CD3 on the T-cells as well as the presence of EBER staining in the liver will essentially rule out lymphoma in most cases. Needless to say, immunocompromised patients with EBV-related immunoproliferative disease always have an EBV infection!

Review of the Literature/Treatment Options (if applicable):
Epstein Barr virus is an almost ubiquitous infection of humans; greater than 90% of all adults have evidence of infection [1, 2]. Transmission appears to be mainly oral, and many patients develop no symptoms, especially those who acquire the infection in childhood; rare cases may be sexually transmitted or transmitted via blood transfusions. In developing nations infection typically occurs in childhood and may be asymptomatic. Adolescent exposure, which occurs more commonly in developed nations, is often associated with symptoms of infectious mononucleosis (IM), which may include mild hepatitis as well as pharyngitis, fever, splenomegaly, and a peripheral lymphocytosis. Occasional patients will develop pancytopenia. Older patients tend to present predominantly with hepatitic symptoms including jaundice. The diagnosis is usually made by recognition of the classical symptoms followed by a positive monospot test. The monospot test has a significant rate of false negative and positive tests, however, so it needs to be interpreted with caution. More specific confirmatory serological tests include tests for early antigen (EA), viral capsid antigen (VCA, looking at both IgM and IgG) and EBV associated nuclear antigen (EBNA) antibody. As typical for viral infection, EA and VCA-IgM occur early with the later development of VCA-IgG and then EBNA. EA appears early and usually disappears after 3 to 6 months; VCA-IgM disappears within 4 to 6 weeks. VCA-IgG also develops in the acute phase, peaks at 2 to 4 weeks and therefore overlaps with VCA-IgM, then declines but persists throughout life. EBNA develops 2 to 4 months after onset of disease, usually after the patient is no longer symptomatic (although there are considerable difference in detection timing depending on the specific test performed) and persist throughout life. Therefore, the presence of EA and VCA-IgM (with or without VCA-IgG but without EBNA) indicates an acute primary infection and the presence of EBNA with VCA-IgG indicates remote infection. VCA-IgG without EBNA would indicate a recovered infection with the acute infection occurring less than 2 months ago. The presence of EA with EBNA can indicate reactivation infection, although there are some patients (apparently about 20%) who have persistence of EA for their lifetime and therefore the presence of EA with EBNA needs to be evaluated in the context of clinical symptoms and other diseases need to be considered as the cause of the symptoms. In cases with liver biopsies in which there is a doubt as to the diagnosis the presence of EBER indicates active infection. Patients with symptoms suggestive of IM but with negative serological testing should be evaluated for CMV, which can present in a very similar fashion and has essentially identical findings on liver biopsy. EBV infects mainly B-lymphocytes and develops into a latent infection. A variety of tumors have been associated with EBV infections, including lymphomas (Burkitts lymphoma, Hodgkins disease, T/NK lymphoma [18], primary effusion lymphoma and B-cell immunoproliferative disease in immunosuppressed patients) as well as carcinomas with a "nasopharyngeal carcinoma" pattern (primarily in the nasopharynx, of course, but also in the stomach and liver, where the tumor may be either an HCC or cholangiocellular carcinoma with nasopharyngeal features) and leiomyosarcomas in HIV-infected patients. Although the virus resides and becomes latent in B-cells, there may be transient infection of epithelial cells of the oropharynx during the primary infection, although this apparently is a matter of debate. Aside from this possible oropharyngeal epithelial infection, infection of other epithelium including epithelial cells in the liver has not been documented. The manifestations of infectious mononucleosis (IM) and of EBV hepatitis seem to be due to a T-cell response to the virally infected B-cells with the damage to hepatocytes being an "innocent bystander" effect; hence, most of the circulating lymphocytes and lymphocytes in the liver are T-cells, as in the current case, and not B-cells. It may be that the phenomenon of abundant hepatocellular mitoses in the face of only minimal necrosis, which is very characteristic of this infection, is the effect of some type of growth stimulation caused by cytokines from these reactive T-cells. An exception to this preponderance of T-cells, of course, is in immunocompromised patients with EBV-related B-cell lymphoproliferative disorder (BLPD, also known as post transplant lymphoproliferative disorder (PTLP) in transplant patients) in which there is a diminished T-cell response and hence the B-cells proliferate in an uncontrolled fashion. It is not at all certain why the liver becomes a target of attack in IM, since the liver is not host to a large population of B-cells. The reason for the peculiar sinusoidal infiltrate of lymphocytes is also completely unknown although it may be a simple "transit" process for the T-cells. It is of interest that the other viral infection which is histologically indistinguishable from EBV hepatitis is CMV hepatitis, a virus well known to infect a variety of cell types in the liver (including hepatocytes, bile duct epithelium, Kupffer cells and endothelial cells) [3, 4, 5, 6]. Why these two viruses with different targets of infection should cause the same characteristic histological pattern is a mystery. Both viruses can also cause granulomatous hepatitis (including cases with fibrin ring granulomas) and both viruses tend to have only mild transaminase elevations in most immunocompetent individuals with the disease, with more cholestatic features in the elderly [7, 8, 9, 10]. Rare cases of severe or even on occasion fatal EBV-related hepatitis have been reported [11, 12, 13]. Both EBV and CMV have also on occasion been associated with ductopenia and neonatal hepatitis [14, 15]. One difference between these two viruses, however, is that CMV in the immunocompromised host does not lead to a B-cell lymphoproliferative process; rather, it leads to CMV inclusion disease with only mild liver dysfunction but with potentially fatal consequences because of the systemic nature of the infection. Another difference is that patients with X-linked lymphoproliferative syndrome are prone to severe liver disease when infected with EBV but not with CMV [16]. Finally, there are occasional reports of EBV causing chronic hepatitis mimicking autoimmune hepatitis and of autoimmune hepatitis being "triggered" by EBV infections [17]. Generally these are all single case reports and there are generally other alternative explanations (e.g. reactivation of latent EBV infection, coincidental infection, etc) and hence at the current time this remains an area of speculation rather than established fact.

Conclusion(s):
When confronted with an unusual pattern of hepatitis in the liver, consider viruses other than the typical hepatotrophic viruses. These include, perhaps most commonly, CMV and EBV, but other systemic viruses (e.g. adenovirus and rhinovirus) can also cause mild self-limited hepatitis in the immunocompetent host.

References:
  1. Macsween KF and Crawford DH. Epstein-Barr virus—recent advances. Lancet Infect Dis 2003; 3: 131–40

  2. Negro F. The paradox of Epstein–Barr virus-associated hepatitis Journal of Hepatology 2006;44:839–841

  3. Kanno A, Abe M, Yamada M, Murakami K. Clinical and histological features of cytomegalovirus hepatitis in previously healthy adults. Liver 1997: 17: 129-132.

  4. Markin RS. Manifestations of Epstein–Barr virus-associated disorders in liver. Liver 1994;14:1–13.

  5. Drebber U, Kasper HU, Krupacz1 J, et al. The role of Epstein–Barr virus in acute and chronic hepatitis. Journal of Hepatology 2006;44: 879–885

  6. Snover DC and Horwitz CA: Liver disease in cytomegalovirus mononucleosis: A light microscopical and immunoperoxidase study of 6 cases. Hepatology 4:408-412, 1984.

  7. Hinedi TB, Koff RS. Cholestatic hepatitis induced by Epstein–Barr virus infection in an adult. Dig Dis Sci 2003;48:539–541.

  8. Biest S, Schubert TT. Chronic Epstein–Barr virus infection: a cause of granulomatous hepatitis? J Clin Gastroenterol 1989;11:343–346.

  9. Shizuma T. A case of infectious mononucleosis complicated with severe jaundice. Kansenshogaku Zasshi 2005;79:149–152.

  10. Shaukat A, Tsai HT, Rutherford R, Anania FA. Epstein–Barr virus induced hepatitis: an important cause of cholestasis. Hepatol Res 2005.

  11. Petersen-Benz C, Hoffmann N, Beckurts T, Goeser T, Steffen HM, Dries V. Fulminant liver failure induced by hepatosplenic alphabeta T-cell lymphoma. Z Gastroenterol 2003;41:1083–1086.

  12. Okano M, Matsumoto S, Osato T, Sakiyama Y, Thiele GM, Purtilo DT. Severe chronic active Epstein–Barr virus infection syndrome. Clin Microbiol Rev 1991;4:129–135.

  13. Chiba T, Goto S, Yokosuka O, et al. Fatal chronic active Epstein–Barr virus infection mimicking autoimmune hepatitis Eur J Gastroenterol Hepatol 2004;16:225–228

  14. Kikuchi K, Miyakawa H, Abe K, Fujikawa H, Horiuchi T, Nagai K, et al. Vanishing bile duct syndrome associated with chronic EBV infection. Dig Dis Sci 2000;45:160–165.

  15. Ozkan TB, Mistik R, Dikici B and Nazlioglu HO. Antiviral therapy in neonatal cholestatic cytomegalovirus hepatitis. BMC Gastroenterology 2007, 7:9

  16. Gilmour KC, Gaspar HB. Pathogenesis and diagnosis of X-linked lymphoproliferative disease. Expert Rev Mol Diagn 2003;3: 549–561.

  17. Cabibi D, Scaduti S, Cacciatore M and Di Gaudio F. Epstein-Barr Virus Infection as A Trigger of Autoimmune Hepatitis: Case Report. Am J Inf Dis 2008;4:200-203

  18. Khan WA, Yu L, Eisenbrey AB, et al. Hepatosplenic gamma/delta T-Cell Lymphoma in Immunocompromised Patients Report of Two Cases and Review of Literature. Am J Clin Pathol 2001;116:41-50