Introduction

Hepatitis-associated aplastic anemia (HAAA), first described by Lorenz and Quaiser in 1955 [1], is a rare phenomenon which usually affects adolescents and young men [2] and is more common in the East than in the West [3–5]. It represents a minor proportion of all aplastic anemia cases [6] and may occur in 28% of young adults after liver transplantation for non-A, non-B hepatitis [7]. Although a limited number of aplastic anemia cases have been reported in association with hepatitis A, B, and G, parvovirus B19, Epstein-Barr virus (EBV), transfusion transmitted virus (TTV), and echovirus [8–20], the causative agent in most of the cases still remains unknown [21, 23]. Patients typically develop severe aplastic anemia 2 to 3 months after an episode of acute hepatitis. Aplastic anemia is always fatal if untreated [2, 11] and the hepatitis associated with this may be mild and transient [2, 22], fulminant [7, 23], or chronic [2]. There is no known relation to blood transfusion, drugs, or toxins. Though not clearly defined thus far, immunologic mechanisms seem to be responsible for HAAA, particularly because immunosuppressive therapy has shown some promising results. Nonetheless, most successful outcomes have been achieved by HLA-matched bone marrow transplantation (BMT) [21].

We report an adolescent male admitted with acute hepatitis of unknown etiology whose course was subsequently complicated by severe aplastic anemia. As an original observation, many CD8-expressing Kupffer cells were detected in the liver biopsy specimen. Despite continuing immunosuppressive therapy, the patient died after an intracranial hemorrhage. Etiopathogenesis of and treatment options for this intriguing case are also discussed.

Table 1 Initial laboratory values
Full size table

Case report

A 16-year-old male student was admitted to an inner-city state hospital with 4 weeks of generalized weakness, malaise, and jaundice. His initial biochemical tests showed profound transaminitis and hyperbilirubinemia, with ALT of 1186 U/L, AST of 948 U/L, and total bilirubin of 8.94 mg/dl. Initial viral markers including HBs, anti-HBc IgM, anti-HBc IgG, anti-HBs, anti-HAV IgM, and anti-HCV were negative. The patient was referred to a central state hospital with an initial diagnosis of non-A, non-B, non-C hepatitis. There he was tested again for hepatitis A, B, C, and E as well as EBV and cytomegalovirus (CMV) serology and only anti-HAV IgG was found to be positive. The tests for Wilson disease, primary biliary cirrhosis, and α-1 antitrypsin deficiency were unremarkable. Meanwhile, his complete blood count started to drop and he developed pancytopenia, upon which he was referred to our hospital's gastroenterology clinic for further evaluation.

On admission to our unit, the patient did not give any history of blood transfusion, drug intake, or toxic substance exposure. On physical examination, vital signs were normal, however, jaundice and conjunctival pallor were noted. There was no sign of hepatic decompensation. His initial laboratory values (Table 1) revealed the following: ALT, 1026 U/L (normal: 0–41 U/L); AST, 679 U/L (normal: 0–38 U/L); ALP, 902 U/L (normal: 0–270 U/L); GGT, 119 U/L (normal: 0–61 U/L); total bilirubin, 8.63 mg/dl (normal: 0.1–1.2 mg/dl); direct, 5.76 mg/dl (normal: 0.0–0.8 mg/dl); Hb, 9.4 mg/dl (normal: 13–17 g/dl); Hct, 26.9% (normal: 40–52%); total WBC count, 1350/mm3 (normal: 3800–10,600/mm3); neutrophil count, 648/mm3; and platelet count, 26,000/mm3 (normal: 150,000–440,000).

Serologic tests for hepatitis B and C as well as HBV DNA PCR and HCV RNA PCR were negative. EBV, CMV, and HSV-1 serology ruled out acute infection. Serum copper and seruloplasmin values were in normal limits. Autoimmune markers including antinuclear antibody, antimitochondrial antibody, anti-smooth muscle antibody, and Coombs tests, serum haptoglobin, and acid-sucrose lysis test were unremarkable.

Serum IgG level was increased, but serum IgM and IgA were within normal limits. Bone marrow biopsy showed a severely hypoplastic marrow (Fig. 1). Liver biopsy revealed chronic active hepatitis (Fig. 2). HBs immunohistochemistry was negative. Based on these findings, the diagnosis of HAAA was made and other rarely seen viruses that could be responsible for this disease entity were investigated. However, HGV RNA PCR, parvovirus B19 DNA PCR, and TTV DNA PCR were all negative.

Fig. 1
figure 1

Bone marrow biopsy showing severe marrow hypoplasia

Full size image
Fig. 2
figure 2

Liver biopsy demonstrating features of chronic active hepatitis, with lymphocyte infiltration and mild fibrosis in portal tracts with mild interface hepatitis. There is mild parencymal cholestasis and a few spotty necroses, in addition to sinusoidal dilatation and focal hyperplasia of Kupffer cells. (H&E.)

Full size image

On day 22 of the patient's admission, immunosuppressive therapy with antithymocyte globulin (ATG; Lymphoglobulin; Aventis Pasteur) at 20 mg/kg per day and cyclosporine A (Cy A) at 5 mg/kg per day was initiated for aplastic anemia since blood counts continued to decline. He was also pretreated with 40 mg of methylprednisolone before ATG infusions. He was further supported with G-CSF, 30 MU per day, and platelet and washed red blood cell infusions as needed. His liver function recovered promptly, with ALT and AST levels normalizing on day 35, 13 days after the initiation of therapy with ATG and Cy A (Fig. 3). However, the patient's complete blood count still continued to decline, and by day 32 his WBC count was down to 400/mm3; Hb, 4.9 mg/dl; and platelet count, 3000/mm3. At this point, he developed febrile neutropenia, bilateral retinal hemorrhage, and bleeding gums. He was treated with intravenous broad-spectrum antibiotics, blood products, and G-CSF as well. The patient clinically recovered within 18 days, although the complete blood count was still at very low levels (Fig. 4). Cy A was continued without any hematologic response and therefore a search for a BMT donor was carried out and one of the patient's siblings was found to have 100% HLA-matched bone marrow. He was listed for BMT with priority. Unfortunatelly, on day 104, 82 days after starting therapy with ATG and Cy A, he died suddenly of acute massive intracerebral hemorrhage as evidenced by physical examination and cranial CT.

Fig. 3
figure 3

The course of liver enzymes before and after admission to our hospital (Day 0). Slowly declining liver enzymes were promptly reduced to normal after starting ATG + Cy A (Day 22)

Full size image
Fig. 4
figure 4

Blood count showing no response to immunosuppressive therapy

Full size image

Discussion

HAAA is typically a severe aplastic anemia which usually ensues within 2 to 3 months after the onset of acute hepatitis. In our case this interval was almost 6 weeks. Except for the rare cases that might be associated with hepatitis A, B, and G, parvovirus B19, and EBV, hepatitis in the majority is thought to be caused by a still unknown virus. Hashimato et al. [12] showed some structures resembling the nuclear capsids of paramyxovirus and some particles by electron microscope in liver specimens at autopsy. Brown et al. suggested that seronegative acute viral hepatitis and HAAA might share a single infectious cause that could be a member of Flaviviridae [24]. We performed various serologic tests to investigate the possible etiologic agent responsible for hepatitis, however, we could not detect any known virus or any other disease entity that might be relevant to such a clinical condition.

Chronic active hepatitis was observed on liver biopsy, though it was performed at the time of marked transaminitis. Interestingly, many Kupffer cells were observed, crowding particularly in periportal areas (Fig. 5). We also immunostained the sections to see if there is any role of CD8(+) cytotoxic lymphocytes in the pathogenesis of hepatitis in HAAA as shown previously for aplastic anemia. Scant intralobular CD8(+) lymphocytes were noted (Fig. 6). Surprisingly, however, a lot of Kupffer cells were demonstrated expressing CD8. We hypothesize that this may be related to inefficient clearing and therefore, possibly, persistence of a virus and consequent chronic hepatitis in this patient as shown very recently by Tang et al. in chronic hepatitis B [25]. They revealed that the number of intralobular CD8(+) T cells was significantly decreased and that of Fas-L(+) Kupffer cells was upregulated in biopsies of patients with high ALT, and conversely low viral replication was associated with increased intralobular CD8(+) T cells. Interestingly, CD8(+) macrophages, a population of monocytes/macrophages with a cytotoxic phenotype, were also reported recently in rat experimental models of human autoimmune diseases such as multiple sclerosis infiltrating at inflammatory sites [26, 27]. The authors suggested a specific role of CD8(+) macrophages in the pathogenesis of inflammatory tissue damage. Likewise, they may have a specific, yet unidentified role in the immunopathogenesis of hepatitis in HAAA.

Fig. 5
figure 5

Increased Kupffer cells, more frequent in periportal areas. (H&E.)

Full size image
Fig. 6
figure 6

A section showing a lymphocyte (thick arrow) and a Kupffer cell (thin arrow), both of which are positively immunostained for CD8

Full size image

Several immunologic abnormalities and a good response to immunosuppressive therapy have been described in some previous studies [24, 28, 29], leading to a consideration that the underlying pathologic mechanism in HAAA is immune-mediated. Brown et al. [24] reported that these patients had a decreased CD4/CD8 ratio in the peripheral blood, in association with increased activated cytotoxic T cells. Our findings support those of Brown et al. in that the patient had a very low CD4/CD8 ratio (0.04; normal range, 0.8–2.8) and a high percentage of HLA-DR-positive CD8 cells (74.6%). Kagan et al. [30] demonstrated in vitro that in the case of aplastic anemia these activated CD8-positive lymphocytes could be cytotoxic to myelopoietic cells in the bone marrow itself. In addition, several other studies showed that activated cytotoxic lymphocytes settle and produce increased amounts of interferon-γ in the bone marrow in patients with aplastic anemia [31, 32]. The same events may also hold true for the liver inflammation associated with HAAA as in the case of hepatitis B and C [33, 34]. In a recent report, Bowen and colleagues concluded that intrahepatic accumulation of activated CD8-positive T cells with hepatitis could result from primary activation of these cells by the liver resident bone marrow-derived cells, inducing bystander damage to non-antigen-bearing hepatocytes in the transgenic mouse model [35]. They also showed that this bystander hepatitis was dependent on tumor necrosis factor α and interferon-γ, suggesting that this mechanism may play a role in some forms of biologically significant hepatitis. Thus, whatever the offending agent, it may lead to chronic changes in the liver despite normalization of liver enzymes by immunosuppressive therapy as was the case in our patient.

The poor prognosis of HAAA justifies a search for an HLA-identical sibling donor to be able to proceed with a prompt BMT. Long-term survival after transplantation from an HLA-matched sibling donor is 60% to 90% [36–38] and the success rate is increased with an immunosuppressive conditioning regimen. Moreover, the survival rate after BMT with stem cells from an HLA-matched sibling is similar to that for patients with non-hepatitis-associated aplastic anemia [21]. Our experience supports the previous reports favoring early BMT [38–40], although the analysis of a 10-patient series by Brown et al. [24] and some other case reports [28, 29) showed good response to ATG and Cy A combination therapy in HAAA. Brown et al. continued Cy A therapy for at least 6 months, with a complete response in the majority of cases. Early treatment with this combination seems to be more effective.

Overall, aplastic anemia may follow cryptogenic hepatitis, and although we could not determine the etiologic agent, our observations clearly indicate an immune-mediated process leading to both aplastic anemia and hepatitis. Aplastic anemia may have a grave prognosis. Hepatitis may vary from mild to acute or chronic forms and may recover with immunosuppressive therapy. CD8-expressing Kupffer cells may be important mediators of hepatitis associated with aplastic anemia. We suggest an early attempt at BMT if aplastic anemia does not respond to short-term (i.e., 1-month) immunosuppressive therapy.