Prevalence of Hepatitis G Virus in Patients With Aplastic Anemia

To the Editor:

Hepatitis G virus (HGV) is a recently discovered flavivirus with sequence homology to hepatitis C virus (HCV).1 The role of HGV in causing liver disease, however, is in doubt, raising the question as to whether there are other illnesses caused by this virus.2,3 Several reports have noted an association between HGV and hepatitis-associated aplastic anemia (HA-aplastic anemia).4-6 In a recent study, Brown et al7 reported the presence of HGV in 3 of 10 patients with HA-aplastic anemia. However, all three patients had received multiple blood transfusions, and the investigators concluded that HGV was not a likely cause of their HA-aplastic anemia.

We firstly have examined the prevalence of HGV RNA in sera (stored at −70°C) from a well-characterized cohort of 16 patients with HA-aplastic anemia8 as well as from 29 concurrent patients with idiopathic aplastic anemia not related to hepatitis, who were matched for age, year of transplant, and transfusion status. The diagnosis of HA-aplastic anemia was made on the basis of acute hepatitis (elevated serum aminotransferase enzymes, jaundice) and lack of evidence of drug-induced hepatitis, toxin exposure or biliary disease, followed by the development of a hypocellular marrow and at least two of these parameters: reticulocyte count <1%, neutrophil count <500/μL, platelet count <20,000/μL. RNA extracted from 100 μL of serum was tested for HGV sequences by polymerase chain reaction. HGV RNA was amplified with two primer pairs for the 5′UTR and NS5 regions (supplied by Genelabs Technologies, Inc, Redwood City, CA). All reaction products were tested by Southern hybridization with digoxigenin-labeled internal probes (sequences were supplied by Genelabs Technologies, Inc). Samples were run in duplicate. Results are shown in Table 1. Ten of the 16 samples from patients with HA-aplastic anemia and 16 of 29 samples from control patients with idiopathic aplastic anemia tested positive for HGV RNA. Logistic regression analysis showed that HA-aplastic anemia patients were not significantly more likely than idiopathic aplastic anemia patients to be HGV RNA–positive (odds ratio = 1.4; P = .63).

Next, to compare results in transfused versus untransfused patients, we combined data from patients with HA-aplastic anemia and idiopathic aplastic anemia given that we had not seen a difference in the prevalence of HGV between the two groups. To ensure that an adequate number of untransfused patients were included and comparison groups were matched for age and year of transplant, additional matched patients were selected (Table 2). Logistic regression analysis comparing the proportion of HGV-positive transfused and untransfused patients showed that the odds of being positive for HGV RNA for transfused patients was 5.9 times that for untransfused patients (P = .001). This result implicated transfusions as a major source of HGV. Even so, 26.1% of untransfused patients were HGV RNA–positive (2 of 4 patients with HA-aplastic anemia and 4 of 19 patients with idiopathic aplastic anemia), a prevalence that was significantly higher than the 1.7% of 779 consecutively screened volunteer blood donors with normal aminotransferase values reported by Linnen et al1 (P < .001, Fisher's Exact test).

We conclude that the prevalence of HGV RNA was significantly higher in transfused patients than in untransfused patients with idiopathic aplastic anemia or HA-aplastic anemia, suggesting that transfusions were a major source of HGV RNA in the serum of these patients. However, the increased prevalence of HGV RNA in untransfused patients with aplastic anemia, whether associated with hepatitis or not, compared with normal blood donors1 suggests involvement of HGV in the development of aplastic anemia in some patients, although not necessarily through the preceding manifestations of active hepatitis. These results point to the need for prospective studies of patients with newly diagnosed aplastic anemia — before transfusion support — to further explore a possible etiologic role for HGV.