A case-controlled study on the use of HBsAg-positive donors for allogeneic hematopoietic cell transplantation

As in many other parts of the world, allogeneic hematopoietic cell transplantation (HCT) is increasingly being used for treatment of various hematological and oncological diseases in Hong Kong.1,2 Hong Kong is also endemic for hepatitis B virus (HBV) infection, with a carrier rate of approximately 10%.3 With the very stringent human leukocyte antigen (HLA) matching of donor and recipient required for allogeneic HCT, there are often not many choices of suitable donors available for patients. A donor positive for hepatitis B surface antigen (HBsAg) may be the only option. Before the era of routine HBV vaccination in Hong Kong, contact with an HBV-infected mother during the perinatal period was the usual mode of transmission of the infection. Therefore, HBV infection often runs in a families.4 As a result, sibling marrow donors of Hong Kong Chinese patients with HBV infection are also commonly either HBsAg positive or hepatitis B surface antibody (HBsAb) positive.

There is always a concern as to whether the use of HBsAg-positive marrow is associated with an increased rate of morbidity and mortality after allogeneic HCT. It has been well recognized that severe liver-related complications frequently occur after HCT is performed for HBsAg-positive patients.5-12 On the other hand, only very scanty data are currently available in the literature on the clinical and serological outcomes of patients who receive a transplant in which HBsAg-positive marrow is used. The precise risk of using such donors for transplant is uncertain.13 In a recent multicenter retrospective survey in Italy, it was shown that the use of HBsAg-positive donors, particularly if they are anti-HBe–positive, significantly increases the risk of severe liver disease in the recipients following transplant.14 

To address this issue further, we have performed a case-control study to investigate the clinical and serological outcomes of our patients who have received an HBsAg-positive marrow transplant.

From May 1990 to May 1998, we studied 353 consecutive patients who underwent allogeneic HCT at Queen Mary Hospital, Hong Kong. In accordance with our standard protocols, all donors and recipients for HCT were screened for HBsAg, HBsAb, and human immunodeficiency virus antibody (HIV Ab). In addition, all donors and recipients were screened for hepatitis C antibody (anti-HCV) after July 1993. Pretransplant serum of both donors and recipients collected before July 1993 were retrospectively tested for anti-HCV. For all HBsAg-positive donors and recipients, further serological testing for Hepatitis B e antigen (HBeAg), Hepatitis B e antibody (HBeAb), and serum HBV DNA by polymerase chain reaction (PCR) were performed within 4 to 8 weeks of marrow harvest or HCT. In addition, all HBsAg-positive pretransplant serum of both donors and recipients collected after December 1995 was tested for HBV DNA by branched DNA (bDNA) assay within 4 to 8 weeks of marrow harvest of HCT. All HBsAg-positive pretransplant serum of both donors and recipients collected before December 1995 was tested retrospectively for HBV DNA by bDNA assay. The precore and core promoter HBV DNA sequences for all HBsAg-positive marrow donors and recipients were also determined by direct DNA sequencing.

Eighteen of these 353 allogeneic HCTs (5.1%) were performed with the use of HBsAg-positive marrow (group 1). These donors were used despite their hepatitis B status, as they were the only HLA-matched donors available. At 4 to 8 weeks prior to the harvesting of the marrow, all HBsAg-positive donors were assessed by experienced hepatologists with a complete physical examination and liver function test. None of the HBsAg-positive donors had clinical evidence of decompensated cirrhosis (ankle edema, ascites, jaundice, and hepatic encephalopathy), and their liver function tests (serum aminotransferase, albumin, bilirubin, and prothrombin time) were all within normal range. Informed consent was obtained from all donors and recipients of the marrow harvest for the use of HBsAg-positive donors' marrow. Careful explanation and counseling were given to these recipients concerning the risk of contracting hepatitis B infection with the use of HBsAg-positive marrow. It was also explained to them that hepatitis B infection after HCT might cause significant morbidity and mortality.

No donors or recipients received any antiviral therapy with significant activities against HBV (famciclovir or lamivudine), intravenous immunoglobulin, or HBV vaccination within 12 weeks of transplant. Another 18 recipients (group 2) receiving HBsAg-negative marrow were selected from the cohort, and they were matched with the 18 patients in group 1 in terms of hepatitis B serology (HBsAg, HBeAg, HBeAb, HBsAb status), sex, age, underlying hematological malignancies, conditioning regimen, and prophylaxis against graft-versus-host disease (GVHD) (Table 1). The titers of HBsAb and hepatitis B core antibody (HBcAb) were determined for all HBsAb-positive marrow donors and recipients in both groups. Some HBsAg-positive recipients in groups 1 and 2 were included in our previous study.11,15 16 

The transplant recipients of both groups were tested at least once weekly for liver function (including serum alanine aminotransferase, serum albumin, and bilirubin) during the first 12 weeks immediately after transplant, and then at 2- to 8-week intervals until 52 weeks after transplant, or their last follow-up. Hepatitis serology (HBsAg, HBeAg, HBeAb, HBV DNA by bDNA assay and PCR, and HCV RNA by PCR) were performed whenever there was any clinical suspicion of liver damage from hepatitis B infection. The patients were followed up for a median duration of 24.5 months (range, 1-91 months). The occurrence of hepatic events (acute hepatitis, chronic hepatitis, anicteric and icteric hepatitis, hepatic failure, veno-occlusive disease, hepatic GVHD) and death were recorded. At least 1 pretransplant serum sample was available from all transplant recipients included in this study (ranging from 1 to 6 weeks before transplant). During the posttransplant period, serum samples were available from all recipients at an interval time of 1 to 8 weeks, from the day of transplant until their last follow-up or death. At least 6 posttransplant serum samples were available from each patient in the study. All serum samples were stored at −70°C.

Hepatitis was defined as a more than 3-fold elevation of serum aminotransferase on 2 consecutive determinations at least 5 days apart, in the absence of clinical features suggestive of veno-occlusive disease, GVHD, or infection by cytomegalovirus or herpes simplex virus. Icteric hepatitis was defined as hepatitis associated with a clinical jaundice and serum bilirubin level that exceeded 30 μmol/L (normal is below 19 μmol/L). Hepatitis was defined as acute if the duration was less than 6 months and as chronic if the duration was more than 6 months. Hepatitis was defined as HBV-related when it was preceded by an elevation of serum HBV DNA to more than 10 times that of the preexacerbation baseline, the serum HBV DNA turned from negative to positive, or the HBsAg became positive and remained so for 2 consecutive readings 5 days apart. Hepatic failure was defined as the presence of hepatic encephalopathy and deranged blood coagulation (prothrombin time prolonged for more than 10 seconds). Veno-occlusive disease was defined and graded as described by McDonald et al,17 and acute GVHD of the liver was defined and graded in accordance with the Glucksberg criteria.18 

Hepatitis B serological markers, including HBsAg, HBsAb, HBeAg, HBeAb, anti-HCV (enzyme immunoassay II), anti–hepatitis-D antibody, and anti-HIV antibody, were tested by commercially available enzyme immunoassays (Abbott Laboratories, Chicago, IL). HBcAb was tested by radioimmunoassay (Corab; Abbott Laboratories). Anti-HBs antibody titers were determined by enzyme-linked immunoabsorbent assay (Abbott Laboratories) according to the manufacturer's recommendations. Anti-HBs levels were converted to international units with the use of a standard serum sample. Serum HBV DNA level was quantified by a bDNA assay (Abbott Laboratories).19Serial serum samples from the same patient were tested within a single run to minimize interassay variation. A nested PCR assay for the detection of serum HBV DNA was performed with the use of primer sets from the HBV surface antigen and the core antigen coding region.15 By serial dilution of the EuroHep-2 HBV DNA plasma standards, the detection limits of the bDNA assay and PCR were estimated to be 7 × 105 and 8 × 10² genomes/mL, respectively. Reverse-transcription–nested PCR for HCV and direct sequencing of precore HBV were also performed as described previously.16 20 

The precore HBV sequence was determined by direct sequencing of HBV DNA amplified by PCR with the use of primers that flanked the precore region, namely P1 (5′-TCCTCTGCCGATCCATACTG, position 1254-1273) and BC1 (5′GGAAAGAAGTCAGAAGGCAA, position 1974-1955), at a concentration of 50 pmol. Direct sequencing of the PCR products was carried out with the fmol sequencing Kit (Promega Co, Southampton, England). Sequencing primer BC1 was end-labeled with 25μCi of³²P-adenosine triphosphase (Amersham International, Amersham, England) in a volume of 10 μL with the use of 10 U of polynucleotide kinase (Boehringer Mannheim, Penzberg, Germany) for 20 minutes at 37°C. We used 1.5μL of the reaction mix directly in sequencing reactions. We precipitated 90μL of PCR product with an equal volume of 4 mol/L sodium acetate and 2 volumes of isopropanol for 10 minutes at room temperature, pelleted it by centrifugation for 10 minutes, and then washed it twice with 70% ethanol. After resuspension in 20μL of water, 9.5 μL of DNA was added to the reaction buffer with the end-labeled primer and 5 units of Taq enzyme. Dideoxynucleotide termination sequencing was performed according to the manufacturer's instructions. Cycle sequencing was performed for 30 rounds. Denaturation, annealing, and extension steps were 30 seconds, 30 seconds, and 1 minute respectively.20 

The chi-square test or the Fisher exact test was used to compare the hepatic events (acute hepatitis, chronic hepatitis, and hepatic failure), serological outcome, and death between HCT recipients in groups 1 and 2 and to compare the various clinical and serological characteristics of those within group 1 who had and those in group 1 who did not have HBV-related hepatitis. The Statistical Package of Social Sciences was used, and P < .05 was considered significant.

Fourteen of the 18 (77.8%) recipients receiving HBsAg-positive marrow (group 1) and 8 of the 18 (44.4%) recipients receiving HBsAg-negative marrow (group 2) developed clinical hepatitis after transplant (P = .04) (Table 2). The cause of the hepatitis is listed in Table 2. None (0%) of the 14 episodes of hepatitis of group 1, but 4 of the 8 episodes (50%) of group 2 were related to serological clearance of serum HBsAg and/or HBeAg in the recipients after the use of HBsAb- and HBcAb-positive marrow for transplant (P = .01).

There were no significant differences in the incidence rates of acute, chronic, anicteric, and icteric hepatitis between the 2 groups. Six of the 18 (33.3%) recipients in group 1 and none (0%) of group 2 developed hepatic failure (P = .02). Seven of the 18 patients of group 1 and 5 of the 18 patients of group 2 developed hepatic GVHD (P = not significant [NS]). Fourteen of the 18 patients of group 1 and 10 of the 18 of group 2 suffered from veno-occlusive diseases clinically (P = NS). Following transplant, there was a total of 11 (61.1%) deaths from all causes in group 1 and 4 (22.2%) in group 2 (P = .02) (Table 2).

Eight (44.4%) patients in group 1 and 2 (11.1%) in group 2 suffered from HBV-related hepatitis (P = .03). The time of onset of HBV-related hepatitis was similar for both groups of patients (2 ± 12 weeks versus 19 ± 10 weeks, P = NS). The incidence of hepatitis B–related acute, chronic, anicteric, and icteric hepatitis was not statistically different in the 2 groups (Table 2). Six (33.3%) patients in group 1 and none in group 2 developed hepatitis B–related hepatic failure (P = .02); 4 of these patients died.

Four (44.4%) of the 9 HBsAg-positive recipients in group 1 and 2 (22.2%) of the 9 HBsAg-positive recipients in group 2 had post-HCT HBV-related hepatitis (P = NS). None (0%) of the HBsAg-negative recipients in group 2 and 4 (44.4%) of the 9 HBsAg-negative recipients in group 1 suffered from post-HCT HBV-related hepatitis (P = .08) (Table 3).

The median age for those group 1 recipients who developed HBV-related hepatitis was 25 years (13-45) and did not differ from those who did not develop HBV-related hepatitis (29 years; 16-36) (P = NS). Six (75%) who developed HBV-related hepatitis and 9 (90%) who did not develop HBV-related hepatitis were male (P = NS). There were also no differences according to the preconditioning regimen; 8 (100%) of those with HBV-related hepatitis and 7 (70%) without HBV-related hepatitis had total body irradiation–conditioning regimen (P = NS).

There was a trend suggesting that the group 1 patients developing HBV-related hepatitis posttransplant were more likely to have had evidence of clinical hepatitis (50% versus 20%, P = .30) and to have been HBsAg positive (62.5% versus 40%, P = .64) pretransplant. Also, they were less likely to be HBsAb positive (12.5% versus 50%, P = .15). However, because of the small number of patients studied, none of these observed differences reached statistical significance (Table 4).

The HBV virological characteristics of the donors were also investigated. In group 1, patients who had HBV-related hepatitis after receiving HBsAg-positive marrow were more likely to have a donor carrying a precore A₁₈₉₆ and/or core promoter T₁₇₆₂/A₁₇₆₄ HBV variants (62.5% versus 0%,P = .007) and detectable pretransplant serum HBV DNA (by bDNA assay) (62.5% versus 10%, P = .02) (Table 4).

Of the 9 HBsAg-negative recipients of group 1, 5 (2 HBV negative and 3 HBsAb positive) became HBsAg positive after receiving HBsAg-positive marrow, all within 4 weeks of the marrow infusion. Of the seroconversions, 3 were sustainable, while 2 were transient (Table 3). On the other hand, none of the 9 HBsAg-negative recipients of group 2 became HBsAg positive after transplant (P = .03). In group 1, all 5 HBsAg-negative recipients became HBsAg positive after receiving marrow from donors with detectable pre-HCT serum HBV DNA (by bDNA assay). In contrast, none of the donors of the other 4 who remained HBsAg negative had detectable serum HBV DNA (P = .008).

After transplantation, 5 anti-HBs–negative recipients in group 2 and none in group 1 acquired anti-HBs positivity (P = .045). All the recipients in group 2 who acquired anti-HBs positivity had received anti-HBs donors' marrow. Among them, 3 HBsAg-positive recipients who had received anti-HBs and anti-HBc–positive marrow had serological clearance HBsAg (Table 3).

Of the 6 recipients of group 1 who were HBsAb positive before transplant, 3 (50%) remained HBsAb positive serologically after a median posttransplant follow-up time of 19 months (range, 5-64 months). The remaining 3 became HBsAg positive and HBeAg positive after a median follow-up time of 25 months (range, 1-87 months) after transplant. There were no significant differences in pretransplant anti-HBs antibody titers, between those who remained HBsAb and those who became HBsAg positive (986.7 ± 1570.5 IU/mL versus 1566.2 ± 472.6 IU/mL, P = NS).

Limited data are currently available on the use of HBsAg-positive donors for allogeneic HCT.13,14 21 The present study has demonstrated a high incidence of posttransplant HBV-related hepatitis (44%) for patients receiving HBsAg-positive donors' marrow. Attempts were made in the study to determine the risk factors predicting development of HBV-related hepatitis posttransplant. Various pretransplant characteristics of the transplant recipients were investigated. There was a trend suggesting that the risk was higher for patients who were already HBsAg positive and had clinical evidence of hepatitis before transplant. However, with a relatively small sample size, the differences observed did not reach statistical significance. It was therefore uncertain whether reactivation of preexisting HBV infection in the recipients might in some cases be contributing significantly to the development of HBV-related hepatitis posttransplant.

Of the transplant recipients in the study, 6 were HBsAb positive before receiving the HBsAg-positive marrow. Despite the intensive myeloablation and immunosuppression of the transplant, 3 of them (50%) remained HBsAb positive posttransplant. The protective effect of anti-HBV immunity seemed to be preserved in at least some of these HBsAb-positive recipients. The other 3 patients became HBsAg positive and 1 of them had HBV-related hepatitis posttransplant. This observation demonstrated that pretransplant immunity against HBV in some of the transplant recipients might confer adequate protection against HBV infection after the receipt of HBsAg-positive marrow.22-24 The persistence of anti-HBV activity in the recipient after transplant might be related to the presence of residual recipients' lymphocytes, which escaped the intensive conditioning given for the transplant. Long-term persistence of hemopoietic chimerism has been shown to be present following allogeneic HCT.25 

Donor factors were also investigated in this study. High HBV viral load, as reflected by a high serum HBV DNA level in the donors, was a significant factor in determining the development of HBV-related hepatitis in the recipients posttransplant. The Italian study14 showed that use of HBeAb-positive marrow for transplant is associated with a higher incidence of liver-related morbidity in the recipients. However, the observation was not confirmed by the present study. Instead, it was possible that the presence of precore A₁₈₉₆ and/or core promoter T₁₇₆₂/A₁₇₆₄ variants in the HBsAg-positive donors played an important role. It is known that these virus variants are associated with rapid replication of HBV DNA, despite the absence of HBeAg.26,27 It has also been reported that HBeAg-negative patients with the core promoter T₁₇₆₂/A₁₇₆₄ variant may have enhanced HBV replication, leading to severe liver damage.28 In the present study, recipients with HBV-related hepatitis posttransplant were more likely to have received marrow from donors with the precore A₁₈₉₆ and/or core promoter T₁₇₆₂/A₁₇₆₄ variants, than those without hepatitis (62.5% versus 0%, P = .007). The exact role of these HBV variants is uncertain from these data. Five of the 9 HBsAg-negative recipients (3 HBsAb positive and 2 HBV negative) in this study who became HBsAg positive were given serum HBV DNA–positive (by bDNA assay) marrow from their HBsAg-positive donors. On the other hand, the other 4 HBsAg-negative recipients, who remained HBsAg negative, received only HBsAg-positive, but serum HBV DNA–negative, marrow. It was apparent that active HBV replication and a high viral load, as reflected by serum HBV DNA positivity of the donors, was a major factor determining transmission of HBV infection posttransplant. Because of its extremely high sensitivity and ability to detect very low levels of HBV DNA, the PCR technique is probably not as useful in this setting.29-31 

For the patients who became HBsAg positive, the seroconversion usually occurred within 4 weeks of receiving the HBsAg-positive marrow. The HBV-related hepatitis, however, did not occur until a median time of 20 weeks after transplant. This was probably related to the immune-mediated mechanism of HBV-induced liver damage. A period of intensive immunosuppression usually occurs after transplant as a result of the myeloablative and immunosuppressive agents given to the recipients. During this time, HBV replication is greatly enhanced.32 Immune reconstitution starts to occur in the recipients at around 12 to 24 weeks after transplant.33With the gradual withdrawal of immunosuppression, immune-mediated damage on HBV-laden hepatocytes may take place, resulting in hepatitis or even hepatic failure.5-12 The events are similar to the case of HBV reactivation following the withdrawal of chemotherapy in cancer patients.34-39 

A strategy should be developed to minimize the risk of using HBsAg-positive marrow transplant when there are no better alternatives. Attempts should probably be made to reduce the HBV viral load in the donor's marrow. Several antiviral agents have been shown to be active in suppressing HBV replication. Treatment with interferon-α40,41 or nucleoside analogues, such as famciclovir and lamivudine,42-46 can dramatically reduce the amount of circulating HBV DNA in most patients with chronic HBV infection. However, nucleoside analogues are preferable as they do not cause significant myelosuppression, a potential concern with the use of interferon-α.13 It may be advisable to treat all HBsAg-positive donors with nucleoside analogues to render them HBV DNA negative (bDNA assay) before transplant. At the same time, the recipients may also be given at least 1 of the nucleoside analogues after transplant to reduce the chance of being infected by the HBsAg-positive marrow. As up to 50% of these recipients were HBsAg positive themselves prior to transplant, the use of nucleoside analogues may minimize the risk of HBV reactivation posttransplant.32 

An alternative strategy would be to induce or enhance anti-HBV immunity in the recipients. This, however, is technically more difficult. The efficacy of the HBV vaccines currently available is doubtful in this setting. The responsiveness of transplant recipients to vaccination is very unpredictable because of the primary disease of the patients and the heavy immunosuppression. Giving intramuscular injections of hepatitis B immune globulin to these patients is not always possible because of the presence of thrombocytopenia.13 Intravenous immunoglobulin with significant anti-HBV activity content has been used, but its effectiveness needs further investigation.47 

Several hypotheses have been generated from this study concerning the use of HBsAg-positive marrow for allogeneic HCT. However, owing to the small number of patients available to be studied, more detailed statistical analysis to discern the risk factors of HBV-related hepatic events was not possible. Further studies would be required to clarify the interaction between the various host factors, such as anti-HBV immunity and immunosuppression, and the donor factors, such as viral load and HBV variants, in this setting. Future clinical trials would also be necessary to determine the optimal approach to the management of this problem.

The authors thank Yuen Wing Sze and Amy Kwok for their assistance in the data management.