1. ¹Division of Hematology and Oncology, JA Aichi Showa Hospital, Konan, Japan
2. ²Department of Hematology, Toyohashi Municipal Hospital, Toyohashi, Japan E-mail: hiro-na@d4.dion.ne.jp

Graft failure represents a serious complication after allogeneic stem cell transplantation (allo-SCT).¹ Such failure comprises two clinical entities: primary graft failure and graft rejection.² Three treatment options are available for primary graft failure: administration of hematopoietic growth factor; booster infusion of donor hematopoietic stem cells; or infusion of previously harvested autologous hematopoietic stem cells.¹ In the management of graft rejection, further immunosuppressive conditioning is necessary before second infusion of hematopoietic stem cells.¹

Some recent reports have demonstrated the feasibility of cord blood transplantation (CBT) using reduced-intensity regimens (RI-CBT) for adult patients with advanced hematological disease.^{3, 4} Graft failure is a significant complication associated with high mortality in both RI-CBT.⁵ Narimatsu et al.⁵ reported that nine of 123 patients developed graft failure and four patients underwent second RI-CBT and developed severe regimen-related toxicities (RRTs) following a conditioning regimen of fludarabine 125 mg/m², either melphalan 80 mg/m² or busulfan 8 mg/kg, and 2–4 Gy total body irradiation (TBI).

To reduce RRT for patients with graft failure, we employed a conditioning regimen for second RI-CBT in which TBI was omitted. We believe that this report is informative for the establishment of optimal conditioning regimens with second RI-CBT.

A 65-year-old man with myelodysplastic syndrome underwent RI-CBT in August 2003 after providing written informed consent. He had developed myelodysplastic syndrome in March 2001. Progression of the underlying disease had begun in April 2003. No HLA-identical family donors were available, and insufficient time was available to wait for a match from the Japan marrow donor program owing to disease progression. The patient had not received pretransplantation chemotherapy. Congestive heart failure developed in March 2003, and was successfully treated using diuretics.

The transplantation procedure is shown in Table 1. HLA antibody was not detected in serum before RI-CBT. Regimen-related toxicities are shown in Table 1. Peripheral cytopenia persisted after RI-CBT. Bone marrow examination on day 26 revealed severe marrow hypoplasia and 100% host chimerism. Graft rejection was diagnosed.

Table 1 - Characteristics of first and second RI-CBT.

Full table

Second RI-CBT was performed for the treatment of graft rejection, with an interval of 34 days. The patient received a conditioning regimen in which TBI was omitted (Table 1). The same prophylaxis against graft-versus-host disease (GVHD) was used as in the first RI-CBT. Severe RRTs (Table 1) and hepatic veno-occlusive disease developed. Primary engraftment of neutrophils >500/l was achieved on day 28. The patient did not receive granulocyte transfusions. Bone marrow examination showed complete donor chimerism. The patient died of multiorgan failure due to RRTs on day 30 after second RI-CBT.

This case indicates that TBI might not be necessary in conditioning regimens for second RI-CBT to achieve engraftment in patients with graft rejection. A reduced dose of melphalan, such as 60 mg/m², could have ameliorated RRTs while ensuring engraftment, as our patient rapidly achieved full donor chimerism. Dosage of melphalan warrants further investigation.

The safety of second RI-CBT must be improved. Severe RRTs following second RI-CBT directly contributed to the death of our patient. Less toxic, immunosuppressive conditioning regimens, such as cyclophosphamide and anti-thymocyte globulin (ATG), are recommended in second transplantation for graft rejection after conventional allo-SCT.⁶ However, cyclophosphamide was avoided in this case owing to concerns regarding cardiotoxicity for this patient with impaired cardiac function. Anti-thymocyte globulin strongly suppresses host immunity in addition to donor immune cells in the graft, and remains in the recipient circulation owing to a long half-life. Furthermore, ATG in RI-CBT may increase the risk of graft failure. Owing to these considerations, we avoided ATG and selected agents with short half-lives, such as fludarabine and melphalan. In contrast, a high rate of engraftment has been reported in CBT using ATG.⁷ Further studies on ATG for second RI-CBT are warranted. Alemtuzumab might also be worth investigating.

Cyclosporine alone was used for GVHD prophylaxis owing to concerns that additional immunosuppressive agents, such as methotrexate, steroids or mycophenolate mofetil, might suppress donor immune cell function in the graft. However, GVHD tends to be severe after second transplantation.⁸ Graft-versus-host disease prophylaxis could be intensified for these mostly 4/6 HLA-matched unrelated cord blood transplants. We are planning a further study on RI-CBT using intensive GVHD prophylaxis.

The etiology of graft failure in RI-CBT remains unclear. Infused total cell count represents a significant factor for primary engraftment.⁹ However, this might not have been an important factor for our patient, as total cell dose was higher in the first RI-CBT than in second RI-CBT. According to the definitions previously described,⁵ both our patient and previously reported patients⁵ were classifiable as experiencing graft rejection. This indicates that immunological mechanisms are important in graft failure after RI-CBT. Several possible etiologies of graft rejection after RI-CBT can be suggested. First, HLA antibody might be involved in graft rejection. Most patients who receive RI-CBT have high-risk underlying disease and had received a high amount of transfusion. These patients are considered at high-risk of developing HLA antibody. HLA antibody was not detected in serum from our patient. However, no studies have yet examined the presence of HLA antibody following RI-CBT. Further investigations are required. Second, our patient did not receive any pretransplantation chemotherapy. Residual host T cells might thus be involved in the immunological mechanisms of graft rejection. This possibility is supported by a report that graft failure after RI-CBT is restricted to patients without recent pretransplantation combination chemotherapy or prior autologous transplantation.⁴ Evaluation of alloreactivity in residual host T cells might also be useful in identifying patients at high risk of graft failure. Third, killer immunoglobulin-like receptors might be involved in graft failure. HLA-C disparity is associated with graft failure in conventional allo-SCT.¹⁰ However, information on HLA-C disparity was unavailable for both our patient and previously reported patients.⁵ This avenue might be worth exploring.

In summary, we described the case of a patient with graft failure who achieved engraftment after second RI-CBT; RI-CBT is an optimal treatment. However, safety must be improved. Identifying the etiology of graft failure in RI-CBT would be helpful in establishing an optimal procedure.