Transplantation of allogeneic hematopoietic stem cells for adult T-cell leukemia: a nationwide retrospective study

Adult T-cell leukemia (ATL) is a mature T-cell neoplasm developing in a minority of persons infected with human T-cell leukemia virus type I (HTLV-I), the first retrovirus isolated from a human malignant disease.^1-4  HTLV-I is estimated to infect 10 to 20 million people worldwide and is endemic in some areas of Japan, sub-Saharan Africa, the Caribbean Basin, and South America.^5,6  The area with the highest HTLV-I prevalence is the Kyushu district in southwestern Japan, where more than 10% of the general population is infected and the cumulative incidence of developing ATL among adult virus carriers is estimated at approximately 6.6% for males and 2.1% for females.⁷  The onset of ATL after HTLV-I infection appears to require a long latency period because the median age at diagnosis ranges from 40 to 60 years in most endemic regions where mother-to-child viral transmission had been previously common.^4-6 

Clinical manifestation of ATL is heterogeneous and characterized by various degrees of lymphadenopathy, abnormal lymphocytosis, hepatosplenomegaly, skin lesions, and hypercalcemia, dividing the disease into 4 subtypes: acute, lymphomatous, chronic, and smoldering.⁸  Patients with acute or lymphomatous type had extremely poor prognosis, mainly because of resistance to a variety of cytotoxic agents and susceptibility to opportunistic infections. Chronic and smoldering forms have relatively indolent clinical courses but can transform into more aggressive subtypes. During the past 3 decades since the clinical discovery of ATL,¹  the results of conventional cytotoxic chemotherapy remain dismal because of low response rates and lack of long-term efficacy. The median survival time that followed the best clinical results to date is approximately 13 months^9,10 ; complete response can only be achieved in 25%-40% of treated cases and most of them eventually relapsed with the median progression-free survival time of 5 to 7 months, whereas available treatment options are extremely limited in those who failed initial chemotherapy.^11-14 

Although the early experience of ablative chemoradiotherapy with autologous hematopoietic stem cell rescue for ATL resulted in a high incidence of relapse and fatal toxicities,¹⁵  allogeneic hematopoietic stem cell transplantation (HSCT) has been explored as a promising alternative that can provide long-term remission in a proportion of patients with ATL.^16-19  Although the mechanisms by which allografting can eradicate HTLV-I–infected neoplastic T cells are not fully elucidated, several reports have suggested the role of graft-versus-HTLV-I or graft-versus-ATL effects.^20-23  Over the past decade, improved access to alternative stem cell sources and the development of less toxic conditioning regimens have led to a rapid increase in the number of cases of ATL treated with allogeneic HSCT, albeit without consistent efficacy.^24-30  Therefore, we conducted a nationwide retrospective cohort study to identify pretransplantation factors that affect survival after allografting for ATL, with special emphasis on the effect of graft source: we compared the outcomes of human leukocyte antigen (HLA)–mismatched related bone marrow or peripheral blood transplantation, unrelated bone marrow transplantation, and unrelated cord blood transplantation with those of HLA-matched related bone marrow or peripheral blood transplantation as treatment for ATL. We also evaluated the effect of donor HTLV-I serostatus on outcomes among patients who received transplants from related donors.

Data on 417 patients with acute or lymphomatous type ATL who had received T-cell–replete allogeneic bone marrow, peripheral blood, or cord blood transplantation between January 1, 1996, and December 31, 2005, were collected through the 3 largest hematopoietic cell transplant registries in our country: the Japan Society for Hematopoietic Cell Transplantation (JSHCT), the Japan Marrow Donor Program (JMDP), and the Japan Cord Blood Bank Network (JCBBN). The patients were included from 102 transplant centers; the data were updated as of December 2008. To evaluate the effect of HTLV-I infection in donors on transplantation outcomes, additional questionnaires were sent to 77 centers in January 2010 to retrieve data on donor HTLV-I serostatus in 217 related transplants registered with the JSHCT. Our analysis included patients for whom there was data on age at transplantation, sex, donor type, stem cell source, and agents used in the conditioning regimen and graft-versus-host disease (GVHD) prophylaxis. Twenty-two patients who missed any of these data, and 8 patients who had a history of prior autologous or allogeneic stem cell transplantation were excluded from the analysis. One patient who had received an ex vivo T-cell–depleted graft was also excluded. Two independent physicians reviewed the quality of collected data, and a total of 386 patients (209 males and 177 females), with a median age of 51 years (range, 18-79 years), were found to fulfill the inclusion criteria: 197 patients from JSHCT, 99 from JMDP, and 90 from JCBBN. No overlapping cases were identified. Data on engraftment or graft failure were missing in 23 patients. Data on acute GVHD were not available in 53 patients because of early death or missing data.

The JSHCT registry currently includes more than 390 transplant centers variously located in Japan and collects data on transplantation by use of autologous or related stem cell grafts. The JMDP includes more than 190 centers and collects data on unrelated bone marrow transplantation. The JCBBN, a national network of 11 cord blood banks, collects data on unrelated cord blood transplantations reported individually from more than 220 transplant centers to each bank. Participating centers to these registries are requested to report each type of transplantation consecutively and longitudinally. Until 2005, the 3 registries were operated separately from one another: however, a project attempting to unify them has been launched via development of the Transplant Registry Unified Management Program, which enables participating centers to use a shared format for data submission to each registry.³¹  All unrelated donor transplants in Japan were facilitated through the JMDP and JCBBN, although peripheral blood donation from unrelated volunteers has not yet been instituted as of March 2010. The study was approved by the data management committees of the JSHCT, JMDP, and JCBBN, as well as by the institutional review boards of Kyoto University, Graduate School of Medicine, where this study was organized.

The primary end point of the study was overall survival, defined as the time from the date of transplantation until date of death from any cause. Patients who remained alive at the time of last follow-up were censored. Reported causes of death were reviewed and categorized into disease-associated or treatment-associated deaths. Disease-associated deaths were defined as deaths from relapse or progression of ATL among patients who survived for at least 30 days after transplantation. Treatment-related deaths were defined as any death other than disease-associated deaths. Neutrophil recovery was considered to have occurred when an absolute neutrophil count exceeded 0.5 × 10⁹/L for 3 consecutive days after transplantation. Primary graft failure was evaluated in patients who survived at least 30 days and was defined as no evidence of neutrophil recovery after transplantation. Acute and chronic GVHD were diagnosed and graded using traditional criteria by the physicians who performed transplantations at each center.^32,33  The incidence of acute GVHD was evaluated in patients who survived for at least 7 days, and that of chronic GVHD was evaluated in patients who survived for at least 100 days.

Descriptive statistics were used for summarizing variables related to patient demographics and transplant characteristics. Comparisons among the groups were performed by use of the χ² statistic or extended Fisher exact test as appropriate for categorical variables, and the Kruskal-Wallis test for continuous variables. The probability of overall survival was estimated according to the Kaplan-Meier method, and univariable comparisons among the groups were made using the log-rank test. Probabilities of acute and chronic GVHD, treatment-related mortality, and disease-associated mortality were estimated with the use of cumulative incidence curves to accommodate the following competing events³⁴ : death without GVHD for acute and chronic GVHD, disease-associated death for treatment-related mortality, and treatment-related death for disease-associated mortality. Data on patients who were alive at the time of last follow-up were censored. Cox proportional-hazards regression was used to evaluate variables potentially affecting overall survival, whereas Fine and Gray proportional-hazard model was used to evaluate variables affecting other outcomes.³⁵  The variables considered were recipient age group (≤ 50 years or > 50 years at transplantation); recipient sex; disease status before transplantation; type of conditioning regimen; type of GVHD prophylaxis; type of graft source; time from diagnosis to transplantation (within 6 months or longer than 6 months); and year of transplantation. Only factors differing in distribution among the graft source groups and factors associated with outcomes by univariable comparison were included in the final models. The effect of donor HTLV-I seropositivity on outcomes after related donor transplantation was also evaluated by univariable and multivariable analysis with the use of data on 156 patients given transplants from siblings or other related family members for whom data on the HTLV-I serostatus were available. Results were expressed as hazard ratios and their 95% confidence interval (CI). All tests were 2-sided, and a P value of less than .05 was considered to indicate statistical significance. All statistical analyses were performed with STATA software (Version 11; Stata Corporation).

Table 1 shows characteristics of the patients and transplantation procedures. Compared with HLA-matched related bone marrow or peripheral blood recipients, HLA-mismatched bone marrow or peripheral blood recipients were more likely to receive tacrolimus for GVHD prophylaxis; unrelated bone marrow recipients took a longer time from diagnosis to transplantation, were more likely to have attained complete remission at transplantation, and were more likely to receive tacrolimus for GVHD prophylaxis; unrelated cord blood recipients were older, underwent transplantation more recently, and were more likely to receive purine analog–containing conditioning regimens. All unrelated cord blood recipients received a single cord blood unit that was not manipulated ex vivo. The median weight of unrelated cord blood recipients was 52.0 kg (range, 31.0-90.2 kg); the median dose of nucleated cells and CD34⁺ progenitor cells in the grafts, measured before freezing, was 2.55 × 10⁷ (range, 1.39-5.34 × 10⁷) and 0.79 × 10⁵ (range, 0.07-3.15 × 10⁵) per kg of recipient body weight, respectively.

Of 310 patients who survived 30 days after transplantation and were evaluable for engraftment, primary graft failure was reported in 2 (6%) of 35 recipients of HLA-mismatched related grafts and in 12 (17%) of 70 recipients of unrelated cord blood, whereas the remaining 296 patients had evidence of initial engraftment. Acute GVHD of grades II, III, or IV occurred in 158 (47%) of 333 evaluable patients; 69 (49%) of 140 HLA-matched related bone marrow or peripheral blood recipients, 20 (56%) of 36 HLA-mismatched related bone marrow or peripheral blood recipients, 40 (44%) of 91 unrelated bone marrow recipients, and 29 (44%) of 66 unrelated cord blood recipients. In a multivariable analysis, rates of grades II to IV acute GVHD did not significantly differ among the 4 groups (supplemental Table 1; available on the Blood Web site; see the Supplemental Materials link at the top of the online article). Chronic GVHD occurred in 94 (48%) of 195 evaluable patients at a significantly lower rate among the unrelated cord blood recipients than among HLA-matched graft recipients (hazard ratio, 0.25; 95% CI, 0.10-0.61, P = .002).

Of 333 patients who survived 30 days after transplantation, 136 patients experienced relapse or progression of ATL at a median of 76 days (range, 1-1964 days) after transplantation. ATL recurred or progressed in 52 (37%) of 141 recipients of HLA-matched related grafts, in 19 (51%) of 37 recipients of HLA-mismatched related grafts, in 27 (32%) of 85 recipients of unrelated bone marrow, and 38 (54%) of 70 recipients of unrelated cord blood. Of 113 patients who were evaluable for the date of relapse or disease progression, the median time from transplantation to relapse or progression of ATL was 65.5 days (range, 1-1964 days) for HLA-matched related bone marrow or peripheral blood recipients, 63 days (range, 22-269 days) for HLA-mismatched related bone marrow or peripheral blood recipients, 152 days (range, 42-819 days) for unrelated bone marrow recipients, and 83 days (range, 7-596 days) for unrelated cord blood recipients.

Of 386 patients included in the study, a total of 125 patients were alive and 101 patients were alive in continuous complete remission after a median follow-up of 41 months (range, 1.5-102 months). The unadjusted 3-year probability of overall survival was 33% (95% CI, 28%-38%) for the whole cohort; 41% (95% CI, 33%-49%) in HLA-matched related graft recipients; 24% (95% CI, 12%-38%) in HLA-mismatched related graft recipients; 39% (95% CI, 29%-49%) in unrelated bone marrow recipients; and 17% (95% CI, 9%-25%) in unrelated cord blood recipients (Figure 1). The median overall survival time after transplantation was 9.8 months for HLA-matched related bone marrow or peripheral blood recipients, 2.5 months for HLA-mismatched related bone marrow or peripheral blood recipients, 9.6 months for unrelated bone marrow recipients, and 2.6 months for unrelated cord blood recipients. Patients who received transplants in complete remission had a higher probability of survival than those who received transplants not in complete remission (51% [95% CI, 41%-60%] vs 26% [95% CI, 20%-31%], P < .001). Multivariable analyses revealed 4 factors that adversely affected overall survival: older recipient age (> 50 years; hazard ratio, 1.56; 95% CI, 1.14-2.12, P = .005), male recipient (hazard ratio, 1.37; 95% CI, 1.07-1.77, P = .014), lack of complete remission at transplantation (hazard ratio, 2.01; 95% CI, 1.50-2.71, P < .001), and transplantation of unrelated cord blood. Hazard ratios for death among recipients of HLA-mismatched related transplants, unrelated bone marrow transplants, and unrelated cord blood transplants, compared with that among recipients of HLA-matched related transplants, were 1.55 (95% CI, 0.98-2.45, P = .063), 1.24 (95% CI, 0.82-1.88, P = .312), and 2.08 (95% CI, 1.43-3.02, P < .001), respectively (Table 2).

Overall, 161 (43%) of 376 evaluable patients succumbed to treatment-related complications. Cumulative incidence of treatment-related mortality at 3 years after transplantation was 37% (95% CI, 29%-45%) in HLA-matched related bone marrow or peripheral blood recipients, 43% (95% CI, 28%-57%) in HLA-mismatched related bone marrow or peripheral blood recipients, 42% (95% CI, 32%-51%) in unrelated bone marrow recipients, and 52% (95% CI, 41%-62%) in unrelated cord blood recipients (Figure 2A). When adjusted by multivariable analysis, patients given unrelated cord blood (hazard ratio, 1.77; 95% CI, 1.10-2.86, P = .019) had higher treatment-related mortality rates (Table 2).

Deaths from progression of ATL occurred in 90 (24%) patients. Cumulative incidence of disease-associated mortality at 3 years after transplantation was 21% (95% CI, 14%-28%) in HLA-matched related bone marrow or peripheral blood recipients, 32% (95% CI, 19%-47%) in HLA-mismatched related bone marrow or peripheral blood recipients, 19% (95% CI, 12%-28%) in unrelated bone marrow recipients, and 30% (95% CI, 21%-40%) in unrelated cord blood recipients (Figure 2B). In multivariable analysis, patients given transplants not in remission (hazard ratio, 2.55; 95% CI 1.50-4.33, P = .001) or male recipients (hazard ratio, 1.86; 95% CI, 1.17-2.95, P = .008) had higher rates of disease-associated mortality (Table 2).

Causes of death after transplantation are summarized in Table 3. Of the 161 patients who died of treatment-related complications, 51 (32%) succumbed to infection and 53 (33%) to organ failure. Treatment-related events were principal causes of early death, whereas death from relapse or progression of ATL was more common later than 100 days after transplantation, irrespective of types of graft source.

Data on donor HTLV-I serostatus were available for analysis in 156 of 197 patients given related transplants; 68 received transplants from an HTLV-I–seropositive donor and 88 from an HTLV-I–seronegative donor. Patients who received transplants from HTLV-I–seropositive donors and those from HTLV-I–seronegative donors had similar background characteristics (supplemental Table 2). Among 113 patients who had data on donor HTLV-I serostatus and maintained or attained complete remission after transplantation, relapse of ATL was observed in 18 (38%) of 48 patients who received transplants from an HTLV-I–seropositive donor, and 16 (25%) of 65 patients who received transplants from an HTLV-I–seronegative donor with a median follow-up time for survivors of 40 months (range, 7.3-102 months). In univariable and multivariable analysis, patients who received transplants from an HTLV-I–seropositive donor had a higher risk of disease-associated mortality compared with those who received transplants from an HTLV-I–seronegative donor, whereas they had similar overall survival and treatment-related mortality rates (Table 4).

The aim of this nationwide registry-based study was to compare overall survival after allogeneic HSCT with the use of various graft sources as treatment for ATL, and to identify factors that may influence transplantation outcomes. Despite the retrospective nature of the study, the validity of our analysis is strengthened by the fact that our cohort included most of the related transplants and nearly all unrelated transplants for ATL performed over a decade in our country.

We found that a substantial proportion of patients with ATL, including those who did not achieve complete remission, could enjoy long-term survival after allogeneic HSCT, validating the results of earlier observations.^18,19  However, our analysis in this cohort also revealed a high rate of treatment-related mortality. In particular, frequent incidence of fatal infectious complications may reflect preexisting profound immunodeficiency observed in patients with ATL.^4,5  Improved supportive care for opportunistic infection might be especially important for reducing treatment-related mortality in allografting for ATL.

Multivariable analysis revealed 4 factors that affected survival: recipient age, recipient sex, disease status before transplantation, and type of graft source. Although higher age of the recipient was associated with lower posttransplantation survival, most of the patients with ATL were older than age 50 years and were less likely to be candidates for fully ablative conditioning. Recently, 2 small prospective trials have demonstrated the feasibility and efficacy of allogeneic stem cell transplantations using reduced-intensity conditioning.^26,29  Although we observed no significant differences in overall survival between patients who received conventional conditioning regimens and those who received purine analog–based regimens in the present study, it was difficult to evaluate the effect of conditioning dose intensity because data on doses of agents or total-body irradiation used in these regimens were not fully available in our cohort. Further studies are warranted to identify unfit or elderly ATL patients who can benefit from allogeneic stem cell transplantation with the use of less toxic conditioning.

A further novel finding in this study was that female patients with ATL had a more favorable outcome after allogeneic stem cell transplantation compared with male patients. Incidence of ATL in Japan is generally higher in male than in female populations, which was partly explained by the difference in routes of HTLV-I transmission between males and females. Sexual transmission of the virus can also occur, predominantly from males to females in adult life, thereby lowering the apparent incidence of ATL among female HTLV-I carriers.⁷  However, the estimated ATL mortality among a prospective cohort of perinatally infected HTLV-I carriers was still higher for male patients,³⁶  suggesting that female sex itself might have a protective role against ATL development. Although much of the underlying mechanism for male predominance in ATL remains to be elucidated,³⁷  unidentified biologic or immunologic aspects of sex difference may contribute not only to development of ATL in HTLV-I carriers, but also to outcomes in allografted patients with ATL.

Despite the high risk for relapse after transplantation, survival rates observed in patients who received transplants not in complete remission were encouraging. Intriguingly, withdrawal of immunosuppressive agents or donor lymphocyte infusion can induce remission in relapse of ATL after allogeneic HSCT, implying the presence of a graft-versus-ATL effect.^19-23  Because several antigens have recently been identified as putative targets for cytotoxic T-cell responses against ATL,^38,39  future development of cellular immunotherapy targeting these molecules would reduce the incidence of relapse and improve survival in patients with residual ATL after allogeneic transplantation. Further investigations are warranted to elucidate the association between the occurrence of GVHD and disease response among allografted patients with ATL because our preliminary analysis using a similar cohort⁴⁰  suggested that patients who developed mild acute GVHD had a better posttransplantation survival compared with those who did not develop acute GVHD (J.K., M. Hishizawa, A.U., S.T., T.E., Y. Moriuchi, R.T., F.K., Y. Miyazaki, M.M., K.N., M. Hara, M.T., S. Kai, Y.A., R.S., T.K., K.M., T.N.-I., S. Kato, H.S., Y. Morishima, J.O., T.I., and T.U., manuscript in preparation).

Finally, the use of unrelated cord blood was associated with lower survival, most likely a result of higher treatment-related mortality. Two major causes of early treatment-related death were infection and organ failure. Because the development of ATL usually worsens preceding immunodeficiency associated with HTLV-I infection, it is imperative to establish effective measures to manage posttransplantation infections in allografted patients with ATL. In addition, the use of more intense conditioning for refractory disease in relatively elderly recipients may increase the risk of regimen-related toxicities especially in the setting of unrelated donor transplantation. However, direct comparison of transplantation outcomes by graft source was not feasible because the selection of graft source is an individual process strongly influenced by donor availability and disease status of patients. It should also be noted that the study period encompassed the developmental phase of cord blood transplantation in adults. Because rates of disease-associated death were similar irrespective of type of graft source, new strategies to reduce early treatment-related mortality would improve the results of alternative donor transplantations for ATL.

Another concern related to selection of graft source involves the use of HTLV-I-seropositive-related donors. Sibling donors for patients with ATL are frequently infected with HTLV-I, because mother-to-child transmission by breastfeeding is a major route of HTLV-I acquisition.^5,6  The use of HTLV-I–seropositive donors raises the risk of ATL development in donor-derived HTLV-I–infected cells under immunosuppressive conditions after transplantation,⁴¹  whereas it may enhance the therapeutic effect by the adoptive transfer of viral-specific immunocompetent cells.²¹  However, the latter possibility seems less likely because transplantation from HTLV-I–seropositive donors was associated with higher risk for disease-associated mortality in our study cohort. Given that donor-derived HTLV-I–specific cytotoxic T-cell response can be observed in transplantation from an HTLV-I–seronegative donor,²¹  it is important to note that the magnitude of specific T-cell responsiveness to HTLV-I might widely differ among healthy HTLV-I carriers. The impairment of HTLV-1–specific T-cell responses was observed not only in patients with advanced ATL but also in a subpopulation of asymptomatic carriers, which was associated with insufficient control of HTLV-I.⁴²  Although whether donor anti–HTLV-I immunity can harness graft-versus-ATL responses is still elusive, further investigations are clearly needed to resolve this issue.

This study had inherent limitations that are common among observational studies: eligibility for transplantation, as well as choice of transplantation protocol, including the selection of graft source, was determined by the treating physicians of each institution; the confounding effect of some variables, such as disease subtype, could not be fully evaluated because of missing data, although adjustment for other key risk factors enabled as controlled a comparison as possible.

In conclusion, allogeneic HSCT is an effective treatment that confers long-term survival in selected patients with ATL, but at the cost of substantial risk of treatment-related mortality. Posttransplantation outcomes are influenced by recipient age, recipient sex, and disease status at transplantation, as well as by type of graft source. More definitive conclusions on the role of allografting in the therapeutic algorithm for ATL will be drawn from future prospective studies that aim to compare the survival outcomes after transplantation with those after conventional chemotherapy.

We are indebted to all the physicians and data managers at the centers who contributed valuable data on transplantation for ATL to the Japan Society for Hematopoietic Cell Transplantation (JSHCT), the Japan Marrow Donor Program (JMDP), and the Japan Cord Blood Bank Network (JCBBN). We also thank all the members of the data management committees of JSHCT, JMDP, and JCBBN for management of data, James de Witt and Takao Nakanishi (ASCA Corporation) for editorial assistance, and Dr Takayuki Ishikawa (Kyoto University) for critical reading of the manuscript.

This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (T.U.).

The views expressed in this report are those of authors and do not indicate the views of the JSHCT, JMDP, or JCBBN.

Contribution: M. Hishizawa, J.K., T.I., and T.U. reviewed and analyzed data and wrote the paper; J.K., K.M., and T.I. performed statistical analysis; A.U., S.T., T.E., Y. Moriuchi, R.T., F.K., Y. Miyazaki, M.M., K.N., M. Hara, M.T., S. Kai, and J.O. interpreted data and reviewed and approved the final manuscript; Y.A., R.S., and H.S. collected data from the JSHCT; T.K. and Y. Morishima collected data from the JMDP; T.N.-I. and S. Kato collected data from the JCBBN; and T.I. and T.U. designed the research and organized the project.

T.U., the senior author, died after acceptance of the final manuscript.

In addition to authors, other members who contributed data on allogeneic hematopoietic stem cell transplantation for adult T-cell leukemia to the JSHCT, JMDP, and JCBBN are listed in the supplemental Appendix.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Tatsuo Ichinohe, Department of Hematology and Oncology, Graduate School of Medicine, Kyoto University, 54 Shogoin Kawaharacho Sakyo-ku, Kyoto 606-8507, Japan; e-mail: nohe@kuhp.kyoto-u.ac.jp.