|Home | About | Journals | Submit | Contact Us | Français|
Identify prognostic factors that influence outcome after unrelated donor bone marrow transplantation in children with acute myeloid leukemia (AML).
Included are 268 patients (age ≤ 18 years) with AML in second complete remission (n = 142), relapse (n = 90), or primary induction failure (n = 36) at transplantation. All patients received bone marrow grafts from an unrelated donor and a myeloablative conditioning regimen. Cox regression models were constructed to identify risk factors that influence outcome after transplantation.
In this analysis, the only risk factor that predicted leukemia recurrence and overall and leukemia-free survival was disease status at transplantation. The 5-year probabilities of leukemia-free survival were 45%, 20%, and 12% for patients who underwent transplantation at second complete remission, relapse, and primary induction failure, respectively. As expected, risk of acute but not chronic graft-versus-host disease (GVHD) was lower with T-cell–depleted bone marrow grafts; T-cell–depleted grafts were not associated with higher risks of leukemia recurrence. We observed similar risks of leukemia relapse in patients with and without acute and chronic GVHD.
Children who underwent transplantation in remission had a superior outcome compared with children who underwent transplantation during relapse or persistent disease. Nevertheless, 20% of children who underwent transplantation in relapse are long-term survivors, suggesting that unrelated donor bone marrow transplantation is an effective therapy in a significant proportion of children with recurrent or primary refractory AML.
Eighty to ninety percent of children with acute myeloid leukemia (AML) treated on current chemotherapeutic trials achieve a complete remission.1-6 However, 30% to 40% of patients achieving a first remission experience relapse, and less than a third of these patients with recurrent leukemia survive long term.7-12 Although most reports identify the length of first remission as the best predictor of survival, others have reported that sex and French-American-British classification are predictors of achieving a second remission and long-term survival.7-12 Therapies for patients who experience relapse are variable and often include allogeneic hematopoietic stem-cell transplantation when a suitable donor is available. In this report, we sought to identify prognostic factors that influence outcome after unrelated donor bone marrow transplantation in children with AML who experience leukemia recurrence after achieving a first complete remission or who received transplantation for primary induction failure.
The National Marrow Donor Program (NMDP) collects detailed demographic, disease, and transplantation characteristics and outcome data on all unrelated donor transplantations it facilitates in the United States. All patients are observed longitudinally, and computerized error checks, physician review of submitted data, and on-site audits of participating centers ensure data quality. The NMDP retrospectively obtained consent for data submission and study participation from surviving patients or their parent/legal guardian for transplantations it facilitated in the United States before 2002. Thereafter, informed consent was obtained prospectively. The Institutional Review Board of the NMDP waived consent for patients who had died before soliciting consent (transplantations facilitated before 2002). To overcome the bias caused by the inclusion of a proportion of surviving patients (those consenting) but all deceased recipients, and hence their over-representation, a sample of deceased patients was selected using a weighted randomized scheme that adjusts for over-representation of deceased patients in the consented cohort.13 This weighted randomized scheme was developed based on all survivors in the NMDP database. A logistic regression model was fit to identify the factors that predicted whether a patient had consented or not consented to use of data collected by the NMDP. This analysis found that the following factors were associated with the likelihood of a patient consenting: age, disease type, race, sex, cytomegalovirus serostatus, and country of transplantation (United States v not United States). Using estimated consenting probabilities from this model based on the characteristics of dead patients, the biased coin method of randomization was performed to determine which of the dead patients likely would have consented to participate had they been alive. Thus, this procedure includes the dead patients at the same probability as surviving patients who consented to participate. Approximately 13% of surviving patients declined to consent, and 12% of dead patients were deleted by the weighted randomized method. This method was tested several times, and on every occasion, the proportion of deleted dead patients was similar.
The study population includes 268 recipients of unrelated donor bone marrow transplantations performed in the United States between 1990 and 2003. Patients (age ≤ 18 years at transplantation) with AML who underwent transplantation in second complete remission or first or subsequent relapse and patients with primary induction failure are included. Complete remission was defined as neutrophil count more than 1.0 × 109/L; platelets more than 100 × 109/L; RBC transfusion independent; less than 5% blasts in bone marrow with absence of cells with Auer rods; normal maturation of the erythrocytic, granulocytic, and megakaryocytic series; and absence of extramedullary disease. All patients received bone marrow grafts and a myeloablative transplantation conditioning regimen. Patients who received an unrelated donor bone marrow transplantation in first complete remission, patients in third or subsequent remission, children with Down's syndrome, and recipients of peripheral-blood or umbilical cord blood grafts were excluded because risk factors were likely to vary in these groups.
The primary outcomes studied were neutrophil and platelet recovery, acute and chronic graft-versus-host disease (GVHD), and early and overall mortality. Neutrophil recovery was defined as achieving an absolute neutrophil count of ≥ 0.5 × 109/L and platelets more than 20 × 109/L unsupported for 7 days. Failure to achieve an absolute neutrophil count of ≥ 0.5 × 109/L or a decline to less than 0.5 × 109/L after an initial recovery and without a subsequent recovery was considered graft failure. Incidence of grades 2, 3, and 4 acute GVHD and chronic GVHD were determined in all patients. Diagnosis of acute14 and chronic GVHD15 was based on local institutional criteria, with overall grade of acute GVHD assigned retrospectively by the NMDP based on stage of involvement reported for each individual organ. Any death occurring during continuous remission was defined as treatment-related mortality. Relapse was defined as morphologic leukemia recurrence at any site, and leukemia-free survival was defined as survival in a state of continuous complete remission.
The probabilities of leukemia-free and overall survival were calculated using the Kaplan-Meier method.16 For analysis of survival, death from any cause was considered an event, and data on patients still alive were censored at date of last follow-up. For analysis of leukemia-free survival, leukemia relapse or death from any cause is considered an event, and patients were censored at last follow-up. The probabilities of neutrophil and platelet recovery, acute and chronic GVHD, transplantation-related mortality, and relapse were calculated with the use of the cumulative incidence function method.16 For neutrophil and platelet recovery and GVHD, death without the event (hematopoietic recovery or GVHD) was the competing event. Data on patients without an event were censored at last follow-up. For relapse, transplantation-related mortality was the competing event, and for transplantation-related mortality, relapse was the competing event. CIs were calculated using log transformation.
Cox regression models were built for analysis of risk factors for GVHD, transplantation-related mortality, relapse, treatment failure, and overall mortality.17 Multivariate models were built with the use of stepwise forward selection, with P ≤ .01 considered to indicate statistical significance. The variable for cytogenetic risk group did not attain the level of significance. Given the reported prognostic significance of cytogenetics in AML, analyses for transplantation-related mortality, relapse, treatment failure, and overall mortality were stratified by cytogenetic risk group. All variables met the proportional hazards assumptions. Variables considered in multivariate model building are listed in Table 1. We tested for an effect of transplantation center on outcome and found none.18 P values are two sided. Analyses were performed using SAS software, version 9.1 (SAS Institute, Cary, NC).
Patient, disease, and transplantation characteristics by disease status at transplantation are listed in Table 1. Median age at transplantation was 10 years (range, < 1 to 18 years), and median time from diagnosis to transplantation was 13 months (range, < 1 to 88 months). Twenty patients (7%) had myelodysplastic syndrome that evolved to AML before transplantation. Fifty-three percent of patients were in second complete remission, 34% were in first or subsequent relapse at transplantation, and 13% had primary induction failure. Sixty-eight (76%) of 90 patients who underwent transplantation in relapse received chemotherapy before transplantation but did not achieve clinical remission. Most patients (87%) received total-body irradiation–containing conditioning regimens. All patients received bone marrow grafts, and 38% of bone marrow grafts were T-cell depleted. Median follow-up time of surviving patients is 5 years (range, 5 to 156 months).
Most patients achieved neutrophil and platelet recovery. The probability of neutrophil recovery at day 28 was 95% (95% CI, 83% to 98%), and the probability of platelet recovery at day 28 was 70% (95% CI, 64% to 76%).
Grade 2 to 4 acute GVHD rates were lower after transplantation of T-cell–depleted bone marrow grafts (hazard ratio [HR] = 0.55; 95% CI, 0.37 to 0.82; P = .003); the probability of grade 2 to 4 acute GVHD at day 100 was 31% (95% CI, 22% to 40%) after transplantation of T-cell–depleted bone marrow compared with 52% (95% CI, 44% to 59%) after transplantation of non–T-cell–depleted grafts. In the current analysis, none of the factors tested was predictive for chronic GVHD. The 5-year probability of chronic GVHD was 34% (95% CI, 28% to 40%).
Transplant-related mortality rates were higher in older patients (age 11 to 18 years) compared with those aged ≤ 10 years. The probabilities of early (day 100) and late (5-year) transplantation-related mortality in patients aged ≤ 10 years were 14% (95% CI, 9% to 21%) and 19% (95% CI, 12% to 26%), respectively; the corresponding probabilities in older patients were 25% (95% CI, 18% to 32%) and 41% (95% CI, 33% to 50%), respectively.
Relapse rates were higher in patients who underwent transplantation in first or second relapse and at primary induction failure compared with patients who underwent transplantation in second complete remission, stratified for cytogenetic risk (Table 2). Rates were similar when transplantation was performed at relapse or primary induction failure (HR = 1.33; 95% CI, 0.73 to 2.40; P = .355). The 5-year probabilities of leukemia recurrence were 22% (95% CI, 15% to 29%), 57% (95% CI, 46% to 67%), and 51% (95% CI, 34% to 66%) for patients who underwent transplantation in second complete remission, relapse, and primary induction failure, respectively (Fig 1). The duration of first complete remission was not associated with leukemia recurrence after transplantation (HR = 1.61; 95% CI, 0.91 to 2.86; P = .106). In the current study, we observed similar rates of relapse in patients with and without acute and chronic GVHD (data not shown).
Treatment failure rates (relapse or death; inverse of leukemia-free survival) were higher in patients who underwent transplantation in first or second relapse and primary induction failure, stratified for cytogenetic risk (Table 2). Failure rates were similar when transplantation was performed in primary induction failure and in relapse (HR = 1.56; 95% CI, 0.98 to 2.48; P = .059). Treatment failure rates were not associated with duration of first complete remission (HR = 1.43; 95% CI, 0.95 to 2.13; P = .084). The 5-year probabilities of leukemia-free survival were 45% (95% CI, 37% to 54%), 20% (95% CI, 13% to 30%), and 12% (95% CI, 3% to 26%) for patients who underwent transplantation in second clinical remission, relapse, and primary induction failure, respectively.
One hundred sixty-three patients died; overall mortality rates were higher in patients who underwent transplantation in primary induction failure compared with patients who underwent transplantation in second complete remission, stratified for cytogenetic risk (Table 2). The 5-year probabilities of overall survival were 47% (95% CI, 38% to 55%), 28% (95% CI, 19% to 38%), and 17% (95% CI, 7% to 32%) for patients who underwent transplantation in second complete remission, relapse, and primary induction failure, respectively (Fig 2). Mortality rates were similar in patients who underwent transplantation in primary induction failure compared with patients in relapse (HR = 1.48; 95% CI, 0.91 to 2.39; P = .111). Recurrent leukemia was the cause of death in 85 of 163 patients who died. Other causes of mortality included GVHD (n = 18), infection (n = 30), adult respiratory distress syndrome or interstitial pneumonitis (n = 9), hemorrhage (n = 8), organ failure (n = 11), and graft failure (n = 2).
Sibling donor transplantation in first remission has been used for many years as a primary approach to treatment of AML, with survival rates of 60% to 70% reported from many centers and cooperative groups.2,3,5,6 Improvements in risk stratification on the basis of genetic abnormalities in leukemic blasts and more effective chemotherapy protocols now allow the identification of subgroups of children with AML for whom transplantation is deemed unnecessary in first remission because cure with chemotherapy is equally likely.2 In parallel with interest in limiting use of transplantation in children with a good prognosis, there is increased interest in investigating whether unrelated donor transplantation can improve outcomes for children with particularly poor prognoses, such as those with primary induction failure and those who relapse after a first remission.11,19,20 In this study, we have explored outcomes in a large group of children receiving unrelated donor transplantations facilitated by the NMDP in the United States to determine how successful this therapy is in rescuing children for whom chemotherapy has been ineffective and to identify risk factors that predict a good outcome after transplantation.
The majority of children included in this study underwent transplantation in second remission. Overall, outcomes were encouraging, with almost half of the children receiving transplantation in second complete remission surviving 5 years later and significant numbers of survivors among the children with refractory disease receiving transplantation. In the current report, the only risk factor that predicted relapse, overall survival, and leukemia-free survival was disease status at the time of transplantation, with children who underwent transplantation in second complete remission having superior outcomes to children who underwent transplantation in relapse or with primary refractory disease. Despite this, 28% of children who underwent transplantation in relapse and 17% of children who underwent transplantation with primary induction failure were alive 5 years after transplantation, suggesting that transplantation can cure at least some children with the most resistant disease. It is perhaps surprising that length of first complete remission did not predict outcome in our study. This may be a reflection of the demographics of the patients, the majority of whom had experienced relapse early, with first complete remission of less than 12 months, limiting statistical power to look at this risk factor.
Children receiving transplantation in remission clearly had a superior outcome to those who underwent transplantation in relapse. It is commonly debated whether it is preferable to perform transplantation in children identified in early relapse immediately or to pursue reinduction chemotherapy and attempt to achieve a second remission before performing transplantation. Although our data show better disease control in children who underwent transplantation in remission, these patients had chemotherapy-sensitive disease and would, therefore, be expected to have better outcomes. Importantly, 76% of the patients who underwent transplantation in relapse had received chemotherapy but did not achieve remission; however, these patients had a 5-year probability of overall survival of 28%. This result suggests that transplantation is worthwhile in this group of particularly difficult patients. Our data are unable to definitively answer the question of the efficacy of immediate transplantation without an attempt at reinduction because there were only 22 such patients in this study. However, the 5-year leukemia-free survival rate was 25% in this group, which is similar to the rate in the group for whom reinduction was attempted. Most of the 22 patients reported good performance scores (90 to 100) despite a high tumor burden; nine patients had peripheral blasts, 10 patients had marrow blast counts of more than 10%, and three patients had marrow blast counts of 5% to 10%. We did not observe differences in leukemia-free survival rates after transplantation for patients in first relapse and second relapse, but there were only 24 patients in the latter group, and our inability to observe differences may be explained by the relatively small number of patients (5-year leukemia-free survival rates of 19% and 22%, respectively).
The importance of adverse cytogenetics was challenging to assess in this group. Our analysis failed to show a significant effect of intermediate- or poor-risk cytogenetics on leukemia recurrence, leukemia-free survival, or overall survival. This may be explained by the fact that patients with recurrent leukemia have high-risk disease, and consequently, the relevance of cytogenetics is limited by the relatively small sample size of approximately 260 patients. Data on cytogenetics were not available for approximately 23% of patients, which is a limitation that occurs when using data reported to an observational database and when transplantations are performed over a 10-year period because cytogenetic testing was not routinely performed during the early years. We adjusted for this limitation by stratifying all analysis of risk factors for transplantation outcome by cytogenetic risk group given the prognostic importance of cytogenetics for this disease.
Almost one third of grafts in this study were T-cell depleted. Although T-cell depletion reduced acute GVHD rates, treatment-related mortality was unchanged, as were leukemia-free and overall survival and relapse, indicating a neutral effect of T-cell depletion on overall outcome. As reported by others, we did not observe lower chronic GVHD rates after T-cell–depleted transplantations.21 Age and WBC count at diagnosis were not associated with transplantation outcome after a first relapse. This is similar to observations by Webb et al11 on outcome for children with relapsed AML after treatment on the Medical Research Council AML 10 trial at diagnosis. We did not observe a graft-versus-leukemia effect in our cohort. This may be explained by the inclusion of patients who received T-cell–depleted grafts (38% of patients) and patients with high tumor burden (47% of patients underwent transplantation in relapse or primary induction failure).
This study represents the largest series of children receiving unrelated donor bone marrow transplantation for AML currently in the literature. The strengths of the study are its large size and high-quality audited data. The limitations of the study are its retrospective nature, the heterogeneity inevitable in registry studies describing aggregate outcomes of transplantations performed at multiple centers, and our inability to compare transplantation outcomes to those after chemotherapy alone in a similar group of patients. We did not observe a significant correlation between year of transplantation, HLA mismatch, and survival, and this may be explained by the relatively few patients who received allele-matched bone marrow grafts in this report. Larger studies in unrelated donor bone marrow transplantation clearly demonstrate the negative effect of HLA mismatch on survival, and matching between donor and recipient using allele-level typing at HLA-A, HLA-B, HLA-C, and DRB1 represents the current standard of care.22 Despite these limitations, these data indicate that unrelated donor bone marrow transplantation is an effective therapy for a significant proportion of children with recurrent or refractory AML who are unlikely to be cured with chemotherapy alone.
The author(s) indicated no potential conflicts of interest.
Conception and design: Nancy J. Bunin, Stella M. Davies, Richard Aplenc, Bruce M. Camitta, Kenneth B. DeSantes, Rakesh K. Goyal, Neena Kapoor, Nancy A. Kernan, Joseph Rosenthal, Franklin O. Smith, Mary Eapen
Collection and assembly of data: Mary Eapen
Data analysis and interpretation: Nancy J. Bunin, Stella M. Davies, Richard Aplenc, Bruce M. Camitta, Kenneth B. DeSantes, Rakesh K. Goyal, Neena Kapoor, Nancy A. Kernan, Joseph Rosenthal, Franklin O. Smith, Mary Eapen
Manuscript writing: Nancy J. Bunin, Stella M. Davies, Bruce M. Camitta, Franklin O. Smith, Mary Eapen
Final approval of manuscript: Nancy J. Bunin, Stella M. Davies, Richard Aplenc, Bruce M. Camitta, Kenneth B. DeSantes, Rakesh K. Goyal, Neena Kapoor, Nancy A. Kernan, Joseph Rosenthal, Franklin O. Smith, Mary Eapen
Supported by Public Health Service Grant No. U24-CA76518 from the National Cancer Institute, the National Institute of Allergy and Infectious Diseases, and the National Heart, Lung and Blood Institute; Office of Naval Research; Health Resources and Services Administration (Department of Health and Human Services); and grants from AABB; Aetna; American International Group, Inc; American Society for Blood and Marrow Transplantation; Amgen, Inc; Anonymous donation to the Medical College of Wisconsin; Astellas Pharma US, Inc; Baxter International, Inc; Bayer HealthCare Pharmaceuticals; BioOne Corporation; BloodCenter of Wisconsin; Blue Cross and Blue Shield Association; Bone Marrow Foundation; Bristol-Myers Squibb Company; Cangene Corporation; Celgene Corporation; CellGenix, GmbH; Cerus Corporation; Cubist Pharmaceuticals; Cylex Inc; CytoTherm; DOR BioPharma, Inc; Dynal Biotech, an Invitrogen Company; EKR Therapeutics; Enzon Pharmaceuticals, Inc; Gambro BCT, Inc; Gamida Cell, Ltd; Genzyme Corporation; Gift of Life Bone Marrow Foundation; GlaxoSmithKline, Inc.; Histogenetics, Inc; HKS Medical Information Systems; Hospira, Inc; Infectious Diseases Society of America; Kiadis Pharma; Kirin Brewery Co, Ltd; Merck & Company; The Medical College of Wisconsin; MGI Pharma, Inc; Millennium Pharmaceuticals, Inc; Miller Pharmacal Group; Milliman USA, Inc; Miltenyi Biotec, Inc; MultiPlan, Inc; National Marrow Donor Program; Nature Publishing Group; Oncology Nursing Society; Osiris Therapeutics, Inc; Pall Life Sciences; PDL BioPharma, Inc; Pfizer Inc; Pharmion Corporation; Roche Laboratories; Schering Plough Corporation; Society for Healthcare Epidemiology of America; StemCyte, Inc; StemSoft Software, Inc; SuperGen, Inc; Sysmex; Teva Pharmaceutical Industries; The Marrow Foundation; THERAKOS, Inc; University of Colorado Cord Blood Bank; ViaCell, Inc; Vidacare Corporation; ViraCor Laboratories; ViroPharma, Inc; and Wellpoint, Inc.
The views expressed in this article do not reflect the official policy or position of the National Institute of Health, the Department of the Navy, the Department of Defense, or any other agency of the US Government.
Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.