Hematology and Oncology, The Cleveland Clinic, Cleveland, Ohio; 3 Department of Biostatistics, The Ohio State University Hospitals, Columbus, Ohio

Transcription

1 Biology of Blood and Marrow Transplantation 12:61-67 (2006) 2006 American Society for Blood and Marrow Transplantation /06/ $32.00/0 doi: /j.bbmt High Disease Burden Is Associated with Poor Outcomes for Patients with Acute Myeloid Leukemia Not in Remission Who Undergo Unrelated Donor Cell Transplantation William Blum, 1 Brian J. Bolwell, 2 Gary Phillips, 3 Sherif S. Farag, 1 Thomas S. Lin, 1 Belinda R. Avalos, 1 Sam L. Penza, 1 Guido Marcucci, 1 John C. Byrd, 1 Matt E. Kalaycio, 2 Ronald M. Sobecks, 2 Brad Pohlman, 2 Stacey Brown, 2 Patrick J. Elder, 1 Edward A. Copelan 1 1 Division of Hematology and Oncology, The Ohio State University Hospitals, Columbus, Ohio; 2 Division of Hematology and Oncology, The Cleveland Clinic, Cleveland, Ohio; 3 Department of Biostatistics, The Ohio State University Hospitals, Columbus, Ohio Correspondence and reprint requests: William Blum, MD, Division of Hematology and Oncology and Comprehensive Cancer Center, The Ohio State University, B310 Starling-Loving Hall, 320 W. 10th Ave., Columbus, OH ( blum-1@medctr.osu.edu). Received January 12, 2005; accepted June 27, 2005 ABSTRACT Results were analyzed for 48 consecutive patients with acute myeloid leukemia not in remission who underwent unrelated donor bone marrow or stem cell transplantation between 1991 and February 2003 at 2 transplant centers. Forty-six were adults with a median age of 32 years (range, 4-58 years). Forty-two were HLA-A, -B, and -DR matched with their respective donors, and 6 were mismatched at 1 of these loci. The conditioning regimen was myeloablative in all cases: busulfan/cyclophosphamide/etoposide in 34 patients, busulfan/cyclophosphamide in 10 patients, and total body irradiation based in 4 patients. Median follow-up for survivors was 540 days (range, days). Only patients with <5000 peripheral blood blasts per microliter at the time of transplantation survived 2 years (18% versus 0%; P.003). Similarly, patients with <20% blasts in the marrow at the time of transplantation had superior 2-year survival compared with those who had >20% (33% versus 5%; P.04). Patients with <20% blasts who had >3 prior therapies also fared poorly. Cause of death was more commonly treatment related rather than relapse related. This study confirms that patients with acute myeloid leukemia not in remission can achieve prolonged survival with myeloablative conditioning and unrelated donor cell transplantation. However, sustained survival occurs only in patients with a low disease burden at the time of unrelated donor stem cell transplantation, and patients with a high disease burden may benefit from added counseling regarding the high risk of death due to both treatment-related toxicities and disease relapse American Society for Blood and Marrow Transplantation KEY WORDS AML Unrelated donor cell transplantation Disease burden INTRODUCTION Allogeneic transplantation of patients with acute myeloid leukemia (AML) not in remission has curative potential but is associated with high rates of relapse and transplant-related mortality [1,2]. Patients with advanced disease fare poorly compared with those in first complete remission (CR) [1-8]. The Société Francaise de Greffe de Moelle (SFGM) reported outcomes of 379 patients with AML (including only 28 with unrelated donors) who underwent allogeneic transplantation [4]. Five-year overall survival (OS) was 11% to 14% for patients with primary refractory disease (n 69), untreated relapse (n 94), or refractory relapse (n 67), compared with 35% for patients in CR. Similarly, The City of Hope reported a 3-year OS of 30% for patients with primary refractory AML (n 68) [9]. Variables that influenced survival outcome were identified by these 2 studies. In the SFGM report [4], age 15 years, achievement of CR before transplantation, female donor, low-grade or no acute 61

2 W. Blum et al. graft-versus-host disease (GVHD), and the presence of chronic GVHD were favorable prognostic factors. T-cell depletion (TCD) was an adverse factor. In the City of Hope study of refractory AML [9], only an unrelated donor and unfavorable cytogenetics were deemed adverse risk factors. Transplantation from unrelated donors (UDT) for patients with AML not in remission has also been investigated by the Seattle group, which showed a 5-year survival of 7% and 19% for patients who underwent transplantation in relapse (n 81) or with primary refractory disease (n 16), respectively [5]. More recently, a German study reported 16% leukemia-free survival (LFS) at 3 years for patients with AML not in remission who underwent UDT (n 63) [8]. Given the poor long-term survival results after transplantation for patients with AML not in remission and the recognition that transplantation from an unrelated donor further negatively affects outcomes compared with transplantation from a matched sibling donor, it is important to identify prognostic factors to aid clinical decision making about the risk of UDT for patients with AML not in remission who are eligible for the procedure. This investigation analyzed the effect of disease burden on survival for patients with AML not in remission who received UDT. METHODS Patient Identification A search of institutional bone marrow (BM)/stem cell transplantation databases of patients who underwent any type of transplantation at the James Cancer Hospital or the Cleveland Clinic Foundation between 1991 and February 2003 identified 48 patients with AML not in remission who underwent unrelated donor BM (n 46) or filgrastim-mobilized peripheral blood (PB) stem cell (n 2) transplantation. All patients had morphologic evidence of AML at the time of transplantation. Pretransplantation clinical data and survival data were available for all patients identified. Disease burden was measured by both PB and marrow blasts. PB measurements to determine disease burden were from the day of admission for transplantation. BM measurements were within a month before transplantation (and after any prior salvage chemotherapy regimen). Data on the use of cytoreductive medications such as hydroxyurea after marrow measurement and before admission for transplantation were not available. This retrospective analysis was approved by the institutional human studies and privacy committees. HLA Typing and GVHD/Infection Prophylaxis Thirty-one patients had high-resolution DNA typing for HLA-DRB1 by polymerase chain reaction/ sequence-specific priming. Of these, HLA-A and -B typing was performed by high-resolution DNA typing (n 3), low- or intermediate-resolution DNA typing (n 15), or serologic typing (n 13). Sixteen patients had intermediate-resolution typing for DRB1 by polymerase chain reaction/sequence-specific oligonucleotide priming. Of these, HLA-A and -B typing was performed by low- or intermediate-resolution DNA typing (n 8) or serologic typing (n 8). One patient had serologic typing for HLA-A, -B, and -DR. HLA-C testing was performed in 38 patients by highresolution DNA typing (n 6), low- or intermediateresolution DNA typing (n 24), or serologic typing (n 8). For 20 patients with TCD of grafts, selective CD8 lymphocyte depletion was performed from the mononuclear fraction of the stem cell product with Dynabeads M-450 CD8 (Dynal, Oslo, Norway); cells and beads were separated with the MaxStep Magnetic Cell Separator (Baxter, Deerfield, IL). Outcomes for patients who had TCD in this study relative to the effectiveness of TCD and effect of CD8 lymphocyte levels on GVHD have been previously reported [10]. GVHD prophylaxis included tacrolimus and methotrexate in 18 patients, cyclosporine and methotrexate in 27 patients, and cyclosporine alone in 3 patients. Infectious prophylaxis included trimethoprim/sulfamethoxazole (or substitute), antiviral prophylaxis, and amphotericin or fluconazole in all patients. Twenty-four patients also received antibacterial prophylaxis with penicillin VK plus ciprofloxacin or ciprofloxacin alone. Nine patients received prophylactic intravenous immune globulin after transplantation. Engraftment/GVHD Grading The date of myeloid engraftment was defined as the first of 3 consecutive days with an absolute neutrophil count 500/ L. The date of platelet engraftment was the first of 3 consecutive days with a platelet count / L with no platelet transfusions within the prior 7 days. Graft failure was defined as the absence of myeloid engraftment by day 28 after transplantation. Acute and chronic GVHD were defined and graded according to previously published criteria [11,12]. Statistical Analysis Univariate analysis was used to assess the effect on survival of each of the following variables: PB blasts 5000/ L, BM blasts 20%, patient age, no PB blasts, level of HLA-A/-B/-DRB1 matching (intermediate/high versus serologic/low), HLA-C mismatching, 3 prior treatment regimens, total nucleated cell (TNC) dose of transplanted cells per kilogram, acute GVHD, chronic GVHD, duration of first CR, TCD, and cytogenetic risk group. Criteria for PB blasts and 62

3 AML and Unrelated Donor Cell Transplantation BM blasts as noted were chosen for clinical relevancy in a heterogeneous population with AML not in remission. A multivariate Cox proportional hazard model was developed for both OS and LFS by using PB blasts 5000/ L, BM blasts 20%, 3 prior treatment regimens, and patient age. For descriptive purposes, we refer to PB blasts 5000/ L or BM blasts 20% as markers of high disease burden; these criteria were selected on the basis of clinical relevance. The goal of this study was to find the true relationship between OS and disease burden based on these parameters. For this purpose, variables were identified that were confounders or effect modifiers. Variables were added to the logistic regression model and assessed as to whether or not they changed the disease burden hazard ratio by more than 15% in either direction, regardless of the variable s statistical significance. Once a group of variables were established that affected the disease burden s hazard ratio, variables were then assessed as effect modifiers with the risk factor. Continuous variables were tested for linearity by using the method of fractional polynomials, although markers of disease burden were dichotomized as noted. The proportional hazard assumption that the hazard of not surviving was constant across time in the final model was also tested. RESULTS Patient Characteristics Forty-eight patients were identified; 46 were adults. Forty-two patients were HLA-A, -B, and -DR matched with their donors, and 6 were mismatched at 1 of these loci. Eight patients were mismatched at HLA-C (of 38 who had HLA-C typing performed). Eight patients had AML evolved from myelodysplastic syndrome. All patients received myeloablative conditioning regimens. Thirty-four patients were treated with busulfan 14 mg/kg, cyclophosphamide 120 mg/ kg, and etoposide 50 mg/kg; 10 were treated with busulfan 16 mg/kg and cyclophosphamide 120 mg/kg; and 4 were treated with total body irradiation based regimens (1200 Gy in 6 fractions). Forty-six patients received BM grafts, and 2 received filgrastim-mobilized PB stem cells. Twenty patients received CD8 lymphocyte-depleted grafts. Patient characteristics are listed in Table 1. All patients had AML not in remission at the time of transplantation; 17 patients were in first relapse, 16 were in or beyond the second relapse, and 14 had primary refractory disease; 1 patient whose disease had transformed from myelodysplastic syndrome was untreated. Karyotypes of 35 patients were available. According to Cancer and Leukemia Group B criteria [13], patients were assigned the following cytogenetic risk groups: favorable risk, n 4; intermediate risk, n 20; and adverse risk, n 11. Nine patients received salvage chemotherapy within 2 months before undergoing transplantation. Engraftment Thirty-seven patients were evaluable for engraftment. The median time to neutrophil engraftment was 20 days (range, 9-60 days). Primary graft failure occurred in 1 patient. Seven patients were not evaluable for engraftment because they died before day 28 (infection, n 5; hepatic veno-occlusive disease, n 2). Three patients were not evaluable for engraftment because no data were available; these 3 patients died on day 30, day 32, or day 53 as a result of venoocclusive disease, acute respiratory distress syndrome, or infection, respectively. The median time to platelet engraftment was 32 days (range, days; n 22). Acute and Chronic GVHD Acute GVHD occurred in 58% of patients (23 of 40 patients). The incidence of grade I/II GVHD was 35%, and the incidence of grade III/IV GVHD was 20%. Of 25 patients who survived at least 100 days, chronic GVHD data were available for 22 patients. Limited or extensive chronic GVHD occurred in 64% of patients (14 of 22 patients). Causes of Death Treatment-related mortality was the primary cause of death, occurring in 26 (54%) of 48 patients, with 1 case of graft failure. Death due to relapse occurred in 15 (31%) of 48 patients. For 28 patients who received unmanipulated grafts, causes of death were relapse (n 7), GVHD (n 4), infection (n 9), organ toxicity (n 3), and hemorrhage (n 1). For 20 patients who received CD8 lymphocyte-depleted grafts, causes of death were relapse (n 8), GVHD (n 2), infection (n 2), organ toxicity (n 4), and graft failure (n 1). Survival Analysis Figure 1 shows the Kaplan-Meier survival curves for OS based on PB blasts 5000/ L, BM blasts 20%, the presence of any circulating blasts, and the number of prior treatments. Kaplan-Meier survival curves were tested for equality by using a log-rank test. OS for patients with PB blasts 5000/ L was superior to OS in those with higher counts (P.0034). OS for patients with BM blasts 20% was superior to OS in those with higher marrow blasts (P.0356). OS for patients with 3 prior treatments was superior to OS in more heavily pretreated patients (P.0027). OS based on the presence or absence of circulating blasts was not significantly different (P.23). LFS was significantly different between the groups based on PB blasts 5000/ L or BM blasts 63

4 W. Blum et al. Table 1. Characteristics of Patients With AML Not in Remission Who Received UDT Patient No. Age (y) Duration of First CR No. Prior Regimens Disease Status at Transplantation PB Blasts >5000/ L? BM Blasts >20%? TNC of Transplanted Cells ( 10 8 /kg) Overall Survival (d) Alive? Cause of Death REL2 No No Y REL1 No No Y REL2 No No Y 5433* REL2 No Yes Y 4057* 20 2 P-REF No No N GVHD 5119* 31 1 P-REF No Yes N Relapse 5609* REL1 No No Y REL2 No Yes Y REL2 No Yes N Relapse P-REF No No Y 5741* REL1 No No N GVHD 5101* 28 0 Untreated No Yes N Relapse REL2 No Yes N Relapse REL1 No Yes N Bleeding 4062* 37 3 REL1 No No N GVHD REL2 No No N GVHD REL1 No No N Relapse 2253* REL1 No Yes N Relapse 5861* 50 2 P-REFC No No N Relapse REL2 No N Relapse P-REFC Yes Yes N Relapse 5302* P-REFC No Yes N Organ toxicity REL1 No Yes N GVHD P-REF No Yes N Organ toxicity REL1 No Yes N Infection 1788* 19 NA REL1 No Yes N Relapse REL2 No No N Infection REL1 No Yes N GVHD P-REF No No N Infection 5681* REL3 No Yes N Relapse P-REF No Yes N Relapse 5684* REL3 Yes Yes N Relapse 3059* 26 3 REL1 No Yes N Infection 5545* 34 3 REL2 Yes No N Organ toxicity REL2 No No N Infection 5212* REL2 No Yes N Relapse REL2 No No N Relapse REL1 No No N Organ toxicity NA REL1 Yes Yes N Organ toxicity 5807* P-REFC No No N Graft failure P-REFC No Yes N Infection 5874* REL2 No Yes N Organ toxicity REL1 No Yes N Infection P-REF Yes Yes N Infection 2175* 25 NA P-REF No Yes N Organ toxicity 5107* REL1 No Yes N Infection REL No Yes NA 14 N Infection P-REF Yes Yes N Infection CR indicates complete remission; NA, not available; REL, relapsed AML; P-REF, primary refractory AML; P-REFC, primary refractory AML but subsequent remission and then relapse; TNC, total nucleated cell dose. *T cell depleted transplantation. 20% (P.0013 and P.0448, respectively; figure not shown). In the subset of patients (n 21) who had intermediate- or high-level HLA typing for HLA-A/ -B/-DRB1, PB blasts 5000/ L remained a predictor of OS outcomes (P.0232), but BM blasts 20% did not influence outcome (P.5771). Univariate analyses of the other independent variables, including the TNC dose of transplanted cells per kilogram, mismatching at 1 HLA-A/-B/-DRB1 locus, mismatching at HLA-C, acute GVHD, chronic GVHD, duration of first CR, CD8 TCD, and cytogenetic risk demonstrated that none statistically affected survival. Multivariable Cox proportional hazard models were developed for both OS (Table 2) and LFS (Table 3) and consisted of 4 variables: PB blasts 5000/ L, BM blasts 20%, 3 prior treatment regimens, and age. The continuous variable (age) was linear in the 64

5 AML and Unrelated Donor Cell Transplantation Figure 1. Overall survival was dichotomized in 4 different ways. A, Dichotomization is based on a cutoff of circulating blasts 5000/ L. B, Dichotomization is based on 20% bone marrow blasts. C, Dichotomization is based on whether the subject had any circulating blasts. D, Dichotomization is based on whether the patient had 2 prior treatments. All 4 Kaplan-Meier survival curves were tested for equality by using a log-rank test. The P values are shown on each graph. log hazard with fractional polynomials. The models met the proportional hazard assumption. For OS, the hazard ratio for PB blasts showed that the hazard of death was 2.89 times higher for PB blasts 5000/ L as compared with PB blasts 5000/ L (P.053; 95% confidence interval [CI], ). The hazard of death based on BM blasts was affected by the number of prior treatments. For patients with BM blasts 20% and 3 prior treatments, the hazard ratio was 7.31 for patients compared with those who had 20% blasts and 3 treatments (P.001; 95% CI, ). For patients with BM blasts 20%, there was no difference in the hazard of death according to prior treatments. Among patients who received 3 prior treatments, the hazard ratio was 3.91 for patients with BM blasts 20% compared with those with BM blasts 20% (P.014; 95% CI, ). No differences in survival based on PB or BM blasts were detected among patients who received 3 prior treatments (P.629). For LFS, the hazard ratio was 3.51 for PB blasts 5000/ L (P.024; 95% CI, ). For patients with BM blasts 20% and 3 prior treatments, the hazard ratio was 6.92 compared with those with BM blasts 20% and 3 treatments (P.001; 95% CI, ). Among patients who received 3 prior treatments, the hazard ratio for patients with BM blasts 20% was 3.43 compared with those with BM blasts 20% (P.019; 95% CI, ). DISCUSSION The concept of adverse outcomes for patients who enter transplantation with residual disease as compared with those who enter transplantation in remission is intuitive and has been reported by others [5,14]. Patients with AML not in remission and no available related donor have limited therapeutic options, including conventional UDT, investigational transplantation trials, additional chemotherapy, or supportive care. Only UDT offers a hope for cure, but this must be balanced against a high risk of transplantrelated mortality, morbidity, and relapse for those who survive the immediate posttransplantation period. This investigation examined whether clinically relevant measures of disease burden at the time of UDT would have prognostic value for predicting which patients within the subset of AML patients not in remission might have even more dismal outcomes. These patients may be encouraged to participate in other therapies, such as clinical trials that use novel methods of conditioning/transplantation or posttrans- Table 2. Cox Proportional Hazard Model for Overall Survival 95% Confidence Interval Predictor Variable n Hazard Ratio Lower Upper P Value Circulating blasts > 5000/ L* Bone marrow blasts <20% Prior Trt (<2) Prior Trt (>3) Bone marrow blasts > 20% Prior Trt (<2) Prior Trt (>3) Prior Trt (<2) BM blasts <20% BM blasts >20% Prior Trt (>3) BM blasts <20% BM blasts >20% Age Trt indicates treatment. *Referent group is circulating blasts 5000/ L (n 42). 65

6 W. Blum et al. Table 3. Cox Proportional Hazard Model for Leukemia-Free Survival 95% Confidence Interval Predictor Variable n Hazard Ratio Lower Upper P Value Circulating blasts >5000/ L* Bone marrow blasts <20% Prior Trt (<2) Prior Trt (>3) Bone marrow blasts >20% Prior Trt (<2) Prior Trt (>3) Prior Trt (<2) BM blasts <20% BM blasts >20% Prior Trt (>3) BM blasts <20% BM blasts >20% Age Trt indicates treatment. *Referent group is circulating blasts 5000/ L (n 42). plantation therapies. This prognostic information may affect a patient s choice to pursue potentially curative but highly toxic UDT. Our results showed that a high disease burden adversely affected OS and LFS for this poor-risk cohort of patients, with the high disease burden based on clinically relevant parameters of PB blasts 5000/ L and BM blasts 20%. The number of prior treatments was also an important predictive factor for survival in patients with lower disease burden. Patients with AML not in remission and their physicians should consider the especially low survival of the patients with a high disease burden in this study when discussing treatment options such as conventional UDT, because investigational transplantation trials may be attractive alternatives. The predictive value of end points such as pretransplantation circulating PB blasts of 5000/ L, BM blasts 20%, and the number of prior treatments was examined in this cohort of 48 patients. No patient with PB blasts 5000/ L survived 1 year. Likewise, patients with BM blasts 20% had poor survival. Both of these variables are indicators of high disease burden. Survival for patients with BM blasts 20% was influenced by the number of prior treatments, and the presence of BM blasts 20% did not seem to be as important when patients with outmoded HLA typing were excluded from the analysis. No other factor (except age) including GVHD, TCD, or karyotype significantly influenced survival in this study. The results from this large cohort of patients with AML not in remission confirm that such patients can achieve long-term LFS with UDT [4,5,9,15,16]. Sierra et al. [5], summarizing the Seattle experience, reported that conventional UDT for patients with AML not in remission had a higher risk of relapse than transplantation in CR. In this subset of patients, it was noted that patients with 30% leukemic blasts in the marrow and no circulating blasts before transplantation had a lower risk of relapse than those with 30% blasts in the marrow, circulating blasts, or both (P.02). Our results were similar to these findings, although they show that a circulating PB blast count 5000/ L may be a more clinically relevant end point than the presence of no circulating blasts, because a blast count 5000/ L was a predictive factor even in the smaller subset of patients who had more sophisticated HLA typing performed. Disease burden has been reported to affect outcomes for AML patients who are treated with reduced-intensity conditioning (RIC) as well. Maris et al. [17] reported inferior survival for patients with 5% marrow blasts at the time of transplantation for a mixed cohort of patients (n 89) treated with RIC and UDT. Sayer et al. [18] reported that AML patients with 20% marrow blasts or 5% to 20% marrow blasts had inferior survival after RIC and allogeneic transplantation compared with patients who had 5% marrow blasts (event-free survival of 14%, 24%, and 49%, respectively; P.001). The dose of transplanted cells has previously been shown to affect outcome in patients undergoing allogeneic transplantation [5,19-21]. For patients with AML not in remission undergoing conventional UDT in 1 study [5], those with a transplanted cell dose above the median ( nucleated cells per kilogram) had a 5-year LFS of 13%, versus 2% for those who received cell doses below the median (P.20). By shortening the time to count recovery and potentially increasing a graft-versus-leukemia effect, it is logical that a higher transplanted cell dose may be especially important for patients with advanced disease, who are at the highest risk for relapse [21]. However, the transplanted cell dose did not significantly affect outcome in the present cohort of patients. Furthermore, TCD, which typically leads to a lower TNC dose, did not adversely affect survival, in contrast to the SFGM study [4], which reported an 66

7 AML and Unrelated Donor Cell Transplantation adverse effect of TCD (and included patients in CR and predominantly sibling transplants). This difference may be due to variation in patients or donors or to the use of selective CD8 TCD in the current study. CD8 TCD, as in this study, has been previously reported to deplete the CD8 lymphocyte dose by 2 logs. A higher CD8 cell dose was associated with more severe acute and chronic GVHD in these patients [10]. This study of 48 patients with AML not in remission who underwent myeloablative UDT demonstrates that consideration of objective measures of disease burden and the number of prior treatments may be helpful in the identification of patients who have an exceedingly low chance of a favorable outcome. For heavily pretreated patients or those with a high disease burden, especially those with PB blasts 5000/ L, counseling regarding the high risk of death from not only relapse, but also, more commonly, from treatment-related complications should be provided. Discussion of alternatives, including investigational transplantation trials or other therapies, would add important information for such patients. REFERENCES 1. Biggs JC, Horowitz MM, Gale RP, et al. Bone marrow transplants may cure patients with acute leukemia never achieving remission with chemotherapy. Blood. 1992;80: Mehta J, Powles R, Horton C, et al. Bone marrow transplantation for primary refractory acute leukaemia. Bone Marrow Transplant. 1994;14: Geller RB, Saral R, Piantadosi S, et al. Allogeneic bone marrow transplantation after high-dose busulfan and cyclophosphamide in patients with acute nonlymphocytic leukemia. Blood. 1989; 73: Michallet M, Thomas X, Vernant JP, et al. Long-term outcome after allogeneic hematopoietic stem cell transplantation for advanced stage acute myeloblastic leukemia: a retrospective study of 379 patients reported to the Societe Francaise de Greffe de Moelle (SFGM). Bone Marrow Transplant. 2000;26: Sierra J, Storer B, Hansen JA, et al. Unrelated donor marrow transplantation for acute myeloid leukemia: an update of the Seattle experience. Bone Marrow Transplant. 2000;26: Hallemeier C, Girgis M, Blum W, et al. Outcomes of adults with acute myelogenous leukemia in remission given 550 cgy of single-exposure total body irradiation, cyclophosphamide, and unrelated donor bone marrow transplants. Biol Blood Marrow Transplant. 2004;10: Copelan EA, Biggs JC, Thompson JM, et al. Treatment for acute myelocytic leukemia with allogeneic bone marrow transplantation following preparation with BuCy2. Blood. 1991;78: Basara N, Baurmann H, Kolbe K, et al. Antithymocyte globulin for the prevention of graft-versus-host disease after unrelated hematopoietic stem cell transplantation for acute myeloid leukemia: results from the multicenter German cooperative study group. Bone Marrow Transplant. 2005;35: Fung HC, Stein A, Slovak M, et al. A long-term follow-up report on allogeneic stem cell transplantation for patients with primary refractory acute myelogenous leukemia: impact of cytogenetic characteristics on transplantation outcome. Biol Blood Marrow Transplant. 2003;9: Kalaycio M, Rybicki L, Pohlman B, et al. CD8 T-cell-depleted, matched unrelated donor, allogeneic bone marrow transplantation for advanced AML using busulfan-based preparative regimens. Bone Marrow Transplant. 2005;35: Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation. 1974;18: Shulman HM, Sale GE, Lerner KG, et al. Chronic cutaneous graft-versus-host disease in man. Am J Pathol. 1978;91: Byrd JC, Mrozek K, Dodge RK, et al. Pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid leukemia: results from Cancer and Leukemia Group B (CALGB 8461). Blood. 2002;100: Girgis M, Hallemeier C, Blum W, et al. Chimerism and clinical outcomes of 110 unrelated donor bone marrow transplant recipients conditioned with low dose (550 cgy), single exposure total body irradiation and cyclophosphamide. Blood. 2004;105: Clift RA, Buckner CD, Appelbaum FR, et al. Allogeneic marrow transplantation during untreated first relapse of acute myeloid leukemia. J Clin Oncol. 1992;10: Godder KT, Mehta J, Chiang KY, et al. Partially mismatched related donor bone marrow transplantation as salvage for patients with AML who failed autologous stem cell transplant. Bone Marrow Transplant. 2001;28: Maris MB, Niederwieser D, Sandmaier BM, et al. HLAmatched unrelated donor hematopoietic cell transplantation after nonmyeloablative conditioning for patients with hematologic malignancies. Blood. 2003;102: Sayer HG, Kroger M, Beyer J, et al. Reduced intensity conditioning for allogeneic hematopoietic stem cell transplantation in patients with acute myeloid leukemia: disease status by marrow blasts is the strongest prognostic factor. Bone Marrow Transplant. 2003;31: Mehta J, Powles R, Treleaven J, Kulkarni S, Horton C, Singhal S. Number of nucleated cells infused during allogeneic and autologous bone marrow transplantation: an important modifiable factor influencing outcome. Blood. 1997;90: Singhal S, Powles R, Treleaven J, et al. A low CD34 cell dose results in higher mortality and poorer survival after blood or marrow stem cell transplantation from HLA-identical siblings: should CD34 cells/kg be considered the minimum threshold? Bone Marrow Transplant. 2000;26: Dominietto A, Lamparelli T, Raiola AM, et al. Transplantrelated mortality and long-term graft function are significantly influenced by cell dose in patients undergoing allogeneic marrow transplantation. Blood. 2002;100: