Fludarabine-Based Conditioning Secures Engraftment of Second Hematopoietic Stem Cell Allografts (HSCT) in the Treatment of Initial Graft Failure

Transcription

1 Biology of Blood and Marrow Transplantation 13: (2007) 2007 American Society for Blood and Marrow Transplantation /07/ $32.00/0 doi: /j.bbmt Fludarabine-Based Conditioning Secures Engraftment of Second Hematopoietic Stem Cell Allografts (HSCT) in the Treatment of Initial Graft Failure Joseph H. Chewning, Hugo Castro-Malaspina, Ann Jakubowski, Nancy A. Kernan, Esperanza B. Papadopoulos, Trudy N. Small, Glenn Heller, Katharine C. Hsu, Miguel A. Perales, Marcel R.M. van den Brink, James W. Young, Susan E. Prockop, Nancy H. Collins, Richard J. O Reilly, Farid Boulad Bone Marrow Transplant Service, Memorial Sloan-Kettering Cancer Center, New York, New York Correspondence and reprint requests: Farid Boulad, MD, Bone Marrow Transplant Service, Department of Pediatrics, Memorial Sloan-Kettering Cancer Center, Box 517, 1275 York Avenue, New York, NY ( bouladf@mskcc.org). Received July 8, 2007; accepted July 10, 2007 ABSTRACT Graft failure is associated with a high mortality rate. To date, regimens invoked for second transplants have resulted in inconsistent engraftment with high transplant-related mortality (TRM). We here report 16 consecutive patients, aged 4-59 years, who received second HSCT (HSCT-2) at a median of 45 days following primary or secondary failure of an initial unmodified (N 3) or T cell-depleted (TCD) (N 13) HSCT (HSCT-1). HSCT-1 was administered after myeloablative total body irradiation (TBI)- or alkylator-based conditioning for acute leukemias (N 7), MDS (N 6), CML (N 2), and Fanconi anemia (N 1). All patients experienced 1 or more infectious complications between HSCT-1 and HSCT-2, and 10 patients had active infections at the time of HSCT-2. Cytoreduction regimens used for HSCT-2 included fludarabine (Flu) in combination with cyclophosphamide (CTX) (N 9), or thiotepa (Thio) (N 5). In addition, 1 patient received Flu alone and 1 patient Thio combined with CTX. Antithymocyte globulin (ATG) (N 11) or Alemtuzumab (N 3) was added pretransplant to prevent rejection. For HSCT-2, donors included HLA-matched (N 3) or mismatched (N 8) related, or matched (N 2) or mismatched (N 3) unrelated donors. The primary graft donor was used in 6 of 16 cases. The grafts administered were unmodified peripheral blood stem cell transplantation (PBSCT) (N 5) or bone marrow transplantation (BMT) (N 3), TCD PBSCT (N 8). All patients achieved engraftment at a median of 12 days and evaluable patients achieved complete donor chimerism. Six patients are alive with a median follow-up of 49 months, including 4/9 conditioned with Flu/CTX. In this series, outcome was statistically superior for younger patients (<20 years). In summary, second HSCT using the combination of a fludarabine- and ATG-based, nonmyeloablative regimen and higher numbers of CD34 progenitor cells has been associated with acceptable toxicity and allowed consistent engraftment with hematopoietic reconstitution in patients with previous graft failure American Society for Blood and Marrow Transplantation KEY WORDS Hematopoietic stem cell transplant Graft failure T cell depletion Engraftment INTRODUCTION Hematopoietic stem cell transplant (HSCT) is an important therapeutic modality for the treatment of hematologic and immunologic disorders. Graft failure following SCT is a rare but often fatal complication, resulting in severe, prolonged pancytopenia and immune deficiency. Risk factors associated with graft failure include the type of donor, degree of HLAmismatch, cytoreduction regimen, and T cell-depleted (TCD) stem cell grafts [1,2]. Following graft failure, a second stem cell transplant represents the best chance of long-term, disease-free survival (DFS) for these patients. Achieving stable engraftment in 1313

2 1314 J. H. Chewning et al. patients who undergo a second transplant following graft failure is often difficult with engraftment rates as low as 33% reported in the literature [3]. Fludarabine is a purine antimetabolite, which was initially used for the treatment of acute myelogenous leukemia (AML) [4]. Over the last decade, based on the profound immune suppression associated with fludarabine, a number of cytoreductive regimens include this drug to facilitate engraftment, especially in the context of related HLA-mismatched transplants [5,6]. Fludarabine has also been used for second stem cell transplants for patients with relapsed leukemia [7]. In this retrospective study, we describe 16 consecutive patients who received a second HSCT following failure of an initial HSCT. We were able to achieve stem cell engraftment in all these high-risk patients using higher numbers of CD34 progenitor cells following an immunoablative chemotherapy regimen with or without antithymocyte globulin (ATG). PATIENTS AND METHODS Patient Characteristics Sixteen consecutive patients who suffered graft failure following a first HSCT (HSCT-1) performed between November 1997 and June 2005 were evaluated. All 16 patients received a second HSCT (HSCT-2) at this institution. Two additional patients experienced graft failure during this time period and were evaluated for a second HSCT. Unfortunately, both these patients expired while still being evaluated and prior to selection of donor or cytoreductive regimen. During this same time period, 673 allogeneic transplants were performed at our institution with a cumulative incidence of graft failure of 2.3%. The patient characteristics at the time of first transplant are summarized in Table 1. There were 11 males and 5 females identified. Median age of these patients at HSCT-1 was 22 years (range: 4-59 years), with 8 patients 20 years old or younger. Patient diagnoses included acute leukemias (N 7), myelodysplastic syndrome (MDS) (N 6), chronic myelogenous leukemia (CML) (N 2), and Fanconi anemia (N 1). At the time of the first transplant, 2 of the patients with acute leukemia had relapsed disease. Informed consent, including the discussion of the agents used for cytoreduction and their respective side effects, as well as the overall risks of a second HSCT, was obtained prior to second transplant from all patients. First Transplants Details of the first SCT administered to each patient and the cytoreduction used in each case are described in Table 2. To facilitate analysis of the results of their second transplants, the patients are grouped in this table according to the preparative regimen and stem cell manipulation used for the second transplant (Table 3). Of the 16 patients treated for graft failure, 3 had received unmodified marrow (N 2) or cord blood (N 1) transplants from unrelated donors, whereas 13 patients had received TCD grafts from related (N 5) or unrelated (N 8) donors. The donors for the 16 transplant patients were: HLA-matched related (N 3), HLA-matched unrelated (N 5), HLAmismatched related (N 2) and HLA-mismatched unrelated (N 6). Donor unique allele disparities that could be recognized by residual host T cells (ie, HLA mismatches predisposing to rejection) are specified in Table 2. Conditioning for the first transplants included myeloablative doses of either total body irradiation (TBI) or alkylating agents (busulfan melphalan) in each case. The regimens used were based on IRB-approved transplant protocols targeting specific patient groups. Recipients of unmodified marrow grafts (N 2) received GVHD prophylaxis consisting of cyclosporine or tacrolimus with short course methotrexate, whereas the cord blood graft recipient received cyclosporine, steroids, and ATG. Recipients of TCD grafts received either equine or rabbit ATG prior to transplant to prevent rejection, but did not receive any additional drug prophylaxis post transplant to prevent GVHD. Characteristics of Graft Failures following First Transplants Following the first transplant, patients were monitored with daily blood counts for evidence of engraftment. Time of engraftment was recorded as the first of 3 consecutive days in which neutrophil counts equaled or exceeded /L. Primary graft failure was defined as a failure to recover neutrophil counts by day 25 posttransplant coupled with persistence of marrow aplasia. Secondary graft failure was defined as graft failure occurring after initial partial or complete recovery of donor-type hematopoiesis, and was characterized by recurrent pancytopenia with neutrophil counts /L, and marrow aplasia. Both primary and secondary graft failures were confirmed in each case by cytogenetic and/or molecular demonstration of loss of donor-type blood elements. By these definitions, 11 patients in this series had primary graft failure, and 5 suffered secondary graft failure. In addition to prolonged pancytopenia, 10 patients with graft failure had 1 or more active issues entering HSCT-2. These included: CMV viremia/antigenemia (N 3), HHV-6 viremia (N 2), VRE bacteremia (N 1), Gram-negative sepsis/bacteremia (N 3), toxoplasmosis (N 1), bacterial abscess (N 1), pneumonitis (N 1), pancreatitis (N 1), severe enterocolitis (N 2), and hemorrhagic cystitis (N 1). The remaining 6 patients were stable at the time of

3 Fludarabine-Based Regimen for Graft Failure 1315 Table 1. Patient and Transplant Characteristics N 16 Age (years) Median 22 Range 4-59 Sex Male 11 Female 5 Diagnosis Acute leukemia 7 Myelodysplastic syndrome 6 Fanconi anemia/aplastic anemia 1 Chronic myelogenous leukemia 2 First transplant Donor Matched related 3 Mismatched related 2 Matched unrelated 5 Mismatched unrelated 6 Stem cell manipulation Unmodified 3 T cell depleted 13 Second transplant Donor Same as HSCT-1 6 Different 10 Donor Matching Matched related 3 Mismatched related 8 Matched unrelated 2 Mismatched unrelated 3 Stem cell manipulation Unmodified 8 T cell depleted 8 Cytoreduction Immunosuppression Flu (30 mg/m 2 /day) 5 days ATG 1 Flu (30 mg/m 2 /day) 5 days Cy (60 mg/kg/day) 2 days 2 Flu (30 mg/m 2 /day) 5 days Cy (60 mg/kg/day) 2 days ATG 7 Flu (25 mg/m 2 /day) 5 days Thio (5 mg/kg/day) 2 days ATG 3 Flu (25 mg/m 2 /day) 5 days Thio (5 mg/kg/day) 2 days Campath 2 Cy (60 mg/kg/day) 2 days Thio (5 mg/kg/day) 2 days Campath 1 N indicates number of patients; HSCT-1, first hematopoietic stem cell transplant; ATG, antithymocyte globulin; Flu, fludarabine; Cy, cyclophosphamide; Thio, thiotepa. HSCT-2, but had experienced the following complications in the interval following HSCT-1: bacteremia (N 3), CMV viremia/antigenemia (N 1), pneumonia (N 2), enterocolitis (N 1), and seizures (N 1). Second Transplants The characteristics of the patients, donors, transplants, and cytoreductive regimens used for the second transplants in this series are summarized in Table 3. Second transplants were received at a median time of 45 days (range: days) after first transplant (HSCT2-HSCT1). The donors for the second transplants were the same donors as for HSCT-1 in 6 cases, of whom 4 were HLA-matched, and 2 were HLAmismatched, whereas a different donor was used in the other 10 cases, of whom 1 was HLA-matched and 9 HLA-mismatched. For the 6 patients who received second grafts from the same related or unrelated donor who provided the first transplant, the same donor was selected because that donor was the only available HLA-matched sibling or unrelated donor, or because the HLA disparity between that donor and the patient was limited enough to permit the use of an unmodified hematopoietic cell transplant. For the other 10 patients, a different donor was recruited. Seven patients received TCD transplants from an HLA nonidentical related donor, either a parent (N 6) or a sibling (N 1, UPN 2478). Three patients received transplants from a second unrelated donor (2 unmodified and 1 TCD) because their primary HLA-matched unrelated donor was not readily available or willing to provide a second graft. For 5 of these 10 patients, we intentionally selected a secondary donor whose HLA

4 1316 Table 2. First Hematopoietic Stem Cell Transplant CD34* Dose ( 10 6 /kg) UPN Age (Years) Diagnosis- Stage Donor Relation HLA Mismatch Rejection/GVHD HLA Match (R, GVHD) Stem Cell Source Stem Cell Manipulation TNC Dose Cytoreduction Rejection/GVHD Prophylaxis Graft Rejection MDS-RAEB Unrelated None 10/10, 10/10 BM Unmodified /kg ND Bu, Mel Tacrolimus Primary MTX FA Unrelated A,C,DR,DQ/A,C,DR, 6/10, 6/10 Cord Blood NA /kg ND TBI, Cy* ATG-equine Primary DQ CSA AML Unrelated B,C/b 8/10, 9/10 BM T cell /kg 0.8 TBI, Thio, Cy ATG-equine Primary AML-CR2 Related A,B,DR,DQ/A,B,DR, 6/10, 6/10 BM PBSC T cell /kg 5.4 TBI, Thio, Cy ATG-equine Primary DQ NA 8 AML-CR1 Unrelated None 10/10, 10/10 BM Unmodified /kg ND TBI, Cy CSA MTX Primary ALL-CR3 Unrelated B,C,DQ/B,C,DQ 7/10, 7/10 BM T cell /kg 2.8 TBI, Thio, Cy ATG-equine Primary ALL-relapse Unrelated A/A 9/10, 9/10 BM T cell /kg 1.1 Bu, Mel, Flu ATG-rabbit Primary (2 nd ) MDS-RAEB- Unrelated None 10/10, 10/10 BM T cell /kg 1.0 TBI, Thio, Cy ATG-rabbit Primary IT (CR1) CML-1 st CP Related None 10/10, 10/10 BM T cell /kg 1.3 TBI, Thio, Cy ATG-equine Secondary MDS-RA Related None 10/10, 10/10 PBSC T cell /kg 16.1 TBI, Thio, Cy ATG-equine Secondary CML-1 st CP Related None 10/10, 10/10 BM T cell /kg 0.6 TBI, Thio, Cy ATG-equine Secondary ALL-relapse Unrelated B,C,DQ/B,C 7/10, 8/10 BM T cell /kg 1.6 Bu, Mel, Thio ATG-rabbit Primary (2 nd ) ALL-CR2 Unrelated a/a 9/10, 9/10 PBSC T cell /kg 6.8 TBI, Flu, Thio ATG-equine Secondary MDS-RAEB- Related A,C/A 8/10, 9/10 BM T cell /kg 3.7 Bu, Mel, Flu ATG-rabbit Primary IT MDS-RA Unrelated None 10/10, 10/10 PBSC T cell /kg 5.2 Bu, Mel, Flu ATG-equine Primary /kg 1.6 TBI, Thio, Cy ATG-equine Secondary MDS-RA Unrelated None 10/10, 10/10 BM T cell UPN indicates unique patient number; Age, patient age at time of first stem cell transplant; MDS, myelodysplastic syndrome; RAEB, refractory anemia with excess blasts; FA, Fanconi s anemia; AML, acute myelogenous leukemia; CR, complete remission; ALL, acute lymphoblastic leukemia; CML, chronic myelogenous leukemia; HLA mismatch, mismatched HLA loci that are donor unique disparities (rejection) or host unique disparities (GVHD)-upper case indicate HLA antigen differences whereas lower case indicates allelic differences; HLA match, number of matched loci for donor and host R indicates number of donor matched alleles and GVHD indicates the number of host matched alleles; BM, bone marrow; PBSC, peripheral blood stem cells; NA, not applicable; TNC, total nucleated cell dose; ND, not determined or unavailable; Bu, busulfan; Mel, melphalan; TBI, total body irradiation; Cy, cyclophosphamide; Flu, fludarabine; Thio, thiotepa; MTX, methotrexate; ATG, anti-thymocyte globulin; CSA, cyclosporine A. *TBI, Cy given at reduced dose (see text). J. H. Chewning et al.

5 Fludarabine-Based Regimen for Graft Failure 1317 disparities differed from those of the original donor, based on our prior studies implicating host T cells reactive against donor class I or II alleles as effectors of marrow graft rejection [8,9]. For the second transplant, the preparative regimen was based on the cytoreduction regimen of the first transplant (HSCT-1) and on existing patient toxicity following the first transplant and prior to the second transplant; for example, cyclophosphamide was not used in patients with cardiac toxicity. The preparative conditioning consisted of fludarabine (30 mg/m 2 /day 5 days) alone (N 1) or sequentially given with cyclophosphamide (60 mg/ kg/day 2 days) (N 9), or fludarabine (25 mg/ m 2 /day 5 days) sequentially given with thiotepa (5 mg/kg/day 2 days) (N 5). One patient was prepared with the same doses of thiotepa followed by cyclophosphamide. All but 2 of the patients were treated with in vivo TCD prior to transplant to eliminate residual host T cells. The TCD agents employed included equine antithymocyte globulin (eatg) (N 8), rabbit antithymocyte globulin (ratg) (N 3), or Alemtuzumab (Campath) (N 3). eatg dosing range was mg/kg, and ratg was dosed at mg/ kg. The cytoreduction schema for the most common preparative regimen (fludarabine/cyclophosphamide/ ATG) is shown in Figure 1. Posttransplant prophylaxis against graft-versus-host disease (GVHD) was administered to 13 of the 16 patients. Recipients of TCD second transplants received either no additional treatment (N 3) or tacrolimus (N 5), beginning at a dose that was adjusted according to blood levels and maintained for 90 to 120 days posttransplant then tapered. Recipients of unmodified second grafts received cyclosporine (CSP) with or without addition of mycophenolate mofetil (MMF) or steroids (N 6), methotrexate (MTX) alone (N 1), or steroids alone (N 1). Transplants and Patient Evaluations Bone marrow (BM) or peripheral blood stem cells (PBSC) were administered either unmodified or following TCD. TCD of marrow was performed by sequential soybean lectin agglutination (SBA) followed by sheep erythrocyte rosetting (E-rosetting) [10]. PBSC transplants (PBSC) consisted of cells obtained from the donors after mobilization with granulocyte-colony stimulating factor (G-CSF) and leukapheresis [11]. TCD of PBSC was performed by initial positive selection of CD34 cells on Isolex columns followed by E-rosette depletion [12]. HLA class I and class II typing of all donorrecipient pairs was performed by sequence-specific oligonucleotide probe (SSOP) analysis. Donor-recipient match was determined by analysis of HLA-A, B, C, DRB, and DQB alleles. Degree of matching was reported as the number of matched alleles of 10 possible. Disparities unique to the donor, predisposing to rejection, and to host, predisposing to GVHD, are specified in Tables 2 and 3. The presence and level of donor chimerism were evaluated in each patient by sequential cytogenetic and molecular analyses of peripheral blood and marrow samples by fluorescence in situ hybridization (FISH) in sex-mismatched donor-recipient pairs and by quantitation of donor and recipient unique DNA polymorphisms in sex-matched pairs. GVHD was diagnosed and graded by the standard criteria of Glucksberg et al. [13] as modified by Martin et al. [14]. Acute GVHD (agvhd) in this report represents the highest overall grade (I-IV) observed in each patient based on staging of the skin, liver, and gastrointestinal tract. Statistical Analysis A permutation test based on the log-rank statistic was used to determine potential prognostic factors that would predict survival. The application of the permutation procedure resulted from the small number of events in this data set. The primary endpoint selected for analysis was survival. Because there were no cases of graft failure following HSCT-2, analysis of factors influencing engraftment was not possible. Analyses of stem cell dose were based on the Wilcoxon rank sum statistic. RESULTS Engraftment Stem cell doses received for HSCT-1 and HSCT-2 are shown in Tables 2 and 3, respectively. Stem cell doses were significantly higher in HSCT-2 compared to HSCT-1 (P.01). Following HSCT-2, all of the 16 patients engrafted. The median time to engraftment for these patients was 12 days (range: 9-21 days). There was no correlation between time to engraftment and (1) cytoreductive regimen, (2) stem cell manipulation, (3) stem cell dose, or (4) stem cell donor (same or different donor as HSCT-1). In this small cohort, there was no statistical difference between neutrophil or platelet recovery time following TCD SCT, compared with unmodified transplants. GVHD Three patients developed agvhd, 1 of which developed manifestations of agvhd beyond 100 days posttransplant. Disease was limited to the skin and scored as grade I in 2 cases (UPN 2402 and UPN 2050). A third patient (UPN 2991) had combined grade I skin and gut involvement, resulting in grade II GVHD. Both patients with disease limited to the skin had received TCD SCT, and patient

6 1318 J. H. Chewning et al. Table 3. Second Hematopoietic Stem Cell Transplant HSCT1 HSCT2 (days) Stem Cell Source UPN Age (years) Donor HSCT-2 Donor Relation HLA mismatch- Rejection/GVHD HLA Match (R, GVHD) Stem Cell Manipulation TNC Dose Different Related B,C,DR,DQ/B,C, 6/10, 6/10 PBSC T cell depletion /kg DR,DQ Different Related DR,DQ/B,DQ 8/10, 8/10 PBSC T cell depletion /kg 7/10, 7/10 PBSC T cell depletion /kg Different Related b,dr,dq/b,dr, DQ Different Related a,b,c,dr/a,b,c, 6/10, 5/10 PBSC T cell depletion /kg DR,DQ Different Related A,B,DR/A,B,DR 7/10, 7/10 PBSC T cell depletion /kg Different Related A,B,C/A,B,C 7/10, 7/10 PBSC T cell depletion /kg Different Related A,B,C,DR,dq/A, 5/10, 5/10 PBSC T cell depletion /kg B,C,DR,dq Different Unrelated dr/dr 9/10, 9/10 PBSC T cell depletion /kg Same Related None 10/10, 10/10 PBSC Unmodified /kg Same Related None 10/10, 10/10 BM Unmodified /kg Same Related None 10/10, 10/10 PBSC Unmodified /kg Different Unrelated B,c,DQ/B,c 7/10, 8/10 BM Unmodified /kg Same Unrelated a/a 9/10, 9/10 PBSC Unmodified /kg Same Related A,C/A 8/10, 9/10 PBSC Unmodified /kg Different Unrelated None 10/10 10/10 BM Unmodified /kg Same Unrelated None 10/10 10/10 PBSC Unmodified /kg UPN indicates unique patient number; Age, patient age at time of first stem cell transplant; HSCT1 HSCT2 indicates time between first and second BMT (in days); Different, stem cell donor for HSCT2 is not donor for HSCT1; Same, donor the same for HSCT-1 and HSCT-2; HLA mismatch, mismatched HLA loci which are donor unique (rejection) or host unique (GVHD)-upper case indicate HLA antigen differences while lower case indicates allelic differences; HLA Match, number of matched loci for donor and host; PBSC, peripheral blood stem cell; BM, bone marrow; TNC, total nucleated cell dose; Flu, fludarabine, Cy, cyclophosphamide; Thio, thiotepa; ATG, antithymocyte globulin; CSA, cyclosporine A; MMF, mycophenylate mofitil; MTX, methotrexate; PDN, prednisone; Engraft, posttransplant day of engraftment; ND, not determined; N/A, not applicable; EBV-LPD, Ebstein-Barr virus-associated lymphoproliferative disease; GVHD, graft-versus-host disease. UPN 2402 was also given additional tacrolimus therapy as GVHD prophylaxis. Patient UPN 2991 received an unmodified SCT followed by GVHD prophylaxis with cyclosporine (CSa) and MMF. All 3 patients had complete resolution of symptoms without further complications. Ten patients survived greater than 100 days, and were evaluable for chronic graft-versus-host disease (cgvhd). Of these patients, a single patient (UPN 2194) developed extensive cgvhd with severe manifestations involving the skin, gut, and liver. The patient had received a TCD, related, identical BM-derived SCT (HSCT-1), followed by an unmodified PBSCT from the same donor (HSCT-2) with CSA and MMF as GVHD prophylaxis. This patient s GVHD was refractory to therapy, and his disease progressed to multiple-organ system involvement. The patient ultimately died as a result of complications of cgvhd. Survival Six patients are alive at a median follow-up time of 49 months (range: months), resulting in an overall survival (OS) rate of 35% at 3 years by Kaplan-Meier analysis (95% confidence interval: [10%-60%]) (Figure 2A). The DFS rate for this cohort is 18% at 3 years (95% confidence interval: [1%-41%]) from the development of leukemic relapse in 2 patients 6 and 30 months, respectively, following second SCT (UPN 2402 and 2565) (Figure 2B). Both patients received a third SCT, and are currently in remission 50 and 10 months post HSCT-3. The median survival time was 44 days for the remaining 10 patients. Six patients died within 100 days posttransplant. Causes of death were pneumonia (N 3), sepsis with multiorgan failure (N 2), and encephalopathy (N 1). Four patients died 6-15 months posttransplant from GVHD (N 1), relapse (N 1), pneumonia (N 1), and Epstein-Barr virus (EBV) lymphoproliferative disorder (N 1). There were no patient deaths from acute organ failure related to the cytoreductive regimen. In this small series, the long-term OS of patients aged 20 years or younger was superior to that of older patients (P.02). Five of 8 younger patients are alive,

7 Fludarabine-Based Regimen for Graft Failure 1319 Table 3. (Continued) CD34 Dose Cyto Reduction Rejection Prophylaxis GVHD Prophylaxis Engraft (days) GVHD (Grade) Chimerism Survival Time (days) Cause of Death /kg Flu/Cy ATG-equine Tacrolimus 14 Acute skin (I) 100% donor 1870 N/A /kg Flu/Cy ATG-rabbit Tacrolimus % donor 2900 N/A /kg Flu/Cy None-did not None % donor 400 N/A tolerate ATG-rabbit /kg Flu/Cy ATG-equine None 12 Acute skin (I) 100% donor 202 Infection /kg Flu/Cy ATG-equine Tacrolimus % donor 2700 N/A /kg Flu/Cy ATG-equine Tacrolimus % donor 177 EBV-LPD /kg Flu/Thio Campath None % donor 178 Relapse /kg Flu/Thio ATG-equine Tacrolimus % donor 41 Encephalopathy ND Flu/Cy None CSA MMF 11 Chronic-skin, 100% donor 461 GVHD gut, liver (extensive) ND Flu/Cy ATG-equine MTX 21 0 ND 37 Infection /kg Flu/Cy ATG-equine steroids 9 0 ND 26 Infection ND Flu/Thio ATG-equine CSA MMF % donor 47 Infection /kg Flu/Thio Campath CSA 12 0 ND 27 Infection /kg Thio/Cy Campath CSA % donor 1430 N/A /kg Flu/Thio ATG-rabbit CSA MMF 14 Acute-skin 100% donor 300 N/A and GI (II) /kg Fludarabine alone ATG-rabbit CSA PDN % donor 35 Infection whereas only 1 of the 8 patients older than 20 years of age survived over 6 months. The outcome of patients receiving stem cells for HSCT-2 from a second donor different than that of HSCT-1 was more favorable than the outcome of patients receiving a second SCT from the same donor. Five of 10 patients with different stem cell donors survived compared with only 1 of 6 patients receiving stem cells from the same donor. This difference, however, was not statistically significant (P.10). Similarly, the outcome of patients receiving TCD stem cells in HSCT-2 was somewhat better than the outcome of those receiving unmodified stem cells. Four of 8 patients survived after receiving TCD SCT compared with only 2 of 8 patients who received an unmodified SCT. However, this difference was also not found to be significant (P.17). Patient 100-day mortality was not statistically different when analyzed for patient age, donor source (same versus different), or stem cell manipulation. Similarly, stem cell source and cytoreduction regimen were not found to be associated with differences in engraftment or survival. DISCUSSION HSCT is an important modality in the treatment of high-risk hematologic and immunologic disorders. However, HSCT has an increased risk of morbidity and mortality because of complications such as organ toxicity, infections, and GVHD. Graft failure represents another rare but potentially lethal complication of HSCT. The incidence of graft failure has been reported to be as low as 0.1% for unmodified HSCT from HLAmatched siblings administered after myeloablative, TBI-containing regimens for leukemia, and as high as 17% following TCD HSCT [1,15]. Risk factors for graft failure reported in the literature include HLA disparity [16-19], TCD [17,20], HSCT for nonmalignant disorders, such as aplastic anemia [17] or hemoglobinopathies [21,22], donor age and sex [8], and cytoreduction using chemotherapy only or nonmyeloablative regimens [23,24]. Umbilical cord blood (UCB) transplants have also been associated with an increased risk of graft failure, with rates of 10%-30% [25,26]. It is important to differentiate graft failure associated with allogeneic HSCT for malignant hematologic disorders where graft failure is an acute process often associated with graft rejection, and graft failure associated with allogeneic HSCT for nonmalignant hematologic disorders such as aplastic anemia and hemoglobinopathies where it presents as a subacute process associated with prolonged mixed chimerism occurring over periods of weeks to months. The patients described in our series belong to the first patient group. A review of

8 1320 J. H. Chewning et al. FLU FLU FLU FLU FLU 30 mg/m2/day x 5 CY CY HSCT-2 60 mg/kg/day x 2 ATG ATG ATG ATG Rabbit ATG 2.5 mg/kg/day x 4 Day Figure 1. Schema for the most frequently administered cytoreduction regimen. Chemotherapeutic agents and ATG are displayed as a box and shown above the corresponding days these agents were administered. FLU indicates fludarabine and CY indicates cyclophosphamide. The HSCT-2 box indicates the administration of the second stem cell graft and is shown above the day of transplant (day 0). the literature revealed 6 retrospective descriptions of such patients [3,8,19,27-29]. The treatment of patients with graft failure following HSCT has included the reinfusion of autologous cryopreserved progenitor cells [30] or high-dose hematopoietic growth factors [31], both of which have been associated with a poor outcome for those patients with true graft failure. Second HSCT represents the therapeutic option with the best chance for long-term survival for patients with primary or secondary graft failure. In this setting, stem cells from the same or a different donor are infused to the patient either with or without additional cytoreduction. In the previously cited series, OS and DFS rates have varied from 6% to 43% [3,8,19,27-29]. There are few studies analyzing engraftment or survival outcomes based on donor choice for second HSCT. Following failure of an HLA-matched graft, the same donor is most commonly recruited because a similarly matched secondary donor is rarely available. In addition, because of the tenuous condition of these patients, the time required to identify and recruit a second donor limits this option. Nevertheless, in a series by Grandage et al. [19], 12 pediatric patients ( 18 years) suffered graft failure a median of 5 months following HSCT from an unrelated donor. Second unmodified HSCT using a different unrelated donor was attempted in 7 patients. Donor source did not affect outcome in this study, with 3 of 6 patients surviving after receiving a second HSCT from the same donor compared to 2 of 6 surviving recipients of second HSCT from a different donor. Overall, 6 of 9 patients achieved engraftment and 5 were long-term survivors. However, there was an increased risk of agvhd, with 67% grade II-IV GVHD [19]. In our limited series recipients of second HSCT from a different donor tended to have a better outcome (5 of 10 long-term survivors) than those who received HSCT from the same donor (1 of 6 long-term survivors), although this was not statistically significant (P.11). Strikingly, in our series using TCD second transplants, the incidence of grade II-IV agvhd or extensive cgvhd was low. The sole purpose of a cytoreductive regimen in the setting of second HSCT is immunoablation in order to secure stem cell engraftment. There is usually no further need for antileukemic or myeloablative effects, as patients are pancytopenic with aplastic marrow. Engraftment rates following second HSCT have varied widely in the literature, reflecting the heterogeneity of the reported patient populations [3,27]. One of the first reported studies using a second HSCT was performed by Kernan et al. [8] at this institution in patients with graft failure following a TCD HSCT. Four initial patients received a second HSCT from the same donor without immune suppression, and all 4 experienced a second rejection. Subsequently, 5 patients received another dose of donor stem cells after pretreatment with ATG and steroids with 3 of 5 engrafting and 1 long-term survivor. In a study by Guardiola et al. [27], 82 patients received a second SCT following primary or secondary graft failure of a first HSCT administered for acute or chronic leukemia and aplastic anemia. Of these 82 patients, 51 achieved engraftment, with a 73% overall probability of engraftment. Engraftment occurred at a mean of 17 days following HSCT. No study has specifically addressed the use of fludarabine-containing regimens for the stable engraftment of donor stem cells following a second SCT after graft failure. We achieved engraftment in all patients in this series using fludarabine-based chemotherapy regimens, with or without ATG, at a median of 12 days after stem cell infusion. Several centers have favored G-CSF-mobilized PBSCs over bone marrow-derived stem cells because the former yields increased numbers of stem cells. A report by Zecca et al. [32] showed engraftment of 2

9 Fludarabine-Based Regimen for Graft Failure 1321 Figure 2. Kaplan-Meier probability of overall (A) and disease-free (B) survival by age following second HSCT. patients following graft failure using PBSCs. The only larger study to address PBSC versus bone marrow stem cells reported higher engraftment rates but equivalent OS following PBSC transplants administered for graft failure. The authors of this study did not report stem cell numbers for PBSCT and BM [27]. In our series, increased stem cell numbers were used for HSCT-2 to improve engraftment rates. The dose of CD34 cells was statistically higher in HSCT-2 compared to HSCT-1 (P.01). The use of increased CD34 stem cell numbers in HSCT-2 likely contributed to the high engraftment rate in this patient population. The use of megadose CD34 has also been associated with low graft failure rates and low GVHD incidence, even in the setting of HLAmismatch [33]. There is no literature on the use of megadose stem cells in the setting of graft failure. All patients in this study engrafted, and no statistical difference was seen in disease-free or overall survival according to stem cell source. The use of TCD stem cells for second transplant following graft failure has not been well studied. Most of the literature includes only unmodified transplants for these patients. The largest series of patients included 12% who received TCD second grafts following primary graft failure, but no separate analysis of these patients was performed [27]. In our study, 8 of 16 patients received TCD PBSCs. Four of these patients are alive versus only 2 of 8 patients receiving conventional BM or PBSC, although this result was not statistically significant (P.13). Fifteen patients survived greater than 30 days and 10 survived 100 days posttransplant, resulting in a 25% overall incidence of agvhd (grade I or II) and a 10% incidence of cgvhd. The patients in this study received a second HSCT following primary graft failure at a median interval time (HSCT2-HSCT1) of 45 days, with all patients transplanted within 3 months of HSCT-1. The patients in this series therefore represent a highrisk population based on the short interval time between the first and second transplant. Previous studies have shown an increased interval between SCT is associated with improved patient survival [19,27]. Specifically, patients receiving HSCT-2 80 days following primary graft failure were shown to have a lower 100 day transplant-related mortality and increased 3-year survival [27]. Infectious etiologies were the principal cause of mortality in 6 patients in this series. Other causes of death included encephalopathy, relapse, GVHD, and EBV lymphoma. Mortality was highest in the peritransplant time period, with 6 patients dying within 100 days of transplant. As previously described, these patients represented a high-risk population. Thirteen patients in this study had infectious complications following HSCT-1, and 10 of these patients had active infectious issues entering HSCT-2. It is possible that immunosuppression secondary to fludarabine resulted in the increased mortality in this at-risk population. Other factors, however, such as immunosuppression from the previous transplant (HSCT-1), prolonged pancytopenia from graft failure, and the high frequency of preceding and ongoing infections at HSCT-2 may also have contributed to the mortality from infectious etiologies. Ongoing progress in the management of peritransplant complications, such as adoptive T cell therapy for EBV and cytomegalovirus (CMV) and improved immunomodulators for GVHD treatment, will continue to improve survival of these patients following stem cell engraftment of second grafts. We have shown that achieving stem cell engraftment with a second HSCT is possible following an

10 1322 J. H. Chewning et al. initial graft failure by using an immunosuppressive, minimally toxic secondary cytoreduction, and increased stem cell numbers. In this study, we were able to obtain stem cell engraftment in all of these highrisk patients. Fewer patients received a TCD HSCT-2 compared to HSCT-1, and it is possible that this contributed to less graft failure in this cohort. Nevertheless, all 8 recipients of TCD transplants engrafted. Cytoreduction was relatively well tolerated with no acute organ failure from toxicity, including no cases of veno-occlusive disease or grade III-IV mucositis. We report an OS rate of 37%, with a median follow-up of 49 months. Younger patients ( 20 years) had a superior survival rate (5 of 8 surviving) than those older than 20 years (1 of 8 patients surviving) (P.01). In addition, there was a trend toward better outcome for recipients of TCD HSCT and transplants from different donors. We have found that a fludarabinebased regimen including thiotepa or cyclophosphamide along with ATG, followed by a TCD PBSC transplant containing higher numbers of progenitor cells from a different donor resulted in consistent engraftment. In these patients, there is a need for particular attention to infectious risk during the peritransplant and early posttransplant periods. A prospective multicenter study that would include larger patient numbers, using a fludarabine-containing cytoreductive regimen followed by TCD PBSCT with aggressive posttransplant infectious prophylaxis for patients with graft failure could be beneficial in studying this promising approach to graft failure posttransplant. ACKNOWLEDGMENTS We wish to thank the nurses and nurse practitioners, including Heidi Abendroth, Joanne Torok-Castanza, Anne Casson, Mary K. Conlon, and Christine Iovino, of the adult and pediatric inpatient bone marrow transplant services who helped care for these patients. REFERENCES 1. Ash RC, Casper JT, Chitambar CR, et al. Successful allogeneic transplantation of T-cell-depleted bone marrow from closely HLA-matched unrelated donors. N Engl J Med. 1990;322: Green A, Clarke E, Hunt L, et al. Children with acute lymphoblastic leukemia who receive T-cell-depleted HLA mismatched marrow allografts from unrelated donors have an increased incidence of primary graft failure but a similar overall transplant outcome. Blood. 1999;94: Davies SM, Weisdorf DJ, Haake RJ, et al. Second infusion of bone marrow for treatment of graft failure after allogeneic bone marrow transplantation. Bone Marrow Transplant. 1994; 14: Gandhi V, Estey E, Keating MJ, Plunkett W. Fludarabine potentiates metabolism of cytarabine in patients with acute myelogenous leukemia during therapy. J Clin Oncol. 1993;11: Terenzi A, Aristei C, Aversa F, et al. Efficacy of fludarabine as an immunosuppressor for bone marrow transplantation conditioning: preliminary results. Transplant Proc. 1996;28: Aversa F, Tabilio A, Velardi A, et al. Treatment of high-risk acute leukemia with T-cell-depleted stem cells from related donors with one fully mismatched HLA haplotype. N Engl J Med. 1998;339: Devine SM, Sanborn R, Jessop E, et al. Fludarabine and melphalan-based conditioning for patients with advanced hematological malignancies relapsing after a previous hematopoietic stem cell transplant. Bone Marrow Transplant. 2001;28: Kernan NA, Bordignon C, Heller G, et al. Graft failure after T-cell-depleted human leukocyte antigen identical marrow transplants for leukemia: I. Analysis of risk factors and results of secondary transplants. Blood. 1989;74: Bordignon C, Keever CA, Small TN, et al. Graft failure after T-cell-depleted human leukocyte antigen identical marrow transplants for leukemia: II. In vitro analyses of host effector mechanisms. Blood. 1989;74: O Reilly RJ, Kapoor N, Pollack M, et al. Reconstitution of immunologic function in a patient with severe combined immunodeficiency following transplantation of marrow from an HLA-A,B,C nonidentical but MLC-compatible paternal donor. Transplant Proc. 1979;11: Young JW, Papadopoulos EB, Cunningham I, et al. T-celldepleted allogeneic bone marrow transplantation in adults with acute nonlymphocytic leukemia in first remission. Blood. 1992; 79: Collins NH BS, Kernan NA, et al. T cell-depletion of bone marrow by treatment with soybean agglutinin and sheep red blood cell rosetting. In: Areman HJ DH, Sacher RA, eds. Bone Marrow and Stem Cell Processing: A Manual of Current Techniques. Philadelphia, PA: FA Davis, 1992: 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: Martin PJ, Schoch G, Fisher L, et al. A retrospective analysis of therapy for acute graft-versus-host disease: initial treatment. Blood. 1990;76: Beatty PG, Clift RA, Mickelson EM, et al. Marrow transplantation from related donors other than HLA-identical siblings. N Engl J Med. 1985;313: Davies SM, Kollman C, Anasetti C, et al. Engraftment and survival after unrelated-donor bone marrow transplantation: a report from the national marrow donor program. Blood. 2000; 96: Ash RC, Horowitz MM, Gale RP, et al. Bone marrow transplantation from related donors other than HLA-identical siblings: effect of T cell depletion. Bone Marrow Transplant. 1991; 7: Anasetti C, Amos D, Beatty PG, et al. Effect of HLA compatibility on engraftment of bone marrow transplants in patients with leukemia or lymphoma. N Engl J Med. 1989; 320: Grandage VL, Cornish JM, Pamphilon DH, et al. Second allogeneic bone marrow transplants from unrelated donors for

11 Fludarabine-Based Regimen for Graft Failure 1323 graft failure following initial unrelated donor bone marrow transplantation. Bone Marrow Transplant. 1998;21: Kernan NA, Bartsch G, Ash RC, et al. Analysis of 462 transplantations from unrelated donors facilitated by the National Marrow Donor Program. N Engl J Med. 1993;328: Woodard P, Tong X, Richardson S, et al. Etiology and outcome of graft failure in pediatric hematopoietic stem cell transplant recipients. J Pediatr Hematol Oncol. 2003;25: Miniero R, Rocha V, Saracco P, et al. Cord blood transplantation (CBT) in hemoglobinopathies. Eurocord. Bone Marrow Transplant. 1998;22(Suppl 1):S78-S 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: Baron F, Maris MB, Storer BE, et al. HLA-matched unrelated donor hematopoietic cell transplantation after nonmyeloablative conditioning for patients with chronic myeloid leukemia. Biol Blood Marrow Transplant. 2005;11: Wagner JE, Barker JN, DeFor TE, et al. Transplantation of unrelated donor umbilical cord blood in 102 patients with malignant and nonmalignant diseases: influence of CD34 cell dose and HLA disparity on treatment-related mortality and survival. Blood. 2002;100: Gluckman E, Rocha V, Chevret S. Results of unrelated umbilical cord blood hematopoietic stem cell transplantation. Rev Clin Exp Hematol. 2001;5: Guardiola P, Kuentz M, Garban F, et al. Second early allogeneic stem cell transplantations for graft failure in acute leukaemia, chronic myeloid leukaemia and aplastic anaemia. French Society of Bone Marrow Transplantation. Br J Haematol. 2000;111: Shah AJ, Kapoor N, Weinberg KI, et al. Second hematopoietic stem cell transplantation in pediatric patients: overall survival and long-term follow-up. Biol Blood Marrow Transplant. 2002; 8: Remberger M, Ringden O, Ljungman P, et al. Booster marrow or blood cells for graft failure after allogeneic bone marrow transplantation. Bone Marrow Transplant. 1998;22: Mehta J, Powles R, Singhal S, Horton C, Treleaven J. Outcome of autologous rescue after failed engraftment of allogeneic marrow. Bone Marrow Transplant. 1996;17: Nemunaitis J, Singer JW, Buckner CD, et al. Use of recombinant human granulocyte-macrophage colony-stimulating factor in graft failure after bone marrow transplantation. Blood. 1990; 76: Zecca M, Perotti C, Marradi P, et al. Recombinant human G-CSF-mobilized peripheral blood stem cells for second allogeneic transplant after bone marrow graft rejection in children. Br J Haematol. 1996;92: Reisner Y, Gur H, Reich-Zeliger S, Martelli MF, Bachar- Lustig E. Hematopoietic stem cell transplantation across major genetic barriers: tolerance induction by megadose CD34 cells and other veto cells. Ann N Y Acad Sci. 2005; 1044:70-83.