Scholarly article on topic 'Higher CD34+ and CD3+ Cell Doses in the Graft Promote Long-Term Survival, and Have No Impact on the Incidence of Severe Acute or Chronic Graft-versus-Host Disease after In Vivo T Cell-Depleted Unrelated Donor Hematopoietic Stem Cell Transplantation in Children'

Higher CD34+ and CD3+ Cell Doses in the Graft Promote Long-Term Survival, and Have No Impact on the Incidence of Severe Acute or Chronic Graft-versus-Host Disease after In Vivo T Cell-Depleted Unrelated Donor Hematopoietic Stem Cell Transplantation in Children Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Hematopoietic stem cell transplantation" / "Unrelated donor" / "Graft-versus-host disease" / Children / "CD34+ cell dose" / "CD3+ cell dose" / "Graft content"}

Abstract of research paper on Clinical medicine, author of scientific article — Krzysztof Kałwak, Julita Porwolik, Monika Mielcarek, Ewa Gorczyńska, Joanna Owoc-Lempach, et al.

The aim of our study was to compare the results of unrelated donor (UD) peripheral blood stem cell transplantation versus UD bone marrow transplantation and to analyze the impact of infused CD34+ and CD3+ cell doses on survival and incidence of severe graft-versus-host disease (GVHD) in 187 children who underwent UD hematopoietic cell transplantation with the use of in vivo T cell depletion (antithymocyte globulin or CAMPATH-1H). HLA typing was performed at the “high-resolution” level. Patients receiving ≥10 × 106 CD34+ cells/kg and ≥4 × 108 CD3+ cells/kg had better overall and disease-free survival. Multivariate analysis has shown that both infused CD34+ cell dose <10 × 106/kg and CD3+ cell dose <4 × 108/kg were independent risk factors for mortality (relative risk [RR] 1.8 and 1.71, P = .009 and .016, respectively). Regarding disease-free survival, multivariate analysis has revealed another independent risk factor for poor outcome apart from the 2 earlier-mentioned cell doses, which was the use of donors mismatched at 2 HLA antigens or 3 HLA allele/antigens (RR 2.5, P = .004). In age groups 0-10 years and 10-20 years, CD34+ cell doses higher than the age-adjusted median dose clearly favored survival. Higher infused doses of CD34+ and CD3+ cells did not result in an increased rate of severe GVHD. The use of mismatched donors was the only independent risk factor for the incidence of severe acute GVHD (RR 2.2, P = .046). The report demonstrates for the first time in a pediatric cohort, that higher doses of transplanted CD34+ and CD3+ cells lead to an improved survival without an increased risk of severe GVHD. The study findings may be limited to the population of patients receiving in vivo T cell depletion, which is now broadly used in unrelated donor setting in Europe.

Academic research paper on topic "Higher CD34+ and CD3+ Cell Doses in the Graft Promote Long-Term Survival, and Have No Impact on the Incidence of Severe Acute or Chronic Graft-versus-Host Disease after In Vivo T Cell-Depleted Unrelated Donor Hematopoietic Stem Cell Transplantation in Children"

ASBMI

American Society for Blood and Marrow Transplantation

Higher CD341 and CD31 Cell Doses in the Graft Promote Long-Term Survival, and Have No Impact on the Incidence of Severe Acute or Chronic Graft-versus-Host Disease after In Vivo T Cell-Depleted Unrelated Donor Hematopoietic Stem Cell Transplantation in Children

Krzysztof Kalwak,1 Julita Porwolik,1 Monika Mielcarek,1 Ewa Gorczynska,1

Joanna Owoc-Lempach,1 Marek Ussowicz,1 Agnieszka Dyla,1 Jakub Musial,1

1 12 1 Dominika Pazzdzior, Dominik Turkiewicz, ' Alicja Chybicka

The aim of our study was to compare the results of unrelated donor (UD) peripheral blood stem cell transplantation versus UD bone marrow transplantation and to analyze the impact of infused CD341 and CD31 cell doses on survival and incidence of severe graft-versus-host disease (GVHD) in 187 children who underwent UD hematopoietic cell transplantation with the use of in vivo T cell depletion (antithymocyte globulin or CAMPATH-IH). HLA typing was performed at the "high-resolution" level. Patients receiving $10 x I06 CD341 cells/kg and $4 x I08 CD31 cells/kg had better overall and disease-free survival. Multivariate analysis has shown that both infused CD341 cell dose <10 x I06/kgand CD31 cell dose <4 x I08/kgwere independent risk factors for mortality (relative risk [RR] I.8 and I.7I, P = .009 and .0I6, respectively). Regarding disease-free survival, multivariate analysis has revealed another independent risk factor for poor outcome apart from the 2 earlier-mentioned cell doses, which was the use of donors mismatched at 2 HLA antigens or 3 HLAallele/antigens (RR 2.5, P = .004). In age groups 0-I0 years and I0-20 years, CD34 cell doses higher than the age-adjusted median dose clearly favored survival. Higher infused doses of CD341 and CD31 cells did not result in an increased rate of severe GVHD. The use of mismatched donors was the only independent risk factor for the incidence of severe acute GVHD (RR 2.2, P = .046). The report demonstrates for the first time in a pediatric cohort, that higher doses of transplanted CD34 and CD3 cells lead to an improved survival without an increased riskof severe GVHD. The study findings may be limited to the population of patients receiving in vivo T cell depletion, which is now broadly used in unrelated donor setting in Europe. Biol Blood Marrow Transplant 16: 1388-1401 (2010) © 2010 American Society for Blood and Marrow Transplantation

KEY WORDS: Hematopoietic stem cell transplantation, Unrelated donor, Graft-versus-host disease, Children, CD341 cell dose, CD31 cell dose, Graft content

From the 'Department of Pediatric Hematology, Oncology and Bone Marrow Transplantation, Wroclaw Medical University, Poland; and 2Department of Pediatric Oncology, University Hospital, Lund, Sweden. Financial disclosure: See Acknowledgments on page 1400. Correspondence and reprint requests: Krzysztof Kalwak, MD, PhD, Department of Pediatric Hematology, Oncology and Bone Marrow Transplantation, Wroclaw Medical University, 44 Buj-wida St., 50-345 Wroclaw, Poland (e-mail: kk@pedhemat.am. wroc.pl).

Received December 14, 2009; accepted April 1, 2010

© 2010 American Society for Blood and Marrow Transplantation

1083-8791/$36.00

doi:10.1016/j.bbmt.2010.04.001

INTRODUCTION

Allogeneic hematopoietic stem cell transplantation (HSCT) has become a standard therapy for children with a variety of both malignant and nonmalignant diseases [1]. In Poland, only approximately 20% of children have an HLA matched sibling donor (MSD) available. The remaining 80% of patients have to be transplanted from alternative donors. For children with rare HLA haplotypes, it remains extremely difficult and time consuming to find a fully matched unrelated donor (MUD). It is generally acceptable to use unrelated donors (UD) mismatched at 1 HLA allele or antigen. Such UDs are categorized as MUD, together with fully matched ones (10/10-allele matched).

According to the acute lymphoblastic leukemia (ALL) international BFM SCT 2008 (Vienna, Austria) transplant protocol, UD HSCT from donors mismatched at more than 1 HLA antigen or allele, requires extensive in vitro T cell depletion. There are, however, two published studies confirming low mortality of children undergoing HSCT from 7 to 8/10 HLA allele-matched UD with the use of antithymocyte globulin (ATG), but not T cell depletion in vitro [2,3].

There is an unresolved question whether unmanipulated UD-peripheral blood stem cell transplantation (PBSCT) leads to a worse overall outcome, and a higher incidence of severe graft-versus-host disease (GVHD) in both adults and children, in comparison with well-established UD bone marrow transplantation (BMT). It is well known that PBSC grafts contain approximately 1 log more mature T cells, which may result in a higher rate of severe (often fatal) acute or/and chronic GVHD (aGVHD, cGVHD) posttransplant.

The aim of our study was to compare the results of UD PBSCT versus UD BMT and analyze the impact of CD341 and CD31 cell doses in the grafts in a large pediatric patient population. PBSC grafts with larger amounts of CD341 and CD31 cells may lead to a prompter engraftment and improved outcome because of a more pronounced graft-versus-malignancy effect [4,5]. However, studies on both adults and children show discrepant results with regard to overall survival (OS) and occurrence of aGVHD and cGVHD after UD PBSCT [6-12]. Based on the fact that the majority of volunteer donors in Germany and Poland prefer (sometimes exclusively) PBSC donation, the number of UD PBSCT in our center has increased steadily over the last decade. To our knowledge, there is no published study with a higher number of UD PBSCTs performed in children and adolescents. Furthermore, data on the impact of CD341 and CD31 cell doses in the UD graft on the OS and disease-free survival (DFS) and aGVHD/ cGVHD is scarce for adults [13-17] and virtually nonexistent for children. In this multivariate detailed analysis of different factors influencing the outcome and the incidence of severe GVHD post-UD HSCT, we demonstrate for the first time in a large pediatric cohort, that higher doses of transplanted CD341 and CD31 cells lead to improved survival without an increased risk of severe GVHD, provided that in vivo T cell depletion is used.

MATERIALS AND METHODS

Patients

Between the years 2000 and 2008,187 children and adolescents (112 males and 75 females) with malignant (n = 155) and nonmalignant disorders (n = 32)

underwent unmanipulated allogeneic UD HSCT at our center (Table 1). Their median age at transplant was 9.6 years (range: 0.6-20.2 years). All patients with malignant disorders were given disease-specific mye-loablative (MA) conditioning regimens according to European protocols, except for those receiving a second transplant, who were grafted with disease-adapted re-duced-toxicity regimens including treosulfan (instead of busulfan). Total body irradiation (TBI) was used in 51 patients. Patients with nonmalignant disorders were conditioned according to the guidelines of the European Blood and Marrow Transplant (EBMT) Inborn Errors Working Party except for patients with severe aplastic anemia (SAA) conditioned mostly with treosulfan and cyclophosphamide. For GVHD prophylaxis, rabbit ATG (Fresenius, median dose 30 mg/ kg) was used in 160 patients, whereas 22 children received Thymoglobuline (Genzyme, median dose 10 mg/kg) and 5 other patients Campath-1H (median dose 1 mg/kg). Standard Cyclosporine A (CsA) and short methotrexate (MTX) were added to GVHD prophylaxis. The risk status of the patients at transplant was assessed by our own modification of the classification proposed by Meisel et al. [12]: patients with acute lymphoblastic leukemia (ALL)/acute myelogenous leukemia (AML) in first and second complete remission (CR), chronic myelogenous leukemia (CML) in the first chronic phase, myelodysplastic syndrome (MDS)-refractory cytopenia, non-Hodgkin lymphoma (NHL) in CR, and any nonmalignant disease were considered standard risk (n = 118), whereas children with ALL/AML in equal to or greater than the third CR or nonremission, CML equal to or greater than the accelerated phase, MDS-refractory anemia with excess of blasts, juvenile myelomonocytic leukemia (JMML), NHL in nonremission, and patients undergoing any second transplant were classified as high risk (n = 69).

The grafted cells were obtained from HLA-allele matched or mismatched UDs (age 19-56 years, median

Table 1. Primary Diagnosis (n = 187 Patients)

Malignant N Nonmalignant N

ALL 67 SAA 13

AML 33 Fanconi Anemia 5

CML 25 X-ALD 4

MDS 18 SCID 3

NHL 5 WAS 2

JMML 4 Omenn Syndrome 2

HD 2 X-CGD 2

LCAL 1 Griscelli Syndrome 1

Total 155 Total 32

ALL indicates acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CML, chronic myelogenous leukemia; MDS, myelodysplastic syndrome; NHL, non-Hodgkin lymphoma; JMML, juvenile myelomonocytic leukemia; HD, Hodgkin disease; LCAL, large cell anaplastic lymphoma; SAA, severe aplastic anemia; X-ALD, X-linked adrenoleukodystrophy; SCID, severe combined immunodeficiency disease; WAS, Wiskott-Aldrich syndrome; X-CGD, X-linked chronic granulomatous disease.

34 years). HLA typing was performed at the highresolution level (4 digits) in A*,B*, Cw*, DRB1 *, and DQB1* alleles. There were 159 matched- and 28 mismatched donor-recipient pairs, respectively (according to ALL international BFM SCT 2008 criteria, all 10 alleles were high-resolution typed). Table 2 presents the detailed distribution of different combinations of HLA-mismatched donor-recipient pairs. Table 3 provides details on the patient sex, age, diagnosis, risk status at transplant, donor age, donor/recipient CMV-status and blood group compatibility, year of transplant (2000-2004 versus 2005-2008), TBI use, degree of HLA-match and data on neutrophil and platelet recovery within the 3 groups divided according to stem cell source (peripheral blood [PB] versus BM), and number of infused CD341 and CD31 cells in the graft. Median number of transplanted CD341 cells was 11.3 x 106/kg (range: 0.8-57.8) in the PBSC group and 3.4 x 106/kg (range: 0.8-11.0) in the BM group (P < .001). Median number of transplanted CD31 cells was 5.2 x 108/kg (range: 0.2-20.0) in the PBSC group and 0.4 x 108/kg (range: 0.04-10.1) in the BM group (P < .001).

The study was approved by the local Ethics Committee, and patients over 18 years old or parents of patients under 18 years signed informed consents.

Aims of the Study

As the endpoints of this prospective study, hematologic recovery, OS, DFS, the incidence of severe acute

Table 2. Types of HLA-Mismatched Donor-Recipient Pairs

Full allele match (10/10) -A*B*Cw*DRBI *DQBI * 94/187 (50,26%)

One antigen or allele mismatch (9/10) 65 (47+18) /187 (34,76%)

A*-mismatch 14 (9+5)

B*-mismatch 8 (2+6)

Cw*-mismatch 33 (30+3)

DRBI *-mismatch 2 (0+2)

DQBI *-mismatch 8 (6+2)

Two allele/antigen mismatch (8/10) 23/187 (12.29%)

Two antigen mismatch (8 ant)

B* +Cw* 3

A*+Cw* 3

Cw* +Cw* 1

Two allele or two antigen + allele mismatch

B* +Cw* 6

C* + DQBI * 3

B* +DQBI * 2 (1 + 1)

A* +B* 2 (1 + 1)

Cw* +Cw* 1

A*+Cw* 1

A*+A* 1

Three allele/antigen mismatch (7/I0) 5/187 (2.67%)

B* +Cw* +Cw* 1

B* +Cw* +Cw* 1

A* +B * +Cw * 1

B* +B* +DQBI * 1

B* +B* +Cw* 1

8 ant indicates abbreviation for the group of two HLA antigen-mismatched recipient pairs; antigen mismatches are depicted as first number in brackets.

GVHD (aGVHD grade III-IV) and extensive chronic GVHD (cGVHD) were compared within 3 study cohorts of patients:

1. Recipients of PBSC versus BM;

2. Patients grafted with higher CD341 cell dose ($10 x 106/kg) versus those grafted with lower dose (<10 x 106/kg).

3. Patients grafted with higher CD31 cell dose ($4 x 108/kg) versus those grafted with lower dose (<4 x 108/kg).

The cell dose cutoffs were chosen by taking 2 approximately similar (in numbers) cohorts of children. The cell dose of 10 x 106 CD34/kg was slightly higher than median (9.72 x 106 CD34/kg), so we rounded the division value up to 10 x 106. Similarly, the cell dose of 4 x 108 CD3/kg was slightly lower than the median (4.26 x 108/kg), so we rounded the division value down to 4 x 108/kg. We concluded that values close to medians were good cutoff points for cell doses (as in NMDP study) [17].

To avoid a possible bias in favor of higher CD341 cell doses in younger patients, we analyzed separately the age groups 0-10 years and 10-20 years and then 0-5 years, 5-10 years, 10-15 years, and older. We looked at the influence of CD341 or CD31 cell dose (above and below age-adjusted medians) on OS, DFS, and both severe aGVHD or extensive cGVHD in relatively homogenous smaller age groups. Ten years was chosen to be the age cutoff because it was close to the median (9.6 years), so we rounded up the cutoff value to 10.

In addition, a parallel subanalysis of patients grafted with different CD341 and/or CD31 cell doses was performed using a method (analysis of quartiles) similar to the one described in a recently published NMDP study [16].

Furthermore, the secondary aim of our study was to analyze the impact of HLA mismatch on the outcomes.

Hematologic recovery was defined as the achievement of an absolute neutrophil count (ANC) >0.5 x 109/L and/or platelet count >50 x 109/L for 3 consecutive days, respectively. For the analysis ofcGVHD, only patients surviving beyond day 1100 were taken into consideration. GVHD was assessed according to well-known clinical symptoms, laboratory tests, and histopathology, as applicable. aGVHD and cGVHD were graded according to the criteria described previously and adapted for children [18,19]. aGVHD was treated mostly with methylprednisolone and CsA. In case of steroid resistance nonstandardized fashion including ATG, etanercept, or mycophenolate mofetil (MMF) or a combination of 3-4 immunosuppressive drugs was used. Extensive cGVHD was treated with methylprednisolone, CsA, etanercept, and/or MMF.

Table 3. Characteristics of the Study Cohorts: Univariate Analysis of Multiple Variables

Variables Stem Cell Source P* CD34+ Cell Dose Pf CD3+ Cell Dose Pf

PB N (%) BM N (%) N Median (SE) N Median (SE)

Recipient sex .680 .487 .535

male 88 (60.7) 24(57.1) 75 9.42 (0.86) 73 4.33 (0.35)

female 57 (39.3) I8 (42.9) II2 9.78 (1.27) II0 4.11 (0.48)

Recipient age .256 .000 .000

$ 10 years 65 (44.8) 23 (54.8) 89 7.20 (0.56) 86 3.07 (0.27)

<10 years 80 (55.2) I9 (45.2) 98 12.13 (1.17) 97 5.64 (0.44)

Donor age .883 .870 .741

$35 years 64 (44.I) I8 (42.9) 82 9.90 (1.09) 8I 3.70 (0.45)

<35 years 8I (55.9) 24 (57.l) I05 9.37 (0.98) I02 4.36 (0.37)

Sex match .399 .013 .813

female to male 23 (I5.9) 9 (21.4) 32 6.77 (1.26) 3I 4.97 (0.73)

other I22 (84.I) 33 (78.6) I55 9.97 (0.83) I52 4.13 (0.3l)

Recipient CMV-status .579 .883 .859

negative 57 (43.5) I8 (48.7) 75 9.81 (1.11) 72 4.49 (0.43)

positive 74 (56.5) 19 (51.4) 93 9.47 (1.00) 92 4.27 (0.39)

Donor CMV-status .178 .912 .389

negative 76 (58.0) 26 (70.3) I02 9.55 (0.96) 99 3.55 (0.41)

positive 55 (42.0) II (29.7) 66 9.64 (1.17) 65 4.94 (0.36)

ABO match .174 .915 .966

matched 43 (29.7) 8 (I9.I) 5I 9.37 (1.28) 49 3.70(0.51)

mismatched I02 (70.3) 34(81.0) I36 9.76 (0.88) I34 4.53 (0.34)

Diagnosis .309 .001 .014

nonmalignant 27 (18.6) 5(11.9) 32 15.64(1.88) 3I 5.24 (0.79)

malignant 118 (8I.4) 37 (88.1) I55 8.98 (0.77) I52 3.80 (0.29)

Risk status^ .855 .480 .614

standard 92 (63.5) 26 (61.9) II8 9.84 (0.83) II5 3.94 (0.37)

high risk 53 (36.6) 16 (38.1) 69 9.17 (1.37) 68 4.81 (0.45)

Degree of HLA match .701 .465 .678

I0\I0 7I (49.0) 23 (54.8) 94 9.18 (0.89) 93 3.75 (0.40)

9\I0 50 (34.5) 15(35.7) 65 9.72 (1.21) 63 4.46 (0.52)

8\I0 20 (13.8) 3(7.1) 23 9.11 (2.80) 23 4.97 (0.67)

7\I0 4 (2.8) 1 (2.4) 5 11.27 (6.82) 4 5.71 (1.03)

BFM match status .261 .525 .346

matched ($9/I0) I2I (83.5) 38 (90.5) I59 9.47 (0.72) I56 4.02 (0.32)

mismatched (#8/I0) 24 (16.6) 4 (9.5) 28 10.19 (2.58) 27 4.97 (0.59)

Stem cell source .000 .000

peripheral blood I45 11.27 (0.84) I43 5.15 (0.30)

bone marrow 42 3.43 (0.39) 40 0.40 (0.29)

Year of transplant .000 .038 .002

2000-2004 43 (62.3) 26 (37.7) 69 7.00 (1.58) 65 2.45 (0.45)

2005-2008 I02 (86.4) 16 (13.6) II8 10.06 (0.69) II8 4.94 (0.35)

TBI .317 .014 .103

TBI 37 (72.5) 14(27.4) 5I 8.63 (0.73) 5I 3.55 (0.28)

non-TBI I08 (79.4) 28 (20.6) I36 10.57 (0.93) I32 4.72 (0.36)

PB indicates peripheral blood; BM, bone marrow; CMV, cytomegalovirus; ABO, blood group; HLA, human leukocyte antigen; BFM, stem cell transplant protocol from Berlin-Frankfurt-Munster group; TBI, total body irradiation; ANC, day of absolute neutrophil count recovery >0.5 x I09/L; PLT, day of platelet count recovery >50 x I09/L.

Univariate analysis was performed by c2-test* and Mann-Whitney U/Kruskal-Wallis testf used for nonparametric variables.

^Standard risk: ALL/AML in first and second complete remission (CR), CML in first chronic phase, MDS-refractory cytopenia, NHL in CR, any nonma-lignant disease; high risk: ALL/AML in equal to or greater than third CR or nonremission, CML equal to or greater than accelerated phase, MDS-refractory anemia with excess of blasts, JMML, NHL in nonremission, any second transplant.

Statistical Analysis

For statistical comparisons between the 3 studied cohorts of patients, a chi-square test was used for categoric parameters, whereas Mann-Whitney U or Kruskal-Wallis tests were performed for nonparametric continuous variables. Kaplan-Meier/cumulative incidence methods and log-rank/Gray tests were used for comparison of time-dependent endpoint outcome parameters within the studied patient groups described earlier [20,21]. Variables were analyzed for prognostic

impact on OS, DFS, and the incidence of severe GVHD (aGVHD grade III-IV and extensive cGVHD) and included: recipient sex, sex match, donor and recipient age, donor and recipient CMV status, ABO blood group compatibility, patients' diagnosis and risk group, stem cell source (PB versus BM), infused CD341 and CD31 cell dose per kg, year of transplant (2000-2004 versus 2005-2008), conditioning regimen (TBI versus non-TBI) and degree of donor-recipient HLA match. Only variables with

a P-value <.2 in univariate analysis were included into the final multivariate analysis-proportional hazard (Cox) regression with a backward stepwise approach [22]. For aGVHD and cGVHD, death without an event, as well as graft rejection were the competing risks. Because of a strong correlation between CD341 and CD31 cell doses, two possible models of multivariate analysis for OS and DFS are presented. Furthermore, two possible models of multivariate analysis for aGVHD grade III-IV are shown, one including BFM match status (matched versus mismatched) and the other including HLA match variant (graft mismatched at 2 HLA antigens or 3 antigen/al-leles versus other). For the cell dose variables, the optimal cut point was determined by taking age-adjusted medians. Because of a well-known correlation of CD341 cell doses with age, probabilities of survival were calculated within groups of patients 0-5 years, 5-10 years, 10-15 years, and 15-20 years old, and 0-10 years and 10-20 years old. The P-values are 2-sided and a significance level of .05 was used as relevant. Statistical analyses were performed using Statistica for Windows 8.0 and R-2.10.1 statistical software package [23].

RESULTS

Study Cohorts

One hundred eighty-seven patients were prospec-tively included in the study: 145 in the PB, and 42 in the BM groups. The decision about the type of transplant material was made by the donor together with the local harvest center, and was only in part influenced by the transplant center's preference. Patient, donor, and transplant characteristics were fully comparable between the 2 cohorts except for the expected higher infused CD341 and CD31 cell dose in the PB group and the year of transplant (Table 3). Patients in the PB group were given statistically significant 3.3 times more CD341 cells/kg and 12.9 times more (>1 log) CD3 1 cells/kg when compared with the BM group (P < .001). The cohorts of patients analyzed according to infused CD341 cell dose/kg were comparable in terms of patient, donor, and transplant characteristics except for sex match, diagnosis, year of transplant, the use of TBI, and, as expected, recipient age and stem cell source (Table 3). The cohorts of patients analyzed according to infused CD31 cell dose were comparable except for diagnosis, year of transplant and, as expected, recipient age and stem cell source (Table 3). The cohorts of patients analyzed according to diagnosis (malignant versus nonmalignant) were comparable except for, as expected, the age (P < .05), CD341 and CD31 cell doses (P < .05), risk group (p < .01), and the use of TBI (P < .001) (data not shown).

Hematologic Recovery

The median day of ANC recovery (>0.5 x 109/L) was 14 for PBSC and 18 for BM recipients (P < .001). Five patients died without achieving neutrophil reconstitution on days 16, 110, 111, +15, and 126 post-transplant, respectively, because of sepsis in 3 cases and heart failure or renal insufficiency in the remaining 2 patients. The median day of platelet recovery was 18 for PBSC and 30 for BM recipients (P < .001). Time to neutrophil engraftment was significantly shorter in patients receiving higher doses of both CD341 ($10 x 106) and CD31 ($4 x 108) cells/kg (median day 14 and 16; P < .001 and P = .003, respectively). Similarly, platelet recovery in patients who received higher doses of both CD341 and CD31 cells/kg was achieved 9 and 6 days earlier, respectively (median day 16 and 18 versus 25 and 24; P < .001 and P = .002, respectively).

05 and DFS (Tables 4 and 5)

With a median follow-up of 3.2 years (range: 0.87.6 years, PB group) and 5.4 years (range: 2.3-8.8 years, BM group) there was a statistically significant advantage in DFS of patients given PBSC versus BM (0.55 6 0.04 versus 0.4 6 0.08, P < .05, log-rank, Figure 1). In terms of OS, there was a trend toward better OS in the PB group (0.56 6 0.04 versus. 0.43

6 0.08, P = 0.065, log-rank, Figure 1). As expected, patients at high risk had a poorer DFS in comparison with standard risk patients (0.41 6 0.06 versus 0.58 6 0.05, P < .05). There was no difference in OS and DFS between patients transplanted from MUDs or mismatched UDs (including separately 10/10, 9/10, 8/10, allele matched transplants, Table 4) despite the fact that patients transplanted from mismatched donors were mostly (57.1%) high risk patients according to BFM criteria, whereas the group of children transplanted from matched donors included significantly fewer high risk individuals (33.3%, P = .016, c2-test).

The group of 7/10-allele matched transplanted children was inconclusive in terms of statistics because of the low number of patients. The most striking statistically significant differences in OS and DFS were observed when comparing groups receiving higher and lower doses of both CD341 and CD31 cells in the graft. Patients receiving $10 x 106 and $15.4 x 106 (>75th percentile) CD341 cells/kg had OS (and DFS) probability of 0.64 and 0.69 (0.60 and 0.67), respectively versus 0.44 and 0.47 (0.44 and 0.46) when given less than the previously mentioned amounts. Similarly patients given $4 x 108 and $6.9 x 108 (>75th percentile) CD31 cells/kg had OS (and DFS) probability of 0.61 and 0.71 (0.60 and 0.69), respectively, versus 0.45 and 0.48 (0.43 and 0.46) when infused less than the earlier-mentioned amounts.

Table 4. Influence of Multiple Variables on Clinical Outcome by Univariate Analysis (Log-Rank* or Gray Testf)

Overall Survival* Disease-Free Survival* aGVHD Grades Ill-IVj Extensive cGVHDj

Variables N % SE P N % SE P N % SE P N % SE P

Recipient sex .658 .552 .297 .504

male 112 0.51 0.05 112 0.49 0.05 112 0.25 0.04 93 0.16 0.04

female 75 0.56 0.06 75 0.55 0.06 75 0.18 0.05 63 0.20 0.05

Recipient age .028 .050 .085 .755

$ 10 years 89 0.44 0.06 89 0.43 0.05 89 0.16 0.04 70 0.16 0.05

<10 years 98 0.62 0.05 98 0.59 0.05 98 0.27 0.05 86 0.14 0.04

Donor age .142 .202 .903 .578

$35 years 82 0.49 0.06 82 0.48 0.06 82 0.22 0.05 62 0.13 0.05

<35 years 105 0.56 0.05 105 0.54 0.05 105 0.21 0.04 94 0.16 0.04

Sex match .775 .994 .267 .542

female to male 32 0.45 0.12 32 0.48 0.11 32 0.28 0.08 28 0.11 0.07

other 155 0.54 0.04 155 0.52 0.04 155 0.20 0.03 128 0.16 0.04

Recipient CMV status .796 .742 .896 .391

negative 75 0.55 0.06 75 0.53 0.06 75 0.21 0.05 64 0.18 0.05

positive 93 0.55 0.05 93 0.52 0.05 93 0.22 0.04 80 0.13 0.04

Donor CMV status .541 .486 .107 .780

negative 102 0.54 0.05 102 0.51 0.05 102 0.25 0.04 86 0.15 0.04

positive 66 0.56 0.07 66 0.55 0.07 66 0.15 0.05 58 0.14 0.05

ABO match .774 .965 .260 .201

matched 51 0.50 0.07 51 0.50 0.07 51 0.27 0.06 44 0.21 0.07

mismatched 136 0.54 0.05 136 0.52 0.04 136 0.19 0.04 112 0.13 0.04

Diagnosis .273 .165 .172 .240

nonmalignant 32 0.66 0.08 32 0.66 0.08 32 0.12 0.06 27 0.08 0.06

malignant 155 0.50 0.04 155 0.48 0.04 155 0.23 0.03 129 0.16 0.04

Riskgroup^ .082 .046 .575 .820

standard 118 0.59 0.05 118 0.58 0.05 118 0.20 0.04 99 0.15 0.04

high risk 69 0.43 0.06 69 0.41 0.06 69 0.23 0.05 57 0.14 0.05

Year of transplant .184 .214 .425 .372

2000-2004 69 0.50 0.06 69 0.46 0.06 69 0.25 0.05 55 0.18 0.06

2005-2008 118 0.60 0.05 118 0.54 0.05 118 0.19 0.04 101 0.13 0.04

TBI .858 .850 .330 .461

TBI 51 0.50 0.04 51 0.51 0.04 51 0.25 0.06 44 0.11 0.03

non-TBI 136 0.50 0.04 136 0.52 0.04 136 0.20 0.04 112 0.16 0.04

Degree of HLA-match .628 .228 .055 .720

I0\I0 94 0.54 0.06 94 0.55 0.05 94 0.16 0.04 78 0.22 0.04

9\I0 65 0.51 0.06 65 0.48 0.06 65 0.23 0.05 56 0.16 0.05

8\I0 23 0.56 0.10 23 0.52 0.10 23 0.30 0.10 19 0.22 0.11

7\I0 5 0.40 0.22 5 0.20 0.18 5 0.60 0.22 3 no ext cGVHD

BFM HLA-match status .692 .334 .039 .645

matched ($9/I0) 159 0.53 0.04 159 0.52 0.04 159 0.19 0.03 134 0.14 0.03

mismatched (#8/I0) 28 0.53 0.10 28 0.46 0.09 28 0.36 0.09 22 0.19 0.10

HLA-match variants .189 .024 .008 .439

8/I0ant + 7/I0 12 0.42 0.14 12 0.25 0.13 12 0.50 0.15 8 0.25 0.20

other 175 0.54 0.04 175 0.53 0.04 175 0.19 0.03 148 0.14 0.03 (Continued)

w «0 w

u> «o ■u

ai wak

Table 4. (Continued) J1

Overall Survival* Disease-Free Survival* aGVHD Grades iii-IVf Extensive cGVHDf .

Variables N % SE P N % SE P N % SE P N % SE

Stem cell source .065 .047 .922 .275

peripheral blood l45 0.56 0.04 l45 0.55 0.04 l45 0.2l 0.04 l27 0.l4 0.04

bone marrow 42 0.43 0.08 42 0.40 0.08 42 0.2l 0.07 29 0.2l 0.08

Infused CD34+ cell dose .009 .018 .889 .780

$10 x l06/kg 86 0.64 0.05 86 0.60 0.05 86 0.22 0.05 77 0.l4 0.05

<10 x l06/kg l0l 0.44 0.05 l0l 0.44 0.05 l0l 0.2l 0.04 79 0.l5 0.05

CD34+ dose per kg .030 .038 0.93 .842

Ql 47 0.4l 0.07 47 0.4l 0.07 47 0.l9 0.06 34 0.l8 0.07

Q2 47 0.43 0.08 47 0.43 0.08 47 0.23 0.06 39 0.l3 0.06

Q3 47 0.59 0.08 47 0.55 0.08 47 0.23 0.06 4l 0.l8 0.07

Q4 46 0.69 0.07 46 0.67 0.07 47 0.20 0.06 42 0.l2 0.05

Infused CD3+ cell dose .016 .012 .609 .981

$4 x l08/kg 96 0.6l 0.06 96 0.60 0.05 96 0.20 0.04 83 0.l5 0.04

<4 x l08/kg 87 0.45 0.05 87 0.43 0.05 87 0.23 0.05 7l 0.l4 0.05

CD3+ dose per kg .062 .035 .888 .632

Ql 46 0.4l 0.07 46 0.39 0.07 46 0.22 0.06 34 0.l2 0.07

Q2 46 0.52 0.08 46 0.49 0.08 46 0.22 0.06 4l 0.l5 0.06

Q3 46 0.50 0.09 46 0.49 0.09 46 0.l7 0.06 40 0.20 0.07

Q4 45 0.7l 0.07 45 0.68 0.07 45 0.24 0.06 39 0.l0 0.05

aGVHD indicates acute graft-versus-host disease; cGVHD, chronic graft-versus-host-disease; HLA, human leukocyte antigen; TBI, total body irradiation; BFM, stem cell transplant protocol from Berlin-FrankfurtMunster group; 8/10 ant, two HLA-antigen mismatch; ext, extensive; Q, quartiles of CD34+ dose per kg: Ql <5.33, Q2 5.33-9.72, Q3 9.73-15.42, Q4 >15.42; quartiles of CD3+ dose per kg: Ql <1.71, Q2 1.714.26, Q3 4.27-6.85, Q4 >6.85.

^Standard risk: ALL/AML in first and second complete remission, CML in first chronic phase, MDS-refractory cytopenia, NHL in complete remission, any nonmalignant disease; high risk: ALL/AML in equal to or greater than third complete remission or nonremission, CML equal to or greater than accelerated phase, MDS-refractory anemia with excess of blasts, JMML, NHL in nonremission, any second transplant. *Log-rank test. fGray test.

Ui 00 00

Table 5. Survival Analyses in Age Groups According to Age-Adjusted Medians of Infused CD34+ and CD3+ Cell Doses per kg (LogRank)

Infused CD34+ Cell Dose/kg

Infused CD3+ Cell Dose/kg

<Me for Age Group

$Me for Age Group

< Me for Age Group

$Me for Age Group

Survival Age Group N % SE N % SE P N % SE N % SE P

OS 0-5 23 0.47 0.11 24 0.83 0.08 .006 23 0.61 0.10 23 0.74 0.09 .364

5-10 26 0.50 0.11 25 0.62 0.10 .560 26 0.55 0.10 25 0.56 0.11 .965

10-15 25 0.46 0.10 25 0.59 0.10 .421 25 0.43 0.10 24 0.59 0.11 .290

15-20 20 0.15 0.08 19 0.58 0.13 .001 19 0.31 0.11 18 0.44 0.14 .225

DFS 0-5 23 0.38 0.10 24 0.83 0.08 .001 23 0.52 0.10 23 0.74 0.09 .150

5-10 26 0.52 0.10 25 0.58 0.10 .703 26 0.56 0.10 25 0.54 0.10 .774

10-15 25 0.46 0.10 25 0.59 0.10 .435 25 0.43 0.10 24 0.60 0.11 .276

15-20 20 0.15 0.08 19 0.54 0.13 .002 19 0.31 0.11 18 0.40 0.13 .344

aGvHD grade I II-IV 0-5 23 0.31 0.10 24 0.21 0.08 .359 19 0.06 0.05 21 0.10 0.07 .898

5-10 25 0.36 0.10 24 0.22 0.09 .218 25 0.33 0.10 20 0.11 0.07 .303

10-15 24 0.13 0.07 25 0.12 0.06 .944 22 0.14 0.07 21 0.16 0.09 .391

15-20 19 0.32 0.11 18 0.12 0.08 .093 12 0.36 0.15 14 0.16 0.10 .669

ext cGvHD 0-5 17 0.07 0.06 24 0.13 0.07 .538 19 0.06 0.05 21 0.10 0.07 .618

5-10 24 0.29 0.09 21 0.15 0.08 .236 25 0.33 0.10 20 0.11 0.07 .076

10-15 21 0.20 0.09 23 0.11 0.08 .297 22 0.14 0.07 21 0.16 0.09 .964

15-20 10 0.33 0.16 16 0.20 0.10 .308 12 0.36 0.15 14 0.16 0.10 .258

Me indicates median for age groups (CD34+ cell dose: 0-5 years, I6.02 x I06; 5-I0 years, II.68 x I06; I0-I5 years, 8.77 x I06; I5-20 years, 6.4I x I06; CD3+cell dose: 0-5 years, 6.96 x I08; 5-I0years, 5.I5 x I08; I0-I5 years, 3.60 x I08; I5-20 years, 2.I4 x I08); OS, overall survival; DFS, disease-free survival; aGVHD, acute graft-versus-host disease; ext cGVHD, extensive chronic graft-versus-host disease.

In addition, a subanalysis of different CD341 and CD31 doses using quartiles revealed the best OS and DFS rates in the fourth quartile and the worst in the first one. Furthermore, we analyzed the impact of CD341 and CD31 cell doses (higher and lower than age-adjusted medians) on OS and DFS in smaller age cohorts, showing the survival advantage when using higher doses of CD341 cells/kg (Table 5 and Figure 2). In age groups 0-10 years and 10-20 years old, the survival advantage of patients who received higher doses of CD341 (Figure 2) and CD31 cells (data not shown) was clearly statistically significant.

These clinically important results were confirmed in a multivariate analysis showing that both infused CD341 and CD31 cell doses lower than 10 x 106/kg

1.0 0.9 0.8

and 4 x 108/kg, respectively, were significant, independent risk factors for mortality (relative risk, RR 1.8, 95% confidence interval [CI] 1.2-2.8, P = .009, and RR 1.7, CI 1.1-2.7, respectively; Table 6 and Figure 3A). Regarding DFS, the multivariate analysis has revealed another independent risk factor for poor outcome apart from the 2 just-mentioned grafted cell doses, which was the use of donors mismatched at 2 HLA antigens or 3 HLA allele/antigens (Table 6 and Figure 3B).

With a median follow-up of 3.5 years (range: 0.8-8.8 years), 103 patients (55.1%) remain alive. Eighty-four patients died because of relapse or early/

0.5 0.4 0.3 0.2 0.1 0.0

[ PB (N=145): - OS: 0.56 ±0.04

H ---DFS: 0.55 ± 0.04

i\ BM (N=42): ......... OS: 0.43+ 0.08

.......... DFS: 0.40 ± 0.08

OS: p=0.0651

DFS: p=0.0472

Figure 1. Probability of OS and DFS according to stem cell source. DFS and OS are depicted for the PB group (upper lines) and BM group (lower

lines). N, number of patients in each arm of the study cohort, P (logrank).

Figure 2. Probability of OS according to infused age-adjusted median dose of CD34 cells/kg. Dotted and short-dashed lines depict patients younger than I0 years given CD341 cell dose below and above age-adjusted median (I2.I3 x I06 CD341 cells/kg), whereas dashed and solid lines depict patients older than I0 years given CD341 cell dose above and below age-adjusted median (7.2 x I06 cells/kg). N, number of patients in each arm of the study cohort, P (log-rank).

Table 6. Multivariate Analysis of Risk Factors for Overall-, Disease-Free Survival, and the Incidence of aGVHD Grades III-IV

Clinical Outcome Variable* Relative Risk 95% CI P (Partial) P (Model)

Overall survival I or II I CD34+ cell dose (<10 x l06/kg) 1.80 1.15 2.81 .010 .009

Disease-free survival CD3+ cell dose (<4 x l08/kg) 8/l0ant + 7/10 CD34+ cell dose (<10 x l06/kg) 1.71 2.54 1.68 1.10 1.27 1.09 2.67 5.08 2.59 .017 .008 .019 .016 .004

Incidence of acute GVHD grades III-IV or II I or II 8/l0ant + 7/10 CD3+ cell dose (<4 x l08/kg) BFM match status (8/l0 + 7/l0) 8/l0ant + 7/l0 2.44 1.76 2.19 3.72 1.18 1.14 1.07 1.56 5.07 2.71 4.48 8.90 .017 .010 .032 .003 .005 .046 .011

CI indicates confidence interval; GVHD, graft-versus-host disease

*Variables analyzed for prognostic impact on survival and incidence of aGVHD grade III-IV: recipient sex, sex match, donor and recipient age, donor and recipient cytomegalovirus status, ABO blood group compatibility match, patients' diagnosis and risk group, stem cell source (peripheral blood versus bone marrow), infused CD34+ and CD3+ cell dose per kg, donor-recipient HLA matching: degree of HLA-match or BFM match status (matched versus mismatched) or 8/10 ant (2 antigen mismatch) + 7/10 versus other. Only variables with a P-value < .2 in univariate analysis (log-ranktest) were included in the final multivariate analysis: proportional hazard (Cox) regression with a backward stepwise approach.

late treatment-related mortality (TRM). TRM until day +100 was 16.0% and overall TRM reached 31.0%. The reasons for death were: relapse n = 21 (11.2%), secondary malignancies n = 2 (1.1%), sepsis n = 14 (7.5%), other infection (viral) n = 7 (3.7%), aGVHD n = 12 (6.4%), cGVHD n = 9 (4.8%), thrombotic thrombocytopenic purpura (TTP) n = 6 (3.2%), invasive fungal infection n = 3 (1.6%), pulmonary embolism n = 3 (1.6%), veno-occlusive liver disease n = 2 (1.1%), brain infarct n = 2 (1.1%), aorta thrombosis n = 1 (0.5%), heart failure n = 1 (0.5%) and renal insufficiency n = 1 (0.5%). There were no statistically significant differences in the TRM causes between the studied patients' cohorts. Year of transplant (2000-2004 versus 2005-2008) influenced neither OS, DFS, or the incidence of severe GVHD.

Severe aGVHD Grade III-IV (Table 4)

An analysis of the incidence of moderate-to-severe aGVHD grade II-IV (59.9%) was not the aim of this study, because moderate aGVHD grade II no longer seems to be a problem in treatment when compared with severe aGVHD grade III-IV. In a recently published large pediatric study, there was absolutely no difference in TRM between children developing aGVHD grade II and those with none or grade I aGVHD [24].

The cumulative incidence of clinically important severe aGVHD grade III-IV was comparable between the recipients of PB versus BM (Figure 4A). As shown in Tables 4 and 5, infused CD34+ or CD3+ cell dose had absolutely no impact on the occurrence of severe aGVHD (Figure 4B and C). As expected, the more HLA mismatches were in the donor-recipient pairs, the more frequent the cumulative incidence of severe aGVHD was (Figure 5A). There was a significantly higher incidence of severe aGVHD in the mismatched group (0.36 6 0.09) versus matched group (0.19 6 0.03, P = .039, log-rank) when using ALL international

BFM SCT 2008 study criteria. However, this did not adversely affect OS or DFS, as shown previously (Table 4). Within the mismatched group the highest risk of severe aGVHD was associated with the use of donor-recipient pairs mismatched at 2 HLA antigens (not 2 alleles or 1 antigen + 1 allele) and 3 HLA allele/antigens (Figure 5B). These clinically relevant results were confirmed in a multivariate analysis showing that either the use of mismatched donors according to BFM criteria (RR 2.2, 95% CI 1.1-4.5, P = .046, Table 6) or, in another model, the use of donors mismatched at 2 HLA antigens or 3 HLA allele/antigens (RR 3.7, 95% CI 1.6-8.9, P = .011) were significant, independent risk factors for incidence of severe aGVHD.

cGVHD (Table 4)

The cumulative incidence of clinically relevant extensive cGVHD was comparable between the recipients of PB and BM (0.14 6 0.04 for PB versus 0.21 6 0.08 for BM groups, Figure 6A). As shown in Tables 4 and 5, infused CD34+ or CD3+ cell dose had no impact on the occurrence of extensive cGVHD (Figure 6B and C). There was a visible trend toward a higher cumulative incidence of extensive cGVHD in patients whose donors were mismatched at 2 HLA antigens (not 2 alleles or 1 antigen + 1 allele) and 3 HLA allele/antigens (P = .074). However, the multi-variate analysis disclosed no significant risk factors for the development of extensive cGVHD.

DISCUSSION

There is a worldwide concern whether to limit the dose of infused CD341 and CD31 cells/kg in recipients of both related and unrelated donor grafts. It is well recognized that doses of different cellular subsets within grafts can affect the clinical outcome following

Figure 3. independent risk factors associated with probability of OS and DFS. (A) Dashed line depicts lower OS in patients grafted with <I0 x I06 CD341 cells/kg, whereas solid line shows better survival in children who are free from the above-mentioned independent risk factor. (B) Dashed line depicts lower probability of DFS in patients with both independent risk factors: CD341 cell dose per kg <I0 x I06 and the graft mismatched at 2 HLA antigens or 3 antigen/alleles, whereas solid line presents patients without the above-mentioned risk factors. RR, relative risk, CI, confidence interval, P (hazard Cox regression model).

allogeneic HSCT. In MA setting in adults, a clear association between higher CD341 cell dose and improved survival after HLA-identical BMT has been reported [25]. The impact of higher CD341 (and presumably CD31) cell dose in the graft on the risk of cGVHD is controversial [26], as some studies have reported only distinct correlations [27] or none at all [16,28].

There are only a few studies analyzing the impact of cell dose in the graft on the outcome of UD PBSCT. In a large single-center study Nakamura et al. [13] showed a correlation between higher CD341 cell doses and faster lymphocyte recovery leading to a decrease in relapse rate. A recently published large multicenter prospective study clearly showed that NMDP-facilitated UD PBSCTs [16] result in rapid engraftment and survival comparable to previous studies on UD BMT [29]. The key finding of that study was

O (0 .b

c <D 4

£ -j 1

I 9 i 8 & 7

o .7 6 5 4 3 2 1 0

— BM: 0.21 ±0.07: N=42

— PB: 0,21 ± 0.04: N=145

10 20 30 40 50 60 70 80 90 100 Days

CD34+ cells/kg — Q1: 0.19±0.06 --- Q2: 0.23±0.06

..... Q3: 0.23 ±0.06

- - Q4: 0.20 ±0.06

N=47 N=47 N=47 N=47

p=.9333

10 20 30 40 50 60 70 80 90 100 Days

CD3+ cells/kg -Q1: 0.22 ± 0.0b; N=46

--- Q2: 0.22 ±0.06; N=46

..... Q3: 0.17 ±0.06; N=46

- - Q4: 0.24 ±0.06; N=45

i—i-—1 J" " _P--J

...... p=.8881

10 20 30 40 50 60 70 80 90 100 Days

Figure 4. Incidence of severe aGVHD grade III-IV. (A) Comparison between PB (solid line) and BM recipients (dotted line). (B) Comparison between patients grafted with Ql (solid line), Q2 (short-dashed line), Q3 (dotted line), and Q4 (long-dashed line); Q, quartile dose of CD341 cells/kg. (C) Comparison between patients grafted with Ql (solid line), Q2 (short-dashed line), Q3 (dotted line), and Q4 (long-dashed line); Q, quartile dose of CD3 cells/kg. N, number of patients in each group of patients; quartiles of CD341 dose per kg x l06: Ql <5.33, Q2 5.33-9.72, Q3 9.73-15.42, Q4 >15.42; quartiles of CD31 dose per kg x l08: Ql <1.71, Q2 1.71-4.26, Q3 4.27-6.85, Q4 >6.85, P (log-rank, cumulative incidence).

the independent predictive value of the higher CD341 cell dose for significant improvement in patients' outcome including faster hematologic recovery and better OS [16]. CD341 cell doses between 4.5 and 9.5 x 106/kg

1.0 .9 .8 .7 .6 .5 .4 .3 .2 .1 .0

1 -10/10: 0.16 ± 0.04; N=94 ------ 9/10: 0.23 ±0.05; N-65 ....... 8/10. 0.30 ± 0.10. N-23 ---7/10: 0.60 ± 0.22; N=5

1 . >■< ..--'................:

p=.0554

10 20 30 40 50 60 70 80 90 100 Days

.9 .8 .7 .6 .5 .4 .3 .2 .1 .0

-----8/10 ant + 7/10: 0.50 ± 0.15: N=12

-other: 0.19 ±0.03: N=175

p=.0078

10 20 30 40 50 60 70 80 90 100 Days

Figure 5. Incidence of severe aGVHD grade III-IV, N, number of patients in each arm of the study cohort. (A) Incidence of severe aGVHD according to degree of HLA match, 10/10, full allele match; 9/10, single HLA allele/antigen mismatch; 8/10, 2 HLA allele/antigen mismatches; 7/ 10, three HLA allele/antigen mismatches. (B) Incidence of severe aGVHD in patients transplanted from donors mismatched (dotted line) at 2 HLA antigens (8/10 ant) or 3 HLA allele/antigens (7/10) versus others, P (log-rank).

recipient weight resulted in a significant 12% improvement in 3-year OS in recipients of MA conditioning [16]. Doses greater than 9.5 x 106/kg did not further improve OS in the earlier mentioned study [16], which stands in contrast to our data, showing best OS and DFS in patients receiving CD341 cell doses above the 75th percentile ($15.4 x 106/kg), but it should be noted that the highest CD341 cell doses in our study were given mostly to younger patients. These patients generally have better prognosis, but there were still a few high-risk patients in the youngest cohort including children after previous unsuccessful haploidentical transplants. Another confirmation of our results may be found in the most recently published study by Collins et al. [17], showing higher survival rates in patients grafted with PBSCs from UD containing >5 x 106 CD341 cells/kg. There is, however, one important point that cannot be missed: all previously cited studies included patients who were not given in vivo T cell depletion as

« r .7

<U ,t>

......BM 0.21 ± 0.08; N=29

-PB; 0.14 ± 0.04; N=127

p=.2755

'i/l r .7

■11 .4

1.0 .9 .8 .7 .6 ,5 .4 .3 .2 1 .0

3 4 5 6 Years

■ CD34<10x10 0.15 ± 0.05; N=79

■ CD34>10x10s: 0.14 ± 0.05; N=77

p=.7795

4 5 6 Years

-----CD3<4x10s 0.14 ± 0.05; N=71

-CD3>4x10s 0.15 ±0.04; N=83

J p=.9812

4 5 6 7 Years

Figure 6. Incidence of extensive cGVHD, N, number of patients in each arm of the study cohort. (A) Comparison between PB (solid line) and BM recipients (dotted line). (B) Comparison between patients grafted with $10 x 106 CD341 cells/kg (solid line) and <10 x 106 CD341 cells/kg (dotted line). (C) Comparison between patients grafted with $4 x 108 CD31 cells/kg (solid line) and <4 x 108 CD31 cells/kg (dotted line), P (log-rank).

GVHD prophylaxis. On the contrary, our study is based on patients who received rabbit ATG or CAM-PATH. The dose of rabbit ATG we used was not high: the majority of our patients were given ATG-Fresenius at a dose of 30 mg/kg, which is lower than the dose recommended by BFM group (60 mg/kg) or Sedlacek et al. [2,3] (40 mg/kg). Another point is

the use of rabbit ATG (Thymoglobuline). Here, the recommended dose is around 7.5 mg/kg, and our dosage of Thymoglobuline in those few patients was similar or slightly higher (10 mg/kg). The majority of children, however, were given the standard (rather low) dose of ATG-Fresenius (30 mg/kg).

In our study, we demonstrated a higher survival rate among the patients who had received higher doses: the more CD341 cells were given, the better survival rates were achieved. Of course, the just-mentioned result is slightly biased by the fact that younger children with more chances to receive higher dose of CD341 cells/kg usually had a better outcome. To avoid this possible bias, we analyzed the age groups 0-10 years and 10-20 years, and 0-5years, 5-10 years, 10-15 years, and older separately, looking at the impact of high and low doses in different age cohorts. Our age-matched analysis (Table 5) proved that the megadose concept was working within almost all age cohorts in terms of survival. Because of a lower number of cases, in age groups 5-10 years and 10-15 years, the survival advantage was not statistically significant, yet visible. When looking at the larger age groups of 0-10 years and 10-20 years old, the survival advantage of patients who received higher doses was clearly statistically significant (Figure 2). We analyzed lower dose cutoffs to find a cell amount that is more achievable for adolescents and adults by looking at the median values for different age groups (Table 5 and Figure 2). For the age group of 0-10 years, the dose over 12 x 106 CD341 cells/kg favored outcome, whereas in older children >10 years old the dose >7 x 106 CD34/kg promoted survival. Those cell doses are achievable in pediatric patients when PBSC is used. The dose of 5 x 106 CD34+/kg obtainable in adult studies [16,17] might be equivocal to the dose of 10 x 106 CD34/kg in children. In conclusion, we think that the easily achievable dose of 10-12 x 106 CD34/kg for children <10 years or 7 x 106 for older children may be desired in pediatric patients undergoing unrelated donor PBSCT (provided that in vivo T cell depletion is used). Multicenter randomized studies are required, however, to further prove this concept.

One may ask whether the year of transplant or diagnosis influenced our results. Indeed, more BM than PBSC transplants were performed in the years 2000-2004 (Table 3); however, there was absolutely no difference in outcomes between the studied periods (Table 4). The major reason for this might be a well-standardized "high-resolution" HLA typing used in our center since the beginning and a uniform use of in vivo T cell depletion.

Despite some differences between patients with nonmalignant and malignant disorders, the majority of variables in both groups were comparable and those 2 groups can be included together. In the nonmalignant

group, we had both older patients at risk of transplant-related complications (children with SAA, Fanconi anemia, or metabolic disorders) and some very young patients (severe combined immunodeficiency [SCID]), who had their second transplants after previous early haploidentical graft rejection or late graft failure. Patients with nonmalignant diseases tended to be younger, but had the same HLA matching levels accepted, the same GVHD prophylaxis and the same use of BM versus PBSC.

The major concern when using higher doses of CD341 and CD31 cells is the increased risk of GVHD. We have found absolutely no correlation between the CD341 and CD31 cell dose and the incidence of severe aGVHD and cGVHD, which remains in line with the NMDP studies [16,17]. In an adult study on UD HSCT using reduced-intensity conditioning (RIC) patients receiving a CD341 cell dose $17 x 106 per kg had a higher incidence of GVHD [14]. In our study, such a high dose of grafted CD341 cells ($15.4 x 106/kg, ie, >75th percentile) promoted a significantly better OS and DFS, and did not influence GVHD. The majority of our patients (except for patients with SAA or Fanconi anemia) were conditioned, however, with an MA regimen, which may make an important difference and they were children, not adults (median age in Remberger et al. [14] study was 50 years). It was not the first time that PBSC appeared to have similar cGVHD outcomes to BM in children: Meisel et al. [12] showed comparable risk of both aGVHD and cGVHD in a small pediatric cohort of PBSC and BM recipients. Most patients in that study received ATG/ALG serotherapy as GVHD prophylaxis. All children in our study received ATG or CAMPATH as GVHD prophylaxis, and beside the well-known lower risk for GVHD in children it appeared to be the major factor for comparable rates of cGVHD among PBSC and BM recipients.

To the best of our knowledge, there are virtually no studies on the effect of cell dose on the outcome in pediatric patients undergoing UD HSCT. Furthermore, little is known about the impact of CD31 cell dose on OS, DFS, and the incidence of GVHD even in adult UD transplant patients.

In our opinion, the use of ATG is the most important factor enabling successful use of UD PBSCT in children. The results of our study, showing improved OS and DFS and no difference in severe GVHD in patients receiving higher CD341 and CD31 cell doses, make PBSC transplants with high CD341 and CD31 cell doses a feasible and perhaps desirable option. The dose above our 25th percentile, that is, 5 x 106 CD34/kg may be easily achieved for any children when using PBSC. Even in an adult study of Remberger et al. [14] median CD341 cell dose in UD PBSC was 7.5 x 106/kg. Our study results remain in

line with 2 very recent NMDP papers, confirming the beneficial effects of higher CD341 cell dose in unrelated donor transplants in prevalent adult populations [16,17]. Our data comes from a pediatric cohort of patients and cannot be easily compared with adult data, but the message remains the same. It is important to note that PBSCs containing 1 log more T cells and higher amounts of CD341 cells are not worse (if not better) for the outcome. Furthermore, there seems to be no need to drastically limit the infused dose of CD341 or CD31 cells/kg in case a transplant center is given high amounts of CD341 cells from an unrelated good mobilizer.

HLA disparity is one of the most important factors governing the severity of both aGVHD and cGVHD [29,30]. An important finding of our study is the similar outcome of transplants using HLA allele mismatched (7-8/10) and matched (9-10/10) UDs (according to BFM criteria). Our concluding remarks reflect the results of the Sedlacek study [2,3]. The use of in vivo T cell depletion is certainly a prerequisite in GVHD prophylaxis, as shown by Sedlacek et al. [2].

In summary, we have demonstrated for the first time in a large pediatric cohort that higher doses of transplanted CD341 and CD31 cells lead to improved survival without an increased risk of severe aGVHD or cGVHD. The study findings may be limited to the population of patients receiving in vivo T cell depletion, which is now broadly used in unrelated donor settings in Europe. With regard to the degree of HLA match, it seems that transplants from 8 of 10 allele matched donors (with maximum 1 mismatched antigen) do not influence OS, whereas one should avoid performing transplants from 2 HLA-antigens or 3 allele/antigen mismatched donors.

ACKNOWLEDGMENTS

Fiwawcia/ disc/oiMre: This work was supported by a national grant of the Ministry for Science and Higher Education (Grant N406 063 31/2352) and a grant No 1491 from Wroclaw Medical University. The authors thank MEDIGEN Warsaw (Monika Sankowska and Leszek Kauc, PhD) for highest quality HLA typing, Dorota Noworolska-Sauren, Renata Ryczan, and Blanka Rybka for excellent lab work, the nursing staff of the clinic for the continuous patient care, and Maja Kowalski and Bogumil Ucherek for editorial help.

AUTHORSHIP STATEMENT

Contribution: K.K. had primary responsibility for study design, data analysis, data interpretation, and manuscript writing, and had primary responsibility for the entire article as an accurate and verifiable report.

J.P. had primary responsibility for data collection, data file preparation, data analysis, and manuscript writing; M.M. and D.T. had primary responsibility for data analysis and interpretation, including statistical analysis and for manuscript writing; for example, had primary responsibility for provision of study patients, data collection, data analysis, and manuscript writing; J.O-L participated in extensive data collection, data file preparation, and data analysis; M.U., A.D., J.M., and D.P. participated in data collection, data analysis, and interpretation and manuscript writing, A.C. had responsibility for study design and for the entire article as an accurate and verifiable report.

REFERENCES

1. Miano M, Labopin M, Hartmann O, et al. Haematopoietic stem cell transplantation trends in children over the last three decades: a survey by the paediatric diseases working party of the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant. 2007;39:89-99.

2. Sedlacek P, Formankova R, Keslova P, et al. Low mortality of children undergoing hematopoietic stem cell transplantation from 7 to 8/10 human leukocyte antigen allele-matched unrelated donors with the use of antithymocyte globulin. Bone Marrow Transplant. 2006;38:745-750.

3. Sedlacek P, Mejstrikova E, Formankova R, et al. Allo-SCT in children with high-risk leukemia using unmanipulated grafts from alternative donors. Bone Marrow Transplant. 2008;42: S10-S15.

4. Bensinger WI, Clift R, Martin P, et al. Allogeneic peripheral blood progenitor cell transplantation in patients with advanced hematologic malignancies: a retrospective comparison with marrow transplantation. Blood. 1996;88:2794-2800.

5. Ottinger HD, Beelen DW, Scheulen B, Schaefer UW, GrosseWilde H. Improved immune reconstitution after allotransplantation of peripheral blood stem cells instead of bone marrow. Blood. 1996;88:2775-2779.

6. Stem cell Trialists' Collaborative Group. Allogeneic peripheral blood stem cell compared bone marrow transplantation in the management of hematologic malignancies: an individual patient data meta-analysis of9 randomized trials. J Cl/'n Oncol. 2005;23: 5074-5087.

7. Ringden O, Remberger M, Runde V, et al. Peripheral blood stem cell transplantation from unrelated donors; a comparison with marrow transplantation. Blood. 1999;94:455-464.

8. Remberger M, Ringden O, Blau IW, et al. No difference in graft-versus-host disease, relapse, and survival comparing peripheral stem cells to bone marrow using unrelated donors. Blood. 2001;98:1739-1745.

9. Remberger M, Beelen DW, Fauser A, Basara N, Basu O, Ringden O. Increased risk of extensive chronic graft-versus-host disease after allogeneic peripheral blood progenitor cell transplantation using unrelated donors. Blood. 2005;105: 548-551.

10. Elmaagacli A, Basoglu S, Peceny R, et al. Improved disease-free-survival after transplantation of peripheral blood stem cells as compared with bone marrow from HLA-identical unrelated donors in patients with first chronic phase chronic myeloid leukemia. Blood. 2002;99:1130-1135.

11. Garderet L, Labopin M, Gorin NC, et al. Patients with acute lymphoblastic leukemia allografted with a matched unrelated donor may have a lower survival with a peripheral blood stem cell graft compared to bone marrow. Bone Marrow Transplant. 2003;31:23-29.

12. Meisel R, Laws H, Balzer S, et al. Comparable long-term survival after bone marrow versus peripheral blood progenitor

cell transplantation from matched unrelated donors in children with hematologic malignancies. Biol Blood Marrow Transplant. 2007;13:1338-1345.

13. Nakamura R, Auayporn N, Smith DD, et al. Impact of graft cell dose on transplant outcomes following unrelated donor alloge-neic peripheral blood stem cell transplantation: Higher CD34+ cell doses are associated with decreased relapse rates. Biol Blood Marrow Transplant. 2008;14:449-457.

14. Remberger M, Mattson J, Hassan Z, et al. Risk factors for acute graft-versus-host disease grade II-IV after reduced intensity conditioning allogeneic stem cell transplantation with unrelated donors: a single centre study. Bone Marrow Transplant. 2008;41: 399-405.

15. Baron F, Maris MB, Storer BE, et al. High doses oftransplanted CD34+ cells are associated with rapid T-cell engraftment and lessened risk of graft rejection, but not more graft-versus-host disease after nonmyeloablative conditioning and unrelated hematopoietic cell transplantation. Leukemia. 2005;19:822-828.

16. Pulsipher MA, Chitphakdithai P, Logan BR, et al. Donor, recipient, and transplant characteristics as risk factors after unrelated donor PBSC transplantation: beneficial effects of higher CD34+ cell dose. Blood. 2009;114:2606-2616.

17. Collins NH, Gee AP, Durett AG, et al. The effect of the composition of unrelated donor bone marrow and peripheral blood progenitor cell grafts on transplantation outcomes. Biol Blood Marrow Transplant. 2010;16:253-262.

18. Jacobsohn DA. Acute graft-versus-host disease in children. Bone Marrow Transplant. 2008;41:215-221.

19. Akpek G. Clinical grading in chronic graft-versus-host disease: is it time for change? Leuk Lymphoma. 2002;43:1211-1220.

20. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1858;53:457-481.

21. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1141-1154.

22. Cox DR. Regression models and life tables. JRStatSoc. 1972;34: 187-220.

23. Scrucca L, Santucci A, Aversa F. Regression modeling of competing risk using R: an in depth guide for clinicians. Bone Marrow Transplant. 2010 Jan 11 [Epub ahead of print].

24. Davies SM, Wang T, Wang T, et al. Recent decrease in acute graft-versus-host disease in children with leukemia receiving unrelated donor bone marrow transplants. Biol Blood Marrow Transplant. 2009;15:360-366.

25. Heimfeld S. HLA-identical stem cell transplantation: is there an optimal CD34 cell dose? Bone Marrow Transplant. 2003;31: 839-845.

26. Zaucha JM, Gooley T, Bensinger WI, et al. CD34 cell dose in granulocyte colony-stimulating factor-mobilized peripheral blood mononuclear cell grafts affects engraftment kinetics and development of extensive chronic graft-versus-host disease after human leukocyte antigen-identical sibling transplantation. Blood. 2001;98:3221-3227.

27. Przepiorka D, Anderlini P, Saliba R, et al. Chronic graft-versus host disease after allogeneic blood stem cell transplantation. Blood. 2001;98:1695-1700.

28. Cao TM, Wong RM, Sheehan K, et al. CD34, CD4, and CD8 cell doses do not influence engraftment, graft-versus-host disease, or survival following myeloablative human leukocyte antigen-identical peripheral blood allografting for hematologic malignancies. Exp Hematol. 2005;33:279-285.

29. 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;328:593-602.

30. Flomenberg N, Baxter-Lowe LA, Confer D, et al. Impact of HLA class I and class II high resolution matching on unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood. 2004;104:1923-1930.