ORIGINAL ARTICLE. Patients A total of 327 patients who received hematopoietic stem cell grafts from volunteer URDs were included in this analysis.

Transcription

1 (211), & 211 Macmillan Publishers Limited All rights reserved /11 ORIGINAL ARTICLE Predictive impact of allele-matching and EBMT risk score for outcome after T-cell depleted unrelated donor transplantation in poor-risk acute leukemia and myelodysplasia T Lodewyck 1, M Oudshoorn 2, B van der Holt, E Petersen, E Spierings 5, PA von dem Borne 6, A Schattenberg 7, W Allebes 8, M Groenendijk-Sijnke 9, L Duinhouwer 1, R Willemze 6, B Lowenberg 1, LF Verdonck 1, E Meijer 1 and JJ Cornelissen 1 1 Department of Hematology, Erasmus University Medical Center, Rotterdam, The Netherlands; 2 Europdonor Foundation and Department of Immunohematology & Blood Transfusion Service, Leiden University Medical Center, Leiden, The Netherlands; Trials & StatisticsFHOVON Data Center, Erasmus University Medical Center/Daniel den Hoed, Rotterdam, The Netherlands; Department of Hematology, University Medical Center Utrecht, Utrecht, The Netherlands; 5 Department of Immunology, University Medical Center Utrecht, Utrecht, The Netherlands; 6 Department of Hematology, Leiden University Medical Center, Leiden, The Netherlands; 7 Department of Hematology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; 8 Laboratory for Medical Immunology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands; 9 HOVON Data Center, Erasmus University Medical Center, Rotterdam, The Netherlands and 1 Department of Internal Medicine, Isala Clinics, Zwolle, The Netherlands Many parameters predict for outcome after unrelated donor (URD) allogeneic hematopoietic stem cell transplantation (allosct). High-resolution HLA-matching significantly impacts outcome and also the European Group of Blood and Marrow Transplantation (EBMT) risk score, based on patient age, disease stage, donor type, time from diagnosis to SCT and gender combination, may predict for non-relapse mortality and overall survival (OS). We evaluated the individual and combined effects of allele-matching and the EBMT risk score in 27 patients with poor-risk acute leukemia or myelodysplasia, who received a T-cell depleted URD allosct. Matching for HLA-A, -B, -C and -DRB1 alleles ( match) was associated with a 5-year OS of % compared with % for mismatched (p7/8) pairs (P ¼.2). Patients with EBMT risk scores of 1 2,, and 5 7 had 5-year OS estimates of 5,, and 2%, respectively (Po.1). The favorable prognostic impact of an donor was most pronounced if the EBMT risk score was low (1 2). Fiveyear OS was 7±8% vs 9±11% for fully matched patients with a low-risk EBMT score as compared with EBMT low-risk patients with p7/8 donors. These data underscore the importance of incorporating both the EBMT risk score and the degree of high-resolution HLA-matching in the risk assessment prior to URD allosct. (211), ; doi:1.18/leu ; published online 2 May 211 Keywords: acute leukemia; allele-matching; unrelated donor transplantation Introduction Allogeneic hematopoietic stem cell transplantation (allosct) is an established treatment modality to consolidate remission in patients with acute leukemia. 1 However, non-relapse mortality (NRM) may partially offset the favorable immunotherapeutic effect of allosct and, as a result, only well-selected patients may benefit. Recent meta-analyses of prospective sibling donor versus no-donor studies have shown a survival advantage for Correspondence: Professor JJ Cornelissen, Department of Hematology, Erasmus University Medical Center, Daniel den Hoed Cancer Cliniques (DDHK), Groene Hilledijk 1, EA Rotterdam, The Netherlands. j.cornelissen@erasmusmc.nl Received 2 November 21; revised 2 January 211; accepted 2 February 211; published online 2 May 211 donor-availability in acute myeloid leukemia (AML) patients in first complete remission (CR1), who lack favorable cytogenetics and acute lymphoblastic leukemia (ALL) patients in CR1 with a standard or high-risk profile. 5 9 In addition, allosct is generally considered the treatment of choice in acute leukemia patients in second or subsequent CR. Whereas allosct is preferably performed with sibling donors, matched unrelated donors (URDs) are increasingly applied. 1 AlloSCT using an URD may be associated with a somewhat higher risk of NRM as compared with allosct from HLA-identical sibling donors, although results using molecularly matched donors may approximate those using sibling donors Apart from HLAmatch, NRM after allosct depends on many parameters, including age, cytomegalovirus (CMV) serostatus, disease stage, donor-recipient gender combination, performance status and comorbidity of the patient. Two different risk scores (the European Group of Blood and Marrow Transplantation (EBMT) risk score and the hematopoietic cell transplantation specific comorbidity index (HCT-CI)) have been developed incorporating selected parameters, and have been demonstrated to predict NRM after allosct The EBMT risk score is based on 5 parameters, including patient age, disease stage, time between diagnosis and transplantation, type of donor and gender combination. The score was initially designed to predict outcome after allosct for chronic myeloid leukemia (CML) and has recently been validated in acute leukemia and myelodysplastic syndrome (MDS). 18,19 The score accurately predicts for overall survival (OS) and NRM, including a NRM of less than 2% in patients with a score of, a NRM of 2 % for a score of 1 2, a NRM of % for a score of and a NRM % in patients with or more points. We set out to evaluate the relative prognostic value of high-resolution HLA-matching status and EBMT risk score on OS and NRM in a cohort of patients with high-risk acute leukemia or MDS receiving an URD allosct. Patients and methods Patients A total of 27 patients who received hematopoietic stem cell grafts from volunteer URDs were included in this analysis.

2 Transplantations took place in the Netherlands between June 1987 and December 26. In 27 out of the 27 donorrecipient pairs, information on allele-matching for HLA-A, -B, -C and -DRB1 between donor and recipient was available for analysis. Patients were completely matched () with their donors (n ¼ 17) or had 1 or multiple mismatches (p7/8; n ¼ ). Table 1 shows patient, donor and transplantation Table 1 Patient characteristics, of all patients and according to allele-matching status All patients, no. (%), no. (%) p7/8, no. (%) Number Age at transplantation, years (16 67) 6 (16 67) (16 65) median (range) o2 8 (12) 18 (11) 1 (1) (5) 86 (51) 56 (56) 112 () 66 (9) 1 (1) Sex, male 176 (5) 9 (5) 5 (5) Karnofsky X8% 19 (98) 167 (98) 95 (95) Male patient/female donor 2 (1) 19 (11) 1 (1) CMV patient/donor / 16 (2) 6 (7) 26 (26) Other 216 (66) 1 (61) 7 (7)? 5 (2) (2) F Disease AML 12 () 7 () 9 (9) ALL 92 (28) () () MDS 7 (2) (2) 2 (2) CMML 1 () 9 (5) F AUL 9 () 5 () 1 (1) Disease stage CR1 129 (9) 72 (2) 7 (7) CR2 91 (28) 8 (28) 2 (2) CR2/not in remission 17 () (29) 1 (1) EBMT score (22) 8 (22) 2 (2) 77 (2) 5 (26) () 76 (2) 7 (22) 22 (22) (1) (29) () Conditioning Regimen TBI based 25 (72) 18 (6) 78 (78) TCD 7 (9) 155 (91) 97 (97) In-vivo 27 (8) 15 (79) 9 (9) In-vitro 229 (7) 1 (61) 76 (76) MA 6 (78) 122 (72) 8 (8) Graft Bone Marrow 192 (59) 85 () 62 (62) PBSC 15 (1) 85 () 8 (8) Year of transplantation p (1) 22 (1) 1 (1) (26) 2 () 1 (1) (6) 16 (62) 55 (55) Median follow-up survivors () 8 (1 25) 7 (1 29) 51 (2 25) Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; AUL, acute undifferentiated leukemia; CMML, chronic myelomonocytic leukemia; CMV, cytomegalovirus; CR, complete remission; MA, myeloablative; MDS, myelodysplastic syndrome; PBSC, peripheral blood stem cells; TBI, total body irradiation; TCD, T-cell depletion. characteristics, for the entire study population and according to allele-matching status. The median age of the patients was years (range, 16 67). Diagnoses included AML (n ¼ 12), ALL (n ¼ 92), acute undifferentiated leukemia (n ¼ 9), chronic myelomonocytic leukemia (n ¼ 1) and MDS (n ¼ 7). All patients were considered to have a high risk of disease relapse, based on the presence of adverse cytogenetics in case of AML or ALL in CR1, an IPSS X1,5 in case of MDS or a more advanced disease stage (CR1). Adverse cytogenetics included complex karyotypes (X abnormalities), del(5q)/ 5, del(7q)/ 7, abn(q), t(6;9), t(9;22), abn(11q2) for AML and t(;11), t(9;22) for ALL. Clinical data were collected from the Dutch Transplantation Registry database (HOVON SCT working group). HLA typing data were performed in three HLA laboratories located in Leiden, Nijmegen and Utrecht, and provided by Europdonor. Patients were grouped into EBMT-risk groups (Table 1) on the basis of points attributed for the following risk factors: age: 2 yrs: points (pts), 2 yrs: 1 pt, yrs: 2 pts; disease stage: early (CR1): pts, intermediate (CR2): 1 pt, advanced (CR2, not in remission): 2 pts; time from diagnosis to transplantation: p12 : pts, 12 : 1 pt; donor type: HLAidentical sibling donor: pts, all other: 1 pt; donor recipient gender combination: all other: pts, female donor male recipient: 1 pt. Distribution into the EBMT categories was evenly among matched and mismatched patients (Table 1). HLA typing High-resolution typing was performed for HLA-A, -B, -C and -DRB1 in the majority (8%) of the donor-recipient pairs. HLAdata included original HLA-typings at the time of transplant for patients transplanted since 1996, completed with retrospective typing of lacking HLA-data that was performed in more recent years or at the time of analysis. 2 Methods for allele typing included PCR techniques using sequence-specific oligo-nucleotide probes or sequence-specific primers and/or sequence-based typing. High-resolution HLA-matching at HLA-A, -B, -C and -DRB1 was considered an fully matched pair. Mismatching was defined as the presence of donor alleles not shared by the recipient, or the presence of recipient alleles not shared by the donor. Mismatches at homozygous alleles were considered single mismatches. Pairs not fully typed for all four loci, but in which at least a single mismatch was present, were grouped into the p7/8 matched category. Endpoints The primary endpoint of the analysis was OS, defined as time from transplantation until death from any cause. Patients still alive at the date of last contact were censored. Secondary endpoints were engraftment, acute and chronic graft-versus-host disease (GVHD), NRM, relapse and relapse mortality. Engraftment was defined as the first of consecutive days when the absolute neutrophil count exceeded /l. Primary graft failure was defined by a lack of neutrophil engraftment in patients surviving at least 28 days. Secondary graft failure was defined as neutrophil recovery followed by a decline in absolute neutrophil count to below /l for consecutive days or loss of donor chimaerism. Acute (agvhd) and chronic GVHD (cgvhd) were diagnosed and graded according to consensus criteria.,26 Chronic GVHD was evaluated among patients who survived at least days after transplantation. NRM and relapse were defined according to standard criteria

3 15 Statistical analysis For analysis of GVHD, mismatching was defined as the presence of recipient alleles not shared by the donor (donor disparity). For analysis of graft failure, mismatching was defined as the presence of donor alleles not shared by the recipient (recipient disparity). For analysis of survival, mismatching included either type of disparity. Categorical data were summarized as frequencies and percentages, continuous variables as median values and range. Univariate Cox regression analysis was used to evaluate the prognostic value of HLA match ( vs p7/8) and the EBMT risk score (1 2 vs vs vs 5 7) on OS, NRM and relapse mortality. The hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were determined. Actuarial probabilities with 95% CIs were determined for appropriate time points. Competing risk analysis was used to calculate cumulative competing risks of NRM and relapse mortality. Kaplan- Meier curves were generated to illustrate survival. In addition, the impact of disease stage (CR1 vs CR2 vs other), age, year of transplant, conditioning regimen (myeloablative vs reduced intensity), stem cell source (bone marrow vs peripheral blood), application of T-cell depletion (ex-vivo TCD no vs yes, in-vivo TCD no vs yes) and CMV seropositivity of the patient on OS, NRM and relapse mortality was evaluated with univariate Cox regression analysis. All reported P-values are two-sided and have not been adjusted for multiple testing. A significance level a ¼.5 was used. Results The study included 27 patients. Currently, 12 patients are still alive with a median follow up of 8 (range, 1 25). As described in Table 1, patients were in CR1 (n ¼ 129), CR2 (n ¼ 91), or beyond CR2 or not in remission (n ¼ 17). T-cell depletion was applied in the majority (9%) of patients, using in vitro (7%) and/or in vivo (8%) methods incorporating antithymocyte globulin (ATG) or alemtuzumab in the preparative regimens. Most of the patients (78%) were treated with myeloablative conditioning regimens and 59% received a bone marrow transplant. There were no differences in distribution of EBMT risk score nor in distribution of individual parameters of the EBMT risk score between the en p7/8 group. HLA matching status Complete high-resolution typing of HLA-A, -B, -C and -DRB1 was available in out of 27 donor-recipient pairs. These pairs as well as the numbers of matched and mismatched pairs were equally distributed between the different EBMT risk groups, as shown in Table 1. Patients and donors were completely () matched for HLA-A, -B, -C and -DRB1 alleles (n ¼ 17) or were mismatched at 1 allele (n ¼ 62), 2 alleles (n ¼ 17) or alleles (n ¼ ), as shown in Table 2. All pairs in the Table 2 Characteristics of HLA-mismatches Single MM (n ¼ 62) 2 MM (n ¼ 17) MM (n ¼ ) A 17 B+C 6 A+B+C 1 B 1 C+DRB1 5 B+B+C 1 C 28 A+C C+C+DRB1 1 DRB1 7 A+B 1 B+DRB1+DRB1 1 C+C 1 DRB1+DRB1 1 Abbreviation: MM, mismatch. group were typed at high-resolution level for all loci. Seventeen pairs, who were not fully typed for all four loci, but in whom at least a single allele-mismatch was present, were grouped into the p7/8 matched cohort. The majority (n ¼ 19) of patients in the group were also matched for DQB1 (1/1 matched). EBMT risk score The EBMT risk score was estimated for the 27 patients in the study. Patients had EBMT risk scores of 1 (n ¼ 9), 2 (n ¼ 62), (n ¼ 77), (n ¼ 76), 5 (n ¼ 81), 6 (n ¼ 19) and 7 (n ¼ ). Because only few patients had risk scores of 1, 6 or 7, the patients with risk scores 1 and 2 (n ¼ 71) and 5 7 (n ¼ 1) were combined in the analysis. Engraftment and GVHD The incidence of primary graft failure was % and similar in the and the p7/8 group. Stable engraftment occurred in 9% of patients. The overall incidence of grade I IV agvhd was 57% for the entire study population. Acute GVHD was reported to be grade I in 1%, grade II in 16% and grade III IV in 9% of the patients. The cumulative incidence of grade II IV agvhd at after allosct was 2±% in completely matched patients and 26±% in the p7/8 group. Overall, % of patients developed cgvhd. Chronic GVHD was extensive in 12% of the patients. The cumulative incidence of overall and extensive cgvhd at 2 years after allosct was 5±5% and 17±% in matched patients as compared with 57±7% and 2±6% in the p7/8 group. Non-relapse mortality The cumulative incidence of NRM at 5 years from transplantation was ±% vs 9±5% for patients with and p7/8 donors, respectively (HR ¼.55; 95% CI.5.85, P ¼.8), as shown in Figure 1b. Due to non-relapse causes of death, 9 of 17 patients with donors and of recipients with p7/8 donors died. Principal causes of death, classified according to Copelan et al., 27,28 were GVHD (18 patients in the group and 18 patients in the p7/8 group) and infections (12 patients in the group and 1 patients in the p7/8 group). NRM at 5 years for patients with EBMT risk scores of 1 2,, and 5 7 was 22±6%, ±5% (HR ¼ 1.77, 95% CI:.9.9), ±6% (HR ¼ 2.11, 95% CI: 1.9.9) and 7±5% (HR ¼ 2.68, 95% CI: ), respectively (Figure 2b). A higher rate of NRM (52±7%) was observed in patients who received their transplant before 1996 (n ¼ 6) as compared with patients transplanted between 1996 and 21 (2±5%, n ¼ 85) and after 21 (26±%, n ¼ 196). Other unfavorable prognostic factors for NRM were advanced disease stage (HR ¼ 2.18, 95% CI: ), myeloablative conditioning (HR ¼ 1.79, 95% CI: 1..22), in-vitro TCD (HR ¼ 1.6, 95% CI: ) and CMV seropositivity (HR ¼ 1.92, 95%: ) (Supplementary Table). Relapse and relapse mortality The cumulative incidence of relapse was 5±% at 5 years from allosct and similar for patients with and p7/8 donors (HR ¼.8; 95% CI: , P ¼.). Overall relapse mortality rate was %± and was not significantly different between patients with and p7/8 donors (Figure 1c). Relapse incidence at 5 years from transplantation for patients with EBMT risk scores of 1 2,, and 5 7 was 28±5%, 29±5%

4 1551 N D Cox LR P=. > N D Cox LR P<.1 > N NRM 17 9 Cox LR P=.8 N REL 17 5 Cox LR P=.62 Figure 1 OS (a), NRM (b), and relapse mortality (c), as determined by degree of matching. (a) Allele-matching () for HLA-A, -B, -C, -DRB1 was associated with significantly superior OS compared with mismatching (p7/8) (P ¼.). (b) Allele-matching () for HLA-A, -B, -C, -DRB1 was associated with significantly lower NRM compared with mismatching (p7/8) (P ¼.8). (c) No significant difference in relapse mortality was observed between matched () compared with mismatched (p7/8) pairs. N > 1 Cox LR P<.9 N REL > 1 9 Cox LR P=. NRM > > Figure 2 OS (a), NRM (b), and relapse mortality (c), as determined by EBMT risk score. (a) A higher EBMT risk score was associated with inferior OS (Po.1). (b) A higher EBMT risk score was associated with a significant increase in NRM (Po.1). (c) A higher EBMT risk score was associated with increased relapse mortality (P ¼.1). (HR ¼ 1.1; 95%CI:.71 2.), 7±6% (HR ¼ 1.76; 95% CI:.98.15) and ±5% (HR ¼ 2.2; 95% CI: 1..87), respectively. Relapse mortality rate according to the EBMT risk score was 2±5% (score 1 2), 28±5% (score ), 7±6% (score ) and ±5% (score 5 7) (Po.1) as shown in Figure 2c. In addition, advanced disease stage was adversely associated with relapse mortality (HR ¼ 2.22, 95% CI: ) (Supplementary Table). Overall survival OS at 5 years from transplantation was 5% (95% CI 29 %) for the entire study population. Currently, 12 patients are alive with a median follow up of 8 (range 1 25), of whom 11 are progression free. As shown in Figure 1a, OS at 5 years was % (95% CI 2 8%) for patients with donors and % (95% CI 21 9%) for those with p7/8 donors (HR ¼.7; 95% CI:.51.95, P ¼.). Patients with EBMT risk scores of 1 2,, and 5 7 had 5-year OS estimates of 5%, % (HR ¼ 1.59; 95% CI: ), % (HR ¼ 2.5; 95%CI: ) and 2% (HR ¼ 2.61; 95% CI: ), respectively (Po.1) (Figure 2a). We also investigated the relative prognostic value of allele-matching and EBMT risk score. As shown in Table, the favorable effect of full allele-matching on OS was most obvious in patients with an EBMT low-risk score. EBMT low-risk (1 2) patients with matched donors had significantly superior OS as compared with EBMT low-risk

5 1552 Table 5-year OS with 95% CI and HR with 95% CI by EBMT risk score and allele-matching EBMT risk score donors p7/8 donors HR a (95% CI) N 5-year OS (%) (95% CI) N 5-year OS (%) (95% CI) (5 8) 2 9 (19 6).5 (.15.82) 5 8 (2 62) 5 (17 5).6 (. 1.21) 7 (18 9) 22 1 (1 51).86 (.5 1.6) ( 27) 19 (8 ).97 ( ) Abbreviation: OS, overall survival. a HR of donors compared to p7/8 donors. patients with p7/8 matched donors, 72% (95% CI 5 81%) vs 9% (95% CI 19 6%) (Table ). Patients with the highest EBMT risk scores (5 7), on the other hand, had a dismal outcome, even if they had a fully matched donor. Other unfavorable prognostic factors for OS were advanced disease stage (HR ¼ 2.1, 95% CI: ) and CMV seropositivity (HR ¼ 1., 95%: ) (Supplementary Table). Discussion Apart from HLA-allele matching, outcome after URD allosct may be determined by a number of other risk factors. 29 Recently, the predictive power of the EBMT risk score was demonstrated in a number of hematological malignancies, including acute leukemia. 19 The present analysis focused on the interaction and individual effects of both allele-matching and EBMT risk score on outcome in patients with high-risk acute leukemia or MDS. It is shown that a lower EBMT risk score and allele-matching for HLA-A, -B, -C and -DRB1 are independently associated with less NRM and superior OS in patients with highrisk acute leukemia or MDS, receiving an URD, T-cell depleted (TCD) hematopoietic stem cell graft. The favorable effect of allele-matching appeared especially evident in EBMT low-risk patients, whereas allele-matching was not associated with better outcome in EBMT high-risk patients. A number of studies have demonstrated that high-resolution HLA-matching between donors and recipients improves outcome after URD allosct. 2 Most studies reported improved survival and less NRM as a result of less GVHD in well-matched patients receiving T-cell replete stem cell grafts. 2 A limited number of studies investigated the impact of HLA-matching on outcome after URD allosct using T-cell depletion. 2 These studies are less unanimous about the prognostic significance of allele-matching. Although Kröger et al. showed limited impact of mismatching on outcome, a recent large Italian analysis of 85 patients who received URD transplants, of which 6% were TCD, concluded that a single allele-mismatch resulted in an inferior OS as compared with a 1/1 match for patients with early stage disease. 1 We also observed a favorable impact of allele-matching on OS and NRM. Whereas the incidence of chronic extensive and severe acute GVHD appeared rather limited in our study, the cumulative incidence of NRM did not differ from what has been observed using T-cell replete URD allosct. 2 Instead of GVHD as the major risk factor for NRM, rather an impaired immune reconstitution and higher incidence of lethal opportunistic infections may have contributed to a similar incidence of NRM in our patients. Lee et al. reported on 857 patients with CML, AML, MDS and ALL who received URD stem cell grafts, which were TCD in 22% of the cases. Allele-matching of HLA-A, -B, -C and -DRB1 appeared to be the preferred degree of matching. A single allele-mismatch was associated with a 1-year OS of % as compared with 52% for matched pairs. Each additional mismatch resulted in a 9 1% loss of survival, but mismatches at HLA-B and -C were better tolerated than mismatches at HLA-A and -DRB1. By incorporating stage of disease and age, it was also shown that the favorable effect of allele-matching was especially apparent in younger patients with low-risk disease. The latter result compares well with our observation of a differential effect of matching depending on the presence of additional risk factors. A comparable observation was reported by Petersdorf et al., who studied 98 patients with CML, acute leukemia and MDS, who received a T-cell replete URD transplant. Although patients with allele-matched donors had superior outcome, that beneficial effect appeared restricted to patients with low-risk disease, defined as chronic phase CML transplanted within 2 years of diagnosis. Collectively, these studies and our results suggest that allele-matching may only modestly contribute to less NRM in the presence of additional risk factors. Apparently, against a background of an a priori high mortality rate due to unfavorable clinical comorbid conditions, the added value of allelematching is lost. On the other hand, allele-matching may be associated with very low rates of NRM in the absence of additional risk factors. Moreover, NRM in those patients approximates what can be observed after sibling allosct, thereby evoking the question whether an URD search should more routinely be considered in leukemia patients in CR1. As discussed before, the beneficial effect of allosct is the net result of the favorable immunotherapeutic effect on relapse and the counterbalancing adverse effect of NRM. As the reduction of relapse was estimated approximately at a % reduction (overall HR of.56), it was argued that even in the context of a NRM of 2 %, allosct may be associated with improved outcome, if the risk of relapse exceeds % after alternative chemotherapeutic consolidation. 7 Therefore, URD allosct using well-matched donors could be considered for leukemia patients, for whom the risk of relapse exceeds % and for whom no additional risk factors for NRM can be identified. Although most cooperative study groups have incorporated the search for an URD in high-risk leukemia patients in CR1, URD allosct is generally not accepted in patients with a standard or intermediate risk profile of their leukemia. However, a number of recent studies in patients with acute leukemia have suggested that donor type (URD versus sibling) does no longer affect outcome if allele matched donors are applied Our results as well as those by Lee et al. would suggest that well-matched URD may also be applied in intermediate risk acute leukemia, provided the risk of NRM, as can be estimated by the EBMT risk score, does not exceed 2. The EBMT risk score was originally established to predict survival and NRM after allosct for patients with CML and was recently validated in a variety of

6 hematological disorders including acute leukemia and MDS. 18,19 In our analysis, patients with EBMT risk scores of 1 2,, and 5 7 had 5-year OS estimates of 5,, and 2%, respectively (Po.1). Both the rate of NRM and relapse mortality increased with increasing EBMT risk scores. Therefore, this score not only selectively predicts for NRM but also impacts on relapse as a result of incorporating stage of disease, age, and time to transplantation. While being a significant prognostic score for overall outcome and NRM, the impact on relapse may hamper the use of the score when selectively weighing the opposite favorable and toxic effects of allosct. A score that more selectively predicts for NRM is the Seattle comorbidity index (HCT-CI), that does not impact on relapse. 2 2 The HCT- CI selectively takes the cumulative number of comorbidities into account and thereby may add important prognostic information that also can be used prior to an URD search and decision making for URD allosct. In conclusion, allele-matching for HLA-A, -B, -C, and -DRB1 of donor and recipients appears also important in T-celldepleted URD allosct as earlier shown in the T-cell replete setting. The favorable effect of allele-matching is most pronounced in patients without additional risk factors predicting for NRM, as was done by the EBMT risk score. That score strongly predicted for overall outcome and NRM, with EBMT low-risk patients who received well-matched grafts experiencing excellent survival. These results emphasize the importance of incorporating a risk score that predicts for NRM in decision making prior to URD allosct that enables a careful weighing of risk for NRM versus the risk of relapse. Conflict of interest The authors declare no conflict of interest. References 1 Appelbaum FR. Incorporating hematopoietic cell transplantation (HCT) into the management of adults aged under 6 years with acute myeloid leukemia (AML). Best Pract Res Clin Haematol 28; 21: Rowe JM. Graft-versus-disease effect following allogeneic transplantation for acute leukaemia. Best Pract Res Clin Haematol 28; 21: Lodewyck T, Cornelissen JJ. Allogeneic stem cell transplantation in acute myeloid leukemia: a risk-adapted approach. Blood Rev 28; 22: Rowe JM. Optimal management of adults with ALL. Br J Haematol 29; 1: Yanada M, Matsuo K, Emi N, Naoe T. Efficacy of allogeneic hematopoietic stem cell transplantation depends on cytogenetic risk for acute myeloid leukemia in first disease remission: a metaanalysis. Cancer 25; 1: Koreth J, Schlenk R, Kopecky KJ, Honda S, Sierra J, Djulbegovic BJ et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. JAMA 29; 1: Cornelissen JJ, van Putten WLJ, Verdonck LF, Theobald M, Jacky E, Daenen SM et al. Results of a HOVON/SAKK donor versus nodonor analysis of myeloablative HLA-identical sibling stem cell transplantation in first remission acute myeloid leukemia in young and middle-aged adults: benefits for whom? Blood 27; 19: Goldstone AH, Richards SM, Lazarus HM, Tallman MS, Buck G, Fielding AK et al. In adults with standard-risk acute lymphoblastic leukemia, the greatest benefit is achieved from a matched sibling allogeneic transplantation in first complete remission, and an autologous transplantation is less effective than conventional consolidation/maintenance chemotherapy in all patients: final results of the International ALL Trial (MRC UKALL XII/ECOG E299). Blood 28; 111: Hahn T, Wall D, Camitta B, Davies S, Dillon H, Gaynon P et al. The role of cytotoxic therapy with hematopoietic stem cell transplantation in the therapy of acute lymphoblastic leukemia in adults: an evidence-based review. Biol Blood Marrow Transplant 26; 12: 1. 1 Appelbaum FR. Hematopoietic cell transplantation from unrelated donors for treatment of patients with acute myeloid leukemia in first complete remission. Best Pract Res Clin Haematol 27; 2: Krauter J, Heil G, Hoelzer D, Ottmann O, Martin H, Lübbert M et al. Role of consolidation therapy in the treatment of patients up to 6 years with high risk AML. Blood 25; 16: abstract Yakoub-Agha I, Mesnil F, Kuentz M, Boiron JM, Ifrah N, Milpied N et al. Allogeneic marrow stem-cell transplantation from human leukocyte antigen-identical siblings versus human leukocyte antigen-allelic-matched unrelated donors (1/1) in patients with standard-risk hematologic malignancy: a prospective study from the French Society of Bone Marrow Transplantation and Cell Therapy. J Clin Oncol 26; 2: Schetelig J, Bornhäuser M, Schmid Ch, Hertenstein B, Schwerdtfeger R, Martin H et al. Matched unrelated or matched sibling donors result in comparable survival after allogeneic stemcell transplantation in elderly patients with acute myeloid leukemia: a report from the cooperative german transplant study group. J Clin Oncol 28; 26: Basara N, Schulze A, Wedding U, Mohren M, Gerhardt A, Junghanss C et al. Early related or unrelated haematopoietic cell transplantation results in higher overall survival and leukaemiafree survival compared with conventional chemotherapy in highrisk acute myeloid leukaemia patients in first complete remission. 29; 2: Gupta V, Tallman MS, He W, Logan BR, Copelan E, Gale RP et al. Comparable survival after HLA-well-matched unrelated or matched sibling donor transplantation for acute myeloid leukemia in first remission with unfavorable cytogenetics at diagnosis. Blood 21; 116: Federmann B, Faul Ch, Vogel W, Kanz L, Bethge WA. Allogeneic hematopoietic cell transplantation in AML: Comparable results after matched or mismatched unrelated versus matched related transplantation. Blood 29; 11: abstract Walter RB, Pagel JM, Gooley TA, Petersdorf EW, Sorror ML, Woolfrey AE et al. Comparison of matched unrelated and matched related donor myeloablative hematopoietic cell transplantation for adults with acute myeloid leukemia in first remission. 21; 2: Gratwohl A, Hermans J, Goldman JM, Arcese W, Carreras E, Devergie A et al. Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation. Lancet 1998; 52: Gratwohl A, Stern M, Brand R, Apperley J, Baldomero H, de Witte T et al. Risk score for outcome after allogeneic hematopoietic stem cell transplantation: a Retrospective Analysis. Cancer 29; 115: Sorror ML, Maris MB, Storb R, Storer BE, Stuart MJ, Hegenbart U et al. Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood 25; 16: Sorror ML, Giralt S, Sandmaier BM, De Lima M, Shahjahan M, Maloney DG et al. Hematopoietic cell transplantation specific comorbidity index as an outcome predictor for patients with acute myeloid leukemia in first remission: combined FHCRC and MDACC experiences. Blood 27; 11: Majhail NS, Brunstein CG, McAvoy S, DeFor TE, Al-Hazzouri A, Setubal D et al. Does the hematopoietic cell transplantation specific comorbidity index predict transplant outcomes? A validation study in a large cohort of umbilical cord blood and matched related donor transplants. Biol Blood Marrow Transplant 28; 1: Farina L, Bruno B, Patriarca F, Spina F, Sorasio R, Morelli M et al. The hematopoietic cell transplantation comorbidity index 155

7 155 (HCT-CI) predicts clinical outcomes in lymphoma and myeloma patients after reduced-intensity or non-myeloablative allogeneic stem cell transplantation. 29; 2: Heemskerk M, van Walraven S, Cornelissen JJ, Barge R, Bredius R, Egeler R et al. How to improve the search for an unrelated haematopoietic stem cell donor. Faster is better than more! Bone Marrow Transplant 25; 5: Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 199 consensus conference on acute GVHD grading. Bone Marrow Transplant 1995; 15: Shulman HM, Sullivan KM, Weiden PL, McDonald GB, Striker GE, Sale GE et al. Chronic graft-versus-host syndrome in man. A longterm clinicopathologic study of 2 Seattle patients. Am J Med 198; 69: Clift R, Goldman J, Gratwohl A, Horowitz M. Proposals for standardized reporting of results of bone marrow transplantation for leukaemia. Bone Marrow Transplant 1989; : Copelan E, Casper JT, Carter SL, van Burik JA, Hurd D, Mendizabal AM et al. A Scheme for Defining Cause of Death and Its Application in the T Cell Depletion Trial. Biol Blood Marrow Transplant 27; 1: Shaw BE, Arguello R, Garcia-Sepulveda CA, Madrigal JA. The impact of HLA genotyping on survival following unrelated donor haematopoietic stem cell transplantation. Br J Haematol 21; 1: 1 8. Kröger N, Zabelina T, Binder T, Ayuk F, Bacher U, Amtsfeld G et al. HLA-mismatched unrelated donors as an alternative graft source for allogeneic stem cell transplantation after antithymocyte globulin-containing conditioning regimen. Biol Blood Marrow Transplant 29; 15: Crocchiolo R, Ciceri F, Fleischhauer K, Oneto R, Bruno B, Pollichieni S et al. HLA matching affects clinical outcome of adult patients undergoing haematopoietic SCT from unrelated donors: a study from the Gruppo Italiano Trapianto di Midollo Osseo and Italian Bone Marrow Donor Registry. Bone Marrow Transplant 29; : Shaw BE, Mayor NP, Russell NH, Apperley JF, Clark RE, Cornish J et al. Diverging effects of HLA-DPB1 matching status on outcome following unrelated donor transplantation depending on disease stage and the degree of matching for other HLA alleles. 21; 2: Lee SJ, Klein J, Haagenson M, Baxter-Lowe LA, Confer DL, Eapen M et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood 27; 11: Petersdorf EW, Anasetti C, Martin PJ, Gooley T, Radich J, Malkki M et al. Limits of HLA mismatching in unrelated hematopoietic cell transplantation. Blood 2; 1: Woolfrey AE, Klein J, Haagenson MD, Spellman SR, Battiwalla M, Gajewski JL et al. Evaluation of Human Leukocyte Antigen (HLA) Matching Requirements for Unrelated Peripheral Blood Stem Cell (PBSC) Transplantation. Blood 28; 112: abstract Morishima Y, Sasazuki T, Inoko H, Juji T, Akaza T, Yamamoto K et al. The clinical significance of human leukocyte antigen (HLA) allele compatibility in patients receiving a marrow transplant from serologically HLA-A, HLA-B, and HLA-DR matched unrelated donors. Blood 22; 99: Flomenberg N, Baxter-Lowe LA, Confer D, Fernandez-Vina M, Filipovich A, Horowitz M et al. Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood 2; 1: Petersdorf EW, Gooley TA, Anasetti C, Martin PJ, Smith AG, Mickelson EM et al. Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient. Blood 1998; 92: Sasazuki T, Juji T, Morishima Y, Kinukawa N, Kashiwabara H, Inoko H et al. Effect of matching of class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor. Japan Marrow Donor Program. N Engl J Med 1998; 9: Petersdorf EW, Kollman C, Hurley CK, Dupont B, Nademanee A, Begovich AB et al. Effect of HLA class II gene disparity on clinical outcome in unrelated donor hematopoietic cell transplantation for chronic myeloid leukemia: the US National Marrow Donor Program Experience. Blood 21; 98: Greinix HT, Faé I, Schneider B, Rosenmayr A, Mitterschiffthaler A, Pelzmann B et al. Impact of HLA class I high-resolution mismatches on chronic graft-versus-host disease and survival of patients given hematopoietic stem cell grafts from unrelated donors. Bone Marrow Transplant 25; 5: Ho VT, Kim HT, Liney D, Milford E, Gribben J, Cutler C et al. HLA- C mismatch is associated with inferior survival after unrelated donor non-myeloablative hematopoietic stem cell transplantation. Bone Marrow Transplant 26; 7: Meijer E, Cornelissen JJ. Allogeneic stem cell transplantation in acute myeloid leukemia in first or subsequent remission: weighing prognostic markers predicting relapse and risk factors for nonrelapse mortality. Semin Oncol 28; 5: Supplementary Information accompanies the paper on the website (