HLA Matching Affects Clinical Outcome of Adult Patients

ORIGINAL ARTICLE 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

R Crocchiolo1,5, F Ciceri1, K Fleischhauer2, R Oneto3, B Bruno3, S Pollichieni4, N Sacchi4, MP Sormani5, R Fanin6, G Bandini7, F Bonifazi7, A Bosi8, A Rambaldi9, PE Alessandrino10, M Falda11 and A Bacigalupo3

1Department of Oncology, Hematology and Bone Marrow Transplantation Unit, S Raffaele Scientiﬁc Institute, Milano, Italy; 2Immunogenetics Laboratory, S Raffaele Scientiﬁc Institute, Milano, Italy; 3Department of Hemato-Oncology, Division of Hematology, Ospedale San Martino, Genova, Italy; 4Italian Bone Marrow Donor Registry, Ospedale Galliera, Genova, Italy; 5Biostatistics Unit, Department of Health Sciences (DISSAL), University of Genova, Genova, Italy; 6Division of Hematology and Bone Marrow Transplantation, University of Udine, Udine, Italy; 7Institute of Hematology and Clinical Oncology ‘LA Seragnoli’, Bologna, Italy; 8Department of Hematology, University of Florence, Florence, Italy; 9Department of Oncology and Hematology, Division of Hematology, Ospedali Riuniti, Bergamo, Italy; 10Division of Hematology, University of Pavia, Pavia, Italy and 11Division of Hematology, Ospedale San Giovanni Battista, Torino, Italy

The importance of HLA donor–recipient matching in Keywords: haematopoietic SCT; haematological malig- unrelated haematopoietic SCT (HSCT) is the subject of nancy; unrelated donor; HLA matching. debate. In this retrospective study, we analyzed 805 adult patients from the Italian Registry receiving HSCT for a haematological malignancy from January 1999 to June 2006 and correlated the degree of HLA matching with Introduction transplant outcome. All patient–donor pairs had high- resolution typing at HLA-A, -B, -C, -DRB1 and -DQB1. Haematopoietic SCT (HSCT) represents a curative option There was a significantly higher risk of overall mortality, for many haematological malignancies.1 A HLA-identical non-relapse mortality, graft failure and acute GVHD sibling is considered the best HSC donor; however, the (aGVHD) for patients receiving HSCT from an unrelated search for an unrelated donor is currently performed when donor with one or more low- or high-resolution mismatch/ a patient lacks such an HLA-matched sibling. Searching es (Mm/s). When only a single HLA Mm is present (9/10 algorithms aim at finding the optimal match between the matched pairs), mortality risk is higher than among 10/10 patient and the potential donor.2 As the ‘10/10 matched’ matched pairs in patients transplanted with acute (that is, high-resolution identity at loci HLA-A, -B, -C, leukaemia in the first CR (‘early’ patients) but not in -DRB1 and -DQB1) donor is not always available, some the other patients (advanced patients): HR ¼ 1.69, 95% extent of HLA mismatch (Mm) can be accepted; however, CI ¼ 0.94–3.02, P ¼ 0.08; HR ¼ 1.03, 95% CI ¼ 0.80– there is still some controversy about the precise impact of 1.32, P ¼ 0.82, for early and advanced patients, respec- HLA mismatching on HSCT outcomes, because published tively. These results confirm that the advantage of a 10/10 reports indicate discordant results about the relative match has a greater effect in early patients, thus importance of loci involved.3–14 A recent study by Lee suggesting that a 9/10 matched donor can be chosen in et al.3 indicated that both class I and class II Mms affect patients with advanced disease lacking a rapidly available survival, suggesting that each HLA Mm is associated with 10/10 matched one. an additional unadjusted survival impairment of 9–10%. Bone Marrow Transplantation (2009) 44, 571–577; The same study did not find a significant role played by doi:10.1038/bmt.2009.67; published online 13 April 2009 HLA-DP or -DQ on survival. Other analyses reported a higher risk of graft failure correlated with class I Mms and a superior rate of mortality in the presence of two or more Mms (regardless of loci involved), whereas one Mm did not Correspondence: Dr R Crocchiolo, Hematology and Bone Marrow confer survival impairment.5,14 Single Mms at HLA-B or -C Transplantation Unit, San Raffaele Scientific Institute, Via Olgettina, 60, seem to be better tolerated than those at HLA-A or -DRB1 Milan 20132, Italy. 3 E-mail: [email protected] in American patients, whereas in Japanese patients the Received 21 January 2009; revised 20 February 2009; accepted 21 presence of HLA-A and -B Mm significantly reduced February 2009; published online 13 April 2009 survival and Mm at HLA-C or -DRB1/DQB1 did not.12 Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 572 The relative importance of antigenic (detected by serological Table 1 Patients’ and donors’ characteristics or low-resolution typing) or allelic (high resolution) Mms Characteristics N % Median has also been studied, and the results are controversial (range) because the studies report both an equivalent3,11 and a different7 effect on outcome. Finally, a recent analysis Number of transplants 805 100 provided by Kawase et al.4 identified 16 significant high- Median patients’age at HSCT 41 (18–65) Patients’ gender: (M/F) 452/353 56/44 risk HLA Mm combinations for severe acute GVHD Disease: (aGVHD), whereas the selection of a cross-reactive group- ALL 154 19 compatible donor does not seem to improve transplant AML 234 29 outcome.9 MDS 70 8 secAL 24 39 This study was designed to assess the role of patient– MPS other than CML 22 3 donor HLA matching on the outcomes of unrelated HSCTs MM 63 8 performed for haematological malignancies in 805 adult NHL 76 9 patients in Italy. HL 52 6 CLL 14 2 CML 96 13

Disease status at HSCT: Patients and methods CML-CP1 40 5 Early 157 20 Patients and treatments Advanced 608 75 A total of 805 patients aged 18 years or more who received HSCTs from unrelated donors in Italy from 1 January 1999 ATG as GVHD prophylaxis: Yes 549 68 to 30 June 2006 were included in the analysis (Table 1). No 216 27 Eligible diagnoses were ALL, AML, CML, myelodysplastic NK 40 5 syndrome, myeloproliferative syndromes other than CML, secondary acute leukaemia, multiple myeloma, Hodgkin’s HSCT conditioning: RIC 230 28 lymphoma, non-Hodgkin’s lymphoma and CLL. Informed MAC 538 67 consent was obtained from each transplantation centre. NK 37 5 All clinical data were obtained from the Gruppo Italiano Trapianto di Midollo Osseo (GITMO) registry, and all Source of stem cells: patient–donor typing results were collected from the Italian BM 455 57 15 PBSC 348 43 Bone Marrow Donor Registry (IBMDR). According to NK 2 0 disease status at HSCT, patients were categorized into three groups: CML in first chronic phase (‘CML-CP1’), acute Total body irradiation: leukaemia in first complete remission (‘early’) and all other Yes 457 57 No 336 42 diseases (‘advanced’). HSCT conditioning regimens were NK 12 1 both myeloablative and reduced intensity, according to local or cooperative protocols. Induction therapy and HSCT Patients’ CMV status: indication was decided by treating physicians in each centre, Positive 501 62 according to the risk/benefit profile based on the most Negative 112 14 NK 192 24 updated knowledge at the time of therapy. Information on in vivo T-cell depletion was collected; aGVHD prophylaxis was Median donors’ age 35 (19–61) mostly performed with CYA and MTX. Donors’ gender: (M/F) 561/244 69/31

Donors’ CMV status: Positive 386 48 HLA typing and HSCT Negative 411 51 All Mms were classified according to the loci involved and NK 8 1 according to whether they were detected at low- or high- Abbreviations: BM ¼ bone marrow; HL ¼ Hodgkin’s lymphoma; resolution level. MAC ¼ myeloablative conditioning; MDS ¼ myelodysplastic syndrome; Originally, only low-resolution typing of loci A, B and MM ¼ multiple myeloma; MPS ¼ myeloproliferative syndrome; NK ¼ not high-resolution at locus DRB1 were used to screen known; NHL ¼ non-Hodgkin’s lymphoma; RIC ¼ reduced-intensity con- potential donors, and the matching at these three loci was ditioning; secAL ¼ secondary acute leukaemia; PBSC ¼ peripheral blood 16 stem cells. the minimum requested for performing HSCT; in the Characteristics of patients, donors and HSCT are here shown. absence of such a ‘6/6 matched’ donor, some exceptions were allowed; therefore, transplants using a ‘5/6 matched’ donor were performed. Clinical end point definitions As the degree of typing was heterogeneous among Overall survival (OS), non-relapse mortality, aGVHD and patient–donor pairs, this analysis includes only 10/10 engraftment were defined according to European Group high-resolution typed pairs, to exclude any confoun- for Blood and Marrow Transplantation (EBMT) criteria;17 ding effect eventually played by the match at loci not relapse-related death is defined as death because of relapse fully typed. after HSCT, irrespective of further treatment performed for

Bone Marrow Transplantation Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 573 relapse. All outcomes were evaluated from the day of stem Results cell infusion. Grading of aGVHD was performed according to current criteria.18 Patient–donor pairs A total of 805 pairs typed at high resolution for HLA-A, -B, -C, DRB1 and -DQB1 were included in the analysis. Statistical analysis Overall, three quarters of patients (611 out of 805, 75%) Continuous variables were expressed as median (range), received HSCT for an advanced-stage haematological whereas categorical ones were expressed as proportions and malignancy; of these, 205 were lymphoid malignancies percentages. Probabilities of OS were calculated using the other than ALL: 76 non-Hodgkin’s lymphoma, 63 multiple Kaplan–Meier estimator19 and survival curves were com- myeloma, 52 Hodgkin’s lymphoma and 14 CLL. Median pared using the logrank test.20 Univariate analysis tested follow-up was 25 months (range: 3–85). Patients’ and any association between HLA matching and the outcome donors’ characteristics are shown in Table 1. of interest. Analysis for each outcome was performed comparing the pairs mismatched at a specific locus with HSCT outcome HLA 10/10 matched pairs, thus avoiding any interference All pairs were classified according to the numbers of Mms played by Mms at other loci. This approach has been and loci involved. Heterogeneity tests disclosed significant successfully used in earlier studies.3,11 Pairs with two Mms differences in the probability of OS between pairs with a were grouped together irrespective of the loci involved; single low-resolution (antigenic) or high-resolution (allelic) pairs with three and four Mms were grouped together. Mms; therefore, they were considered equivalent and A heterogeneity test was performed between pairs with one grouped together in all analyses. antigenic or allelic Mm to verify whether they had a different impact on survival; it was performed within pairs with a single Mm between the different loci involved in Overall survival Mm. Multivariate models were built after the addition of A slight but not statistically significant survival difference all non-HLA factors that were significantly associated with emerged from the analysis between pairs with a single Mm outcome in univariate analysis. Po0.20 and Po0.05 were compared with 10/10 matched pairs. This is true for any considered as cutoff values in univariate and multivariate locus analyzed separately and also after grouping single analyses, respectively. Both analyses were performed using Mm pairs together, as no heterogeneity exists between loci Cox regression21 for OS, relapse-related death and non- A, B, C and DQB1 (data not shown). Multivariate hazard relapse mortality and logistic regression for aGVHD and ratio of OS was 1.12 (95% CI ¼ 0.90–1.40, P ¼ 0.31) for engraftment. Non-HLA factors included: donor’s gender, single Mm compared with 10/10 matched pairs; a higher age (continuous variable) and CMV status; patient’s mortality risk was observed when the number of total Mms gender, age (continuous variable), CMV status; year of is two or more (HR ¼ 1.42, 95% CI ¼ 1.08–1.87, P ¼ 0.01, transplantation (continuous variable), use of anti-thymo- Table 2). Non-HLA factors significantly associated with cyte globulin, disease group (advanced vs early vs CML- survival were: patient’s age at HSCT, donor gender, year of CP1), source of HSCs (PBSCs vs BM), conditioning transplantation, disease status at HSCT, conditioning regimen (myeloablative vs reduced intensity), use of TBI- regimen (Table 3). Survival curves according to HLA based conditioning. Centre effect (440 HSCT vs p40) was matching are shown in Figure 1a. Although interaction test tested and was NS. Disease status at transplant and use of between one HLA Mm and disease status at HSCT did not anti-thymocyte globulin were tested for interaction with yield a statistically significant association for OS (P ¼ 0.12), HLA matching. No significant interactions were identified. we observed a greater effect played by a single Mm in

Table 2 Adjusted hazard ratios of HSCT outcomes according to total number of HLA Mms

No. of patients Overall survival Graft failure aGVHD grade II–IV Non-relapse mortality

HR 95% CI P-value HR 95% CI P-value HR 95% CI P-value HR 95% CI P-value

10/10 matched 340 1 1 1 1

1 Mm at any locus 280 1.12 0.90–1.40 0.31 1.21 0.65–2.25 0.55 1.45 0.95–2.20 0.08 1.32 0.99–1.75 0.06

1 Mm locus A 54 1.47 0.96–2.27 0.08 2.05 0.66–6.43 0.21 1.56 0.76–3.20 0.23 1.53 0.94–2.50 0.09

1 Mm locus B 47 1.12 0.70–1.79 0.62 1.50 0.39–5.71 0.55 2.25 1.10–4.61 0.03a 1.12 0.65–1.91 0.68

1 Mm locus C 135 1.17 0.85–1.60 0.33 1.23 0.46–3.28 0.67 1.27 0.75–2.16 0.37 1.26 0.88–1.79 0.20

1 Mm locus DQB1 44 1.03 0.62–1.72 0.90 3.26 1.02–10.38 0.04a 0.89 0.37–2.14 0.79 1.37 0.81–2.30 0.24

2–4 Mms 185 1.42 1.08–1.87 0.01 1.94 0.88–4.28 0.10 1.68 1.05–2.69 0.03 1.59 1.17–2.16 0.003

Hazard ratios of OS, graft failure, aGVHD grade 2–4 and NRM were calculated for total number of HLA Mms after adjustment with non-HLA variables. aHeterogeneity test showed no signiﬁcant differences among pairs with single Mm. Bold values indicate Po0.05.

Bone Marrow Transplantation Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 574 Table 3 Multivariate analysis on overall survival a 1.0 Variable HR 95% CI P-value

HLA matching 0.8 10/10 matched 1 1 Mm 1.12 0.90–1.40 0.31 2–4 Mms 1.42 1.08–1.87 0.01 0.6 Patient’s age 1.01 1.006–1.024 0.001 Donor’s gender (F/M) 1.33 1.09–1.63 0.005 10/10-matched 0.4 HSCT year 0.91 0.85–0.96 0.001 1 Mm

Conditioning regimen (RIC vs MAC) 0.65 0.51–0.82 o0.0001 Survival probability 2-4 Mms Disease status at HSCT 0.2 Early 1 Advanced 1.82 1.39–2.39 o0.0001 CML-CP1 0.38 0.18–0.80 0.01 0.0

Abbreviations: RIC ¼ reduced-intensity conditioning; MAC ¼ myeloabla- 0 20 40 60 80 100 tive conditioning. Months from HSCT HLA matching and non-HLA variables signiﬁcantly associated with b survival were included in the ﬁnal multivariate model. 1.0 Bold values indicate Po0.05.

patients undergoing HSCT with acute leukaemia in ﬁrst 0.8 CR (‘early’) compared to patients with advanced disease (Figures 1b and c). Adjusted hazard ratios of OS for single 0.6 10/10-matched Mm pairs compared with 10/10 matched ones in early and 2-4 Mms advanced disease were 1.69 (95% CI 0.94–3.02, P 0.08) ¼ ¼ 0.4 1 Mm and 1.03 (95% CI ¼ 0.80–1.32, P ¼ 0.82), respectively. Survival probability 0.2 Engraftment The engraftment rate was 91% (694 out of 759 evaluable transplants). The median time to engraftment was 18 days 0.0

(range: 1–76). Univariate logistic regression using the 0 20 40 60 80 100 number of Mms as a continuous variable showed an Months from HSCT increased risk of graft failure with increasing number of Mms (HR ¼ 1.32, 95% CI ¼ 1.01–1.73, P ¼ 0.04). When c evaluated in multivariate analysis, only the presence of Mm 1.0 at locus DQB1 seemed to be strongly associated with graft failure (HR ¼ 3.26, 95% CI ¼ 1.02–10.38, P ¼ 0.04), whereas 0.8 two or more Mms did not (Table 2); however, heterogeneity tests showed no significant difference between loci DQB1 vs 0.6 A, B or C (P ¼ 0.45). Other factors significantly associated with graft failure were donor’s age, use of anti-thymocyte 0.4 10/10-matched globulin and infusion of BM stem cells. 1 Mm Survival probability 2-4 Mms Acute GVHD grade 2–4 0.2 Overall rates of aGVHD grades 2–4 were 27, 40 and 39% for 10/10, 9/10 and p8/10 matched pairs. Incidence of 0.0 aGVHD grades 3–4 were 9, 21 and 16%, respectively. 0 20 40 60 80 100 Testing for HLA matching and occurrence of aGVHD Months from HSCT resulted in a higher risk for pairs with two or more Mms (HR ¼ 1.68, 95% CI ¼ 1.05–2.69, P ¼ 0.03). Among pairs with Figure 1 Survival estimates according to HLA-matching status. Kaplan– Meier probabilities of OS in patients transplanted from 10/10, 9/10 and a single Mm, only locus B resulted in a significant association p8/10 matched donors were analyzed in all patients (a) and in patients with a higher risk of grades II–IV aGVHD (HR ¼ 2.25, 95% with ‘early’ (b) or ‘advanced’ (c) disease status. CI ¼ 1.10–4.61, P ¼ 0.03); in addition, no significant heterogeneity exists between loci B vs A, C or DQB1. Other significant non-HLA factors were: donor’s age, patient’s CMV Non-relapse mortality status and myeloablative conditioning regimen. Similar to survival analysis, no significant differences in non-relapse mortality were observed between pairs with Relapse-related death one Mm compared with 10/10 matched ones. A signifi- No association was present between HLA matching and cantly higher HR is associated with two or more Mms death because of relapse; therefore, no multivariate model (Table 2). Other significant non-HLA factors present in the was created for this end point. multivariate model were: donor’s gender, patient’s age at

Bone Marrow Transplantation Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 575 HSCT, year of HSCT, disease status at HSCT and disease (including lymphoma), a 9/10-matched donor could conditioning regimen. be suitable for these patients. Even in the presence of one or multiple HLA Mms, the survival of early patients seems to be superior compared with that of advanced patients, Discussion showing that performing HSCT earlier in the course of the disease could improve patients’ prognosis and that when a In this study, we evaluated the impact of patient–donor clinical indication for transplantation exists, the lack of a HLA compatibility on the HSCT outcome of a large 10/10 matched donor should not preclude HSCT and the number of adult patients mostly receiving in vivo T-cell choice of a partially mismatched donor is warranted. depleted allografts. Two levels of information emerge from In conclusion, this study adds useful information about the analysis. the role of HLA matching on the outcome of adult patients First, the relative value of low-resolution Mm seemed to undergoing HSCT from a unrelated donor, thus helping be similar to that of high-resolution Mm, as no difference physicians facing the search of a suitable donor for a of HSCT outcome emerged when a single low-resolution haematological malignancy; it could also represent the Mm was compared with a high-resolution one. Studies basis for new emerging studies, possibly providing HLA- published earlier indicate conflicting results; in fact, both matching algorithms and/or new antigens that better equivalent3,11 and different7 effects on outcome were correlate with prognosis and more precisely define the reported; our results suggest that both low- and high- association with HSCT outcomes.4,9,22–26 resolution Mms should be considered equivalent in the search for the optimal donor–patient HLA matching at present. However, we cannot exclude that we did not find Conflict of interest any statistical difference because of the relatively small numbers (a total of 171 vs 109 pairs with single low- The authors declare no conflict of interest. resolution and high-resolution Mm, respectively). Second, the presence of one or more Mms was associated Acknowledgements with impaired survival as compared with 10/10 matched pairs, both in unadjusted and adjusted models (Table 2). The following institutions (GITMO centers) contributed to the The predictive role of one single Mm was different in early study: Division of Hematology, Ospedale ‘SS Antonio e Biagio’ as compared with advanced disease. In fact, in the subset of Alessandria (A Levis); Division of Hematology, Nuovo Ospe- advanced disease, the choice of a one-Mm donor led to dale ‘Torrette’ Ancona (P Leoni); Division of Hematology, survival rates similar to 10/10 matched pairs in the same Ospedali Riuniti, Bergamo (A Rambaldi); Institute of Hematol- group (Figure 1c), with 2-year survival estimates of 40 and ogy and Clinical Oncology ‘L A Seragnoli,’ Ospedale ‘S Orsola- 39%, respectively. On the other hand, 2-year OS was 63 Malpighi,’ University of Bologna, Bologna (G Bandini, and 54% for those with a 10/10 matched and 9/10 matched F Bonifazi); Department of Hematology, Ospedale Regionale, donor, respectively, among patients with early disease. This Bolzano (M Casini); Division of Hematology and Bone Marrow Transplantation Center, Ospedale Oncologico ‘A Businco,’ is probably explained by the fact that disease status at Cagliari (E Angelucci, D Baronciani); Bone Marrow Transplan- HSCT represents the strongest predictor of survival, and tation Unit, Ospedale ‘R Binagli,’ University of Cagliari, that HLA matching acquires a smaller impact on prognosis Cagliari (G La Nasa); Division of Hematology and Bone in advanced patients. Interestingly, our registry reported a Marrow Transplantation, Ospedale ‘Ferrarotto,’ Catania (G higher number of lymphoproliferative diseases compared Milone); Division of Hematology, Ospedale ‘S Croce e Carlo,’ with other studies (here 205 out of 805 patients, 25%), and Cuneo (N Mordini); Department of Hematology, Ospedale an analysis of this subgroup revealed that no effect on ‘Careggi,’ University of Florence, Firenze (S Guidi, A Bosi); outcome is played by HLA matching (data not shown). Division of Hematology, Ospedale ‘S Martino,’ Genova Although the relatively small number of these cases and the (A Bacigalupo, M T Van Lint); Division of Hematology and retrospective nature of the subgroup analysis do not allow Oncology, IRCCS ‘G Gaslini’, Genova (G Dini); Hematology– Bone Marrow Transplantation Unit, Istituto Nazionale dei us to draw definite conclusions, this could partially explain Tumori, University of Milano, Milano (P Corradini, R Milani), why we did not detect a significant impairment of survival Division of Hematology Ospedale ‘Ca` Granda’ Niguarda, played by a single Mm overall. Other relevant differences Milano (E Morra, P Marenco); Department of Hematology, from reported large analyses3,12 are the wider use of anti- Fondazione IRCCS Ospedale Maggiore Policlinico, Mangiagalli thymocyte globulin (64% of patients in our series) and the e Regina Elena, Milano (G Lambretenghi Deliliers, F Onida); higher rate of RIC transplants, together with a lesser extent Hematology and Bone Marrow Transplantation Unit, of TBI use, factors all known to affect GVHD occurrence S Raffaele Scientific Institute, Milano (F Ciceri, M Marcatti); and survival. It is worth noting that all patient–donor pairs Transplantation Unit Department of Oncology–Hematology, on this series were high-resolution matched at locus DRB1, IRCCS Clinica Humanitas, Rozzano (L Castagna); Department different from most of the published studies: for this of Oncology and Hematology University of Modena and Reggio Emilia, Modena (F Narni); Division of Hematology and reason, no comparisons regarding this locus can be made. Transplant Unit, Ospedale ‘S Gerardo,’ University of Milano- Our findings confirm that the best matching for adult Bicocca, Monza (P Pioltelli), Division of Hematology, patients with acute leukaemia in first CR undergoing University of Napoli ‘Federico II’ Medical School, Napoli HSCT from unrelated donor is the 10/10 allele-matched; (C Selleri); Clinica Onco-Ematologica, University of Padova, however, since a single patient-donor HLA Mm did not Padova (L Zanesco); Division of Hematology and Transplant significantly affect survival of patients with advanced Unit, Ospedale ‘V Cervello,’ Palermo (R Scime` ); Department

Bone Marrow Transplantation Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 576 of Oncology, Hematology Unit, Ospedale ‘La Maddalena,’ 7 Flomenberg N, Baxter-Lowe LA, Confer D, Fernandez-Vina Palermo (M Musso), Division of Hematology, University of M, Filipovich A, Horowitz M et al. Impact of HLA class I and Pavia, Fondazione IRCCS Policlinico ‘S Matteo,’ Pavia (EP class II high-resolution matching on outcomes of unrelated Alessandrino); Pediatric Hematology–Oncology University of donor bone marrow transplantation: HLA-C mismatching is Pavia, Fondazione IRCCS Policlinico ‘S Matteo,’ Pavia associated with a strong adverse effect on transplantation (F Locatelli); Hematology and Transplant Center, Ospedale outcome. Blood 2004; 104: 1923–1930. ‘S Salvatore,’ Pesaro (G Visani); Department of Hematology, 8 Greinix HT, Fae´ I, Schneider B, Rosenmayr A, Ospedale Civile, Pescara (P Di Bartolomeo); Division of Mitterschiffthaler A, Pelzmann B et al. Impact of HLA class Hematology, University of Pisa, Pisa (F Papineschi); Division I high-resolution mismatches on chronic graft-versus-host of Onco-Hematology, Azienda Ospedaliera Universitaria Pisa- disease and survival of patients given hematopoietic stem cell na, Pisa (C Favre); Transplant Unit ‘A Neri,’ Ospedale ‘Bianchi- grafts from unrelated donors. Bone Marrow Transplant 2005; Melacrino-Morelli,’ Reggio Calabria (G Messina); Division of 35: 57–62. Hematology, Department of Cellular Biotechnologies and 9 Wade JA, Hurley CK, Takemoto SK, Thompson J, Davies Hematology, University ‘La Sapienza’ (A P Iori, R Foa` ); SM, Fuller TC et al. HLA mismatching within or outside of Hematology and Stem Cell Transplantation Unit Ospedale cross-reactive groups (CREGs) is associated with similar ‘S Camillo,’ Roma (A Locasciulli, I Majolino); Hematology, outcomes after unrelated hematopoietic stem cell transplanta- University ‘S Cuore,’ Roma (P Chiusolo, G Leone); Hemato- tion. Blood 2007; 109: 4064–4070. Oncology Transplant Unit, University ‘Tor Vergata,’ Roma (W 10 Loiseau P, Busson M, Balere ML, Dormoy A, Bignon JD, Arcese, R Cerretti); Unit of Hematology and Bone Marrow Gagne K et al. HLA Association with hematopoietic stem Transplantation, IRCCS, ‘Casa Sollievo della Sofferenza,’ S cell transplantation outcome: the number of mismatches at Giovanni Rotondo (A M Carella, N Cascavilla); Division of HLA-A, -B, -C, -DRB1, or -DQB1 is strongly associated Hematology and Bone Marrow Transplantation, Policlinico ‘Le with overall survival. Biol Blood Marrow Transplant 2007; 13: Scotte’, Siena (F Lauria); Institute of Hematology, Ospedale 965–974. Nord, Taranto (P Mazza); Division of Hematology, Ospedale ‘S 11 Petersdorf EW, Anasetti C, Martin PJ, Gooley T, Radich J, Giovanni Battista,’ Torino (M Falda); Division of Hematology, Malkki M et al. Limits of HLA mismatching in unrelated University of Torino, Ospedale ‘S Giovanni Battista,’ Torino (B hematopoietic cell transplantation. Blood 2004; 104: 2976–2980. Bruno, M Boccadoro); Center of Stem Cell Transplantation and 12 Morishima Y, Sasazuki T, Inoko H, Juji T, Akaza T, Cellular Therapy, Ospedale Infantile Regina Margherita, Torino Yamamoto K et al. The clinical significance of human (F Fagioli); Division of Hematology and Bone Marrow leukocyte antigen (HLA) allele compatibility in patients Transplantation, University of Udine, Udine (R Fanin, M receiving a marrow transplant from serologically HLA-A, Cerno); Division of Hematology-Bone Marrow Transplantation HLA-B, and HLA-DR matched unrelated donors. Blood 2002; Unit, Policlinico ‘GB Rossi’, Verona (F Benedetti) and 99: 4200–4206. Department of Hematology, Ospedale ‘S Bortolo,’ Vicenza 13 Sasazuki T, Juji T, Morishima Y, Kinukawa N, Kashiwabara (R Raimondi). 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. References N Engl J Med 1998; 339: 1177–1185. 14 Petersdorf EW, Hansen JA, Martin PJ, Woolfrey A, Malkki 1 Ljungman P, Urbano-Ispizua A, Cavazzana-Calvo M, M, Gooley T et al. Major-histocompatibility-complex class I Demirer T, Dini G, Einsele H et al. Allogeneic and autologous alleles and antigens in hematopoietic-cell transplantation. transplantation for haematological diseases, solid tumours and N Engl J Med 2001; 345: 1794–1800. immune disorders: definitions and current practice in Europe. 15 Dini G, Cancedda R, Locatelli F, Bosi A, Bandini G, Bone Marrow Transplant 2006; 37: 439–449. Alessandrino EP et al. Italian Bone Marrow Transplant 2 Hurley CK, Setterholm M, Lau M, Pollack MS, Noreen H, Group (GITMO). Unrelated donor marrow transplantation: Howard A et al. Hematopoietic stem cell donor registry an update of the experience of the Italian Bone Marrow strategies for assigning search determinants and matching Transplant Group (GITMO). Haematologica 2000; 85: relationships. Bone Marrow Transplant 2004; 33: 443–450. 30–36. 3 Lee SJ, Klein J, Haagenson M, Baxter-Lowe LA, Confer DL, 16 Lanino E, Lamparelli T, Dini G, Alessandrino EP, Aversa F, Eapen M et al. High-resolution donor-recipient HLA match- Calori E et al. Bone marrow transplantation from unrelated ing contributes to the success of unrelated donor marrow donors: the Italian experience. GITMO, AIEOP and IBMDR. transplantation. Blood 2007; 110: 4576–4583. Bone Marrow Transplant 1993; 11(Suppl 1): 88–89. 4 Kawase T, Morishima Y, Matsuo K, Kashiwase K, Inoko H, 17 Available from: http://www.ebmt.org/1whatisebmt/Op_Manual/ Saji H et al. Japan Marrow Donor Program. High-risk HLA OPMAN_StatGuidelines_oct2003.pdf. allele mismatch combinations responsible for severe acute 18 Glucksberg H, Storb R, Fefer A, Buckner CD, Neiman PE, graft-versus-host disease and implication for its molecular Clift RA et al. Clinical manifestations of graft-versus-host mechanism. Blood 2007; 110: 2235–2241. disease in human recipients of marrow from HL-A-matched 5 Petersdorf EW, Gooley TA, Anasetti C, Martin PJ, Smith AG, sibling donors. Transplantation 1974; 18: 295–304. Mickelson EM et al. Optimizing outcome after unrelated 19 Kaplan EL, Meier P. Non-parametric estimation from marrow transplantation by comprehensive matching of HLA incomplete observations. J Am Stat Assoc 1958; 53: 457–481. class I and II alleles in the donor and recipient. Blood 1998; 92: 20 Mantel N. Evaluation of survival data and two new rank order 3515–3520. statistics arising in its consideration. Cancer Chemother Rep 6 Petersdorf EW, Gooley T, Malkki M, Horowitz M., for The 1966; 50: 163–170. International Histocompatibility Working Group in Hemato- 21 Cox DR. Regression models and life tables. J Royal Stat Soc poietic Cell Transplantation. Clinical significance of donor– 1972; 34 (Series B): 187–220. recipient HLA matching on survival after myeloablative 22 Fleischhauer K, Locatelli F, Zecca M, Orofino MG, Giardini hematopoietic cell transplantation from unrelated donors. C, De Stefano P et al. Graft rejection after unrelated donor Tissue Antigens 2007; 69(Suppl 1): 25–30. hematopoietic stem cell transplantation for thalassemia is

Bone Marrow Transplantation Donor–patient HLA matching and outcome of UD HSCT R Crocchiolo et al 577 associated with nonpermissive HLA-DPB1 disparity in host- 25 Miller JS, Cooley S, Parham P, Farag SS, Verneris MR, versus-graft direction. Blood 2006; 107: 2984–2992. McQueen KL et al. Missing KIR ligands are associated with 23 Shaw BE, Marsh SG, Mayor NP, Russell NH, Madrigal JA. less relapse and increased graft-versus-host disease (GVHD) HLA-DPB1 matching status has signiﬁcant implications for following unrelated donor allogeneic HCT. Blood 2007; 109: recipients of unrelated donor stem cell transplants. Blood 2006; 5058–5061. 107: 1220–1226. 26 Le Maux A, Noe¨l G, Birebent B, Grosset JM, Vu N, De 24 Tseng LH, Lin MT, Hansen JA, Gooley T, Pei J, Smith AG Guibert S et al. Soluble human leucocyte antigen-G molecules et al. Correlation between disparity for the minor histocom- in peripheral blood haematopoietic stem cell transplantation: a patibility antigen HA-1 and the development of acute graft- speciﬁc role to prevent acute graft-versus-host disease and a versus-host disease after allogeneic marrow transplantation. link with regulatory T cells. Clin Exp Immunol 2008; 152: Blood 1999; 94: 2911–2914. 50–56.

Bone Marrow Transplantation