Scholarly article on topic 'Impact on Outcomes of Human Leukocyte Antigen Matching by Allele-Level Typing in Adults with Acute Myeloid Leukemia Undergoing Umbilical Cord Blood Transplantation'

Impact on Outcomes of Human Leukocyte Antigen Matching by Allele-Level Typing in Adults with Acute Myeloid Leukemia Undergoing Umbilical Cord Blood Transplantation Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Umbilical cord blood" / "Human leukocyte antigen (HLA)" / "Acute myeloid leukemia"}

Abstract of research paper on Clinical medicine, author of scientific article — Jaime Sanz, Francisco J. Jaramillo, Dolores Planelles, Pau Montesinos, Ignacio Lorenzo, et al.

Abstract This retrospective study analyzed the impact of directional donor-recipient human leukocyte antigen (HLA) disparity using allele-level typing at HLA-A, -B, -C, and -DRB1 in 79 adults with acute myeloid leukemia (AML) who received single-unit umbilical cord blood (UCB) transplant at a single institution. With extended high-resolution HLA typing, the donor-recipient compatibility ranged from 2/8 to 8/8. HLA disparity showed no negative impact on nonrelapse mortality (NRM), graft-versus-host (GVH) disease or engraftment. Considering disparities in the GVH direction, the 5-year cumulative incidence of relapse was 44% and 22% for patients receiving an UCB unit matched ≥ 6/8 and < 6/8, respectively (P = .04). In multivariable analysis, a higher HLA disparity in the GVH direction using extended high-resolution typing (Risk ratio [RR] 2.8; 95% confidence interval [CI], 1.5 to 5.1; P = .0009) and first complete remission at time of transplantation (RR 2.1; 95% CI, 1.2 to 3.8; P = .01) were the only variables significantly associated with an improved disease-free survival. In conclusion, we found that in adults with AML undergoing single-unit UCBT, an increased number of HLA disparities at allele-level typing improved disease-free survival by decreasing the relapse rate without a negative effect on NRM.

Academic research paper on topic "Impact on Outcomes of Human Leukocyte Antigen Matching by Allele-Level Typing in Adults with Acute Myeloid Leukemia Undergoing Umbilical Cord Blood Transplantation"

Impact on Outcomes of Human Leukocyte Antigen Matching by Allele-Level Typing in Adults with Acute Myeloid Leukemia Undergoing Umbilical Cord Blood Transplantation

Jaime Sanz*, Francisco J. Jaramillo, Dolores Planelles, Pau Montesinos, Ignacio Lorenzo, Federico Moscardó, Guillermo Martin, Francisca López, Jesús Martínez, Isidro Jarque, Javier de la Rubia, Luis Larrea, Miguel A. Sanz, Guillermo F. Sanz

Department of Hematology, Hospital Universitari i politecnic La Fe, Spain

American Society for Blood and Marrow Transplantation

Article history: Received 22 July 2013 Accepted 21 October 2013

Key Words: Umbilical cord blood Human leukocyte antigen (HLA)

Acute myeloid leukemia

ABSTRACT

This retrospective study analyzed the impact of directional donor-recipient human leukocyte antigen (HLA) disparity using allele-level typing at HLA-A, -B, -C, and -DRB1 in 79 adults with acute myeloid leukemia (AML) who received single-unit umbilical cord blood (UCB) transplant at a single institution. With extended highresolution HLA typing, the donor-recipient compatibility ranged from 2/8 to 8/8. HLA disparity showed no negative impact on nonrelapse mortality (NRM), graft-versus-host (GVH) disease or engraftment. Considering disparities in the GVH direction, the 5-year cumulative incidence of relapse was 44% and 22% for patients receiving an UCB unit matched > 6/8 and < 6/8, respectively (P = .04). In multivariable analysis, a higher HLA disparity in the GVH direction using extended high-resolution typing (Risk ratio [RR] 2.8; 95% confidence interval [CI], 1.5 to 5.1; P = .0009) and first complete remission at time of transplantation (RR 2.1; 95% CI, 1.2 to 3.8; P = .01) were the only variables significantly associated with an improved disease-free survival. In conclusion, we found that in adults with AML undergoing single-unit UCBT, an increased number of HLA disparities at allele-level typing improved disease-free survival by decreasing the relapse rate without a negative effect on NRM.

© 2014 American Society for Blood and Marrow Transplantation.

INTRODUCTION

One intrinsic advantage of umbilical cord blood (UCB) over other stem cell sources for allogeneic stem cell transplantation is the ability to tolerate greater degrees of human leukocyte antigen (HLA) disparity. However, the impact of donor-recipient HLA mismatch on umbilical cord blood transplantation (UCBT) outcomes is still not well established, particularly in adults.

Most registry-based studies that included a majority of pediatric population and a variety of diseases have reported an increased nonrelapse mortality (NRM) rate with HLA disparity considering HLA-A and -B by low-resolution and DRB1 by high-resolution techniques [1-4], although this negative effect may be restricted to mismatches in the rejection direction [5]. However, the impact on outcomes of HLA disparity seems different in the adults, with no increase in NRM and a possible benefit in relapse [6]. Although closer match at allelic level has consistently shown improved outcomes after unrelated bone marrow or peripheral blood stem cell transplantation [7], very few studies have evaluated the importance of extended HLA typing at allele level in UCBT.

We hypothesized that a higher degree of HLA mismatch could enhance the graft-versus-leukemia (GVL) effect, decreasing the relapse risk and potentially affecting other outcomes. We, therefore, analyzed the short- and long-term

Financial disclosure: See Acknowledgments on page 109.

* Correspondence and reprint requests: Jaime Sanz, MD, Hospital Universitario La Fe, Hematology, Bulevar S/N, Valencia, Spain. E-mail address: sanz_jai@gva.es (J. Sanz).

outcomes of a series of adults with acute myeloid leukemia (AML) who received UCBT from unrelated donors at a single institution, focusing on the impact on outcomes of donor-recipient HLA match, using extended allele-level typing.

PATIENTS AND METHODS Eligibility Criteria

Between June 2000 and December 2012, 79 consecutive adult patients with AML underwent myeloablative UCBT from unrelated donors at our institution. Patients were eligible for enrollment if they met the following criteria: (1) high-risk AML, defined by either high-risk cytogenetics or poor prognosis gene markers, more than 1 cycle to achieve complete remission (CR), or disease status beyond first CR; and (2) no suitable related donor (HLA-identical or 1-antigen—mismatched). From 2005, patients were included in 2 subsequent prospective trials: TSCU-GETH2005 (21 patients) and TSCU-GETH/GITM02008 (42 patients). The institutional review board approved the protocol and written informed consent was obtained from all patients according to the Declaration of Helsinki. TSCU-GETH/GITM02008 clinical trial was registered on the EudraCT with code 2008-000927-24.

The treatment plans, including of graft selection, conditioning regimen, immune suppression, and supportive care have been reported in detail previously [8,9], and are summarized below.

Umbilical Cord Blood Unit Selection and Management

The search of UCB units was conducted by the Spanish Registry of Bone Marrow Donors (Registro Español de Donantes de Médula Ósea). Graft selection algorithm required UCB > 4/6 HLA matched with the recipient (HLA class I antigens [A and B] considering the antigen level and class II antigens [DRB1] considering allele-level resolution DNA typing). Minimum total nucleated cell (TNC) and CD34+ cell counts from the information provided by the different UCB banks were required. Minimum cell dose criteria changed over time. A TNC dose > 1.5 x 107/kg recipient's body weight was required until 2005, whereas TNC > 2 x 107/kg and CD34+ cell dose > 1 x 105/kg recipient's body weight were required in 2006 and 2007. From 2008 to present, minimum cell dose criteria considered total cell dose of the UCB units without taking into account recipient's body weight as follows: TNC >

1083-8791/$ — see front matter © 2014 American Society for Blood and Marrow Transplantation. http://dx.doi.Org/10.1016/j.bbmt.2013.10.016

150 x 107 and CD34+ cells > 70 x 105. Cell dose was considered the most important criteria for unit selection.

Conditioning Regimen

Two subsequent myeloablative conditioning regimens were used, all based on the combination of thiotepa, busulfan, cyclophospamide or flu-darabine, and antithymocyte globulin (ATG). Until March 2005,16 patients received thiotepa (10 mg/kg), busulfan (12 mg/kg orally in 8 patients or 9.6 mg/kg i.v. in 8 patients), cyclophosphamide (120 mg/kg) and ATG (Lym-phoglobuline, Merieux, Lyon, France; 60 mg/kg in the first 6 patients or Thymoglobulin, Genzyme Transplant, Cambridge, MA; 8 mg/kg in the remaining 10 patients). From March 2005, the remaining 63 patients received thiotepa (10 mg/kg), i.v. busulfan as a single dose (9.6 mg/kg), fludarabine (150 mg/m2), and ATG (Thymoglobulin 8 mg/kg in the first 21 patients and 6 mg/kg in the remaining 42 patients).

Graft-versus-host-disease Prophylaxis and Treatment

All patients received cyclosporine 1.5 mg/kg/12 hours i.v., followed by 3 to 5 mg/kg/12 hours orally when oral intake was possible, with slow tapering starting between day +90 and +180 and discontinuation on day +180, or before if feasible. Cyclosporine was combined with long-course prednisone in the first 38 patients (.5 mg/kg/day on days +7 to +14,1 mg/ kg/day on days +14 to +28, with slow tapering until discontinuation on day +180), micofenolate-mofetil (15 mg/kg/12 hours until day +28) in the following 26 patients, or a short course of prednisone in the last 15 patients (1 mg/kg/day on days +14 to +28). Patients developing acute graft-versus-host disease (GVHD) received high-dose methylprednisolone as initial therapy (20 mg/kg/day; halving the dose every 3 days until reaching 1 mg/ kg/day, and then gradually tapered), followed by ATG in refractory cases. Chronic GVHD was treated with prednisone 1 mg/kg/day.

HLA Typing

For confirmation purposes, all patients and UCB units were HLA typed in our laboratory (D.P.). High-resolution typing was performed for HLA-A, -B, -C, and DRB1 in all patient and donor pairs as previously described [10]. High-resolution (allele level) and low-resolution (serologic or antigen level) HLA matching were considered in the analysis of clinical outcomes. Low-resolution disparities involved conversion of the DNA-based typing to its lower level serologic equivalent, by collapsing the 4-digit typing result back to its first 2 digits. Directional mismatches were considered in the analysis. Mismatch in the graft-versus-host (GVH) direction was defined in homozygous donors at an HLA locus when the recipient shared 1 HLA antigen with the donor. Mismatch in the host-versus-graft (HVG) direction was defined in homozygous recipients at an HLA locus when the donor shared 1 HLA antigen with the recipient. Mismatches at homozygous alleles were considered single mismatches.

Definitions

Treatment outcomes were assessed according to the revised criteria by Cheson et al. [11]. Myeloid engraftment was defined as an absolute neutrophil count (ANC) of .5 x 109/L or greater on 3 consecutive days. Platelet engraftment was defined as a platelet count of20 x 109/L or higher, without transfusion support, for 7 consecutive days. Patients who survived more than 28 days after transplantation and who failed to achieve myeloid engraftment were considered graft failures. Time to myeloid or platelet engraftment was defined as the time required to reach the first day of engraftment. Secondary graft failure was defined as the loss of the engraftment. Acute and chronic GVHD were defined and graded according to standard criteria [12-14]. Killer cell immunoglobulin-like receptor (KIR)-ligand compatibility was defined as described by Willemze et al. [15]. Briefly, KIR-ligand compatibility in the GVH direction was determined according to whether or not they expressed HLA-C group 1 or 2, HLA-Bw4, or HLA-A3/-A11. NRM was defined as death from any cause without evidence of relapse. Disease-free survival was defined as survival from the time of transplantation without evidence of disease relapse.

Statistical Analysis

The probabilities of engraftment, NRM, GVHD, and relapse were estimated by the cumulative incidence method (marginal probability) [16,17]. For cumulative incidence analyses of engraftment, GVHD, and relapse, death in CR was considered as a competing cause of failure, whereas relapse was the competing event for NRM. Unadjusted time-to-event analyses were performed using the Kaplan-Meier estimate [18], and, for comparisons, the log-rank tests [19]. Disease-free survival was calculated from the date of UCBT. In the analysis of leukemia-free survival (LFS), relapse, or death in CR, whichever occurred first, was considered an uncensored event. The follow-up of the patients was updated on December 1, 2012. A Cox proportional hazards model [20] or the Fine and Gray method for competing events [21]

were used for multivariable analysis using variables with a P value < .10 for each endpoint. The variables considered for prognostic factor analysis were age, gender, recipient weight, recipient CMV serology, disease status at transplantation, HLA compatibility considering HLA-A, -B, -C, and DRB1 by high- and low-resolution typing, ABO blood group mismatch, KIR-ligand incompatibility, conditioning regimen, GVHD prophylaxis, and TNC and CD34+ cells before freezing and infused. Continuous variables were dichotomized at the most discriminative cutoff point for each outcome. The slight modifications in the conditioning regimen and GVHD prophylaxis performed over the years did not have an impact on any outcome and. Therefore, subgroup analysis was not performed, although these variables were included in the multivariable analysis. Statistical analyses were conducted using R version 2.12.2 (The CRAN project) with packages, survival v2.37-4, rms 3.6-3, prodlim v1.3.3, and cmprsk v2.2-4 [22].

RESULTS

Patient, Graft, and Transplantation Characteristics

Patient characteristics are shown in Table 1. Briefly, median age was 37 years (range, 16 to 55) and 55 patients (72%) were in CR1 or CR2.

Graft and transplantation characteristics are shown in Table 2. Briefly, 26 (33%) males received a graft from a female donor and KIR-ligand mismatch was present in 28 (35%) patients.

Donor-recipient HLA Matching

Table 3 shows the number of HLA disparities between donors and recipients using the standard HLA match criteria used for unit selection, as well as the extended allele-level typing at HLA-A, -B, -C, and DRB1. The number of HLA disparities was considered in both GVH and HVG direction.

With the standard criteria, 6 patients (8%) received a fully matched 6/6 UCB unit whereas 19 patients (24%) and 54 patients (68%) received a 5/6 and 4/6 matched graft, respectively. With extended high-resolution HLA typing, the donor-recipient compatibility ranged from 2/8 to 8/8. Ten donors had at least 1 homozygous allele resulting in a unique vector in the GVH direction, whereas 7 patients had at least 1

Table 1

Characteristics of Patients with AML Undergoing UCBT

Characteristic n

No. of patients 79

Age, median (range), yr 37(16-55)

Gender, n (%)

Male 50 (63)

Female 29 (37)

Weight, median (range) kg 71 (37-112)

Disease status at transplantation

First complete remission 46 (58)

Second complete remission 11(14)

More advanced phase 22 (28)

Diagnosis

De novo 70 (89)

Therapy- or MDS-related secondary AML 9(11)

Cytogenetics*

Favorable 2 (3)

Intermediate 42 (53)

Poor 24 (30)

Unknown 11 (14)

Previous autologous transplantation 7(9)y

Cytomegalovirus serologic status before transplantation

Positive 60 (76)

Negative 19 (24)

AML indicates acute myeloid leukemia; UCBT, umbilical cord blood transplantation; MDS, myelodysplastic syndrome.

Percentages may total 100 because of rounding. Data presented are n (%), unless otherwise indicated. * Cytogenetic risk according to European Leukemia Net classification. y One patient underwent 2 autologous transplantations.

Table 2

Graft and Transplantation Characteristics

Characteristic n (%) Donor-recipient sex match

Male-male 24 (30)

Male-female 13 (l6)

Female-male 26 (33)

Female-female 16 (20) KIR-ligand incompatibility in graft-versus-host direction

Yes 28 (35)

No 51(65) Conditioning regimen

TT + BU + CY + ATG 16 (20)

TT + BU + FLU + ATG 63 (80) Graft-versus-host disease prophylaxis

Cyclosporine A + prednisone 53 (67)

Cyclosporine A + MMF 26 (33) No. of nucleated cells before freezing, x 107/kg

Median 3

Range 1.4-6.8 No. of nucleated cells infused, x 107/kg

Median 2.4

Range 1.0-4.9 No. of CD34+ cells before freezing, x 105/kg

Median 1.6

Range .2-4.1 No. of CD34+cells infused, x 105/kg

Median 1.3

Range .1-6.1

TT indicates thiotepa; BU, busulfan; CY, cyclophosphamide; FLU, fludar-abine; ATG, antithymocyte globulin; MMF, micofenolate-mofetil. Percentages may total 100 because of rounding. Data presented are n (%), unless otherwise indicated.

homozygous allele resulting in a unique vector in the HVG direction.

Engraftment and GVHD

The cumulative incidence of myeloid engraftment at 57 days was 95% (95% confidence interval [CI], 89 to 99) at a median time to neutrophil recovery of 20 days (range, 11 to 57). The cumulative incidence of platelet engraftment at 180 days was 77% (95% CI, 68% to 87%) at a median time to platelet recovery of 47 days (range, 23 to 138).

The cumulative incidence of acute GVHD at 100 days grade II to IV and III to IV was 32% (95% CI, 22% to 42%) and 17% (95% CI, 8% to 25%), respectively. The cumulative incidence of chronic extensive GVHD at 3 years was 38% (95% CI, 26% to 50%).

HLA disparity had no significant impact in multivariable analysis on myeloid and platelet engraftment, or acute and chronic GVHD.

Nonrelapse Mortality

Thirty-one patients died without relapse at a median of 181 days (range, 19 to 2535). The 5-year cumulative incidence of NRM for the entire cohort was 35% (95% CI, 25% to

Table 3

Number of HLA Mismatches between Donor and Recipient

46%). HLA disparity had no significant impact on NRM. In multivariate analysis TNC < 2.4 x 107/kg was the only variable significantly associated with a higher NRM (relative risk [RR], 2.7; 95% CI, 1.3 to 5.9; P = .009).

Impact of HLA Mismatch on Relapse

Overall, 21 patients relapsed at a median time of 202 days (range, 32 to 1372). The 5-year cumulative incidence of relapse of the entire cohort was 28% (95% CI, 18% to 39%). For patients who underwent transplantation in CR, cumulative incidence of relapse was 24% compared with 41% for patients who underwent transplantation while in relapse (P = .08). With extended high-resolution HLA typing, bidirectional mismatches showed no impact on the risk of relapse; however, the 5-year cumulative incidence of relapse was 44% and 22% for patients receiving an UCB unit matched > 6/8 and < 6/8, respectively (P = .04) (Figure 1).

In multivariable analysis, a higher HLA disparity in the GVH direction using extended high-resolution typing (RR, .4; 95% CI, .2 to .9; P = .0004) and CR at time of transplantation (RR, .4; 95% CI, .2 to .9; P = .02) were the only variables significantly associated with a reduced risk of relapse.

Leukemia-free Survival

After a median follow-up of 66 months (range, 21 to 132) for surviving patients, 27 patients remained alive and disease free at last follow-up. The 5-year LFS for the entire cohort was 35% (95% CI, 23% to 46%). For patients who underwent transplantation in first CR, LFS was 41%, compared with 25% for patients who underwent transplantation in more advanced phases of the disease (P = .04).

Considering standard bidirectional HLA typing, the 5-year LFS was 40%, 26%, and 17% for patients who underwent transplantation with a 4/6, 5/6, and 6/6 UCB unit, respectively (P = .04). With standard HLA typing and considering mismatches in the GVH direction, the 5-year LFS was 45%, 21%, and 14% for patients who underwent transplantation with a 4/6, 5/6, and 6/6 UCB unit, respectively (P = .03).

With extended high-resolution HLA typing, bidirectional incompatibilities showed no impact on LFS, but considering mismatches in the GVH direction, the 5-year LFS was 45% and 10% for patients receiving an UCB unit matched 2 to 5/8 and 6 to 8/8, respectively (P = .003) (Figure 2).

In multivariable analysis, a higher HLA disparity in the GVH direction using extended high-resolution typing (RR, 2.8; 95% CI, 1.5 to 5.1; P = .0009) and CR1 at time of transplantation (RR, 2.1; 95% CI, 1.2 to 3.8; P = .01) were the only variables significantly associated with an improved LFS.

DISCUSSION

This study shows that a higher donor-recipient HLA disparity in myeloablative single-unit UCBT for adults with

Allele level HLA match for HLA-A, -B, -C and DRB1, n

Bidirectional Mismatches Mismatches in the Graft-versus-hosty Mismatches in the Host-versus-graftz

Direction Direction

2/8 3/8 4/8 5/8 6/8 7/8 8/8 2/8 3/8 4/8 5/8 6/8 7/8 8/8 2/8 3/8 4/8 5/8 6/8 7/8 8/8

Standard HLA match*

6/6 0 0 0 0 2 2 2 0 0 0 0 1 3 2 0 0 0 0 2 2 2

5/6 0 0 2 5 5 7 0 0 0 2 5 5 7 0 0 0 2 5 4 7 1

4/6 2 4 13 33 2 0 0 2 3 12 31 3 2 1 1 3 14 30 5 1 0

Total 2 4 15 38 9 9 2 2 3 14 36 9 12 3 1 3 16 35 11 10 3

* HLA-A and -B at the antigen level and DRB1 at the allele level.

Figure 1. Cumulative incidence of relapse after UCBT according to the number of HLA disparities in the GVH direction.

AML decreases the risk of relapse, suggesting an enhanced GVL effect, which translates to a benefit on long-term LFS. We could also identify donor-recipient pairs with a high number of discrepancies by extended high-resolution HLA typing without an evident deleterious effect. A separate analysis of HLA disparity according to GVH and HVG direction improved the predictive value on relapse rate. These data may provide clinically useful information for a better selection of cord blood units in adults with poor-risk AML.

The number of patients and events in this study was smaller than in registry-based studies and did not allow an analysis of the impact of specific HLA disparities. Results should, therefore, be interpreted with caution. However, this retrospective study included a series of adults with AML treated with a homogenous strategy of single-unit UCBT after busulfan-based myeloablative conditioning regimen at a single institution. Of note, most patients were included in 2 subsequent prospective trials (TSCU-GETH2005 and TSCU-GETH/G1TM02008).

With extended high-resolution typing, we identified a significant proportion of patients that received highly

Figure 2. Kaplan-Meier estimate of leukemia-free survival after UCBT according to the number of HLA disparities in the GVH direction.

mismatched units. In fact, 70% of the donor-recipient pairs had 3 or more disparities at the allelic level. Compared to the standard HLA typing criteria, directional high-resolution typing was able to further discriminate disparities. For example, UCB units matched 4/6 with standard criteria had disparities that were reclassified ranging from 2/8 to 8/8 in the GVH direction at high resolution.

Our study showed no negative impact of HLA disparity on NRM, GVHD, or engraftment, whereas TNC content was the only variable associated with the risk of NRM. In contrast to pediatric patients, HLA disparity in adults has never shown a negative impact on NRM in large registry-based studies [6,23,24]. The different outcome in children compared with adults may be explained, at least in part, by the important differences in cell dose between these populations. Another important factor to consider is the importance of mismatch direction on outcomes. However, only few registry-based studies have evaluated its impact by serologic data with inconsistent results, maybe because of the heterogeneity of the population. Some studies have suggested a deleterious effect on engraftment with mismatches in the HVG direction [5,25], whereas others found an opposite effect [26]. UCBT performed with mismatches in the GVH direction had outcomes that were similar to those that were fully matched [5,25], but this effect was restricted to children in 1 of the studies [25]. No mismatch category has been previously associated with relapse.

We observed an impressive reduction in risk of relapse, suggesting an enhanced GVL effect, for patients who underwent transplantation with UCB units with a higher HLA disparity in the GVH direction. This observation was made in the context of a very high-risk population and a specific highly immunosuppressive treatment platform. Future studies are needed to determine whether this finding is true for all or only specific scenarios. In this regard, previous registry-based studies from Eurocord [27] and the Japanese Society for Hematopoietic Cell Transplantation group [6] have also found that UCBTs with a higher HLA disparity had a lower probability of relapse. However, we could not see any favorable effect of KIR-ligand incompatibility in the graft-versus-host direction as previously described [15]. A reduced relapse rate, together with the lack of a negative impact on NRM, translated into improved LFS after UCBT performed with more mismatched units in the GVH direction. To our knowledge, this has not been previously reported, maybe because of the fact that this was this was the first disease-specific analysis of an adult-restricted cohort evaluating the impact directional HLA mismatches by highresolution typing.

In conclusion, we found that in adults with AML undergoing single-unit UCBT with thiotepa, busulfan, fludarabine, and ATG, UCB units with high number of discrepancies seem acceptable for transplantation. In fact, an increased number of HLA disparities using allele-level typing improved disease-free survival by decreasing the relapse without increasing mortality rate. Whether these findings will apply in other settings, such as double-unit UCBT or using reduced-intensity conditioning, is unknown and should be further investigated. Further studies addressing the potential allor-eactivity and GVL effect of HLA disparity are warranted.

ACKNOWLEDGMENTS

The authors thank David Pellicer and Shirley Weiss for data collection and management.

Financial disclosure: The authors have nothing to disclose.

Conflict of interest statement: There are no conflicts of interest to report.

Authorship statement: J.S., F.J. and G.F.S. conceived the study and interpreted the data; J.S., M.A.S., and G.F.S. wrote the paper; J.S. performed the statistical analyses; D.P. performed the HLA typing; F.J., P.M., I.L., F.M., G.M., F.L., J.M., I.J., J.R., and L.L. reviewed the manuscript and contributed to the final draft.

REFERENCES

1. Eapen M, Rubinstein P, Zhang M-J, et al. Outcomes of transplantation of unrelated donor umbilical cord blood and bone marrow in children with acute leukaemia: a comparison study. Lancet. 2007;369: 1947-1954.

2. Kurtzberg J, Prasad VK, Carter SL, et al. Results of the Cord Blood Transplantation Study (COBLT): clinical outcomes of unrelated donor umbilical cord blood transplantation in pediatric patients with hema-tologic malignancies. Blood. 2008;112:4318-4327.

3. Barker JN, Scaradavou A, Stevens CE. Combined effect of total nucleated cell dose and HLA match on transplantation outcome in 1061 cord blood recipients with hematologic malignancies. Blood. 2010;115: 1843-1849.

4. Rubinstein P, Carrier C, Scaradavou A, et al. Outcomes among 562 Recipients of Placental-Blood Transplants from Unrelated Donors. N Engl J Med. 1998;339:1565-1577.

5. Stevens CE, Carrier C, Carpenter C, et al. HLA mismatch direction in cord blood transplantation: impact on outcome and implications for cord blood unit selection. Blood. 2011;118:3969-3978.

6. Atsuta Y, Kanda J, Takanashi M, et al. Different effects of HLA disparity on transplant outcomes after single-unit cord blood transplantation between pediatric and adult patients with leukemia. Haematologica. 2013;98:814-822.

7. Lee SJ, Klein J, Haagenson M, et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110:4576-4583.

8. Sanz J, Montesinos P, Saavedra S, et al. Single-unit umbilical cord blood transplantation from unrelated donors in adult patients with chronic myelogenous leukemia. Biol Blood Marrow Transplant. 2010;16: 1589-1595.

9. Sanz J, Boluda JCH, Martín C, et al. Single-unit umbilical cord blood transplantation from unrelated donors in patients with hematological malignancy using busulfan, thiotepa, fludarabine and ATG as myeloa-blative conditioning regimen. Bone Marrow Transplant. 2012;47: 1287-1293.

10. Sanz GF, Saavedra S, Planelles D, et al. Standardized, unrelated donor cord blood transplantation in adults with hematologic malignancies. Blood. 2001;98:2332-2338.

11. Cheson BD, Bennett JM, Kopecky KJ. Revised recommendations of the 1nternational Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in acute myeloid leukemia. J Clin Oncol. 2003;21: 4642-4649.

12. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation. 1974;18:295-304.

13. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant. 1995;15: 825-828.

14. Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graft-versus-host syndrome in man. A long-term clinicopathologic study of 20 Seattle patients. Am J Med. 1980;69:204-217.

15. Willemze R, Rodrigues CA, Labopin M, et al. KIR-ligand incompatibility in the graft-versus-host direction improves outcomes after umbilical cord blood transplantation for acute leukemia. Leukemia. 2009;23: 492-500.

16. Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695-706.

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

18. Kaplan EL, Meier P. Nonparametric Estimation from 1ncomplete Observations. J Am Statistical Assoc. 1958;53:457-481.

19. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966;50:163-170.

20. Cox DR. Regression models and life tables (with discussion). JRStatSoc B. 1972;34:187-220.

21. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Statistical Assoc. 1999;94:496-509.

22. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: The R Foundation for Statistical Computing; 2012.

23. Cohen YC, Scaradavou A, Stevens CE, et al. Factors affecting mortality following myeloablative cord blood transplantation in adults: a pooled analysis of three international registries. Bone Marrow Transplant. 2011;46:70-76.

24. Arcese W, Rocha V, Labopin M, et al. Unrelated cord blood transplants in adults with hematologic malignancies. Haematologica. 2006;91: 223-230.

25. Kanda J, Atsuta Y, Wake A, et al. Impact of the direction of HLA mismatch on transplantation outcomes in single unrelated cord blood transplantation. Biol Blood Marrow Transplant. 2013;19:247-254.

26. Matsuno N, Wake A, Uchida N, et al. Impact of HLA disparity in the graft-versus-host direction on engraftment in adult patients receiving reduced-intensity cord blood transplantation. Blood. 2009;114: 1689-1695.

27. Gluckman E, Rocha V, Arcese W, et al. Factors associated with outcomes of unrelated cord blood transplant: guidelines for donor choice. Exp Hematol. 2004;32:397-407.