In the current era, imatinib therapy has significantly altered the paradigm for selection of both nontransplantation and transplantation strategies. Initial reports of frequent cytogenetic responses with imatinib were published in 1999, and the drug was approved by the United States Food and Drug Administration in May 2001, followed by its widespread use.13,14
This was followed by a dramatic decrease in allogeneic transplantation procedures for CML.31
Currently, most centers now recommend allogeneic transplantation after failing imatinib or in later stages of the disease.13,31
Allogeneic transplantation, although curative, is associated with considerable mortality and morbidity with risks including GVHD, veno-occlusive disease of liver, infections, risks of secondary malignancy, and overall poorer quality of life.32
URD compared with MSD transplants have been associated with both higher incidence of GVHD as well as higher TRM.3
The gold standard donor for allogeneic transplantation remains an MSD. Most prior studies evaluating comparative outcomes between MSD and URD transplant recipients have been limited by either lack of molecular typing in the URD or by small sample sizes.3,4,33,34
We present results comparing outcomes in a large cohort of HLA matched and mismatched URD transplant recipients with those in MSD transplant recipients, while controlling for other major factors that may affect outcomes (disease, conditioning, disease stage, stem cell source). Majority of patients selected received transplants in the preimatinib era, with only 3% (n = 145; 89 MSD, 56 URD) receiving transplants after 2001, hence the impact of prior imatinib therapy was not evaluated.
Two recent studies from the CIBMTR15,24
evaluated the impact of locus specific mismatching on outcomes. In both these analyses, mismatching at either HLA-A, -B, -C, or -DRB1 was associated with worse outcomes. The analysis by Lee et al24
demonstrated that there was no difference between low- and high-resolution mismatch at a particular locus, hence the two were considered together for the purpose of this analysis. They also demonstrated no impact of mismatching at DP or DQ loci in their studies. We evaluated the impact of an DQ mismatch and found no independent influence on survival. However, a significant association was observed when HLA-DQB1 mismatching was present with additional HLA class I mismatches (data not shown); this observation confirms and extends those of previous studies which suggest that the additive effect of multiple mismatches is detrimental.24,30
In our analysis, overall survival after receipt of 8/8 matched URD transplant was closest to that after receipt of MSD transplant (63% v 55%) and declined with greater degrees of mismatch. Patients who received MSD and matched and mismatched URD transplants had similar risks of relapse; importantly, we did not observe lower risks of relapse with greater degrees of HLA mismatching. The risk of TRM was significantly higher in 8/8 matched URD than in MSD transplant recipients; the risk almost doubled in the presence of a single class I mismatch and more than tripled in the presence of mixed mismatches, yielding lower LFS in URD transplant recipients. LFS in 8/8 matched URD transplant recipients was closest to that in MSD transplant recipients but was progressively worse in the presence of mismatch in class I loci, due predominantly to higher TRM. T-cell depletion was used more frequently in URD transplant recipients, but was not frequent enough to allow comparison in each category of mismatch. It was, however, adjusted for in the multivariate models.
Although the current study did not address the outcomes in patients with more advanced stages of CML who received HLA matched or mismatched URD transplants, the results of the current analysis of CML CP1 patients suggest that overall transplantation outcome is defined by a balance of risks contributed by HLA disparity and by disease progression.
There have been few studies comparing outcomes in patients who received MSD and URD transplants. In an earlier report from NMDP3
, survival after receipt of MSD transplant was compared with that after receipt of URD transplant in patients with CML. Similar to our study, overall survival and disease-free survival (DFS) were only slightly (although significantly) lower in the cohort that received URD transplants, compared with the cohort that received MSD transplants. However, the population that received URD transplants in the earlier study was only serologically matched for HLA-A and -B, and matched by molecular typing only at HLA-DRB1. In that report, similar DFS in HCT involving MSDs and URDs was observed only in younger patients (< 30 years of age) undergoing transplantation within 1 year from diagnosis. In another study,33
outcomes in 55 10/10 allele-matched URD (HLA-A, -B, -C, -DRB1, and -DQB1) transplant recipients were compared with those in 181 MSD transplant recipients for standard-risk hematologic malignancies, and similar outcomes were reported in the two cohorts. This study included 43 patients with CML (30 MSD transplant recipients and 13 URD transplant recipients) who were in either chronic or accelerated (n = 4) phase. The Australian Bone Marrow Transplant Registry reported a case control analysis of outcomes in 105 URD transplant recipients and 105 MSD transplant recipients with acute myelogenous leukemia. The URDs were serologically matched at HLA-A and -B, and molecular typing was used for HLA-DRB1 only. Five-year DFS was similar in the two cohorts.34
In 1997, Szydlo et al4
reported outcomes using IBMTR data in 2,055 MSD, partially matched related donor, or matched or mismatched URD transplant recipients. Matching, however, was defined only by serological criteria at HLA-A, -B, and -DRB1. Similar to our results, they reported a higher TRM and lower DFS in the cohort that received URD transplants.
This study confirms that in good-risk patients with CML in CP1 who lack a MSD, survival and LFS using 8/8 allele-matched URDs, although statistically slightly inferior, approach that of MSD HCT especially in the first year after diagnosis. When neither MSDs or 8/8 matched URDs are available, the judicious use of mismatched URDs requires balancing risks and benefits for individual patients, as graft versus leukemia potency was not afforded by HLA disparity.