|Home | About | Journals | Submit | Contact Us | Français|
We describe 70 children with myelodysplastic syndrome [refractory cytopenia (n=31) and refractory anemia with excess blasts (n=30) or blasts in transformation (n=9)] who received umbilical cord blood (UCB) transplantation with a single UCB unit and a myeloablative conditioning regimen. Approximately 20% of children had secondary myelodysplastic syndrome. Median age at transplantation was 7 years and the median follow-up, 3 years. The day-60 probability of neutrophil recovery was 76%; recovery was faster after transplantation of matched or 1-locus mismatched UCB, irradiation-containing conditioning regimen, cell dose >6 × 107/kg and monosomy 7. Risks of treatment failure (recurrent disease or death) were lower in patients with monosomy 7 and transplantations after 2001. The 3-year disease-free survival (DFS) was 50% for transplantations after 2001 compared to 27% for the earlier period (p=0.018). Transplantations after 2001 occurred within 6 months after diagnosis and used UCB units with higher cell dose. DFS was highest in patients with monosomy 7 (61%) compared to other karyotypes (30%), p=0.017. These data suggest transplantation of mismatched UCB graft is an acceptable alternative for children without a matched sibling or suitably matched unrelated adult donor.
Myelodysplastic syndromes (MDS) other than juvenile myelomonocytic leukemia (JMML) are a heterogeneous group of clonal disorders accounting for less than 5% of all hematological malignancies of childhood.(1) Time to progression from refractory cytopenia (RC) to a more advanced form of MDS (refractory anemia with excess blasts [RAEB], refractory anemia with excess blasts in transformation [RAEB-t]) or acute myeloid leukemia (AML) varies widely. In general, patients with RAEB, RAEB-t, or RC with monosomy 7 are more likely to progress to leukemic transformation (1, 2) and are considered, together with MDS secondary to radiotherapy and/or chemotherapy, candidates for hematopoietic stem cell transplantation (HSCT), which has the potential to cure MDS.(3-7)
Data on the outcome of children with MDS other than JMML after allogeneic HSCT are scanty; disease-free survival (DFS) rates are reported to be approximately 60% when the donor is a HLA-matched sibling.(7) However, approximately two-third of patients who may benefit from HSCT lack a matched family donor and a mismatched related or unrelated donor HSCT is considered. In the last few years, the outcome of children with acute leukemia given unrelated cord blood transplantation (UCBT) has been reported to be similar to that of patients transplanted with bone marrow from HLA-matched unrelated donor(8, 9) suggesting UCBT may be a suitable option for children with MDS who do not have a HLA-matched sibling. The objective of the current analysis was to describe outcomes after UCBT and identify prognostic factors for children with MDS.
Data on patients who received UCBT in the U.S were identified from the Center for International Blood and Marrow Transplant Research (CIBMTR) and for UCBT facilitated by Netcord (excluding the U.S), from the Eurocord-European Blood and Marrow Transplant Group (Eurocord-EBMT) and the European Working Group on childhood MDS (EWOG-MDS). Patient, transplant and donor characteristics between the registries were compared to ensure patients were not duplicated.
Patients with MDS, aged less than 18 years at transplantation, received a single unrelated cord blood (UCB) unit between 1995 and 2005 and myeloablative conditioning regimens were eligible. Patients who had received prior HSCT, those with Down syndrome, Fanconi anemia, JMML or MDS with progression to acute myeloid leukemia (AML) were excluded.
Patient and disease characteristics of the 70 eligible children (3 patients have been reported previously (10)) are shown in Table 1. Disease status (RC, n=31; RAEB, n=30; RAEB-t, n=9) was assigned according to the EWOG-MDS criteria for childhood MDS either by the transplant center or assigned retrospectively by two investigators (ABMM and ME) applying published criteria to submitted pathology reports and/or data on collection Forms (11). The subtype assigned is the highest disease subtype at any time prior to transplantation. Patients with prior inherited bone marrow failure or severe aplastic anemia were considered as having secondary MDS if they evolved to RAEB/RAEB-t or abnormal karyotype.(11) Sixteen patients had MDS secondary to either chemotherapy and/or radiotherapy for a prior malignancy (n=9) or an acquired or congenital bone marrow failure other than Fanconi anemia (n=7). Forty-nine patients had karyotype abnormalities with monosomy 7 being the most common abnormality (Table 1). As with the assignment of subtype, the worst cytogentic abnormality at any time prior to UCBT was assigned (i.e. if a patient had normal cytogenetics at diagnosis and subsequently monosomy 7, the patient was classified as having monosomy 7).
Transplant characteristics are shown in Table 1. The median time from diagnosis to transplantation was longer at 8 months for patients with RC compared to 5 months for those with RAEB/RAEB-t. Considering all subtypes, the median time from diagnosis to UCBT was 6 months after 2000 compared to 9 months for the earlier period (p=0.04). Donor-recipient histocompatibility was determined by low resolution typing for HLA-A and -B in all pairs and allele-level typing for DRB1 in all pairs expect one were typing was low resolution. Four patients and their donors were HLA-matched, 34 were mismatched at a single locus, 26 at 2-loci and 5 at 3-loci. The pre-freeze median total nucleated cell dose was 5.75×107/kg (range 1-28) and post-thaw, 4.7×107/kg (range 1-24). Transplantations in 2001 and thereafter received UCB units containing higher infused cell doses, median cell dose of 6 × 107/kg compared to 4 × 107/kg prior to 2001 (p<0.01). Thirty-three patients received total body irradiation (TBI) with cyclophosphamide or melphalan alone (n=26) or in combination with etoposide (n=3), thiotepa (n=2) or fludarabine (n=2); 29 of these patients received anti-thymocyte globulin (ATG). Thirty-seven patients received busulfan-containing regimens with ATG; busulfan was given with cyclophosphamide (n=17), melphalan (n=11). fludarabine (n=7) and without another chemotherapeutic agent (n=1). The predominant graft-versus-host disease (GVHD) prophylaxis was cyclosporine-A and steroids.
Neutrophil recovery was defined as the first of 3 consecutive days with an absolute neutrophil count (ANC) ≥0.5 ×109/L. Patients who failed to achieve ANC ≥0.5 ×109/L within day +60 after UCBT or experienced a sustained decline in ANC (<0.5 ×109/L) after initial recovery were considered to have graft failure. Platelet recovery was defined as the first of 7 consecutive days with an unsupported platelet count ≥20 ×109/L. Grade II-IV acute and chronic GVHD were diagnosed and graded by the transplant center according to published criteria.(12, 13) Recurrent disease was determined morphologically; patients with RC evolved to RAEB post-transplantation and those with RAEB/RAEB-t had recurrence of blasts or AML. Death from any cause in the absence of recurrent disease was defined as non-relapse mortality.
The probabilities of neutrophil and platelet recovery and acute and chronic GVHD were calculated using the cumulative incidence function estimator.(14) For neutrophil and platelet recovery and GVHD, death was the competing event. All surviving patients were censored at last follow-up. The probability of DFS and overall survival were calculated using the Kaplan-Meier estimator.(15) 95% confidence intervals (CI) were calculated using log transformation. Multivariate models for hematopoietic recovery was performed using the Fine and Gray method and for non-relapse mortality and treatment failure (relapse or death from any cause; the inverse of DFS) the Cox regression. Multivariate models were not constructed for acute or chronic GVHD and recurrent disease as there were too few events. Models were built using a forward selection procedure and only variables that attained a p-value ≤0.05 in univariate analysis were retained in the final model (14, 16). The following variables were considered: 1) patient characteristics: gender, human cytomegalovirus serology (positive versus negative), age at transplantation (≤7 versus >7 years, the median age of the study population), 2) disease characteristics: monosomy 7 versus other abnormalities versus none, de novo versus secondary MDS, time from diagnosis to UCBT (≤6 versus >6 months), and 3) transplant characteristics: donor-recipient HLA disparity [0-1 versus 2-3], nucleated cell dose infused/kg (<6 × 107/kg versus ≥6 × 107/kg), year of transplantation (≤2001 versus >2001), and conditioning regimen (TBI-containing versus chemotherapy only regimens). All continuous variables were categorized into two groups (above and below the median) except nucleated cell dose which was divided into five groups at approximately the 20th, 40th, 60th and 80th percentiles. The appropriate “cut-off” for cell dose was determined by testing the cell dose groups in the model for neutrophil recovery (Fine and Gray method) and for overall survival (Cox regression method). A cut-off of 6×107/kg (80th percentile) was found to be associated with neutrophil recovery. We did not find a statistically significant cell dose cut-off for overall survival. All p-values are two-sided. Analyses were performed with SPSS (Inc., Chicago, IL), Splus (MathSoft, INC, Seattle) and SAS version 9.0 (Cary, NC).
Fifty-three of 70 patients achieved neutrophil recovery and the median time to recovery, 23 days (range 12-59). The day-28 and day-60 probabilities of neutrophil recovery were 51% (95% CI 46-57) and 76% (95% CI 64-84), respectively. In multivariate analysis, neutrophil recovery was faster with transplantation of UCB units matched or mismatched at 1-locus (hazard ratio [HR] 0.47, 95% CI 0.25-0.90, p=0.017), pre-freeze cell dose ≥6×107/kg (HR 0.55, 95% CI 0.33-0.93, p=0.024), TBI-containing conditioning regimen (HR 0.47, 95% CI 0.25-0.85, p=0.014) and monosomy 7 (HR 0.58, 95% CI 0.33-0.99, p=0.045). Only 40 of 70 patients achieved platelet recovery; the median time to recovery was 53 days (range 16-152) and the day-180 probability of platelet recovery, 57% (95% CI 45-68). In multivariate analysis, year of transplantation (after 2001) was the only factor associated with faster recovery (HR 0.49, 95% CI 0.25-0.94, p=0.033).
Chimerism data were available for 55 of 70 patients (measurements were within the first 100 days after transplantation). Forty-two patients achieved full chimerism; chimerism was mixed in 3 patients, 8 patients had autologous reconstitution and 2 patients, secondary graft failure. Of the three patients with mixed chimerism, one died at 1 month of bacterial infection; one relapsed, received a second transplant and died early after second transplantation and the remaining patient converted to full donor chimerism and is alive and in remission 28 months after transplantation. Five patients with primary graft failure (3 of 5 were transplanted for secondary MDS) received a second transplant but only one patient is alive.
Twenty-two patients developed acute GVHD (grade II n=13; grade III n=6; grade IV n=3). The 100-day cumulative incidence of grade II-IV acute GVHD was 30% (95%CI 20-41). Sixteen patients of 48 patients at risk developed chronic GVHD and the 3-year cumulative incidence of chronic GVHD was 23% (95% CI 14-33). Severity of chronic GVHD was reported as limited in 5 patients and extensive in 11. Only one patient with recurrent disease reported chronic GVHD. The relatively small sample size and few events prevented us from performing risk factor analyses for GVHD.
Thirty patients died from a transplantation-related complication and 13 had recurrent disease. Three of the 13 patients with recurrent disease were transplanted in RC and progressed to RAEB after transplantation. In multivariate analysis, risks of non-relapse mortality were lower when transplantation was performed after 2001 (HR 0.41, 95% CI 0.20-0.84, p=0.015), (Figure 1). Though lower cell dose and HLA-disparity at ≥2-loci were associated with slower neutrophil recovery, a statistically significant effect of cell dose and HLA disparity on non-relapse mortality was not found. We did not identify any factors predictive for recurrent disease. Recurrence appears to be lower in patients with monosomy 7 though this did not reach statistical significance (HR 0.16, 95%CI 0.02-1.20, p=0.074).
Twenty-nine of 70 children are alive and disease-free with a median follow-up of 39 months (range 10-105). The 3-year probabilities of overall survival and DFS were 42% (95% CI 35-53) and 39% (95% CI 33-45), respectively. In multivariate analysis, transplantations prior to 2001 and karyotypes other than monosomy 7 were independent risk factors for higher risks of treatment failure (recurrent disease or death, inverse of DFS), (HR 2.38, 95% CI 1.14-5.00, p=0.018) and HR 2.04, 95% CI 1.11-3.70, p=0.017), respectively (Figure 2). The 3-year DFS was 61% (95% CI 51-71) in patients with monosomy 7 compared to 26% (95% CI 16-36) and 37% (95% CI 26-48) in patients with normal karyotpye and karotypes other than monosomy 7, respectively. To further explore the association of monosony 7 and DFS, in subset analysis, we examined the effect of karyotype in patients with RAEB/RAEB-t. Three-year DFS rates were 75% in patients with monosomy 7 (n=8) compared to 31% in patients with other karyotypes (n=31). The 3-year DFS was 50% (95% CI 42-58) for transplants after 2001 and 27% (95% CI 19-34) for the earlier period. We did not find a statistically significant association between disease status (RC versus RAEB/RAEB-t) and treatment failure. Similarly, the effect of cell dose and HLA disparity on treatment failure was not statistically significant.
Eleven patients died from recurrent disease and 30 patients died from transplant-related complications: infection (n=7), GVHD (n=3), graft failure (n=5), sinusoidal occlusive syndrome (n=1), post-transplantation lymphoproliferative disease (n=1), organ toxicity (n=11) and the cause of death was not reported for 2 patients (died in remission). Most deaths occurred within the first year after transplantation. There were 9 deaths beyond the first year, 4 from recurrent disease and 5 from infections (infection was secondary to chronic GVHD in 4 patients).
HSCT remains the only treatment with curative potential for MDS.(1, 17) UCB as an alternative source of hematopoietic stem cells for unrelated donor transplantation may offer an acceptable option in children without an HLA-matched sibling. While outcomes and risk factors affecting UCBT have been reported in children with acute leukemia,(18-20) and the results comparable to that after matched unrelated bone marrow transplantation,(8, 9, 21) there are few reports detailing risk factors and outcomes after UCBT for MDS. Published reports are limited to very few patients,(4-7, 22-25) and include patients with JMML and/or secondary AML and some reports have included children with adults.(2, 26-28) This analysis, in contrast, to other reports, report outcomes after UCBT and prognostic factors associated with this treatment in children with RC, RAEB and RAEB-t.
Neutrophil and platelet recovery were lower than that reported after UCBT for acute leukemia or unrelated donor bone marrow transplantation for MDS in children.(7, 23, 24) Consistent with reports in children with acute leukemia (9, 18, 19, 29) transplantation of UCB units containing higher cell doses was associated with faster neutrophil recovery. However, the current analysis observed faster neutrophil with transplantation of UCB units containing 6×107/kg total nucleated cells or higher. This “cut-off” is almost twice that reported for children with acute leukemia who received UCBT. Lower rates of hematopoietic recovery and the apparent requirement for UCB grafts with higher nucleated cell doses may be explained by abnormalities of bone marrow stroma associated with MDS. In this regard, it is noteworthy that UCB cells co-cultured with bone marrow stroma from patients with MDS exhibit a lower proliferative capacity than when these cells are co-cultured with normal bone marrow stroma.(30, 31) Given the relatively small number of patients in the current analysis (albeit the largest to-date) the observed optimal cell dose in the current analysis requires validation in a larger cohort. Consistent with other reports, we observed slower hematopoietic recovery with increasing donor-recipient HLA disparity and non-TBI conditioning regimens.(29, 32, 33) We also observed a marginal advantage for neutrophil recovery in patients with monosomy 7. Most patients with monosomy 7 were transplanted in RC; transplantation prior to the more advanced disease phase (RAEB/RAEB-t) may explain the observed higher likelihood of neutrophil recovery.
Transplantation period and monosomy 7 influenced DFS. Higher failure rates were observed prior to 2001. During the early period UCB units with low cell dose and donor-recipient HLA disparities at 3-loci were selected for transplantation. Whereas UCBT performed during the later period (2001-2005) have generally adhered to selection of units with minimum pre-freeze cell dose >3 × 107/kg, donor-recipient HLA disparity limited to 1 or 2-loci and selection of patients with disease status in remission (29). Better unit and patient selection would have enhanced hematopoietic recovery and consequently fewer transplant-related complications and death. Improvements in post-HSCT supportive care, independent of graft source would have also contributed to the success of UCBT in recent years. The waiting period from diagnosis to transplantation was shorter after 2001 and may also have contributed to the observed higher DFS. Shorter waiting times to transplantation included patients with RAEB/RAEB-t and RC with monosomy 7, the groups at highest risk for treatment failure. As with all observational reports, unmeasured and/or unknown factors may have also contributed to the observed improvement after 2001.
DFS rates in patients with monosomy 7 were higher compared to patients with normal or other cytogenetic abnormalities. In a recent EWOG-MDS analysis, complex karyotype defined as 3 or more structural aberrations was an independent prognostic factor, regardless of treatment (C. Niemeyer, personal communication, June 2008). In that analysis, DFS was similar amongst those with normal karyotype, monosomy 7 and karyotypes other than complex. Only 3 patients in the current analysis had complex karyotypes which explain our inability to detect an adverse effect of this karyotype on transplant-outcome. The observed higher DFS with monosomy 7 in the current analysis warrants further examination. However, other two small series of HSCT recipients described similar encouraging results (34, 35). We explored whether timing of transplantation and chemotherapy prior to UCBT may have influenced DFS. The timing of transplantation in patients with monosomy 7 was not significantly different from those without this abnormality (5 versus 6 months, respectively). We also tested for an effect of treatment prior to transplantation on DFS and found none (univariate analysis; data not shown). Surprisingly, disease status was not associated with DFS. The relatively small sample size may have prevented us from observing a significant effect. Most patients with monosomy 7 had RC and this karyotype might be a surrogate for disease status. . We adhered to the EWOG-MDS classification instead of the WHO or IPSS classification given the former is more appropriate for children.(3, 11, 28) A separate analysis using the WHO classification (the standard for adults with MDS) was not undertaken. Given the study period, the WHO classification would have had to be assigned retrospectively and the relevant data for retrospective assignment were not available for all patients.
Our report has several limitations. Given the rarity of childhood MDS, we used data reported to two transplant registries to obtain a relatively large sample for a rare disease. As all patients received a myeloablative conditioning regimen we are unable to comment on the use of reduced intensity conditioning regimens especially for children with RC.(36) UCB grafts are readily available and should be considered when transplantation is required urgently or in the absence of a HLA-matched related or unrelated donor. Recent advances, such as use of double cord blood units may circumvent the cell dose limitation, lead to faster hematopoietic recovery (37) and lower transplant-related mortality.
Funding: Public Health Service Grant U24-CA 756518-11 (M.E., M-J.Z.) from the National Cancer Institute, the National Institute of Allergy and Infectious Diseases and the National Heart Lung and Blood Institute and European grant QLK3-CT-1999-00380 (E.G., V.R).
These data were presented in part at the American Society of Hematology annual meeting, San Francisco, 2008