|Home | About | Journals | Submit | Contact Us | Français|
Umbilical cord blood (UBC) stem cells are a useful stem cell source for patients without matched related or unrelated donors. Adult transplantation with single UBC units is associated with high transplantation-related mortality (TRM). In most cases, mortality is due to infection related to slow engraftment and immunoincompetence. In this study, we used a reduced-intensity conditioning regimen of fludarabine, melphalan, and antithymocyte globulin followed by 2 partially matched UBC units. The UBC units were a 4/6 HLA match or better with each other and with the patient and achieved a minimum precryopreservation cell dose of 3.7 × 107 nucleated cells/kg. A total of 21 patients (median age, 49 years) were treated. The median time to an absolute neutrophil count > 0.5 × 109/L was 20 days, and the median time to an unsupported platelet count > 20 × 109/L was 41 days. Two patients experienced primary graft failure and underwent a second UBC transplantation. One patient had a late graft failure. Acute graft-versus-host disease (GVHD) grade II–IV occurred in 40% of patients. The 100-day TRM was 14%, and the 1-year disease-free survival was 67%. Mixed chimerism was associated with a higher risk of chronic GVHD. Our findings indicate that adult patients can tolerate double UBC transplantation well and achieve sustained antitumor responses using this reduced-intensity conditioning regimen.
Umbilical cord blood (UCB) has been shown to contain sufficient progenitor cells to provide durable engraftment in children, and it provides an alternative stem cell source for patients without matched related or unrelated donors [1,2]. Trials of UCB transplantation in children have shown a 30%–80% disease-free survival (DFS), depending on patient age and disease status [3,4]. However, in adults, the results of single UCB transplantation have been less promising because of high transplantation-related mortality due to a low infused cell dose and the subsequent risk of infection associated with prolonged cytopenia [5–7].
The results of both pediatric and adult UBC transplantation studies suggest improved survival with UBC grafts of higher cell dose [3,4,8]. Barker and colleagues reported that outcomes may be improved with 2 UBC units in both the nonmyeloablative and ablative setting [9,10]. Preliminary data from the same group indicates a reduction in relapse with transplantation of 2 UBC units . In an effort to increase the cell dose and decrease transplantation-related toxicity, we treated 21 patients with a reduced-intensity conditioning regimen followed by sequential infusion of 2 partially matched UBC units.
Patients were eligible for UCB transplant if they had no 5/6 or 6/6 HLA-A, -B, or -DR allele-matched related donor or no 6/6 HLA-A, -B, or -DR allele-matched unrelated donor available. Patients were age 18–65 years and met standard transplantation eligibility criteria, including ejection fraction > 45%, diffusing capacity of the lung for carbon monoxide > 50% predicted, and European Cooperative Oncology Group performance status of 0–2. Patients with acute leukemia were eligible if they were in second or subsequent complete remission or in first remission with high-risk cytogenetics or an antecedent hematologic disorder. Patients with refractory leukemia or leukemia in relapse were not eligible. Patients with chronic lymphocytic leukemia progressive after at least 2 chemotherapy regimens, myelodysplasia with life-threatening cytopenia, aplastic anemia, and non-Hodgkin’s lymphoma or Hodgkin’s lymphoma refractory to chemotherapy or relapse were eligible. We limited our study to a reduced-intensity conditioning regimen, because it was anticipated that the lower risk of mucositis would be preferable in patients with a high probability for delayed engraftment. The protocol was reviewed and approved by the Institutional Review Board of the Dana-Farber/Harvard Cancer Center. Written informed consent was obtained from all patients before participation.
Confirmatory HLA typing was performed on all UCB units before transplantation. Molecular typing was performed for both class I (HLA-A, -B, -C) and class II (HLA-DR and -DQ) alleles. Information was recorded on all of these alleles, but only HLA-A, -B, and -DR were used in the search strategy. HLA typing was performed by polymerase chain reaction and sequence-specific primer technology (One Lambda, Canoga Park, CA).
UCB units were obtained from a various national and international registries. UBC units had to meet a minimum combined precryopreservation cell dose of 3.7 × 107 nucleated cells/kg recipient. A minimum cell dose per UCB unit was not required. UCB units were required to be a 4/6 or better allele-level HLA-A, -B, and -DR match with each other and with the patient. Typing at HLA-C and -DQ was obtained but not used in the match strategy. The choice of UCB when multiple units were available was based on higher cell dose and greater HLA compatibility, in that order.
Patients received conditioning with fludarabine 30 mg/m2/day on days −8 through −3 (total dose, 180 mg/m2), melphalan 100 mg/m2/day on day −2, and rabbit antithymocyte globulin (Sangstat, Fremont, CA) 1.5 mg/kg/day on days −7, −5, −3, and −1 (total dose, 6.0 mg/kg). GVHD prophylaxis consisted of continuous intravenous infusion cyclosporine starting on day −3 and mycophenolate mofetil (MMF) 15 mg/kg intravenously twice daily starting on day 0. Cyclosporine and MMF were given orally when the patient was able to tolerate oral medications. In the absence of GVHD, the MMF was tapered between days +30 and +60 posttransplantation, and the cyclosporine was tapered between days +100 and +180. Stem cells were infused on day 0. UCB units were thawed according to the methods of Rubinstein et al , and administered sequentially on the same day 1–6 hours apart. After transplantation, patients received supportive care with transfusion support; prophylactic antibacterial, antifungal, antiviral, and antipneumocystis therapy; and filgrastim 5 μg/kg/day from day +5 until the absolute neutrophil count was > 2.0 × 109 cells/L for 2 consecutive days. Neutrophil engraftment was defined as the first of 3 consecutive days with neutrophil recovery to at least 0.5 × 109 cells/L. Platelet engraftment was defined as the first day of an unsupported platelet count > 20 × 109 cells/L (no platelet transfusion within 7 days). Early graft failure was defined as absence of neutrophil recovery by day +35 after transplantation. Late graft failure was defined as the presence of neutrophil engraftment by day +35, followed by a decline of the neutrophil count to < 0.5 × 109 cells/L after day +42, and the loss of donor chimerism. GVHD was graded according to the consensus criteria .
Donor chimerism was determined from the peripheral blood at posttransplantation weeks 2, 4, 6, 8, 10, and 12 and months 6, 12, and 24. Bone marrow samples were also obtained at posttransplantation months 3, 6, 12, and 24. Chimerism assays were performed by short tandem repeat analysis using a multiplex kit with primers for 10 different loci (Profiler Plus, Applied Biosystems, Foster City, CA) and an Applied Biosystems ABI-310 genetic analyzer. Pre-transplantation samples from the recipient and both UBC donors were tested to identify alleles at each locus that would be specific for each of them. When possible, recipient blood was tested with unseparated cells as well as CD3+- and CD33+-enriched populations, separated using a negative selection assay (RosetteSep; StemCell Technologies, Vancouver, Canada) .
Time for the engraftment and survival outcomes was measured from day 0 of stem cell infusion and estimated by the Kaplan-Meier method. The engraftment time of a hematopoietic parameter was censored at the time of relapse, or time of death if the event occurred before engraftment. Patients who had not died were censored at their last follow-up date in the analysis of overall survival (OS). Patients who were still alive without relapse were censored at the same time with respect to DFS. Wilcoxon’s rank-sum test was used to determine the association between CD4 count at day +100 and likelihood of infection. Mc-Nemar’s test and Wilcoxon’s signed-rank test were used to identify UBC factors associated with the predominating unit at day +100 among the evaluable patients, excluding primary graft failures and early deaths during the first month posttransplantation. The rates of acute grade II–IV GVHD and chronic GVHD were analyzed actuarially, treating graft failure and non-GVHD death as independent competing risks. The log-rank test was used to compare OS and GVHD incidence between groups. All P values were based on 2-sided hypothesis tests, with the exact inference computed using StatXact Version 6 (Cytel Software Corp, Cambridge, MA).
During the study period, UBC searches were undertaken for 60 patients, who lacked a suitably matched related or unrelated donor. In 48 patients (80%), the searches resulted in identification of an appropriate matched pair. Twenty-seven patients did not proceed to reduced-intensity double UBC transplantation because another donor source was identified, the patient elected not to undergo transplantation, the disease relapsed, or insurance approval could not be obtained. UBC searches were unsuccessful for 12 patients; 9 of these patients were non-Caucasian. Five patients would have had 4/6 UBC matches if antigen level, rather than allele-level matching, were used for class I. The median number of confirmatory testings performed to locate a suitable UBC pair was 6 (range, 2–37).
Patient characteristics are listed in Table 1. A total of 21 patients (median age, 49 years) underwent transplantation in this study between October 2003 and May 2005. The median recipient weight was 78 kg (range, 58–111 kg). Four patients were non-Caucasian. Acute myelogeneous leukemia was the most common diagnosis. Five patients (24%) had undergone previous autologous stem cell transplantation.
Characteristics of the UBC graft are displayed in Table 2. Allele-level typing for HLA-A, -B, and -DR were used in the matching strategy. Fourteen patients (67%) received 2 4/6 matched UBC units, 5 patients received 1 4/6 and 1 5/6 HLA-matched unit, and 2 patients received 2 5/6 matched units. Most of the UBC pairs (76%) were 4/6 matches. Allele-level typing for HLA-C and -DQ were performed for 15 patients and corresponding UBC pairs and are displayed in Table 2. Matching ranged from 8/10 alleles to 5/10 alleles for patients and UBC units, and 10/10 to 5/10 alleles between UBC units; for example, 4/6 HLA-A, -B, and -DR allele matches corresponded to 5/10, 6/10, or 7/10 HLA-A, -B, -C, -DR, and -DQ allele matches. The median infused cell doses were 4.0 × 107 nucleated cells/kg (range, 2.9–5.1) and 1.9 × 105 CD34+ cells/kg (range, 0.6–9.7) for the 2 UBC units combined. The median infused cell doses for each individual UBC unit were 2.0 × 107 nucleated cells/kg (range, 1.1–3.4) and 0.9 × 105 CD34+ cells/kg (range, 0.1–8.6).
Two patients had primary graft failure and received a second double UBC transplant with a different conditioning regimen of low-dose total body radiation, cyclophosphamide, and fludarabine on days +39 and +50 after the first transplantation. Both of these patients had aplastic anemia with late transformation to myelodysplastic syndrome after failing to respond to antithymocyte serum and calcineurin inhibitor therapy. Neither had received previous cytotoxic therapy. Before the second transplantation, the patients remained neutropenic and thrombocytopenic, with no evidence of autologous reconstitution. One patient died of zygomycetes infection 8 months after the first transplantation, and the other patient was surviving at 19 months after the first transplantation. Among the remaining 19 patients, the median time to an absolute neutrophil count > 0.5 × 109 cells/L was 20 days (range, 15–34 days). Two patients did not recover to a platelet count > 100 × 109 cells/L by day +100. Among the remaining patients, the median time to a platelet count > 20 × 109 cells/L unsupported was 41 days (range, 21–55 days), and the median time to a platelet count > 100 × 109 cells/L was 63 days (range, 34–91 days). One patient experienced late graft failure 3 months posttransplantation, after a rapid early taper of cyclosporine in the context of severe adenovirus infection. He underwent infusion of previously stored autologous stem cells at 4 months posttransplantation and relapsed 18 months posttransplantation. Acute GVHD grade I–IV was seen in 58% of patients (40% grade II–IV), but only 1 patient had grade III GVHD and no patients had grade IV GVHD. There were no deaths attributable to acute GVHD. Sixteen patients were evaluable for chronic GVHD; 5 of these 16 patients developed chronic GVHD, 2 with extensive disease. There was 1 death from chronic GVHD complicated by vancomycin-resistant enterococcus (VRE) sepsis.
Transplantation-related (nonrelapse) mortality was 14% at 100 days and 19% at 6 months. One patient with CLL had progressive disease and 2 patients (1 with AML as discussed above and 1 with NHL) relapsed. The median follow-up of survivors was 18 months (range, 11–30 months). OS was 71% and DFS was 67%, as illustrated in Figure 1. The projected 2-year OS and DFS are 71% and 55%, respectively. Causes of death occurring before day +100 included central nervous system hemorrhage, staphylococcal sepsis with cerebral edema, and posttransplantation lymphoproliferative disease. There were 3 deaths after day +100, 1 from chronic GVHD with VRE sepsis, 1 from disseminated fungal infection after a second double UBC transplantation in a patient with primary graft failure, and 1 from posttransplantation lymphoproliferative disorder. Other nonfatal infections included cytomegalovirus reactivation, disseminated adenovirus, mucor, and human herpes virus 6 encephalitis.
Peripheral blood was analyzed by flow cytometry for T-cell, B-cell, and natural killer cell markers at 100 days and 1 year posttransplantation; the results are displayed in Table 3. Early graft failures were excluded from these analyses, and the patient with late graft failure was excluded from the 1-year analysis. The median total CD4 count at 100 days posttransplantation was 90 cells/μL (range, 18–425 cells/μL). By 1-year posttransplantation, the median CD4 count had risen to 518 cells/μL (range, 21–1785 cells/μL), consistent with immune system recovery. Severe and life-threatening infections were more common in the patients with a lower CD4 count at day +100 but did not reach statistical significance.
Excluding the 2 early graft failures and 2 early deaths within 1 month of transplantation, chimerism was evaluable in 17 patients. By day +28, a single UBC unit predominated in 8 patients. In 6 patients, a combination of both UBC units was detected, ranging from 50%5– 0% to 80%–20%. Three patients had persistent host hematopoiesis; 2 of these patients had detectable donor cells from only 1 UBC unit. Three patterns of chimerism emerged by 12 weeks posttransplantation (Fig 2). Group 1 comprised 10 patients with evidence of only a single UBC engraftment by 6 weeks posttransplantation; the same pattern was seen at week 12. Group 2 comprised 4 patients in whom both UBC donors were initially identified, but in 3 of 4 patients, only a single donor predominated by week 12. In the other patient, both donors were present in a 69/31 ratio at week 12. Group 3 (3 patients) consisted of patients with a mixed chimera of recipient blood and a single UBC unit.
In 76% of patients (13 of 17), the predominant UBC unit at 3 months posttransplantation was the first UBC unit infused (P = .049). There was a trend toward the predominant UBC unit having a higher nucleated cell dose (P = .071) and CD34 count (P = .120). The median cell doses of the predominant UBC unit were higher than those of the minor UBC unit by 0.4 × 107 nucleated cells/kg and 0.3 × 105 CD34 cells/kg. A closer HLA match, ABO match, or sex agreement did not correlate with the predominant cord. Some ninety percent of the UBC unit pairs were infused 3.5–4.5 hours apart. There was no correlation between time between unit infusions and graft failure.
Patient age, HLA match, cell doses infused, and chimerism pattern at day +100 were analyzed to determine predictive factors for OS and DFS, as well as acute GVHD (grade II–IV) and chronic GVHD (Table 4). The limited patient numbers precluded the fitting of a multivariate model; thus, the univariate numbers must be interpreted cautiously. At 1 year posttransplantation, DFS was 44% and OS was 56% respectively for patients over age 50 (n = 9). Younger patients achieved 83% DFS and OS. There was a trend toward improved survival in patients with a higher CD34 count in the predominant UBC unit. No patients have died among those who received a predominant UBC unit with a CD34 dose/kg above the median of 1.9 × 105. A higher nucleated cell dose or closer HLA match between the patient and the 2 UBC units was not associated with better survival. The HLA match of the predominating UBC unit did not correlate with survival or with GVHD. There was a trend toward less chronic GVHD in the patients infused with lower CD34 doses; however, there were only 5 patients with chronic GVHD, and the median follow-up of the study is 18 months. In addition, the chimerism pattern (chimerism groups 2 and 3) was also strongly associated with the presence of chronic GVHD. The actuarial rate of chronic GVHD rate at 1 year posttransplantation was 100% in patients who had mixed chimera of either 2 UBC units or 1 UBC unit and recipient blood at 6 weeks posttransplantation, but 0% in patients who had evidence of only 1 cord engraftment (P < .001). The strong association between chimerism and GVHD remained highly significant even after adjusting for potential confounding by each of the factors in Table 4. However, these numbers should be interpreted cautiously, given the small number of patients with chronic GVHD.
In this study, we report a low TRM and encouraging DFS using a reduced-intensity regimen followed by sequential UBC transplantation. These findings are consistent with the favorable experiences of nonmyeloablative and reduced-intensity regimens after related and unrelated adult donor transplantation [15–17]. Reduced-intensity regimens have been used with single UBC transplantation; however, TRM was high even with a reduced-intensity regimen [18,19].
The sequential or double UBC transplantation approach was built on the observation that cell dose is an important prognostic factor for engraftment and survival [6,20]. An early report of patients infused with multiple mismatched units suggested that crossed immunologic rejection would not occur . Moreover, recent studies by Barker et al  at the University of Minnesota demonstrated that reliable engraftment can occur using 2 UCB products.
In our study, the 100-day TRM was remarkably low, at 14%. The low TRM might be attributable to patient selection (patients with active leukemia were excluded), high infused cell dose (all patients received 2 UBC units, with a median nucleated cell dose of 4.0 × 107 cells/kg), allele-level matching, and a low incidence of GVHD disease. Younger patients had a better outcome than patients over age 50. Because there are no objective data demonstrating better outcomes for myeloablative therapy in UBC transplantation, we chose to focus our attention on reduced-intensity transplantation, because it was anticipated that the lower incidence of mucositis would be preferable in patients with a high probability of having prolonged neutrophil nadirs.
The relapse rate was low, suggesting preservation of the graft-versus-leukemia after UBC transplantation. The 2 patients with graft failure had aplastic anemia/MDS and had not received previous cytotoxic chemotherapy. Each had primary graft failure but engrafted satisfactorily after the subsequent administration of a more intensive conditioning regimen, followed by a second double UBC transplant. We have previously shown that lack of previous intensive therapy is a risk factor for graft failure in nonmyeloablative unrelated transplants or related donor transplants with very-low-dose conditioning [22,23]. Two patients died of posttransplantation lymphoproliferative disorder (PTLD), on days 41 and 216 posttransplantation. In both of these patients, recurrent disease was suspected, and posttransplantation lymphoma was diagnosed late in the course. The greater immunosuppressive activity of thymoglobulin may also be a contributing factor to the PTLD seen in our study.
Future studies include monthly Epstein-Barr virus (EBV) monitoring. The risk of EBV viremia or EBV posttransplantation lymphoproliferative disorder has been reported to be 20% in UBC recipients conditioned with nonmyeloablative regimens containing antithymocyte globulin and 7% in UBC recipients conditioned with an ablative regimen containing antithymocyte globulin . The median time to development of PTLD was 133 days, and 5 of 9 patients with PTLD survived.
In all of our patients, chimerism analyses indicated the predominance of 1 cord by 3 months posttransplantation. Patients who were mixed chimeras at 6 weeks posttransplantation had a higher incidence of developing chronic GVHD. The explanation for this observation is unclear; a hypothesis may be ongoing graft-versus-graft interactions. This finding will need to be confirmed in larger trials.
In our series, with a limited number of patients, the first UBC unit infused was more often the chimerism “winner.” One possible explanation for this is that there may be limited opportunities to fill the hematopoietic stem cell niche after conditioning, and that the first UBC infused may be more likely to fill this niche . Scadden et al have demonstrated that the capacity of the stem cell niche is limited; in a murine system, osteoblastic cells form a regulatory component of the stem cell niche , and the osteoblast product, osteopontin, may provide a constraint against stem cell numbers in this niche . Also in a murine model, Nilsson has shown that nonlineage-committed hematopoetic stem cells redistributed to the endosteal hematopoietic niche region by 4 hours . Our UCB units were infused 1–6 hours apart, with most units infused 4 hours apart. It is possible that in a competitive niche repopulation model, even a short interval (2–4 hours) may be sufficient to confer an advantage to the first infused unit. The Minnesota group did not report this finding; however, in that analysis, UBC units were administrated directly after one another, which may not have allowed sufficient time for the first unit to populate the niche [9,10]. These observations need to be tested in a larger patient population.
Our study is the first double UBC study to carefully evaluate high-resolution HLA typing and matching between the 2 UBC units. Preliminary data using antigen-level typing for class I suggest higher posten-graftment mortality in mismatched UBC transplantation recipients . The use of allele-level typing has resulted in improved outcomes in the unrelated bone marrow transplantation setting compared with antigen-level typing . Therefore, selected UBC products had to be a 4/6 HLA-A, -B, and -DR allele-level match or better with each other and with the patient. There was no obvious relationship between the degree of histocompatibility between the patient and UBC units and the risk of acute or chronic GVHD, but small numbers limited the ability to detect such a relationship. Mismatching at HLA-C has been associated with worse outcomes after adult unrelated-donor reduced-intensity transplantation . HLA-C and -DQ typing were obtained but were not used in the matching strategy. In the UBC setting, the use of allele-level typing may have limited the degree of mismatch and excluded 5 patients from transplantation. In contrast, the NETCORD group did not find a benefit of high-resolution typing for HLA-A, -B, -C, -DR, and -DQ in single UBC transplantation outcomes . Although allele-level matching might not be important for single UBC transplantation, allele-level typing might decrease the risk of GVHD or graft failure in the double UBC setting. The role of allele-level typing in double UBC transplantation will need to be clarified as more of these transplants are performed.
We have demonstrated that the double UBC transplantation strategy is an important advance in adult UBC transplantation. TRM is low, suggesting that this approach is an acceptable alternative approach for patients without sibling donors.
This work was supported in part by National Heart, Lung, and Blood Institute grant PO1 HL070149.