Abstract
Umbilical cord blood (UCB) transplantation yields survival comparable to adult stem cell transplantation, but there is significant variability among UCB products, in large part because of differences in processing conditions from collection to cryopreservation at cord banks. While impact of processing conditions on the actual UCB product has been reported, there is a little information regarding impact on patient-level outcomes. We report a retrospective exploratory analysis of processing performed at cord banks prior to freezing of UCB units and the impact on clinical outcomes such as engraftment, cord dominance, transplant-related mortality (TRM) and overall survival (OS) in 133 UCB recipients.
Methods
All adult recipients of unmanipulated double UCB transplantation (dUCBT) for hematologic malignancy from 2003 to 2011 at the 3 Harvard Cancer Center sites (Dana-Farber Cancer Institute, Massachusetts General Hospital, and Beth Israel Deaconess Medical Center) were included. All UCB units were thawed and washed prior to infusion. Multivariate analyses controlled for prognostic factors including age, malignancy, conditioning intensity, degree of HLA matching, presence of anti-HLA antibodies, order of cord infusion, TNC/kg, and CD34+/kg infused. Time to engraftment and treatment-related mortality were analyzed in the competing risks regression setting and survival was analyzed using proportional hazards models.
98 recipients underwent reduced-intensity conditioning, primarily fludarabine, melphalan and anti-thymocyte globulin. 35 underwent myeloablative conditioning, primarily cyclophosphamide and total body irradiation. Of the 48 banks contributing cords to this study, 42% employed simple cryopreservation and 23% employed plasma/volume reduction only, at some point during operation. These 2 methods were considered “RBC replete”. Of the 34 banks sharing their current practices, simple cryopreservation is no longer practiced by any; 12% practice plasma/volume reduction alone. 88% of the banks now employ RBC depletion, of which 67% use hydroxyethyl starch for RBC sedimentation and 67% use automated processing systems. Engraftment – Neutrophil and platelet engraftment were not impacted by RBC depletion, sedimentation with hydroxyethyl starch, automated processing, HLA matching or CD34+/kg dose in multivariate analyses. An anti-HLA antibody against one or more cords (p<0.001), myeloablative conditioning (p=0.001) and lower TNC/kg doses (p=0.027) were associated with longer time to engraftment. Cord Dominance –Cord dominance was not significantly affected by RBC depletion, RBC sedimentation, use of an automated system, cord bank inventory or years in existence, time in storage, or viability less than 90% at infusion in a multivariate model. Transplant-related Mortality – None of the processing conditions had a significant effect on TRM at 100 days or 1 year. Overall Survival –Although recipient numbers were small, an interesting observation was the improved survival of the 17 recipients of 2 RBC replete units versus the 115 who received 1 or more RBC-depleted units. This was significant in univariate modeling (p=0.022, see Figure) and just significant in multivariate modeling (HR 3.20, p=0.049, 100day OS of 72% vs 100% and 1year OS of 47% vs 82%). Age over 50 and myeloablative conditioning were the only other cord or patient factors to correlate with increased TRM (p=0.033 and 0.001) and decreased overall survival (p=0.062 and 0.003).
1) UCB bank practices have changed with time and currently employ primarily RBC depletion by sedimentation via automated systems prior to cryopreservation. 2) Neutrophil and platelet engraftment and TRM were not significantly affected by processing conditions at the cord bank in our models. 3) Receipt of 1 or more RBC-depleted units seemed to be associated with inferior survival at 100 days and 1 year vs receipt of 2 RBC-replete cords in multivariate modeling. A clear limitation to our interpretation is sample size, with only 17 recipients receiving 2 RBC-replete cords. While our observations require validation in a larger population, processing conditions at the collection banks and their potential impact on clinical outcomes merit further investigation to determine which aspects of processing should be considered in UCB unit selection.
Koreth:Takeda Pharmaceuticals: Consultancy; Millennium pharmaceuticals: Research Funding.
This icon denotes a clinically relevant abstract
