|
|
Background Data regarding the outcome of cord-blood transplantation in adults are scant, despite the fact that these grafts are increasingly used in adults.
Methods We compared the outcomes of the transplantation of hematopoietic stem cells from unrelated donors in adults with leukemia who had received cord blood that was mismatched for one HLA antigen (34 patients) or two antigens (116 patients), bone marrow that had one HLA mismatch (83 patients), and HLA-matched bone marrow (367 patients). We used Cox proportional-hazards models to analyze the data.
Results Cord-blood recipients were younger and more likely to have advanced leukemia than were bone marrow recipients, and they received lower doses of nucleated cells. Hematopoietic recovery was slower with transplantation of mismatched bone marrow and cord blood than with matched marrow transplantations. Acute graft-versus-host disease (GVHD) was more likely to occur after mismatched marrow transplantation, and chronic GVHD was more likely to occur after cord-blood transplantation. The rates of treatment-related mortality, treatment failure, and overall mortality were lowest among patients who received matched marrow transplants. Patients who received mismatched bone marrow transplants and those who received mismatched cord-blood transplants had similar rates of treatment-related mortality (P=0.96), treatment failure (P=0.69), and overall mortality (P=0.62). There were no differences in the rate of recurrence of leukemia among the groups. There were no differences in outcome after cord-blood transplantation between patients with one HLA mismatch and those with two HLA mismatches.
Conclusions HLA-mismatched cord blood should be considered an acceptable source of hematopoietic stem-cell grafts for adults in the absence of an HLA-matched adult donor.
Cord-blood grafts from unrelated donors have been used successfully, primarily in children.6,7,8,9,10,11 In children, these grafts reconstitute hematopoiesis more slowly than do bone marrow grafts and thereby contribute to relatively high rates of post-transplantation infection and early death.11,12,13,14 The incidence and severity of GVHD are not excessive, however, even with cord-blood grafts mismatched for more than one HLA antigen, and graft-versus-leukemia effects are well maintained.11,12,13,15 Reports of cord-blood transplantation in adults also suggest slower hematopoietic recovery and variable rates of leukemia-free and overall survival, but these studies have been small and did not compare cord-blood with bone marrow transplantation.14,15,16,17,18
We analyzed data on a large number of adult patients that were reported to the International Bone Marrow Transplant Registry (IBMTR) and the National Cord Blood Program (NCBP) of the New York Blood Center. We compared the outcomes after transplantation of hematopoietic stem cells from cord blood from unrelated donors that was mismatched for one or two HLA antigens (150 patients), from bone marrow that was mismatched for one HLA antigen (83 patients), or from HLA-matched bone marrow (367 patients).
Methods
Collection of Data
Data on patients who underwent bone marrow and cord-blood transplantation were obtained from the IBMTR and the NCBP of the New York Blood Center. The IBMTR is a working group of more than 400 transplantation centers worldwide. Participating centers register basic information on consecutive transplantations at a statistical center located at the Medical College of Wisconsin. Detailed demographic and clinical data are collected on a representative sample of patients in the registry. Centers that obtain cord-blood grafts from the NCBP are required, under the Investigational New Drug rules of the Food and Drug Administration, to report data on the outcome of transplantation procedures.
Inclusion Criteria
The study included patients 16 to 60 years of age who had received either an HLA-matched marrow transplant or a marrow transplant with a single HLA mismatch from an unrelated donor or had received a cord-blood transplant with one or two HLA mismatches. HLA matching was performed with the use of serologic or low-resolution molecular typing methods for HLA-A and HLA-B and high-resolution molecular typing for HLA-DRB1. The matching process identified all specificities that were recognized by the World Health Organization at the time of transplantation. There were not enough transplantations performed with HLA-matched cord blood (5 patients) or cord blood with three mismatches (19 patients) to include these in the comparison. Also excluded were recipients of T-cell–depleted marrow, peripheral blood, reduced-intensity preparative regimens, ex vivo expanded grafts, or multiple cord-blood units and those in whom prior transplantation had failed. Transplantations were performed between January 1, 1996, and December 31, 2001, in various locations throughout the United States. Eligible for the study were 450 bone marrow recipients and 150 cord-blood recipients. The patients had been given a diagnosis of acute lymphoblastic leukemia, acute myeloid leukemia, chronic myeloid leukemia, or myelodysplastic syndrome.
End Points
Neutrophil recovery was defined by an absolute neutrophil count of at least 500 cells per cubic millimeter for three consecutive days; platelet recovery was defined by a count of at least 20,000 platelets per cubic millimeter, unsupported by transfusions, for seven days. The incidence of grade 2, 3, or 4 acute GVHD was determined in all patients,19 and the incidence of chronic GVHD was determined in patients who survived for at least 90 days.20 Treatment-related death was defined as death during a continuous remission. Relapse was defined as a recurrence of leukemia; patients in whom a remission failed to occur after transplantation were considered to have had a recurrence at day 1. Leukemia-free survival was defined as survival in a state of continuous complete remission.
Statistical Analysis
Variables related to patients, disease, and transplants were compared among the groups with the use of the chi-square statistic for categorical variables and the Kruskal–Wallis test for continuous variables. The probabilities of overall and leukemia-free survival were calculated with the use of the Kaplan–Meier estimator.21 For analyses of survival rates, death from any cause was considered an event, and data on patients who were alive at the last follow-up contact were censored. For leukemia-free survival, relapse or death (i.e., treatment failure) was considered an event, and data on patients who were alive and in continuous complete remission were censored at the last follow-up. The probabilities of neutrophil and platelet recovery, acute and chronic GVHD, treatment-related death, and relapse were calculated with the use of the cumulative-incidence–function method.21 For neutrophil and platelet recovery and for GVHD, death without an event (hematopoietic recovery or GVHD) was the competing event; for treatment-related mortality, relapse was the competing event; and, for relapse, treatment-related death was the competing event. Data on patients who were alive without an event were censored at the last follow-up.
Confidence intervals were calculated with the use of a log transformation.21 Adjusted probabilities of overall and leukemia-free survival were estimated with the use of the Cox proportional-hazards regression model, with consideration of the variables in the final multivariate models.22 Multivariate models were built with the use of stepwise forward selection, with a P value of 0.05 or less considered to indicate statistical significance; all variables met the proportional-hazards assumption. The primary objective was to compare outcomes according to graft type: HLA-matched bone marrow, single-antigen–mismatched bone marrow, and cord blood with one or two antigen mismatches. Results were expressed as hazard ratios — the relative rate of occurrence of the event with one graft type as compared with another.
Before comparing bone marrow and cord-blood grafts, we analyzed the effect of HLA matching in cord-blood transplantation. There were no statistically significant differences in outcome between patients who had received transplants with one antigen mismatch and those who received transplants with two antigen mismatches, and these groups were combined for further analyses. The variable for graft type was retained in all steps of model building. Other variables that were considered were the age of the transplant recipient, the serologic status with respect to cytomegalovirus in both donor and recipient before transplantation, the type of leukemia and the status of the disease at transplantation, the sex of the recipient and the donor, the conditioning regimen (i.e., with irradiation vs. without it), the regimen for GVHD prophylaxis (e.g., treatment with cyclosporine and methotrexate vs. treatment with alternative regimens), and the total dose of nucleated cells (for each graft type separately). There were no first-order interactions between the graft type and the other variables studied. No statistically significant effects according to treatment center were noted.23 P values are two-sided. Analyses were completed with the use of PROC PHREG in SAS software, version 8.2 (SAS Institute).
Results
Patients
Table 1 shows features of the patients, the types of leukemia they had, and the types of transplant they received. As compared with bone marrow recipients, cord-blood recipients were younger, were less likely to be white, were more likely to have acute leukemia, were more likely to have advanced leukemia at transplantation, and weighed less. The average total dose of nucleated cells for marrow recipients was generally 10 times as high as that for cord-blood recipients. All cord-blood grafts were HLA-mismatched — 23 percent for one antigen, and 77 percent for two antigens. The median period of follow-up for survivors after marrow and cord-blood transplantation was 48 months and 40 months, respectively. Completeness of follow-up for the study population was 91 percent.24
|
Among patients who had neutrophil and platelet recovery, the recovery times were longer after cord-blood transplantation than after bone marrow transplantation. Median times to neutrophil recovery were 18 days (95 percent confidence interval, 18 to 19) after HLA-matched bone marrow transplantation, 20 days (95 percent confidence interval, 18 to 22) after mismatched marrow transplantation, and 27 days (95 percent confidence interval, 25 to 29) after cord-blood transplantation (P<0.001). Corresponding times to platelet recovery were 29 days (95 percent confidence interval, 27 to 30), 29 days (95 percent confidence interval, 27 to 34), and 60 days (95 percent confidence interval, 54 to 71), respectively (P<0.001). Despite early differences, the cumulative incidence of neutrophil recovery at day 100 (P=0.29) and that of platelet recovery at one year (P=0.16) were similar after transplantations of mismatched bone marrow and of cord blood. Corresponding cumulative incidence rates after transplantation of HLA-matched bone marrow were significantly higher (P<0.01) (Figure 1).
|
Acute GVHD of grade 2, 3, or 4 developed in 176 of 367 recipients of HLA-matched bone marrow, in 43 of 83 recipients of mismatched bone marrow, and in 61 of 150 recipients of mismatched cord blood. The rate of acute GVHD grade 2, 3, or 4 was similar between patients who received mismatched cord blood and those who received HLA-matched bone marrow (hazard ratio, 0.81; 95 percent confidence interval, 0.59 to 1.10; P=0.17). However, acute GVHD was less likely after transplantation of mismatched cord blood than after that of mismatched bone marrow (hazard ratio, 0.66; 95 percent confidence interval, 0.44 to 0.99; P=0.04).
Among patients who survived for 90 days or longer, chronic GVHD developed in 86 of 243 recipients of HLA-matched bone marrow, in 17 of 43 recipients of mismatched bone marrow, and in 35 of 69 recipients of mismatched cord blood. The rate of chronic GVHD was higher among patients who received cord blood than among those who received HLA-matched marrow (hazard ratio, 1.62; 95 percent confidence interval, 1.08 to 2.42; P=0.02) but similar to the rate among recipients of mismatched marrow (hazard ratio, 1.12; 95 percent confidence interval, 0.63 to 2.02; P=0.69). Among patients with chronic GVHD, the proportion with extensive disease was lower among cord-blood recipients than among HLA-matched or mismatched marrow recipients, with rates of 33 percent, 52 percent, and 71 percent, respectively (P=0.03).
Treatment-Related Mortality
Death from treatment-related complications occurred in 169 of 367 recipients of HLA-matched bone marrow, in 54 of 83 recipients of mismatched marrow, and in 95 of 150 recipients of cord blood. Rates of treatment-related death were similar among patients who received mismatched bone marrow and those who received mismatched cord blood; the rate was significantly lower among patients who received HLA-matched bone marrow (Table 2).
|
Leukemia recurred after transplantation in 83 of 367 recipients of HLA-matched marrow, in 12 of 83 recipients of mismatched marrow, and in 26 of 150 recipients of mismatched cord blood. All of these rates were similar (Table 2).
Leukemia-free Survival
Treatment failure (i.e., relapse or death) occurred in 252 of 367 recipients of HLA-matched marrow, in 66 of 83 recipients of mismatched marrow, and in 121 of 150 recipients of mismatched cord blood. Rates of treatment failure were similar among patients who had received mismatched marrow and those who had received cord blood, but the rates were significantly lower among patients who had received HLA-matched marrow (Table 2). The probability of surviving for three years without a recurrence of leukemia, after adjustments were made for other significant factors, was 19 percent (95 percent confidence interval, 12 to 27 percent) for recipients of mismatched marrow and 23 percent (95 percent confidence interval, 17 to 30 percent) for recipients of cord blood (P=0.69); the probability of such survival after transplantation with HLA-matched marrow was higher, at 33 percent (95 percent confidence interval, 28 to 37 percent; P=0.001) (Figure 2A).
|
Death due to any cause after transplantation occurred in 247 of 367 recipients of HLA-matched marrow, in 65 of 83 recipients of mismatched marrow, and in 117 of 150 recipients of mismatched cord blood (Table 3). Mortality rates were similar after transplantation of mismatched marrow and of cord blood but significantly lower after transplantation of HLA-matched marrow (Table 2). Adjusted probabilities of three-year survival were 20 percent (95 percent confidence interval, 12 to 28 percent) for recipients of mismatched marrow, 26 percent (95 percent confidence interval, 19 to 32 percent) for recipients of cord blood (P=0.62), and 35 percent (95 percent confidence interval, 30 to 39 percent) for recipients of HLA-matched marrow (P<0.001) (Figure 2B).
|
Discussion
Our objective was to compare the effectiveness of hematopoietic stem-cell transplantation with the use of cord-blood grafts mismatched for one or two HLA antigens, HLA-matched bone marrow, and bone marrow mismatched for one antigen, all from unrelated donors, in adults with leukemia. Significant prognostic factors were older age (over 40 years), advanced leukemia, and positivity for cytomegalovirus in the donor, recipient, or both. After we adjusted for these factors, treatment-related mortality, the rate of treatment failure, and overall mortality were significantly lower after the transplantation of HLA-matched marrow than after the transplantation of HLA-mismatched marrow or cord blood. The rate of recurrence of leukemia was similar among the three groups. It is important to note that there were no significant differences in treatment-related mortality, the rate of treatment failure, or overall mortality between recipients of HLA-mismatched marrow and those who received mismatched cord blood despite the fact that 77 percent of cord-blood grafts were mismatched for two HLA antigens, whereas all mismatched-marrow grafts were mismatched for only one antigen. The relative efficacy of HLA-matched and mismatched bone marrow and cord-blood grafts did not differ according to age, type of leukemia, or other prognostic variables we studied.
An HLA mismatch is perhaps the strongest risk factor for GVHD and other outcomes after bone marrow transplantation. The association of HLA disparity with GVHD after cord-blood transplantation is less clear.8,9,10,18 We did not observe differences in outcome between transplantations that involved cord blood mismatched for one and two HLA antigens; this finding contrasts with our recent analysis in which an increasing HLA disparity in children was associated with worse outcomes after transplantation of cord blood (unpublished data). However, in that study, overall outcomes were better, the number of patients was larger, doses of nucleated cells were higher, and the range of HLA disparity was wider (from 0 to 3 antigens). All of these factors may have produced an effect that was easier to detect. Most of the cord-blood grafts (77 percent) in the current study were mismatched at two HLA loci; the relatively small number mismatched at one HLA locus gave the study low statistical power to detect an effect of the number of mismatches. Moreover, the definition of an HLA match in this study included only six HLA loci and low-resolution typing for class I antigens. This is how cord-blood units were typed and selected during the study period, but bone marrow donors were often selected with the use of criteria that included high-resolution typing for 8 or 10 HLA antigens. It may be that undetected mismatches also decreased our ability to find an effect of the HLA matching. Further analysis of larger numbers of patients, with the use of more sensitive typing techniques, will be necessary before we can conclude that increasing HLA disparity does not affect the outcome of cord-blood transplantation in adults.
A major limitation to the use of cord-blood grafts in adults is the concern that these grafts have an insufficient number of precursor cells as compared with bone-marrow grafts. Reports of cord-blood transplantation in adults suggest that hematopoietic recovery is faster with higher cell doses.16,17,18 The median dose of nucleated cells in our study was 2.2x107 per kilogram of body weight, which is similar to that reported in other studies, and the rates of neutrophil and platelet recovery were lower with cord blood than with bone marrow. However, we did not find a significant effect of the cell dose on the rates of mortality and treatment failure among cord-blood recipients. The range of doses was relatively narrow (80 percent of patients received a dose of less than 3.0x107 per kilogram), which may have precluded the detection of a cell-dose effect.
We observed, as others have, that the proportion of deaths related to infection was higher soon after cord-blood transplantation than after transplantation of HLA-matched and mismatched bone marrow — a difference that is possibly the result of slower myeloid recovery.14,15 Although infections were more common with transplantation of cord blood, the distribution of types of infection did not differ significantly among recipients of HLA-matched marrow, mismatched marrow, and cord blood. Others have also reported delayed neutrophil and lymphocyte recovery and higher rates of bacterial infections in the early post-transplantation period (i.e., less than 50 days) in the case of cord blood, with overall rates of infection at later times being similar to those for marrow transplantation.15 Several ongoing studies are attempting to improve hematopoietic recovery and thereby decrease infection in the early period after cord-blood transplantation.25,26,27,28
All aspects of the transplantation regimen, including the choice of graft type, were determined by the transplantation centers. A higher proportion of cord-blood recipients than marrow recipients had advanced leukemia at the time of transplantation, which probably reflects a practice of using these grafts in patients with limited treatment options. Any observational study that compares different interventions is subject to bias owing to the complex criteria for selection that underlie the choice of intervention, and our study is no exception. However, our ability to adjust for key risk factors made a controlled (though not randomized) comparison possible. The results support a standard of care in which transplantation of HLA-matched marrow is performed when a donor is available in a timely manner. In the absence of such a donor, cord blood mismatched for one or two antigens or marrow grafts mismatched for one antigen are acceptable alternatives and have similar outcomes. The rapid availability of a unit of cord blood (i.e., after a median of 13.5 days)29 may be a particular advantage for patients who require urgent transplantation. The slow rate of hematopoietic recovery remains a major deterrent to the use and success of cord-blood transplantation in adults, and novel strategies to overcome this obstacle are needed.
Supported by a Public Health Service grant (U24-CA76518) from the National Cancer Institute, the National Institute of Allergy and Infectious Diseases, the National Heart, Lung, and Blood Institute, the Abraham J. and Phyllis Katz Foundation, the Leukemia and Lymphoma Society of America, the American Society of Clinical Oncology, and the Children's Leukemia Research Association. Dr. Laughlin is a Leukemia Scholar in Clinical Research, the Leukemia and Lymphoma Society of America. Dr. Eapen is the recipient of a Clinical Research Career Development award from the American Society of Clinical Oncology.
Dr. Champlin reports having received consulting fees from StemCyte.
Source Information
From the Case Comprehensive Cancer Center and University Hospitals of Cleveland Ireland Cancer Center, Cleveland (M.J.L., H.M.L.); the International Bone Marrow Transplant Registry, Health Policy Institute, Medical College of Wisconsin, Milwaukee (M.E., M.-J.Z., M.M.H); the National Cord Blood Program, New York Blood Center, New York (P.R., C.S., A.S.); the University of Minnesota Medical School, Minneapolis (J.E.W., J.N.B.); the M.D. Anderson Cancer Research Center, Houston (R.E.C.); the Center for Advanced Studies in Leukemia, Los Angeles (R.P.G.); the Adult Blood and Marrow Transplant Unit, United Bristol Health Care Trust, Bristol, United Kingdom (D.I.M.); and Leiden University Medical Center, Leiden, the Netherlands (J.J.R.).
Address reprint requests to Dr. Horowitz at the International Bone Marrow Transplant Registry, Medical College of Wisconsin, 8701 Watertown Plank Rd., P.O. Box 26509, Milwaukee, WI 53226, or at marymh{at}mcw.edu.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Umbilical-Cord Blood for Transplantation in Adults
Fernandez M. N., Regidor C., Cabrera R., Bandini G., Bonifazi F., Stanzani M., Laughlin M. J., Eapen M., Rubinstein P., Rocha V., Labopin M., Gluckman E., the Acute Leukemia Working Party of the European Blood and Marrow Transplant Group and the Eurocord–Netcord Registry
Extract |
Full Text |
PDF
N Engl J Med 2005;
352:935-937, Mar 3, 2005.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
