Abstract
The Blood and Marrow Transplant Clinical Trials Network (BMT CTN 0501) randomized children with hematologic malignancy to one or two cord blood unit (UCB) transplantation between 2006 and 2012. While the trial concluded that survival was similar regardless of number of units infused, survival was better than previously reported. This prompted a comparison of survival of trial versus non-trial patients to determine the generalizability of trial results and whether survival was better because of the trial treatment regimen. During the trial period, 396 recipients of a single UCB unit met trial eligibility but were not enrolled. Trial patients (n=100) received total body irradiation (TBI) 1320 cGy, cyclophosphamide 120 mg/kg, and fludarabine 75 mg/m2 (TCF). Non-trial patients either received the same regimen (N=62; non-trial TCF) or alternative regimens (N=334; non-trial regimens). Five-year survival between trial and non-trial patients conditioned with TCF was similar (70% vs. 62%). However, 5-year survival was significantly lower with non-trial TBI-containing (47% [HR 1.97, p=0.001]) and chemotherapy only regimens (49% [HR 1.87, p=0.007]). The results of BMT CTN 0501 appear generalizable to the population of trial eligible patients. The survival difference between the trial-specified regimen and other regimens indicate the importance of conditioning regimen for UCB transplantation.
Keywords: Cord Blood Transplant, Acute Leukemia, Survival, Conditioning Regimen
Introduction
Most children and adolescents with acute leukemia can be cured by conventional chemotherapy. However, a subset of patients is at particularly high risk of disease recurrence and is frequently offered allogeneic hematopoietic cell transplantation (HCT) as a treatment option. For those without an HLA matched related or unrelated donor, partially HLA matched umbilical cord blood (UCB) is a suitable alternative for HCT. A study by the Blood and Marrow Transplant Clinical Trials Network (BMT CTN 0501; NCT00412360) randomized children with hematologic malignancy to one or two UCB unit transplantation between December 2006 and February 2012.1 While it was hypothesized that transplantation of two UCB units would result in better survival based on higher cell doses, survival was similar in the two treatment arms (p=0.17). However, survival in both arms was higher than that reported in prior large studies. For example, survival was substantially better in single UCB transplant patients enrolled in the BMT CTN 0501 trial relative to those in a similar high risk pediatric malignancy population in an earlier multi-center trial of Cord Blood Transplantation (COBLT, 73% vs. 57% at one year, respectively, p=0.01).2,3 Notable differences between the trials included transplant conditioning, immunoprophylaxis for graft-versus-host disease (GVHD) and transplant period. The BMT CTN 0501 conditioning regimen consisted of total body irradiation (TBI) 1320 cGy, cyclophosphamide 120 mg/kg, and fludarabine 75 mg/m2 (TCF) and cyclosporine and mycophenolate for GVHD prophylaxis 1; the transplant-conditioning regimen for the phase II trial2 consisted of 1350 cGy of total body irradiation (TBI), cyclophosphamide 120 mg/kg and antithymocyte globulin (ATG, equine) 90 mg and cyclosporine and methylprednisolone for GVHD prophylaxis. Lastly, patients in BMT CTN 0501 were enrolled between 2006 and 2012 as compared to 1999 and 2003 in the prior study.
Examination of non-trial single UCB unit transplants reported to the Center for International Blood and Marrow Transplant Research (CIBMTR) during the trial period (December 2006 to February 2012), revealed that 80% met the broad eligibility criteria for BMT CTN 0501 (i.e. aged 1–21 years, high-risk acute leukemia, performance score ≥70). The potentially improved survival in recipients of BMT CTN 0501 compared to prior studies in children transplanted with a single UCB unit prompted a comparison of trial versus non-trial treatment outcomes to determine the generalizability of treatment offered on trial and whether trial participation per se or some other aspect like the trial specified transplant conditioning regimen, led to the apparent improved survival reported for patients in BMT CTN 0501.
Materials and Methods
Data Source
The CIBMTR is a working group of transplant centers that contributes data on consecutive allogeneic and autologous transplants. Participating centers report consecutive transplants and compliance is monitored by on-site audits. The CIBMTR, together with the EMMES Corporation and the National Marrow Donor Program (NMDP) serve as the Data Coordinating Center for the BMT CTN. Consent is sought from patients enrolled on BMT CTN trials for data sharing with the CIBMTR for the conduct of research and for longitudinal follow up beyond the trial period. The Institutional Review Board of the NMDP approved the study.
Patients
Eligibility criteria for BMT CTN 0501 included: age 1–21 years, high risk leukemia, performance score of 70 or higher, adequate organ function and availability of two cord blood units with adequate cell dose (pre-cryopreservation TNC ≥2.5 × 107/kg in recipients of a single unit) and were HLA-matched to the patient and each other at least four of six HLA-loci (i.e., HLA match score of 6/6, 5/6 or 4/6) considering HLA-A, and –B at the antigen level and HLA-DRB1 at the allele-level.
The current study population includes two cohorts of patients with acute myeloid or lymphoblastic leukemia for whom data were retrieved from the CIBMTR’s database: patients treated on the single UCB arm of the BMT CTN 0501 protocol (n=100 of 113 enrolled; 28 centers) and those who underwent a single UCB unit transplant during the study period (n=396; 72 centers) in the United States and Canada (Figure 1). Twenty-five of these 72 centers (35%) also enrolled patients on BMT CTN 0501. Excluded were 13 patients on BMT CTN 0501 randomized to receive a single UCB unit (one did not proceed to transplantation after randomization and 12 had other malignant diseases with too few non-trial patients for comparison).
Figure 1.
Consort Diagram
Based on information reported to the CIBMTR the comparator group of 396 patients met inclusion criteria for BMT CTN 0501. While information on organ function was not collected on CIBMTR data collection Form, HCT-CI index, with known effect on survival after transplantation.4,5 was used as surrogate. The HCT-CI index for non-trial patients was comparable to that for patients enrolled on BMT CTN 0501.
Outcomes
Definitions for each outcome are detailed in the BMT CTN 0501 study protocol.1 The primary outcome for this analysis was overall survival; death from any cause was considered an event and surviving patients were censored at last follow-up. Secondary outcomes were leukemia-free survival (relapse or death as cause of treatment failure), non-relapse mortality (death in continuous remission), relapse (morphologic or cytogenetic evidence of AML or ALL consistent with pre-transplant features), hematopoietic recovery (neutrophil recovery ≥0.5 × 109/L and platelets ≥20 × 109/L unsupported for 7 days) and acute6 and chronic7 graft-versus-host disease (GVHD).
Statistical Methods
The characteristics of patients who received a single UCB unit on trial and the non-trial groups were compared using the chi-square test for categorical variables. The probabilities of hematopoietic recovery, infection and GVHD were calculated using the cumulative incidence estimator to accommodate competing risks.8 Comparison of overall survival, treatment failure, non-relapse mortality and relapse between treatment groups adjusting for variables associated with outcomes were performed using the Cox proportional hazards model.9 The probabilities of leukemia-free and overall survival, relapse and non-relapse mortality were generated from final Cox regression models.10,11
Variables considered included: treatment regimen (Trial TCF, non-Trial TCF, and non-Trial alternative regimens that did or didn’t include TBI); age (≤10 vs. >10 years); sex (male vs. female); performance score (90 and 100 vs. 70 and 80); recipient CMV serostatus (positive vs. negative); disease (AML vs. ALL); disease status (1st complete remission vs. 2nd complete remission vs. relapse); and cytogenetic risk (intermediate vs. poor risk). Models were built using step-wise forward selection and variables that met a significance level of ≥0.05 were held in the final model. The variable for treatment type was held in all steps of model building regardless of level of significance. All variables met the assumption of proportional hazards and there were no first order interactions between the variable for treatment type and other variables held in the final model. Transplant center effect was tested using the frailty model and a p-value ≤0.01 was considered significant.12 All p-values are two-sided. All analyses were done using SAS version 9.3 (Cary, NC).
Results
Patient and Disease Characteristics
Table 1A shows patients, disease and transplant characteristics of patients treated on BMT CTN 0501 (n=100), those who received the trial regimen (non-trial TCF, n=62) and those who received different myeloablative transplant-conditioning regimens (with or without TBI, n=334). The patient and disease characteristics of those treated on BMT CTN 0501 and non-trial TCF regimen were similar except that non-trial TCF patients were more likely to report performance scores of 80 or 70 (11% vs. 24%, p=0.04) and more likely to be transplanted in relapse (5% vs. 18%, p=0.02). The patient and disease characteristics of those treated on BMT CTN 0501 and non-trial alternative regimens were also similar except for age with recipients of non-trial regimens being more likely to be aged 1–10 years (53% vs. 76%, p<0.0001). For ALL transplants that occurred beyond CR1, the duration of CR1 was <36 months for 74% of patients on trial, 70% of those who received the TCF regimen off trial and 84% of those who received non-trial regimens. There were no differences in cytogenetic risk across treatment groups. For AML transplants, 9 of 41 (22%) patients treated on BMT CTN 0501 and 30 of 148 patients (20%) treated off trial with non-TCF regimens had MLL rearrangement. With regard to other factors, donor-recipient HLA-match, median infused TNC and median follow up was similar across the treatment groups. Of patients treated with TCF off-trial, transplants occurred at 23 centers of which 14 (60%) were centers enrolling patients on BMT CTN 0501. Of patients treated with non-TCF regimens, transplants occurred at 68 centers of which 23 centers (34%) also enrolled patients on BMT CTN 0501.
Table 1A.
Characteristics of patients and their disease and transplant
| Characteristics | BMT CTN 0501 TCF regimen | Non-Trial TCF regimen | Non-Trial non-TCF regimens* |
|---|---|---|---|
| Number | 100 | 62 | 334 |
| Transplant centers | 28 | 23 | 68 |
| Age, years | |||
| 1 – 5 | 18 (18) | 18 (29) | 166 (50) |
| 6 – 10 | 33 (33) | 21 (34) | 88 (26) |
| 11 – 16 | 32 (32) | 19 (31) | 57 (17) |
| 17 – 22 | 17 (17) | 4 (6) | 23 (7) |
| Sex | |||
| Male | 51 (51) | 37 (60) | 173 (52) |
| Female | 49 (49) | 25 (40) | 161 (48) |
| Race | |||
| Caucasian | 79 (79) | 47 (76) | 260 (78) |
| Other | 21 (21) | 15 (24) | 74 (22) |
| Performance score | |||
| 90 – 100 | 88 (88) | 47 (76) | 289 (87) |
| 80 | 8 (8) | 15 (24) | 27 (8) |
| 70 | 4 (4) | __ | 14 (4) |
| HCT-CI score | |||
| 0 – 2 | 84 (84) | 55 (89) | 291 (87) |
| ≥3 | 4 (4) | 3 (5) | 11 (3) |
| Not reported | 12 (12) | 4 (6) | 32 (10) |
| CMV seropositivity | |||
| Positive | 52 (52) | 43 (69) | 171 (51) |
| Negative | 39 (39) | 19 (31) | 148 (44) |
| Not reported | 9 (9) | 15 (4) | |
| Disease | |||
| AML | 41 (41) | 26 (42) | 148 (44) |
| ALL | 59 (59) | 36 (58) | 186 (56) |
| Extramedullary disease | |||
| None | 79 (79) | 48 (78) | 251 (75) |
| CNS | 15 (15) | 14 (22) | 75 (22) |
| Other sites | 6 (6) | __ | 8 (2) |
| White blood cell count | |||
| <20 × 109/L | 49 (49) | 24 (39) | 142 (43) |
| 20 – 100 × 109/L | 20 (20) | 17 (27) | 81 (24) |
| > 100 × 109/L | 14 (14) | 9 (15) | 49 (15) |
| Not reported | 17 (17) | 12 (19) | 62 (19) |
| Cytogenetic risk | |||
| Good risk | 4 (4) | __ | 9 (3) |
| Intermediate risk | 59 (59) | 31 (50) | 198 (59) |
| Poor risk | 31 (31) | 17 (27) | 97 (29) |
| Not reported | 6 (6) | 14 (23) | 30 (9) |
| Disease status | |||
| 1st complete remission | 41 (41) | 27 (44) | 126 (38) |
| 2nd complete remission | 43 (43) | 22 (35) | 147 (44) |
| 3rd complete remission | 11 (11) | 2 (3) | 26 (8) |
| Relapse | 5 (5) | 11 (18) | 35 (10) |
| Follow up, median (range), months | 48 (24 – 80) | 50 (8 – 75) | 61 (4 – 90) |
Abbreviations:
TCF = total body irradiation (1320 cGy), cyclophosphamide (120 mg/kg), fludarabine (75 mg/m2)
Includes TBI containing and chemotherapy only regimens
HCT-CI: hematopoietic cell transplant co-morbidity index
AML = acute myeloid leukemia
ALL = acute lymphoblastic leukemia
CNS = central nervous system
Sixty-three percent of non-trial alternative regimens included TBI and most of these regimens included cyclophosphamide (Table 1B). Busulfan with cyclophosphamide was the predominant chemotherapy only regimen (Table 1B). All patients received calcineurin inhibitor containing prophylaxis. Sixty-six percent of transplantations included anti-thymocyte globulin (ATG) which was rabbit-derived (72 of 219, 33%) or horse derived (106 of 219, 48%). The median dose of rabbit ATG was 8 mg/kg (range 6–12) and horse ATG, 90 mg/kg (range 60 – 120). The type of ATG was not reported for 41 patients (19%). Further exploration of recipients of non-trial regimens, revealed the predominant population who received TBI-containing compared to non-TBI regimens were older (median age 6 years vs. 3 years), more likely to have ALL compared to AML (80% vs. 20%) and were less likely to be in 1st compared to 2nd CR (38% vs. 59%). There were no differences in the proportion of patients transplanted in relapse (9% vs. 11%). The characteristics of patients who received non-trial non-TCF regimens by leukemia type and inclusion of ATG to their conditioning regimen are shown in Supplemental Table S1A, S1B.
Table 1B.
Transplant Characteristics
| Characteristics | BMT CTN 0501 TCF regimen | Non-Trial TCF regimen | Non-Trial non-TCF regimens* |
|---|---|---|---|
| Number of patients | 100 | 62 | 334 |
| TBI-containing regimens | |||
| TBI (1320 cGy) + Cy + fludarabine | 100 | 62 | __ |
| TBI (1350 cGy) + Cy + fludarabine | 5 (<1%) | ||
| TBI + Cy | __ | __ | 69 (21) |
| TBI + Cy + thiotepa | __ | __ | 50 (15) |
| TBI + Cy + etoposide | __ | __ | 34 (10) |
| TBI + Cy + Ara-C | __ | __ | 13 (4) |
| TBI + melphalan | __ | __ | 21 (6) |
| TBI + other agents | __ | __ | 18 (5) |
| Chemotherapy only regimens | |||
| Busulfan + Cy | __ | __ | 87 (26) |
| Busulfan + melphalan | __ | __ | 19 (6) |
| Busulfan + fludarabine | __ | __ | 18 (5) |
| In vivo T-cell depletion | |||
| Anti-thymocyte globulin | __ | __ | 219 (66) |
| None | 100 (100) | 62 (100) | 115 (34) |
| GVHD Prophylaxis | |||
| Cyclosporine + mycophenolate | 100 (100) | 61 (98) | 91 (27) |
| Cyclosporine + methotrexate | __ | 1 (2) | 25 (7) |
| Cyclosporine alone | __ | __ | 107 (32) |
| Tacrolimus + mycophenolate | __ | __ | 28 (8) |
| Tacrolimus + methotrexate | __ | __ | 68 (20) |
| Tacrolimus alone | __ | __ | 15 (4) |
| Donor-recipient HLA match | |||
| 6/6 | 15 (15) | 7 (11) | 70 (21) |
| 5/6 | 49 (49) | 36 (58) | 170 (51) |
| 4/6 | 36 (36) | 19 (31) | 94 (28) |
| Median infused TNC | 3.9 × 107/kg | 4.5 × 107/kg | 5.9 × 107/kg |
Abbreviation:
TCF = total body irradiation (1320 cGy), cyclophosphamide (120 mg/kg), fludarabine (75 mg/m2)
Includes TBI containing and chemotherapy only regimens
TBI = total body irradiation (dose ≥1000 cGy)
Cy = cyclophosphamide
TNC = total nucleated cell
Outcomes
Comparison between BMT CTN 0501 (TCF) and Non-trial TCF Transplants
Transplantation outcomes are shown in Tables 2A and 2B, Figures 2, 3A, 3B. Compared to patients treated on BMT CTN 0501, non-trial TCF patients had similar 5-year survival rates adjusted for age, recipient CMV serostatus and disease status at transplantation. Cytogenetic risk was not associated with survival after transplantation. Similarly, there were no differences in leukemia-free survival, non-relapse mortality, relapse and hematopoietic recovery between patients treated on BMT CTN 0501 and non-trial TCF. The 5-year probabilities of leukemia-free survival for patient treated on BMT CTN 0501 was 61% (95% CI 51–70) compared to 57% (95% CI 43–68) for those treated with the TCF regimen off-trial. Although the GVHD prophylaxis regimen for non-trial TCF patients with the exception of a single patient was identical to that for BMT CTN 0501 the 3-month incidence of grade II–IV acute GVHD was higher but not grade III–IV acute or chronic GVHD. There were no differences in day 100 incidences of bacterial, viral or fungal infections; 83% (95% CI 75 – 90%) and 77% (95% CI 66 – 87) after BMT CTN 0501 and non-trial TCF transplants, respectively, p=0.46.
Table 2A.
Analysis of Risk Factors: BMT CTN 0501 (TCF regimen) compared to Non-trial TCF regimen and Non-trial regimens
| Outcome | Hazard Ratio (95% confidence interval) | P-value |
|---|---|---|
| Overall Mortality | ||
| TCF regimen | 1.00 | |
| Non-trial TCF regimen | 1.27 (0.74 – 2.18) | 0.37 |
| Non-trial TBI regimen | 1.97 (1.30 – 2.99) | 0.001 |
| Non-trial chemotherapy only regimen | 1.88 (1.18 – 2.98) | 0.007 |
| Other factors associated with mortality | ||
| Age | ||
| 1 – 5 years | 1.00 | |
| 6 – 21 years | 1.29 (0.96 – 1.72) | 0.08 |
| Cytomegalovirus serostatus | ||
| Serostatus negative | 1.00 | |
| Serostatus positive | 1.44 (1.08 – 1.90) | 0.011 |
| Disease | ||
| Acute lymphoblastic leukemia | 1.00 | |
| Acute myeloid leukemia | 1.65 (1.19 – 2.28) | 0.002 |
| Disease status | ||
| 1st complete remission | 1.00 | |
| 2nd complete remission/relapse | 1.79 (1.34 – 2.40) | <0.001 |
| Relapse | ||
| TCF regimen | 1.00 | |
| Non-trial TCF regimen | 1.15 (0.59 – 2.25) | 0.66 |
| Non-trial TBI regimen | 1.17 (0.71 – 1.96) | 0.52 |
| Non-trial chemotherapy only regimen | 1.88 (1.08 – 3.27) | 0.025 |
| Other factors associated with relapse | ||
| Age | ||
| 1 – 5 years | 1.00 | |
| 6 – 21 years | 0.65 (0.45 – 0.93) | 0.020 |
| Disease status | ||
| 1st complete remission | 1.00 | |
| 2nd complete remission/relapse | 2.69 (1.80 – 4.01) | <0.001 |
| Non-relapse Mortality | ||
| TCF regimen | 1.00 | |
| Non-trial TCF regimen | 1.09 (0.52 – 2.28) | 0.81 |
| Non-trial TBI regimens | 1.69 (0.99 – 2.90) | 0.053 |
| Non-trial chemotherapy only regimens | 1.75 (0.91 – 3.33) | 0.09 |
| Other factors associated with non-relapse mortality | ||
| Age | ||
| 1 – 5 years | 1.00 | |
| 6 – 21 years | 2.10 (1.32 – 3.35) | 0.002 |
| Cytomegalovirus serostatus | ||
| Serostatus negative | 1.00 | |
| Serostatus positive | 1.66 (1.09 – 2.52) | 0.016 |
| Treatment failure (inverse of LFS) | ||
| TCF regimen | 1.00 | |
| Non-trial TCF regimen | 1.12 (0.68 – 1.85) | 0.63 |
| Non-trial TBI regimens | 1.41 (0.97 – 2.04) | 0.07 |
| Non-trial chemotherapy only regimens | 1.86 (1.22 – 2.83) | 0.003 |
| Other factors associated with treatment failure | ||
| Cytomegalovirus serostatus | ||
| Serostatus negative | 1.00 | |
| Serostatus positive | 1.45 (1.11 – 3.39) | 0.019 |
| Disease status | ||
| 1st complete remission | 1.00 | |
| 2nd complete remission/relapse | 1.76 (1.34 – 2.32) | <0.001 |
Abbreviation:
TCF = total body irradiation (1320 cGy), cyclophosphamide (120 mg/kg), fludarabine (75 mg/m2)
TBI = total body irradiation
LFS = leukemia-free failure
Table 2B.
Hematopoietic recovery, Acute and Chronic GVHD: BMT CTN 0501 (TCF regimen) compared to Non-trial TCF regimen and Non-trial regimens
| Outcome | Trial TCF regimen | Non-trial TCF regimen | Non-trial TBI regimens | Non-trial chemotherapy only |
|---|---|---|---|---|
| Neutrophil recovery | 69% (60 – 78) | 74% (63 – 84) p-value NS |
65% (59 – 72) p-value NS |
68% (60 – 76) p-value NS |
| Platelet recovery | 75% (65 – 83) | 80% (68 – 89) p-value NS |
68% (62 – 75) p-value NS |
64% (55 – 73) p-value NS |
| Acute GVHD | ||||
| Grade II – IV | 47% (37 – 57) | 66% (54 – 78) p=0.01 |
34% (28 – 41) p=0.04 |
31% (23 – 40) p=0.02 |
| Grade III – IV | 17% (10 – 25) | 28% (17 – 40) p-value NS |
16% (12 – 22) p-value NS |
18% (12 – 25) p-value NS |
| Chronic GVHD | 43% (33 – 53) | 37% (25 – 50) p-value NS |
32% (25 – 38) p-value NS |
31% (23 – 40) p-value NS |
Abbreviation:
GVHD = graft versus host disease
TCF = total body irradiation (1320 cGy), cyclophosphamide (120 mg/kg), fludarabine (75 mg/m2)
Figure 2. Overall Survival.
The 5-year probabilities of overall survival adjusted for age, recipient CMV serostatus, disease and disease status for patients enrolled on BMT CTN 0501 TCF regimen (A, 70% [95% CI 60–77]), non-trial TCF regimen (B, 62% [95% CI 48–73]), non-trial chemotherapy only (C, 49% [95% CI 39–59]) and non-trial TBI regimens (D, 47% [95% CI 40–54])
Figure 3.
Figure 3A. Relapse
The 5-year probabilities of relapse adjusted for age and disease status for patients enrolled on BMT CTN 0501 TCF regimen (A, 24% [95% CI 15–33]) non-trial TCF regimen (B, 27% [95% CI 16–38]), non-trial chemotherapy only (C, 37% [95% CI 28–45]) and non-trial TBI regimens (D, 26% [95% CI 21–32])
Figure 3B. Non-relapse mortality
The 5-year probabilities of non-relapse mortality adjusted for age and disease status for patients enrolled on BMT CTN 0501 TCF regimen (A, 16% [95% CI 10–23]) non-trial TCF regimen (B, 18% [95% CI 10–28]), non-trial chemotherapy only (C, 24% [95% CI 16–33]) and non-trial TBI regimens (D, 25% [95% CI 19–31])
Comparison between BMT CTN 0501 (TCF) and Non-trial alternative regimens: Overall and Leukemia-free Survival
Compared to patients treated on BMT CTN 0501, overall mortality risks were significantly higher after alternative regimen transplants regardless of whether the regimen included TBI or not (Tables 2A, Figure 2). A similar trend was seen for treatment failure (inverse of leukemia-free survival). The 5-year probability of leukemia-free survival adjusted for recipient CMV serostatus and disease status at transplantation for BMT CTN 0501, non-trial, TBI-containing and chemotherapy only regimens were 61% (95% CI 51–70), 48% (95% CI 41–55) and 40% (95% CI 31–49), respectively. The only disease characteristic associated with overall and leukemia-free survival was disease status at transplantation; leukemia type or cytogenetic risk was not associated with either outcome. We examined carefully for a random effect of transplant center on survival using the frailty model (p=0.06) and a frailty Cox model and found none (Supplemental Table S2). We also performed a subset analysis that compared patients enrolled on BMT CTN 0501 to patients who received alternative regimens at BMT CTN 0501 centers; consistent with the main analysis, mortality risks were higher with the use of non-trial regimens (HR 1.63, 95% CI 1.05 – 2.55, p=0.03).
Other subset analysis was undertaken to confirm our findings of the main analysis. In recipients of non-trial regimens, we compared survival results in those treated with chemotherapy only to TBI-containing regimens. After adjusting for age, disease, disease status at transplantation and CMV seropositivity, survival was similar (HR 0.95, 95% CI 0.64 – 1.41, p=0.80. Consistent with general clinical practice ATG was included in the conditioning regimen for 66% of non-TCF transplants. We found no effect of ATG on survival (HR 1.01, 95% CI 0.73 – 1.39; p=0.95) consistent with other reports13–15 or an interaction between regimen and inclusion of ATG (p=0.78). The 5-year probabilities of leukemia-free and overall survival are shown in Supplemental Table S3.
Relapse and Non-Relapse Mortality
Compared to patients treated on BMT CTN 0501, relapse risks were significantly higher after chemotherapy only conditioning regimens (Table 2A, Figure 3A). In contrast, there were no significant differences in relapse risks comparing patients treated on BMT CTN 0501 and non-trial TBI-containing conditioning regimens. Cytogenetic risk was not associated with relapse after transplantation. Among non-trial patients, relapse risk was significantly higher for the chemotherapy alone regimens compared to TBI-containing regimens (HR 1.61, 95% CI 1.06 – 2.42, p=0.02).
Compared to BMT CTN 0501, non-relapse mortality risks were not significantly different for non-trial patients (Table 2A, Figure 3B). Similarly, among non-trial patients, non-relapse mortality risk was not significantly different with chemotherapy only compared to TBI-containing regimens (HR 0.91, 95% CI 0.54 – 1.55, p=0.73). The day 100 incidences of bacterial, viral or fungal infections were 83% (95% CI 75 – 90%), 70% (95% CI 64 – 76) and 75% (95% 67 – 83) after BMT CTN 0501, non-trial TBI and chemotherapy only regimens, respectively (p=0.45).
As with leukemia-free and overall survival we found no effect of ATG on relapse or non-relapse mortality. The 5-year probabilities of relapse and non-relapse mortality are shown in Supplemental Table S3.
Hematopoietic recovery and Graft-versus-host disease
Compared to BMT CTN 0501 there were no differences in the incidence of neutrophil and platelet recovery between the treatment groups (Table 2B). Compared to BMT CTN 0501 the incidence of grade II–IV acute GVHD was lower with non-trial regimens. We hypothesized that the lower rate of acute GVHD with non-trial regimens may be explained by the inclusion of ATG to these regimens (Table 2B). However, there were no differences in the incidence of grade III–IV acute or chronic GVHD between the treatment groups. The day 100 incidences of bacterial, viral or fungal infections was lower with non-trial, non-TCF regimens compared to BMT CTN 0501, 72% (95% CI 67 – 77) and 83% (95% CI 75 – 90%), respectively, p=0.007.
Discussion
While it was postulated that higher cell dose in recipients of two UCB units would be associated with improved engraftment and survival, 16 these outcomes in patients enrolled in the BMT CTN 0501 trial were improved in both arms when compared to published reports.1,2,3,17 Leveraging data reported to an observational registry and after carefully controlled analyses, our findings support the hypothesis that transplant-conditioning/immunoprophylaxis treatment package may indeed be an important factor for overall and leukemia-free survival in pediatric recipients with acute leukemia transplanted with UCB. Our observations also support the generalizability of the treatment package as overall and leukemia-free survival of recipients of non-trial TCF regimen was comparable to that observed on trial. Importantly, this analysis also extends the follow-up in recipients of a single UCB transplant with acute leukemia enrolled in BMT CTN 0501. At a median of 4 years (range, 2–6.7 years) of follow up, the 5-year adjusted probability of survival is 70% and that of relapse is 24% compared to 73% and 12% at 1 year in the original intent-to-treat report.1
Randomized clinical trials (RCT) offer the highest quality data for modifying clinical practice.18 Randomization provides unbiased allocation to the treatment arms being studied and ensures that even if the groups being studied are not identical with respect to relevant prognostic factors, such differences will be due to chance, allowing statistical theory based on random sampling to calculate confidence intervals.19,20 Limitations of RCTs, however, include eligibility criteria that typically exclude higher comorbidity patients and an assessment of the generalizability of their results. Still a discussion about generalizability is a quality indicator for RCT reporting within the Consolidated Standards of Reporting Trials.21 To our knowledge, the current analysis is the first study to evaluate the generalizability of the results of a RCT of hematopoietic cell transplantation in children with acute leukemia. Here, we provide data that confirm the generalizability of the primary endpoint, overall survival, of the standard of care arm of BMT CTN 0501 to a population of trial-eligible patients who received an identical transplant-conditioning region as offered on trial. However, when trial and non-trial transplant-conditioning regimens differed in a population of trial-eligible patients, those treated on trial had better overall and leukemia-free survival after adjustment for risk factors such as age, disease, disease status, cytogenetic risk and CMV serostatus. In addition, the survival difference occurred beyond the first year after transplantation, suggesting that the apparent survival advantage cannot be fully explained by the potential exclusion of sicker patients from the trial.22 The higher mortality risk seen with non-trial alternative regimen transplants is in part attributed to higher leukemia recurrence with chemotherapy only regimens and a higher though not statistically significant non-relapse mortality with non-trial TBI-containing regimens.
The predominant non-trial TBI-containing regimen was TBI with cyclophosphamide and predominantly used for ALL. The inclusion of another agent in addition to cyclophosphamide may explain the observed higher non-relapse mortality risk.23 The role of TBI-containing versus non-TBI regimens offer mixed results, with one recent report that showed no significant differences in outcome for children with AML receiving TBI with cyclophosphamide or busulfan and cyclophosphamide24 and the other, fewer relapse with busulfan, cyclophosphamide and melphalan.25 None of these reports included recipients of UCB transplants. The unique characteristics of the UCB graft may require a regimen that is different from others to optimize transplant-outcomes. About half of non-trial participants who received a different transplant conditioning regimen were aged 1 – 5 years compared to 18% of patients enrolled on trial and AML effect of conditioning regimen on outcomes was independent of age and leukemia type. Consequently, we recommend the use of the TCF treatment regimen for all children with ALL and AML transplanted with UCB and acknowledge these recommendations do not apply to other donor or graft sources. We also recognize the desire to avoid irradiation-containing regimens in very young children and the results of the current analysis should prompt careful consideration of the relative benefit for fewer relapse versus the longer-term burden of morbidity of using TBI in very young children when considering UCB transplantation.
Of note, off-trial recipients of TCF/immunoprophylaxis package had a higher incidence of grade II acute GVHD compared to those enrolled on trial. A plausible explanation may be related to the standardized retrospective rescoring algorithm for acute GVHD for trial recipients; e.g., assignment of engraftment syndrome rather than acute GVHD when skin involvement occurred in isolation and with symptoms consistent with engraftment syndrome. The level of detailed data required for rescoring acute grade II GVHD were not available for off-trial recipients. The incidence of grade II acute GVHD was significantly lower for recipients of non-trial alternative regimens due to the frequent inclusion of ATG in those regimens. This may in part explain the lower incidence of infections at day 100 although we were unable to detect differences in early survival.
The most reliable way to establish the association between trial participation and survival would be to randomize patients to be offered a clinical trial or not and among those assigned to the clinical trial arm, a second randomization to assign treatment arms.26 Such a study would be ethically and practically challenging. Consequently, comparison of outcomes of trial versus non-trial participation has thus relied on observational studies such as ours.27 Another potential source of bias is data acquisition for non-trial patients that we addressed by using the same data source for trial and non-trial patients ensuring consistent patterns of endpoint assessment. The validity of data reported to the CIBMTR and that reported for the trial were verified and consistency confirmed. Finally, we do not know the reasons for non-participation. Willingness to participate in a clinical trial, physician bias with regard to treatment offered or bias with regard to prognosis or regulatory burden associated with trial participation might have influenced non-participation. It is unlikely patients were excluded on trial eligibility as there were no differences in trial versus non-trial representation with regard to patient demographics,28 performance score, co-morbidity index, disease, cytogenetic risk and disease status. Access is unlikely to have been a barrier as the trial was available to all U.S. sites with an allogeneic transplant program within the Children’s Oncology Group. In conclusion, the results of BMT CTN 0501 appear generalizable to the population of trial eligible patients and the survival differences between trial-specified and other regimens indicate the importance of conditioning regimen for UCB transplantation.
Supplementary Material
Highlights.
Generalizability of results of a phase III trial for cord blood transplantation
Superior survival with trial-specified regimen compared to other regimens
The relative importance of conditioning regimen for UCB transplant
Acknowledgments
Support for BMT CTN 0501 was provided by U10-HL069294 from the National Heart, Lung, and Blood Institute and National Cancer Institute to the Blood and Marrow Transplant Clinical Trials Network and U10CA098543 from the National Cancer Institute to the Children’s Oncology Group; the Center for International Blood and Marrow Transplant Research is supported primarily by U24-CA76518 from the National Cancer Institute, the National Heart, Lung, and Blood Institute, and the National Institute of Allergy and Infectious Diseases; HHSH250201200016C (HRSA/DHHS) to the Center for International Blood and Marrow Transplant Research. The content is solely the responsibility of the authors and do not represent the official policy of the National Institutes of Health or the Health Resources and Services Administration or any other agency of the United States Government.
Footnotes
Financial Disclosure Statement:
The authors declare they have no relevant or material financial interests that relate to the research described in this paper.
Author Contribution
ME, GH and MJZ designed the study. GH and MF prepared the dataset and MJZ analyzed the data. ME drafted the first manuscript. JK, MJZ, AM, GH, MF, KWC, SDO, KRS, DW, MMH and JEW critically reviewed the manuscript. All authors approved the final manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Wagner JE, Eapen M, Carter S, et al. One-unit versus two-unit cord blood transplantation for hematologic cancers. N Engl J Med. 2014;371:1685–1694. doi: 10.1056/NEJMoa1405584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kurtzberg J, Prasad VK, Carter SL, et al. Results of the Cord Blood Transplantation Study (COBLT): clinical outcomes of unrelated donor umbilical cord blood transplantation in pediatric patients with hematologic malignancies. Blood. 2008;112:4318–4327. doi: 10.1182/blood-2007-06-098020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kurtzberg J, Carter SL, Mendizabal AM, et al. Superior Survival after Single Unit Umbilical Cord Blood Transplantation (UCBT) in Children with Hematological Malignancies Treated on Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0501 Relative to the Cord Blood Transplantation (COBLT) Trial. Biol Blood Marrow Transplant. 2013;19(suppl):S121. abstr 23. [Google Scholar]
- 4.Sorror M, Maris M, Storb R, et al. Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106:2912–2919. doi: 10.1182/blood-2005-05-2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Smith AR, Majhail NS, MacMillan ML, et al. Hematopoietic cell transplantation comorbidity index predicts transplantation outcomes in paediatric patients. Blood. 2011;117:2728–2734. doi: 10.1182/blood-2010-08-303263. [DOI] [PubMed] [Google Scholar]
- 6.Przepiorka D, Weisdorf D, Martin P, et al. 1994 consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15:825–828. [PubMed] [Google Scholar]
- 7.Weisdorf DJ, Hurd D, Carter S, et al. Prospective grading of graft-versus-host disease after unrelated donor marrow transplantation: a grading algorithm versus blinded expert panel review. Biol Blood Marrow Transplant. 2003;9:512–518. doi: 10.1016/s1083-8791(03)00162-9. [DOI] [PubMed] [Google Scholar]
- 8.Lin DY. Non-parametric inference for cumulative incidence functions in competing risks. Stat Med. 1997;16:901–910. doi: 10.1002/(sici)1097-0258(19970430)16:8<901::aid-sim543>3.0.co;2-m. [DOI] [PubMed] [Google Scholar]
- 9.Cox DR. Regression model and life-tables. J R Stat Soc Series B. 1972;34:187–200. [Google Scholar]
- 10.Zhang X, Loberiza FR, Klein J, Zhang MJ. A SAS macro for estimation of direct adjusted survival curves based on a stratified Cox regression model. Comput Methods Programs Biomed. 2007;88:95–101. doi: 10.1016/j.cmpb.2007.07.010. [DOI] [PubMed] [Google Scholar]
- 11.Zhang X, Zhang MJ. SAS macros for estimation of direct adjusted cumulative incidence curves under proportional subdistribution hazards models. Comput Methods Programs Biomed. 2001;101:87–93. doi: 10.1016/j.cmpb.2010.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Andersen PK, Klein JP, Zhang MJ. Testing for center effects in multi-center survival studies: A Monte Carlo comparison of fixed and random effects tests. Stats Med. 1999;18:1489–500. doi: 10.1002/(sici)1097-0258(19990630)18:12<1489::aid-sim140>3.0.co;2-#. [DOI] [PubMed] [Google Scholar]
- 13.Veys P, Wynn RF, Ahn KW, et al. Impact of immune modulation with in vivo T-cell depletion and myeloablative conditioning on outcomes after unrelated donor transplantation for childhood acute lymphoblastic leukemia. Blood. 2012;119:6155–6161. doi: 10.1182/blood-2012-01-405795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ponce DM, Eapen M, Sparapani R, et al. In vivo T-cell depletion with myeloablative regimens on outcomes after cord blood transplantation for acute lymphoblastic leukemia in children. Biol Blood Marrow Transplant. 2015;21:2173–2179. doi: 10.1016/j.bbmt.2015.08.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lindemans CA, Chiesa R, Amrolia PJ, et al. Impact of thymoglobulin prior to pediatric unrelated umbilical cord blood transplantation on immune reconstitution and clinical outcome. Blood. 2014;123:126–132. doi: 10.1182/blood-2013-05-502385. [DOI] [PubMed] [Google Scholar]
- 16.Wagner JE, Barker JN, DeFor TE, et al. Transplantation of unrelated donor umbilical cord blood in 102 patients with malignant and non-malignant diseases: influence of CD34 cell dose and HLA disparity on treatment-related mortality and survival. Blood. 2002;100:1611–1618. doi: 10.1182/blood-2002-01-0294. [DOI] [PubMed] [Google Scholar]
- 17.Eapen M, Rubinstein P, Zhang M-J, et al. Comparison of outcomes after transplantation of unrelated donor umbilical cord blood and bone marrow in children with acute leukemia. Lancet. 2007;369:1947–1954. doi: 10.1016/S0140-6736(07)60915-5. [DOI] [PubMed] [Google Scholar]
- 18.Sniderman AD, LaChapelle KJ, Rachon NA, et al. The necessity for clinical reasoning in the era of evidenced-based medicine. Mayo Clin Proc. 2013;88:1108–1114. doi: 10.1016/j.mayocp.2013.07.012. [DOI] [PubMed] [Google Scholar]
- 19.Vandenbroucke JP. When are observational studies as credible as randomised trials? Lancet. 2004;363:1728–1731. doi: 10.1016/S0140-6736(04)16261-2. [DOI] [PubMed] [Google Scholar]
- 20.Fisher RA. Statistical methods, experimental design, and scientific inference. Oxford, United Kingdom: Oxford University Press; 1990. [Google Scholar]
- 21.Moher D, Schulz KF, Altman DG. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. Ann Intern Med. 2001;134:657–662. doi: 10.7326/0003-4819-134-8-200104170-00011. [DOI] [PubMed] [Google Scholar]
- 22.Unger JM, Barlow WE, Martin DP, et al. Comparison of survival outcomes among cancer patients treated in and out of clinical trials. J Natl Cancer Inst. 2014;106:1–13. doi: 10.1093/jnci/dju002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tracey J, Zhnag MJ, Theil E, Sobocinski KA, Eapen M. Transplantation conditioning regimens and outcomes after allogeneic hematopoietic cell transplantation in children and adolescents with acute lymphoblastic leukemia. Biol Blood Marrow Transplant. 2013;19:255–259. doi: 10.1016/j.bbmt.2012.09.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ishida H, Kato M, Kudo K, et al. Comparison of outcomes for pediatric patients with acute myeloid leukemia in remission and undergoing allogeneic hematopoietic cell transplantation with myeloablative conditioning regimens based on either intravenous busulfan or total body irradiation: a report from the Japanese Society for Hematopoietic Cell Transplantation. Biol Blood Marrow Transplant. 2015;21:2141–2147. doi: 10.1016/j.bbmt.2015.08.011. [DOI] [PubMed] [Google Scholar]
- 25.Lucchini G, Labopin M, Beohou E, et al. Impact of conditioning regimen on outcomes for children with acute myeloid leukemia undergoing transplantation in first complete remission. An analysis on behalf the pediatric disease working party of the European Group for Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2017;23:467–474. doi: 10.1016/j.bbmt.2016.11.022. [DOI] [PubMed] [Google Scholar]
- 26.Peppercorn JM, Weeks JC, Cook EF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet. 2004;363:263–270. doi: 10.1016/S0140-6736(03)15383-4. [DOI] [PubMed] [Google Scholar]
- 27.Edwards SJ, Lilford RJ, Braunholtz DA, Jackson JC, Hewison J, Thornton J. Ethical issues in the design and conduct of randomized controlled trials. Health Technol Assess. 1998;2:i–vi. 1–132. [PubMed] [Google Scholar]
- 28.Murthy VH, Krumholz HM, Gross CP. Participation in clinical trials. JAMA. 2004;291:2720–2726. doi: 10.1001/jama.291.22.2720. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.