TO THE EDITOR:
Allogeneic hematopoietic cell transplantation (allo-HCT) is used as consolidation therapy in children with high-risk hematological malignancies. This curative treatment option comes with high morbidity and mortality. One of the critical challenges in improving clinical outcome is designing an effective conditioning regimen with limited toxicity. The core of standard conditioning regimens for allo-HCT in children consists of total body irradiation (TBI) or alkylating cytotoxic agents. Conventionally, the nucleoside analog fludarabine (Flu) has been included in most reduced-intensity regimens together with the alkylating agent IV busulfan (Bu). The second-generation purine analog clofarabine (Clo) was recently added to the conditioning1 due to its synergistic antileukemic effect in combination with FluBu.2,3 A recently published multicenter prospective study in pediatric patients with acute lymphoblastic leukemia (ALL) showed that TBI-based conditioning results in better outcome than chemotherapy-based conditioning.4 Nevertheless, these findings should be weighed against the well-known late toxic effects of TBI.5 We recently reported that CloFluBu is a good alternative for TBI-based conditioning in ALL in first complete remission and an effective, less toxic strategy in patients with acute myeloid leukemia (AML).6
Optimizing the pharmacokinetic (PK) exposures of conditioning agents can improve clinical outcomes, and the need for an optimal exposure has so far been reported for Bu,7,8 Flu,9,10 and antithymocyte globulin (ATG).11,12 PK studies on Clo in children undergoing allo-HCT show that dosing based on body weight and renal function would lead to the best predictable exposure.13,14 Studies associating Clo PK with clinical outcome after allo-HCT have not been performed so far. Furthermore, the reduced Flu dose in the CloFluBu conditioning compared with the dose in the FluBu conditioning leads to the question whether the previously reported risk for therapy-related mortality (TRM), associated with CD4 immune reconstitution (IR), and graft failure also applies in the context of CloFluBu conditioning.9,10 Therefore, we conducted a retrospective cohort analysis to examine the effect of Flu and Clo exposures on clinical outcome after allo-HCT.
All patients with high-risk hematological malignancies receiving an allo-HCT in the University Medical Center Utrecht/Princess Máxima Center between October 2011 and October 2019 were included after written informed consent in accordance with the Declaration of Helsinki. The study was approved under trial number 11/063-k.
Patients were treated as previously described.6 Conditioning consisted of Clo and Flu on 4 consecutive days at a daily dosage of 30 mg/m2 and 10 mg/m2, respectively, in combination with IV-administered Bu, targeted at an area under the concentration time curve (AUCT0−∞) of 85 to 95 mg∗hour/L.7,8 ATG was administered, and subsequent total exposures were determined as previously reported.6,11,12,15,16 Absolute numbers of immune cells were measured by flow cytometry at least every other week up to 12 weeks after allo-HCT and monthly thereafter.
Concentrations were measured using a liquid chromatography-tandem mass spectrometry method validated according to European Medicines Agency guidelines, with a lower limit of quantification of 0.001 μg/L.17 The total exposures (AUCT0−∞) were determined by applying population PK models13,18 with maximum a posteriori Bayesian estimation using the software package NONMEM version 7.5.0.
The main outcome of interest was event-free survival (EFS), defined as survival without the events of graft failure and relapse. Additional outcomes of interest were overall survival (OS), graft-versus-host disease (GVHD)–free relapse-free survival (GFRS), and the cumulative incidence of TRM, CD4 IR, relapse, acute GVHD grade 2 to 4, and chronic GVHD (cGVHD). For OS, death from any cause was considered an event. For GFRS, acute GVHD grade 3 to 4 and extensive cGVHD, TRM, and relapse were considered events. CD4 IR was defined as measurement above a value of >50 per microliter CD4+ T cells within 100 days after HCT.16,19
The association of exposures with clinical outcome was explored using martingale residuals and further fitted by Cox proportional hazards and Fine and Gray competing risk models. Relapse was considered a competing event for TRM. For relapse, graft failure and TRM were considered as competing events. Competing events considered for GVHD were graft failure, TRM, and relapse. Analyses were performed using R 4.3.1 with packages survival, cmprsk, and survminer.
A total of 103 children receiving an allo-HCT for ALL (44%), AML (37%), or other high-risk hematological malignancy (19%; of which 55% was diagnosed with myelodysplastic syndrome) were included (Table 1). The median age was 11.0 years (range, 0.5-18.6) at the time of allo-HCT. Most donors were unrelated (78%). Grafts were derived from bone marrow (43%), cord blood (56%), and peripheral blood (1%). The median follow-up time was 1.4 years (range, 0.04-8.2). Thirty seven of 46 patients (80%) treated with ATG received model-based dosing.12 IR data of 87 patients (84%) and ATG exposures of 22 patients (46%) were available. The missing ATG exposures were predicted using a previous validated population pharmacokinetic/pharmacodynamic model.15,16 Neutrophil and monocyte counts normalized within weeks after allo-HCT. The 5-year EFS was 56%, 64%, and 50%, and the 5-year OS was 67%, 71%, and 55% for the whole cohort, patients with ALL, and patients with AML, respectively (supplemental Figure 1).
. | All patients . |
---|---|
No. of patients | 103 |
Female | 44 (43%) |
Age at transplant, median (range), y | 11.0 (0.5-18.6) |
Renal function, median (range), mL/min per 1.73 m2 | 140 (69.3-140) |
Cell source | |
Bone marrow | 44 (43%) |
Cord blood | 58 (56%) |
Peripheral blood | 1 (1%) |
Donor type | |
Unrelated | 80 (78%) |
Related | 23 (22%) |
Diagnosis type | |
ALL | 45 (44%) |
AML | 38 (37%) |
Other | 20 (19%) |
Year of transplantation | |
2011-2013 | 16 (15%) |
2014-2016 | 45 (44%) |
2017-2019 | 42 (41%) |
No. of transplant | |
First | 99 (96%) |
Second | 3 (3%) |
Third | 1 (1%) |
ATG | |
Yes | 46 (45%) |
No | 57 (55%) |
. | All patients . |
---|---|
No. of patients | 103 |
Female | 44 (43%) |
Age at transplant, median (range), y | 11.0 (0.5-18.6) |
Renal function, median (range), mL/min per 1.73 m2 | 140 (69.3-140) |
Cell source | |
Bone marrow | 44 (43%) |
Cord blood | 58 (56%) |
Peripheral blood | 1 (1%) |
Donor type | |
Unrelated | 80 (78%) |
Related | 23 (22%) |
Diagnosis type | |
ALL | 45 (44%) |
AML | 38 (37%) |
Other | 20 (19%) |
Year of transplantation | |
2011-2013 | 16 (15%) |
2014-2016 | 45 (44%) |
2017-2019 | 42 (41%) |
No. of transplant | |
First | 99 (96%) |
Second | 3 (3%) |
Third | 1 (1%) |
ATG | |
Yes | 46 (45%) |
No | 57 (55%) |
The AUC0−∞ ranged from 2.9 to 12.7 mg∗h/L for Flu and from 1.8 to 6.1 mg∗h/L for Clo. Exposures of Flu and Clo as continuous variables were not shown to be predictive for clinical outcome (Table 2). In addition, martingale residuals of the null Cox proportional hazard models were plotted against the Flu and Clo AUCT0−∞, and the patterns of the martingale residuals did not suggest a relationship with clinical outcome. It should be noted that an effect on graft failure might be there but could not be detected due to the limited number of events (n = 5). When analyzing patients with ALL and those with AML separately, the AUC0−∞ ranges for Flu and Clo were similar, and no effect on OS for both Flu (P = .61 and P = .84) and Clo (P = .80 and P = .12) was found. In addition, Flu and Clo exposures were not predictors for EFS in either patients with ALL (Flu, P = .38; Clo, P = .55) or those with AML (Flu, P = .95; Clo, P = .06). The effect on TRM, graft failure, and cGVHD in ALL vs AML could not adequately be assessed due to the limited number of events. The cohort is, however, too small to consider all “disease risk index” and very high-risk cytogenetic changes. It cannot be excluded that exposure may have an impact on outcomes in certain subgroups. Notably, almost all patients for whom IR data were available (80/87 [92%]) reached the value of >50 CD4+ T cells per μL within 100 days. Because CD4 IR is an important driver for outcome after allo-HCT,16,19 the observation that all patients reach this threshold is in line with the lack of effect for Flu and Clo exposures on clinical outcome.
. | Clo . | Flu . | ||||
---|---|---|---|---|---|---|
P value univariable . | HR (95% CI) . | P value univariable . | HR (95% CI) . | P value multivariable . | HR (95% CI) . | |
OS | .20 | 1.32 (0.88-1.97) | .50 | 1.06 (0.90-1.25) | - | - |
EFS | .13 | 1.31 (0.94-1.83) | .37 | 1.07 (0.93-1.23) | - | - |
GRFS | .20 | 1.26 (0.88-1.79) | .58 | 1.04 (0.90-1.2) | - | - |
TRM | .98 | 1.01 (0.43-2.4) | .49 | 0.90 (0.65-1.24) | - | - |
Relapse | .43 | 1.17 (0.79-1.74) | .27 | 1.09 (0.94-1.26) | - | - |
Graft failure | .07 | 2.12 (1.00-4.49) | .21 | 1.27 (0.90-1.80) | - | - |
aGvHD | .86 | 1.05 (0.62-1.78) | .93 | 0.99 (0.84-1.17) | - | - |
cGvHD | .30 | 0.70 (0.36-1.37) | .18 | 0.72 (0.44-1.17) | - | - |
CD4 IR | .41 | 1.14 (0.84-1.55) | .03∗ | 1.14 (1.02-1.276) | .18 | 1.08 (0.96-1.22) |
. | Clo . | Flu . | ||||
---|---|---|---|---|---|---|
P value univariable . | HR (95% CI) . | P value univariable . | HR (95% CI) . | P value multivariable . | HR (95% CI) . | |
OS | .20 | 1.32 (0.88-1.97) | .50 | 1.06 (0.90-1.25) | - | - |
EFS | .13 | 1.31 (0.94-1.83) | .37 | 1.07 (0.93-1.23) | - | - |
GRFS | .20 | 1.26 (0.88-1.79) | .58 | 1.04 (0.90-1.2) | - | - |
TRM | .98 | 1.01 (0.43-2.4) | .49 | 0.90 (0.65-1.24) | - | - |
Relapse | .43 | 1.17 (0.79-1.74) | .27 | 1.09 (0.94-1.26) | - | - |
Graft failure | .07 | 2.12 (1.00-4.49) | .21 | 1.27 (0.90-1.80) | - | - |
aGvHD | .86 | 1.05 (0.62-1.78) | .93 | 0.99 (0.84-1.17) | - | - |
cGvHD | .30 | 0.70 (0.36-1.37) | .18 | 0.72 (0.44-1.17) | - | - |
CD4 IR | .41 | 1.14 (0.84-1.55) | .03∗ | 1.14 (1.02-1.276) | .18 | 1.08 (0.96-1.22) |
Cox proportional hazards and Fine and Gray competing risk models were used for the evaluation of time-to-event data. Factors were entered into multivariable models if P value < .05. In multivariable analyses, ATG exposure was considered as covariable for CD4 IR. CD4 IR was defined as >50 CD4+ T cells per μL within 100 days after allo-HCT. P values < .05 were considered statistically significant.
OS, overall survival; EFS, event-free survival; GRFS, GVHD-free relapse-free survival, TRM, therapy-related mortality; aGvHD, acute GvHD; cGvHD, chronic GvHD; CD4 IR, CD4 immune reconstitution; CI, confidence interval; HR, hazard ratio; ∗, P < .05.
Here, we show that Flu and Clo exposures do not affect outcome after allo-HCT. Although the Flu exposures were lower than previously reported in the FluBu regimen,9 this did not correlate with graft failure and TRM as previously observed,9 probably because of the addition of Clo to the conditioning. This indicates that individualized dosing is not necessary in children when using a cumulative Clo dosage of 120 mg/m2 together with a cumulative Flu dosage of 40 mg/m2, given that Bu is targeted at an AUCT0−∞ of 85-95 mg∗hour/L. In addition, individualized ATG dosing20 and normal renal function (estimated globular filtration rate 60-140 mL/min/1.73 m2) is necessary. It is important to emphasize that these results cannot directly be translated to other conditioning regimens or patient groups.
Although TBI-based conditioning is the European standard in allo-HCT for pediatric ALL,4 CloFluBu might be a good alternative for patients with ALL ineligible for TBI.6 In pediatric AML, chemotherapy-based conditioning yields better outcomes than TBI,21,22 with so far no difference between CloFluBu and Bu-cyclophosphamide-melphalan.23 This points toward the use of CloFluBu in pediatric AML, especially taking the toxic effects of TBI and alkylators into consideration.5,24 The data presented here add to the optimization of this TBI-free, single-alkylator conditioning regimen, showing that neither Clo nor Flu needs to be targeted before allo-HCT in this treatment protocol because this will not aid improved clinical outcome.
Acknowledgment: J.J.B. acknowledges support of the NIH/NCI Cancer Center Support Grant P30 CA008748.
Contribution: C.A.L., S.N., C.C.H.d.K., J.J.B., and L.D. were involved in the study design; L.D. performed statistical analyses and wrote the manuscript; K.C.M.v.d.E. provided the Flu and Clo concentrations and reviewed the manuscript; A.L.N. and A.D.R.H. calculated the total Flu and Clo exposures and reviewed the manuscript; R.A. provided the ATG exposures and reviewed the manuscript; and A.B.V., C.C.H.d.K., J.J.B., S.N., and C.A.L. reviewed the manuscript.
Conflict-of-interest disclosure: J.J.B. reports consulting for Sanofi, Sobi, Merck, Immunosoft, SmartImmune, and Advanced Clinical. The remaining authors declare no competing financial interests.
Correspondence: Stefan Nierkens, Princess Máxima Center for Pediatric Oncology, Heidelberglaan 25, 3584 CS Utrecht, The Netherlands; email: [email protected].
References
Author notes
C.A.L. and S.N. contributed equally to this study.
Data are available on request from the corresponding author, Stefan Nierkens ([email protected]).
The full-text version of this article contains a data supplement.