EXPERIMENTAL AND CLINICAL TRANSPLANTATION

Objectives: Allogeneic hematopoietic stem cell transplantation has been used for several decades to treat patients with acute lymphoblastic leukemia. Total body irradiation has been promoted as an important component of conditioning regimens for this process; however, recent reports of chemotherapy-based conditioning regimens have shown comparable outcomes.

Materials and Methods: We report our experience with radiation-free conditioning using busulfan and cyclophosphamide in 127 pediatric patients with acute lymphoblastic leukemia who were treated between 1997 and 2014. The median age was 11 years (range, < 1 to 15 y), 70% of patients were male, 81.1% received transplants from HLA-matched siblings, 83% received peripheral blood stem cells, 41% were in second complete remission at the time of transplant, and 83% had B-lineage immunophenotype.

Results: In patients who were in complete remission at the time of transplant, 5-year overall survival, leukemia-free survival, and relapse rates were 62.48% (95% confidence interval, 52.29-71.09%), 49.43% (95% confidence interval, 39.57-58.53%), and 45.64% (95% confidence interval, 35.85-54.88%), respectively. We observed significant differences between outcomes in patients by time of transplant, presence of chronic graft-versus-host disease, and remission status.

Conclusions: Our relapse rates were comparable to those shown in recent studies, although the transplant-related mortality rate was lower. The results of our study showed that a busulfan/cyclophosphamide conditioning regimen has acceptable outcomes without the undesirable adverse effects of total body irradiation, particularly in pediatric patients. Large multicenter studies are needed to assess less toxic conditioning regimens with fewer adverse effects in these patients.

Key words : Busulfan, Cyclophosphamide, Graft-versus-host disease, Remission status

Allogeneic hematopoietic stem cell transplantation (HSCT) has been promoted as a means of therapy in pediatric patients with relapsed, refractory, or high-risk acute lymphoblastic leukemia (ALL).^1,2

After presentation of the seminal work of Davies and associates3 in 2000, total body irradiation (TBI) has been accepted as the main component of preparative regimens for HSCT in patients with acute leukemia throughout most transplant centers. Although regimens containing TBI have been for some time considered to be superior due to better leukemia-free survival (LFS),4 the introduction of conditioning regimens using busulfan and cyclophosphamide for leukemia5 has led to less emphasis on TBI as the conditioning regimen for patients with acute myeloid leukemia.^6-9 Similar studies have concluded that TBI-free conditioning regimens for chronic myeloid leukemia are equal or superior to conditioning regimens that include TBI.^9-11 This overall trend has not yet transformed the treatment of pediatric ALL, and most investigators still emphasize the role of TBI in conditioning regimens for HSCT in ALL; however, some studies have explored the feasibility of conditioning regimens that exclude TBI for both adult and pediatric patients with ALL.^12-19

Total body irradiation is associated with a wide range of adverse effects, including cardiac, pulmonary, and renal complications; cataract; and increased risk of developing new malignancies and endocrinopathies. In pediatric patients, cognitive impairments and development abnormalities are more concerning.^20-27

There have been contradicting results for and against TBI-free conditioning regimens in pediatric patients with ALL.^16,28,29 Due to the lack of necessary equipment for TBI, our center has been exclusively performing HSCT using TBI-free conditioning regimens since the beginning of its activity.³⁰ Although this fact was initially recognized as a shortcoming of HSCT for our patients, today, we can present it as an opportunity. Our results could be included in the ongoing debate on the desirability of TBI-free conditioning regimens for HSCT in pediatric ALL.

Materials and Methods

Patients and data collection
All patients who had undergone HSCT from May 1997 until February 2014 and were 15 years old or younger were included in this study.

To determine the availability of a human leukocyte antigen (HLA)-identical related donor before the scheduled HSCT, histocompatibility typing procedures of patients and donors were done by low-resolution molecular typing for HLA-A and HLA-B antigens and allelic typing for DRB1 (for patients receiving HSCT from siblings). High-resolution HLA typing was applied when other related donors or unrelated donors were considered, which included HLA-DR and HLA-C.

The institutional review board and the ethics committee of the Hematology-Oncology and Stem Cell Transplantation Research Center (Tehran, Iran) reviewed and approved the study design. Written informed consent was obtained before HSCT from legal guardians of the patients.

Transplant process
The day of HSCT was used as a reference point (day 0). The conditioning regimen consisted of oral busulfan at 4 mg/kg/day (4 times per day) from day -7 for 4 consecutive days or intravenous busulfan (doses of 1 mg/kg for patients weighing 9 kg and less, 1.2 mg/kg for those weighing 9-16 kg, 1.1 mg/kg for those weighing 16-23 kg, 0.95 mg/kg for those weighing 23-34 kg, and 0.8 mg/kg for those weighing more than 34 kg) at 4 times per day from day -7 for 4 consecutive days. This was followed by once daily intravenous cyclophosphamide (60 mg/kg/day) on days -3 and -2. Up to 2010, patients received only oral busulfan; recently, we have been able to administer intravenous busulfan (Busilvex; Pierre Fabre Medicament, Boulogne, France) to our patients. All patients had received an identical regimen during the course of the study. Recipients of HSCT from other related, unrelated, or HLA-mismatched donors were additionally treated with 2.5 mg/kg/day of rabbit antithymocyte globulin (Thymoglobulin, Sanofi-Aventis, Bridgewater, NJ, USA) for 2 days (for HSCT from matched other related or unrelated donors) or 3 days (for HSCT from HLA-mismatched donors) starting on day -3. Marrow harvests were obtained from iliac crests of donors under general anesthesia. Peripheral blood stem cells were mobilized from donors with granulocyte colony-stimulating factor. For graft-versus-host disease (GVHD) prophylaxis, patients received daily intravenous cyclosporine (initiated at 1.5 mg/kg from day -2), which was then increased to 3 mg/kg/day starting on day +7 in peripheral blood stem cells or on day +11 in bone marrow or cord blood stem cell transplant. A short course of methotrexate (10 mg/m2 on day +1 and 6 mg/m² on days +3, +6, and +11) was also administered, except for patients who received HSCT from cord blood.

Supportive care
All patients were hospitalized under strict isolation in rooms with high-efficiency particulate arresting filters. The same supportive care was provided for all patients, and nutritional supplementation was introduced by hyperalimentation. Prophylaxis for busulfan-mediated seizures was provided in the form of phenytoin. Monitoring for cytomegalovirus infection was performed by screening for cytomegalovirus DNA polymerase chain reaction or cytomegalovirus PP65 antigen twice per week. Screening for fungal infections was performed once per week by checking for galactomannan antigen. Antimicrobial prophylaxis with acyclovir, fluconazole, and sulfamethoxazole/trimethoprim was provided against viral, fungal, and Pneumocystis jiroveci infections. In the event that a blood product transfusion was needed, irradiated products were used. The level of cyclosporine was measured twice per week. Levels between 100 and 250 ng/dL were in the therapeutic range.

Follow-up
The short tandem repeat method or fluorescence in situ hybridization was conducted on days 15 and 30 post-HSCT on bone marrow and on days 60, 90, and 180 post-HSCT on peripheral blood samples. These tests were done to monitor hematopoietic chimerism. Thereafter, patients were evaluated at months 12, 18, and 24 after HSCT and eventually every year. After discharge, patients were followed up in our post-HSCT clinic at weekly intervals during the first 60 days, which was reduced to every 2 weeks by day 100 after transplant. Eventually, follow-up would be conducted per the needs of each patient.

Definitions
Myeloid engraftment was defined as at least 3 consecutive daily absolute neutrophilic counts of ≥ 500/μL without granulocyte colony-stimulating factor administration and 7 consecutive daily platelet counts of ≥ 20 000/μL without transfusion.³¹

Overall survival (OS) was measured from the time of HSCT to time of death or to last contact in surviving patients. Leukemia-free survival was defined as the time between HSCT and relapse, death, or last contact in patients without relapse or death event. Relapse rate was defined as the probability of having a relapse at a given time; death without a relapse was considered a competing event.

Acute and chronic GVHD were defined according to consensus criteria.^32,33Transplant-related mortality (TRM) was defined as death from causes other than relapse or progression of the underlying medical condition.

Remission status was determined within 2 weeks before HSCT by analysis of bone marrow aspirations and biopsies, cytogenetic studies, and cerebrospinal fluid analyses. Patients were in complete remission (CR) if there were no lymphoblasts in the cerebrospinal fluid or peripheral blood and fewer than 5% blasts in bone marrow samples. Partial remission was defined as having no lymphoblasts in the circulation but ≥ 5% and ≤ 25% lymphoblasts in a bone marrow aspirate during the 2-week interval before HSCT.34 Relapse was defined as recurrence of blasts or leukemic infiltrates in bone marrow or extramedullary sites.³⁵

Statistical analyses
Continuous variables are presented as mean values and standard deviations, and categorical variables are shown as frequencies and percentages.

Overall survival and LFS were calculated using the Kaplan-Meier method and were compared between the levels of categorical variables using log-rank test and accompanied by relevant 95% confidence intervals.

Univariate and multivariate Cox proportional hazards models (with stepwise exclusion and inclusion of factors, using a criteria of P < .05 for retention and P > .1 for deletion of factors in the model) were used to identify factors that were independently associated with OS and LFS. We applied a stepwise selection method due to our small sample size. The effects of predictors on OS and LFS were described as hazard ratios (HR) with 95% confidence intervals.

All analyses were done using STATA version 11.2 (StataCorp, College Station, TX, USA) and R language and environment software (R package “cmprsk”).^36,37 P < .05 was considered as significant in all statistical analyses.

Patient characteristics
This was a retrospective study of 127 pediatric ALL patients (90 boys and 37 girls) who received HSCT; the median age was 11 years (range, < 1 to 15 y). Patients received treatment at our center between 1997 and 2014, with most receiving transplants from HLA-matched siblings. At diagnosis, all eligible patients were treated with a chemotherapy regimen adapted from the ALL IC-Berlin-Frankfurt-Munster trial protocol. Except for 6 patients, all other patients showed morphologic CR (< 5% bone marrow blast) at the time of HSCT. These 6 patients were not excluded from the analysis. The characteristics of patients and donors are provided in Table 1.

The median follow-up time was 59.89 months (range, 13.5-214.4 mo). Although both 5-year LFS and OS were significantly different among patients transplanted at first, second, or more advanced CR (P < .001), no patients who were transplanted at partial remission survived more than 7 months after HSCT (4 had relapse-related mortality and 2 had TRM). In patients who had CR at the time of transplant, 5-year OS and LFS rates were 62.48% and 49.43%, respectively. Five-year OS was significantly different between patients who had received HSCT before 2007 (45.71%) and those who underwent HSCT in 2007 or later (66.47%; P = .03).

Impact of donor source
Most patients (83.46%) received hematopoietic stem cells from peripheral blood. Distribution of different hematopoietic stem cell sources among other subcategories was uneven. Furthermore, we were not able to calculate some variables in survival statistics for patients in the cord blood subgroup. A combination of the above hindered meaningful analysis of these results; however, we found that the source of hematopoietic stem cells did not influence overall outcomes.

Impact of graft-versus-host disease
Of total patients, 86 (67.72%) developed acute GVHD, with 64 patients (74.4%) having grade I/II and 22 (25.6%) having grade III/IV acute GVHD. No significant differences in incidence or severity of acute GVHD were observed regarding the source of HSCT, relation to donor, and degree of HLA disparity.

Chronic GVHD occurred in 49 patients (38.58%), with 37 patients (75.5%) developing limited and 12 patients (24.5%) developing extensive chronic GVHD. The cumulative incidence of chronic GVHD was significantly different among patients in various relation subgroups (P = .001). Patients who experienced chronic GVHD had a statistically higher 5-year OS rate (67.61%) than those without chronic GVHD (55.16%; P = .02) (Table 2).

No differences in LFS or OS were identified between patients with and without high grades of both acute and chronic GVHD.

Transplant-related mortality
We observed significant differences in TRM among patients transplanted at first, second, or more advanced CR (P < .001); the main cause of death was relapse (39 deaths, 83.0%).

Multivariate analyses
Due to skewness of the data, only patients who had received HSCTs from an HLA-matched related donor were included in the regression models. Based on our regression model, occurrence of chronic GVHD (HR = 0.44), being female (HR = 0.31), and being at second remission at the time of HSCT (HR = 0.47) predicted an increase in OS. With regard to LFS, chronic GVHD (HR = 0.36) and being female (HR = 0.39) predicted more favorable outcomes (Table 3 and Table 4).

The tailoring of conditioning regimens is vital to improving outcomes of pediatric patients with ALL who receive HSCTs. Up until recently, there was a common consensus that TBI-based conditioning regimens were superior to TBI-free regimens.^14-16,38 An ongoing international study with the aim of assessing outcomes of TBI-free conditioning regimens for pediatric ALL HSCT is challenging this common perception.³⁹ Although our center has been constrained by lack of equipment necessary for TBI, it is interesting to note that this trend has changed in more recent years.^8-10

Although some previous studies have not found any differences in outcomes of HSCT among patients transplanted at first CR, second CR, or more advanced CR,^40,41 other studies have demonstrated significant differences.^42,43 In our study, we also found LFS, relapse rate, and TRM to be significantly different among patients transplanted at first, second, or more advanced CR. We also found that patients who were transplanted at partial remission did not benefit from HSCT, as none survived beyond the first year after HSCT. These findings should be weighed against another HSCT study on relapse or induction failure in patients with acute leukemia43 who received TBI and stem cells from unrelated donors. The 3-year OS for patients with ALL who had undergone HSCT and were not in CR was 16%.

We found that patients who developed chronic GVHD had significantly better 5-year OS, LFS, and relapse rates. These patients also had lower TRM, although differences were not statistically significant. Our findings have been corroborated by publications by Zikos and associates43 and Thepot and associates.⁴² Our regression models verified that chronic GVHD is a predictor of higher OS and disease-free survival. However, we did not find any differences in outcomes by grade of acute GVHD, and no association was found between outcomes and stem cell source. In reports from MacMillan and associates44 and Garderet and associates,⁴⁵ HSCT from peripheral blood samples was suggested to be not as effective as bone marrow. However, Madero and associates46 and Vicent and associates⁴⁷ did not report such a difference between these 2 sources. Another study on HSCT from unrelated cord blood in pediatric ALL suggested that using TBI as a preparative regimen was associated with lower relapse rates; because most patients in our study received HSCT from HLA-matched peripheral blood sources, a comparison was not meaningful.

Our analyses showed that 5-year OS and relapse rates were considerably superior in patients who had received HSCT more recently (between 2006 and 2014), since an independent pediatric ward was created in our center. As staff and group members gain more experience with pediatric HSCT, we expect even better outcomes over time. An additional factor may be improvements in supportive care and closer monitoring of patients.

Overall, although the 5-year relapse rate in our patients was slightly more than some other studies,^48,49 the 5-year LFS and TRM rates were comparable.^50-55 Although differences between busulfan-cyclophosphamide and TBI-cyclophosphamide regimens are mainly the higher TRM with the former, various undesirable adverse effects of TBI in children may make the busulfan-cyclophosphamide regimen more desirable. Because our study showed higher relapse rates, we were not able to determine the significance of this difference without controlling for some confounders. A closer examination of all existing data in this field is required. Of note, on the basis of our unpublished findings, there were no significant differences in OS and LFS between oral and intravenous busulfan.

Some studies have shown that TBI-based conditioning regimens may result in more frequent late-onset complications such as metabolic syndrome, premature arterial vascular disease, poor cardiac function, endocrinological dysfunction, and secondary malignancies^.30,31,48 A busulfan-based conditioning regimen as a less toxic myeloablative conditioning alternative to TBI has been suggested by other reports.^{6,21,22,48,49,56}

We found that, although incidence of relapse rate was higher among our patients, OS and LFS were comparable with other studies. We suggest that further investigations on the short-term and long-term efficacy and toxicity of busulfan-cyclophosphamide conditioning regimens, without TBI, in pediatric ALL HSCT and 2-arm multicenter studies are warranted.

Volume: 17 Issue: 2 April 2019

FULL TEXT