Role of cytarabine in paediatric acute promyelocytic leukemia treated with the combination of all-trans retinoic acid and arsenic trioxide: a randomized controlled trial

Eligible patients were those who were less than 14 years old, were newly diagnosed with APL, and had not previously received chemotherapy. A molecular diagnosis was not required for enrollment, but a subsequent molecular confirmation, including the demonstration of PML-RARA transcripts, was required for inclusion in the analysis. A genetic diagnosis was established by detecting the PML-RARA fusion gene using polymerase -chain -reaction (PCR) assays [19, 20] or by demonstrating t (15; 17) translocation using conventional karyotyping or fluorescence in situ hybridization (FISH) [21]. Written informed consent was obtained from all patients before study entry.

The study was a prospective, randomized, single-centre trial. It was designed to determine whether the combination of ATRA and ATO is safe and effective in paediatric APL and whether Ara-C can be omitted when ATO is added for 2 courses. Patients were assigned to receive ATRA plus ATO for induction followed by 1 consolidation course of idarubicin (IDA) and 1 consolidation course of a 28-day cycle of ATO. The patients were then randomly assigned using a computer-generated random allocation schedule to receive 2 courses of either daunorubicin (DNR) or DNR + Ara-C. Patients who were treated with DNR alone were included as the no-Ara-C group. Patients who were treated with DNR + Ara-C were included as the Ara-C group. The patients were subsequently treated with maintenance therapy consisting of oral ATRA, 6-mercaptopurine, and methotrexate for 1.5 years. When CR was achieved, all patients received a prophylactic intrathecal injection (cytarabine, methotrexate, and dexamethasone) for the first time. The patients with an initial white blood cell count > 10 × 10⁹/L then received intrathecal injection once every course. Patients with an indication of CNS leukaemia received intrathecal injection once every other day until normal results were achieved. The regimen is shown in Fig. 1. This trial was conducted in accordance with the Declaration of Helsinki and was retrospectively registered at Clinical- Trials.gov (identifier: NCT01191541).

All children were monitored using reverse transcription polymerase chain reaction (RT-PCR) of bone marrow samples [19]. To amplify the PML/RARa fusion gene, a two-step qualitative RT-PCR analysis was performed as previously described [19]. From January 2011, real-time quantitative PCR (RQ-PCR) was used to identify the PML/RARa fusion transcript [20]. In the RQ-PCR method, established in our laboratory based on cDNA, a dilution of the NB4 cell line reached a sensitivity of 1 × 10^− 5 for PML-RARa. Bone-marrow morphology, cytogenetics, and RT-PCR/RQ-PCR for PML-RARA were assessed after induction and each consolidation cycle. After consolidation, the patients were assessed every 3 months for 1 year and then every 6 months for 1 year. No pharmaceutical company was involved in the design of the study, data collection or analysis, or the writing of the manuscript.

Haematological complete remission (HCR) and haematologic relapse were defined as described in previous publications [1, 9]. Molecular remission was defined as undetectable PML/RARa fusion transcripts. Molecular relapse was defined as the detection of the fusion oncogene PML/RARa in multiple samples within 2 weeks in the same patient. Early death (ED) was considered a death that occurred within two weeks of the beginning of treatment.

The follow-up of the patients was updated in May 2017. The overall survival (OS) durations were calculated as the date of diagnosis to the date of last follow-up or death. Event-free survival (EFS) was defined as the time from diagnosis to the time at last follow-up or an event (i.e., relapse or death). Disease-free survival (DFS) was calculated as the time from the day HCR was achieved to the date of the last follow-up or an event (i.e., relapse). Death at any time and relapse were considered events for the EFS curve, while death in HCR and relapse were considered for DFS curves. Due to the open label character of the study, survival analysis was performed on an intention-to-treat (ITT) and a per-protocol (PP) basis.

Coagulopathy was treated using fresh frozen plasma or fibrinogen. Platelet transfusions were administered to maintain a platelet count above 50 × 10⁹/L until any significant sign of coagulopathy was resolved. The patients were administered hydroxyurea (1–1.5 g per day), or homoharringtonine (HHT, 1–2 mg per day for 5–10 days) when their peripheral white blood cell (WBC) counts were greater than 25 × 10⁹/L. At the earliest manifestation of suspected differentiation syndrome, ATRA, arsenic trioxide, or both were temporarily discontinued, and intravenous dexamethasone was administered at a dose of 5–10 mg/m² until these signs and symptoms disappeared. Antibiotics and antifungal drugs were administered for fever when required.

Forty-one patients and 11 healthy children as controls were included in the study to analyse the arsenic concentrations. All samples were collected on Oct 10, 2016. The arsenic concentration in collected plasma, urine, hair, and nail samples was determined using inductively coupled plasma mass spectrometry (ICP-MS). For each assay, 2 mL of plasma, 5 mL of urine or 0.1–0.5 g of nails or hair was collected. Plasma and urine specimens were stored at 4 °C and analysed within 2 weeks. Other specimens were collected in polypropylene tubes. An Agilent 7700× ICP-MS (Agilent technology, USA) equipped with a pure He octopole reaction system (ORS) was used for the total arsenic analysis. No polyatomic interference or argon chloride interference was observed while using this system. The ICP-MS instrument operating conditions are shown in Table 1. A 1.0 mL volume of blood (or urine) or 0.1 g of hair (or nails) was digested in 2 mL of HNO₃ (65%) and 1 mL of H₂O₂ (30%) in a microwave digestion system and then diluted to a total volume of 8 mL using deionized water. (Nitric acid (UP, China), BV-III grade). A blank digest was performed using the same method. All sample solutions were clear. The following digestion conditions were used for the microwave system: 5 min at 1300 W and 160 °C, 5 min at 1300 W and 200 °C, and 20 min at 1300 W and 200 °C. The digested samples were filled to the final volume using ultrapure water and then analysed using ICP-MS. A standard curve was generated for a linear range of 0 to 20 ng/ml and a detection limit of 0.01 μg/L.

The primary objective was to demonstrate the noninferiority of DNR alone compared to DNR + Ara-C in terms of the DFS rate at 2 years. Assuming a 95% rate of DFS in the two groups, a margin of − 14%, 5% type 1 error, and 80% power, 31 evaluable patients per group were required to draw a noninferiority conclusion.

The characteristics of all of the included patients were summarized using cross-tabulations (for categorical variables) and quantiles (e.g., the median; for continuous variables). Nonparametric tests were used to analyse comparisons between groups (i.e., χ² and Fisher’s exact tests for categorical variables). EFS, DFS and OS were estimated using the Kaplan -Meier method, and log-rank tests were used for comparisons. All P values were two-sided, and those with values of 0.05 or less were considered to be statistically significant. All statistical analyses were performed using SPSS 16.0 software.

Among the 66 patients, some had severe symptoms at presentation. These included intracranial bleeding in 4 (6.1%), intraocular bleeding in 6 (9.1%), and mild partial splenic embolization in 2 (3.0%). There were no significant difference in the rate of severe symptoms between the patients with WBC > 10 × 10⁹/L and WBC ≤10 × 10⁹/L (P = 0.225). No early deaths occurred. One patient ended treatment for economic reasons. A total of 65 patients were evaluated to determine their response to induction therapy. Haematologic complete remission was achieved in all of these patients.

During induction, hyperleukocytosis (> 10 × 10⁹/L) developed in 59 (90.8%) of the 65 patients with peak WBC counts ranging from 12.8 to 267.8 × 10⁹/L(median, 38.0 × 10⁹/L). In addition, 24 (36.9%) of the 65 patients exhibited an increase in peak WBC counts to more than 50 × 10⁹/L. HHT was used in 28 patients. The dosage of HHT was 1–2 mg/d, and it was administered for 2 to 15 days (median, 7 days). After CR was achieved, 11 (11/28, 39.3%) of the HHT-treated patients tested negative for PML-RARA fusion transcripts, whereas of the patients without HHT, 16 (16/37, 43.2%) tested negative for PML-RARA fusion transcripts. There was no significant difference in the proportion of patients who were negative for PML-RARA fusion transcripts between those who were treated with or without HHT (P = 0.749). There was also no significant difference in initial WBC, Hb, and PLT counts and outcomes between the two groups.

During induction therapy, retinoic acid syndrome (RAS) was diagnosed in 9 (13.8%) patients, but it did not contribute to any deaths. Four (6.2%) of the 65 patients suffered Common Terminology Criteria for Adverse Events (CTCAE V.4.0) grade 1–2 hepatotoxicity. Other ATO-associated adverse reactions included extremity oedema in 9 (13.8%) cases, nausea in 2 (3.1%) cases, skin pigmentation in 2 (3.1%) cases, bone ache in 2 (3.1%) cases, cardiac arrhythmia in 1 (1.5%) case and asymptomatic QTc prolongation on electrocardiography in 2 (3.1%) cases. Additional ATRA-associated adverse reactions included headache in 24 (36.9%) cases, skin rash in 3 (4.6%) cases, nausea in 7 (10.8%) cases, abdominal pain in 2 (3.1%) cases, bone ache in 8 (12.3%) cases and skin desquamation in 4 (6.2%) cases. All toxicity events were reversed by appropriate management.

Consolidation therapy was administered in all patients except for the patient who ended therapy early. No patient died after CR was achieved. Side-effects included sepsis in 6 (9.2%) cases, and hepatotoxicity in 3 (4.6%) cases. No secondary malignancies have so far been reported in our patients.

According to the results of our regimens, which were applied to 2 groups using random selection, 31 patients were included in the Ara-C group, and 34 patients were included in the no-Ara-C group. One patient in the Ara-C group voluntarily transferred to the no-Ara-C group before the random treatment was administered. After the first course of random treatment was administered, three patients in the Ara-C group voluntarily transferred to the no-Ara-C group due to haematologic toxicity. Finally, 30 patients were included in the Ara-C group with 57 courses of treatment and 35 patients in the no-Ara-C group with 73 courses of treatment. There were no significant differences in baseline characteristics between the Ara-C and no-Ara-C groups on a PP basis analysis (Table 2). Also, there was no significant difference (P ≥ 0.05) between the two groups on an ITT basis analysis in baseline characteristics (data not shown). In addition, there was no difference in EFS, DFS and OS between the two groups on an ITT and a PP basis analysis. Based on the actual application of the treatment, we compared the hematology toxicity between the two groups. The percentages of courses that included platelet and red blood cell (RBC) transfusions in the Ara-C group were 91.2% (52/57) and 24.6% (14/57), respectively. During consolidation, no blood product was required in the no-Ara-C group. A total of 84.2% (48/57) and 5.5% (4/73) of the patients in the Ara-C and no-Ara-C groups, respectively, had WBC counts <1.0 × 10⁹/L (P = 0.000). In the Ara-C group, the median lowest WBC count was 0.62 × 10⁹/L (range, 0.02 to 1.82 × 10⁹/L). The median days of neutropenia was 0 day (range, 0 to 9 days) in the no-Ara-C group and 6 days (range, 0 to 13 days) in the Ara-C group, respectively (P = 0.000). There were 6 cases of sepsis, including five in the Ara-C group and one in the no-Ara-C group. No deaths occurred during consolidation therapy.

Twenty-seven (41.5%) of the 65 patients who were tested after induction were negative for PML-RARA fusion transcripts. After the first consolidation cycle, 58 (89.2%) of the 65 patients were negative, and after the second ATO treatment cycle, 64 (98.5%) of the 65 patients tested negative. After the third consolidation cycle (i.e. the first cycle of DNR/DA), a complete remission (molecular) was achieved in all patients (Fig. 1). There were no significant differences in baseline characteristics between the cohorts that were positive or negative after induction and IDA chemotherapy.

Of the 65 patients who entered haematologic CR, the median follow-up time was 36 months (range, 5 to 83 months). Only one patient in the no Ara-C group relapsed. After a median follow-up of 36 months, the EFS was 97.3 ± 2.7%, and the OS was 100%. No factor impacted relapse or survival in our study.

Arsenic concentrations were assayed in plasma, urine, hair, and nail samples during and after the cessation of arsenic treatment in forty-one patients. Eleven healthy children were used as the control group. In our patients, 5 ceased ATO treatment after fewer than 3 months, 9 ceased ATO treatment after 3–12 months, 7 ceased ATO treatment after 12–24 months, and 20 ceased ATO treatment after more than 24 months.

Figure 2 shows the arsenic concentrations in the plasma, urine, hair, and nail samples at different time points and compared to the control group. Patients who had been off of the arsenic-containing treatment for less than 3 months had higher arsenic concentrations in the plasma, urine, hair and nail samples than the patients in the control group. The arsenic levels in the nails obtained from patients who had ceased treatment for 3–12 months (median, 264.2 ng/g; range, 117–24,240 ng/g) were higher than the levels in the controls (median, 198.8 ng/g; range, 33.6–588.1 ng/g). But statistical analyses revealed no significant difference between the two groups (P = 0.215). There was no difference in the median arsenic concentrations in plasma, urine, hair, and nail samples between patients in whom arsenic treatment had been ceased for more than 12 months and normal controls.