Randomized Trial and Pharmacokinetic Study of Pegfilgrastim versus Filgrastim after Dose-Intensive Chemotherapy in Young Adults and Children with Sarcomas Free

Pegfilgrastim (filgrastim-SD/01, Neulasta), which is a sustained-duration formulation of filgrastim, is Food and Drug Administration approved at a fixed dose of 6 mg s.c. injection once per treatment cycle for adults and children >45 kg to shorten the duration of chemotherapy-induced neutropenia. Pegfilgrastim has a 20-kDa polyethylene glycol molecule covalently bound to the NH₂-terminal methionyl residue of filgrastim. Polyethylene glycol is inert, has low antigenicity, has minimal toxicity, and alters the clearance of the agent. Filgrastim (clearance, 40 mL/h/kg) is eliminated primarily by renal excretion, whereas pegfilgrastim elimination (clearance, 14 mL/h/kg) is dependent on receptor binding to neutrophils (1). As a result, pegfilgrastim is self-regulating because it persists in the body during neutropenia but is cleared rapidly when the neutrophil count recovers (2, 3).

The safety and efficacy of pegfilgrastim has been shown in large randomized trials in adults with nonmyeloid cancers that were treated with myelosuppressive chemotherapy (1, 4–7). Incidence and duration of neutropenia and side effects were similar to daily filgrastim. The pharmacokinetic profile of pegfilgrastim in adults is nonlinear (8). In 379 adults with cancer, the pegfilgrastim apparent clearance was highly variable and dependent on dose and neutrophil count (2). The terminal half-life ranges from 46 to 62 hours in normal volunteers.

Primary prophylaxis with myeloid growth factors is frequently used in children and adolescents receiving dose-intensive chemotherapy (9–11). In children, several small, uncontrolled studies and case series describing the use of pegfilgrastim have been reported (12–15). Spunt et al. (16) have reported preliminary results from a trial conducted in children ages <21 years with sarcoma randomized to receive pegfilgrastim (n = 14) or filgrastim (n = 2) after dose-intensive chemotherapy. Preliminary data suggest that once per cycle dosing with pegfilgrastim (100 mcg/kg) resulted in similar severity and duration of postchemotherapy neutropenia to daily filgrastim (5 mcg/kg).

We conducted a trial of pegfilgrastim in children and young adults with newly diagnosed sarcoma. Patients were randomized (1:1) at the time of study enrollment to receive pegfilgrastim or filgrastim after multiagent, dose-intensive chemotherapy to determine if the time to neutrophil recovery was equivalent between the two arms. The study was also designed to assess the tolerance and toxicity of pegfilgrastim, compare the pharmacokinetics of pegfilgrastim and filgrastim, and compare CD34+ stem cell mobilization between the two arms.

This trial was done under an Investigator IND. Pegfilgrastim and filgrastim were manufactured and provided by Amgen. Pegfilgrastim was provided in single-use, preservative-free vials as a sterile 1.2-mL solution at concentration of 10 mg/mL. This pegfilgrastim formulation allowed for individualized dosing based on body weight.

The chemotherapy treatment plan is described in Table 1. Patients were randomized at study entry to receive pegfilgrastim (100 mcg/kg s.c. as a single dose 24 to 36 h after completion of each cycle of chemotherapy) or filgrastim (5 mcg/kg/d s.c., daily starting 24 h and continuing until the postnadir neutrophil count was ≥10,000/mcL after each cycle of chemotherapy). Randomization was done centrally and was not stratified for age, diagnosis, or other baseline characteristics. For each patient, the treatment assignment (pegfilgrastim versus filgrastim) was maintained throughout all treatment courses. Supportive care measures also included mesna as a uroprotectant with cyclophosphamide and ifosfamide, and dexrazoxane as a cardioprotectant before doxorubicin. Surgery or radiation directed against the primary tumor commenced after cycle 5. Fourteen cycles [six vincristine, doxorubicin, and cyclophosphamide (VDC) and eight etoposide and ifosfamide (IE)] were planned.

If a treatment cycle was delayed for >7 d to allow for recovery from myelosuppression caused by the prior cycle's treatment, subsequent doses of the myelotoxic drugs that caused the prolonged myelosuppression were reduced by 25% on subsequent cycles of the same type of therapy. Dose modifications were not based on the nadir of the neutrophil or platelet counts or on the duration of grade 4 neutropenia and thrombocytopenia, unless it delayed the next treatment cycle by >7 d.

Patients ages <26 y with Ewing sarcoma family of tumors, alveolar rhabdomyosarcoma, stage 3 or 4 embryonal rhabdomyosarcoma, and unresectable or metastatic malignant peripheral nerve sheath tumor or synovial sarcoma were eligible. Normal cardiac (left ventricular ejection fraction) and renal function, absolute neutrophil count (ANC) of ≥1,500/mcL, hemoglobin of ≥9 g/dL, and platelet count of ≥100,000/mcL were required. Patients who had received prior chemotherapy or radiation therapy, patients with tumor infiltration of bone marrow determined by biopsies, and patients who were pregnant or breastfeeding were excluded. All patients or their parent or guardian signed an informed consent document. Assent of minors was obtained according to institutional guidelines. The study was approved by Institutional Review Boards at the participating sites.

The primary objective of this trial was to determine whether the duration of chemotherapy-induced severe neutropenia (ANC, <500/mcL) was noninferior for the pegfilgrastim arm compared with the filgrastim arm for the first four cycles of chemotherapy (no more than one SD longer for pegfilgrastim). The sample size was estimated based on standard methods for a two group t test of equivalence of mean with equal SDs and sample sizes per arm, but controlling for two comparisons as noted below. Specifically, separately for two cycles of vincristine (three weekly doses), doxorubicin, cyclophosphamide (V₃DC) averaged together and two IE cycles averaged together, when the sample size in each arm is 17, for each set of averaged cycles, a two-group 0.025 one-sided t test will have 80% power to reject the null hypothesis of inferiority (that the average duration of severe neutropenia on the pegfilgrastim arm was at least one SD longer than that on the filgrastim arm) in favor of an alternative that the mean of the pegfilgrastim arm is not inferior to that of filgrastim, that is, that the average durations of neutropenia on the two arms do not differ by >1.0 SD (which is presumed to be equal in both arms). Based on data from previous institutional experience using these chemotherapy regimens in this population, the variance in the duration of severe neutropenia was 2 d (17, 18). Although this determination was based on an assumption that the durations of severe neutropenia would be normally distributed, this was not the case for the IE cycles. Thus, for simplicity, a Wilcoxon rank-sum test was used to test if the duration of neutropenia differed significantly (overall P < 0.05) between the arms when the durations from the two V₃DC cycles and the two IE cycles were averaged and tested separately. Differences in toxicity, pharmacokinetic parameters, CD34 stem cell mobilization, and days of febrile neutropenia, were also compared by the Wilcoxon rank-sum test. Differences in duration of neutropenia between the V₃DC and IE cycles were tested for statistical significance by a Wilcoxon signed-rank test, separately for the two treatment arms. All P values are two tailed and are presented without adjustment for multiple comparisons.

Severe neutropenia was defined as an ANC of <500/mcL and recovery was defined as a postnadir ANC ≥500/mcL. The primary end point was determined during the first four treatment cycles (two V₃DC cycles followed by two IE cycles) to avoid potential confounding effects of radiation therapy on bone marrow function. Complete blood counts were done at least thrice week during cycles 1 to 4 and at least twice weekly during cycles 5 to 14. Cycle duration was calculated for each cycle as the number of days from day 1 of the cycle to day 1 of the subsequent cycle. Cycles commenced at 21-d intervals if the patient had recovered from treatment-related toxicity from the prior cycle. For nonhematologic toxicities, recovery was defined as National Cancer Institute Common Toxicity Criteria (v.2) grade ≤1; and for hematologic toxicity, the ANC had to be ≥500/mcL and the platelet count had to be ≥75,000/mcL. Criteria for the administration of RBC and platelet transfusions were not specified in the protocol. The use of other growth factors (granulocyte macrophage colony-stimulating factor, erythropoietin, interleukin-11) was prohibited.

Toxicity was graded using the National Cancer Institute Common Toxicity Criteria v.2.⁴

Serum samples were obtained before and 1, 2, 4, 6, 8, 24, and 28 h after first dose of filgrastim and pegfilgrastim in all patients. For pegfilgrastim, samples were also obtained at 96, 144, 192, 240, and 504 h. For filgrastim, additional trough samples were obtained on days 4, 6, 8, 10, and 4 h after the day 10 dose. Serum granulocyte colony-stimulating factor concentrations were measured using an ELISA (R&D Systems; ref. 19). The lower limit of quantification was 0.04 ng/mL. Pharmacokinetic parameters were calculated using noncompartmental methods.⁵

Quantification of CD34+ stem cells was done using flow cytometry of EDTA whole blood drawn on the day that the postnadir neutrophil count first exceeded 500/mcL during cycle 1.

Binding antibody immunoassay (Biacore 3000, BIAcore AB) was used to detect antibodies capable of binding to pegfilgrastim in patients' serum at baseline and before each cycle (postexposure) of therapy by exposure of patient serum samples to biosensor with drug covantely immobilized to a surface. Antibodies were detected by accumulation of mass on the surface and confirmed by anti-immunoglobulin reagent. Postexposure response in the binding antibody immunoassay of ≥2-fold higher than baseline was considered positive. In samples positive for binding antibody, a cell-based neutralizing antibody assay was used to determine if serum contained antibodies capable of neutralizing pegfilgrastim. Using a cell line that required granulocyte growth factor and interleukin-3 for growth, neutralizing antibody was measured in patient's serum when cell growth was inhibited in the presence of patients serum.

Thirty-four total patients were enrolled, 29 at the Pediatric Oncology Branch, National Cancer Institute and 5 at Seattle Children's Hospital, Seattle, WA between December 2000 and December 2005. Patient characteristics are listed in Table 2. Two patients in the pegfilgrastim arm did not complete the initial four cycles of chemotherapy. Neither patient discontinued therapy due to adverse events. One patient discontinued therapy after cycle 2 due to progressive tumor, and one withdrew consent after cycle 1 to transfer care to another institution.

For patients on the filgrastim arm, the median (range) number of daily doses of filgrastim was 13 (7-27) per cycle for the two V₃DC cycles and 10 (6-24) per cycle for the two IE cycles. Two children enrolled on the study were ages <5 years, both randomized to filgrastim.

The duration of severe neutropenia was significantly longer after V₃DC (cycles 1 and 2) than after IE (cycles 3 and 4) for both the pegfilgrastim (P < 0.001) and filgrastim (P < 0.001) treatment arms (Fig. 1). Six of 15 patients on the pegfilgrastim arm and 5 of 14 patients on the filgrastim arm did not have an ANC of <500/mcL documented during their initial two IE cycles.

The duration of severe neutropenia (ANC, <500/mcL) for the mean of cycles 1 and 2 of V₃DC and the mean of cycles 3 and 4 of IE are presented in Fig. 1 by treatment arm (pegfilgrastim versus filgrastim). The duration of neutropenia could be definitively determined on 28 of the 34 enrolled patients during the two initial V₃DC cycles and on 29 patients during the two initial IE cycles. The inevaluable patients either did not complete the first four cycles of treatment (n = 2) or did not have blood counts performed frequently enough to accurately determine the duration of severe neutropenia (n = 4). During the first four cycles, the number of days of severe neutropenia was not significantly different between the two treatment arms for either V₃DC cycles (median of 5.5 and range of 3-8 for pegfilgrastim versus median of 6.0 and range of 0-9 for filgrastim; P = 0.76) or IE cycles (median of 1.5 and range of 0-4 for pegfilgrastim versus median of 3.75 and range of 0-6.5 for filgrastim; P = 0.11).

The cycle duration for all delivered cycles of therapy is presented in Table 2 and was a median of 21 days for both treatment arms for vincristine, doxorubicin, and cyclophosphamide cycles and for IE cycles. No patient required a dose reduction due to delayed recovery of blood counts. The majority of cycle delays were due to scheduling or complications of surgery or radiation on cycles 5 to 7.

The ANC profile after pegfilgrastim, which was characterized by higher prenadir than postnadir ANC peaks, differed from that observed after filgrastim, which resulted in higher postnadir ANC peaks (Fig. 2). During cycles 1 and 2 and cycles 3 and 4, there were statistically significant differences in prenadir and postnadir peak ANCs on the treatment arms. The median (range) prenadir ANC peak was 20,100/mcL (2,300-94,900/mcL) in the pegfilgrastim group and 10,700/mcL (1,400-39,400/mcL) in the filgrastim group (P = 0.024); the median (range) postnadir ANC peak was 8,000/mcL (2,400-28,200/mcL) in the pegfilgrastim arm and 20,400/mcL (2,200-47,400/mcL) in the filgrastim arm (P < 0.001).

Twelve of 17 patients on the pegfilgrastim arm experienced 18 episodes (29% of cycles) of grade 3 fever and neutropenia requiring hospitalization during the first four cycles of chemotherapy compared with 15 of 17 patients and 32 episodes (47% of cycles) on the filgrastim arm. There were four documented infections on the pegfilgrastim arm, including one episode each of S. aureus bacteremia, S. aureus skin infection, P. jiroveci pneumonia, and Enterobacter sp. urinary tract infection. Eight documented infections occurred on the filgrastim arm, including one Escherichia coli bacteremia, one S. epidermitis bacteremia, three episodes of mucocutaneous candidiasis, one group A streptococcal pharyngitis, one Enterbacter sp. urinary tract infection, and one episode of pityrosporum folliculitis.

Mobilization of CD34+ stem cells did not differ between the two arms (P = 0.97). The median (range) CD34+ cell count at the time of neutrophil recovery on cycle 1 (V₃DC) was 165/mcL (10-739/mcL) on the pegfilgrastim arm and 53/mcL (35-865/mcL) on the filgrastim arm.

No neutralizing antibody to pegfilgrastim was detected in 16 patients (163 cycles) who were enrolled on the pegfilgrastim arm.

Pegfilgrastim and filgrastim were well tolerated in this patient population. Adverse events associated with growth factor administration on the first four cycles of chemotherapy were similar on the two treatment arms (Table 3). No dose modifications to growth factor therapy were required because of adverse events on either treatment arm. Twenty months after completion of chemotherapy, one patient on the pegfilgrastim arm developed acute leukemia with t(3;21), +21, −8, and del 7 chromosomal abnormalities.

The mean (n = 17) plasma concentration time profile of pegfilgrastim is shown in Fig. 2A. The serum concentration peaked 2 hours after the dose and then declined in concert with the prenadir ANC peak. A secondary pegfilgrastim peak was observed after day 7 when the ANC approached its nadir. The serum pharmacokinetic parameters for pegfilgrastim and filgrastim are reported in Table 4. The terminal half-life of pegfilgrastim derived from the mean serum concentrations (n = 17) was 49 hours. Absorption (T_max) and apparent clearance (CL/F) were significantly different (P < 0.001) in the pegfilgrastim compared with filgrastim arm. Substantial intersubject variability was observed with both growth factors.

The duration of severe neutropenia after dose-intensive, combination chemotherapy for pediatric and young adult patients with sarcoma was similar for a single dose of pegfilgrastim (100 mcg/kg s.c.) and daily filgrastim injections (5 mcg/kg/d s.c.) in this relatively small study. This trial was designed to show that pegfilgrastim is not inferior to filgrastim, but, in fact, the parameters measured to assess the therapeutic effect of the two growth factors favored pegfilgrastim. The duration of severe neutropenia was shorter with pegfilgrastim, especially with IE, and there were fewer episodes of febrile neutropenia and fewer documented infections in patients receiving pegfilgrastim. These observations are similar to results in the meta-analysis of randomized trials in adults (20).

Pegfilgrastim was well tolerated and had a toxicity profile similar to daily filgrastim. Growth factor–related bone pain and transient elevations in hepatic transaminases were observed in <14% of cycles on both arms. Episodes of mucositis and number of transfusions were also similar.

Pegfilgrastim's single injection per cycle dosing schedule provides a significant advantage over daily filgrastim injections for younger children, for whom needles and injections can cause substantial discomfort and distress. The single dosing schedule is also likely to result in improved compliance. These advantages reinforce the need to develop a formulation of pegfilgrastim that can be administered to children who weigh under 45 kg.

The combination of ifosfamide and etoposide was significantly less myelosuppressive than vincristine, cyclophosphamide, and doxorubicin. The duration of severe neutropenia after IE was shorter than V₃DC on both treatment arms, and a substantial proportion (38%) of patients across both arms did not have a documented ANC of <500/mcL after IE with thrice weekly monitoring of the blood counts. This suggests that higher doses of IE may be tolerable if it is administered with growth factor support.

The ANC profile with pegfilgrastim differed from the profile with filgrastim in that the peak ANC before the nadir was higher in patients receiving pegfilgrastim than in those receiving filgrastim, and the peak ANC after the nadir was higher on the filgrastim arm. Serum pharmacokinetics and neutrophil kinetics indicate that administration of pegfilgrastim results in higher serum granulocyte colony-stimulating factor concentrations and neutrophil count early in the cycle, but this did not impair the rate of neutrophil recovery after the nadir. The absence of high peak in postnadir neutrophil counts with pegfilgrastim reflects the self-regulatory, receptor-mediated clearance of pegfilgrastim. In contrast, high postnadir ANC peaks are associated with daily filgrastim (21), in part related to the dosing algorithm, which requires continued daily dosing until the neutrophil count recovers to 10,000/mcL. Mobilization of CD34+ cells after V₃DC was similar in both cohorts and was similar to mobilization with pegfilgrastim after cyclophosphamide in adults (22).

Pegfilgrastim serum concentrations were highly variable, which may reflect variability in the neutrophil count. The serum concentration peaked 24 hours after the dose and then declined in concert with the prenadir peak in the neutrophil count due to receptor binding. The secondary peak in the pegfilgrastim serum concentration around day 7 (Fig. 2A) may reflect release of the drug from receptors as the neutrophil count decreases. The clearance of pegfilgrastim in this population (11 mL/h/kg) was similar to that reported in adults (14 mL/h/kg; ref. 1). Clearance did not differ in patients receiving <6 mg compared with patients receiving ≥6 mg of pegfilgrastim. Filgrastim was absorbed more rapidly and cleared more rapidly than pegfilgrastim.

In conclusion, on this small randomized study, single dose per cycle of pegfilgrastim was well tolerated and was as effective as daily filgrastim based on the duration of severe neutropenia and number of episodes of febrile neutropenia and documented infections after dose-intensive treatment with V₃DC and IE. Future pediatric studies of pegfilgrastim should focus on the development and clinical testing of a formulation for patients <45 kg.

No potential conflicts of interest were disclosed.