Biosimilar granulocyte–colony-stimulating factor for mobilization of autologous peripheral blood stem cells in pediatric hematology-oncology patients




Recently biosimilars of granulocyte–colony-stimulating factor (G-CSF) became available for prophylaxis and treatment of postchemotherapy neutropenia and for mobilization of peripheral blood CD34+ cells for either autologous or allogeneic hematopoietic stem cell transplant. Most of the data on the mobilization efficacy and safety of biosimilar G-CSF are from adult patients, whereas no data are available in pediatric patients.

Study Design and Methods

This was a retrospective study on cases treated at three Italian pediatric transplant centers, from January 2011 to October 2013. Data were collected on all children undergoing first peripheral blood stem cell (PBSC) mobilization after stimulation with biosimilar G-CSF and chemotherapy. The results were compared with a historical control group.


Twenty-nine children underwent mobilization with biosimilar G-CSF. Peak peripheral blood CD34+ cell count of 20 × 106/L was achieved in 90% of patients, with a median value of 71 × 106/L. Eighty-three percent reached the desired target (CD34+/kg) dose. The median number of collected CD34+ cells was 10 × 106/kg (range, 4.8 × 106-68.8 × 106/kg). No difference was observed in comparison with historical control group mobilized with originator filgrastim. Moreover, no major and/or unexpected side effects were reported.


Biosimilar G-CSF resulted as effective and safe as originator filgrastim molecule in mobilizing PBSCs in children, with the advantage of a reduced cost.


fever of unknown origin


peripheral blood stem cell(s)

High-dose chemotherapy with autologous hematopoietic stem cell transplant is still an essential step of the standard care for selected types of cancer in children at high risk of relapse with conventional chemotherapy.[1] The use of mobilized peripheral blood stem cells (PBSCs) is preferred in the autologous transplant setting because of the easier collection procedure and the faster hematologic recovery.[2, 3] Collection may be achieved by using the administration of granulocyte–colony stimulating factor (G-CSF), with or without chemotherapy.[2]

A biosimilar drug is a copy of an approved original biologic substance whose data protection has expired.[4] Since the manufacturing processes remain proprietary, biosimilars are manufactured using independently developed proprietary processes. Thus, the new, biosimilar drug is a complex molecule of biologic origin, produced with different methods, not chemically identical to the originator. It must be assessed no less rigorously than the original compound or generics, and thus not surprisingly the regulatory requirements of the European Medicines Agency for the approval of biosimilars are even more demanding.[5] On the other hand, the use of biosimilars has been advocated to reduce the costs of drug expenditures, thus contributing to the financial sustainability of treatment programs as well as to save resources to be allocated in innovative therapies.[6, 7]

Studies on the use of biosimilars in pediatric patients are available almost exclusively for growth hormone: efficacy and safety data (growth rates, insulin growth factor-1 generation) for up to 7 years measure up favorably to previously approved growth hormones, which served as reference comparators.[8] Yet, to our knowledge no studies are available on the use of G-CSF biosimilars in pediatric patients. We describe our experience with G-CSF biosimilars in children for mobilization of PBSCs before autologous stem cell transplant.

Patients and Methods


This is a retrospective study performed at three Italian pediatric transplant centers, from January 2011 to October 2013. In all three centers, the local hospital policy was to use biosimilar G-CSF instead of the originator filgrastim both for treatment of postchemotherapy neutropenia and for PBSC mobilization. This policy was part of the measures adopted to contain the pharmaceutical costs.

Eligible patients were aged between 0 and 18 years, affected by solid tumor or Hodgkin's lymphoma, candidates for autologous transplant, and thus undergoing first PBSC mobilization with G-CSF and chemotherapy. All parents or patients (where applicable) gave their informed consent for data collection. Patient follow-up was updated as of November 2013.

Patients receiving biosimilar G-CSF were compared with a historical control group who received the originator filgrastim after mobilizing chemotherapy. The control patients were selected from a previous series,[3] by 1:1 matching with case patients on the basis of the following characteristics: age (±3 years), sex, and underlying disease.

Mobilization and collection procedures

The policy of PBSC collection in the three centers was similar with a target threshold of peripheral blood CD34+ cell count for starting leukapheresis of 20 × 106/L. The mean blood volumes processed during leukapheresis were 3. Biosimilar G-CSF was given subcutaneously or intravenously, according to patient preference.

Biosimilar G-CSF was started on Days +1 to +5 from the end of chemotherapy, at the dose of 10 μg/kg, not to exceed 300 μg/day, for 6 or more days. In these patients, white blood cell (WBC) count was checked every 2 to 3 days, until their nadir was reached. At the start of the recovery of WBCs, peripheral blood CD34+ cells were counted every 1 to 2 days until a blood peak count of CD34+ of at least 20 × 106 cells/L was reached. PBSC collection by leukapheresis was then scheduled on the same or the following day. Patients failing to achieve this blood peak count were classified as poor mobilizers with biosimilar G-CSF. The failure of mobilization was managed according to individual center policy, which comprised a subsequent course of mobilization or a marrow harvest. The monitoring of side effects of biosimilar G-CSF was performed by the prescribing physician at daily visit for admitted patients or at follow-up visit every 2 to 3 days for the discharged patients. Moreover, in all centers a pediatric hematologist was available on call 24 hours a day.

Leukapheresis was performed by standard volume procedure using an apheresis system (Cobe Spectra, Cobe BCT, Inc., Lakewood, CO) in two centers or a separator (Com.Tec, Fresenius Kabi AG, Bad Homburg, Germany) in one center and was continued, if possible, until the target dose of CD34+ cells/kg was achieved.

Aim of the study

For the purposes of this study, we considered as primary endpoint the number of patients who reached the threshold of at least 20 × 106 CD34+ cells/L in peripheral blood; the secondary endpoint was the number of patients who collected the target CD34+ cell dose required as per treatment protocol or by the transplant physician request.

We also collected information on the following items: number of leukapheresis procedures required to reach the target CD34+ cell dose, use of chemotherapy for the mobilization, underlying status of the disease at mobilization, type of biosimilar G-CSF, duration of severe neutropenia, incidence of febrile neutropenia and proven infection, time to polymorphonuclear leukocytes (PMNs) and platelet (PLT) recovery after the mobilization chemotherapy course, type, and severity of adverse events. The remission status of disease was defined as follows: complete remission as the disappearance of any clinical and radiologic signs of disease at sites involved at diagnosis; very good partial remission as the reduction of more than 90% but less than 100% of the disease at the sites involved at diagnosis; partial remission as the reduction of more than 50% but less than 90% of the disease at the sites involved at diagnosis; stable disease as the reduction of less than 50% of the disease at the sites involved at diagnosis. Proven infections were considered as all infections documented by the isolation of a germ from the blood or from a sterile site of the body.

Quality of PBSCs collected

As an indirect measure of the quality of the PBSCs collected with biosimilar G-CSF, we monitored the time to PMN and PLT recovery, number of febrile episodes (fever of unknown origin [FUO]), number of proven infections, duration of antibiotic therapy, duration of G-CSF, duration of hospitalization, and 100-day survival of patients who received PBSC transplant.

Autologous PBSC procedure

Myeloablation followed by autologous PBSC infusion was performed in high-efficiency particulate–filtered air rooms or isolated rooms, according to the policy of the center. Standard supportive care and preventive measures adopted to prevent infectious complications during the neutropenic phase, included fluconazole for antifungal prophylaxis, acyclovir for prophylaxis of herpes simplex virus, and cotrimoxazole for Pneumocystis jirovecil infections. Fever, defined as the presence of an oral or axillary temperature of at least 38.5°C in a single measurement or at least 38°C on two or more occasions taken at least 1 hour apart, was treated empirically with broad-spectrum antibiotics. Mucositis severity was defined according to WHO grading and painful swallowing or chewing requiring nonopioid (paracetamol) or opioid (tramadol, morphine) analgesic therapy. Red blood cell and PLT products were filtered to remove WBCs and irradiated (25 Gy). PMN and PLT recoveries were defined as the first of 3 and 7 consecutive days in which the counts were more than 0.5 × 109 and 50 × 109/L (unsupported by transfusion), respectively.

Statistical analysis

Patients' data were collected retrospectively by a specific case report form containing information on demographics (sex, age), disease (type, date of diagnosis, remission status), type of biosimilar G-CSF, type of mobilizing chemotherapy and complications (occurrence and duration of severe neutropenia, mucositis, infections), and PBSC collection (date of CD34+ cell peak, number of leukapheresis procedures to achieve the target PBSC dose); for patients who underwent PBSC transplant during the study period, type of conditioning regimen, number of CD34+ cells infused, early posttransplant complications (neutropenia, mucositis, infection, duration of parenteral nutrition, and antibiotic therapy needed), and date of last follow-up; in case of death by Day 100, date and cause of death were recorded. Descriptive statistics were reported as percentages for categorical variables and median and ranges for continuous variables. Probability of survival on Day +100 was estimated using the Kaplan-Meier method. Comparisons between case and control groups were performed using chi-square or Fisher exact test for categorical variables and using Mann-Whitney test for continuous variables. All analyses were performed using computer software (SAS, Version 9.2, SAS Institute, Cary, NC).


During the study period, a total of 38 consecutive patients were treated with biosimilar G-CSF for PBSC mobilization purpose: nine patients were mobilized from the steady state, while 29 patients were mobilized after a high-dose chemotherapy course. The last group represented the study group, and main characteristics of study patients, compared with those of control group, are shown in Table 1. Table 2 details the dose and schedule of chemotherapy administered to study patients.

Table 1. Main features of the study populationa
Patient characteristicStudy group, biosimilar G-CSF (n = 29)Control group, filgrastim (n = 29)p value
  1. aData are reported as number (%) or median (range). Definition of remission: complete remission = disappearance of any clinical and radiologic signs of disease at sites involved at diagnosis; very good partial remission = reduction of more than 90% but less than 100% of the disease at the sites involved at diagnosis; partial remission = reduction of more than 50% but less than 90% of the disease at the sites involved at diagnosis; stable disease = reduction of less than 50% of the disease at the sites involved at diagnosis. Disease status at mobilization CR + VGPR versus other.
  2. bNeuroblastoma, 22; Ewing sarcoma/peripheral neuroectodermic tumor, nine; rhabdomyosarcoma, two; osteosarcoma, two; Wilms tumor, one.
  3. cNumber of planned infusions one versus more than one.
Male19 (66)19 (66)1
Female10 (34)10 (34) 
Leukemia/lymphoma2 (7)2 (7)0.6
Intracranial solid tumor10 (34)8 (28) 
Extracranial solid tumorb17 (59)19 (66) 
Age at diagnosis (years)3.7 (0-18.7)4.2 (0.5-17.7)0.5
Age at mobilization (years)4.6 (0.4-23.0)4.4 (0.8-18.4)0.9
Disease status at mobilization
Complete remission3 (10)2 (7)0.9
Very good partial remission4 (14)4 (14) 
Partial remission20 (69)14 (48) 
Stable disease2 (7)6 (21) 
Not evaluable 3 (10) 
Required dose of CD34+ (×106/kg)5 (3.5-16)5 (2-10)0.7
Number of planned infusions
119 (66)20 (69)0.8c
210 (34)8 (28) 
3 1 (3) 
Body weight (kg)15 (4-79)16 (9-86)0.7
Table 2. Chemotherapy regimens used for mobilization
Number of drugs usedType of drugs and dosesNumber of patients
1Etoposide 2000-3000 mg/m25
Ifosfamide 1500 mg/m21
  Subtotal 6
2Vincristine 1.5 mg/m2, cisplatin 80-96 mg/m28
Cyclophosphamide 4000 mg/m2, etoposide 600 mg/m24
Ifosfamide 15 g/m2, etoposide 750 mg/m21
Carboplatin 800 mg/m2, etoposide 300 mg/m21
Cyclophosphamide 4000 mg/m2, vincristine 1.5 mg/m21
Doxorubicin 60 mg/m2, cyclophosphamide 1500 mg/m21
  Subtotal 16
3Ifosfamide 6000-9000 mg/m2, etoposide 300-450 mg/m2, carboplatin 500-800 mg/m25
Topotecan 2.25 mg/kg, vincristine 1.5 mg/m2, doxorubicin 45 mg/m21
Vincristine 1.5 mg/m2 , cyclophosphamide 1500 mg/m2, actinomycin D 1.5 mg/m21
  Subtotal 7

The biosimilar G-CSF used was Zarzio (Sandoz, GmbH, Kundl, Austria). The median duration of biosimilar G-CSF administration was 8 days (range, 3-15 days), with a median dose of 10 μg/kg/day (range, 4-10.5 μg/kg/day). It was started after a median of 5 days (range, 1-26 days) from the end of chemotherapy. It was well tolerated and no withdrawal or severe side effects were recorded, only mild bone pain and headache in one patient each, respectively. After mobilizing chemotherapy, the recorded adverse events included FUO in 13 patients, proven infection in two, mucositis in one; moreover, 25 patients had severe neutropenia (PMN < 500 × 106/L), with a median duration of 4 days (range, 1-73 days). This information is summarized in Table 3, and no difference was found with the historical control group except for a major incidence of episodes of FUO, 45% versus 17% (p = 0.02).

Table 3. Side effects of mobilizing chemotherapy regimens in 29 patients of study group and in 29 patients of control group*
Clinical featuresStudy group, biosimilar G-CSFControl group, filgrastimp value
  1. * Data are reported as median (range) or number (%).
  2. CT = chemotherapy.
Time from CT to biosimilar G-CSF5 (1-26)6 (2-16)0.4
Duration of biosimilar G-CSF8 (3-15)8 (3-19)0.5
Mucositis post-CT
No28 (97)25 (86)0.4
Yes1 (3)4 (14) 
FUO post-CT
No16 (55)24 (83)0.02
Yes13 (45)5 (17) 
Proven infections
No27 (93)28 (97)1
Yes2 (7)1 (3) 
PMNs < 500 × 106/L
No4 (14)8 (28)0.2
Yes25 (86)21 (72) 
Duration of PMNs < 500 × 106/L4 (1-73)5 (2-16)0.06

Regarding the primary endpoint, 26 patients (90%) reached a peak CD34+ cell count of more than 20 × 106/L on peripheral blood, median value being of 71 × 106/L (range, 25 × 106-826 × 106/L). Twenty-four (83%) reached the target cell dose of 5 × 106 CD34+/kg desired by transplant physician. In these patients the median number of CD34+ collected cells was 10 × 106/kg (range, 4.8 × 106-68.8 × 106/kg). This target dose was achieved by one or two leukapheresis procedures in 18 and six patients, respectively. The time interval from the start of chemotherapy to PBSC collection was 13.5 days (range, 6-31 days). No difference was observed in comparison with the historical control group mobilized with originator filgrastim molecule (Table 4).

Table 4. Summary of the results of peripheral stem cell mobilization with biosimilar compared with a matched historical control group who used filgrastim*
ResultsStudy group, biosimilar G-CSF (n = 29)Control group, filgrastim (n = 29)p value
  1. * Data are reported as median (range) or number (%).
Patients achieving blood CD34+ cell peak ≥20 × 106/L
No3 (10)6 (21)0.5
Yes26 (90)23 (79) 
Value of CD34+ blood count at first mobilization (×106/L) in patients successfully mobilized71 (25-826)98 (28-804)0.3
Days to CD34+ peak7 (3-14)6 (2-18)0.5
Patients achieving the target CD34+/kg dose
No5 (17)11 (38)0.08
Yes24 (83)18 (62) 
Yield of CD34+ cells (106/kg) in patients achieving the target10 (4.8-68.8)12.3 (4-46.1)1
Time interval from the start of mobilizing chemotherapy to CD34+ collection (days)13.5 (6-31)12 (8-22)0.8
Number of daily leukapheresis procedures needed for achieving the target CD34+ dose
118 (75)14 (78)1
26 (25)4 (22) 

Of the five patients who failed to reach the target CD34+ cell dose, two underwent a marrow harvest, one underwent a second mobilization course, one lost the eligibility to PBSC transplant due to disease progression, and one proceeded to PBSC transplant despite a suboptimal CD34+ cell dose (3.7 × 106/kg vs. 5 × 106/kg).

As of November 2013, a total of 19 of the 29 patients who were successfully mobilized with biosimilar G-CSF, underwent autologous transplant after a median time of 124 days from PBSC collection (range, 56-220 days). The median number of CD34+ infused was 5 × 106/kg (range, 3.2 × 106-20 × 106/kg). All patients engrafted for PMNs in a median time of 12 days (range, 8-22 days), and 95% engrafted for PLTs in a median time of 16 days (range, 10-39 days). One patient (5%) proceeded with a second PBSC transplant before PLT engraftment was achieved. Mucositis occurred in 84% of patients and lasted a median of 7 days (range, 3-26 days), whereas FUO and proven infection occurred in 89 and 16% of patients, respectively. The median duration of hospitalization was 17 days (range, 12-28 days). With a median follow-up of 167 days (range, 70-727 days), the resulting 100-day overall survival was 100%. The only differences with historical control group were a shorter duration of mucositis (median, 7 days vs. 10 days; p = 0.04), and a less frequent use of preengraftment G-CSF (63% vs. 95%; p = 0.04) in the study group. Table 5 summarizes the main clinical and transplant data.

Table 5. Main transplant characteristics of the study patients who underwent hematopoietic stem cell transplantation after PBSC collection with biosimilar G-CSF compared with a group of 19 historical controls mobilized with filgrastima
Clinical characteristics of transplanted patientsStudy group biosimilar G-CSF (n = 19)Control group, filgrastim (n = 19)p value
  1. aData are reported as median (range) or number (%).
  2. bNeuroblastoma, 19; Ewing sarcoma/peripheral neuroectodermic tumor, seven; rhabdomyosarcoma, one; Wilms tumor, one.
Intracranial solid tumor6 (32)4 (21)0.5
Extracranial solid tumorb13 (68)15 (79) 
Time from mobilizing course to autologous transplant (days)124 (56-220)109 (45-218)0.3
CD34+ infused, ×106/kg5.0 (3.2-20.0)6.3 (1.1-46.1)0.6
PMN engraftment19 (100)19 (100) 
Time to PMN engraftment (days)12 (8-22)11 (8-23)0.2
PLT engraftment (>20 × 109/L)
Yes18 (95)19 (100)1
NA1 (5)0 (0) 
Time to PLT engraftment (days)13 (9-23)13 (5-58)1
PLT engraftment (>50 × 109/L)
Yes18 (95)18 (95)1
NA1 (5)1 (5) 
Time to PLT engraftment (days)15.5 (10-39)21.5 (7-58)0.5
No2 (11)4 (21)0.7
Yes17 (89)15 (79) 
Proven infections
No16 (84)11 (58)0.07
Yes3 (16)8 (42) 
No3 (16)0 (0)0.2
Yes16 (84)19 (100) 
Duration of mucositis (days)7 (3-26)10 (3-20)0.04
No7 (37)1 (5)0.04
Yes12 (63)18 (95) 
Duration of G-CSF (days)8.5 (5-14)8.5 (1-18)0.8
Days of hospitalization (days)17 (12-28)18 (13-43)0.2
Follow-up from transplant (days)167 (70-572)295 (123-456)0.07
100-day survival (days)19 (100)19 (100)


To the best of our knowledge, this is the first report of the use of biosimilar G-CSF for the mobilization of PBSC in children committed to autologous transplantation. Biosimilar G-CSF is a cheaper alternative to originator molecule filgrastim. The patent for filgrastim expired in Europe in 2006 and in the United States in 2013. European Medicines Agency has approved several biosimilar versions, three of which are commercially available.[9] Yet, concerns have been raised in 2011 among hematologists on the use of biosimilar G-CSF in stem cell transplant, until sufficient efficacy and safety data are available.[10] The first report of the use of a biosimilar G-CSF for stem cell mobilization for autologous transplantation was published by Lefrère and colleagues in 2011,[11] in whom 40 patients achieved similar mobilization yield and safety profile to the originator G-CSF. Moreover, the biosimilar G-CSF–mobilized cells were able to rescue the patient following high-dose chemotherapy. Speed of both myeloid and PLT engraftment was similar to the engraftment with stem cells mobilized with the originator G-CSF. Additional favorable data have been recently reported in adults.[12] Thus, currently available evidence for equivalence between biosimilar G-CSF and the originator products comprises identical or very similar amino acid sequence and production in Escherichia coli,[8] results of trials in patients and volunteers,[13] in autologous[12, 14, 15] as well as in allogeneic stem cell mobilization.[16] A recent review found in the published literature a total of 904 patients and healthy related or unrelated donors who were mobilized successfully with biosimilar G-CSF with side effects and a posttransplant engraftment similar to cases treated with the originator G-CSF.[17]

In this study, 90% of patients achieved the primary endpoint, that is, a peak CD34 cell count of at least 20 × 106/L. Although this was not a randomized controlled study, the comparison with the results of a historical matched control group may be informative. This group was selected from a database containing the results of a recent prospective, observational study conducted in 10 Italian pediatric hematology-oncology centers on the efficacy of mobilization of autologous PBSCs with the originator molecule of filgrastim in a large population of 145 children.[3] In this control group, the proportion of patients achieving the current primary endpoint, that is, the peak CD34 cell count of at least 20 × 106/L, was 79%. When additional qualifying variables were compared, such as the number of CD34 cells mobilized, the time required to reach the peak CD34 count, and the number of daily apheresis procedures required, all of them were fully comparable in the two cohorts. Moreover, the engraftment efficiency of PBSCs collected by biosimilar G-CSF and the time of hospitalization for autologous PBSC transplant were similar in the study and control groups and were comparable to that reported in another study using pegfilgrastim for mobilization.[18] Thus, the use of a biosimilar G-CSF seems to be at least as effective as the originating filgrastim molecule, in a comparable setting of children with cancer undergoing PBSC mobilization. Moreover, its safety profile was very favorable and consistent with the expected side effects for this class of drugs, such as mild bone pain and transient headache. Although we noted a shorter duration of mucositis in the study group after PBSC transplant, we think that this finding deserves a confirmation in a prospective study.

The use of biosimilar G-CSF has an impact on the cost of mobilization. Assuming that one vial containing 300 μg was used for each day of treatment, the market cost of the biosimilar G-CSF in Italy ranged between 95.97 euros and 99.80 euros; this compared favorably with the 127.96 euros cost of the originator molecule filgrastim.

This study has limitations, since it is not a randomized comparative study of the biosimilar G-CSF versus the originating drug, but a retrospective data collection, conducted in a limited number of centers. Nevertheless, the setting was highly comparable to that of the previous prospective study conducted by our cooperative group. The apparent noninferiority of biosimilar G-CSF to the originating filgrastim molecule is consistent with its pharmacodynamic properties and make it appealing for the favorable cost-effect ratio. These results would deserve a confirmation by a prospective, randomized study.

Conflict of Interest

The authors have disclosed no conflicts of interest.