Introduction

Autologous peripheral blood progenitor cells (PBPC) provide a rapid and sustained hematopoïetic recovery after the administration of high-dose therapy in patients with hematological malignancies and certain solid tumors.1, 2, 3, 4, 5 The mobilization of PBPC is usually achieved with the use of hematopoietic growth factors (HGFs) alone or in combination with chemotherapy, in which case a higher yield of CD34+ cells can be reached and collected.6 It is generally recognized that granulocyte colony-stimulating factor (G-CSF) as a single agent mobilizes more CD34+ cells than does granulocyte–macrophage colony stimulating factor (GM-CSF).7, 8, 9 Other HGFs, such as interleukin-3,10 Flt3 ligand,11 stem cell factor12, 13 and erythropoietin 14 or antagonists of SDF-1-CXCR415 have been tested generally in combination for mobilizing PBPC. The recommended dose of G-CSF after chemotherapy is 5 μg/kg/d.16 Increasing the dose of G-CSF to more than 5 μg/kg/d17, 18 or changing the administration schedule19 has not been proven to improve PBPC mobilization. The use of a single dose of pegfilgrastim was demonstrated as equivalent to a daily administration of filgrastim for mobilizing PBPC.20 The infusion of CD34+ cells higher then 3–5 × 106 CD34/kg results in rapid engraftment, reduces the transfusion need and may have a clinical influence.21, 22, 23 The synergistic effect of coadministration of HGFs like G-CSF and GM-CSF on PBPC mobilization has been suggested in phase I/II study,24 for improving the harvest of CD34 cells.25 Several non-randomized or randomized clinical trials have been performed, showing a little or no benefit for sequential administration of standard doses (5–10 μg/kg/d) of GM-CSF and G-CSF.26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37 Neverthelesss, the use of these HGFs was not well explored and particularly the minimal efficient dose for their concomitant administration, following high dose cyclophosphamide (CY). We report here two randomized studies comparing G-CSF to the association of G-CSF and GM-CSF.

Materials and methods

Study design

As shown in Figure 1, this was a randomized, open-label, unicenter study, including two parts (study A and study B). The primary objective of this study was to achieve a maximal of 10 × 106 CD34 cells/kg in a minimal number of leukapheresis. Calculation of the number of subjects was based on the fact that the percentage of the subjects reaching more than 10 × 106 CD34 cells/kg in the combination arm will be twice than that observed in the single-cytokine arm. With a power of 90% and a 5% risk, the number of patients is 25 per arm. For that reason, we decide to include 30 patients per arm due to the possibility of unevaluable patients. There was no stratification. The secondary objectives of this study include tolerance, factors influencing the mobilization of progenitor cells, percentage of myeloma cells mobilized in PB, and difference of mobilization at the level of 5 × 106 CD34 cells/kg between the two arms. In study A, 60 patients were randomized to receive after CY (4 g/m2) 5 μg/kg/day (d) of rHu-GM-CSF (Molgramostim, Schering-Plough, Kenilworth, NJ, USA) and 5 μg/kg/d of rHu-G-CSF (Filgrastim, Amgen, Thousand Oaks, CA, USA) (G+GM-CSF, total HGF dose=10 μg/kg/d) compared to 5 μg/kg/d of G-CSF (Filgrastim). In study B, 60 subsequent patients have received the same design of treatment with 2.5 μg/kg/d of GM-CSF and 2.5 μg/kg/d of G-CSF (G+GM-CSF, total dose of HGF=5 μg/kg/d) compared to 5 μg/kg/d of G-CSF. The 4 g/m2 CY was administered by intravenous (i.v.) route to all the patients followed 24 h later by the HGF administration subcutaneously until the last day of leukapheresis (LK). For patients receiving the two HGFs, they were injected in separate sites.

Figure 1
figure 1

Design of the study: mobilization and collection of CD34 cells.

Full size image

Patient eligibility

Written informed consent was obtained from all patients enrolled in the study. Patients with histologically confirmed history of cancer and requiring a high dose myeloablative chemotherapy with PBPC rescue were eligible. Disease status and demographic characteristics are detailed in Table 1. Inclusion criteria were: age between 18 and 65 years, performance status ⩽2, half-life expectancy of at least 6 months, normal organ function as defined by serum creatinine, transaminase <2 times normal range, cardiac ejection fraction within the institution's normal range, and normal blood count (ANC>1.5 × 109/l, platelets >100 × 109/l) at the day of mobilization chemotherapy. Exclusion criteria were: patients who had received HGF within 2 weeks before study entry, patients with other malignancies within 5 past years, HIV1 or 2, HTLV 1 or 2, hepatitis C sero-positivity, hepatitis B positive virological status, patients with psychiatric, addictive or any disorder which compromised their ability to give truly informed consent.

Table 1 Demographic and disease status of patients at diagnosis
Full size table

Stem cell collection

White blood counts (WBC) were assessed daily after the beginning of chemotherapy. When it reaches 0.5 × 109/l, the percentage of circulating CD34 cells was monitored daily according to the ISHAGE protocol, by using a phycoerythrin-conjugated (PE) antibody to CD34 (Q-Bend 10, Immunotech, Marseille, France) using a FACScan (Becton-Dickinson, San-Jose, CA, USA).38 The estimated number of CD34 cells/kg that can be collected in one LK was calculated by the central laboratory according to the following formula: (%CD34 × WBC × 8/kg body weight (eight being a coefficient corresponding approximately to 2 blood volumes in liter). Then, a LK was started if the estimation of harvestable CD34 number was ⩾106 CD34/kg. If the estimation of harvestable CD34 cell number was below 106 CD34/kg the collection was delayed. A maximum of three LK was attempted to reach a minimum target of 5 × 106 CD34/kg. For MM patients, at this time, we performed a CD34 selection in order to decrease contaminating tumoral cells. For that reason, due to the loss of cells during this procedure, we attempt to collect 10 × 106 CD34/kg. Large volume LK was performed through a dual peripheral venous puncture using a Cobe Spectra separator version 4 (CS 3000; Baxter Healthcare Corp Lakewood, CO, USA). Median time of processing was 5 h and the mean volume of processed blood ranged between 12 and 16 l.

Analysis of product of leukapheresis

The determination of CD34 cell content in the LK was carried out according to the following procedure. Incubation of 106 cells with 30% AB serum (100 μl) for 10 min, two washings with PBS, suspension in a final volume of 100 μl, incubation with 20 μl of monoclonal antibody CD34 conjugated with PE, or 20 μl of murine IgG recognizing no human antigen and conjugated with PE as a negative control, or 5 μl anti-CD 45 conjugated with PE as a positive control for 30 mn at 4°C, cell lysis followed by two washings and resuspension in 200 μl PBS. At least 30 000 events were acquired for each sample. In patients with multiple myeloma (MM), myeloma cells were enumerated by FACS analysis using the FITC-conjugated MI15 (anti-CD138 mAb.39 106 cells were incubated with 1 μg MI15FITC (10 μl) or to 1 μg IgGFITC negative control (20 μl) and fluorescence was analyzed with FACSCAN cytofluorometer. No assesment of CD34 cells was performed on the LK product.

Autologous stem cell transplantation

After a resting time of a maximum of 8 weeks, patients received high-dose chemotherapy (HDC) followed by autologous PBPC transplant. MM patients were scheduled to receive two subsequent HDC and PBPC transplants. The first HDC was 140 mg/m2 Melphalan. At least 3 months later, patients received the second HDC consisting of 200 mg/m2 melphalan or 140 mg/m2 melphalan and total body irradiation delivering 12 Gy followed by 5 μg/kg/d G-CSF support administrated from day 2 after HDC and until ANC >1.5 × 109/l on two consecutive counts. Patients having malignant lymphoma received a BEAM regimen consisting on 300 mg/m2 BCNU, Etoposide 200 mg/m2/d for 4 days, Cytarabine 200 mg/m2/d for 4 days and Melphalan 140 mg/m2 without post-transplant G-CSF support. Breast cancer, testis and ovarian cancer patients were scheduled to receive one or two HDC and PBPC transplant. The following parameters of hemotologic recovery were analyzed: day of ANC ⩾0.5 × 109/l, day of platelet count ⩾50 × 109/l, cost estimation of the whole transfusion procedure and duration of hospitalization.

Statistical analysis

An intent-to-treat analysis was done taking into account all patients who received CY and HGFs. A Wilcoxon non-parametric test was carried out to assess the comparability of group at randomization. The comparison of number of LK per group were evaluated by an exact Fischer's test. The median values of CD34 cells harvested in the different groups were compared by a Mann–Withney non-parametric test. The comparison of distribution of patients with 1 to 3 LK to reach the pre-defined threshold was done by the χ2 test. If no LK were performed it is considered as mobilization failure and not evaluated. Data were analyzed using Statistical Analysis System (SAS) software version 6.08.

Results

Patient population: demographics and distribution (Table 1)

One hundred twenty consecutive patients were enrolled and randomized in the Hematology-Oncology department, University hospital of Montpellier. All patients had a disease requiring HDC and PBPC transplantation as salvage therapy. They were 58 (48%) MM patients (stage II or III), 32 (27%) non-Hodgkin Lymphoma (NHL) (large cell type with IPI⩾3 or sensitive relapse), 8 (7%) Hodgkin disease (HD) (sensitive relapse or refractory), and 4 (3%) chronic lymphocytic leukemia (CLL) (Binet stage B and C), and 18 (15%) solid tumors (ST) including 13 metastatic (or high-risk) breast cancer four metastatic testis cancer, 1 metastatic ovarian cancer. Sixty patients were included in study A (5 μg/kg/d G-CSF plus 5 μg/kg/d GM-CSF versus 5 μg/kg/d G-CSF): 26 MM, 12 NHL, 6 HD, 2 CLL, 14 ST. Sixty subsequent patients were included in study B (2.5 μg/kg/d G-CSF plus 2.5 μg/kg/d GM-CSF versus 5 μg/kg/d G-CSF): 32 MM, 20 NHL, two HD, two CLL, four ST. Four patients were not evaluable for the following events at time of chemotherapy: infection at time of chemotherapy (one case), disease progression or relapses (two cases), death before mobilization (one case). Patients characteristics were comparable, particularly age, sex ratio, body weight and blood cell count before mobilization.

Kinetics of mobilization of circulating CD34 PBPC

In study A, the circulating CD34 cells appear in the peripheral blood (PB) respectively between day 9 to 12 in the G-CSF group and day 9 to 13 in G-CSF plus GM-CSF group. In study B, the circulating CD34 cells appear between day 9–11 in the G-CSF group and day 9–13 in G-CSF plus GM-CSF group (data not shown).

Collection of PB CD34 cell

There was a good correlation between the estimated number of CD34 cells and the collected number of CD34 cells measured (Figure 2). Parameters of LK procedures are shown in Table 2. A total of 232 cytaphereses were performed: 123 in study A, and 109 in study B. One to three large volume LK were performed to harvest the required number of CD34 cells. The following parameters of LK were analyzed: number of LK processed per patient (LK/patient), the day of first LK after CY, WBC at the first LK, and the median number of blood volumes treated per LK (BV/LK). None of these parameters were statistically different between the randomized groups.

Figure 2
figure 2

Correlation between the estimated number of CD34 cells (%CD34 × WBC × 8)/kg body weight (eight being a coefficient corresponding approximately to two blood volumes in liter) and the number of collected CD34 cells.

Full size image
Table 2 Parameters of leukapheresis (LK) procedures
Full size table

Number of patients achieving a harvest of 5 × 106 and 10 × 106 CD34 cells/kg

In study A, 123 LK were performed (61 in the G-CSF group, 62 in the G+GM-CSF group). The distribution of patients undergoing one, two or three LK were identical in both groups. Among the patients who have undergone LK, the proportion of patients reaching a minimum target of 5 × 106 CD34 cells/kg was not statistically different between the G-CSF arm (21/27, 77.8%) and the G-CSF plus GM-CSF arm (25/29, 86.2%). When the objective of CD34 collection was 10 × 106 cells/kg, the proportion of patients reaching this target was significantly higher in the G-CSF plus GM-CSF arm (21/29, 72.4%) compared to the G-CSF arm (11/27, 40.7%: P=0.00006).

In study B, a total of 109 LK were performed (56 in G-CSF arm and 53 in G-CSF plus GM-CSF arm). The distributions of patients undergoing one, two or three LK are identical in both groups. Among the patients who have undergone LK, the proportion of patients reaching a minimum target of 5 × 106 CD34 cells/kg was not statistically different between the G-CSF arm (20/27, 74.1%) and the G-CSF plus GM-CSF arm (20/26, 76.9%). When the objective of CD34 collection was 10 × 106 cells/kg, the proportion of patients reaching this target was significantly higher in the G-CSF plus GM-CSF (arm 15/26, 57.7%) compared to the G-CSF arm (10/27, 37%: P=0.003) (Figure 3).

Figure 3
figure 3

Percentage of patients reaching 5 or 10 × 106 CD34 cells/kg in study B.

Full size image

CD34 cell harvest (Table 3)

Table 3 Study A and B: Collection of CD34+ PBSC (× 106/kg) for harvested patients (n) in each group
Full size table

In study A, the median number of CD34 cells × 106/kg collected in LK1, LK2 or LK3 in the G-CSF and G+GM-CSF groups is respectively, 3.1. (range 0.9–44.7), 3.7 (1.2–13.1), 2.7 (1.6–7.6) vs 7.3 (1–30.9), 6.28 (0.8–59), 5.72 (0.8–12.5). The total number of CD34 cells × 106/kg harvested is twofold higher in the G+GM-CSF group 18.3 (3–90) compared to 9 (3–45) in the G-CSF group (P=0.09).

In study B, the median number of CD34 cells × 106/kg collected at LK1, LK2, LK3 in the G-CSF and G+GM-CSF groups is respectively, 3.7 (range 0.6–30.1), 3.0 (0.6–14.3), 3.3 (2.9–7) vs 7.07 (0.9–24.5), 8.02 (0.7–29.8), 6.71 (0.6–9.3). The total number of CD34 × 106/kg harvested is about twofold higher in the G+GM-CSF group 15.85 (2.5–34.9) compared to 8.1 (1.6–33.9) in the G-CSF group (P=0.09).

To avoid a bias due to the variable number of LK, and due to the variable blood volume treated during one LK, we have evaluated the efficacy of CD34 PBPC collection by calculating the number of CD34 cells × 106/kg collected during the first LK according to the number of blood volume processed (CD34 LK1/BV). In study A, this parameter was again increased about twofold in the G-CSF plus GM-CSF group 2.4 (0.3–13) compared to 1.1 (0.3–12) in the G-CSF group (P=0.02). In study B, the same observation could be done in the G+GM-CSF group: 2.52 (0.3–10.2) vs 1.37 (0.2–11.6) in the G-CSF group (P=0.04).

Patients with MM

In the study A: 14 MM patients received G-CSF and 12 received G-CSF plus GM-CSF (11 are evaluable). The total CD34 cells × 106/kg harvested is 2.2-fold higher in the G-CSF plus GM-CSF group compared to the G-CSF group respectively 23.8 (11.1–89.9) vs 11.02 (3.2–29.2) (P=0.005). The CD34 LK1/BV is 2.4-fold higher in the G-CSF plus GM-CSF group 2.81 (0.9–13.3) vs 1.17 (0.3–8.1) in the G-CSF group (P=0.02). In the MM patients, we recommended to collect a minimum target of 10 × 106 CD34 cells/kg before the CD34 positive selection. In these conditions, 11/11 (100%) of patients receiving the G-CSF plus GM-CSF combination reach this target compared to only 8/14 (57%) in the G-CSF group (P=0.02) (Figure 4). The same observations were carried out in patients of study B, 17 MM patients received G-CSF (15 are evaluable) and 15 received G-CSF plus GM-CSF (14 are evaluable). The total CD34 cells × 106/kg harvested is about 2.5-fold higher in the G-CSF plus GM-CSF group (24.6, range: 2.5–34.9) compared to the G-CSF group (9.26, range: 2.8–33.9) (P=0.03). The CD34 LK1/BV is 2.2-fold higher in the G-CSF plus GM-CSF group 2.95 × 106/kg (0.4–10.2) compared to 1.35 × 106/kg (0.1–11.6) (P=0.014). Patients (12/14 (86%)) receiving the G-CSF plus GM-CSF combination reach the defined target of 10 × 106 CD34+ cells/kg compared to only 7/15 (44%) in the G-CSF group (P=0.03) (Figure 4).

Figure 4
figure 4

Percentage of patients with multiple myeloma and reaching 10 × 106 CD34 cells/kg. ▪ Patients receiving G-CSF only. □ Patients receiving G-CSF plus GM-CSF.

Full size image

Patients with other lymphoid malignancies (NHL, HD and CLL)

Mobilization by CY and G-CSF plus GM-CSF did not significantly improve the amount of collected CD34 cells compared to the CY and G-CSF mobilization especially the CD34 LK1/BV was not significantly increased (Table 3). The comparison between myeloma and lymphoma patients is detailed in Table 4. MM patients received low number of lines of chemotherapy (all the patients were in front line therapy with a median number of chemotherapy courses of 3, range: 2–6), while lymphoma patients were in first- or second-line of therapy and received a median of 6 chemotherapy courses (median 2–16, P=0.02). As shown in Figure 5, no differences were seen between the two arms (G-CSF and G+GM-CSF) in the group of patients with more than 1 year of chemotherapy before mobilization.

Table 4 Differences between MM and Lymphoma
Full size table
Figure 5
figure 5

Kinetics of circulating CD34 cells in patients with less (a) and more (b) than 1 year of prior chemotherapy. ▪ Patients receiving G-CSF only. □ Patients receiving G-CSF plus GM-CSF.

Full size image

In study A, among patients receiving G-CSF, the CD34 LK1/BV is not significantly higher in MM 1.2 × 106 (0.3–8) than in Lymphoma 0.9 × 106 (0.4–12) (P=0.12) while among patients receiving G-CSF plus GM-CSF, the CD34 LK1/BV is 1.5 higher in MM (2.8 × 106, range: 1–13) than in Lymphoma (1.9 × 106, range: 0.3–7) (P=0.047). In study B, same observations was made, the CD34 LK1/BV is not significantly higher in MM than in Lymphoma (P=0.8) while among patients receiving G-CSF plus GM-CSF, the CD34 LK1/BV is 2.1 higher in MM (2.95 × 106, range: 0.4–10) than in Lymphoma (1.4 × 106, range: 0.3–5) (P=0.01).

Patients with solid tumors

Owing to of low number of patients with solids tumors and their heterogenicity, none comparison was made between the two groups of treatment.

Mobilization failure

The percentage of patient that failed to mobilize sufficient PBSC was similar in study A (4/56, 7%) and in study B (3/53, 6%). None difference has been observed according to the HGF treatment group (Table 5).

Table 5 Mobilization failure in patients receiving the hematopoietic growth factor
Full size table

Adverse events

No serious adverse events were associated with the administration of cyclophosphamide at 4 g/m2. Grade 1 and 2 nausea and vomitis were easily controlled by systematic administration of andosetron. In study A, three patients with full dose of G+GM-CSF experienced grade 2 symptoms (chill, bone and muscle pains, tachycardia) within the first 2 days. These patients stopped the GM-CSF themselves. No patient experienced grade 1 and 2 toxicities during HGF treatment in the study B.

Effects of HGF on residual plasma cells mobilization in MM patients

In study A, an estimation of residual PC in product of LK was possible in 19 patients. The number of PC was 4.6 × 106/kg in the G-CSF group and 5.39 × 106/kg in the G-CSF plus GM-CSF group (P=NS).

In study B, the numeration of residual PC in LK was carried out in 21 patients with MM. The number of PC was 2.58 × 106/kg in the G-CSF group and 6.19 × 106/kg in the G+GM-CSF group (P=NS). When comparing the total PC collected to the total CD34 cells collected in each patient, the PC/CD34 ratio is in study A 1 for 2.4 in G-CSF group and 1 for 4.42 in the G-CSF plus GM-CSF group (P=NS). In study B, the PC/CD34 ratio is 1 for 3.6 in G-CSF group and 1 for 3.97 in the G-CSF plus GM-CSF group (P=NS) showing no enhancement of PC mobilization by G-CSF plus GM-CSF regimen.

Hematologic recovery after myeloablative treatment

Ninety-seven patients were transplanted: 49 patients in Study A, 48 in study B. All MM patients were scheduled to receive transplantation with a graft purged by CD34 positive selection. In study A, the CD34 positive selection was carried out in only 9/14 (64.3%) MM patients receiving G-CSF and 11/11 (100%) in patients receiving G-CSF plus GM-CSF (P=0.03). In study B, this regimen was done in only 8/15 (53.3%) patients receiving G-CSF and 13/14 (92.8%) patients receiving G-CSF plus GM-CSF (P=0.02).

As a result of purification and double HDC in MM, the median number of CD34 cells infused was similar in all treatment groups and ranged between 3.78 to 4.32 × 106/kg. As expected, all patients had an engraftment, the median times to reach an ANC of ⩾0.5 × 109/l and platelet ⩾50 × 109/l were, respectively, 11 days (range 9–17) and 15 days (8–34). The transfusion cost is estimated at 1619 $ (265–8752). The median duration of hospitalization was 24 days (16–94).

Discussion

These two randomized studies demonstrate two major points:

  • the combination of GM-CSF plus G-CSF following CY results in a twofold enhancement of the number of PBPC, in a safe and efficient manner for an homogenous group of patients, particularly seen in myeloma patients,

  • the concomitant administration of these two HGFs, had additive effects, as observed by the superiority of the combination at equivalent doses (5 μg/kg vs 2.5 plus 2.5 μg/kg).

Most of the patients included in these studies, achieved a minimal target of 5 × 106 CD34/kg, but more patients receiving G-CSF plus GM-CSF achieved a target of 10 × 106 CD34/kg. This result was obtained with a median of 2 LK. No adverse effects related to LK procedures have been reported. Although CD34 cell dose is currently considered to be the prefered indicator of satisfactory engraftment, the optimal CD34 cell dose for autologous transplantation remains to be defined. A collection of a threshold value of at least 3 to 5 × 106 CD34/kg is recommended to insure rapid and successful engraftment and long-term hematological recovery. A dose–response relationship is evident between the number of CD34 cells/kg infused and neutrophil and platelet engraftment kinetics, decrease hospital stay and cost.40, 41, 42 Infusion of very high dose superior to 15 × 106 CD34/kg was shown to be associated with a shorter time for platelet recovery compared to that observed for patients receiving between 2.5 and 15 × 106 CD34/kg.43 However, Dercksen et al.44 found no difference for neutrophil recovery when the number of CD34+ cells were superior to 6 × 106 cells/kg. Recently, different groups observed a linkage between the number of CD34 harvested and reinfused, and survival or duration of the response in different diseases including breast cancer and multiple myeloma.45, 46 The mechanism of such effect remains uncertain, but may be linked to a high number of immunocompetent cells collected in addition to the stem cells, as suggested by Porrata et al.46 who observed a correlation between progression-free survival and the number of lymphocytes readministered. In addition, GM-CSF had several immune activities, including differentiation of dendritic cells from monocytes and it has been associated to different strategies of immune therapy including anti-tumoral vaccination.46, 47 These observations may represent a rationale for re-discovering the role of GM-CSF for mobilizing both hematopoietic stem cells and immune cell effectors.48, 49 The response we made through this study appears important to define the minimal efficient dose for a combination of G-CSF and GM-CSF, prior to add other compounds for mobilization such as selective CXCR4 antagonists.15

We observed that the efficacy of dual HGFs was more evident in patients having MM, where 2.5-fold higher PBSC are collected with the combination of HGFs. However, no differences have been observed in the group of lymphoma patients. The impressive results observed with MM compared to the lymphoma group, are probably due to prior chemotherapy compared to the lymphoma group. Some investigations have highlighted the importance of prior exposure to the specific stem cell toxicity of chemotherapy containing alkylating agent and/or anthracycline and their impact on mobilization efficacy.50 Haas et al.51 suggest the same findings showing a loss of 0.2 × 106 CD34+ cells/kg per chemotherapy course including alkylating agent. This feature is probably not corrected by the HGF treatment, including the co-administration of G-CSF and GM-CSF. The dexamethasone administred preferentially to MM patients should also stimulate hematopoïesis in synergy to HGFs. The role of glucocorticoids, espacially dexamethasone, on proliferation of hematopoïetic progenitors has been suggested in vitro.52, 53 This particular benefit observed in MM patients and the fact that we use a concomitant administration, may represent the differences observed between the literature and our results.

The timing of harvest is one of the major parameter of successfull yield of CD34+ PBSC. The criteria used to define the optimal harvesting day was the peripheral WBC or mononuclear cells. But, these criteria are associated with a weak correlation to the CD34 cell amounts in the leukapheresis product.54 At the present time, the absolute number of circulating CD34 cells is correlated closely with the CD34 cell yield of the corresponding leukapheresis product. The usual critical threshold of circulating CD34 cells as indicator of onset of cytapheresis is 40 CD34 cells per μl.55 But, this method does not provide an estimation of CD34 cells collection. We use an estimation of the CD34 PBSC yield determined just before each cytapheresis to give directly an estimation of the final collect of CD34+ PBSCs. In addition to PBSC mobilization, the mobilization of tumor cells remains likely. The role of these residual tumor cells in the graft for the relapse after transplantation is still controversial.56, 57 In this study, we evaluate the tumors cells co-mobilization in MM patients. We have shown no trend to mobilize more tumoral cells with G-CSF plus GM-CSF regimen than that observed with G-CSF alone.

As expected, no clear benefit appears in terms of engraftment, transfusion cost according to HGF regimen. The median number of CD34 cells infused to the patients in all treatment group is comparable (3.8–4.32 × 106 CD34+/kg). The majority of the patients having MM or breast cancer were scheduled to receive a double intensive chemotherapy followed by a CD34 immunoselected graft rescue. This schedule was possible more frequently in the G-CSF plus GM-CSF mobilization regimen (93–100% of patients) than in G-CSF group (53 to 64% of patients).

Future insight of graft manipulation as high level tumor cell purging, genetic manipulation, monocytes mobilization in view of anti-tumoral vaccination by dendritic cells, ex vivo expansion will invariably result in loss of a substantial fraction of hematopoietic progenitor cells harvested, a situation that needs to collect higher number of hematopoietic stem cells. The combination of cytokines, including GM-CSF may represent a particular interesting combination, in addition to the effects of GM-CSF on the immune system, as demonstrated in vaccination programs.