A Prospective Randomized Trial of Granulocyte Colony-Stimulating Factor Therapy After Autologous Blood Stem Cell Transplantation in Adults

Home , Autotransplantation

Bone Marrow Transplantation, (1999) 24, 601–607  1999 Stockton Press All rights reserved 0268–3369/99 $15.00 http://www.stockton-press.co.uk/bmt A prospective randomized trial of granulocyte colony-stimulating factor therapy after autologous blood stem cell transplantation in adults

E Ojeda, J Garcia-Bustos, MJ Aguado, R Arrieta, E Quevedo, V Jimenez Yuste, M Canales and F Hernandez-Navarro

Servicio de Hematologı´a y Hemoterapia, Hospital ‘La Paz’, Madrid, Spain

Summary: Despite the large number of PBSC autotransplants performed nowadays and the ample experience using recombi- In order to assess the potential clinical benefit of filgra- nant cytokines in this particular setting, very few published stim (G-CSF) after peripheral blood stem cell (PBSC) randomized studies explore the real clinical impact of the autotransplantation a randomized study was begun in use of filgrastim (G-CSF) in PBSC autotransplantation. our center in July 1997: 62 patients were involved (30 Only five recent reports have analyzed in a prospective received filgrastim after PBSC infusion and 32, the con- manner the possible benefit of the administration of G-CSF trol group, received no cytokines). All were adults after PBSC autotransplantation since 1997 (one of them (median 40 years, range 18–65). Patients with one of used the glycosylated form, lenograstim).1–5 In four of these three different pathologies were recruited: 28 had studies the authors recommended the use of G-CSF on the advanced breast carcinoma, 23 had lymphomas (12 basis of the favourable clinical results. A recent Japanese Hodgkin’s disease and 11 non-Hodgkin’s lymphoma) study performed in children has questioned this support for and 11 had de novo AML. All of them were transplanted the use of G-CSF and reported a limited clinical benefit using myeloablative chemotherapy conditioning regi- from this cytokine in PBSC autotransplantation.5 mens. G-CSF was administered subcutaneously from This article summarizes our clinical experience using G- day +5 in the treated group at a dose of 5 ␮g/kg body CSF (filgrastim) after PBSC transplantation in adults, weight/day. The numbers of CD34+ and mononuclear reporting the results of a prospective and randomized trial (MNC) cells infused were similar in each group. Only performed in our hospital in the past months. minor differences regarding the use of G-CSF could be inferred from the analysis of the data. Faster granulocyte engraftment was evident in the treated group Materials and methods (mean of 10 vs 12 days to achieve Ͼ0.5 × 109/l granulo- = cytes, P 0.0008), without differences in incidence and Patients severity of infections, days of fever or duration of antibiotic treatment between groups. There was slightly Between July 1997 and November 1998, 62 patients were slower platelet engraftment (mean of 15 days in the referred to our service and recruited for the study. The group with G-CSF vs 12 days in the other group to ach- patients had different neoplastic diseases: 28 with breast ieve Ͼ20 × 109/l platelets, P = NS) in this series, but carcinoma in advanced stages, 12 with Hodgkin’s disease there were no differences in incidence and severity of (HD), 11 with non-Hodgkin lymphomas (NHL) and 11 with haemorrhage or platelet transfusion support. Consider- de novo AML who were autotransplanted using PBSC. ing the economical costs, the median expenditure per Clinical data at the time of transplantation and the charac- inpatient stay was Eur5961 (range Eur4386–Eur17186) teristics of both treatment groups (with and without G-CSF) in the G-CSF group compared with Eur5751 (range are given in Table 1. There were no statistical differences Eur3676–Eur15640) in the control group (P = 0.47). for age, sex, diagnosis, remission status, interval between From our data it could be concluded that for adult diagnosis and autotransplantation, number of mononuclear patients transplanted with PBSC there is no clear ben- and CD34+ cells infused by body weight, platelet count eficial impact of post-infusion G-CSF administration. before transplantation or conditioning employed between Keywords: randomized study; G-CSF; filgrastim; PBSC the different groups. The control group had more irradiated autotransplantation patients. The time period analyzed in this study ran from PBSC infusion to discharge. The median follow-up after transplantation was 9 months (range 3.5–19) for the whole series. Multiple variables in the patients’ post-transplantation evolution were recorded and analyzed in order to Correspondence: Dr E Ojeda, Servicio Hematologıá y Hemoterapia, Hos- reveal any possible difference between the groups. PBSC pital La Paz, Paseo de la Castellana 261, 28046 Madrid, Spain were collected from patients after intensive chemo- and/or Received 8 March 1999; accepted 26 April 1999 radiotherapy. Harvest was scheduled at least 1 month after G-CSF after PBSC transplantation E Ojeda et al 602 Table 1 Patient characteristics at transplantation Conditioning regimens

Characteristic G-CSF Control P value Patients received different marrow-ablative regimens, (n = 30) (n = 32) depending on their diagnoses. The patients diagnosed as having AML received busulfan and cyclophosphamide as Age (years) conditioning.8 Carboplatin, cyclophosphamide and thiotepa Median 38 41 0.88 were used in women with breast carcinoma.9 The BEAC Range 18–60 19–64 regimen was followed in the lymphoma cases.10 Sex The lowest numbers for peripheral blood counts after Female 21 19 0.54 conditioning ranged between 0 and 0.7 × 109/l (median 0.1) Male 9 13 for leucocytes and 2 to 27 × 109/l (median 11) for platelets. Diagnosis No differences were observed between the two study Breast carcinoma 15 13 0.21 groups. Lymphoma 8 15 AML 7 4 Remission status Clinical management and support measures Complete 24 23 0.44 Partial 4 8 Before myeloablative therapy, the patients were nursed in No remission 2 1 single rooms using simple reverse isolation measures. Previous radiotherapy PBSC were reinfused after thawing on day 0 (considering Yes 2 10 0.034 as day 0 the first day of PBSC infusion). RBC concentrates No 28 22 were transfused when the hemoglobin value was less than Time from diagnosis to transplantation (months) 9 g/dl and random platelet transfusions were employed Median 8.5 8.9 0.12 when platelet count was less than 15 × 109/l or bleeding Range 4–45 6–146 had occurred. No. of aphereses required Total parenteral nutrition (TPN) was initiated on day +2 Median 2 2 0.63 Range 1–4 1–6 in all patients and discontinued when oral nutrition was resumed. All patients received the same prophylactic anti- MNC ×108/kg infused Median 11 11 0.79 biotherapy, consisting of acyclovir, fluconazol and cotri- Range 4–24.6 4.4–20.3 moxazol. Intravenous broad-spectrum antibiotic treament CD34+ ×106/kg infused was initiated when the axillary temperature was higher than Median 3.52 3.92 0.59 38.5°C with a third-generation cephalosporin, amikacin and Range 1.6–25.7 1.8–10.9 vancomicin. Amphotericin B (liposome complex, Abelcet; Esteve Hospital, Barcelona, Spain) was added after 3 days MNC = mononuclear cells. of persistent fever. In the treated group, filgrastim (5 ␮g/kg body weight) (Amgen, F Hoffman-La Roche), administered daily as a the last chemotherapy session. PBSC were mobilized using single s.c. injection, was initiated on day +5, and continued rhG-CSF (filgrastim ICD; Amgen, F Hoffman-La Roche, until the absolute neutrophil count was higher than Basel, Switzerland); 10 ␮g/kg/day were administered s.c. 1.5 × 109/l. The day of initiation was decided after a pre- on the 4 days prior to harvest and for the duration of the vious randomized study in our centre demonstrated that the procedure. Harvesting was scheduled to terminate when the results of delayed administration were similar to those of total number of CD34+ cells obtained reached 2.5 × 106/kg early filgrastim administration (reported by others).11,12 The body weight. This threshold was achieved in most of the control group received no colony-stimulating factors. patients. The procedures for harvesting and cryopreserv- Discharge occurred when the patient was engrafted, free ation were as previously described.6 All patients gave infor- of fever, independent of RBC transfusion or any other med written consent. Each patient was randomly allocated eventual complication (complete platelet engraftment was to one of the study groups at the beginning of the transplan- not necessary for discharge). The decision was jointly tation procedure prior to conditioning. agreed upon by the three physicians (always the same) An estimate of the stem cell number in the inoculate was responsible for the clinical unit. determined by analysis of the CD34+ mononuclear cells, employing a direct immunofluorescence flow cytometry Cost analysis assay.7 Cells were thawed by rapid heating of the cryopreserv- Costs for filgrastim (if appropriate), i.v. antibiotics, platelet ation bags through immersion in a 40°C water bath. The and red cell concentrates, parenteral nutrition and occupied- bags were thawed one by one, and a new bag was not bed days from day 0 were totalled for each patient. Prophy- thawed until the last bag had been completely infused. lactic antibiotics were not included. The calculated unit cost Before infusion, 3 ml of the thawed solution was obtained of an occupied-bed day in our transplantation unit for the for viability, cytologic and microbiological control tests. year 1998 was Eur232.60 and includes personal expenses, Viability was assessed by trypan blue exclusion. Immedi- laboratory, radiologic and other special tests. The total cost ately after thawing, the PBPC were rapidly infused via a for each patient was found by summing the number of Hickman’s central line. resources consumed multiplied by their respective unit G-CSF after PBSC transplantation E Ojeda et al 603 prices. Only transplantation in-patient time (from day 0 to G-CSF group than for the control group (median 10 days, discharge) was considered. Incidentals after the first dis- range 7–19, and median 12 days, range 9–20, respectively). charge were not considered. Charges for conditioning The median lapse time in aplasia (days between less than chemotherapy, cell mobilization and harvest and cryopre- 500 granulocytes and more than 500 granulocytes) was 9 servation were not considered in this analysis. days for the G-CSF-treated patients and 10 days for control patients. In the G-CSF group, the median time required for Statistical analysis recovery of more than 20 × 109/l platelets was 13 days (range 10–47), slightly slower than for the control group All data were collected using a custom-adapted File-Maker (median of 11 days, range 5–33). This difference was more Pro (Macintosh, Apple Computer, Cupertino, CA, USA) + evident but not significant with the number of patients database and analysed with Stat-View SE Graphics × 9 software (Macintosh, Apple Computer). The differences available when the time to recover more than 50 10 /l between the medians for different variables were analyzed platelets was considered (median of 18 days in G-CSF using the Mann–Whitney U test (quantitative variables) or group and 14 in the controls). chi-square test (qualitative variables). Actuarial survival The number of patients that received sub-optimal num- + Ͻ × 6 curves were constructed according to the method of Kaplan bers of CD34 cells (fixed at 2.5 10 /kg) was the same = and Meier. Survival was calculated from the date of day 0 in the two groups: six patients in each group (P 0.2). The engraftment kinetics for these patients with sub-optimal to the date of death from any cause or end of the study + period (February 1999). numbers of CD34 cells was similar in both groups and to the patients with optimal numbers of CD34+ cells infused. Complete engraftment (defined as Ͼ1.5 × 109/l granulo- Results cytes, Ͼ10 g/dl Hb and Ͼ50 × 109/l platelets without stimulating factors or transfusional support) was evaluated on Engraftment kinetics and haematopoietic recovery day +100 post-transplant and was achieved by the majority Data on engraftment kinetics and haematopoietic recovery of patients in both groups. In the G-CSF-treated group, two are shown in Table 2. patients did not engraft: a woman diagnosed with AML and Granulocyte recovery over 0.5 × 109/l was faster for the transplanted in partial remission eventually relapsed and

Table 2 Engraftment kinetics and haematopoietic recovery at day +100

G-CSF (n = 28)a median (range) Control (n = 31)b median (range) P value

Days to reach Ͼ0.5 × 109/l granulocytes 10 (7–19) 12 (9–20) 0.0008 Days to reach Ͼ1 × 109/l granulocytes 11 (9–20) 13 (10–26) 0.0003 Days to reach Ͼ20 × 109/l platelets 13 (10–47) 11 (5–33) 0.11 Days to reach Ͼ50 × 109/l platelets 18 (12–59) 14 (10–93) 0.069 Days with less than 0.5 × 109/l granulocytes 9 (6–11) 10 (7–17) 0.0001

G-CSF (n = 27)a median (range) Control (n = 29)b median (range) P value

Hb at day +100 (g/dl) 12.4 (9.9–15.2) 12.3 (8.3–15.1) 0.59 Leucocytes at day +100 (×109/l) 4.6 (1.8–7.1) 4.5 (1.7–8.6) 0.95 Platelets at day +100 (×109/l) 165 (62–302) 175 (26–338) 0.71 aOne patient died before day +100 and two others did not achieve complete engraftment. bOne patient died before day +100, one underwent myelodysplastic syndrome before day +100 (with a previous complete engraftment) and one did not achieve complete engraftment.

Table 3 Clinical outcome

G-CSF Control P value (n = 30) (n = 32) median (range) median (range)

Febrile days (Ͼ38.5°C) 1 (0–10) 1 (0–14) 0.24 Days on i.v. antibiotics 8 (0–44) 8 (0–23) 0.97 Days on i.v. amphotericin 0 (0–29)a 0 (0–23)a 0.79 Red blood cell concentrates 4 (0–18) 3 (0–17) 0.81 Platelet cell concentrates 18 (6–58) 14 (0–137) 0.17 Days with i.v. morphine 0 (0–11)b 0 (0–13)b 0.93 Days on G-CSF treatment 6 (4–11) — — Days on TPN 10 (4–23) 10 (2–16) 0.76 Days until discharge (from 0) 15 (11–51) 17 (12–37) 0.46 aTen patients required i.v. amphotericin in each group. bEleven patients required i.v. morphine in the G-CSF group and 12 patients in the control group. G-CSF after PBSC transplantation E Ojeda et al 604 died 3 months after transplantation; the other, a man diag- Table 4 Mucositis and G-CSF nosed with refractory HD, did not achieve erythrocyte or platelet engraftment and was diagnosed with secondary WHO grading for mucositis G-CSF Control myelodysplastic syndrome; he died after an allogeneic total (%) total (%) PBSC transplantation 1 year after the ﬁrst transplantation. In the control group without G-CSF, only one patient did No mucositis 4 (13) 6 (19) (grade 0) not achieve a complete engraftment: a woman with AML Soreness/Erythema 7 (23) 9 (28) transplanted at partial remission who has achieved no (grade 1) erythrocyte or platelet engraftment and is alive, but trans- Erythema, ulcers, can eat solids 9 (31) 9 (28) fusion-dependent. In the same group, a woman with HD (grade 2) Ulcers, requires liquid diet only 4 (13) 3 (9) achieved complete engraftment, displayed progressive pan- (grade 3) cytopenia 2 months after transplantation and was diagnosed Alimentation not possible 6 (20) 5 (16) with secondary myelodysplastic syndrome. (grade 4)

P (chi-square) = 0.93. Hospital course and toxicity after PBSC autograft

No patients died during the transplantation process and all Table 5 Infectious complications left the hospital after the procedure. Two patients died before day +100 post-infusion, one in each group, although Infections grades G-CSF Control both had achieved complete engraftment: the first, in the total (%) total (%) G-CSF group, was a patient with refractory HD. He died of aspergillosis pneumonia 2 months after the PBSC No infections 11 (37) 6 (19) (grade 0) infusion. The other, in the control group, died due to relapse One isolated febrile episode 10 (33) 16 (50) of his NHL 3 months after day 0. At the time of analysis, (grade 1) six patients have died: four in the G-CSF group at 2, 3.4, Two or more febrile episodes 6 (20) 6 (19) 3.8 and 12.3 months after PBSC infusion (one from asper- (grade 2) Pneumonia 2 (7) 3 (9) gillosis, two from relapse of their AML and the last one (grade 3) during a later allogeneic transplantation); and two in the Septic shock or respiratory distress 1 (3) 1 (3) control group from relapse of their disease at 3 and 11 syndrome (grade 4) months after infusion (with NHL and HD, respectively). No significant statistical differences in survival were observed P (chi-square) = 0.55. between the two groups (median of 8.5 for the G-CSF group vs 10.3 months for the control group, P = 0.21). Severe haemorrhagic complications appeared in eight No differences in terms of days with TPN, i.v. anti- patients: in the G-CSF group one patient experienced an biotics, amphotericin B, or transfusional support were evi- autolimited haematuria due to prostatitis, one had a severe dent in our series between the two groups (Table 3). Days epistaxis and a third displayed oral mucosal and bilateral until discharge were similar between the two groups, with- conjunctive haemorhage. In the control group, two patients out statistically relevant differences (mean of 18.6 Ϯ 1.6 experienced menorrhagia, one had cutaneous purpura, a (s.e.) days for the G-CSF group and 18.5 Ϯ 1 (s.e.) days fourth displayed severe epistaxis and a fifth (without plate- for the control group, P = 0.46). let engraftment) experienced long-lasting haemorrhagic Because of the possible impact of the use of G-CSF on cystitis that ultimately resolved. There were no differences the incidence and severity of mucositis, the relevance of between the two groups in the number of transfused platelet this complication was investigated in our patients by means concentrates (Table 3). of the WHO scale for mucositis13 and the consumption of i.v. morphine (directly related to the duration of the most Cost analysis severe forms of mucositis) was recorded. No relevant differences in incidence, severity nor morphine consumption Taking into account the costs in our institution of occupied- were evident in our series (Tables 3 and 4). bed days from day 0, antimicrobials used, G-CSF usage, total parenteral nutrition and transfusional support, the Infectious and haemorrhagic complications median expenditure per inpatient stay was Eur5961 (range Eur4386–Eur17186) in the G-CSF group compared with In order to evaluate the incidence and severity of infectious Eur5751 (range Eur3676–Eur15640) in the control group complications in our series, an original grading score sys- (P = 0.47) (Table 6). The two groups of patients had the tem was used (Table 5). It showed no significant differ- same economical expenses for all the parameters investi- ences between the two study groups. Only a few patients gated (excluding the G-CSF). in either group experienced severe or life-threatening infections (grades 3 and 4, 10% and 12% of patients, respectively). In all cases the infections resolved without Discussion sequelae. The treatment with G-CSF did not imply a lower consumption of i.v. antibiotics and/or amphotericin or The major benefit of using PBSC instead of bone marrow fewer febrile days for these patients (Table 3). for autotransplantation, in addition to the advantage of a G-CSF after PBSC transplantation E Ojeda et al 605 Table 6 Cost analysis

Variables G-CSF (n = 30) Control (n = 32) P mean Ϯ s.e. mean Ϯ s.e. (Mann–Whitney U) median (range) median (range)

Occupied-bed days (from day 0) 4362 Ϯ 387 4317 Ϯ 238 0.23 3372 (2558–11862) 3954 (2791–8606) Non-prophylactic antibiotics 711 Ϯ 140 601 Ϯ 76 0.90 501 (0–2760) 501 (0–1443) Amphotericin 351 Ϯ 116 292 Ϯ 93 0.88 0a (0–2828) 0a (0–2242) TPN 511 Ϯ 38 483 Ϯ 32 0.90 480 (192–1105) 480 (96–769) RBC concentrates 423 Ϯ 70 382 Ϯ 61 0.69 339 (0–1529) 269 (0–1457) Platelet concentrates 572 Ϯ 63 611 Ϯ 135 0.23 507 (105–1635) 399 (0–3738) G-CSF 553 Ϯ 37 — — 508 (240–1016)

Total costs 7449 Ϯ 645 6689 Ϯ 480 0.47 5961 (4386–17186) 5751 (3676–15640)

All values are expressed in Eur. aTen patients required i.v. amphotericin in each group. TPN = total parenteral nutrition. simpler and less invasive method of obtaining progenitor carefully examined and we cannot demonstrate a clear cells, is the faster haematopoietic recovery. There is contro- benefit for G-CSF on the incidence and severity of infec- versy as to whether the use of G-CSF after myeloablative tions or antibiotic support. With regard to platelet conditioning also has a remarkably beneficial impact on engraftment, most of the studies observed no differences in PBSC autotransplantation. A few recent studies have days to recovery of more than 20 × 109/l platelets or to explored this aspect in a prospective and randomized man- being transfusion independent. Only the study of Kawano ner (Table 7). First of all it should be noted that extracting et al and our study note an unfavourable trend in platelet relevant and exactly comparable information from these recovery when G-CSF is used after transplantation; this has different studies is difficult because of the many design also been noted in another retrospective study on allogeneic variables. Diagnosis, age of patients (children in the study transplantation.14 This effect is more evident in our series of Kawano et al5), conditioning and state of remission at when the days until reaching more than 50 × 109/l platelets transplantation vary considerably in all these reports. More- is analysed, and almost, reaches statistical significance over, the type of G-CSF and schedule of treatment are also (P = 0.069). Nevertheless, and this is the most relevant diverse: Linch et al2 used a glycosylated form of G-CSF finding, this delay was not reflected in increased platelet in their study and McQuaker et al4 employed a low dose transfusional support or increased haemorrhagic compli- of G-CSF. The time of initiating filgrastim therapy also var- cations in the G-CSF-treated group. In our series we cannot ies from the same day as PBSC infusion to the day +5. demonstrate a beneficial impact from the use of G-CSF on Added difficulties in some multicentric studies are the hospitalization time (mean of 18 days in each group). Ana- many centres involved with a comparativly low number of lysing the other reports, only three of them show that the patients enrolled in each one (as is the case for the studies use of G-CSF seems to shorten hospitalization (Table 7). of Linch et al and Kawano et al), thus decreasing the value Nevertheless, these data are misleading in the study of of possible differences among some variables regarding the Linch et al, in which the number of centres involved and use of G-CSF; particularly the days of hospitalization and the nonexistence of common criteria for patient discharge, the cost analysis. Despite these problems, all these publi- diminish the potential relevance of their results. Also, the cations, and our study is not an exception, demonstrate that studies of Lee et al3 and McQuaker et al, each performed the addition of G-CSF significantly accelerates granulocyte in a single centre, do not have enough patients to confirm recovery after PBSC rescue (Table 7). Unfortunately, this the possible positive clinical effect of G-CSF without finding does not imply any clear benefit for the use of G- doubt. CSF in terms of a lower incidence of infections, days of Only three studies perform a cost analysis that demon- fever or antibiotic consumption. Three studies1,3,4 showed strates a potentially favourable economic impact of the use that the administration of filgrastim to autotransplanted of G-CSF after PBSC transplantation. In contrast with the patients resulted in a decreased use of intravenous anti- reports of Lee et al and McQuaker et al, our results do not biotics or amphotericin. In our study, this aspect has been support a clear advantage from the use of filgrastim. In G-CSF after PBSC transplantation E Ojeda et al 606 marginal 36 Yes 9 Yes vs vs support amphotericin Less transfusional to G-CSF to G-CSF Yes favourableYes favourable Abs 28 Less use of Yes 15 16 17 Yes No differences No 14 No Abs 5 15.4 No No differences Yes, but vs vs vs vs vs 12 13 analysed 13 /l 9 1611 Not 15 No No differences No 10 13 9 10 13.5 × vs vs vs vs vs 20 differences Ͼ /l reach support of G-CSF 9 14 12 1212 22 13 13.2 No 10 12.5 10 12.5 9 × vs vs vs vs vs vs 0.05). 10 11 10 0.5 9 10 9.7 Ͻ Ͼ P 5 gg Lymphoma Lymphoma +

␮ ␮ g/m2 ALL, neuroblastoma and 1 other 5 5 lymphoma and AML No differences 1 1 myeloma g/m2 Lymphoma and multiple g/kg Breast carcinoma g/kg Breast carcinoma ␮ + + + + + ␮ ␮ ␮ beginning platelets involved and day of granulocytes 80 2 5 patients centres dose/day Randomized prospective studies using G-CSF after PBSC transplantation 1997 38 1 50 1998 63 15 300 4 1 1997 90 9 263 5b i.v. antibiotics. 2a 1997 23 1 300 = 3 Children. Glycosylated G-CSF (lenograstim). a b study 1999 1997 0 and Kawano Table 7 Author/Ref./Year No.Hornedo No.Linch G-CSFMcQuaker Diagnosis Days to reach DaysValues to are expressed as median. Shaded values express statistically Days significant stay differences ( Cost analysis Infections and Advantages Ojeda, this 62 1 5 Lee Abs G-CSF after PBSC transplantation E Ojeda et al 607 contrast, the use of G-CSF in our patients slightly raises 4 McQuaker IG, Hunter AE, Pacey S et al. Low-dose filgrastim the costs of transplantation. significantly enhances neutrophil recovery following autolog- In conclusion, whilst this study is not a placebo-con- ous peripheral-blood stem-cell transplantation in patients with trolled study, our data indicate an absence of a clear clinical lymphoproliferative disorders: evidence for clinical and econ- benefit from the post-PBSC autotransplantation use of G- omic benefit. J Clin Oncol 1997; 15: 451–457. 5 Kawano Y, Takaue Y, Mimaya J et al for the Japanese Coop- CSF in adults, and this accords with another recently pub- 5 erative Study Group of PBSCT. Marginal benefit/disadvantage lished study in children. Nevertheless, we feel a larger of granulocyte colony-stimulating factor therapy after autolog- study is required to confirm this important question or deny ous blood stem cell transplantation in children: results of a this conclusion. prospective randomized trial. Blood 1998; 92: 4040–4046. 6 Hernańdez-Navarro F, Ojeda E, Arrieta R et al. Single-centre experience of peripheral blood stem cell transplantation using cryopreservation by immersion in a methanol bath. Bone Mar- Acknowledgements row Transplant 1995; 16: 71–77. 7 Siena S, Bregni M, Brando B et al. Flow cytometry for clinical We thank Maria Moreno and the rest of the personnel from the estimation of circulating hematopoietic progenitors for auto- Bone Marrow Transplantation Unit of La Paz, since without their logous transplantation in cancer patients. Blood 1991; 77: invaluable contribution, this study would have not been possible. 400–409. We are indebted to CF Warren for help in preparing the manu- 8 Tutschka PJ, Copelan EA, Klein JP. Bone marrow transplan- script. tation for leukemia following a new busulfan and cyclophosphamide regimen. Blood 1987; 70: 1382–1388. 9 Eder JP, Elias A, Shea TC et al. A phase I–II study of cyclophosphamide, thiotepa, and carboplatin with autologous bone References marrow transplantation in solid tumor patients. J Clin Oncol 1990; 8: 1239–1245. 1 Hornedo J, Sola C, Solano C et al. Multicentric, randomized 10 McMillan AK, Goldstone AH. Autologous bone marrow and prospective study about the role of Filgrastim (G-CSF) transplantation for non-Hodgkin’s lymphoma. Eur J Hematol post-transplantation of hematopoietic progenitors cell mobil- 1991; 46: 129–135. ized with G-CSF in patients with breast cancer. Sangre 1997; 11 Faucher C, Le Corroller AG, Chabannon C et al. Adminis- 42 (Suppl. 1): 54–56. tration of G-CSF can be delayed after transplantation of auto- 2 Linch DC, Milligan DW, Winfield DA et al. G-CSF after per- logous G-CSF primed blood stem cells: a randomized study. ipheral blood stem cell transplantation in lymphoma patients Bone Marrow Transplant 1996; 17: 533–536. significantly accelerated neutrophil recovery and shortened 12 Ojeda E, Garcia-Bustos J, Aguado MJ et al. G-CSF after time in hospital: results of a randomized BNLI trial. Br J PBSC transplantation. Br J Haematol 1998; 101: 594–595. Haematol 1998; 99: 933–938. 13 Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting 3 Lee SM, Radford JA, Dobson L et al. Recombinant human results of cancer treatment. Cancer 1981; 47: 207–214. granulocyte colony-stimulating factor (filgrastim) following 14 Bernstein SH, Nademanee AP, Vose JM et al. A multicenter high-dose chemotherapy and peripheral blood progenitor cell study of platelet recovery and utilization in patients after mye- rescue in high-grade non-Hodgkin’s lymphoma: clinical bene- loablative therapy and hematopoietic stem cell transplantation. fits at no extra cost. Br J Cancer 1998; 77: 1294–1299. Blood 1998; 91: 3509–3517.