Leukemia Research 29 (2005) 159–163

Bone marrow CD34+ cell count is predictive for adequate peripheralprogenitor cell collection

Alessandro Corsoa,∗, Marzia Varettonia, Silvia Mangiacavallia, Patrizia Zappasodia,Catherine Klersyb, Chiara Rusconia, Emile Coluccia, Angela Lorenzia,

Daniela Trolettia, Erica Consensia, Mario Lazzarinoa

a Division of Hematology, IRCCS Policlinico San Matteo, University of Pavia, Via Golgi 2, 27100 Pavia, Italyb Division of Biometry and Clinical Epidemiology, IRCCS Policlinico San Matteo, University of Pavia, Via Golgi 2, 27100 Pavia, Italy

Received 26 April 2004; accepted 10 May 2004Available online 5 August 2004

Abstract

nitor cells( er of PBPCy myeloidl ilisationw s 833( l target of 4× s than 41 ilisation,o aB©

K

1

pam(cmtfc

ave

ili-diest. Aentstorselyad aeasepro-uallybur-

apy,ents,

0d

Several studies have investigated the influence of clinical and biological variables on mobilisation of peripheral blood progePBPCs). The aim of this study was to evaluate the role of steady-state bone marrow (BM) CD34+ cells as a predictive parametield. We studied 90 patients with multiple myeloma (MM) (41 patients), non-Hodgkin’s lymphoma (NHL) (25 patients) or acuteeukaemia (AML) (24 patients), mobilised with chemotherapy and growth factor. The median time from first treatment to mobas 5 months. Only one patient was previously exposed to alkylating agents. The median BM CD34+ count at mobilisation wa�l−1

range:1.4–15.540) corresponding to 1.51% of mononuclear cells (range: 0.02–8.6). Sixty-six patients (73%) reached the optima106 kg−1 CD34+ cells with 1 (18 patients), 2 (42 patients) or 3 leukaphereses (6 patients). Eleven patients (12%) mobilised les×

06 kg−1 CD34+ cells and 13 (15%) failed mobilisation. Among the laboratory and clinical parameters evaluated at the time of mobnly BM CD34+ count was a predictive factor for adequate collection (P = 0.04), particularly in MM patients (P = 0.003). In this setting,M concentration of CD34+ cells lower than 66�L−1 was associated with a higher probability of inadequate collection.2004 Elsevier Ltd. All rights reserved.

eywords:CD34+ cells; Lymphoma; Multiple myeloma; Peripheral stem cell collection; Transplant

. Introduction

High-dose chemotherapy with autologous stem cell trans-lantation has proved to be an effective therapy in patientsffected by multiple myeloma[1,2], lymphomas[3] or acuteyeloid leukaemia[4]. Peripheral blood progenitor cells

PBPCs) are usually preferred to bone marrow (BM) as stemell source due to the easier collection and faster engraft-ent [5]. Hematopoietic recovery following myeloablative

herapy depends on the number of CD34+ progenitor cells in-used. A dose ranging from 2× 106 to 5× 106 kg−1 CD34+ells seems adequate to ensure a rapid and sustained engraft-

∗ Corresponding author. Fax: +34 382 502250.E-mail address:a.corso@smatteo.pv.it (A. Corso).

ment [6–10], and however no significant advantages hbeen observed using more than 5× 106 kg−1 stem cells[11].Unfortunately, a variable proportion of patients fail mobsation or yield insufficient numbers of PBPC. Several stuhave tried to identify factors affecting the CD34+ harveslong therapeutic history, prior exposure to alkylating agor radiation therapy, and bone marrow infiltration are facmore frequently reported to affect mobilisation negativ[6,8,12–20]. Patients reported in these studies, anyway, hlong previous therapeutic history or present extensive disat the time of mobilisation. The more recent high-dosetocols are applied earlier in the course of disease, and usafter an induction therapy aimed at reducing the diseaseden. In this setting of patients with a short period of thera low disease burden and no exposure to alkylating ag

145-2126/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.oi:10.1016/j.leukres.2004.05.018

160 A. Corso et al. / Leukemia Research 29 (2005) 159–163

alternative predictive parameters of adequate stem cell col-lection are needed. Few authors have investigated the relationbetween bone marrow CD34+ cell count before mobilisationand the peripheral stem cell yield, reporting contrasting re-sults[21,22].

We performed this study in order to better define the roleof steady-state BM CD34+ cell count as predictive parameterof mobilisation.

2. Patients and methods

2.1. Patients

We retrospectively studied 90 patients affected by hemato-logic malignancies who underwent PBPC mobilisation fromSeptember 1997 to April 2003. The characteristics of patientsare shown inTable 1. Forty-nine were males and 41 femaleswith a median age of 51 years (range: 26–65); 41 patientshad multiple myeloma (MM), 25 non-Hodgkin’s lymphoma(NHL) and 24 acute myeloid leukemia (AML); the mediantime from the first treatment to mobilisation was 5 months(range: 1–212); only one patient was previously exposed toalkylating agents.

2.2. PBPC mobilisation

as-s ereuw dw );3 ex-a tinu-o -s -p edi pa-t1 1a am-

TC

s

NM /18M 3 (36–6M 12 (8.1 16.3)M (2.6– .4)M 8 (68– )B (7.3)M 3 (1–2P 1 (2.4S

1.6 (0 2–8.6)9 (4.4– 540)

ethasone 20 mg/day on days 1–3) was used in 5/25 (20%)patients with NHL. Granulocyte colony-stimulating growthfactor (G-CSF) was given at the dose of 5�g/kg day starting48 h after the end of chemotherapy until the day of last leuka-pheresis. PBPC collection was started when peripheral bloodCD34+ cell count reached the threshold of 20�L−1. Leuka-pheresis was performed using a continuous-flow blood cellseparator Spectra (COBE BCT) processing a total volume of2–3 blood mass volumes per leukapheresis. Leukaphereticprocedures where repeated, whenever indicated, in order tocollect at least 4× 106 kg−1 CD34+ cells. Patients who failedmobilisation or collected less than 4× 106 kg−1 CD34+ cellswere considered poor mobilisers.

2.3. Flow cytometry

Aliquots of both bone marrow and leukapheretic productswere incubated with phycoerythrin-conjugated anti-CD34monoclonal antibody HPCA-2 (Becton Dickinson, San Jose,CA) for 15 min at 20◦C, then lysed with a lyse-no-wash stan-dard assay and finally incubated for 10 min at 4◦C. Cells wereprocessed with a FACSCalibur analyser (Becton-Dickinson).Data acquisition was performed with CellQuest 3.1 software(Becton Dickinson).

2

(S.D.)o ari-a oricalv ng tog1 heirr t tom PLTa per-c cells�

torsf ter-

All patients were mobilised with chemotherapy inociation to G-CSF. Different mobilising regimens wsed according to underlying disease (Table 2): all patientsith AML and 19/25 (76%) with NHL were mobiliseith high-dose Ara-C (1–3 g/m2 every 12 h for 2–6 doses0/41 patients (73%) affected by MM received D-CEP (dmethasone 40 mg daily for 4 days and 4-day conus infusion of cyclophosphamide 400 mg/m2 day, etopoide 40 mg/m2 day and cisplatin 10 mg/m2 day); cyclophoshamide 6 g/m2 given in two fractions over 24 h was us

n the remaining 11 patients (27%) with MM and in oneient with NHL (4%); IPAD (idarubicin 12 mg/m2 on day, continuous infusion of cisplatin 50 mg/m2 day on daysnd 2, Ara-C 4 g given in two fractions on day 3, dex

able 1linical and hematological characteristics of patients

MM

o. of patients 41ale/female 23edian age (range) 5ean Hb value at mobilisation (g/dl) (range)ean WBC count at mobilisation (×109 l−1) (range) 6.3ean PLT count at mobilisation (×109 l−1) (range) 24M disease infiltration≥50% (number of cases, %) 3edian time from first treatment, to mobilisation, months (range)rior exposure to alkylating agents (number of cases, %)teady-state BM CD34+ cell countPercentage of mononuclear cells (range)Concentration (�L) (range) 45

NHL AML Whole case serie

25 24 9015/10 11/13 49/41

5) 49 (26–59) 47 (28–59) 51 (26–65)–15.4) 13.2 (8.8–16.3) 11.7 (5.6–15.2) 12.2 (5.6–14.4) 5.96 (3.3–11.3) 5.9 (0.96–11.6) 6.1 (0.96–14403) 213 (129–316) 231 (23–605) 234 (23–605

0 0 3 (3.3)3) 17 (4–212) 5 (2–31) 5 (1–212)) 0 0 1 (1.1)

.02–3.9) 0.97 (0.02–3.2) 2.67 (0.04–8.6) 1.51 (0.02808) 475 (1.4–4005) 1844 (27–15540) 833 (1.4–15

.4. Statistical analysis

Data are presented as mean and standard deviationr median and inter-quartile range (IQR) for continuous vbles, and as frequencies and percentages for categariables, both for the whole case series and accordiroups defined based on CD34+ harvest (≤ versus >4×06 kg−1). The following parameters were analysed for telation to CD34+ harvest: time from the first treatmenobilisation, bone marrow disease infiltration, Hb value,nd WBC count, BM CD34+ count expressed either asentage of mononuclear cells or as number of CD34+L−1.Logistic modeling was used to identify prognostic fac

or adequate harvest, while controlling for diagnosis. In

A. Corso et al. / Leukemia Research 29 (2005) 159–163 161

Table 2Characteristics of PBPC collection

MM/NHL/AML (90 patients) MM (41 patients) NHL (25 patients) AML (24 patients)

Total number of leukapheretic procedures 140 59 41 40Median number of leukaphereses per patient (range) 2 (1–3) 2 (1–3) 2 (1–3) 2 (1–3)Median days from mobilisation to PBPC collection (range) 13 (9–17) 12 (9–17) 12 (10–14) 13 (11–15)Median number of CD34+ (× 106 kg−1) (range) 7.9 (1.28–73.2) 5.8 (1.3–19.6) 12 (2–73.2) 6.8 (1.9–33.5)Poor mobilisers (CD34+ <4× 106 kg−1) 20 (22%) 9 (22%) 4 (16%) 7 (29%)Good mobilisers (CD34+≥4 × 106 kg−1) 70 (78%) 32 (78%) 21 (84%) 17 (71%)

With one leukapheresis 18 8 6 4With two leukaphereses 42 21 14 10With three leukaphereses 6 3 1 3

action between risk factor and diagnosis has been assessedas well. Subgroup analysis on myeloma patients has beensimilarly performed.

Receiver operating characteristic (ROC) analysis was ap-plied to identify the optimal cut-off for CD34+ that wouldmaximize sensitivity and specificity.

Stata 8 (StataCorp, College Station, TX) was used for com-putation. A two-sidedP-value <0.05 was considered as sta-tistically significant.

3. Results

Characteristics of PBPC collection are listed inTable 2. Atotal of 140 leukapheres were performed in 90 patients, witha median of two procedures per patient (range: 1–3). The me-dian time elapsed from mobilisation to the first leukapheresiswas 13 days (range: 9–17). The median number of CD34+cells kg−1 collected was 7.88× 106 kg−1 (range: 1.28–73.2).Sixty-six patients (73%) collected more than 4× 106 kg−1

CD34+ cells, with 1 leukapheresis in 18 cases, 2 leukaphere-ses in 42 cases and 3 leukaphereses in 6 cases. Eleven patients(12%) mobilised less than 4× 106 kg−1. Thirteen patients(15%) failed mobilisation.

Statistical analysis did not show significant correlationb d thep theb

cellc de-

TF

U 9)W .507P .381B ) 0.294B ) 0.006B 03T

.896

quate PBPC (P = 0.04). This significance was much higherwhen considering only multiple myeloma patients (P =0.006). By contrast, for NHL and AML patients the corre-lation did not reached the statistical significance. A furtheranalysis performed in multiple myeloma patients showed thata steady-state bone marrow CD34 count≤66�L−1 can pre-dict an inadequate collection with a sensibility of 66.6% anda specificity of 87.1%. In particular, 5/31 MM patients (16%)with CD34 count >66�L−1 collected <4× 106 kg−1 CD34+cells, while 5/10 patients (50%) with CD34 count≤66�L−1

harvested <4× 106 kg−1 CD34+ cells.

4. Discussion

PBPC have nearly completely replaced bone marrow stemcells as source for the hemopoietic rescue after high-dosetherapy. In the last decade, many studies have been performedto identify factors predicting the outcome of mobilisation.Exposure to alkylating agents or radiations, and duration ofprevious chemotherapies emerged as the main adverse prog-nostic factors[6–8,12,14,16–20]. Most of these studies, how-ever, reported on patients previously undergone to several cy-cles of therapy, often including alkyating agents which haveproved to be a stem cell poison[23]. On the contrary, the cur-r rseo news ess-f etersn

etween the clinical and hematological parameters anrobability of an adequate PBPC collection apart fromone marrow CD34 cell count. Details are reported inTable 3.

As shown, a high number of bone marrow CD34+ount has a statistically significant predictive value for a

able 3actors influencing CD34+ yield

Whole case series

OR (95% CI)

nderlying disease 0.33 (0.057–1.8BC count 0.79 (0.61–1.04)LT count 0.97 (0.91–1.04)one marrow disease infiltration 0.91 (0.52–1.58one marrow CD34+ cell count 0.64 (0.43–0.93one marrow CD34+ cell count�l−1) 0.7 (0.5–0.9)ime from first treatment≥6 months 0.50 (0.20–1.28)≥4 months 0.95 (0.85–1.06)

Multiple myeloma

P OR (95% CI) P

0.2600.084 1.92 (0.29–12.76) 00.469 3.75 (0.21–66.76) 0

0.749 2.3 (0.49–10.7)0.020 0.43 (0.22–0.84)

0.044 0.48 (0.28–0.82) 0.0

0.104 0.857 (0.081–8.96) 00.415

ent strategy is to perform mobilisation earlier in the couf disease avoiding the use of alkylating agents. In thisetting of patients the known prognostic factors for succul PBPC collection are not applicable and so new parameed to be found.

162 A. Corso et al. / Leukemia Research 29 (2005) 159–163

We perform this study on 90 haematologic patients, can-didates to high-dose therapy, in order to evaluate the possiblecorrelation between the steady-state BM CD34+ counts andthe outcome of stem cell harvest. Based on the data of theliterature we also considered for the analysis other clinicaland hematologic parameters which had demonstrated to cor-relate with CD34+ yield in previous studies[12,15,17–20]. Inour hands, the only predictive parameter for PBPC collectionwas the bone marrow CD34+ cell count. Actually, most ofthe patients included in this study received mobilising therapyafter a short period of debulking therapy with a median timeto mobilisation of 5 months. As shown inTable 1, the groupof NHL patients have a longer period from first treatmentto mobilisation. This was due to the fact that most of NHLpatients were follicular lymphomas, which were included inhigh-dose strategies at relapse.

Few authors have investigated the role of steady-stateCD34+ cells as a predictive parameter for mobilisation, re-porting contrasting results[21,22]. Osma et al.[22], in a ret-rospective study on 15 patients with hematologic and solidmalignancies, found a strong correlation between peripheralblood CD34+ cell count during steady-state hematopoiesisand the CD34+ yield. On the contrary, no correlation wasfound with BM CD34+ cell count, even considering dif-ferent CD34+ subsets. Even though the sample was verys eady-s thatp riph-e ut-t tateB ndC 71s pro-g aly-s ncerp

tateB de-qT withM MLa bero mag forem

ofBf gherp ,w sup-p use-f ert . Re-ct reaseo e, a

possibility could be represented by a short period of thalido-mide before attempting mobilisation.

References

[1] Attal M, Harousseau JL, Stoppa AM, et al. A prospective, ran-domised trial of autologous bone marrow transplantation andchemotherapy in multiple myeloma. Intergroupe Francais duMyelome. N Engl J Med 1996;335:91–7.

[2] Barlogie B, Jagannath S, Vesole D, et al. Superiority of tandem autol-ogous transplantation over standard therapy of previously untreatedmultiple myeloma. Blood 1997;89:789–93.

[3] Shipp MA, Abeloff MD, Antman KH, et al. International consen-sus conference on high-dose therapy with hematopoietic stem celltransplantation in aggressive non-Hodgkin’s lymphomas: report ofthe jury. J Clin Oncol 1999;17:423–9.

[4] Burnett AK, Goldstone AH, Stevens RM, et al. Randomised com-parison of addition of autologous bone marrow transplantation tointensive chemotherapy for acute myeloid leukaemia in first remis-sion: results of MRC AML 10 trial. Lancet 1998;351:700–8.

[5] Gratwohl A, Passweg J, Baldomero H, Hermans J. Blood and mar-row transplantation activity in Europe 1997: European Group forBlood and Marrow Transplantation (EBMT). Bone Marrow Trans-plant 1999;24:231.

[6] Bensinger WI, Appelbaum F, Rowley S, et al. Factors that influencecollection and engraftment of autologous peripheral blood stem cells.J Clin Oncol 1995;13:2547–55.

entpro-n of

cellari-–96.

nceousplant

[ntlyplan-

[ cellsom-heraldose.

[ ctionrow

[ m-toxic

riph-ica

[ sh-blood

[ llec-ma.

[ s ofcore

mall and heterogeneous, they concluded that the sttate CD34+ peripheral population could better reflectart of bone marrow stem cell which is released to peral blood after mobilising therapy. More recently, Van P

en et al.[23] found a correlation between the steady-sM CD34+/CD71−, but not CD34+ cell concentration, aD34+ cells in apheresis. The reason is that CD34+/CD−eems to be a subset rich of multipotent hematopoieticenitors and most immature myeloid progenitors. The anis, however, was performed on small cohort of breast caatients mobilised with growth factor alone.

By contrast, in our study we found that a high steady-sM CD34+ cell count was significantly correlated with auate peripheral blood progenitor cell collections (P= 0.04).his correlation was even more significant in patientsM (P= 0.003), whereas it was lost in the subgroups of And NHL patients. This could be related to a higher numf MM patients and to a greater homogeneity in myeloroup with regard to the therapeutic approach beobilisation.Within the MM group we tried to identify a cut-off level

M CD34+ cell count. We found that a value≤66�L−1, be-ore starting mobilising therapy, was associated with a hirobability of collecting less than 4× 106 kg−1 CD34+ cellshich is considered the optimal target for an autologousort after high-dose therapy. This information might be

ul in identifying the so-called “poor mobilisers”, in ordo adopt alternative strategies to overcome this problemently on 17 patients[24] with refractory MM treated withhalidomide, we observed in responsive patients an incf BM CD34+ cells after 3 months of therapy. Therefor

[7] Weaver CH, Hazelton B, Birch R, et al. An analysis of engraftmkinetics as a function of the CD34 content of peripheral bloodgenitor cell collections in 692 patients after the administratiomyeloablative chemotherapy. Blood 1995;86:3961–9.

[8] Tricot G, Jagannath S, Vesole D, et al. Peripheral blood stemtransplants for multiple myeloma: identification of favourable vables for rapid engraftment in 225 patients. Blood 1995;85:588

[9] Perez-Simon JA, Martin A, Caballero D, et al. Clinical significaof CD34+ cell dose in long-term engraftment following autologperipheral blood stem cell transplantation. Bone Marrow Trans1999;24:1279–83.

10] Scheid C, Draube A, Reiser M, et al. Using at least 5× 106 kg−1

CD34+ cells for autologous stem cell transplantation significareduces febrile complications and use of antibiotics after transtation. Bone Marrow Transplant 1999;23(1):9–13.

11] Lefrere F, Delarue R, Somme D, et al. High-dose CD34+are not clinically relevant in reducing cytopenia and blood cponent consumption following myeloablative therapy and peripblood progenitor cell transplantation as compared with standardTransfusion 2002;42:443–50.

12] Corso A, Caberlon S, Pagnucco G, et al. Blood stem cell collein multiple myeloma: definition of a scoring system. Bone MarTransplant 2000;26:283–6.

13] Corso A, Arcaini L, Caberlon S, et al. A combination of dexaethasone, cyclophosphamide, etoposide and cisplatin is lessand more effective than high-dose cyclophosphamide for peeral stem cell mobilization in multiple myeloma. Haematolog2002;87:1041–5.

14] Gandhi MK, Jestice K, Scott MA, et al. The minimum CD34 threold depends on prior chemotherapy in autologous peripheralstem cell recipients. Bone Marrow Transplant 1999;23:9–13.

15] Demirer T, Buknener CD, Gooley T, et al. Factors influencing cotion of peripheral blood stem cells in patients with multiple myeloBone Marrow Transplant 1996;17:937–41.

16] Drake M, Ranaghan L, Morris TCM, et al. Analysis of the effectprior therapy on progenitor cell yield: use of a chemotherapy ssystem. Br J Haematol 1997;98:745–9.

A. Corso et al. / Leukemia Research 29 (2005) 159–163 163

[17] Haas R, Mohle R, Frohauf S, et al. Patients characteristics associ-ated with successful mobilizing and autografting of peripheral bloodprogenitor cells in malignant lymphoma. Blood 1994;83:3787–94.

[18] Ketterer N, Salles G, Moullet I, et al. Factors associated with suc-cessful mobilization of peripheral blood progenitor cells in 200 pa-tients with lymphoid malignancies. Br J Haematol 1998;103:235–42.

[19] Perea G, Sureda A, Martino R, et al. Predictive factors for a suc-cessful mobilization of peripheral blood CD34+ cells in multiplemyeloma. Ann Hematol 2001;80:592–7.

[20] Canales MA, Fernande-Jimenez MC, Martin A, et al. Identificationof factors associated with poor peripheral blood progenitor cell mo-bilization in Hodgkin’s disease. Haematologica 2001;86:494–8.

[21] Fruehauf S, Haas R, Conradt C, et al. Peripheral blood progeni-tor cell (PBPC) counts during steady-state hemtopoiesis allow toestimate the yield of mobilized PBPC after Filgrastim (R-metHuG-CSF)-supported cytotoxic chemotherapy. Blood 1995;85:2619–26.

[22] Osma MM, Ortuno F, de Arriba F, et al. Bone marrow steady-state CD34+/CD71− cell content is a predictive value of rG-CSF-mobilized CD34+ cells. Blood 1998;21:983–5.

[23] Van Putten LM, Lelieveld P, Kram-Idsenga LK. Cell-cycle specificityand therapeutic effectiveness of cytostatic agents. Cancer ChemotherRep 1972;56(6):691–700.

[24] Corso A, Lorenzi A, Zappasodi P, et al. Early changes in bone mar-row morphology induced by thalidomide in patients with refractorymyeloma. Haematologica 2003;88(8):958–60.