British Journal of Haematology, 1999, 106, 538544

Use of the haemopoietic progenitor cell count of the Sysmex SE-9500 to rene apheresis timing of peripheral blood stem cells
Y. P O LL A RD , M. J. WATTS , D. G R AN T, N. C HAVDA , D. C. L I N C H A N D S. J. M AC H I N Department of Haematology, University College London Hospitals, London Received 22 February 1999; accepted for publication 14 May 1999

Summary. The Sysmex SE-9500 automated cell counter provides an estimate of immature cells referred to as `haemopoietic progenitor cells' (HPC). The aim of this study was to relate the HPC count to CD34 cell levels in mobilized peripheral blood and to determine whether the HPC count was valuable in predicting apheresis yields of CD34 cells. Studies were performed on 114 samples from 67 patients undergoing progenitor cell mobilization. HPC cells were undetectable in the steady state. On the day of apheresis the HPC and CD34 counts were weakly correlated, with the median HPC count being 23-fold greater than the CD34 cell count. The HPC count did not include the CD34 cells as immunomagnetic depletion of CD34 cells did not signicantly reduce the HPC count.

CD34 cell counts predicted for apheresis yield (r 0773) on that day as did the HPC count (r 0623). The optimal strategy to prevent unnecessary harvesting while minimizing the risk of missing an adequate harvest, and minimizing laboratory investigations, was to screen all samples for HPC and limit CD34 cell measurements to those with an HPC count <10 106/l (19/114 samples). If the CD34 cell count was also <10 106/l then harvesting should not be carried out. Keywords: haemopoietic progenitor cells, Sysmex SE-9500 analyser, CD34 cells, apheresis timing, peripheral blood stem cells.

There is widespread agreement that there is a threshold requirement of CD34 cells or granulocyte/monocytecolony forming cells (GM-CFC) for rapid haematological engraftment after peripheral blood stem cells (PBSC) infusion following high-dose chemotherapy and/or radiotherapy (Demirer & Besinger, 1995). In a recent study at our institution of 258 patients transplanted with autologous PBSC, delayed platelet independence, dened as >28 d to achieve a sustained platelet count >15 109/l, occurred in about half of the 37 patients who failed to achieve a total apheresis yield of 1 106/kg CD34 cells. The 105 patients who received between 1 and 35 106/kg CD34 cells had a risk of delayed platelet recovery of 8% and >35 106/kg delayed recovery was very rare (Watts et al, 1998). We would be reluctant to commence apheresis unless a minimum CD34 cell dose of 1 106/kg could be assured within two or three collections, and we would therefore consider the minimum yield worthwhile collecting from a single apheresis to be 05 106/kg.
Correspondence: Dr Mike Watts, Department of Haematology, 98 Chenies Mews, University College London, London WC1E 6HX. e-mail: michael.watts@ucl.ac.uk.

An important problem in optimizing PBSC progenitor yields is knowing when to commence apheresis. In the case of G-CSF alone for mobilization, apheresis usually begins from a xed day according to the dose used (Anderlini et al, 1996). In the case of growth factor use following chemotherapy, the recovery WBC is reliable for harvest guidance with some well-standardized protocols (Dreger et al, 1993; Watts et al, 1995), but both timing methods can be confounded by poor mobilizers and where very diverse protocols are in use. In these instances ow cytometric PB CD34 cell counts have been used either on the day or preceding the day of apheresis to predict yield. These methods have been shown to be reliable in some studies (Schwella et al, 1995; Armitage et al, 1997) but not others (Luider et al, 1997) and have not focused on the problems of patients where yields <2 106/kg CD34 cells are likely. The wider application of blood CD34 cell measurement has been restricted by the time-consuming, expensive and technically challenging nature of the test especially at the very low levels (below 01%) frequently encountered in poorer mobilizers. In the present study we have evaluated the performance of a new parameter, the haemopoietic progenitor cell count,
q 1999 Blackwell Science Ltd

538

HPC Count and Apheresis Timing

539

Fig 1. Principle of the IMI channel.

which can be rapidly and simply obtained as part of the Sysmex SE-9500 analyser automated blood count and compared this with PB CD34 cell counts, to predict when apheresis was feasible to achieve minimal CD34 cell target yields. In addition to the full blood count (FBC) and ve-part white cell differential, the Sysmex SE-9500 provides an immature myeloid information (IMI) channel in which haemopoietic progenitor cells may be measured in selected samples. In the IMI channel all white cells except immature myeloid cells are lysed, leaving only bare nuclei. This lysis is due to the action of surfactants on the lipid components of the cell membrane. The reagent used in the IMI channel contains polyoxyethylene non-ionic surfactant and sulphurcontaining amino acid to x blood cell cytoplasmic membranes (Ishii et al, 1997). It also contains anionic surfactant to damage the blood cell membrane in order to reduce the red cell ghosts and platelets. As white cells mature there is an increase in the lipid content of the membrane with mature neutrophils containing the highest levels of phospholipid (Gottfried, 1972), this means that it is easier for the combination of surfactants (known as super surfactant) in the reagent to lyse mature white cells than to lyse immature cells, as illustrated in Fig 1. In the IMI channel, once the immature cells are separated the cluster is analysed using radio frequency (RF) and direct current (DC). The RF/

DC method detects impedance changes as pulses when blood cells pass through the sensor where the RF and DC are applied. The RF signal conveys information about the cell contents such as the nuclear size and presence of granules, the DC signal reects the size or volume of the cells. Analysis with these two parameters provides detailed morphological information about each cell. When the blood sample is treated with the IMI reagent, immature myeloid cells can be distinguished from other cells (Ishii et al, 1997). Normally, the more immature the cells, the smaller the RF signal strength. They therefore appear lower on the IMI scattergram, as shown in Fig 2. MATERIALS AND METHODS Patients studied, PBSC mobilization and collection. The diagnosis, number of apheresis collections and mobilizing regimens used are summarized in Table I. The nine normal donors were mobilized with G-CSF alone at 10 mg/kg/d and apheresis commenced from day 5. The 58 patients with haematological and non-haematological malignancies were given chemotherapy followed by G-CSF for at least 9 d which continued until harvesting was complete. Apheresis commenced in these patients on a rising WBC recovery count >30 109/l for patients mobilized with ESHAP and G-CSF or 50 109/l for the cyclophosphamide and G-CSF protocol. These WBC thresholds to commence harvesting have been established in previous studies from our unit (Watts et al, 1996, 1997). The remaining patients who received chemotherapy and G-CSF began apheresis when the WBC rst exceeded 50 109/l. 45 patients had a total of 72 apheresis collections performed on a Baxter CS3000 (Baxter Healthcare Ltd, Berkshire) set to process a xed 10 litre blood volume and the remainder had 42 collections performed on a COBE Spectra (COBE Laboratories Ltd, Gloucester) with 1014 litres of blood processed (median 117 litres). Six patients had at least one collection performed on each machine type. We have previously shown in a larger study that the performance of these two machines is very similar (Watts et al, 1997). CD34 cell and GM-CFC evaluation. Peripheral blood and harvest CD34 positive cell numbers were determined by ow cytometry using a modied Milan protocol (Sienna et al, 1990). Apheresis samples were diluted 1/20 in AB serum in

Fig 2. IMI scattergram.

q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544

540

Y. Pollard et al

Table I. Patients studied. No. of patients 1 5 1 8 32 2 4 5 9 67 No. of harvests 1 6 2 22 49 5 4 7 18 114

Diagnosis Acute lymphoblastic leukaemia Acute myeloid leukaemia Chronic lymphoblastic leukaemia Hodgkin's disease High-grade non-Hodgkin's lymphoma n1 Low-grade non-Hodgkin's lymphoma Myeloma Solid tumour Normal donors Total

Mobilizing regimen (all include G-CSF*) Methotrexate ADE n 2, MIDAC n 1, MACE n 1, DAT n 1 Cyclophosphamide Cyclophosphamide n 3, ESHAP n 4, DexaBEAM n 1 Cyclophosphamide n 14, ESHAP n 17, DexaBEAM Cyclophosphamide, DexaBEAM Cyclophosphamide n 3, EDAP n 1 Cyclophosphamide n 2, ifosfamide n 1 Ifosfamide VP-16 n 1, VID n 1 G-CSF only n 9

a 50 ml volume (approximately 05 106 cells) or a 100 ml aliquot of peripheral blood was incubated for at least 30 min with a directly conjugated phycoerythrin HPCA-2 monoclonal antibody or appropriate matched control (Becton Dickinson). This was followed by red cell lysis (Coulter Qprep), a phosphate-buffered saline wash and ow cytometry (Coulter Epics XL-MCL). A minimum of 50 000 white cells were counted for each sample and the percentage of cells which were CD34 positive were identied by the dual characteristics of low light scatter and PE staining properties and the irrelevant control percentage subtracted. The CD34 cell percentages were used to derive absolute counts from the peripheral blood and harvest total white cell counts obtained from the SE-9500 analyser. Granulocyte/monocyte colony forming cells (GM-CFC) were performed using commercially available methylcellulose media (StemCell Technologies Inc., Metachem, Northampton) with added recombinant growth factors (Stem cell factor 10 ng/ ml, IL-3 30 ng/ml, GM-CSF 25 ng/ml G-CSF 25 ng/ml and hrErythropoietin 3 U/ml). Harvest specimens were plated in triplicate at concentrations of 25 104/ml and colonies counted after 14 d incubation at 378C in a humidied CO2 atmosphere. Sysmex SE-9500 HPC analysis. All samples were analysed in the HPC mode of the Sysmex SE-9500. In addition to collecting data on individual patients for PBSC harvest, reproducibility, linearity and carryover were assessed. RESULTS Sysmex SE-9500 HPC count performance: reproducibility To assess reproducibility, samples with a range of HPC counts were consecutively analysed 10 times and the mean, standard deviation (SD) and coefcient of variation (CV) calculated, as shown in Table II. A CV of 318% was seen at the lowest levels of HPC with a progressive fall to 98% at the higher levels of HPC. Considering the very low levels of HPC present (in percentage terms 001054%), this precision was considered acceptable and at least comparable with the

precision attainable using standard ow cytometric CD34 cell counts (Serke et al, 1997). Linearity The linearity of the HPC was assessed by serially diluting a peripheral blood sample with an HPC of 106 106/l in group AB serum. Good linearity was seen down to very low levels of HPC (Fig 3). Carryover Using the method originally described by Broughton et al (1974) and England et al (1994), carryover was performed several times by analysing samples with HPC counts of up to 500 106/l three times (high 1, 2, 3) followed by samples with low HPC counts three times (low 1, 2, 3), Carryover was then calculated as follows: Carryover (%) low 1 low 3 high 3 low 3

There was no detectable carryover at these levels. This is very important since it is at the lowest levels of HPC that the decision to apherese would be made. Correlation with PB CD34 cell count The HPC counts were on average 23-fold higher that of the blood CD34 cell counts with a median (range) of 59 (0 655) and 26 (0371) 106/l respectively and with a limited

Table II. Reproducibility of HPC counts. HPC 106/l Mean SD CV Min. Max. 47 15 318 3 8 111 22 201 8 14 428 53 124 35 53 846 83 98 71 98

q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544

HPC Count and Apheresis Timing

541

Fig 3. HPC linearity. Whole blood diluted in autologous plasma.

correlation (Fig 4, r 052). In some samples, however, the CD34 cell count exceeded the HPC count.

involved extensive manipulation and washing and may have altered the surfactant reactions required for HPC detection. Accordingly we used CD34 magnetic selection beads (Miltenyi Biotec Ltd, Surrey), directly conjugated with a CD34 monoclonal antibody (QBEND10) to deplete CD34 cells from whole blood with minimal manipulation. Three EDTA blood count samples with a CD34 cell concentration above 1% were chosen and 100 ml of the selection beads added to a 2 ml aliquot of whole blood for 30 min at 48C. 1 ml of the labelled sample was passed through a CD34 selection column in a magnetic holder (`VarioMACs', Miltenyi Biotec Ltd) and the CD34 cell depleted fraction collected. The remaining 1 ml was passed through an identical column without the magnetic holder. Flow cytometric measurement of CD34 cells showed magnetic depletion of these cells. In contrast, the HPC count was not signicantly affected by this procedure (Table III).
Table III. HPC count after CD34 cell depletion from whole blood. Patient 1 Patient 2 18 09 19 09 01 06 Patient 3 26 11 25 09 03 10

Fig 4. Correlation between blood HPC and CD34 cell counts prior to apheresis (n 114).

Whole blood (pre) CD34% HPC% CD34 beads, no magnet CD34% HPC%

20 21 18 28

HPC counts following CD34 cell depletion of whole blood The above analysis indicated that different progenitor populations were measured by the HPC and CD34 cell assays. To determine whether CD34 cells were usually measured as HPC we rst assayed samples of puried CD34 cells obtained from PBSC collections with clinical CD34 selection systems. These were obtained by immunoafnity selection in two instances (CellPro, Ceprate, CellPro, Seattle, Wash., U.S.A.), and in the third by immunomagnetic selection (Isolex 300i, Baxter Healthcare, Berks.). The CD34 cell purity of these products was 76% and 78% and 78% respectively with corresponding HPC concentrations of only 98%, 57% and 50%. However, CD34 cell purication

CD34 beads, magnetic depletion CD34% 01 HPC% 28

Apheresis CD34 cell and GM-CFC yields GM-CFC assays were available for 112/114 harvests and yields were highly correlated with CD34 cell doses (r 0994, Fig 5). A CD34 dose above 1 106/kg was equivalent to a GM-CFC yield of at least 1 105/kg in all cases. Amongst the poorer mobilizers a CD34 dose between

q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544

542

Y. Pollard et al
blood progenitor counts was considered. A threshold blood progenitor count of 10 106/l was selected for both tests as this was the lowest reliable limit of detection for these tests at around 01% in the majority of poorest mobilizers. The distribution of the blood progenitor counts with resulting CD34 cell yields is shown in Figs 6A and 6B and the implications of their use if applied to apheresis collection management summarized in Table IV. Four options for the making the decision of when to commence apheresis were considered (Table IV). The rst option was to base the decision to commence apheresis as at present on a recovery WBC or xed collection day with no blood progenitor measurements. The median recovery white cell count at apheresis in the patients receiving post mobilization chemotherapy was 146 (34 630) 109/l and was not correlated with the CD34 cell yield (r 031, data not shown) The normal donors were collected on day 5 and 6 of G-CSF administration. Since all patients undergo apheresis, no worthwhile yields are missed, however 24% (27/114) harvests with very poor yields <05 106/kg were collected. Patients may therefore endure excessive apheresis and resource utilization is poor. In the second option the implications of modifying the collection day based on blood CD34 cell counts on all patients was considered. At a threshold of 10 106/l taken for the decision to apherese, 16% (18/114) of harvests which proved to be poor would be avoided. However, 6% (7/114) of yields between 05 and 1 106/kg CD34 cells would be missed and more seriously 5% (6/114) >1 106/kg CD34 cells would be missed. The third option considered was the use of PB HPC counts as a screening test on all patients. At a threshold of 10 106/l taken for the decision to apherese, 11% (13/114) of harvests which proved to be inadequate would be avoided. However,

Fig 5. Correlation between apheresis CD34 cell and GM-CFC yields (n 112).

05 and 1 106/kg was at least equivalent to 05 105/kg in all but three patients in which the yields were 09, 09 and 06 106/kg CD34 cell compared to 03, 03 and 05 105/kg GM-CFC respectively. Use of blood progenitor measurements to judge apheresis commencement The implications of modifying the decision of when to commence apheresis based on a rising WBC by introducing

Fig 6. (A) Relationship between blood CD34 count and apheresis yield on that day (n 114). (B) Relationship between blood HPC count and apheresis yield on that day (n 114). q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544

HPC Count and Apheresis Timing

Table IV. Implications for apheresis collection where the decision to collect PBSC is modied by blood progenitor counts. CD34 cell apheresis yields 106/kg Progenitor collection threshold 106/l None PB CD34 >10 HPC count >10 HPC >10 (PB CD34 used if HPC is <10) Collect (n) (114) (83) (95) (103) <05 27 9 14 18 051 20 13 16 18 >1 67 61 65 67 Do not collect (n) (31) (19) (11) <05 18 13 9 051 7 4 2

543

>1 6 2 0

4% (4/114) of CD34 cells yields between 05 and 1 106/ kg and 2% (2/114) were >1 106/kg and would not have been collected on that day. The nal option considered was to apply HPC counts as a rapid screening test on all patients with CD34 cell counts as a second discriminator for those with HPC counts <10 106/l. This would avoid the collection of 8% (9/114) of harvests which proved to have very poor yields. Only 2% (2/114) of harvests in patients with these very low blood progenitor counts yields had harvests of 051 106/kg CD34 cells (05 and 07 106/kg). These patients were correctly predicted on the following day when yields >1 106/kg were collected. DISCUSSION The overall incidence of poor mobilizers at our centre dened as a failure to collect 1 106/kg CD34 cells is around 10%, rising to 1520% in some diagnostic groups such as lowgrade lymphoma and myeloma (Watts et al, 1998). We were therefore interested in rening our apheresis strategy to minimize the collection of very poor yields in these patients and to detect the early rise of progenitors where diverse mobilizing chemotherapy protocols had been used. We and others have previously shown that using a standardized mobilization protocol a recovery WBC threshold can be reliably used to predict the optimal timing of rst apheresis (Dreger et al, 1993; Haas et al, 1995; Watts et al, 1995). In the present cohort, where a variety of mobilization schedules were administered, use of a rising WBC threshold for initiating harvesting was clearly unreliable and cost ineffective. Many centres use the peripheral blood CD34 cell count to initiate harvesting and predict yields (Schwella et al, 1995; Armitage et al, 1997; Areman et al, 1997). This should be the most accurate predictor of CD34 cell yields, although in this study the correlation between blood levels and harvest yields was not high (r 077). This is similar to that reported by Areman et al (1997) (r 072) and Luider et al (1997) (r 057), although it should be noted that others have reported better correlation. The relatively poor correlation can reect technical difculties in the harvesting procedure or inaccuracies in the blood CD34 measurements when the

percentage CD34 cells is low (Sovalat et al, 1994) as in many patients in this series. In the present study the relatively simple `Milan protocol' was used for CD34 cell estimation. Improved accuracy can probably be obtained with more complex staining and analytical techniques (Sutherland et al, 1997) but this increases the time and expense of an already costly procedure. For this reason we evaluated the use of the HPC count on the Sysmex SE-9500 which is rapid and incurs no cost beyond that of a blood count. The precise nature of the progenitors measured in the HPC channel is unclear. The HPC count was typically double that of the CD34 cell count and was not affected by CD34 cell depletion, suggesting that a different progenitor population (presumably more mature in view of the higher numbers) is measured. We have previously shown that morphological `blast cell' counts of Romanoski stained blood lms is a good indication of successful mobilization but difcult to quantify (Jones et al, 1994). Visual scoring of progenitors has however been successfully utilized by Mijovic et al (1998) to assess apheresis harvest adequacy for engraftment. Automated HPC counting should give a more precise measurement of early cells than visual counts and allow a more quantitative assessment of the release of progenitor cells into the blood. This study shows that the HPC count alone is inadequate for predicting yield, although it should be noted that the correlation between blood HPC and CD34 cell yields was only a little worse than the correlation between blood CD34 cell levels and CD34 cell harvest yields (r 066 v r 077). Further analysis revealed, however, that the use of an HPC screening assay followed by CD34 cell measurements restricted to those patients with an HPC count <10 106/l was highly predictive and more cost effective than the use of blood CD34 cell measurements alone. In a study of 43 patients undergoing 125 PBSC collections reported by Areman et al (1997) the cost of a peripheral blood CD34 cell assay was estimated at $200 and the cost of each PBSC collection $4000. Applying this data to our own cohort of patients the saving associated with HPC screening and limited blood CD34 cell estimations, compared to the use of the WBC alone, would be approximately $32 000, equating to a saving of about $500 per patient.

q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544

544

Y. Pollard et al
D., Stewart, D., Klassen, J. & Russell, J.A. (1997) Factors inuencing yields of progenitor cells for allogeneic transplantation: optimization of G-CSF dose, day of collection, and duration of leukapheresis. Journal of Hematotherapy, 6, 575580. Mijovic, A., Fishlock, K., Pagliuca, A. & Mufti, G.J. (1998) Blast counts in blood progenitor cell (BPC) correlate with CD34 cells and CFU-GM and are useful predictor of haemopoietic recovery after autologous BPC transplantation collections. Bone Marrow Transplantation, 21, 869872. Schwella, N., Siegert, W., Beyer, J., Rick, O., Zingsem, J., Eckstein, R., Serke, S. & Huhn, D. (1995) Autografting with blood progenitor cells: predictive value of preapheresis blood cell counts on progenitor cell harvest and correlation of the reinfused cell dose with hematopoietic reconstitution. Annals of Hematology, 71, 227234. Serke, S., Arseniv, L., Watts, M., Fritsch, G., Ingles-Esteve, J., Cancelas, J.A., Meyer, O., Kadar, J.G., Huhn, D. & Matcham, J. (1997) Imprecision of counting CFU-GM colonies and CD34expressing cells. Bone Marrow Transplantation, 20, 6771. Sienna, S., Bregni, M. & Brando, B. (1991) Flow cytometry for clinical estimation of circulating haematopoietic progenitors for autologous transplantation in cancer patients. Blood, 77, 400 406. Sovalat, H., Wunder, E., Zimmerman, R. & Serke, S. (1994) Multicentric determination of CD34 cells. Hematopoietic Stem Cells: The Mulhouse Manual (ed. by Eckart Wunder et al), pp. 61 66. AlphaMed Press, Dayton, Ohio. ISBN1-880854 17-1. Sutherland, R., Anderson, L., Keeney, M., Rakash, N. & Chin-Yee, I. (1996) The Ishage guidelines for CD34 cell determination by ow cytometry. Journal of Hematotherapy, 5, 213226. Watts, M.J., Leverett, D., Sullivan, A.M., Peniket, A.J., Perry, A.R., MacMillan, A., Goldstone, A.H. & Linch, D.C. (1996) A comparison of ESHAP G-CSF vs cyclophosphamide 15 g/ m2 G-CSF for PBSC mobilisation in pre-treated lymphoma patients: a matched pairs analysis. Blood, 88, (Suppl. 1), 396a. Watts, M.J., Sullivan, A.M., Jamieson, E., Chavda, N., Fielding, A., Goldstone, A.H. & Linch, D.C. (1995) A single apheresis algorithm for PBSCT. British Journal of Haematology, 89, 22a (abstr. 79). Watts, M.J., Sullivan, A.M., Jamieson, E., Pearce, R., Fielding, A., Goldstone, A.H. & Linch, D.C. (1997) Progenitor cell mobilisation after low dose cyclophosphamide and G-CSF: an analysis of progenitor cell quantity and quality and factors predicting for these parameters in 101 pretreated patients with malignant lymphoma. Journal of Clinical Oncology, 15, 535546. Watts, M.J., Sullivan, A.M., Leverett, D., Peniket, A.J., Perry, A.R., Williams, C.D., Devereux, S., Goldstone, A.H. & Linch, D.C. (1998) Back up bone marrow is frequently ineffective in patients with poor peripheral blood stem cell mobilization. Journal of Clinical Oncology, 4, 15541560.

REFERENCES
Anderlini, P., Przepiorka, D., Huh, Y., Lauppe, J., Miller, P., Sundberg, J., Seong, D., Champlin, R. & Korbling, M. (1996) Duration of lgrastim mobilization and apheresis yield of CD34 progenitor cells and lymphoid subsets in normal donors for allogeneic transplantation. British Journal of Haematology, 93, 940942. Areman, E.M., Rhodes, P.L., Meehan, K.R. & Sacher, R.A. (1997) Costs of autologous peripheral blood stem cell collection can be reduced by PB CD34 screening before apheresis. Blood, 90, (Suppl. 2), 322b, abstract 4198. Armitage, S., Hargreaves, R., Samson, D., Brennan, M., Kanfer, E. & Navarrete, C. (1997) CD34 counts to predict the adequate collection of peripheral blood progenitor cells. Bone Marrow Transplantation, 20, 587591. Broughton, P.M.G., Gowenlock, A.N., McCormack, J.J. & Neil, D.W. (1974) A revised scheme for the evaluation of automated instruments for use in clinical chemistry. Annals of Clinical Biochemistry, 11, 207218. Demirer, T. & Bensinger, W.I. (1995) Optimisation of peripheral blood stem cell collection. Current Opinions in Haematology, 2, 219226. Dreger, P., Marquardt, P., Haferlach, T., Jacobs, S., Mulverstedt, T., Eckstein, V., Suttorp, M., Lofer, H., Muller-Ruchholtz, W. & Schmitz, N. (1993) Effective mobilisation of peripheral blood progenitor cells with `Dexa-BEAM' and G-CSF: timing of harvesting and composition of the leukapheresis product. British Journal of Cancer, 68, 950957. England, J.M., Rowan, R.M., Van Assendelft, O.W., Bull, B.S., Coulter, W.H., Fujimoto, K., Groner, W., Jones, A.R., Koepke, J.A., Lewis, S.M., Shinton, N.K., Tatsumi, N., Thom, R., Verwilghen, R.L. & McLare, C.E. (1994) Guidelines for the evaluation of blood cell analysers including those used for differential leucocyte and reticulocyte counting and cell marker applications: ICSH expert panel on cytometry. Clinical and Laboratory Haematology, 16, 157 174. Gottfried, E.L. (1972) Lipid patterns of leukocytes in health and disease. Seminars in Hematology, 9, No. 3, pp. 241250. Haas, R., Witt, B., Mohle, R., Goldschmidt, H., Hohaus, S., Freuhauf, S., Wannenmacher, M. & Hunstein, W. (1995) Sustained longterm hematopoiesis after myeloablative therapy with peripheral blood progenitor cell support. Blood, 85, 37543761. Ishii, T., Kawasumi, I. & Matsumoto, H. (1997) SE-9000 IMI Channel: focusing on the roles and functions of surfactant. Sysmex Journal International, 7, 123128. Jones, H.M., Jones, S.A., Watts, M.J., Khwaja, A., Mills, W., Fielding, A., Goldstone, A.H. & Linch, D.C. (1994) Development of a simplied single-apheresis approach for peripheral-blood progenitor-cell transplantation in previously treated patients with lymphoma. Journal of Clinical Oncology, 12, 1693702. Luider, J., Brown, C., Sellinger, S., Quinlan, D., Karlsson, L., Ruether,

q 1999 Blackwell Science Ltd, British Journal of Haematology 106: 538544