Peripheral Blood CD34 Cell Count Reliably Predicts Autograft Yield

Home , White blood cell

Bone Marrow Transplantation, (1998) 22, 125–130  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt Peripheral blood CD34+ cell count reliably predicts autograft yield

P Chapple1,2, HM Prince1, M Quinn1, I Bertoncello3, S Juneja1,2, M Wolf1, H Januszewicz1, M Brettell1, J Gardyn1, C Seymour1 and D Venter2

1Blood and Marrow Transplant Service and 2Department of Pathology, Division of Haematology and Medical Oncology and 3Stem Cell Biology Laboratory, Division of Research, Peter MacCallum Cancer Institute, Melbourne, Victoria, Australia

Summary: routinely at many centers with threshold values of 2– 5 × 106 CD34+ cells/kg predicting rapid and durable A reliable measure to predict peripheral blood progeni- engraftment following PBPCT.4 tor cell (PBPC) autograft CD34+ cell content is required However, obtaining PBPCs by leukapheresis is expens- to optimize the timing of PBPC collection. We prospec- ive and there is morbidity, albeit small, associated with the tively examined the peripheral blood (PB) CD34+ cell procedure. Consequently, a reliable assay to predict the count in 59 consecutive patients with various malig- CD34+ cell yield is required to optimize the timing of col- nancies and analyzed the correlation between the PB lections. To determine the optimal time of PBPC collection, CD34+ cell count and various parameters in the PBPC we prospectively examined a number of parameters in the autograft. Two hundred and thirty-five collections were peripheral blood (PB) of patients on the day of planned performed with a median of 4.0 collections per patient leukapheresis, and correlated these with various parameters (range, 2–10). The median PB CD34+ cell count at the in the autograft product. time of collection was 39 × 106/l (range, 0.0–285.6). The PBPC autograft parameters measured were the CD34+ cell, colony-forming unit granulocyte–macrophage Methods (CFU-GM) and mononuclear cell (MNC) content. There was a strong linear correlation between PB CD34+ cells/l Immediately prior to 235 consecutive PBPC collection pro- and autograft CD34+ cells/kg (r = 0.8477). The corre- cedures, the PB white blood cell (WBC), MNC, platelet lation with CFU-GM/kg (r = 0.5512) was weaker. There and CD34+ cell counts were determined and correlated with was no correlation between autograft CD34+ cells/kg various autograft parameters namely, CD34+ cell, CFU-GM and PB WBC (r = 0.0684), PB MNC (r = 0.1518) or PB and MNC content. These autograft parameters were platelet count (r = 0.2010). At our institution we aim to expressed as a function of patient body weight. CD34+ obtain a minimum of 0.5 × 106 CD34+ cells/kg with each cells/kg in the autograft was also compared to the other day of collection. We demonstrate that such a collection autograft parameters. can be reliably obtained if the PB CD34+ cell count exceeds 5.0 × 106/l. Patient characteristics Keywords: CD34; transplantation; apheresis All patients undergoing PBPC harvesting at our institute between June 1996 and June 1997 were included and their disease characteristics are summarized in Table 1. This pro- Rapid and durable hematopoietic recovery is required fol- vided 235 complete data sets for analysis. lowing intensive therapy and autologous peripheral blood progenitor cell transplantation (PBPCT). Although, no definitive assay for the pluripotent hematopoietic stem cell Apheresis procedure 1 exists, a number of surrogate laboratory measures are used The mobilization strategy employed various chemothera- to predict hematopoietic recovery following PBPC trans- peutic agents supplemented with rhGCSF (10 ␮g/kg), or plantation.2 These include the autograft mononuclear cell count (MNC), clonogenic assays in semi-solid media, usu- = ally the colony-forming unit granulocyte–macrophage Table 1 Patient characteristics (n 59) (CFU-GM), and more recently the number of hematopoietic cells marked by the progenitor cell antigen CD34.3 Indeed, Diagnosis Number Number of days + of apheresis as the CD34 cell content in the autograft correlates well mean (range) with the rate of hematopoietic recovery, it is now assessed All patients 59 4.0 (2–10) Non-Hodgkin’s lymphoma 23 3.6 (2–8) Correspondence: Dr HM Prince, Blood and Marrow Transplant Service, Hodgkin’s disease 3 3 (3–3) Division of Haematology and Medical Oncology, Peter MacCallum Breast cancer 18 3.6 (2–9) Cancer Institute, Locked Bag 1, A’Beckett St, Melbourne 3000, Victoria, Multiple myeloma 8 5.1 (2–10) Australia Others 7 5.3 (2–9) Received 21 July 1997; accepted 8 March 1998 Blood CD34 and autograft yield P Chapple et al 126 utilized rhGCSF (10 ␮g/kg) alone (Table 2). Following cytometry using phycoerythrin-conjugated class III anti- chemotherapy and cytokine mobilization, PBPC collection bodies to the CD34 antigen (clone HPCA-2) and peridium was initiated when the patient’s PB WBC returned to chlorophyl protein (PerCP) conjugated anti-CD45 (all anti- Ͼ1.5 × 109/l and/or the PB absolute neutrophil count bodies from Becton Dickinson Immunocytometry Systems, (ANC) reached 1.0 × 109/l following the nadir. A minimum San Jose, CA, USA), using an in-house whole blood lysis platelet count of 30 × 109/l was also required. For patients methodology which has been verified against the Inter- mobilized with cytokine alone, apheresis was commenced national Society of Hematotherapy and Graft Engineering 5 days following initiation of G-CSF. By employing these (ISHAGE) guidelines.5,6 Briefly, 1 × 106 cells, either in standard trigger points for initiating apheresis, patients were whole blood or the apheresis collection product, were added harvested between 5–16 days following mobilization. Dur- to an aliquot of anti-CD34 PE and anti-CD45 PerCP which ing the study, the PB CD34+ cell count was not used as a had previously been titrated to provide a saturating anti- criterion for initiating apheresis. body concentration, and incubated for 20 min at room tem- Apheresis was performed using a Haemonetics V50, perature. This was followed by addition of red cell lysing CS3000plus (Baxter Healthcare Corp, Deerfield, IL, USA) solution (FACS lysing solution, Becton Dickinson), fol- or Spectra (Cobe, Denver, CO, USA) cell separators. Blood lowed by brief vortex mixing and incubation for 10 min at was collected via a central venous access device or from room temperature. The cells were washed once by adding peripheral veins with 1.5–2.5 times the patient’s estimated imidazole buffered saline with 2% bovine serum and 0.1% blood volume processed. Na azide and centrifuging at 600 g. Without delay they were analyzed on a Becton Dickinson FACSCalibur flow cytometer by acquiring florescence and forward light scatter Cell counts data on 120 000 ungated events. Subsequent data analysis Following appropriate dilution with Isoton III (Coulter, was performed on the listmode data files using Becton Hialeah, FL USA), total WBC counts for PB and PBPC Dickinson Cellquest software v3.0 (Becton Dickinson). In samples were determined on a Coulter STKS analyzer all list mode data analysis ‘true’ CD34+ events were defined (Coulter). A 100 cell manual differential count was used to as having the dual characteristics of bright CD34 determine the ANC. The MNC count was calculated by expression and low-to-intermediate CD45 expression subtracting the ANC from the total WBC count. (Figure 1). Using the autograft WBC count, %CD34+ cells and the patient’s actual body weight, CD34+ cells/kg body weight was determined. Colony assays CFU-GM assays were performed by culture in Iscove’s Statistical methods modified Dulbecco’s medium (Gibco Laboratories, New York, USA) containing 3% agar, 15% fetal calf serum A correlation matrix was used to determine all possible per- (FCS) supplemented with rhIL-6 (6.66 ng/ml), rhG-CSF mutations of the available parameters and statistical analy- (3.33 ng/ml), rhGM-CSF (1.33 ng/ml) and rhSCF sis of correlation between the various measured and calcu- (1.0 ng/ml). Cultures were plated in duplicate at 1 × 105 lated parameters was performed using linear regression cells ml and incubated at 37°C for 14 days in a humidified analysis. This included three separate sets of correlations;

environment with 5% CO2. Colonies containing more than a correlation of the PB parameters to one another, corre- 50 cells were scored. The CFU-GM content of the autograft lation of the PB parameters to the PBPC parameters and was then calculated from the colonies scored per ml and correlation of the PBPC parameters with the overall PBPC the total cells/kg. yield (CD34+ cells/kg). Statistics were performed on StatXact 3 for windows (CYTEL Software, Cambridge, MA, USA) and S-PLUS version 3.3 for Windows (StatSci; CD34 analysis MathSoft, Seattle, WA, USA). The PBPCs and their accompanying PB samples were analyzed for the percentage of CD34+ cells by bivariate ﬂow Results Table 2 Mobilization strategies employed Two hundred and thirty-ﬁve consecutive apheresis pro- Mobilization regimen Number of cedures were performed. The median and mean number of patients procedures per patient was 4.0 (range: 2–10) (Table 1). At the time of initiating apheresis, the median peripheral blood 2 + a Cyclophosphamide 1.5 g/m rhGCSF 29 + × 9 Cyclophosphamide 4.0 g/m2 + rhGCSFa 5 WBC, MNC and CD34 cell counts were 27.1 10 /l 9 Ifosfamide 5 g/m2 + doxorubicin 50 mg/m2 + 2 (range, 1.5–139.6), 15.2 × 10 /l (range, 0.1–56.9) and rhGCSFa 39 × 106/l (range, 0–285.6), respectively. The median %PB Ifosfamide 5 g/m2 + paclitaxel 175 mg/m2 + rhGCSFa 4 CD34+ cells was 0.21 (range, 0–2.65). In the 235 PBPC 2 + 2 Cyclophosphamide 1.5 g/m doxorubicin 50 mg/m 5 products, there was a median WBC count of 310.6 × 109/l + rhGCSFa × 8 Other chemotherapy and rhGCSFa 7 (range, 21.0–668.0), MNC/kg of 2.27 10 (range, 0.02– rhGCSF alone 7 20.55), CFU-GM/kg of 7.84 × 104 (range, 0.08–95.67), %CD34+ cells of 0.81 (range, 0–10.67) and CD34+ cells/kg arhGCSF (10 ␮g/kg) commenced 24 h after completion of chemotherapy. of 1.81 × 106 (range, 0–19.15). Blood CD34 and autograft yield P Chapple et al 127 104 PB MNC (r = 0.2911) and PB platelet count (r = 0.0901) was poor. The relationship between the PB and autograft parameters is detailed in Table 4. There was a strong linear 103 relationship between PB CD34+ cells/l and PBPC CD34+ cells/kg (r = 0.8477; P Ͻ 0.005). The PB CD34+ cells/l was less strongly related to the autograft % CD34+ cells (r = 0.5767; P Ͻ 0.05), CFU-GM/kg (r = 0.5512; 102 P Ͻ 0.05) and MNC/kg (r = 0.5001; P Ͻ 0.05). The peripheral blood % CD34+ cells also showed a modest correlation

Anti-CD34 PE with a number of PBPC autograft parameters; namely PBPC % CD34+ cells (r = 0.5938; P Ͻ 0.05) and PBPC 101 CD34+ cells/kg (r = 0.4446; P = 0.06). Figure 2 demonstrates the corresponding regression plots for these correlations. Figure 2a, demonstrates the correlation of peripheral blood CD34+ cells/l and PBPC CD34+ cells/kg 100 harvested, these same data are also represented on a log- 100 101 102 103 104 log scale in Figure 3. The correlation between these two Ͻ Anti-CD45 PerCP parameters is 0.8477 by linear regression (P 0.005). These two parameters showed the strongest correlation of Region statistics all parameters studied. We further performed multi- parameter statistical analysis to examine the relationship File: R896301.002 between various PB parameters (PB CD34+ cells/l, % PB Log data units: linear values CD34+ cells, PB MNC, PB WCC and PB platelets) and Gate: G1 PBPC CD34+ cell yield. By multivariate analysis, the PB Gated events: 115496 CD34+ cells/l was the only independently significant pre- Total events: 120000 dictor of autograft CD34+ cell content (P Ͻ 0.005). The correlation between autograft CD34+ cells/kg and Region Events % Gated % Total other PBPC autograft parameters is detailed in Table 5. CD34+ 1143 0.99 0.95 Low-mid CD45 115496 100.00 96.25 There was a significant correlation with PBPC CFU-GM/kg (r = 0.6429; P Ͻ 0.05). Figure 1 CD34 analysis. The PBPCs and their accompanying PB + On the basis of the significant correlation between peri- samples were analyzed for the percentage of CD34 cells by bivariate flow pheral blood CD34+ cells/l and PBPC CD34+ cells/kg, we cytometry using phycoerythrin (PE)-conjugated antibodies to the CD34 further examined this relationship to determine a threshold antigen and peridium chlorophyl protein (PerCP)-conjugated anti-CD45 + using an in-house whole blood lysis methodology which has been verified value for peripheral blood CD34 cells that would predict against the ISHAGE guidelines. Cells were analyzed by acquiring flor- a satisfactory harvest. In our institution a total autograft escence and forward light scatter data on 120 000 ungated events. In all + × 6 + + CD34 cell content of 2 10 /kg CD34 cells reliably pre- list mode data analysis ‘true’ CD34 events were defined as having the dicts engraftment post-transplant. Consequently, we aim to dual characteristics of bright CD34 expression and low to intermediate 6 + CD45 expression. This representative example demonstrates 0.95% obtain a minimum of 0.5 × 10 /kg CD34 cells per harvest CD34+ events. procedure which translates into a maximum of four procedures per patient. Using the equation of the line of best fit (PBPC CD34+ cells/kg = 0.045 PB CD34+ cells/l + 0.24) The correlations between the various PB parameters are and a target CD34+ cell yield of 0.5 × 106/kg, we demon- detailed in Table 3. Not unexpectedly there were significant strate that a PB CD34+ cell count of Ͼ5.0 × 106/l predicts = correlations between PB WBC and PB MNC (r 0.8079; for a satisfactory harvest (Figure 3). Ͻ + + P 0.005) and between PB % CD34 cells and PB CD34 We next sought to re-examine the original data set and = Ͻ cells/l (r 0.4167; P 0.05). However, the correlation determine the incidence of false-negative and false-positive + = between PB CD34 cell count and PB WBC (r 0.2468), outcomes. False-negatives, ie a good collection (PBPC CD34+ cells Ͼ 0.5 × 106/kg), when PB CD34+ cells pre- + Ͻ × 6 Table 3 Correlation of peripheral blood parameters (r values) dicts a poor collection (PB CD34 cells 5.0 10 /l) occurred with two collections in the data set (0.9%). These + × 6 CD34% CD34 × 106/l PB MNC two cases had PB CD34 cell levels of 3.9 10 /l and 3.1 × 106/l with harvest yields of 0.8 × 106/kg and 6 WBC ×109/l −0.2772 0.2468 0.8079b 0.57 × 10 /kg, respectively. These two cases were therefore + CD34% 0.4167a −0.2700 borderline for both PB and autograft CD34 cell values. CD34 ×106/l 0.4167a 0.2911 A false-positive result, ie a poor collection (PBPC CD34+ − PB MNC 0.2700 0.2911 cells Ͻ 0.5 × 106/kg), when PB CD34+ cell count predicts PB Platelets ×109/l 0.1068 0.0901 0.1782 for an adequate collection (PB CD34+ cells Ͼ5.0 × 106/l) = = = occurred on 34 occasions (14%) in six patients. In 26 of WBC white blood cells; MNC mononuclear cells; PB peripheral + blood. these occasions the autograft CD34 cell content was aP Ͻ 0.05; bP Ͻ 0.005. Ͻ0.3 × 106/kg. This analysis of outliers demonstrates that Blood CD34 and autograft yield P Chapple et al 128 Table 4 Correlation of peripheral blood parameters vs PBPC parameters (r values)

PBPC WBC ×109/l PBPC MNC/kg PBPC CD34% PBPC CD34 ×106/kg PBPC CFU-GM/kg

WBC ×109/l 0.1471 0.3427 −0.1218 0.0684 0.1953 CD34% −0.1154 0.0464 0.5938a 0.4446 0.2097 CD34 ×106/l 0.0932 0.5001a 0.5767a 0.8477b 0.5512a PB MNC/l 0.0961 0.3291 −0.0939 0.1518 0.2423 PB platelets ×109/l 0.1428 0.0601 0.2002 0.2010 0.0148

WBC = white blood cells; MNC = mononuclear cells; PB = peripheral blood; PBPC = peripheral blood progenitor cell; CFU-GM = colony-forming units granulocyte–macrophage; kg = kilogram. aP Ͻ 0.05; bP Ͻ 0.005.

a b 25.00 r = 0.8477 12.00 r = 0.5767 /kg

6 20.00 10.00

10 8.00 × 15.00 6.00 10.00 4.00

5.00 PBSC Cd34% 2.00 0.00 0.00 PBSC CD34 0.00 50.00 100.00 150.00 200.00 250.00300.00250.00 0.00 50.00 100.00150.00 200.00 250.00300.00 PB CD34 ×106/l PB CD34 ×106/l

c d

120 r = 0.5520 25.00 r = 0.5011 100 20.00 80 15.00 60 40 10.00 20 5.00 PBSC MNC/kg

PBSC CFU-GM/kg 0 0.00 0.00 50.00 100.00150.00 200.00 250.00 300.00 0.00 50.00100.00 150.00200.00 250.00300.00

PB CD34 ×106/l PB CD34 ×106/l

Figure 2 Correlation plots. (a) Peripheral blood (PB) CD34 × 106/l vs peripheral blood progenitor cell (PBPC) CD34 × 106/kg; (b) PB CD34 × 106/l vs PBPC CD34%; (c) PB CD34 × 106/l vs PBPC colony-forming units granulocyte–macrophage (CFU-GM)/kg; (d) PB CD34 × 106/l vs PBPC mononuclear cells/kg.

the PB CD34+ cell threshold value of 5.0 × 106/l favors the MNC, platelet and CD34+ cell count and the autograft scenario of inadequate collections. However, we believe MNC, CFU-GM and CD34 content. We were particularly that this feature is of clinical relevance in that very few interested in determining whether the peripheral blood adequate collections are missed. CD34+ cell count could reliably predict the autograft CD34+ cell yield. Indeed, there was a strong correlation between the PB CD34+ cells/l and the autograft CD34+ cells/kg Discussion (r = 0.8409; P Ͻ 0.005). There was also a significant, albeit weaker, correlation with autograft CFU-GM/kg. We also Mobilization of PBPC into the circulation results in an examined the WBC, MNC and platelet count in the PB as increase in measurable CD34+ cells.2 Since this phenom- potential predictors of PBPC autograft yield, however, none enon is transient, the optimal timing of PBPC collection is was found to predict harvest outcome. Indeed, PB CD34+ critical to maximize the efficiency of the procedure. A num- cells/l was the only independently significant predictor of ber of criteria have previously been utilized to determine autograft CD34+ cell content (P = 0.005) (Table 4). the time to initiate collection. These include the number of Based on the significant correlation between the PB days post-mobilization, the PB WBC and the PB MNC. CD34+ cell count and the autograft CD34+ cell content, we More recently, others have examined the role of the PB attempted to determine a threshold PB CD34+ cell value CD34 count as a predictor of autograft yield.7–9 Here, we that would predict a satisfactory harvest (Figure 3). In our report the results of a comprehensive comparison of various institution a total autograft CD34+ cell content of 2 × 106/kg PB and autograft parameters to determine the optimal CD34+ cells reliably predicts engraftment post-transplant. strategy for maximizing PBPC yield. Consequently, we aim to obtain a minimum of 0.5 × 106/kg We examined the correlation between the PB WBC, CD34+ cells per harvest procedure which translates into a Blood CD34 and autograft yield P Chapple et al 129 100.000 threshold value of 5.0 × 106/l results in a 14% incidence of inadequate collections. However, we prefer this low PB 10.000 + /kg

6 CD34 cell threshold and believe it to be of clinical rel-

10 evance in that very few adequate collections are missed × 1.000 0.01 0.10 1.0010.00. 100.00 1000.00 (0.9%). Nonetheless, as there is a linear correlation between + 0.100 PB and PBPC CD34 cell values, individual investigators could increase the PB CD34+ cell threshold to reduce the

PBPC CD34 0.010 incidence of inadequate collections and thus reduce the r = 0.8477 number of aphereses performed. However, the price of this 0.001 strategy is one of inevitably missing days when adequate collections could be obtained. 0.000 Four recently published studies have examined the PB CD34 ×106/l relationship between PB CD34+ cell count and autograft CD34+ cell content.7–10 All but one of these studies demonstrated a clinically signiﬁcant correlation between PB CD34+ cells/l and CD34+ cell content in the autograft.7–9 100.000 A further small study demonstrates a correlation between PB CD34+ cell count and autograft CD34+ Thy-1+ cell con- 11 8 +

/kg 10.000 tent. Remes et al demonstrated that a PB CD34 cell 6 × 6 10 count exceeding 50 10 /l predicts a median autograft × 1.000 yield of 4 × 106/kg following two collections. Elliot et al7 0.01 0.10 1.0010.00 100.00 1000.00 demonstrated that a PB CD34+ cell count of Ͼ10 × 106/l 0.100 on the day prior to collection reliably predicted a yield of 6 + 9 PBPC CD34 Ͼ1.0 × 10 CD34 cells/kg. Knudsen et al demonstrated 0.010 that PB CD34+ cell count of 20 × 106/l reliably predicted a r = 0.8477 Ͼ × 6 + 0.001 yield of 1.0 10 /kg CD34 cells. However, in two of the studies,7,8 the investigators used a pre-determined PB + 0.000 CD34 cell count to initiate the PBPC collection. By utiliz- ing this criterion, they biased their studies for patients with PB CD34 ×106/l a good CD34+ cell count in the autograft product. Our Figure 3 Log peripheral blood (PB) CD34 × 106/l vs log peripheral analysis examines all consecutive patients at our institution 6 blood progenitor cell (PBPC) CD34 × 10 /kg (top panel). From this corre- who underwent PBPC mobilization and collection without lation, the equation for the line-of-best-fit was derived (PBPC + = + using the PB CD34 cell count to initiate collection. Conse- CD34/kg 0.045 PB CD34/l 0.24) (lower panel). The horizontal dotted + line indicates a PBPC CD34 yield of 0.5 × 106/kg. This line intersects the quently, we examine a wider spectrum of PB CD34 cell line-of-best-fit at a PB CD34 count of 5.0 × 106/l. Thus PB CD34 in excess counts. This may partly explain the lower threshold value of 5.0 × 106/l predicts for a PBPC yield of greater than 0.5 × 106/kg. As we describe. Furthermore, it is well recognized that there the correlation is linear, a higher PB CD34 threshold would predict for a higher PBPC yield. A corollary of increasing the threshold is that fewer is substantial variation in CD34 detection methodologies ‘inadequate harvests’ would be obtained, however, this is at the expense between institutions and this may also contribute to the dis- of missing some ‘adequate’ collections (see text). crepancy between our study and others. Importantly, we have previously verified our CD34+ cell enumeration methodology with the ISHAGE guidelines.5,6 Consequently, our Table 5 Correlation of PBPC parameters vs PBPC CD34/kg (r values) findings could potentially find wider applicability. It could be argued that our study was biased because of PBPC our criteria to initiate a harvest based on the ANC and mini- × 6 CD34 10 /kg mum platelet values. However, the multivariate analysis + × 9 demonstrates that the PB CD34 cells/l was the only inde- PBPC WBC 10 /l 0.1654 + PBPC MNC/kg 0.5198 pendently significant predictor of autograft CD34 cell PBPC CD34% 0.6825a content (P Ͻ 0.005). PBPC CFU-GM/kg 0.6429a We conclude that PB CD34+ cell count performed on the morning of planned PBPC collection, is a reliable predictor WBC = white blood cells; MNC = mononuclear cells; PB = peripheral of the success of PBPC collections. We demonstrate that a = = blood; PBPC peripheral blood progenitor cell; CFU-GM colony- PB CD34+ cell count of greater than 5.0 × 106/l is predictive forming units granulocyte–macrophage; kg = kilogram. + × 6 aP Ͻ 0.05. of a PBPC CD34 cell yield of at least 0.5 10 /kg. As a consequence of this study we now initiate apheresis collections only when the peripheral blood CD34+ cell count maximum of four procedures per patient. Using the equ- exceeds this threshold. ation of the line of best fit and a target CD34+ cell yield of 0.5 × 106/kg, we demonstrate that a PB CD34+ cell count References Ͼ × 6 of 5.0 10 /l predicts for a satisfactory harvest. Our 1 Orlic D, Bodine DM. What defines a pluripotent hematopoitic analysis of outliers (false-positive and false-negative PB + + stem cell (PHSC): will the real PHSC please stand up! Blood CD34 cell results) demonstrates that the PB CD34 cell 1994; 84: 3991–3994. Blood CD34 and autograft yield P Chapple et al 130 2 Bender JG, To LB, Williams S et al. Defining a therapeutic of CD34-positive cell yield by preceding day CD34-positive dose of peripheral blood stem cells. J Hematother 1992; 1: concentration in peripheral blood. J Clin Oncol 1996; 14: 329–341. 970–973. 3 Sutherland DR, Stewart AK, Keating A. CD34 antigen: mol- 8 Remes K, Matinlauri I, Grenman S et al. Daily measurements ecular features and potential clinical applications. Stem Cells of blood CD34+ cells after stem cell mobilization predict stem 1995; 11 (Suppl. 3): 50–57. cell yield and post-transplant recovery. J Hematother 1997; 6: 4 To LB, Haylock DN, Simmons PJ et al. The biology and clini- 13–19. cal use of blood stem cells. Blood 1997; 89: 2233–2258. 9 Knudsen LM, Gaarsdal E, Jensen L et al. Improved priming 5 Sutherland DR, Anderson L, Keeney M et al. The ISHAGE for mobilization of and optimal timing for, harvest of periph- guidelines for CD34+ cell determination by flow cytometry. J eral blood stem cells. J Hematother 1996; 5: 399–406. Hematother 1996; 5: 213–226. 10 Benjamin RJ, Linsley L, Fountain D et al. Preapheresis per- 6 Chang A, Ma DDF. The influence of flow cytometric gating ipheral blood CD34+ mononuclear cell counts as predictors of strategy on the standardization of CD34+ cell quantitation: an progenitor cell yield. Transplantation 1997; 37: 79–85. Australian multicenter study. J Hematother 1997; 5: 605–616. 11 Stewart AK, Imrie K, Keating A et al. Optimizing the CD34+ 7 Elliot C, Samson DM, Armitage S et al. When to harvest per- and CD34+ Thy-1+ stem cell content of peripheral blood col- ipheral-blood stem cells after mobilization therapy: prediction lections. Exp Hematol 1995; 23: 1619–1627.