Validation of a Predictive Formula for Collection of Hematopoietic Progenitor Cells (HPC) By Leukapheresis at 2 Institutions Using 4 Different Machine Protocols

Introduction: We evaluated a predictive formula for CD34+ cell yield from HPC collections at 2 institutions using different mobilization and machine parameters on the COBE® Spectra (Spectra) and Spectra Optia® (Optia) devices.

Patients and Methods: Heidelberg: Data were reviewed for all 67 pts who collected autologous HPC from 2/13/14 to 7/11/14. A total of 89 collection events (n) were analyzed (50 male, 39 female). Diagnoses were myeloma (n=64) and non-myeloma (n=25). Age was 17–74 years (median 60). Mobilization was chemotherapy + G-CSF (5 µg/kg SC daily; n=82) or G-CSF (10 µg/kg SC daily, minimum 4 days; n=7) + plerixafor (240 µg/kg SC) added to G-CSF 15 h before collection in poor mobilizers (n=9). Pts were randomly assigned to collect on either Spectra (n=26) or Optia (n=63). Anticoagulant citrate dextrose (ACD) was used at an inlet:anti-coagulant ratio of 15:1 with inlet flow rate of 40–100 mL/min. A central venous line (n=6) or peripheral access (n=83) was used to process 3–4 blood volumes.

UAMS: Data were reviewed for all 52 pts who collected HPC from 8/13/13 to 7/10/14. A total of 155 collection events were analyzed (90 male, 65 female). Diagnoses were myeloma (n=127) and non-myeloma (n=28). Age was 24–79 years (median 64). Mobilization used chemotherapy + G-CSF (5–8 mg/kg SC twice daily; n=116) or G-CSF (5–8 mg/kg SC twice daily; n=39) + plerixafor (240 µg/kg SC) added to G-CSF 15 h before leukapheresis in poor mobilizers (n=41). All pts had central venous lines. Spectra pts (n=133) had 5–40 L blood processed using 1,000 mL ACD and 5,000 U heparin for anticoagulation at an inlet:anti-coagulant ratio of 31:1, inlet flow rate of 150 mL/min. Collection flow rate was set at 1.5 mL/min and 10 mL ACD was added to the component at processed volumes of 10 L, 20 L and 30 L. Optia pts (n=22) had 7–23 L blood processed using ACD at a ratio of 12:1 for the first chamber, changing to 25:1 during first chamber collection phase, with a maximum inlet flow rate of 125 mL/min. Heparinized citrate was used as with the Spectra, with 3 mL of ACD without heparin added to the product bag before start of collection, and 1 mL of ACD was added for every chamber processed during the procedure after every 10 L, with an aim to keep the concentration in the product bag 10% by volume. Pts were collected randomly on Optia one day and Spectra another and if platelet count was ≥80 K/mL were given 81 mg aspirin at start of collection.

Both institutions: continuous CaCl₂ infusion was used to prevent side effects of citrate for both machines, and CD34+ cells were quantified by flow cytometry using the ISHAGE protocol.

Prediction of CD34+ cells collected/L blood processed was calculated using the formula: (peripheral blood CD34+ cells/µL) × (estimated collection efficiency of 30%) / body weight (kg) (Rosenbaum et al. Cytotherapy, 2012;14:461–466). Predicted CD34+ counts/kg were plotted on the x-axis against corresponding observed counts (y-axis) for all collections and the scatter plots assessed for linear relationship between predicted and observed. Linear regression analyses were performed and the linear correlation coefficients (r-values) were calculated; slope and intercepts of the regression lines were evaluated. The same analyses were performed on the following subgroups: combined data for both institutions, data for Spectra and Optia separately by institution, and Spectra and Optia data separately but combining institutions.

Results: Regression analyses performed on all collection events categorized by institution demonstrated very strong linear correlations between predicted and observed values [r=0.88 (y=1.64x - 0.10) and 0.92 (y=1.14x + 0.88), Heidelberg and UAMS, respectively]. For data combined for both institutions, strong linear correlation persisted [r=0.91 (y=1.17x + 1.31); Figure]. Data categorized by machine used (Spectra vs. Optia) both within institution and combining institutional data showed consistently strong linear correlation (r-values all >0.84).

Conclusions: This is the first report confirming inter-institutional reproducibility of this formula for predicting the minimum number of CD34+ cells that may be expected for a given number of liters of blood processed for an HPC collection. We also demonstrate that inter-institutional differences in mobilization regimens and collection parameters do not affect efficacy of the prediction, nor does machine type used (Spectra vs. Optia).

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal