An Algorithm to Facilitate Successful and Cost Effective Hematopoietic Progenitor Cell (HPC) Harvests Based on Peripheral Blood HPC CD34 Counts

Autologous HPC transplants require the collection of sufficient HPCs to ensure engraftment. Inefficient HPC harvests lead to an over utilization of cytokines (GSCF, Plerixafor), impose a limitation on storage space and lead to a waste of resources, material and human.  We report the  use of an algorithm based on a previously demonstrated positive linear correlation between peripheral blood HPC dose and the product HPC dose, to improve the efficiency of HPC collection in patients undergoing autologous HPC harvests.

Methods: HPC collections on 62 consecutive patients were analyzed. All patients received GCSF priming. Peripheral blood CD34 (PHPC) counts were enumerated in 31 cases after the adoption of the predictive algorithm (Cohort 2), and not in the preceding 31 cases (Cohort 1). PHPC counts were performed from day 4 of GCSF administration. Product CD34 count was expressed per kg of recipient weight. In Cohort 2 the predetermined algorithm was used to decide on the use of Plerixafor and timing of collection.  Based on peripheral blood CD 34 counts (cells/µL) patients in Cohort 2 were grouped into three categories (A:< 7, B:7-10,C: >10). Group A received Plerixafor prior to collection. For group B the decision to use Plerixafor was deferred to the physician. Group C did not receive Plerixafor prior to collection. Further dosing of Plerixafor was determined by the product CD34 count. The use of Plerixafor in the two cohorts was assessed retrospectively. The association between testing the PHPC count and the use of Plerixafor was performed using the Chi Square test. The Wilcoxon Rank Sum test was used to compare the median number of collections between the two groups. The proportional difference in patients receiving Plerixafor between the 2 cohorts was assessed using the Fisher Exact Test.

Results: In Cohort 1 (21 Myeloma  and 10 Lymphoma) 80 collections were performed. In Cohort 2 (19 myeloma, 12 Lymphoma) 75 collections were performed. The target collection was achieved in all patients. Forty three doses of Plerixafor were administered in 14 patients in Cohort 1, and 21 doses in 6 patients in Cohort 2. The average daily charge per collection with Plerixafor was $37728.9 and  $11406, without. There was a significant association between the use of a predictive PHPC count and following the algorithm, and the use of Plerixafor (p = 0.0011). The proportion of collections preceded by Plerixafor was higher when a PHPC count was not performed (54% vs28%, p= 0.0009). Proportionately, more patients received Plerixafor when a PHPC count was not performed (45% vs 19%, p = 0.0279). There was no difference in the number of collections between Cohort 1(Median=2, R=6) and Cohort 2 (Median=2, R=4, p=0.99).

Conclusions: The use a predictive algorithm in deciding on the timing of use of cytokines and HPC collection can lead to potentially significant savings on material resources and cost, and improve the efficiency of collection.