Mathematical modeling of platelet survival with implications for optimal transfusion practice in the chronically platelet transfusion-dependent patient


  • J.K Hersh,

    Search for more papers by this author
    • 2

      Clinical Instructor, Tufts Medical School; and Attending Physician, Emergency Medicine, Faulkner Hospital, Boston, MA.

    • 1

      Jeff K. Hersh, PhD, MD, FAAP, Research Consultant, University of North Carolina

  • E.G. Horn,

    Search for more papers by this author
    • 5

      Department of Family Practice, University of Massachusetts, Worcester, MA.

    • 3

      Elaine G. Horn, MD, Transfusion Medicine Fellow, Transfusion Service, University of North Carolina Hospitals

  • M.E. Brecher MD

    Associate Professor, Corresponding authorSearch for more papers by this author

Department of Pathology and Laboratory Medicine, University of North Carolina, and Director, Transfusion Medicine Service, University of North Carolina Hospitals, CB 7600, 101 Manning Drive, Chapel Hill, NC 27514.


BACKGROUND:It is known that in vivo platelet survival varies as the platelet count changes. Previous attempts at curve fitting fail to predict the decreased platelet survival in thrombocythemia. Therefore, mathematical relations that more closely approximate platelet survival were derived and used in models of platelet transfusion practice.

STUDY DESIGN AND METHODS:A differential equation for platelet loss was derived that included a constant (constant homeostatic loss), a first- order term (senescent loss), and a second-order term (one proportional to the square of the platelet concentration and whose contribution is expected to be significant only at higher platelet concentrations). Data derived from this model was compared to platelet survival data in normal, thrombocytopenic, and thrombocythemic patients and to the platelet decay after high-dose chemotherapy. To provide further validation of this model, predicted and actual platelet requirements were calculated or obtained (chart review) in bone marrow patients with uncomplicated thrombocytopenia after ablation and at two platelet- transfusion thresholds (20 and 10 × 10(9)/L).

RESULTS:The equations accurately modeled normal, thrombocytopenic, and thrombocythemic platelet survival. Chart review demonstrated a 12.5 percent reduction in platelet transfusion requirements when the transfusion threshold was reduced from 20 to 10 × 10(9) per L. The model predicted a reduction of 14.0 percent. For 100 days of uncomplicated thrombocytopenia and a transfusion threshold of 10 × 10(9) per L, transfusion of 3 units of platelet concentrates compared to a 6-unit pool of platelet concentrates, resulted in a 22-percent savings of platelet units. CONCLUSION: Platelet survival as a function of platelet concentration can be modeled by use of a differential equation. This model challenges current dogma regarding platelet destruction and predicts decreased platelet survival in thrombocythemic patients. The model illustrates that large doses of platelets would result in greater time between transfusions, however, more units of platelets are used. Consideration should be given to the more frequent use of smaller doses of platelets in patients who chronically require platelet transfusion support.