A model for studying the hemostatic consumption or destruction of platelets

Mark R. Dowling, Emma C. Josefsson, Katya J Henley, Benjamin T. Kile, Philip D Hodgkin

Research output: Contribution to journalArticleResearchpeer-review

6 Citations (Scopus)


A fundamental issue in understanding homeostasis of the hematopoietic system is to what extent intrinsic and extrinsic factors regulate cell fate. We recently revisited this issue for the case of blood platelets and concluded that platelet life span is largely regulated by internal factors, in contrast to the long-held view that accumulated damage from the environment triggers clearance. However, it is known that in humans there is an ongoing fixed requirement for platelets to maintain hemostasis and prevent bleeding; hence a proportion of platelets may be consumed in such processes before the end of their natural life span. Whether it is possible to detect this random loss of platelets in normal individuals at steady-state is unknown. To address this question, we have developed a mathematical model that independently incorporates age-independent random loss and age-dependent natural senescent clearance. By fitting to population survival curves, we illustrate the application of the model in quantifying the fixed requirement for platelets to maintain hemostasis in mice, and discuss the relationship with previous work in humans. Our results suggest a higher requirement for platelets in mice than in humans, however experimental uncertainty in the data limits our ability to constrain this quantity. We then explored the relationship between experimental uncertainty and parameter constraint using simulated data. We conclude that in order to provide useful constraint on the random loss fraction the standard error in the mean of the data must be reduced substantially, either through improving experimental uncertainty or increasing the number of experimental replicates to impractical levels. Finally we find that parameter constraint is improved at higher values of the random loss fraction; thus the model find utility in situations where the random loss fraction is expected to be high, for example during active bleeding or some types of thrombocytopenia.
Original languageEnglish
Article numbere57783
Number of pages14
JournalPLoS ONE
Issue number3
Publication statusPublished - 7 Mar 2013
Externally publishedYes


  • platelets
  • monte carlo method
  • curve fitting
  • simulation and modeling
  • blood
  • cell death
  • confidence intervals
  • gaussian noise

Cite this