Related Experiment Video
Updated: Sep 11, 2025

Preparation and Pathogen Inactivation of Double Dose Buffy Coat Platelet Products using the INTERCEPT Blood System
Published on: December 7, 2012
Three buffy coat platelet concentrates: Use of data modelling to make more with less
Michael Cahillane1, Nicola Pearce1, Christine Saunders1
1Welsh Blood Service, Pontyclun, UK.
Background And Objectives:
The Welsh Blood Service (WBS) produces pooled platelet concentrates from four buffy coats (4BC-PC), achieving consistently high platelet yields exceeding the UK specification. Three buffy coat platelet concentrates (3BC-PC) offer the potential for improved supply chain resilience and increased production capacity. Data modelling was performed to evaluate the local feasibility of 3BC-PC production.
Materials And Methods:
Platelet recovery (%) was calculated from 41 4BC-PC to generate three reference values (maximum, average and minimum). Measurements from the 164 BCs were used to model 3BC-PC production, with further analysis determining the impacts of manufacturing variability. Calculated yields were assessed against UK, European (EU) and UK contingency specifications. Potential production increases for WBS were estimated alongside statistical significance and effect size of the modelling.
Results:
Data modelling with 3BC-PC showed 100% compliance against UK minimum requirements approved as a contingency measure (≥75% achieving ≥150 × 109/unit), while compliance to EU specification (≥90% achieving ≥200 × 109/unit) ranged from 93.2% to 100%. Compliance to routine UK specification (≥75% achieving ≥240 × 109/unit) varied for the three modelled scenarios (54.5%-86.8%). Modelling indicated potential increases in platelet concentrate (PC) production, with notable rises for Groups A and O components of approximately 30%-40%.
Conclusion:
Data modelling supports 3BC-PC as a feasible alternative to 4BC-PC, offering increased production capacity. Modelled 3BC-PC met UK contingency and EU specifications. Compliance with the UK standard specification was variable and suggests the need for further optimization. Real-world validation would confirm whether this is a practical proposition.

