Comprehensive Red Blood Cell Matching

Abstract

Advances in Genotyping technology have opened up the possibility of typing donors and patients on many more antigens than A, B and RhD, which is currently the standard. A future where all patients receive extensively matched Red Blood Cell units is now foreseeable. Matching compatibly on all eleven clinically relevant minor antigens would eliminate almost all alloimmunization among patients and thereby improve the quality of Red Blood Cell transfusions. However, strictly compatible matches for all patients on all relevant antigens are unlikely due to an exponential increase in the number of possible phenotypes. Large scale extended matching should therefore neither decrease the availability nor increase the expiration of RBC units. We propose the MINRAR-Online Integer Linear Programming formulation which considers all AB-RhD compatible matches while minimizing the alloimmunization risk for the patients. This is achieved by allowing minor antigen mismatches at a cost based on the immunogenicity of the antigen. Furthermore, the ILP also contains terms in the objective to prevent shortages, outdating and alloimmunization in the long run. The performance is tested in simulations and results show that shortages can be prevented while the alloimmunization risk is lowered compared to previous work. Lastly, we have investigated how the MINRAR-Online ILP can be extended to prioritize certain patient groups in the matching for which there is a larger incentive to prevent alloimmunization. Simulations of a single hospital show that this prioritization can be implemented effectively with patient group specific weights, while multi-hospital simulations show that the availability of extensively matched RBC units is ≥ 99% for all considered patient groups.