Transnational validation of the Australian algorithm for virtual crossmatch allocation in kidney paired donation.

An independent pool of 16 incompatible live donor-recipient pairs registered at the Vienna transplant unit was applied to test whether virtual crossmatch allocation used in the Australian kidney paired donation (KPD) program reliably predicts negative crossmatches. High resolution HLA data were entered into the computer-matching algorithm and allocation was performed excluding any DSA>2000MFI. CDC and flow crossmatch data of recipients against any of the donors were available for 112 crossmatch combinations. The computer program identified 19 possible pairings in 2-way or 3-way chains in multiple combinations. The top ranked combination included one 3-way and two 2-way ABO-compatible chains. Where crossmatches were available all recipients were CDC crossmatch negative with the computer-matched donor. Excluding allocation of KPD donors in the presence of DSA>2000MFI had a negative predictive of 99.9% for CDC and 96.4% for flow crossmatch. In the 12 pairings with ⩾1 DSA against crossmatched donors there was a negative CDC and flow crossmatch. These results show excellent correlation between matching using virtual crossmatch and actual crossmatch results. Using the 2000MFI cut-off the number of potentially unacceptable CDC and flow crossmatch positive pairings identified by virtual crossmatching is low, but some potential crossmatch negative pairings are missed.