A 6-Year Experience of a Kidney Paired Donation Centralized Lab, Performing Crossmatches Resulting in 220 Transplants.

Abstract

INTRODUCTION: Flow crossmatch (FXM) results with donor cells were compared to computer-based virtual crossmatch (VXM) results in patients stratified by their calculated PRA (cPRA) values to evaluate the accuracy of both crossmatch methods. METHODS: A total of 1949 VXM tests (with HLA-A, B, C, DR, and DQB) were performed. A single antigen bead assay was used for patients with 1-100% cPRA to create a list of unacceptable HLA antigens; VXM results were used to perform daily match runs. Selected donor/recipient pairs were tested by FXM with donor cells. Weekly match runs were performed to identify compatible pairs with negative VXM results. RESULTS: Between 2007 and 2012, there was an increase in the number of sensitized (1-79% cPRA) and highly sensitized (80-100% cPRA) patients (Figure 1). The accuracy of negative VXM ranged from 70.5% in 2007 to 74% in 2012 (Figure 2). False negative VXM results ranged from 26% in 2012 to 29.5% in 2007. When all negative VXM results were divided by cPRA, patients with 0% cPRA had no false negative results, those with 1-79% cPRA had 20% positive FXM, and those with 80-100% cPRA had 45% positive FXM. Analysis of the class I and II false negative VXM results revealed that the majority were falsely negative in both class I and II HLA (47%), rather than in class II alone (40%) or in class I alone (13%).Figure: No Caption available.Figure: No Caption available.CONCLUSION: Significant (˜ 25%) false negative VXM (especially for high PRA and Class II sensitized patients) when assessed by actual FXM demonstrates that data reported to centralized repositories by multiple transplant centers results in higher false negative VXM than can be achieved by single centers.