#42
AN ALGORITHM FOR HLA MISMATCH ANALYSIS.
Andrew L. Rivard 1, Roger M. Christen 2, Shoeb Mohammed 1, Andrea S. Huang 1, Jennifer E. Hrabe 1, Richard W. Bianco 1, Harriet J. Noreen 3 and Leslie A. Miller 4. 1 Department of Surgery, University of Minnesota ; 2 Electrical Engineering and Computer Science ; 3 Fairview-University Medical Center and 4 Department of Medicine, Division of Cardiology .
Due to the short preservation time of donor hearts, HLA data is typically not submitted for a prospective computerized matching by UNOS for heart transplant recipients. To ease data analysis of heart transplant HLA loci, we desired a mismatch algorithm independent of that used for the match run.
To achieve this, a relational database was built using Microsoft Access 2003. Validation and relationship rules were implemented for data consistency and accuracy. Visual Basic was used to create an algorithm based upon unidirectional (donor to recipient) mismatching and equivalency tables from UNOS. Following an equivalency search, the algorithm calculates A, B, and DR loci in two separate logical steps. Calculated mismatches from 90 heart transplants provided proof of concept.
Fifteen pairs were not evaluable, resulting in mismatch data for 75 transplants. Percentages of donor-recipient mismatches are: 1=1%, 2=4%, 3=16%, 4=31%, 5=34%, 6=14% (mean 4.3
1.1). The accuracy of transcription and mismatches (using a 2003 equivalency table) was verified.
We created an accurate and valid mismatch algorithm for donor-recipient HLA typing. This provides a tool for independent analysis of HLA mismatching at a transplant center.