ANALYSIS OF HLA DISCREPANCIES IN THE UNOS/OPTN REGISTRY 1995-2001.

4.1
#61-OR
ANALYSIS OF HLA DISCREPANCIES IN THE UNOS/OPTN REGISTRY 1995-2001.
L. R. Cagle , S. N. Miller , S. Warnell , C. Bryan , B. Colombe , C. Tolleris , A. Ting and B. Freed . University of Colorado Health Sciences Center ClinImmune Labs ; Midwest Transplant Network ; Thomas Jefferson Univ Hosp and UNOS Research Dept .

Accurate typing of HLA-A,B,DR antigens plays a critical role in the allocation of zero antigen mismatched kidneys within the UNOS/OPTN. To evaluate the accuracy of donor HLA data, the discrepancy rate was assessed for the period, 1995-2001. HLA data can be entered in up to three areas within the registry: 1) match run for initial organ placement, 2) donor histocompatibility form for final typing, 3) recipient histocompatibility form for donor re-typing. 25,398 DONOR typings were analyzed giving a total of 152,388 ANTIGEN comparisons (2A, 2B, 2DR x 25,398). Discrepant antigen comparisons were found in 10,732 donor typings; 9,402 of these were identified as equivalent when checked against the UNOS HLA equivalence tables. From the remaining 1,330 DONORS, 1,866 ANTIGEN discrepancies were identified and assigned a severity grade of 1-4. 1=parent/split antigen (not recognized as equivalent), 2=antigen of same CREG, 3=major antigen discrepancy (e.g. antigen vs. blank), and 4=clerical error. Of the 1,866 ANTIGEN discrepancies, 296 (16%) were grade 1; 233 (12%) were grade 2; 780 (42%) were grade 3; and 557 (30%) were grade 4. Grades >1 required follow-up by the HLA labs to verify or correct the data. The total DONOR discrepancy rate was 5.24% and the ANTIGEN discrepancy rate was 1.22%. When compared to donor and antigen discrepancy rates of 14% and 2.7% from 1987-1995 by Colombe and Lou, the discrepancy rates have decreased significantly over time. The quality of the donor HLA typings has improved which provides greater confidence for zero antigen mismatch allocation. To ensure continuous quality improvement, a system has been implemented in UNet to flag discrepant typings, allowing more efficient identification and resolution of typing discrepancies.