COMPARISON OF SSP vs. SEROLOGIC TYPING OF HLA CLASS I.
T Rami, W Schaefer, J Sholander, JM Hart, AA Zachary, MS Leffell, Department of Medicine, Johns Hopkins University, Baltimore, MD

---

We compared results of 195 class I typings performed by both SSP (PelFreez ) and serology on 61 end-stage organ failure patients, 61 hematology patients, and 73 healthy controls. Results of the two tests were discrepant in 119 (61%) cases. Disparity rates for the A and B loci were approximately proportional to the number of antigens tested at each locus (~1:2). However, considering the correlation coefficients (r), 37.5% of A antigens had r>0.9 and 16.7% had r<0.7 while for the B locus, these values were 25.5% and 34.0% (P=0.12). Disparities occurred, most frequently, among the most recently defined antigens which are, by definition, the most difficult to identify immunologically. Increased disparity was seen, also, with the antigens of theA10 and B7 groups. We evaluated the SSP results obtained with the phenotypes of 94 subjects tested at the allele level and found no errors in the SSP results. In contrast, we found that serologic typing results were incorrect or indeterminate in 7.5% of 1579 tests performed. We found the requirement for additional testing to be comparable in the 2 assays. However, in SSP this was almost always necessitated by a technical problem while in serology, the this was due both to technical problems and uninterpretable or ambiguous results, not all of which were resolved by additional immunologic testing. Because serologic equivalents are not known for all class I alleles, some SSP results include Aunknown serologic equivalent@ as a possible interpretation. This occurred in 120 (61%) of the 195 tests evaluated. In our laboratory, the cost (reagents + personnel time) of SSP is ~0.76 that of serologic typing. Based on our findings, we believe that DNA-based testing is advantageous for hematology patients where cell type and number may be altered or where increased resolution is important, but that until the serologic equivalents of all alleles are known, serologic testing remains more relevant for solid organ transplantation where it provides information important to crossmatch interpretation.