Comparison of HLA class I and II molecular and serological typing within clinical laboratory

In order to develop an adequate tissue typing strategy, we performed human leukocyte antigen (HLA)-A, B and DR generic typing on 235 (120 HLA-A, B and 115 HLA-DR) routine clinical samples by polymerase -chain reaction (PCR)-SSP in parallel with conventional serological typing. At the A locus, there were two (1.7%) discrepancies between molecular and serological typing besides 25 (20.8%) serological blanks, which was defined by molecular typing. At the B locus, there were two (1.7%) discrepancies and 30 serological blanks, which were defined by molecular typing. At the DR locus there were two (1.8%) discrepancies and 44 serological blanks, which were defined by molecular typing. We conclude that molecular typing is of substantial benefit in the resolution of poorly defined serological antigens. In view of the low percentage of discrepancy between the serological and molecular typing besides the high cost of molecular typing, our policy is to perform HLA typing first by the serological method and to use PCR-SSP as an adjuvant tool.