Comparison of RFLP-DR beta and serological HLA-DR typing in 1500 individuals.

A group of 1522 individuals were HLA-DR-typed both by the standard serological technique and by the RFLP method. Whereas 11% (n = 164) of the serological typings were technically unsuccessful or doubtful, all typings were successful by RFLP. The results of the remaining 1358 typings revealed a serological error rate of 25%. In 16% a serological "blank" turned out to be a definable allele by RFLP, while in 9% an allele was incorrectly interpreted by serology. Of the individuals tested, 11% were HLA-DR homozygous by RFLP. Our results demonstrate an important clinical potential of RFLP typing for the typing of bone marrow transplant candidates in whom serology often fails, and for kidney transplant candidates with "blanks" or serologically "difficult" HLA antigens.