In the present study PCR primers were designed for detecting all phenotypically expressed DQB1 and DQA1 allelic variability, 19 and 10 alleles, respectively, by PCR amplification with sequence-specific primers (PCR-SSP). For DQB1 typing, each sample was amplified by a first set of 14 PCR primer pairs, followed in some cases by two to six additional PCR reactions. The first 14 primer pairs allowed the identification/separation of all but a few of the recently described DQB1 alleles: DQB1*0504, DQB1*0605, DQB1*0606 and DQB1*0607 would not be identified; DQB1*0603 and DQB1*0608; and DQB1*0301 and DQB1*0304, respectively, would not be distinguished. Therefore an additional set of eight DQB1 primer pairs was used for a complete DQB1 typing, including all homozygous and heterozygous combinations. For DQA1 typing, 12 PCR reactions were performed per sample, 10 for detecting variability within the second exon and two for identifying first exon polymorphism. All homozygous and heterzoygous combinations of DQA1 alleles could be resolved by these primer pairs. In addition, four primer mixes were designed for determining codon 57 of the HLA-DQB1 gene. Thirty cell lines and 120 individuals were investigated by the DQB1 and DQA1 PCR-SSP technique, as well as with the HLA-DQ beta 57 primers. The concordance between PCR-SSP typing and assigning DQB1 and DQA1 alleles from TaqI DRB-DQA-DQB RFLP analysis was 100%. The reproducibility was 100% in 30 samples investigated on two separate occasions. Amplification patterns, investigated in 15 nuclear families, segregated according to dominant Mendelian inheritance. DQB1 and DQA1 PCR-SSP typing can be performed in 2 hours, including DNA extraction, PCR amplification and post-amplification processing. The method is technically simple and the typings are easy to interpret. The cost for typing one individual is low and is independent of the number of samples analyzed simultaneously, i.e. the technique is well-suited for routine clinical use.