Renal donor-recipient HLA DQB1 typing at the DNA level provides a new avenue to study graft survival (GS) and compatibility. HLA DQB1 genotypes of 63 renal donor-recipient pairs were typed simultaneously by assessing patterns from electrophoresed restriction fragments (PCR-RFLP) of amplified DNA (ADNA) and hybridization patterns of sequence specific oligonucleotides (PCR-SSO) of ADNA. Thus the clinical applicability of these two protocols for HLA DQB1 typing was assessed in addition to the compatibility study. Typing results of these two protocols gave overall agreeable results. Sixty-seven per cent of 150 alleles from 75 heterozygotes typed by both protocols had identical allelic type assignments. Serotyping shared more concordant results to PCR-RFLP determined types than to PCR-SSO determined types. The PCR-RFLP protocol can be easily implemented for clinical DNA typing because of its clarity in assigning allelic types and the possible handling of a small number of typing samples, even a single sample, in a single run. The degree of compatibility of these donor-recipient pairs was measured by matching (M) or mismatching (MM) the PCR-RFLP determined DQB1 allelic types. The Kaplan-Meier method was used to estimate GS. Significantly higher GS rates (P < 0.03) were found in donor-recipient pairs with 2 M (GS:74%) as compared with those with 1 M (GS:68%) or 0 M (GS:40%). Higher graft survival was also associated with 0 MM (GS:86%) compared with those with 1 MM (GS:60%) or 2 MM (GS:40%), although the significance level is P = 0.08 for both Mantel-Cox and Breslow tests. These findings indicate the importance of determining HLA DQB1 molecular alleles for assessing GS.