Objective. The analysis of genetic markers of rheumatoid arthritis (RA) in a population in which the DR4 serotype is not strongly associated with the disease.Methods. Chilean RA patients (56 seropositive and 22 seronegative) and 141 controls were studied by serotyping. Southern blot analysis of Bum HI restriction fragment length polymorphism (RFLP) was done in genomic DNA from 46 patients with seropositive RA, 17 patients with seronegative RA, and 45 controls, using a complementary DNA probe specific for DRBl genes.Results. The prevalence of the HLA-DR9 haplotype was strikingly higher in seropositive RA patients (21 %) than in controls (3%) (P,,,, < 0.0008, by Fisher's exact test; relative risk [RR] = 9.34). The prevalence of DR4 and DR1 haplotypes, although slightly increased, did not achieve a significant preponderance. The simul- taneous presence of two Bum HI fragments (3.6 kb and 4.5 kb) was found with higher prevalence in seropositive patients (83%; RR = 9; P , , , < 0.00002) than in controls (36%), and seemed higher in seronegative RA patients as well (71%; RR = 4). Furthermore, its prevalence remained increased in comparisons of DR4 positive controls (36%) with DR4 positive seropositive patients (100%; RR = 67; P,,,, < 0.0002) and DR4 positive seronegative patients (100%; RR = 36; P,,,, < 0.006)' even after excluding the DR9 positive individuals. A tendency toward higher association with DR1 seropositive RA patients (67%; RR = 12), a group with no DR4 or DR9 positive individuals, than in DR1 positive controls (14%)' was also observed.Conclusion. The HLA-DR9 haplotype was definitively consolidated as a very strong genetic marker exclusively for seropositive RA in Chilean patients, as suggested by our previous observations. RFLP analysis showed that the simultaneous presence of 3.6-kb and 4.5-kb Bum HI fragments constituted a better RA marker than did any of the heretofore studied haplotypes. These fragments together would be linked to RA independently of the DR1, DR4, and DR9 haplotypes. The overall evidence indicates that Chilean seropositive RA patients display a genetic background that is different from that underlying RA susceptibility in other populations and suggests the existence of common, as well as distinct, genetic elements predisposing to seronegative and seropositive RA.The genetic basis of the susceptibility to the development of rheumatoid arthritis (RA) has not yet been elucidated. The extent to which similar or distinct HLA class I1 alleles may act as risk-conferring factors in different populations and ethnic groups