Associations of HLA antigens with many of the rheumatic diseases have been established over the last two decades. Although these discoveries provide potential new insights into disease pathogenesis, the clinical utility of HLA typing has been limited. The major exception is that of HLA-B27 in the spondyloarthropathies, where clinical uses of HLA-B27 testing has permitted identification of a large spectrum of disease that was previously misdiagnosed and misclassified. HLA-B27 remains potentially useful in the diagnosis of atypical spondyloarthropathies because of its high frequency in patients with these diseases (yielding good sensitivity) and its relatively low frequencies in most normal populations (yielding good specificity). Its predictive value in individual cases, however, depends on the quality of the physician's assessment of the likelihood of a spondyloarthropathy. In patients with juvenile-onset arthritis, typing for HLA-B27, as well as several HLA-class II alleles (DR5, DR8, DP2, and DP3), may prove to be useful in diagnosis and classification; however, additional studies are necessary. HLA oligotyping of DNA in patients with early rheumatoid arthritis to determine homozygosity versus heterozygosity for the DRB1 susceptibility sequence promises a potential new parameter for predicting clinical disease severity, and thus the possible early initiation of more aggressive therapies. Additional studies are necessary, however, to determine the validity of this approach. Finally, the future diagnosis, prevention, and treatments of these diseases may depend on the identification and manipulation of specific immune responses mediated by HLA molecules, thus making HLA typing for clinical purposes routine.