Seventy-seven patients with Raynaud's disease were studied for a mean of 4 years (range 1-11 years) to determine the relationship between autoantibodies and long-term clinical outcome. Anticentromere antibodies (ACA) were assayed by indirect immunofluorescence and by immunoblotting of HeLa cell chromosome extracts. Antibodies to topoisomerase I (anti-topo I) were assayed by immunodiffusion and immunoblotting. Antibodies to the major centromeric protein, CENP-B, and anti-topo I were studied by enzyme-linked immunosorbent assay (ELISA). Eight patients developed telangiectasias, 4 developed skin tightening, and 4 developed a connective tissue disease other than scleroderma. The presence of ACA at the start of the study was associated with the development of telangiectasias (P less than 0.003). An initial 100-kd band on immunoblot in conjunction with a positive anti-topo I ELISA result was associated with the development of tight skin (P less than 0.0025), while a 100-kd band with a negative anti-topo I ELISA result was associated with the subsequent development of a connective tissue disease other than scleroderma (P less than 0.0073). Patients who were initially ACA positive, had the 100-kd band on immunoblot, or had positive ELISA results for anti-topo I or for anti-CENP-B were 63-fold more likely to develop signs of connective tissue disease by the end of the study (P less than 0.000009). The presence of any of these autoantibodies was more sensitive (100%), although less specific (75%), than were findings from nailfold capillaroscopy (sensitivity 67% and specificity 95%) in predicting subsequent clinical progression. We conclude that findings of assays for anti-topo I and ACA complement the findings from nailfold capillaroscopy in providing useful prognostic information in Raynaud's disease.