In 135 patients with systemic scleroderma, we compared three different methods to determine antinuclear autoantibody (ANA) specificity: indirect immunofluorescence, double immunodiffusion, and, employing recombinant antigens, immunoblotting using both marker autoantigens of this disease. A characteristic Scl-70 antibody pattern was found on HEp-2 cells in 83.8% of the patients, double immunodiffusion was positive for the Scl-70 antibodies in 81.9%, and immunoblot with the recombinant topoisomerase I (Topo I) was positive in 71% of the patients. For the centromere autoantibodies we found a high concordance between the anticentromere antibody (ACA) pattern on HEp-2 cells (27 patients positive) and the detection of recombinant kinetochore in immunoblotting (26 patients positive). The three testing techniques gave comparable results, except that the Topo I recombinant antigen used in immunoblotting reacted strongly with fewer than expected of the known Scl-70-positive sera. However, a method using recombinant antigens expressing all epitopes (rather than one of the epitopes of Topo I) will undoubtedly become the method of choice for detecting antibodies in systemic scleroderma. Using the immunoblotting technique with the recombinant antigens we detected in four patients antibodies against both Topo I and kinetochore. More severe symptoms of systemic scleroderma were found in patients who had both antibodies. The combined presence of both marker autoantibodies is therefore not as rare as previously reported and may predict severe disease.