Test performance in systemic sclerosis: anti-centromere and anti-Scl-70 antibodies.

PURPOSE To determine the sensitivity and specificity of anti-centromere (ACA) and anti-Scl-70 antibodies in systemic sclerosis (SSc). METHODS Four-hundred ninety-seven English language articles published from 1966 to 1994 were identified by structured MEDLINE search. Articles in which either ACA or anti-Scl-70 antibodies were measured in both SSc patients and a non-SSc control group were reviewed and rated using a previously published diagnostic testing scale. Reported sensitivity and specificity from each study was converted into a 2 x 2 table, and combined across studies to calculate summary rates for each antibody. Author's clinical classification criteria for SSc served as the gold standard for disease diagnosis. RESULTS In 30 articles that fulfilled inclusion criteria, ACA were found in 441 of 1,379 SSc patients (sensitivity 32%, range 17% to 56%). This increased to 57% (332 of 585) in patients with the limited cutaneous, or CREST, subset of SSc (IcSSc). Anti-Scl-70 antibodies were found in 366 of 1,074 SSc patients (sensitivity 34%, range 3% to 75%), and this increased slightly to 40% in patients with the diffuse cutaneous form of SSc (dcSSc). Both antibodies were measured in 670 patients, and either test was positive in 58% (range 29% to 86%), but in only 3 patients were both antibodies present. The specificity of each antibody was high, but varied by control group. ACA were present in 5% and anti-Scl-70 antibodies were present in 2% of patients with other connective tissue diseases, but fewer than 1% of disease free controls had either antibody present. CONCLUSIONS As individual diagnostic tests in SSc, both ACA and anti-Scl-70 antibodies are highly specific. Each performs somewhat better as discriminators of clinical subsets for patients in whom a diagnosis of SSc has already been established. Clinicians can rely on a positive test result as being specific in the detection of disease, but 40% of SSc patients are likely to have neither antibody present, and a negative result does not exclude the diagnosis. Measurement of these antibodies should be considered secondary to the clinical features when making a diagnosis of SSc.