Author(s): Slater CA, Davis RB, Shmerling RH, Slater CA, Davis RB, Shmerling RH
Abstract Share this page
Abstract OBJECTIVE: To determine the clinical utility of the antinuclear antibody (ANA) test as ordered in a large teaching hospital. METHODS: Retrospective chart review in a 400-bed teaching hospital that provides care for hospital-based and community-based practices. PATIENTS: A consecutive sample of 1010 patients (including inpatients and outpatients) for whom ANA testing was ordered over 10 months; all patients with positive ANA test results and an equal number of randomly selected patients with negative test results were included. Clinical utility of the ANA in the identification of rheumatic disease was determined by its estimated sensitivity, specificity, and positive and negative predictive values. RESULTS: Of 1010 ANA test results reviewed, 153 were positive. The group with positive ANA test results included more patients aged 65 years or older than the group with negative ANA test results (30\% vs 15\%, P < .003). The diagnosis of systemic lupus erythematosus (SLE) was established in 17 patients, all of whom had positive ANA test results. Other rheumatic diseases were found in an additional 22 patients. The estimated sensitivity and specificity of the ANA test for SLE were 100\% and 86\%, respectively. For other rheumatic diseases, sensitivity and specificity were 42\% and 85\%, respectively. The positive predictive value of the ANA test was 11\% for SLE and 11\% for other rheumatic diseases. Specificity and positive predictive value for ANA testing in the elderly patients were lower than among younger patients. CONCLUSIONS: The sensitivity of the ANA test for SLE was high, but overall the positive predictive value was low for SLE or other rheumatic diseases. Sensitivity was low for ANA testing among patients with non-SLE rheumatic disease. More selective test ordering might improve the clinical utility of this test. Clinicians ordering the ANA test should be aware of the test's low-positive predictive value in settings with a low prevalence of rheumatic disease, particularly among older patients.
This article was published in Arch Intern Med
and referenced in Journal of Drug Metabolism & Toxicology