Evidence-Based Use of Rheumatologic Laboratory Tests

In a recent review, Mourad and colleagues¹ purport to provide an "evidenced-based" approach to fever of unknown origin. However, in a table entitled "Minimal Diagnostic Workup to Qualify as Fever of Unknown Origin," the authors include antinuclear antibodies (ANA) and rheumatoid factor. This is certainly not "evidenced based" and is unfortunate in that it is most likely to yield erroneous diagnoses and unnecessary workup. Recently, systematic analyses of the medical literature were performed in an effort to define the optimal use of the ANA and related tests.^2- 3 When used for the diagnosis of systemic lupus erythematosus (SLE), the generic ANA has a positive likelihood ratio of approximately 2.2 and a negative likelihood ratio of approximately 0.1.³ While virtually all patients with SLE are positive for ANA, this test may also have positive results in many other autoimmune conditions and also among healthy persons (eg, 5%-6% at a titer of 1:160 and 12%-13% at a titer of 1:80). In the setting where the pretest probability of SLE would be expected to be quite low, a positive test result does little to support a diagnosis of SLE. While a negative ANA result militates against a diagnosis of SLE, it does little to exclude other autoimmune conditions. Rheumatoid factor is positive in 80% to 85% of patients with rheumatoid arthritis; by definition, it is positive in 5% of normal controls, with a higher prevalence in normal controls with advancing age. Therefore, nonspecific testing for rheumatoid factor is of limited value in diagnosing or excluding the presence of rheumatoid arthritis or other autoimmune conditions.⁴ Nonselective ordering of rheumatoid factor and ANA tests in the absence of other, generally readily identifiable indicators of autoimmune disease, is of limited value and should be proscribed.⁵