Objective: To evaluate the sensitivity, specificity, and likelihood ratios of anti-dsDNA, C3, C4, and Clq binding tests for predicting flares in systemic lupus erythematosus (SLE) and to evaluate the conflicting results reported in the recent literature.
Methods: Test results from 9 to 3 months preceding 83 flares among 53 patients from a cohort study of 202 patients with SLE were assessed for their ability to predict renal and nonrenal flares and all flares combined. A flare was defined as a 6 point rise in a modified SLE Disease Activity Index.
Results: For all 4 tests, sensitivity approximated 50% and specificity was less than 75%. For anti-dsDNA, Clq binding, and C4, both positive and negative likelihood ratios were close to 1.0, suggesting little clinical value for routine testing. For C3, likelihood ratios for a positive test approximated 2.0, meeting the minimal level for clinical significance. Review of previous studies suggested that sample size, the use of a standardized disease activity measure to define a flare, the severity of the flare, and the study duration could not explain the discrepant results among studies. The 2 studies with testing every 4 to 6 weeks found that a rising anti-dsDNA predicted flares, while those with less frequent testing did not.
Conclusion: While our study provided no major support for routine testing with anti-dsDNA, C3, C4, or Clq binding, additional longitudinal studies are needed to assess monthly testing of anti-dsDNA as a predictor of flare in SLE.