Abstract
The myth persists that only the labor intensive Farr radioimmunoassay and Crithidia luciliae immunofluorescence (CL‐IFA) are systemic lupus erythematosus (SLE)‐specific tests. We compared them to ELISA with bacteriophage λ DNA (EL‐dsDNA) and denatured calf thymus DNA (EL‐ssDNA). By percentile ranking, the specificity cut‐off level was set both out of clinical context (SOCC) on 100 blood bank donors, and in clinical context (SICC) on 100 patients with either rheumatoid arthritis or scleroderma (50/50). Clinical sensitivity was calculated on 100 random SLE patients. At 95% SICC, the sensitivity of Farr, CL‐IFA, EL‐dsDNA, and EL‐ssDNA was similar (95%CI): 76% (66–84), 76% (66–84), 63% (53–72), and 75% (65–83), respectively; 87% of the patients were positive by at least one method and 55%by all methods. At 99% SICC, the sensitivity was also similar (95% CI): 57% (47–67), 47% (37–57), 58% (47–67), and 43% (33–53), respectively. The areas under ROC curve were similar (95% CI) when patients were used as controls for specificity. At 99% SOCC, EL‐ssDNA identified 89% positive, 2 negative but positive by another method at 95% SICC, and 9 negative (i.e. 89/2/9), followed by CL‐IFA (80/6/14), Farr (76/12/12), and EL‐dsDNA (64/23/13). Thus, at relatively low cost and easy automation, under the same conditions of specificity, the two ELISA tests combined were at least as good, if not superior, to CL‐IFA or Farr: they showed similar clinical sensitivity and also identified more patients with anti‐DNA antibodies. J. Clin. Lab. Anal. 24:77–84, 2010. © 2010 Wiley‐Liss, Inc.