Background/Purpose: Anti-dsDNA autoantibodies are highly specific for SLE. Anti-dsDNA positivity is an important eligibility criterion in many SLE studies, and decreases in anti-dsDNA levels are a measure of clinical improvement. Various anti-dsDNA assays have similar performance characteristics in distinguishing SLE patients from controls and those with other diseases. Previous studies have shown that the performance characteristics may change when testing active vs inactive SLE patients. The performance characteristics of a commercially available multilyte assay to the Farr radioimmunoassay (considered a gold standard of anti-dsDNA testing) were compared in adults with active moderate to severe SLE.

Methods: Subjects included adults with moderate-severe active SLE (92% female, median disease duration 6.3 years, mean SLEDAI 2K: 11.3) in an ongoing international, multi-center, double-blind, randomized, placebo-controlled trial. Anti-dsDNA levels were determined by both the Farr radioimmunoassay and the AtheNA Multi-Lyte® ANA-II Plus Test System (AMLII) in the 409/431 randomized subjects who were included in this analysis. Both assays were performed according to manufacturers’ instructions in Clinical Laboratory Improvement Amendments (CLIA) certified laboratories. The sensitivity and specificity of the recommended cutoff values for positive (>120) and indeterminate (>100) AMLII assay were assessed using the recommended values for positive (>10) and indeterminate (>5) and cutoffs of the Farr assay. Alternative optimal cutoff values of AMLII for different Farr values were determined by ROC analysis.

Results: Using the recommended cutoff values for positivity, 24% of the patients were anti-dsDNA positive by the AMLII compared to 66% by the Farr assay. When compared to the Farr assay, the recommended cutoff values for AMLII had high specificity but low sensitivity, leading to an unacceptably low concordance with the gold standard of anti-dsDNA testing in this population. Using the optimal cutoff values determined by ROC analysis identified ≥28 (the lower limit of quantifiable detection) as the optimal cutoff for AMLII, which significantly improved the performance of the AMLII (Table).

	Positive		Indeterminate or positive
Farr cutoff	≥10		≥5
AMLII cutoff	Recommended	ROC determined	Recommended	ROC determined
	≥120	≥28	≥100	≥28
ROC AUC	0.84		0.79
Sensitivity	36%	78%	33%	68%
Specificity	99%	82%	97%	87%
Concordance	57%	79%	45%	72%
ROC: receiver operator curve; AUC: area under the curve. An AUC ≥0.8 represents good diagnostic accuracy

Conclusion: The optimal cutoff value in a population with a high pretest likelihood of anti-dsDNA positivity should be confirmed for any non-gold standard assay if it is being considered to replace a gold standard, such as the Farr assay. Misclassification of a large proportion of subjects with active moderate to severe SLE who are anti-dsDNA positive by the Farr assay as anti-dsDNA negative may lead to unnecessary exclusion of otherwise eligible subjects from clinical trials and may reduce the likelihood of demonstrating a change in the SLEDAI component of composite endpoints in SLE trials.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/decreased-sensitivity-of-a-commercial-anti-dsdna-assay-in-patients-with-moderately-to-severely-active-systemic-lupus-erythematosus/