Relevance of Different Results of Different Anti–Double-Stranded DNA Assays in Reporting Clinical Studies: Comment on the Article by Petri et al
Version of Record online: 27 JAN 2014
Copyright © 2014 by the American College of Rheumatology
Arthritis & Rheumatology
Volume 66, Issue 2, pages 479–480, February 2014
How to Cite
Bonroy, C., Verfaillie, C., De Witte, E. and De Keyser, F. (2014), Relevance of Different Results of Different Anti–Double-Stranded DNA Assays in Reporting Clinical Studies: Comment on the Article by Petri et al. Arthritis & Rheumatology, 66: 479–480. doi: 10.1002/art.38252
- Issue online: 27 JAN 2014
- Version of Record online: 27 JAN 2014
- Accepted manuscript online: 31 OCT 2013 01:20PM EST
We read with interest the recent report by Petri and colleagues on baseline predictors of disease flares in patients with serologically active systemic lupus erythematosus (SLE) treated with standard therapy alone—a study that used the combined placebo group of the phase III belimumab trials (). The definition of serologic activity included the presence of anti–double-stranded DNA (anti-dsDNA) antibodies in concentrations of at least 30 IU/ml. This inclusion criterion was based on a prior phase II placebo-controlled trial showing that belimumab reduced and stabilized disease activity within this subpopulation and was subsequently used in two phase III trials ([2-4]). In our opinion, however, it should be clearly anticipated that the serologic criterion for study inclusion used in the phase II and phase III studies (anti-dsDNA antibody levels ≤30 IU/ml) cannot be extrapolated directly for use in clinical practice to select patients eligible for belimumab therapy. From the same perspective, use of an anti-dsDNA level of >200 IU/ml to predict SLE flares, as reported by Petri et al (), should be adopted with caution.
Our arguments are as follows. First, assays assessing the presence and the level of dsDNA antibodies vary significantly (between techniques as well as between manufacturers) despite the use of international units (IU/ml) based on previous attempts to obtain standardization (). Second, despite the emerging use of enzyme immunoassays (EIAs), it is generally accepted that the Farr assay remains the gold standard for determining SLE-specific DNA antibodies (). Within this context, details are missing regarding the EIA and manufacturer used in the phase II and phase III belimumab studies, as well as their relationship with the results obtained using a Farr immunoassay (as was used in the phase I belimumab study) ([1-4, 7]).
To document this latter relationship as well as the variability between assays, we analyzed anti-dsDNA in a set of 44 patients with SLE, using a Farr assay (Anti-dsDNA Kit; Trinity Biotech) and a selection of commercially available EIAs (Figure 1). All assays included in the study were calibrated against the Wo/80 reference material (). Using the cutoff values for positivity proposed by the manufacturers, moderate to substantial kappa value agreement between the EIAs was observed (κ = 0.462–0.732). For the comparison between EIAs and the Farr assay, only moderate agreement was observed (κ = 0.573–0.657) (results not shown). Nevertheless, it should be stressed that the high variability in test results cannot be reduced solely to an issue of differences in the choice of the cutoff values applied (r = 0.527–0.956) (Figure 1).
For clinical practice, this implies that direct use of criteria for anti-dsDNA antibodies deduced from clinical trials would result in the identification of different patient subsets depending on the assay used (e.g., use of the criterion of an anti-dsDNA antibody level of >30 IU/ml for guidance of belimumab therapy would select 32–100% of the patients in our cohort). We would like to call upon investigators to systematically specify details on the assay and manufacturer when reporting anti-dsDNA results. This has important relevance for interpretation and reproduction of the results reported.