Jump to comment:

• Re:Are autoantibodies to RNA-polymerase III to be incorporated in routine diagnostic laboratory algorithms for systemic autoimmune rheumatic diseases?
  Frank van den Hoogen
  Published on: 17 January 2014
• Are autoantibodies to RNA-polymerase III to be incorporated in routine diagnostic laboratory algorithms for systemic autoimmune rheumatic diseases?
  Jan Damoiseaux
  Published on: 7 January 2014

• Published on: 17 January 2014

  Re:Are autoantibodies to RNA-polymerase III to be incorporated in routine diagnostic laboratory algorithms for systemic autoimmune rheumatic diseases?

  • Frank van den Hoogen, Dr
  • Other Contributors:
    • Jaap Fransen, Janet Pope, Murray Baron et al

  Dear Editor

  In his letter, Jan Damoiseaux plees for testing of anti-RNA polymerase III antibodies (ARA), and also other (novel) SSc-associated antibodies, only for those patients who are really suspected for SSc. Damoiseaux argues that there is a necessity to provide clinical information together with the request for testing for rheumatic systemic autoimmune diseases.

  We very much agree that testing for S...

  Show More

  Dear Editor

  In his letter, Jan Damoiseaux plees for testing of anti-RNA polymerase III antibodies (ARA), and also other (novel) SSc-associated antibodies, only for those patients who are really suspected for SSc. Damoiseaux argues that there is a necessity to provide clinical information together with the request for testing for rheumatic systemic autoimmune diseases.

  We very much agree that testing for SSc-associated antibodies should only be requested for those people who really are suspected to have SSc. We feel that a correct request of testing for these auto-antibodies is first and foremost a responsibility of the requestor, i.e. the rheumatologist. However, immunologists and rheumatologists almost certainly can learn from each other in optimizing requests for antibody testing, which admittedly also may mean fewer requests but therefore for the right people at the right moment.

  Curiously enough, in his letter Damoiseaux says that 'it is to be questioned whether inclusion of ARA was valid' and further he brings forward that '25x more false-positive ARA results are to be expected than true positives'. We do not agree with both of his points.

  Indeed in the derivation cohort of 100 SSc patients and 100 controls with diseases similar to SSc, the occurrence of ARA positivity was extremely low, and consequently there was no difference between SSc patients and controls. While being eyebrow-raising (at least to us), it would have been an error to discard ARA for this reason solely, as the validation set showed later. ARA was included because many other studies showed its relevance in recognizing SSc (external knowledge) and because of the systematically assessed opinion of expert clinicians in SSc. Finding outlying results on a single item in a sample of this size most likely is the result of sampling error; which is uncomfortable, but no reason to question the inclusion of ARA as not valid. In the validation sample, 27/268 (10%) SSc patients were ARA positive, towards zero ARA positive patients in the 137 consecutively collected patients with a SSc- like disorder.

  Second, in his reasoning, Damoiseaux states that in routine laboratory settings it is realistic to expect that only 2 out of each 1000 patients being suspected for systemic autoimmune rheumatic diseases, and thus are tested for autoantibodies according to local algorithms, eventually are diagnosed with SSc. Damoiseaux shows in his calculations persuasively that ARA is not suitable as a screening test. However, while a SSc/non-SSc ratio of 2/1000 probably may be realistic it can be debated whether that is usual in practice. Many referral centres for SSc may have patient domains with higher a priori probabilities for SSc than 0.002.

  If we argue for example that ARA is ordered only when suspecting SSc and that in those situations it may be that non-SSc diseases are 5 times more prevalent and ARA specificity is only99%, then for every 100 SSc cases there would be 500 non-SSc cases with 5 positives while in the SSc cases there would be 11 positives (if we accept 11% sensitivity). Then there would be no where near 25 times more false positives.

  We feel it is the responsibility of the rheumatologist to select those patients with a sufficiently high probability to have SSc for further testing. Notably, the SSc classification criteria for SSc consist of a number of items, including SSc-associated antibodies, that together inform of the probability of SSc being present. In diseases of the syndrome-type like SSc, it is difficult to interpret tests in isolation.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top
• Published on: 7 January 2014

  Are autoantibodies to RNA-polymerase III to be incorporated in routine diagnostic laboratory algorithms for systemic autoimmune rheumatic diseases?

  • Jan Damoiseaux, Medical Immunologist

  Dear Editor,

  Recently a collaborative initiative of the American College of Rheumatology (ACR) and the European League against Rheumatism (EULAR) resulted in novel classification criteria for systemic sclerosis (SSc) [1]. These criteria were first selected in a derivation cohort and, next, confirmed in a validation cohort. From a laboratory perspective it is interesting that the new criteria include, besides th...

  Show More

  Dear Editor,

  Recently a collaborative initiative of the American College of Rheumatology (ACR) and the European League against Rheumatism (EULAR) resulted in novel classification criteria for systemic sclerosis (SSc) [1]. These criteria were first selected in a derivation cohort and, next, confirmed in a validation cohort. From a laboratory perspective it is interesting that the new criteria include, besides the classical anti- centromere and anti-topoisomerase I antibodies, also the anti-RNA polymerase III antibodies (ARA). These ARA have been associated with diffuse cutaneous involvement and increased risk for renal crisis. Although candidate parameters to be included in the novel classification criteria were selected in a multistep process, the respective parameter is expected to significantly discriminate between well-defined SSc patient and a cohort of relevant disease controls. Although ARA eventually appeared to significantly discriminate patients from controls in the validation cohort, this happens not to be the case for the derivation cohort. Therefore, it is to be questioned whether inclusion of ARA was valid.

  More importantly, as a consequence of inclusion of ARA in the classification criteria, ARA will be increasingly incorporated in the routine diagnostic laboratory algorithm for systemic autoimmune rheumatic diseases (SARD). Obviously, classification criteria are not designed for diagnostic purposes, but in clinical practice they are not independent. Integration of the detection of novel SSc-associated antibodies, like ARA, in routine laboratory settings requires novel, multistep strategies to avoid high numbers of false-positive results. Bonroy et al recently calculated that in routine laboratory settings a patient/control ratio of 0.002 is realistic [2], i.e., only 2 out of each 1000 patients being suspected for SARD, and thus are tested for autoantibodies according to local algorithms, eventually are diagnosed with SSc. The ARA prevalence in SSc is about 11% (95% CI: 8-14; being in line with data obtained in the validation cohort) [3], while the specificity of the distinct test-systems varies between 99.0% and 99.5% [4,5]. Taking these data together, at least 25x more false-positive ARA results are to be expected than true positives. These data reinforce the necessity to provide clinical information together with the request for testing for SARD-related autoantibodies [6]. In that case, testing for ARA, and also other novel SSc-associated antibodies, can be restricted to those patients that really are suspected for SSc and may benefit from the result.

  References

  1. Van den Hoogen F, Khanna D, Frasen J, et al. 2013 classification criteria for systemic sclerosis. an American College of Rheumatology/European League against Rheumatism collaborative initiative. Ann Rheum Dis 2013; 65: 2737-2747.

  2. Bonroy C, Smith V, Van Steendam K, et al. The integration of the detection of systemic sclerosis-associated antibodies in a routine laboratory setting: comparison of different strategies. Clin Chem Lab Med 2013; 51: 2151-2160.

  3. Sobanski V, Dauchet L, Lef?vre G, et al. Prevalence of anti-RNA polymerase III antibodies in systemic sclerosis: new data from a French cohort, systematic review and meta-analysis. Arthritis Rheum DOI 10.1002/art.38219.

  4. Villalta D, Imbastaro T, Di Giovanni S, et al. Diagnostic accuracy and predictive value of extended autoantibody profile in systemic sclerosis. Autoimmun Rev 2013; 12: 14-120.

  5. Maes L, Blockmans D, Verschueren P, et al. Anti-PM/Scl-100 and anti-RNA-polymerase III antibodies in scleroderma. Clin Chim Acta 2010; 411: 965-971.

  6. Agmon-Levin N, Damoiseaux J, Shoenfeld Y. Response to: 'Detection of anti-nuclear antibodies, added-value of solid phase assay?' by Bossuyt and Fieuws. Ann Rheum Dis DOI 10.1136/annrheumdis-2013-204797.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top