SLE classification criteria perform well in validation study



– The first European League Against Rheumatism and American College of Rheumatology joint criteria for classifying systemic lupus erythematosus have a sensitivity and a specificity of more than 90%.

This is important because they improve upon the existing ACR and Systemic Lupus International Collaborating Clinics (SLICC) criteria, said Martin Aringer, MD, PhD, who cochaired the Steering Committee that produced the new classification criteria.

The video associated with this article is no longer available on this site. Please view all of our videos on the MDedge YouTube channel.

Most clinicians working with lupus are familiar with the 1997 ACR criteria for the classification of systemic lupus erythematosus (SLE), which “had a relatively simple structure,” Dr. Aringer said during the opening plenary abstract session at the European Congress of Rheumatology. These considered items such as the presence of malar or discoid rash, photosensitivity, oral ulcers and arthritis, among others. These had a high specificity but a lower sensitivity. The development of the SLICC criteria in 2012 improved upon the sensitivity of the ACR criteria (92%-99% vs. 77%-91%), but at a loss in specificity (74%–88% vs. 91%-96%).

The SLICC criteria introduced two novel ideas, said Dr. Aringer, professor of medicine and chief of the division of rheumatology at the Technical University of Dresden (Germany). The first was that there had to be at least one immunologic criterion met, and the second was that biopsy-proven lupus nephritis had to be present with antinuclear antibodies (ANA) and anti-DNA antibodies detected.

One of the goals in developing the joint EULAR/ACR criteria therefore was to try to maintain the respective sensitivity and specificity achieved with the SLICC and ACR criteria. One of the key things that the new criteria looked at was to see if ANA could be used as an entry criterion. Investigations involving more than 13,000 patients with SLE showed that it could, with a antibody titer threshold of 1:80, exhibit a sensitivity of 98% (Arthritis Care Res. 2018;70[3]:428-38). Another goal was to see if histology-proven nephritis was a stronger predictor of SLE than clinical factors, such as oral ulcers, and to identify items that would only be included if there was no other more likely explanation (Lupus. 2016;25[8]:805-11).

Draft SLE classification criteria were developed based on an expert Delphi process and included ANA as an entry criterion and weighted items according to the likelihood of being associated with lupus. Items considered included the presence and severity of lupus nephritis, serology and other antibody tests, skin and central nervous system involvement, and hematologic and immunologic criteria such as the presence of thrombocytopenia and low complement (C3 and/or C4).

Dr. Martin Aringer, Technical University of Dresden, Germany Sara Freeman/MDedge News

Dr. Martin Aringer

Dr. Aringer described how these criteria had been derived and now validated in a large international cohort of individuals with and without SLE. In total, 23 expert centers participated in this process, each contributing up to 100 patients each with SLE or non-SLE diagnoses. Three independent reviewers confirmed each patient’s diagnosis, with 1,160 patients with SLE and 1,058 without SLE finally identified. Of these, 501 and 500 were randomly allocated to a derivation cohort and 696 and 574 to a validation cohort.

The final, simplified draft SLE classification criteria include 22 items in addition to the presence of ANA. A cut-off score of 10 or more is required for a classification of SLE. For example, a patient with an ANA of 1:80 or higher plus class III/IV nephritis (scoring 10) would be classified as having SLE. A patient with class II/V nephritis (scoring 8) would need another factor to be classified as having lupus, such as the presence of arthritis (scoring 6).

“Performance characteristics find sensitivity similar to the SLICC criteria while maintaining the specificity of the ACR 1997 criteria,” Dr. Aringer said, adding that these criteria will now be formally submitted to and reviewed by EULAR and ACR.

The sensitivity and specificity of the new criteria were 98% and 96% in the derivation cohort and 96% and 93% in the validation cohort.

“I was really very pleased and very happy to see that the revised or the new ACR/EULAR classification criteria had sensitivity and specificity of above 90%,” Thomas Dörner, MD, PhD, said in an interview at the congress. Dr. Dörner was a codeveloper of these criteria.

Over the past 10-15 years there have been several therapies that have failed to live up to their early promise as a potential treatment for lupus, said Dr. Dörner, professor of medicine at Charité–Universitätsmedizin Berlin. He noted that the failed treatment trials had led investigators to try to determine ways in which lupus might be best treated, such as by a “treat-to-target” approach to attain remission and low-disease activity. It also led to the reevaluation of how lupus is classified to see if that might be affecting the population of patients recruited into clinical trials.

“We had the feeling, and this is now confirmed by the new classification criteria, that a number of patients studied in earlier trials may have not fulfilled what we think is the classical lupus profile, so-called lupus or SLE mimickers,” Dr. Dörner said. This could have affected the chances of a treatment approach being successful versus placebo.

The new classification criteria are similar to those in other rheumatic diseases in that they give different weight to the effects on different organ systems, Dr. Dörner said. The stipulation that there must be a positive ANA test is also an important step, “really to make sure that we are looking at an autoimmune disease and nothing else,” he observed.

For patients who do not have a positive ANA test, they can of course still be treated, Dr. Dörner reassured, but for the classification criteria and entering patients into clinical trials, it’s really important to have strict classification criteria so that the results may be compared.

Dr. Aringer and Dr. Dörner had no relevant disclosures besides their involvement in developing the new classification criteria.

Next Article: