Conference Coverage

Proposed SLE classification criteria prove highly sensitive, specific


 

AT ACR 2017

The first validation study of proposed new classification criteria for systemic lupus erythematosus (SLE) yielded a sensitivity of 98% and a specificity of 97%.

The development is part of a four-phase joint effort by the American College of Rheumatology and the European League Against Rheumatism, first launched in 2014, to improve classification of patients with SLE for the purposes of clinical trials of new therapies and clinical research into the causes and outcomes of the disease. “The ACR and EULAR have long recognized the importance of classification criteria, so that we have more homogeneous groups in these research studies, and then we can compare results from studies,” Sindhu Johnson, MD, PhD, said during a press briefing at the annual meeting of the American College of Rheumatology.

Dr. Sindhu Johnson of the University of Toronto Doug Brunk/Frontline Medical News

Dr. Sindhu Johnson

“In the last 10-15 years, the standards for developing classification criteria have changed. We have come to appreciate potential biases that may reduce the validity or reliability of classification criteria. As a consequence, both organizations have put out position papers on the standards that we must consider when we develop classification criteria,” she said.

A synopsis of the first three phases from the ongoing effort was presented at the June 2017 EULAR meeting. It included results from a systematic review of the literature and a meta-regression analysis to determine whether antinuclear autoantibodies should be required in the classification of SLE. The analysis showed that an antinuclear antibody titer of greater than or equal to 1:80 by immunofluorescence on human epithelial type 2 cells had a sensitivity of 98.4% for correctly capturing SLE. This prompted the EULAR/ACR steering committee to propose this titer as an “entry criterion” for SLE classification.

At the ACR meeting, Dr. Johnson, a rheumatologist at the University of Toronto who also cochairs the ACR Classification and Response Criteria subcommittee, discussed the fourth phase of the collaboration, which involves fine-tuning and validating the proposed SLE classification criteria.

To date, 189 investigators and more than 3,500 patients have contributed to the effort. “In the most recent phase, 36 international lupus centers were approached to contribute 100 cases and 100 controls,” she explained. “We then had each case independently adjudicated by three lupus experts at three different lupus centers to make sure there was consensus on the diagnosis. Next, we randomly selected 500 cases and 500 controls, resulting in a derivation cohort of 1,000 subjects to test our draft criteria system.”

Dr. Johnson and her associates found that the proposed SLE criteria had a sensitivity of 98% and a specificity of 97%, which exceeds that of the old ACR criteria. “We have defined a system of criteria which produces a measure of the relative probability that a particular case with a combination of clinical symptoms or features has SLE,” she said.

Dr. Johnson emphasized that the proposed criteria are intended for the classification of SLE, not for diagnosing the condition. “The diagnosis of lupus still remains in the hands of the physician, who will take into account all of the symptoms, signs, and other investigations,” she said. “We have identified the highest yield of those, but there will be some people who do not fulfill classification criteria yet do have a diagnosis of SLE.”

The next few weeks is a period to solicit feedback from stakeholders, after which members of the EULAR/ACR steering committee will be weighing feedback on the proposed criteria. “After that, we will see if there are any final revisions that need to be made,” Dr. Johnson said. “If we’re happy with the product, then the final validation will occur. We still have more than 1,000 patients from the validation core that has been reserved for that final validation.”

Data from the final validation are expected to be presented at the June 2018 EULAR meeting and ultimately published in Arthritis and Rheumatism and the Annals of Rheumatic Diseases. “Before we can get there, though, it needs to be formally reviewed by the ACR and EULAR for their formal endorsement. We expect that will take another 6 months.” She reported having no disclosures.

Next Article:

Blindness linked to HCQ use rare in rheumatic patients
   Comments ()

Recommended for You

News & Commentary

Quizzes from MD-IQ

Research Summaries from ClinicalEdge

Related Articles