Conference Coverage

New classification system for systemic lupus erythematosus moves forward

View on the News

The aim is to get patients into trials at earlier stages

This joint ACR/EULAR effort is very large, involving over 40 international SLE experts and centers, including many Systemic Lupus International Collaborating Clinics Criteria members. The goals are to develop new criteria that will be both sensitive and specific for SLE, which is a very heterogeneous and often elusive disease, using newer rigorous expert opinion–based and data-driven methodologies (as has been accomplished recently for rheumatoid arthritis, scleroderma, and gout).

Dr. Karen Costenbader

Dr. Karen Costenbader

The new criteria will include a point system on a continuous scale with a cut-off for “definite SLE” decided upon by SLE expert consensus and many cases used for validation. There is a particular interest in trying to include SLE patients at earlier stages of disease, as there is impetus in SLE clinical trials and studies to test strategies and medications capable of preventing the longer-term sequelae and complications of the disease.

This is a tall order! While genetic and cytokine-based biomarkers are being developed for the identification of SLE, unfortunately, we were not able incorporate them as they are not available for routine use at this time. The SLE classification criteria are used worldwide for inclusion in clinical trials and studies.

The work has been preceding in phases: phase 1, item generation; phase 2, item reduction and definition of criteria; phase 3, multicriteria decision analysis and threshold identification; and phase 4, validation. Phases 1-3 involved many iterative, group discussions, data collection and review, and novel “forced choice” methodologies for arriving at group consensus. In phase 4 (validation), the goal is to compare classification using these criteria with the existing ACR and SLICC criteria, as well as SLE expert physician diagnosis. We will see how they do after all this effort!

Karen H. Costenbader, MD, is the lupus program director at Brigham and Women’s Hospital, Boston. She is a member of the classification criteria steering committee and is the senior author on a paper describing the process to develop the draft classification criteria (Arthritis Care Res. 2017 July 10. doi: 10.1002/acr.23317).



MADRID – A proposed classification scheme for systemic lupus erythematosus (SLE) relies on a combination of antinuclear antibody titer and the weighted scoring of signs and symptoms in seven clinical and three immunologic domains.

Anyone with an ANA titer of at least 1:80 on immunofluorescence and 84 points accumulated from the domains can be classified as having the disease, according to the proposed system. An international collaboration between the American College of Rheumatology and the European League Against Rheumatism, the system is the first classification scheme update since 2012, said Sindhu Johnson, MD, PhD, who discussed it during the European Congress of Rheumatology.

Dr. Sindhu Johnson

Dr. Sindhu Johnson

“This is a work in progress at this point, not the final system,” said Dr. Johnson of the University of Toronto, cochair of the project’s steering committee. “While the prior iterations of lupus classification criteria have served us well, both groups felt it was time for an update that would reflect our current thinking on the disease.”

The proposed system is not intended to be a diagnostic tool, Dr. Johnson said in an interview. Rather, it’s meant to better stratify patients into research studies. “The prior criteria were missing patients. And, since classification criteria are used to decide whether patients can get into a clinical trial, we all felt that we were doing patients an injustice if the criteria were excluding some and denying them an opportunity to receive a novel therapy.”

There are currently two classification criteria in use: the 1982 American College of Rheumatology criteria and the Systemic Lupus International Collaborating Clinics Criteria (SLICC 2012). New understandings of SLE pathogenesis have rendered the 1982 ACR criteria outdated, according to a recently published paper (Arthritis Care Res. 2017 July 10. doi: 10.1002/acr.23317). While the SLICC 2012 criteria incorporated some of the new concepts and have increased sensitivity, compared with the 1982 ACR criteria, their specificity actually declined, partially because the document assigned equal weight to each of the clinical and immunologic criteria. The ACR/EULAR project takes a different tack. Criteria are weighted to reflect the clinical impact of different signs and symptoms, Dr. Johnson said.

“In clinical practice, if someone has class III/IV lupus nephritis, that’s a very different patient than someone who has leukopenia. As clinicians, we weight things differently, and so do these criteria, putting more weight on serious or internal organ manifestations of SLE.”

The ACR/EULAR criteria begin with a confirmed ANA titer of at least 1:80. “This has never been a requirement before, but the consensus now is that you need to have a positive ANA to be classified.”

Once that baseline is established, patients can be assessed in seven clinical domains and three immunologic domains. Each contains a subgroup of weighted signs and symptoms. These are ordered from those with least impact to those with most impact. Within each domain, only the highest-scoring criterion is counted toward the total score. When assessing, clinicians should not score any symptom if a cause other than SLE is more likely. The symptoms are not time-bound either, Dr. Johnson said. A symptom may have occurred only one time in the past, and that’s sufficient to earn a score. At least one clinical criterion must be present to be classified as SLE-positive.

The clinical domains

Constitutional: Fever (13 points). The only symptom in this domain, it’s never before been assessed in SLE criteria, Dr. Johnson said. “Our inclusion of fever is new, but our work in phase 1 of this project found that fever is a feature that can distinguish early lupus from mimickers. We want to identify patients as early in the disease course as possible so we can intervene, and fever appears to improve the ability to detect those patients.”

Cutaneous: Nonscarring alopecia (13), oral ulcers (14), subacute cutaneous or discoid lupus (29), and acute cutaneous lupus (38).

“Skin has long been recognized as an important part of lupus, but it got a lot of weight that some people felt was inappropriate. These criteria still include skin, but a patient can’t be classified on skin findings alone. There is concern that skin findings by themselves may not be lupus but something else, and some people even consider that cutaneous and systemic lupus are two different things.”

Arthritis: Synovitis in at least two joints (34).

Neurologic: Delirium (12), psychosis (20), and seizure (34).

Serositis: Pleural or pericardial effusion (34) and acute pericarditis (38).

Hematologic: Leukopenia (12), thrombocytopenia (26), and autoimmune hemolysis (28).

Renal: Proteinuria more than 0.5 g/24 hours (27), renal biopsy with class II or V lupus nephritis (55), and renal biopsy with class III or IV lupus nephritis (74).

The immunologic domains

Antiphospholipid antibodies: Anticardiolipin immunoglobulin G more than 40 GPL units, anti-beta2GP1 IgG more than 40 units, or lupus anticoagulant positive (13).

Complement proteins: Low C3 or low C4 (19) and low C3 and low C4 (27).

Highly specific antibodies: Anti-dsDNA antibody (38) and anti-Smith antibody (40).

Moving forward

Screening 10 domains with their attendant components may seem a bit clunky now, Dr. Johnson noted, but the final iteration should be more streamlined. Plus, she said, the system will be presented on a computer application that makes calculation much easier. “We’re aiming for feasibility and simplicity, but, at the same time, when you have a complex disease, you don’t want oversimplification. You may lose sensitivity and specificity.”

After further streamlining, Dr. Johnson said, the next step will be validating in a large retrospective patient cohort. “Right now, we are still collecting data for the validation cohort, which will be drawn from 36 centers. We’ll analyze sensitivity and specificity, comparing this system with the other two. We hope to present all these data at the ACR meeting in the fall.”

While research classification is the system’s raison d’être, it will undoubtedly influence diagnosis and clinical assessment as well, Dr. Johnson said. “ACR and EULAR are very clear that they only support the validation of classification criteria. The diagnosis of SLE is still within the hands of the physician. But, we know that classification criteria do inform our concept of the disease, so it’s likely these will shift the way we think about diagnosing lupus as well. We do hope to identify patients with earlier disease, so they have the opportunity to be involved in research” that may modify their disease course and, ultimately, prevent permanent damage and improve quality of life.

Dr. Johnson had no disclosures related to the development of the classification criteria.

On Twitter @alz_gal

Next Article: