Revision of Classification Criteria for Systemic Lupus Erythematosus
Derivation and validation of Systemic Lupus International Collaborating Clinics classification criteria for systemic lupus erythematosus
Petri M, Orbai AM, Alarcón GS, Gordon C, Merrill JT, Fortin PR, Bruce IN, Isenberg D, Wallace DJ, Nived O, Sturfelt G, Ramsey-Goldman R, Bae SC, Hanly JG, Sanchez-Guerrero J, Clarke A, Aranow C, Manzi S, Urowitz M, Gladman D, Kalunian K, Costner M, Werth VP, Zoma A, Bernatsky S, Ruiz-Irastorza G, Khamashta MA, Jacobsen S, Buyon JP, Maddison P, Dooley MA, Vollenhoven RF, Ginzler E, Stoll T, Peschken C, Jorizzo JL, Callen JP, Lim SS, Fessler BJ, Inanc M, Kamen DL, Rahman A, Steinsson K, Franks AG Jr, Sigler L, Hameed S, Fang H, Pham N, Brey R, Weisman MH, McGwin G Jr, and Magder LS. Arthritis & Rheumatism. 2012 May 2. doi: 10.1002/art.34473. [Epub ahead of print]
What is the topic?
In 1982, the American College of Rheumatology (ACR) revised the criteria for the classification of patients with systemic lupus erythematosus to help physicians classify (not diagnose) these patients. In 1997, these criteria were updated, but not validated. Subsequently, multiple groups of investigators have attempted to refine the classification criteria for lupus, but none have been widely accepted or implemented.
In order to improve clinical relevance and incorporate new knowledge to the field of lupus immunology, the Systemic Lupus Erythematosus Collaborating Clinics (SLICC), an international group dedicated to clinical research on lupus, revised the ACR systemic lupus classification criteria. To accomplish this, SLICC used defined clinical scenarios and stringent methodology. These criteria were then validated. This represents the largest validation study to date since the initial conception of the ACR classification criteria for lupus.
What did the researchers hope to learn?
SLICC undertook this study to revise the classification criteria for systemic lupus in order to address multiple concerns that have arisen since the 1982 criteria were developed. These include possible duplication of highly correlated cutaneous lupus terms (such as malar rash and photosensitivity) and the omission of other cutaneous lupus manifestations and many neurological manifestations of lupus.
Who was studied?
The SLICC classification criteria were derived from a set of 702 expert-rated patient scenarios. Subsequent validation of the classification criteria was accomplished in an independent sample comprising 690 lupus patients and controls.
How was the study conducted?
Each participating SLICC center was asked to submit data on 10-12 patients with a clinical diagnosis of systemic lupus erythematosus, as well as 12-15 control patients (for purposes of comparison). The control patients could have any of the following conditions: rheumatoid arthritis, myositis, chronic cutaneous lupus, undifferentiated connective tissue disease, vasculitis, primary anti-phospholipid antibody syndrome, scleroderma, fibromyalgia, Sjögren syndrome, rosacea, psoriasis, sarcoidosis, or juvenile idiopathic arthritis.
For derivation (establishment) of the classification rule, information about each of the 702 patients was summarized in a standardized short narrative, and these were sent to up to 32 rheumatologists from the SLICC group. The SLICC physicians then classified each patient as having systemic lupus or not, while blinded to the original center’s diagnosis. If 80% of the SLICC rheumatologists agreed on the classification, then that diagnosis was considered the “consensus” diagnosis. Scenarios which did not reach consensus in this way were later discussed by a panel of 5, and if 4/5 agreed on the classification, then those diagnoses were also considered a consensus.
Both statistical tests and consensus views were used to determine the final selection of variables to be part of the classification rule. A final SLICC meeting was held to discuss and vote on ratification of the final classification rule. To validate the new classification rule, the second set of scenarios from 690 patients and controls were used. In addition, blood serum from each patient was sent to the coordinating site and was analyzed for anti-double-stranded DNA (anti-ds-DNA) antibodies, anti-Smith antibodies, and complement C3 and C4 levels. A second set of blood samples was used to test for anti-phospholipid antibodies.
In order to arrive at a consensus diagnosis for each patient, 12 SLICC physicians rated all 690 patient scenarios. Those scenarios which did not reach 80% consensus in the initial rating process were edited for clarity and later re-rated by the larger group of 33 SLICC physicians. SLICC physicians were unaware of diagnoses given to patients at the respective SLICC sites. Over 80% consensus was reached on 615 cases while 75 cases remained without a consensus diagnosis of systemic lupus or not. After the second round of ratings, the 75 non-consensus scenarios were classified as either systemic lupus or not based on the majority opinion.
What did the researchers find?
Derivation of the SLICC criteria. Eighteen criteria associated with a diagnosis of systemic lupus were identified and initially considered. These were divided into two groups: “clinical” and “immunological.” Two clinical criteria were combined into one, thus leaving seventeen final criteria. Criteria need not be present simultaneously. The “clinical” criteria included the following: acute cutaneous lupus, chronic cutaneous lupus, oral ulcers, non-scarring alopecia (hair loss), synovitis (inflammation of the lining of the joints), serositis (inflammation of the lining around organs such as the heart and lungs), renal (kidney) involvement, neurological (brain-related) involvement, hemolytic anemia (abnormally low number of red blood cells), leukopenia (abnormally low number of white blood cells), and thrombocytopenia (abnormally low number of platelets in the blood). The “immunological” criteria included the following: elevated anti-nuclear antibodies (ANA), elevated anti-ds-DNA, anti-Smith antibodies, anti-phospholipid antibodies, low complement, and direct Coombs (anti-globulin) test in the absence of hemolytic anemia.
Using the proposed classification rule, a classification of systemic lupus erythematosus is made if: a) a patient satisfied four of the SLICC criteria (including at least one immunological criterion and at least one clinical criteria); or b) if a patient has biopsy-proven nephritis (inflammation of the kidney) compatible with systemic lupus in the presence of either ANA or anti-ds-DNA.
For the derivation of the criteria, data from each patient was reviewed by 26-32 clinicians. Among these clinicians, there was 86% agreement on the patient scenarios (with respect to whether the patients had lupus or not). These classifications were in agreement with the submitted diagnosis 91% of the time. Additional efforts resulted in an ultimate consensus of 98% of the 716 patients submitted for this part of the study. This consensus was in agreement with the submitted diagnosis 95% of the time.
Validation of the SLCC criteria. For the validation of the classification rule, 86% agreement was reached on the patient scenarios (with respect to whether the patients had lupus or not). Additional efforts resulted in an ultimate consensus of 93% of the 690 patients submitted for this part of the study.
The SLICC classification criteria for lupus represent an 8-year effort and can be stated as follows: a person is classified as having lupus if he/she has “lupus nephritis (in the presence of at least one immunological criteria) or four criteria (with at least one criterion being clinical and at least one criterion being immunological).” The clinical criteria included in the SLICC classification criteria improve on the revised ACR classification criteria in several important ways.
What were the limitations of the study?
The SLICC classification criteria have not been tested for purposes of diagnosing lupus.
What do the results mean for you?Overall, the SLICC classification criteria are intended to be more clinically relevant, reflect new knowledge about lupus immunology, include updated and more inclusive definitions of variables, and help overcome concerns with the current criteria. It should be emphasized that the SLICC classification criteria, however, have not yet been tested for purposes of lupus diagnosis. Furthermore, they need to be validated in other settings, and particularly, among patients suspected of having lupus but do not meet the ACR classification criteria.