Definition of SLE

Systemic lupus erythematosus is the prototypic systemic autoimmune disease characterized by heterogeneous, multisystem involvement and the production of an array of autoantibodies. Clinical features in individual patients can be quite variable, ranging from mild joint and skin involvement to severe, life-threatening internal organ disease. Lupus might be confined to the skin, without the presence of systemic involvement. There is no gold standard test for the diagnosis of SLE. Instead, the diagnosis rests on the judgment of an experienced clinician who recognizes the constellation of characteristic symptoms, signs, and laboratory findings in the appropriate clinical context, all other reasonable diagnoses having been excluded.

Development of the SLE Classification Criteria

The first classification criteria for SLE were developed by the American Rheumatism Association (ARA) in 1971.¹ The criteria were derived from a group of 245 patients with SLE contributed by 52 rheumatologists in the United States and Canada. The patients with SLE were compared with 234 patients who had rheumatoid arthritis (RA) and 217 patients without rheumatic disease. Out of 74 SLE manifestations reviewed, 14 were decided upon as the final criteria. Four or more of the 14 criteria had to be fulfilled in order for a person to be classified as having SLE. Importantly, the criteria could occur simultaneously or serially over any period. The final criteria heavily emphasized mucocutaneous features by including malar rash, discoid rash, photosensitivity, and oral ulcers as independent criteria. Notably, the criteria incorporated the presence of LE cells and a false-positive syphilis test result, but did not include tests for autoantibodies such as an antinuclear antibody (ANA) test and anti–double stranded DNA antibody (anti-dsDNA) because these tests were not widely in use at the time the criteria were developed. When tested in the population from which they were derived, the criteria were determined to have a sensitivity of 90% and a specificity of 99% against rheumatoid arthritis. After publication, the 1971 criteria became widely utilized. It has been estimated that approximately 90% of articles on SLE incorporated the criteria by 1978.²

In an effort to improve upon the 1971 criteria and incorporate new immunologic tests, the ARA commissioned a subcommittee in 1979 to update the 1971 criteria. The revised criteria for the classification of SLE were published in 1982.³ As part of the revision process, the subcommittee scrutinized each of the original criteria, and new potential variables were put forward, such as serologic tests and skin and kidney histopathology. In the end, 30 potential variables were assessed. Eighteen academic investigators contributed 177 patients with SLE along with 162 age-, race-, and sex-matched controls with various other connective tissue diseases. The majority of the control patients had RA, with scleroderma being the second most common diagnosis. Cluster analysis and other techniques were used to analyze relationships between variables. Potential sets of criteria were tested on random subsets of patients from the case and control groups. After completion of the process, the final criteria were composed of 11 elements. Consistent with the 1971 criteria, a patient had to fulfill 4 out of 11 criteria in order to be classified as having SLE. Five of the elements were composites of more than one variable: serositis, renal disorder, neurologic disorder, hematologic disorder, and immunologic disorder. Repeated analyses determined that the four original mucocutaneous variables (malar rash, discoid rash, photosensitivity, and oral ulcerations) should remain as independent variables. Raynaud’s phenomenon and alopecia were eliminated from the original criteria because of low sensitivity and specificity. The decision was made not to include skin and renal histopathology because it was the opinion of the investigators that those tests were infrequently performed. The arthritis criterion was revised to include the descriptor “nonerosive” and the necessity for more than two involved joints rather than more than one joint. The definition of proteinuria was altered from a threshold of >3.5 g/day in the 1971 criteria to >0.5 g/day in the revised criteria. The serologic tests for ANA, anti-DNA, and anti-Smith antibody (anti-Sm) were incorporated into the revised criteria. ANA was thought to be the strongest addition to the criteria because of its very high sensitivity, despite a relatively low specificity of 50%. The investigators decided not to include serum complement components because they did not improve accuracy. When tested in the population from which they were derived, the criteria had sensitivity and specificity values of 96%. When the criteria were tested against a separate population of patients with SLE, scleroderma, and dermato/polymyositis, the sensitivity was 83% and the specificity was 89%.

Over the years, several groups have studied the revised classification criteria in other patient populations. When studied in 156 patients with SLE from the University of Connecticut, the sensitivities of the original and revised criteria were 88% and 83%, respectively.⁴ When the “nonerosive” aspect of the arthritis criterion was removed because hand radiographs were not available for all patients, the sensitivity of the revised criteria increased to 91%. The investigators of this study determined that both sets of criteria were more likely to be met in patients with a longer duration of disease. The same group later determined that the specificities of the preliminary and revised criteria were 98% and 99%, respectively, when tested in a group of 207 patients with other rheumatic diseases.⁵ When the revised criteria were tested in SLE and control groups of Japanese patients, the sensitivity was 97% and specificity was 89%.⁶ A study of 135 patients with SLE from Tehran demonstrated a sensitivity of 90% for the revised criteria.⁷ Lastly, in a Zimbabwean study of 18 patients with the disease, the sensitivity of the revised criteria was 94%. When serologic elements were excluded, the sensitivity declined to 78%.⁸

In 1997, the criteria for the classification of SLE were revised for a second time in order to incorporate advancing knowledge about the association of antiphospholipid antibodies with SLE. Under the criterion “immunologic disorder,” the decision was made to exclude LE cells and insert antiphospholipid antibodies. Antiphospholipid antibodies were defined as the presence of immunoglobulin (Ig) G or IgM anticardiolipin antibodies, a positive lupus anticoagulant test result, or a false-positive serologic syphilis test result (Table 1-1).⁹ The changes reflected in the updated 1997 criteria were studied in an inception cohort of 154 patients with SLE in order to determine whether the replacement of LE cells with antiphospholipid antibodies would result in the selection of different groups of patients for inclusion in clinical studies.¹⁰ From the cohort of 154 patients, 36 patients were selected who met four criteria, one of which was the immunologic disease element. When the LE cell criterion was removed from the patients, 2 of 36 were no longer classified as having SLE. Both of these patients tested negative for anticardiolipin antibodies and lupus anticoagulant. To assess the impact of the addition of the anticardiolipin antibodies and lupus anticoagulant criteria, the investigators evaluated those patients who tested positive for anticardiolipin antibodies or lupus anticoagulant, but negative for LE cells. Only 1 patient was identified. Thus, the investigators determined that this alteration in the immunologic criterion would not result in a significant change in the patients classified as having SLE in their cohort. Going forward, it will be important to more fully study and validate the 1997 revised criteria in other cohorts of patients with SLE and related rheumatologic diseases.

TABLE 1-1 The 1997 Update of the 1982 Revised American College of Rheumatology Classification Criteria for SLE

Constraints of the Current SLE Classification Criteria

Despite the fact that the 1997 revised criteria are widely accepted and utilized, several limitations affect their use in clinical practice. SLE can involve virtually any organ system with heterogeneous manifestations; however, only a relatively few potential manifestations are represented in the criteria. In addition, some of the manifestations might be confused with common mimickers. For example, the criteria of malar rash and photosensitivity can be troublesome for clinicians because several common conditions closely mimic these findings. Acne rosacea and flushing can appear similar to a lupus malar rash, and polymorphous light eruption can simulate photosensitivity. The potential difficulty in interpreting malar rash and photosensitivity might lead to decreased specificity of those criteria. The serositis criterion includes pleuritis and pericarditis but not peritonitis. Arthritis is defined as “nonerosive,” implying that a radiograph has been taken. In routine clinical evaluations for SLE, however, hand radiographs are rarely performed. Proteinuria, defined as serum protein higher than 0.5 g/day, and urinary cellular casts are the only two renal criteria. Because many clinical laboratories do not routinely report cellular casts, the usefulness of this criterion is unclear. Notably, a positive renal biopsy result is not included in the criteria. It is possible for a patient with biopsy-proven lupus nephritis to not meet the necessary four criteria for classification as having SLE. Although there are a variety of ways in which SLE can affect the central and peripheral nervous system, psychosis and seizures are the only two manifestations included in the classification criteria. The hematologic criterion is categorized into the four subcomponents of hemolytic anemia, leukopenia, lymphopenia, and thrombocytopenia. Leukopenia and lymphopenia must be present on at least two occasions. The criteria do not specify how leukopenia and lymphopenia secondary to medications should be differentiated from those due to SLE.

Future Directions

Because of the limitations of the criteria as previously described, there has been a concerted effort by the Systemic Lupus International Collaborative Clinics (SLICC) group to further revise the ACR classification criteria.¹¹ During this process, patient scenarios from 716 patients with SLE and control patients were submitted by the SLICC members, and a consensus diagnosis was established for each scenario. The group identified those variables that were most predictive of SLE, and a classification rule was derived on the basis of multiple potential predictor variables. Although this effort is still a work in progress, 11 clinical and 6 immunologic elements have been selected for inclusion in the SLICC revision of the classification criteria. A patient is classified as having SLE if he or she (1) has biopsy-proven lupus nephritis with a positive ANA or anti–double-stranded DNA antibody test result or (2) fulfills four of the criteria including at least one clinical criterion and one immunologic criterion. Thus, one of the ways in which the classification of SLE by the SLICC criteria differs from that of the ACR criteria is by allowing for the stand-alone criterion biopsy-proven lupus nephritis. This alteration corrects a notable problem with the ACR criteria, in which a patient with biopsy-proven lupus nephritis might not meet enough criteria to be classified as having SLE. Also, counter to the ACR criteria, the SLICC criteria require at least one clinical element and one immunologic element for this classification. A patient cannot be classified as having SLE on the basis of purely clinical features. The SLICC criteria significantly expand upon the dermatologic elements by including various types of acute, subacute, and chronic cutaneous lupus lesions, as opposed to the ACR criteria inclusion of only malar rash and discoid rash. Photosensitivity has been removed, and nonscarring alopecia has been added. In the arthritis criterion, the term “inflammatory synovitis” has been substituted for “nonerosive arthritis.” In addition to the original seizures and psychosis elements, the new neurologic criterion incorporates several other neurologic manifestations of SLE, such as mononeuritis multiplex, myelitis, peripheral or cranial neuropathy, and acute confusional state. Within the group of immunologic criteria, low complement levels and positive direct Coombs test result in the absence of hemolytic anemia have been added. At the time of the writing of this chapter, the SLICC criteria were undergoing finalization and validation. It remains to be seen whether these criteria will eventually replace the ACR classification criteria.