Comparative assessment of sensitivity and specificity of three variants of classification criteria for systemic lupus erythematosus in a cohort of Russian patients

Background. The clinical and serologic heterogeneity of systemic lupus erythematosus (SLE) presents challenges for diagnosis, particularly in the earliest stages of the disease when there are insufficient signs to make a reliable diagnosis.

Aim. To make a comparative assessment of sensitivity and specificity of various classification criteria of SLE on a cohort of patients of Nasonova Research Institute of Rheumatology.

Materials and methods. A total of 252 patients were included in the study; 152 (60%) of 252 patients had reliable SLE (mean age 36 [29.5–46] years, duration of disease 9 [3.4–19] years). Of 252 patients, 26 (11%) had PAPS (mean age 36.5 [31–42] years, duration of disease 4.6 [1–10.4] years). Systemic sclerosis was diagnosed in 74/252 (29%) patients, (mean age 51.5 [42–59] years, duration of disease 9 [5–16] years). The quality of the classification function of the criteria was assessed by ROC analysis.

Results. SLE was diagnosed in 131 (86%) of 152 patients using the American College of Rheumatology – ACR)-1997 criteria, in 145 (95%) using the The Systemic Lupus International Collaborating Clinics (SLICC) 2012 criteria, and in 144 (94.7%) using the European League Against Rheumatism (EULAR)/ACR 2019 criteria. ANF positivity was the least statistically significant of all signs in relation to the diagnosis of SLE. The area under the curve (AUC) for ANF≥1/160 titers was AUC 0.654 for the ACR-97 criteria, AUC 0.616 for the SLICC-12 SLE criteria, and AUC 0.609 for the 2019 EULAR/ACR criteria. ROC analysis of the relationship between the number of criteria/points and a reliable diagnosis of SLE revealed a high diagnostic accuracy – the AUC for all SLE criteria was greater than 0.940. In the ROC analysis of patients with SLE and PAFS, indicating the number of diagnostic criteria, sensitivity was 86% for ACR-1997, 95% for SLICC-2012, 95% for EULAR/ACR 2019, and specificity was 100, 62 and 62%, respectively.

Conclusion. The classification criteria SLICC-2012 and EULAR/ACR 2019 are more sensitive for the diagnosis of SLE in the Russian population, and the criteria ACR-1997 are more specific. All three variants of the SLE classification criteria have sufficient sensitivity and specificity for their use in real clinical practice.