• systemic lupus erythematosus
• autoimmune diseases
• epidemiology

The release of the 2019 EULAR/ACR classification criteria of SLE would be considered a landmark work in developing international consensus on SLE classification criteria.1 However, there are some concerns regarding the new criteria. First of all, the criteria did not include clinical manifestations sometimes could occur as initial symptoms in SLE patients. If a patient comes to a rheumatologist with the only presentation, such as repeated vascular thrombosis, mesenteric vasculitis or unexplained liver injury and positive results for both antinuclear antibodies and antibodies to the double-stranded DNA, would they be classified as SLE? With the increasing understanding of the aetiology and pathogenesis of SLE and the continuous development of the testing technique, the updated classification criteria of SLE included complement proteins and more antibodies.1–4 However, the clinical domains of the criteria were not substantially updated. Besides, the non-SLE populations used for the derivation and validation cohorts were mainly other rheumatic disorders and infection except malignancies. Malignancies, especially haematological tumours, could sometimes mimic SLE-like symptoms, and misdiagnosis would delay the treatments, which were contributing to a growing quandary for clinical decision-making. It would be more practicable if patients with malignancies such as lymphoma were included in the non-SLE groups during the derivation and validation process. What is more, it would be interesting if the authors could show the performance of the criteria in different age groups, sex and ethnicity as the authors pointed out that the derivation and validation cohorts were from limited populations and needed to be further validated among different groups.