Determination of four homogeneous subgroups of patients with antiphospholipid syndrome: a cluster analysis based on 509 cases.

Objective: APS is a heterogeneous disease with different phenotypes. Using an unsupervised hierarchical cluster analysis, we aimed to determine distinct homogeneous phenotypes among APS patients.
Methods: We performed an observational, retrospective study of APS patients enrolled in the French multicentre 'APS and SLE' registry who met the Sydney classification criteria. The clustering process involved an unsupervised multiple correspondence analysis followed by a hierarchical ascendant clustering analysis; it used 27 variables selected to cover a broad range of APS clinical and laboratory manifestations.
Results: These analyses included 509 patients, mainly women (77.8%). Mean (s.d.) age at APS diagnosis was 36.2 (14.6) years, and mean follow-up since diagnosis 10.3 (8.5) years. This hierarchical classification cluster analysis yielded four homogeneous groups of patients: cluster 1, mostly with venous thromboembolism without any associated autoimmune disease; cluster 2, older, lowest proportion of women, history of arterial events, and/or with migraines, arterial hypertension, diabetes mellitus, or dyslipidaemia; cluster 3, younger, highest proportion of women, associated SLE or other autoimmune diseases, and a history of venous thromboembolism or pregnancy morbidity; and cluster 4, mainly with a history of catastrophic antiphospholipid syndrome, aPL-associated nephropathy, and pregnancy morbidity, with frequent triple positivity and more deaths (16.7%).
Conclusions: Our study applied an unsupervised clustering method to distinguish four homogeneous APS patient subgroups that were predominantly venous; arterial; associated with SLE or another autoimmune disease; and arterial microthrombotic. Heterogeneous pathophysiological mechanisms may explain these findings.
(© The Author(s) 2022. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)