Phenotypic differences and treatment persistence in a large multicenter cohort of axial psoriatic arthritis

Objective To compare clinical features across three axial psoriatic arthritis (PsA) phenotypes: (i) isolated axial PsA, (ii) axial PsA with oligoarticular involvement (axial PsA+oligo), and (iii) axial PsA with polyarticular disease (axial PsA+poly), and assess persistence of first-line biologic and targeted synthetic DMARDs (b/tsDMARDs). Methods Baseline demographic, clinical, therapeutic, laboratory, and imaging data were retrospectively collected. Kaplan–Meier analysis evaluated treatment persistence, and LASSO Cox regression identified baseline factors associated with discontinuation. Results Among 621 patients, 175 (28.2%) had isolated axial PsA, 252 (40.6%) axial PsA+oligo, and 194 (31.2%) axial PsA+poly. Psoriasis was more frequent in isolated axial PsA and axial PsA+oligo than in axial PsA+poly (82.3% and 82.1% vs 66.1%; p< 0.001), with lower HLA-B27 positivity in axial PsA+poly (10.9% vs 18.9% in isolated axial PsA and 21.6% in axial PsA+oligo; p= 0.02). Sacroiliitis was more frequent in isolated axial PsA (70.3%) and axial PsA+oligo (65.7%) than in axial PsA+poly (53.6%, p< 0.001), whereas spondylitis was more frequent in axial PsA+poly (15.0% vs 8.3% in axial PsA+oligo and 6.3% in isolated axial PsA; p< 0.001). Treatment persistence was longest in axial PsA+oligo (62 months, 95% CI 43–67) and shortest in axial PsA+poly (38 months, 95% CI 24–49) (p= 0.006). Higher VAS global pain was associated with shorter persistence (aHR 1.010, 95% CI 1.003–1.018, p= 0.009), whereas HLA-B27 positivity (aHR 0.56, 95% CI 0.32–0.99, p= 0.045) with longer persistence. Conclusion Treatment persistence in axial PsA varies according to clinical phenotypes, HLA-B27 status and pain, supporting the relevance of phenotypic stratification in clinical practice.