RESUMEN
OBJECTIVES: Patients with ankylosing spondylitis (AS) have a heterogenic disease course and treatment response. Cluster-based phenotypes are useful for predicting AS disease course. Here, we compared drug retention and clinical efficacy of biologic disease-modifying anti-rheumatic drugs (bDMARDs) in AS patients with cluster A and cluster B phenotypes. METHODS: AS patients enrolled in the Korean College of Rheumatology BIOlogics registry were divided into cluster A (axial symptoms predominant) and cluster B (both axial and peripheral symptoms). Retention of bDMARDs was measured using Kaplan-Meier curve and Cox regression analyses. Clinical efficacy (BASDAI50, ASAS20, ASAS40, ASDAS inactive state, and clinically important improvement/major improvement of ASDAS) at 1-year follow-up was measured by logistic regression analysis. Also, propensity score (PS)-matched analyses were conducted. RESULTS: 1600 AS patients (1468 for cluster A, 132 for cluster B) were included. Kaplan-Meier curve analysis revealed that the drug retention rate was lower in cluster B patients (p=0.03). PS-matched analyses showed that the hazard ratio (HR) for drug discontinuation was signi cantly higher in cluster B patients (HR=1.568; 95% con dence interval =1.055-2.329). The odds ratio for BASDAI50 at 1-year was comparable between cluster A and cluster B patients in PS-matched and multivariate logistic regression analyses. A similar result was obtained in other clinical efficacy assessments. CONCLUSIONS: The drug retention rate was lower in cluster B patients than in cluster A patients; clinical efficacy was comparable between the two groups at 1-year follow-up. These results may help predict drug retention and clinical efficacy in AS patients.