Long-term survival of biological therapy in psoriatic arthritis: 18-year analysis of a cohort in a tertiary hospital.

To study retention of biologic disease-modifying anti-rheumatic drugs (bDMARDs) or apremilast and potential predictors of lack of response in patients with psoriatic arthritis (PsA). A single-center retrospective analysis of PsA patients who received  ≥  1 bDMARD or apremilast during 2000-2018. The main endpoint was lack of response (primary or secondary failure). Analyses included retention of DMARDs (Kaplan-Meier curves) and potential predictors of lack of response (bivariate and multivariate logistic regression models). A total of 159 patients with PsA received up to 8 DMARDs: etanercept (34%), adalimumab (30%), infliximab (9%), golimumab (9%), apremilast (7%), ustekinumab (5%), certolizumab (4%), and secukinumab (2%). Therapy was discontinued in 96 cases (60%), mainly owing to secondary failure (37%), followed by primary failure (25%) and adverse effects (24%). Retention was analyzed based on 313 units of analysis. Duration of follow-up was 846.1 treatment-years (maximum 14.8 years, median 2.75 years). A total of 172 DMARDs were discontinued. The probability of continuing the initial treatment was 37% at 5 years, 22% at 10 years, and 12% at 14 years. The longest medium retention time was observed for infliximab (6.2 years) and etanercept (4.5 years). Predictors of lack of response included male sex, number of swollen joints, and, especially, depression (OR  =  35.2). The sensitivity and specificity of the model were 86.4% and 85.7%, respectively, with a coefficient of determination (R2) of 45.6 (ROC, 0.912). Rates of discontinuation due to primary and secondary failure are high in PsA. Retention is better for anti-TNF agents than for other agents.