Background:Biological Disease-Modifying Anti-Rheumatic Drugs (bDMARDs) targeting Tumor Necrosis Factor (TNF) alpha, Interleukin (IL) 12/23 and IL17, have been approved for Psoriatic Arthritis (PsA) treatment, in this chronological order.Objectives:The aims of our study were to evaluate 1) predictors of first bDMARD failure, including mechanism of action 2) factors associated to failure of multiple (>=2) therapies.Methods:Consecutive patients attending our Rheumatology Unit, classified as PsA according to CASPAR criteria and beginning treatment with bDMARDs in the period 2004-2020, were enrolled. Disease characteristics, previous/ongoing treatments, comorbidities and follow-up duration were recorded. Demographic, clinical and laboratory data, disease-activity and functional indexes [including Disease Activity in PSoriatic Arthritis (DAPSA), PASI (Psoriasis Area and Severity Index) and Health Assessment Questionnaire (HAQ)], were recorded at baseline and yearly and were compared between switchers (>=1 switch/swap) and non-switchers. Date and reason for switching were collected. Effectiveness was evaluated over-time with descriptive statistics. A multivariable Cox-Proportional-Hazard (PH) model was built to evaluate the influence of mechanism of action (anti-TNFalpha/anti-IL12/23/anti-IL17) and of negative prognostic factors for drug response on time to first bDMARD discontinuation. Furthermore, a multivariable logistic regression model was built to assess the association between negative prognostic factors for drug response (independent variables) and failure of>=2 bDMARDs (“multifailure”, outcome). Kaplan-Meier curves were used to assess differences in time-to-first bDMARD discontinuation according to the targeted cytokine. P values <=0.05 were considered significant. Infections and adverse events were recorded.Results:Our study included 264 patients, 117 (44.32%) females, mean age 56±12 years, mean PsA duration 15±3 years;117 (44.32%) switched bDMARDs at least once. Switchers were mostly females, with higher PASI and worse HAQ at baseline (Figure 1). Mean time-to-first bDMARD discontinuation was 72 months; 2-year and 5-year retention rate was 75% and 60%, respectively. Survival curves for anti-TNFalpha/anti-IL12/23/anti-IL17 were similar (log-rak test=0.83;p=0.66). Main reasons for switching were inefficacy (79) and adverse events (38). The Cox PH model showed that female sex was independently associated to a higher risk of first bDMARD discontinuation (HR=2.39; 95%CI:1.50-3.81), while initiating therapy before 2015 was protective (HR=0.40; 95%CI:0.22-0.73). Other independent variables, including mechanism of action (HR=0.76; 95%CI:0.30-1.74 for anti-IL17; HR=0.53; 95%CI 0.15-1.86 for anti-IL12/23; reference: anti-TNFalpha), age (HR=1.00; 95%CI:0.99-1.03), baseline DAPSA (HR=0.98; 95%CI:0.96-1.00), PASI (HR=0.95; 95%CI:0.86-1.04), HAQ (HR=1.29; 95%CI:0.91-1.83), Body Mass Index BMI (HR=1.02; 95%CI:0.98-1.07) and comorbidities (HR=1.10; 95%CI:0.92-1.31) were not associated to the outcome. In the logistic regression model, only female sex was significantly associated to failure of>=2 therapies (OR=1.99, 95%CI:1.07-3.69); bDMARD mechanism of action, age, and initiating therapy before 2015 were instead not independently associated.Conclusion:Survival rate was good for anti-TNFalpha and other bDMARDs. Female sex was a predictor of first bDMARD discontinuation, unlike mechanism of action, comorbidities and BMI.Figure 1.Disclosure of Interests:Mariagrazia Lorenzin: None declared., Augusta Ortolan: None declared., Giacomo Cozzi: None declared., Antonia Calligaro: None declared., Maria Favaro: None declared., Teresa Del Ross: None declared., Andrea Doria Grant/research support from: AD has received honoraria and speaker fees from Novartis, Abbvie, Pfizer, MSD, Janssen., Roberta Ramonda Grant/research support from: RR has received honoraria and speaker fees from Novartis, Abbvie, Pfizer, MSD, Janssen.