LAUSR

Background Although tumour necrosis factor-alpha inhibitor (TNFi) therapy has proven efficacy in the management of psoriatic arthritis (PsA), relatively little is known about predictors of TNFi persistence. Such knowledge would assist the application of stratified medicine. Objectives To determine baseline clinical characteristics associated with TNFi persistence in patients receiving their first-line agent, and to compare TNFi persistence in first versus second-line users. Methods A retrospective cohort study was performed of all patients with PsA attending a single-centre between 2003–2016. Demographic, clinical and laboratory characteristics were compared with TNFi persistence, using Kaplan-Meier survival analysis and multivariable Cox proportional hazards models. Results A total of 188 PsA cases had used TNFi therapy as first-line over a period spanning 7,620 person months: 46% male; mean age at TNFi initiation 47.27 (SD 11.36) years; median disease duration at initiation 11 (IQR 7,16) years. Etanercept was used by 102 and adalimumab by 86. Concomitant DMARDs were used by 121/186 (65%) and 87/188 (46%) had metabolic co-morbidities (hypertension, dyslipidaemia, type-2 diabetes, obesity). TNFi therapy was terminated in 65/188 (35%) of cases (35% due to primary inefficacy, 22% secondary inefficacy, 43% adverse events), with a median duration of TNFi persistence of 26.5 (IQR 10.5, 62.0) months. Multivariable Cox proportional hazards modelling found the following parameters at TNFi initiation to be associated with shorter (poorer) TNFi persistence in first line users: female sex (hazard ratio, HR 2.57; 95% CI 1.26, 5.24; p=0.01); presence of metabolic co-morbidity (HR 2.65; 95% CI 1.24, 5.69; p=0.01); with a non-significant statistical trend towards younger age at TNFi initiation (HR 0.94; 95% CI 0.88, 1.00; p=0.06) and older age at PsA onset (HR 1.05; 95% CI 0.99, 1.12; p=0.08). Parameters not statistically associated with TNFi persistence included: choice of TNFi agent (adalimumab vs. etanercept), concomitant DMARD/methotrexate use, tender/swollen joint counts, patient global assessment (PGA) of disease activity, CRP, ESR and disease duration. Of 32 cases proceeding to a second TNFi (19 adalimumab, 13 etanercept), persistence was 14/32 (44%) over 954 person months. TNFi failure was two-fold more likely in second versus first-line users (HR 2.02; 95% CI 1.20, 3.42; p=0.009) [Figure 1], with no significant contribution from other co-variables. Conclusions Patients with PsA who are female and have metabolic co-morbidities appear to be more likely to fail first-line TNFi therapy. Contrary to observations in rheumatoid arthritis, and somewhat challenging the few studies in PsA, choice of TNFi agent (humanised monoclonal vs. chimeric), concomitant methotrexate use, acute-phase response and joint counts prior to TNFi initiation did not influence TNFi persistence. Although TNFi failure was more likely in second-line users, a significant proportion of PsA cases responded to second-line TNFi therapy, advocating this strategy in refractory cases. Acknowledgements TG and WY contributed equally to this study. Disclosure of Interest None declared