Background Although the efficacy of TNF inhibitors (TNFi) in ankylosing spondylitis (AS) is well established, their long term survival (>5 years), baseline prognostic factors for their discontinuation and switching rate have not been adequately studied in daily clinical practice.
Objectives To estimate the long-term survival, switching rate of TNFi as well as to identify baseline predictors for drug survival in a cohort of patients with AS in real life settings.
Methods A retrospective systematic chart review of all AS patients treated with TNFi in our Rheumatology Unit from 1/1/2003 to 31/12/2015 was performed. Demographic, clinical and laboratory data were collected and analyzed. Drug survival was estimated by Kaplan-Meier analysis. The impact of several covariates on drug survival was analysed by Cox regression.
Results 103 AS patients who had received etanercept (ETN, n=41), infliximab (INF, n=33), adalimumab (ADA, n=21), golimumab (GOL, n=7) or certolizumab pegol (CZP, n=1) were included. Patients were predominantly males (67%, n=69) with a mean age of 45±11.8 years and a median disease duration of 7 years. The patients were followed for a median time of 5 years (mean: 5.1±2.6 years). Peripheral arthritis was present in 42% (n=43) and extra-articular manifestations in 26.5% (n=27) of patients. The majority (94%, n=97) received TNFi as monotherapy. The mean overall survival of the first TNFi was 81.2±6.7 months (40% survival after 7 years). ADA demonstrated a longer survival (84±8.5 months) compared both to ETN (53±6 months, p=0.04) and INF (76±11 months, p=0.063). 43 patients (42%) discontinued the 1st TNFi due to adverse events (37%), inadequate response (35%) or other reasons (28%). 10% of patients were lost to follow up. No difference in discontinuation reasons was found among the different TNFi (p=0.125). 33 patients (32%) switched to a second TNFi (ADA: n=23, ETN: n=3, INF: n=3, GOL n=2, CZP n=1). Overall the retention to therapy was better for non-switchers compared to switchers' to a second TNFi (81.2±6.7 months vs. 42.8±7 months, p=0.086). Specifically for ADA, this difference was statistically significant (p=0.03). Among several baseline variables (age, disease duration, extra-articular manifestations, peripheral arthritis, HLA B27 status), none was predictive for therapy discontinuation. Although lower baseline CRP (HR=0.7, 95% CI: 0.5–1.3), longer disease duration (HR=1.1, 95% CI: 0.9–1.3), concurrent smoking (HR=4.9, 95% CI: 0.9–25) and male gender (HR=4.8, 95% CI: 0.8–26) were associated with higher risk for drug discontinuation, these did not prove to be statistically significant by multivariate analysis.
Conclusions Only 40% of AS patients remain on their 1st TNFi 7 years after starting treatment. ADA demonstrated longer survival compared to ETN and INF as the 1st TNFi. Approximately one-third of patients switched to a 2nd TNFi which had a shorter survival compared to the 1st. In our patient cohort we did not identify any baseline predictors for longer drug survival.
Acknowledgement Supported by research grants from the Special Account for Research Grants (S.A.R.G.), National and Kapodistrian University of Athens, Athens, Greece.
Disclosure of Interest None declared