Background To establish better treatment strategies in RA, it is important to identify the factors that are responsible for drug discontinuation. Several studies had compared drug retention rates (RR) among TNF inhibitors but, to date, none has systematically analyzed the long-term survival of all approved biological agents for RA. The objective of this study was to compare the drug RR in daily clinical conditions of all approved BTs in RA, and to identify the risk factors associated with a poor RR.
Objectives The objective of this study was to compare the drug RR in daily clinical conditions of all approved BTs in RA, and to identify the risk factors associated with a poor RR.
Methods RA patients treated with ≥1 BT during the period December'99-March'13 were included. For each patient, a large number of clinical and epidemiological variables were collected. DAS28 score was recorded at baseline and every 3 months. Date of start of therapy as well as date and mean reason of discontinuation were also registered. Drug RR were calculated using the Kaplan–Meier method and compared using the log-rank test. Cox models were used in the multivariate comparison between BTs.
Results This analysis included 452 BT in 222 RA patients (1249.5 patient-years (PY)). The exposure according to each treatment was 30.1 PY for abatacept, 203.5 PY for adalimumab, 6.6 PY for anakinra, 9.8 PY for certolizumab group, 496.5 PY for etanercept, 29.0 PY for golimumab, 312.9 PY for infliximab, 86.9 PY for rituximab and 74.3 PY for tocilizumab. The mean drug survival was 4.0 (±0.2) years and the overall drug survival rates at 1, 2, 3, 4 and 5 years were 62.6% (58.3-67.2), 48.5% (44.1-53.4), 41.4 (37.0-46.3), 36.0% (31.6-40.9) and 32.0% (27.7-37.0) respectively.
When comparing RR of each agent against the overall treatment group, we have found that etanercept (ETN) had the lowest discontinuation risk (P=1.59e-05; HR=0.57[95%CI, 0.44-0.74]). Then we compared each agent against ETN, and we observed that anakinra (ANK) treatment had the highest discontinuation risk (P=2.88e-07; HR=5.91[95% CI, 3.00-11.64]).
In the cox regression model, we have found that a longer disease duration (P=3.26e-02; HR=1.37 [95%CI, 1.03-1.83]), previous treatment with ≥2 DMARDs (P=1.72e-02; HR=1.47 [95%CI, 1.07-2.03]) and ≥2BT (P=4.53e-03; HR=1.60 [95%CI,1.16-2.22]) were associated with reduce drug survival. DAS28≥5,1 (P=3.75e-03; HR=1.64 [95%CI, 1.17-2.29]) was also associated to a higher risk of discontinuation. When we compared the risk of treatment withdrawn due to an adverse event, we found that ETN treatment hat lowest risk of discontinuation (P=3.82E-04, HR=0.48 [95%CI, 0.32-0.72]) and ANK was the BT with significantly higher risk of withdrawn due to an adverse event (P=4.04e-05, HR=6.56 [95%CI, 2.67-16.12]).
Conclusions In our series of patients with RA treated with BT, the discontinuation risk was significantly lower in patients treated with ETN, meanwhile ANK had the lowest RR. Factors associated with the withdrawal of the BT were disease duration, previous treatment with more than 2 BT and DMRADs, as well as DAS28≥5.1. The results of this study contributes to a better understanding of the survival of BT in rheumatoid arthritis patients.
Disclosure of Interest None declared