| Abstract |
Background
Inflammatory arthritides [rheumatoid arthritis (RA) and spondyloarthritis (SpA)] are chronic diseases with long-term consequences which affect 2-3% of the population resulting in a significant individual and societal burden. The advent of biologic agents, such as tumor necrosis factor inhibitors (TNFis), has dramatically transformed the management of these diseases. TNFis were proven effective in randomized controlled studies (RCTs) of both RA and SpA patients. However, questions that influence clinical decision making are insufficiently addressed by RCTs. Longitudinal observation studies and registry data can provide valuable information to optimize clinical use of these novel and expensive drugs. Greek nationwide data regarding effectiveness and safety of TNF inhibitors in RA and SpA patients are lacking. This is important in view of the variations in disease severity of inflammatory arthritides across different ethnic backgrounds and local variations of clinical practice.
Objectives
In the present study we sought to assess the effectiveness and the safety of the TNFi therapy in a nationwide cohort of Greek patients with inflammatory arthritides focusing in patients with RA and SpA. In the study of patients with RA we aimed to compare the effectiveness, the drug survival and the safety between the three TNFis infliximab, adalimumab, and etanercept, and to identify potential predictors of response, drug survival and serious adverse events. In the study of SpA patients, our objective was to evaluate the 10-year drug survival of the first TNFi in patients overall and comparatively between SpA sub-diagnoses and between different TNFis. Predictors of drug retention were also sought among baseline parameters and early major response variables.
Methods
We organized the “Hellenic Registry of Biologic Therapies (HeRBT)”, a prospective observational cohort of patients who receive biologic therapies for inflammatory arthritides in 8 hospitals of Greece. All consecutive patients in the participating centers are included in HeRBT when they start their first biologic agent. According to the protocol, baseline data, response data and events are collected every 6 months for the first 2 years and every year thereafter. For the first study, 1208 adult RA patients starting infliximab, adalimumab or etanercept between January 2004 and December 2009 were identified. The observational period was until May 2011. Clinical responses were assessed by several outcome measures (DAS28, CDAI, EULAR response criteria). Drug survival and serious adverse events during entire follow-up (median 2.9 years) were also monitored. For the second study, 1077 adult spondyloarthritis patients starting their first TNFi between 2004 and the end of 2014 were analyzed. Monitoring period was until May 2015. 10-year drug survival rates and 6- and 12- month rates of response to therapy were calculated applying standard outcome measures (BASDAI50, ASDAS). We used standard descriptive statistics, Kaplan-Meier curves and logistic and Cox regression models. In the second study we used multiple imputation for our main Cox regression analyses, but complete-case analyses were also performed.
Results
Concerning RA patients, EULAR response (good and moderate combined) was achieved by 79% of the patients at 12 months in the three TNFi groups while remission rates were low: 13-16% and 15-23% of patients (DAS28-remission at 6 and 12 months respectively) and was comparable between the three TNFis. In multivariate analysis adalimumb was associated with greater odds for remission [adjusted odds ratio (OR) for EULAR/ACR remission at 12 months (reference: infliximab): 4.1 for adalimumab and 2.7 for etanercept]. Other baseline factors independently predicting remission were male gender (OR 2.2), use of glucocorticoids (OR 2.2) and swollen joint count >7 (OR 0.26). Five-year drug survival was 31%, 43%, and 49% for infliximab, adalimumab and etanercept, respectively (log-rank p=0.010). Although efficacy-related survival was comparable, infliximab was associated with significantly more withdrawals due to adverse events (p<0.001). Lower baseline disease activity, higher baseline CRP and the use of glucocorticoids predicted longer efficacy-related drug survival. Younger age, no use of methotrexate, use of adalimumab and etanercept and less prior DMARDs failures predicted longer safety-related survival. Interestingly, adjusted 5-year drug survival was highest for patients with sustained (both at 6 and 12 months) DAS28 remission compared to patients with poorer clinical responses during the 1st year (p<0.001). The incidence rate of serious adverse events (SAEs) was 8.5, 5.3 and 3.5 per 100 patient-years in the infliximab, adalimumab and etanercept groups respectively (p<0.001). The risk a first serious infection was lower with adalimumab (OR 0.62) or etanercept (OR 0.39) than with infliximab. Other independent predictors of a serious infection at baseline were higher age (OR 1.65 per 10-years), tender joint count >10 (OR 1.86), and glucocorticoids >35 mg/week (OR 1.83).
Concerning SpA patients, we analysed 561 with AS, 375 with PsA, 108 with uSpA and 33 with IBD-related SpA. Five- and 10-year drug survival was 60% and 49% respectively. In the unadjusted analyses, TNFi survival was associated to isolated axial disease (p=0.001). Regarding SpA subdiagnosis, AS patients had longer drug survival compared to uSpA and PsA patients [(significant beyond the first 2.5 (p=0.003) and 7 years respectively (p<0.001)]. In the multivariable analysis, men had a significantly longer TNFi adherence [hazard rate (HR) 0.68], both for efficacy (HR 0.6) and safety-related (HR 0.57) reasons of discontinuation. Use of a monoclonal antibody was associated with a longer overall drug survival (HR 0.64), but etanercept had less safety-related stops compared to infliximab (HR 0.52). Finally, the use of methotrexate was protective, mainly through preventing safety-related stops (HR 0.6). Among patients having axial SpA, 59% and 42% achieved BASDAI50 or had ASDAS-ID respectively within the first year of therapy. Achievement of major responses during the first year of therapy in either axial or peripheral arthritis was the strongest predictor of longer therapy retention (HR 0.33 for ASDAS-ID and HR 0.35 for DAS28 remission respectively). Conclusions
These data based on the largest Greek cohort of patients with systemic arthritides reassured about safety of TNFis in clinical practice. Greek RA patients starting TNF inhibitors have comparable response rates across the 3 different TNFis, while remission rates are low in clinical practice. Overall, 5-year drug survival was below 50%, with infliximab demonstrating increased safety-related discontinuations. The long-term retention of the first TNFi in SpA patients is high, especially for males with axial disease. The strongest predictor of long-term TNFi survival is a major response within the first year of treatment. Strategies to increase effectiveness and long-term survival of TNF inhibitors in RA and SpA are needed.
|
Search
