Background/Purpose: To examine the association of cardiovascular, malignant and serious infection events (SIEs) with certolizumab pegol (CZP) use as compared to other TNFi agents in patients (pts) with rheumatoid arthritis (RA).

Methods: RA pts enrolled in the Corrona national registry who were ≥18 years of age when initiating CZP or other TNFi agents (etanercept, adalimumab, golimumab or infliximab) on or after May 1^st, 2009 with at least 1 follow-up visit post index drug initiation were identified. Physician-reported cardiovascular events, malignancies and SIEs were captured during the follow-up period, which were measured separately for each adverse event type in pt-years (PY), from drug initiation to either: occurrence of the first event, 90 days following pt/physician-reported discontinuation/switch of biologic (censored), or the 12 month Corrona visit (censored). The analysis before propensity score matching (PSM) included all initiators. After PSM, the analysis included CZP and other TNFi pts matched with a propensity score, to control for disease and pt characteristics that differed between the two groups (age, gender, disease duration, Clinical Disease Activity Index [CDAI] at initiation, and line of therapy).

Results: There were 975 RA pts initiating CZP and 5240 RA pts initiating other TNFi. CZP was initiated later in the treatment course in terms of 1^st, 2^nd or ≥3^rd line of therapy (CZP: 26%, 29%, 45% respectively vs other TNFi: 50%, 30%, 20% respectively; p<0.001). Consequently, those initiating CZP were older (median age 58 vs 56 years; p<0.001), had longer median disease duration (9 vs 5 years; p<0.001), more active disease based on median CDAI (19 vs 17; p=0.01), and more functional impairment based on median modified health assessment questionnaire (0.5 vs 0.4; p<0.001). At initiation, history of cardiovascular events (5.8% vs 5.4%; p=0.57), malignancy (4.9% vs 4.3%; p=0.38), SIEs (6.6% in each group; p=0.93), and insurance status (yes/no) were balanced between groups. Before PSM, the incidence rates (IR) per 100 PY for cardiovascular events and malignancies were similar in the two groups, while the IR for SIEs was higher in the CZP group (Table). However, PSM resulted in the resolution of clinically significant baseline differences, and, in the 952 propensity score matched initiators, IRs of cardiovascular events, malignancies and SIEs were similar between the two groups (Table).

Conclusion: Accounting for baseline differences using PSM demonstrated that CZP is not associated with an increased risk of cardiovascular events, malignancies or SIEs compared to other TNFi agents utilized in earlier lines of RA therapy. Failure to account for baseline differences (disease duration, disease severity and functional impairment) may have contributed to the difference in SIE IRs between CZP and the other TNFi groups in the cohort before PSM.

« Back to ACR/ARHP Annual Meeting 2016

ACR Meeting Abstracts - https://acrabstracts.org/abstract/the-real-world-comparative-safety-of-certolizumab-pegol-czp-as-compared-to-other-tnfi-in-a-national-us-cohort/