Background/Purpose: Patients with RA who experience an inadequate response to TNF inhibitor (TNFi) therapy (TNF-IR) may be successfully treated using an alternative TNFi or a biologic with a different mode of action such as rituximab (RTX) or abatacept (ABA). The relative effectiveness of these approaches has not been determined in head-to-head trials. Evidence from real-world experience would help to guide treatment decisions for TNF-IR patients.

Methods: This was a retrospective chart review of patients from a single site (Rebecca MacDonald Center) with ACR-defined RA who initiated RTX or ABA after failure of at least one TNFi. The relative effectiveness of RTX and ABA was evaluated by analyzing drug survival distribution. Kaplan-Meier survival curves were compared using the log rank test (p-values <0.05 indicated statistical significance).

Results: The study cohort comprised 61 patients, of whom 37 and 24 were treated with RTX and ABA, respectively. In the RTX group, 10 patients had also received prior therapy with ABA. Demographics and disease characteristics were generally similar in the two groups, although RTX patients had higher disease activity compared with ABA patients (mean CDAI: RTX 32.5 vs ABA 26.9) and had received more prior TNFis (1.8 vs 1.7) and/or ABA (2.1 vs 1.7). After excluding the 10 RTX patients who had received prior ABA, survival rates over time were generally better with RTX vs ABA (table).

Estimated survival rates

Overall survival distribution (figure) was not significantly different between the RTX and ABA groups (p=0.658); however, RTX patients who had also failed ABA had significantly reduced survival compared with those who had not received ABA (p=0.015).

Overall survival distribution

Stratification of survival data according to number of prior TNFi failures indicated that RTX (excluding patients with prior ABA) was superior to ABA among patients who had failed 3 TNFis. Survival was also numerically greater with RTX vs ABA in patients who had failed 1 prior TNFi. Finally, survival was better on both RTX and ABA when compared with that seen in a separate cohort of patients (n=88) who received a second TNFi after first TNFi failure.

Conclusion: These results from real-life practice suggest that in RA patients who have failed one or more TNFi, RTX may have better long-term survival than either ABA or an alternative TNFi. Prior ABA appears to reduce the efficacy of RTX. Further data are needed.