Background/Purpose:  Anti-TNF biologics have been associated with a reduced risk for acute myocardial infarction (AMI) in some rheumatoid arthritis (RA) populations. The comparative risks for AMI associated with newer biologics with alternative mechanisms of action other than TNF blockade have not been well characterized.

Methods: Using Medicare data from 2006-2011 for patients with RA, we identified new users of abatacept, adalimumab, certolizumab, etanercept, golimumab, infliximab, rituximab and tocilizumab. New users were defined specific to each drug using a 12 month ‘baseline’ period. To increase homogeneity of patients’ characteristics to allow comparison with biologics typically not used as first line agents, patients were required to have used another biologic during baseline.

Eligible subjects must have been continuously enrolled in Medicare Parts A, B and D in baseline and throughout follow up and could not have had a hospitalized AMI during baseline. Follow up started from the drug initiation date and ended at the earliest of: AMI (primary position or non-primary), a 30 day gap in current exposure, death, loss of Medicare coverage or Dec 31, 2011. Confounding was controlled through a person-specific risk score. We calculated the incidence rate of AMI for each biologic and risks during follow-up using Cox regression adjusting for risk score decile, disability, baseline glucocorticoid use, methotrexate use, and dual eligible status, and accounting for clustering of biologic exposure treatment episodes within patients.

Results: Of 27,082 new biologic users, 11.3% initiated etanercept, 15.8% adalimumab, 6.1% certolizumab, 4.4% golimumab, 12.6% infliximab, 29.0% abatacept, 14.6% rituximab and 6.3% tocilizumab. Among 24,237 eligible RA patients, we identified 181 hospitalized AMIs, yielding an overall crude incidence rate of 0.88 / 1000py. After adjustment for potential cofounders and compared to abatacept users, the adjusted hazard ratios of various biologics were not significantly different (Table). Ongoing work is confirming the robustness of the results across a range of sensitivity analyses.

Conclusion: Among patients with RA, the risk for hospitalized myocardial infarction was similar between anti-TNF therapies and biologics with other mechanisms of action. The protective effect previously observed with anti-TNF therapies may apply more broadly to newer biologics and may be a consequence of reduction in inflammation rather than a drug-specific effect.