Background/Purpose: Risk of myocardial infarction (MI) is estimated to be increased up to 60% in ankylosing spondylitis (AS).  Studies in patients with rheumatoid arthritis, psoriatic arthritis and psoriasis have demonstrated reduced risk of MI in users of TNF inhibitors relative to users of other treatments. We sought to evaluate risk of MI in AS patients with use of TNFi, either alone, or in combination with other therapies, in a large US health insurance claims database.

Methods: We conducted a nested case-control study using 1994-2017 data from the OptumLabs®

Data Warehouse which includes de-identified claims data and demographic information on enrollees.  We included adults aged 18-89 with an AS diagnosis after at least 6 months of claims data prior to their AS diagnosis.  Incident MI cases were defined by diagnostic codes.  Four controls without MI were matched to each case based on sex, age (+/- 2 years) and year of AS diagnosis (+/- 2 years).   AS treatment was assessed within the year prior to MI or the matched date in controls.  Treatment was categorized as: NSAID only, symptom-modifying anti-rheumatic drugs (SMARD) only, TNFi only, combinations these categories, or none.  Odds of MI was assessed in each exposure category, relative to NSAID use alone, with adjustment for potential confounders (smoking, obesity, diabetes, chronic kidney disease stage 2 or greater, hypertension and ischemic heart disease at any time prior to AS diagnosis date) using conditional logistic regression. Additionally, a sensitivity analysis was performed assessing for exposure within the 6 months prior to index date.

Results: Among 23249 adults with AS meeting inclusion criteria, we identified 629 MI cases and 2385 matched controls.  The mean age of included subjects was 59.1 (SD 11.9) years, and 48.9% were female. Relative to NSAID use, the RR for MI among TNFi only users was 1.03 (95% CI 0.59-1.78), and for SMARD only users 0.95 (95% CI 0.68-1.32).  No combinations of NSAID, SMARD and TNFi use resulted in a significantly increased or decreased OR after adjustment for potential confounders.  Results were not materially changed with an assessment period of 6 months prior to index date.

Conclusion: In this large health insurance claims database, use of TNFi among AS patients, either alone or in combination with NSAIDs or SMARDs was not associated with a reduced risk of MI.  While this study is limited a small study sample, and there is the potential for bias related to confounding by disease severity, these findings suggest that TNFi are not cardioprotective in AS.  

Table 1. Odds of Myocardial Infarction among ankylosing spondylitis patients within each drug treatment category