The risk of acute coronary syndrome in rheumatoid arthritis in relation to tumour necrosis factor inhibitors and the risk in the general population: a national cohort study

INTRODUCTION: The elevated risk of ischaemic heart disease in patients with rheumatoid arthritis (RA) has been linked to inflammation and disease severity. Treatment with tumour necrosis factor inhibitors (TNFis) is often effective in reducing disease activity and could possibly modify cardiovascular risk. Our objective in the study was to evaluate the risk of acute coronary syndrome (ACS) in patients with RA treated with TNFis compared with the risk among biologic-naïve RA patients and the general population. METHODS: By linkage of the Swedish National Patient Register and the Swedish Biologics Register, we identified a cohort of patients who were started on their first biologic, a TNFi, between 2001 and 2010 (N = 7,704), and a cohort comprising matched biologic-naïve RA patient referents at a 3:1 ratio. Furthermore, a matched comparator cohort (5:1 ratio) was extracted from the Swedish population register. The incidence rates of a first ACS event were calculated and compared between cohorts using Cox proportional hazards regression in three different risk windows: 'ever-exposed', 'actively on TNFi' and 'short-term exposure' (active treatment maximized to 2 years). The models were adjusted for disease duration, joint surgery, comorbidity and socioeconomic factors, and, in a sensitivity analysis including a subpopulation started on therapy beginning 1 January 2006 or later, for dispensed drugs. RESULTS: Based on 221 events in 7,704 patients (comprising 32,621 person-years) treated with TNFi biologics, the hazard ratio ((HR); ever-exposed) for ACS among the TNFi-exposed RA patients compared with biologic-naïve RA patients was 0.8 (95% confidence interval (CI) = 0.7 to 0.95). In comparison with the general population referents, statistical analysis using fully adjusted models resulted in a HR of 2.0 (95% CI = 1.8 to 2.3) for biologic-naïve RA patients and a HR of 1.6 (95% CI = 1.4 to 1.9) for the TNFi-exposed group. Similar risk estimates were obtained using the other two risk windows. A sensitivity analysis in which we compared the TNFi-exposed patients included from 1 January 2006 onward with biologic-naïve patients resulted in a HR (ever-exposed) of 0.7 (95% CI = 0.5 to 1.0). CONCLUSIONS: RA patients treated with TNFi had a lower risk of ACS compared with biologic-naïve RA patients. Compared with the general population, the risk among patients with RA was elevated, although the difference was less pronounced among the TNFi-exposed patients. This finding could be attributable to the TNFi as such, or it could correspond to a lower degree of inflammation in the TNFi-treated group.