Has the excess risk of acute myocardial infarction in rheumatoid arthritis relative to the general population declined? A population study of trends over time

Abstract

ObjectiveTo evaluate secular trend in ten-year risk of incident acute myocardial infarction (AMI) in incident rheumatoid arthritis (RA) relative to the general population. MethodsWe conducted a retrospective study of population-based incident RA cohorts with RA incidence from 1997 to 2004 in British Columbia, Canada, with matched general population comparators, using administrative health data. RA and their matched cohorts were divided according to the year of RA incidence, defined according to the first RA visit of the case definition. Incident AMI was defined as the first event occurring within 10 years from RA incidence. Secular trend was assessed using delayed-entry Cox models with an interaction term between the year of RA onset and indicator of RA vs. general population. Linear, quadratic and spline functions of year of RA onset were compared to assess possibility of nonlinear trends. The model with the lowest AIC was selected to interpret the results. Sensitivity analyses were conducted to account for potential effect of unmeasured (e.g. smoking) or partially measured (e.g. obesity) confounders in administrative data, on the interaction term. ResultsOverall, 23,237 RA and 46,474 general population controls experienced 1,133 and 1,606 incident AMIs, respectively. A linear Cox model was selected as the model best fitting the AMI events. Overall, RA patients were found to have a 21% higher risk of AMI than the matched general population controls [1.21 (1.10, 1.32); p < 0.001]. A significant linear decline in risk of AMI was observed in RA patients [0.94 (95% CI 0.91, 0.97) p = <0.0001], and in the general population [0.93 (0.91, 0.95); p = <0.0001]. The change in AMI risk over time did not differ in RA compared to the general population [p-value of interaction term=0.49]. Our results remained similar after adjusting for the potential effect of confounders on the interaction term, and no difference in the change in risk of AMI over time was observed between RA and the general population. ConclusionOur findings suggest a decline in 10-year risk of AMI in RA, and in the general population. The decline in the risk of AMI over time did not differ between RA and the general population, such that the excess risk of AMI in RA relative to the general population, has remained the same.