Background Non-steroidal anti-inflammatory drugs (NSAIDs) have been reported to increase cardiovascular (CV) risk. Studies on NSAIDs safety in patients with spondyloarthritis (SpA) are rare and contradictory. However, this question is highly relevant for these patients with chronic NSAIDs use and increased CV risk (1).
Objectives The objective of this study was to determine the relationship between NSAIDs use and the occurrence of CV events, including myocardial infarction (MI) and stroke, in patients with SpA.
Methods This is an ancillary study of the observational, cross-sectional, multicenter, international ASAS-COMOSPA study. The inclusion criteria were: age >18 years and SpA diagnosis. In order to overcome prescription bias when comparing patients who ever received/did not receive NSAIDs, a propensity score (PS) to predict NSAIDs intake was calculated. Patients who had never received NSAIDs were matched to patients who ever received an NSAID according to the PS. In this matched population, the probability for a patient to present a CV event was estimated by logistic regression. Furthermore, in the global population of the study, factors associated with the occurrence of CV event were explored by logistic multivariate analysis, here including the NSAIDs score (2) of the last three months, age and gender, and using the PS as an adjustment variable.
Results Of the 3984 patients enrolled in the study, data on CV event occurrence were available for 3961 patients. Among them, 3548 patients received NSAIDs (89.6%) and 378 had never received NSAIDs (10.4%). Patients who had never received NSAIDs had more often inflammatory bowel disease (9.9% vs 5.3%), and a less severe disease (% of bamboo patients: 2.0% vs 7.6%). Among the 756 matched patients, 29 (3.8%) patients reported a CV event (21 MI, 13 strokes and 5 MI + stroke). The number of patients with CV event was not significantly different between the two groups, ever and never NSAIDs exposure respectively (16 (2,1%) vs 13 (1,7%), OR =1.86 [0.496–6.989]). No difference was observed between the two groups for MI and stroke compared separately (12 (1,6%) vs 9 (1,2%), OR =2.96 [0.417–21.013] and 8 (1,1%) vs 5 (0,7%), OR =0.926 [0.288–22.108] respectively). In the second model, where PS was used as a covariate, the occurrence of overall CV event was independently associated with age (OR =2.7 [2.3–3.2]) and male sex (OR =1.9 [1.2–3.0]), but not with the NSAIDs score (OR =1.0 [0.9–1.1]).
Conclusions The use of NSAIDs does not seem to be associated with the occurence of CV event in patients with SpA, however it cannot be excluded that the study is underpowered. Further prospective studies are needed to confirm these results.
Haroon NN, et al. Ann Intern Med. 2015;163(6):409–16.
Dougados M, et al. Ann Rheum Dis. 2011;70(2):249–51.
Moltό A, et al. Ann Rheum Dis. 2016;75(6):1016–23.
Disclosure of Interest None declared