The use of nonsteroidal anti-inflammatory drugs (NSAIDs) by individuals with osteoarthritis (OA) – who are at higher risk for cardiovascular disease (CVD) than persons without OA – substantially mediated the OA-CVD relationship, according to study results published in Arthritis & Rheumatology.

Although the mechanism has yet to be elucidated, OA is an independent risk factor for CV events, and NSAIDs are known to increase CVD risk in frequent users and may play a role in the causal pathway between OA and CVD. As people with OA often use NSAIDs to control pain and inflammation, investigators sought to determine whether and how much NSAID use mediates the OA-CVD relationship.

A longitudinal, population-based study used 1991-2013 health administrative data from 720,055 patients in British Columbia to enroll 30,972 participants (mean age, 64.53; 56.04% women) who were divided 1:3 into persons with (n=7743) and without (n=23,229) a diagnosis of OA, with non-OA control participants age- and sex-matched to their OA counterparts. The primary outcome was composite incident CVD risk, and secondary outcomes included risk for incident ischemic heart disease (IHD), congestive heart failure (CHF), and stroke.

The researchers used Cox proportional hazard modeling to estimate these risks and calculated adjusted hazard ratios (aHRs) and 95% CIs. They used a marginal structural model to estimate the mediating effect of NSAIDs in the OA-CVD association, employing linked PharmaNet data regarding current NSAID use. Multivariable adjustment was performed to account for potential confounders, including diabetes, hypertension, body mass index, socioeconomic status, chronic obstructive pulmonary disease, hyperlipidemia, and Romano comorbidity score.

Continue Reading

Compared with patients without OA, patients with OA demonstrated an increased overall risk of developing CVD, with incidence rates of 37.23% vs 28.28%, corresponding to crude rates of 38.07/1000 PY and 29.05/1000 PY for individuals with and without OA, respectively. After adjustment, patients with OA had a 23% higher risk of developing CVD compared with patients without OA (aHR 1.23; 95% CI, 1.17-1.28). On secondary outcomes, risk of CHF, IHD, and stroke were increased 42% (aHR 1.42; 95% CI, 1.33-1.51), 17% (aHR 1.17; 95% CI, 1.1-1.26), and 14% (aHR 1.14; 95% CI, 1.07-1.22), respectfully, when comparing the OA and non-OA groups.

Related Articles

Significantly more patients with OA currently used NSAIDs than did patients without OA (5.35% vs 1.07%; adjusted odds ratio 5.09; 95% CI, 4.33-5.99), and NSAID use was strongly associated with CVD risk (aHR 4.14; 95% CI, 3.8-4.5). To estimate the mediating effect of NSAIDs on the OA-CVD relationship, the researchers calculated the natural direct effect (HR 1.13; 95% CI, 1.09-1.18) and natural indirect effect (HR 1.09; 95% CI, 1.07-1.09) for the direct OA-CVD and indirect OA-NSAID-CVD associations, respectively, revealing a mediation effect of 40.59% of NSAIDs on the OA-CVD risk relationship. The corresponding mediation effects of NSAIDs on stroke, IHD, and CHF risk were 63.69%, 56.19%, and 23.31%, respectively.

Study strengths included use of data from a large population-based cohort, exclusion of patients with rheumatoid arthritis, and use of state-of-the-art mediation analysis. Study limitations included an observational design, lack of information on over-the-counter medications, possible underestimation or overestimation of the NSAID mediator variable, lack of information regarding all relevant clinical/demographic variables, and potential residual confounding that may have inflated NSAID-mediated CVD risk.

“Our findings suggest that a noteworthy portion of the increased risk of CVD among people with OA is mediated through their NSAIDs use,” noted the authors[KL1] [KT2] .


Atiquzzaman M, Karim ME, Kopec J, Wong H, Anis AH. Role of non‐steroidal anti‐inflammatory drugs (NSAIDs) in the association between osteoarthritis and cardiovascular diseases: a longitudinal study [published online August 6, 2019]. Arthritis Rheumatol. doi:10.1002/art.41027

This article originally appeared on Clinical Pain Advisor