In the first study to examine the possible role of nonsteroidal anti-inflammatory drugs (NSAIDs) in the elevated risk for cardiovascular disease (CVD) of patients with osteoarthritis (OA), it was estimated that NSAIDS might account for 41% of the association between OA and CVD.
This, the authors indicate, has huge public health implications, because more than three quarters of patients with OA are treated with NSAIDs. The study was published online August 6 in Arthritis and Rheumatology.
Senior author Aslam H. Anis, PhD, from the School of Population and Public Health, the University of British Columbia, Vancouver, told Medscape Medical News, “Use of NSAIDs plays a substantial role in developing cardiovascular diseases among people with OA. Patients need counseling so that they know the risk and use NSAIDs cautiously. The findings of this study highlight the importance of monitoring for cardiovascular adverse effects among OA patients. Experts in this area have recommended alternative treatment approaches, such as exercise and physiotherapy for pain control.”
One cardiovascular expert was less convinced. Steven E. Nissen, MD, MACC, chief academic officer of the Heart and Vascular Institute and Lewis and Patricia Dickey Chair in Cardiovascular Medicine at the Cleveland Clinic, Ohio, said that the study “represents an unreliable source of evidence.” According to Nissen, the conclusions could be correct, although the type of observational study that the investigators conducted is not considered to yield high-quality evidence. Only a randomized controlled trial (RCT) can definitively answer the question of what role NSAIDs play in OA-associated CVD, he indicated.
Nissen said, “Physicians should not change their prescribing practices based upon this type of evidence. EHR [electronic health record]–based research is notoriously confounded and cannot be interpreted without better-quality evidence.”
Anis responded that this criticism reflects lack of understanding of the study’s design. Anis said, “This is not an EHR-based study. It is a validated, observational data. It has high external validity, as opposed to RCT data. Both have their strengths.” In their article, the researchers note limitations of the study, which include their having to create the mediator variable, NSAID use, from prescription claims data that did not include over-the-counter (OTC) medications. Another limitation involved the method for handling data for non-OA individuals who did not have CVD, which might have resulted in overestimation of the NSAID mediating effect.
Others Find Data Convincing
Rheumatologist Daniel H. Solomon, MD, chief of clinical sciences, Division of Rheumatology, Brigham and Women’s Hospital, and professor of medicine, Harvard Medical School, Boston, and cardiologist Dominick J. Angiolillo, MD, PhD, FACC, FESC, director of cardiovascular research, University of Florida College of Medicine–Jacksonville, told Medscape Medical News that they find data from Anis and colleagues convincing.
Solomon said, “The article is very well done and provocative. It suggests that a large portion of the CVD risk associated with OA is mediated by NSAID use. This is an important message for patients and providers. Many patients can use NSAIDs without substantial risk. However, chronic use at higher dosages in patients with risk factors for NSAID toxicity should be pursued with great caution. These risk factors include older age, male sex, tobacco use, hypertension, prior CV event, and elevated serum creatinine. If the patient also has RA, this increases the risk of NSAID toxicity further.”
Angiolillo commented, “Overall, I am convinced by the data, as there are well-established safety concerns from a cardiovascular standpoint with the use of NSAIDs. This particular analysis used a very large database, and the researchers made the necessary adjustments. This study confirms the cardiovascular impact of NSAIDs.
“This is another piece of evidence of a safety signal associated with nonsteroidals, and it highlights the need for strategies to reduce the use of these agents and to identify alternative approaches for patients with OA. One of the main issues with the treatment of these patients is that they can have considerable pain, and like any patient who has pain, they’re often more interested in immediate relief than in the long-term consequences of chronic use of NSAIDs,” Angiolillo explained.
Angiolillo agreed that because the analyses are based on administrative data, there is concern regarding the quality of those data, but he sees this article as providing “another important piece of the puzzle” of the cardiovascular concerns associated with NSAIDs.
Anis explained the origin of this study: “Our (and others’) previous research had shown that OA was an independent risk factor for CVD. It is also well known that people with OA are frequently treated with NSAIDs to control their pain and inflammation and that these NSAIDs are known to be associated with cardiovascular adverse effects. In contemplating these interrelationships, we decided to investigate the role of NSAIDs in the observed association between OA and CVD.”
The researchers set out to disentangle the role of NSAIDs in the increased risk for CVD among OA patients by analyzing longitudinal linked health administrative data from British Columbia, Canada. From a population-based cohort of 720,055 patients, they selected 7743 patients with OA and 23,229 control persons who did not have OA and who were matched for age and sex. The researchers used multivariable Cox proportional hazards models to estimate the risk for incident CVD (primary outcome) as well as ischemic heart disease, congestive heart failure, and stroke (secondary outcomes). Adjustments were made for socioeconomic status, body mass index, hypertension, diabetes, hyperlipidemia, chronic obstructive pulmonary disease, and Romano comorbidity score.
The impact of current NSAID use was estimated using linked PharmaNet data and a marginal structural model.
Patients with OA were nearly 25% more likely than patients without OA to develop CVD (adjusted hazard ratio [HR], 1.23; 95% confidence interval [CI], 1.17 – 1.28). The researchers estimate that 41% of the increased CVD risk in those with OA was related to NSAID use.
The increase in overall CVD risk included an adjusted HR of 1.42 for CHF (23% mediated through NSAID). For patients with OA, the adjusted HR was 1.17 for IHD (56% mediated through NSAID) and 1.14 for stroke (64% mediated through NSAID).
Anis explained, “Since it had been previously reported that NSAIDs were associated with 30% to 42% increased risk of cardiovascular diseases, these findings were not totally surprising. Since we were unable to account for OTC NSAIDs use, our risk estimates are likely conservative. In future studies, we hope to find a way to account for OTC medication use.”
Solomon and Angiolillo both emphasized that these data support the importance of exercising greater caution in the use of NSAIDs by patients with OA, along with the need for alternatives for pain control.
Solomon said, “Options for analgesia in high-risk patients include topical agents (topical NSAIDs, capsaicin), braces, or a cane. Weight loss, physical therapy, and tai chi have also been found in some studies to reduce pain.”
Angiolillo advised clinicians to follow the recommendations of the European Society of Cardiology (ESC) regarding NSAIDs. Step 1 of the ESC approach is to optimize treatment of underlying disease with approaches such as physiotherapy, medication, and surgery. Step 2 is to use acetaminophen plus nonpharmacologic therapies, such as weight loss, physiotherapy, and exercise. Step 3 is to use acetaminophen plus weak opioids, such as tramadol or codeine. Only at step 4 are nonselective NSAIDs recommended (naproxen ≤500 mg/day or ibuprofen ≤1200 mg/day), combined if needed with weak opioids to reduce NSAID dose and with proton pump inhibitors to prevent gastrointestinal bleeding. Step 5 is the use of COX-2 selective NSAIDs (diclofenac and the coxibs); step 5 carries the warning that these agents should be avoided in patients who have or are at high risk for cardiovascular disease.
Angiolillo added, “Completely eliminating NSAIDs might be ideal but is an unrealistic goal at present because our patients do not want to live with pain, and we do not yet have adequate alternatives to NSAIDs. Part of the message is to make patients more conscious of what they’re taking and when. I think that all patients taking NSAIDs should be informed about the cardiovascular safety concerns associated with these agents. I recommend to all my patients to minimize as much as possible the duration and the dose they take.”
The authors have disclosed no relevant financial relationships. Angiolillo has consulted for and has received honoraria from Amgen, Aralez, AstraZeneca, Bayer, Biosensors, Bristol-Myers Squibb, Chiesi, Daiichi-Sankyo, Eli Lilly, Janssen, Merck, PLx Pharma, Pfizer, Sanofi, and the Medicines Company; has received honoraria for participation in review activities from CeloNova and St Jude Medical; and has received institutional payments from Amgen, AstraZeneca, Bayer, Biosensors, CeloNova, CSL Behring, Daiichi-Sankyo, Eisai, Eli Lilly, Gilead, Janssen, Matsutani Chemical Industry, Merck, Novartis, Osprey Medical, and Renal Guard Solutions. Solomon has received research grants from Pfizer on non-NSAID-related topics and has received royalties for chapters related to NSAIDs and selective COX-2 inhibitors. Nissen has received grants from Pfizer.
Arthritis Rheum. Published online August 6, 2019. Abstract