Starting NSAIDs May Be Riskier for Some With AS and SpA

Kristensen LE, Jakobsen AK,  AsklingJ, et al.Safety of etoricoxib, celecoxib and non-selective NSAIDs in Ankylosing spondylitis and other Spondyloarthritis patients: A Swedish national population based cohort study.Arthritis Care & Research 2015; Accepted article. doi: 10.1002/acr.22555.

One of the first large studies to look at the risks of nonselective anti-inflammatory drugs (NSAIDs) in patients with ankylosing spondylitis (AS) and spondyloarthritis (SpA) suggests that doctors may be avoiding these drugs for patients who have comorbidities, and that it may be prudent to watch for congestive heart failure if starting NSAIDs in spondyloarthritis patients who never used them before.

The national registry study of almost 22,000 Swedish AS and SpA patients found that first-time users of nonselective NSAIDs (nsNSAIDs) like diclofenac or ibuprofen and selective NSAIDs such as celecoxib (Celebrex) or etoricoxib (Arcoxia) had double the relative risk of heart failure compared to those who’d taken NSAIDs before.

The study was funded by Merck, Sharp & Dohme (MSD), manufacturer of Arcoxia and also of the cyclooxygenase-2 inhibitor rofecoxib (Vioxx), which was withdrawn from the market due to concerns over cardiovascular risk. The publication says that although MSD staff contributed to the study design, the authors (none of them affiliated with the company) had complete control over the content of the report.

The authors reached two main conclusions: it's difficult to establish that any NSAID is more dangerous than any other in this patient group, but the fact that doctors react to what they learn about these drugs makes it a challenge to settle questions about their comparative safety.

Overall, none of the NSAIDs in the study significantly increased risk for serious gastrointestinal or renal events (ulcers, hypertension, or heart or kidney failure), or atherosclerotic events such as heart attacks and stroke. But atherosclerotic events, particularly congestive heart failure, were significantly elevated among those who had never before taken NSAIDs. 

The group of patients previously unexposed to any type of NSAID, who experienced more side effects from NSAIDs, had more comorbidities such as diabetes at baseline.

This appears to be a distinct group more likely to be taking medications for other conditions, the authors suggest. This "confounding by indication," which contrasts with results from some previous clinical trials. It becomes a challenge to estimate drug safety profiles, they observe, when "responsible physicians adapt to the evidence provided to them."

The study involved 21,872 AS and SpA patients in Swedish national patient registers, about equal numbers of men and women whose median age was 46. It examined records of prescription drug use between 2006 and 2009.

During the study period, 7,130 received a first-time diagnosis of AS or SpA. About 20% of the subjects never took any NSAID. More than 71% took nsNSAIDs: 8% of them celecoxib and 4% etoricoxib, which is approved in Sweden but not in the US.

About one-quarter of the patients were also taking anti-rheumatic drugs including corticosteroids and biologics at baseline, perhaps indicating more serious disease. The authors note that the advisibility of combining NSAIDs with biologics is not established.