NSAIDs: Prescribing Grows Even More Complex

For years, GI toxicity and risk of bleeding were the issues of most concern when deciding to prescribe an NSAID. The cardiac effects associated with these drugs were considered a positive in that least some have been shown to provide prophylaxis against myocardial infarction.

However, we have become increasingly aware of cardiovascular (CV) adverse events associated with NSAIDs. Although these at first seemed to be primarily associated with the selective COX-2 inhibitors - - which led to the removal of rofecobix (Vioxx) from the market in the US - - it is clear that adverse events are also associated with the older non-specific NSAIDs.

A recent study by Schmidt and colleagues¹ highlights an additional dangerous CV adverse effect associated with NSAID use: atrial fibrillation. Although this problem was identified in a study published last year,² the new study indicates that it may not only be long-term users of these drugs whom we should be concerned about.

The Schmidt study was performed in Denmark using the database of the national health service. Because of the overlap in risk factors and pathophysiology, atrial fibrillation and flutter were collapsed into a single diagnostic entity. In the study, 32,602 patients with these disorders were compared with a control group of 325,918.

Relative to the control group, there was a 33% increase in cases of atrial fibrillation/flutter in users of the older, nonselective NSAIDs and a 50% increase among those using a selective COX-2 inhibitor. However, when a variety of comorbid medical conditions that could increase the risk for these conditions were factored in, there was a 17% increase in cases among non-selective NSAID users and of 27% for the COX-2 inhibitor users.

New users of the drugs (defined as those who had filled their first prescription for the drugs within 60 days before the index date) had the greatest increase in incidence rate of atrial fibrillation or flutter. There was a relative risk increase of 40% in the non-selective NSAID users and of 70% for the COX-2 inhibitor users when comorbid disorders were taken into account. This means that that there would be 4 additional cases of atrial fibrillation or flutter per year per 1000 users among the former group and 7 additional cases per 1000 users in the latter group.

The 2010 study that had identified the increased incidence of atrial fibrillation among NSAID users reported rates similar to those in the Schmidt study. However, there was one major difference between the earlier report and current study. It found that the greatest increase in risk of atrial fibrillation was among those who were long-term users of NSAIDs. The authors of the current study were unable to explain this apparent disparity.

Of the drugs studied by Schmidt and colleagues, ibuprofen (Motrin) was the NSAID least likely to be associated with atrial fibrillation or flutter among new drug users. This was followed by naproxen (Naprosyn, Aleve), etodolac (Lodine), rofecoxib (Vioxx), diclofenac (Voltaren), and celecoxib (Celebrex). Except for rofecoxib, these agents are all available in the US.

The risk for the cardiac problems was present regardless of whether high or low doses of the drugs were used. For ibuprofen, naproxen, and diclofenac, however, the risk was increased with higher doses. The elderly (not surprisingly) were at highest risk for NSAID-associated cardiac problems.

What do these results mean?

The Schmidt study indicates that NSAIDs-- older non-selective ones and selective COX-2 inhibitors-- increase the risk for atrial fibrillation or flutter. This is an additional factor that needs to be taken into account when you are deciding whether to prescribe or recommend an NSAID. Furthermore, the study suggests that it may not only be long-term users of these drugs for whom we should have heightened concerns.

A few caveats. An editorial accompanying the Schmidt paper points out several study deficiencies, including the failure to include some risk factors for atrial fibrillation that might result in NSAID use. These include obesity, which is associated with both the cardiac problem, and osteoarthritis, for which patients might be using one of these drugs.³ In other words, it is unclear whether the cardiac problems are a direct result of NSAID use or an indirect result of an underlying disorder for which the NSAID was prescribed.

That the selective COX-2 inhibitors appear in general to be more likely than older NSAIDs to be associated with increased risk of atrial fibrillation or flutter certainly fits with our growing body of knowledge of their effects. However, what is unexplained is why, of the older non-specific NSAIDs, diclofenac was by far the most likely to increase the risk of atrial fibrillation or flutter at a rate even greater than that for rofecoxib. Schmidt and colleagues identify diclofenac as a COX-2 inhibitor, but etodolac is usually cited as being more COX-2 specific than diclofenac. I therefore question whether this explanation is valid.

References
1. Schmidt M, Christiansen CF, Mehnert F, et al. Non-steroidal anti-inflammatory drug use and risk of atrial fibrillation or flutter: population based case-control study. BMJ. 2011;343:d3450
2. De Caterina R, Ruigomez A, Rodriguez LA. Long-term use of anti-inflammatory drugs and risk of atrial fibrillation. Arch Intern Med. 2010;170:1450-1455.
3. Gurwitz JH. NSAIDs and atrial fibrillation. BMJ. 2011;343:d2495. http://www.bmj.com/content/343/bmj.d2495.short.