To examine relative cardiovascular risks of NSAIDs, scientists from Bern University in Switzerland analyzed data from 31 randomized, controlled clinical trials involving patients taking either placebo or one of the following NSAIDs for noncancer pain: naproxen, ibuprofen, diclofenac, or the selective COX-2 inhibitor celecoxib (Celebrex®). They also included trials of the selective COX-2 inhibitors etoricoxib (Arcoxia®), rofecoxib (Vioxx®), and lumiracoxib (Prexige®) — none of which are currently marketed in the U.S. or most other countries. Altogether, the trials enrolled 116,429 patients with more than 115,000 patient years of follow-up. The primary outcome examined was myocardial infarction and, secondarily, stroke, death from cardiovascular disease, and death from any cause [Trelle et al. 2011].
Reporting their findings in the British Medical Journal, the investigators found that compared with placebo ibuprofen was associated with the highest risk of stroke (more than a 3-fold increase) followed closely by diclofenac. Rofecoxib was associated with the highest risk of myocardial infarction (more than 2-fold increase) followed closely by lumiracoxib. Diclofenac and etoricoxib were both associated with the highest risk of cardiovascular death (roughly 4-fold increases). Overall, celecoxib had only slightly more favorable risks than the other 3 COX-2 inhibitors. Naproxen was found to have the best safety profile in terms of myocardial infarction or cardiovascular death risks, although results for this agent regarding stroke or death from any cause were somewhat equivocal.
While the absolute numbers of adverse cardiovascular events among patients taking NSAIDs were low, the researchers found that, relative to placebo, two of the most commonly used drugs — ibuprofen and diclofenac — carried important cardiovascular risks. They concluded, “although uncertainty remains, little evidence exists to suggest that any of the investigated drugs are safe in cardiovascular terms. Naproxen seemed least harmful. Cardiovascular risk needs to be taken into account when prescribing any nonsteroidal anti-inflammatory drug.”
COMMENTARY: We have previously discussed concerns with NSAIDs when it comes to adverse gastrointestinal events [here] and cardiovascular risks [here], as well as more broadly in our series on NSAID risks [here]. Therefore, this latest research report comes as no surprise; however, there are some limitations of the evidence.
In their approach, the Swiss researchers used a technique called “network meta-analysis,” a complex approach that allows for a coherent analysis of somewhat disparate data, as well as indirect comparisons of one drug with another or with placebo that are not otherwise possible. More specifically, the approach used statistical methods permitting researchers to indirectly compare the relative safety of different NSAID medications, even if they had never gone head-to-head in the same clinical trial. This is both a strength and a weakness of the technique and results should be carefully examined and cautiously accepted.
Furthermore, as noted above, considering the large number of trials and more than 116,000 enrolled patients in the meta-analysis, the absolute numbers of adverse events reported were relatively low. For example, data from 29 of the trials showed there were 554 heart attacks in total, in 26 trials assessed there were 377 strokes, and 28 trials found a total of 676 deaths. Along with this, Trelle and colleagues  recommend viewing the results in context. In everyday practice, many users of NSIADs will only take them for relatively brief periods of time, or intermittently, to treat acute episodes of pain and might thereby be at minimal increased risk of cardiovascular events.
On the other hand, if NSAIDs are taken regularly for more chronic pain conditions, the data from Trelle et al. suggest relatively small, unfavorable numbers needed to harm (NNH): for example, only 25 to 50 patients would need to take NSAIDs for one year to cause one additional serious event such as a heart attack or stroke. Since millions of people take NSAIDs daily, and many of them are elderly and may already have risk factors for cardiovascular disease, patients with chronic pain who are prescribed NSAIDs for longer periods might be exposed to some excess risk. While naproxen appears to be safest in terms of cardiovascular risk, this advantage must be weighed against the drug’s gastrointestinal toxicity and the need for concomitant gastroprotective therapy (eg, proton pump inhibitor) in many patients.
Many think that topical NSAID gels and patches may relieve pain without as many adverse effects for the stomach and heart, as compared with oral NSAID formulations. Research does suggest that topical NSAIDs may limit, but not eliminate, risks of adverse events; however, there still may be some concern regarding their use in elderly patients [discussed here].
There are other NSAIDs to choose among — such as fenoprofen, indomethacin, meloxicam, piroxicam, etc. — but research evidence to adequately assess comparative cardiovascular and gastrointestinal risks with these is insufficient. Simple aspirin, in analgesic doses, is usually not included in research comparisons of NSAIDs, although gastric toxicity concerns with aspirin are well documented.
Additional alternatives include acetaminophen (paracetamol) or opioids. For many types of pain, acetaminophen may be inadequate and there are concerns about clinically relevant hepatotoxicity in some patients, even at recommended dosages. Analgesic effects of opioids are the most pronounced but associated with their own risks of potential adverse events, including recently alleged, albeit questionable, cardiac and bone-fracture risks [discussed in blogpost here].
Controversies and confusion surrounding the cardiovascular safety profiles of commonly prescribed analgesics are problematic. These agents must be evaluated in terms of the symptoms each addresses along with their infrequent yet serious adverse events, and how those might relate to individual patient health factors. The perfect analgesic with respect to safety and effectiveness still awaits discovery; meanwhile, healthcare providers need to be aware of potential cardiovascular risks of the various analgesics and take those into account when prescribing. And, patients need to be made aware that many NSAIDs readily available without prescription, over-the-counter, can be seriously harmful if taken regularly.
ADDENDUM: An earlier report examined a large study population in Denmark comprised of 1,028,427 apparently healthy individuals with a median age of 39 years [Fosbøl et al. 2010]. At least 1 prescription of an NSAID was claimed by roughly 45% of the study population from 1997 to 2005. Analyses, using data mining techniques, found that diclofenac and, to a lesser extent, ibuprofen were significantly associated (in a dose-dependent fashion) with increased rates of cardiovascular complications, including death, and, in high doses, fatal or nonfatal stroke. The selective COX-2 inhibitor rofecoxib was significantly related to increased cardiovascular risks but there was only a nonsignificant trend for increased stroke risks. Celecoxib was not associated with excess cardiovascular death or stroke risks, but there was a trend toward increased MI and coronary death risks. Finally, use of naproxen was neutral in terms of risks except for a trend toward increased risk of stroke. In sum, this study appears supportive of the findings in the above report by Trelle et al. , but it also has many of the same limitations.REFERENCES:
> Fosbøl EL, Folke F, Jacobsen S, et al. Cause-Specific Cardiovascular Risk Associated With Nonsteroidal Antiinflammatory Drugs Among Healthy Individuals. Circ Cardiovasc Qual Outcomes. 2010;3(4):395-405 [full text article here]
> Trelle S, Reichenbach S, Wandel S, et al. Cardiovascular safety of non-steroidal anti-inflammatory drugs: network meta-analysis. BMJ. 2011;342:c7086 [full text article here].