Tuesday, March 6, 2012

Cardiac Risks of NSAIDs Revisited

NSAIDs In agreement with previous results from randomized trials, a new systematic review of observational studies covering nearly 3-million persons corroborates that nonsteroidal anti-inflammatory drugs (NSAIDs) may increase the risks of serious cardiovascular events. This latest review confirms that among widely used NSAIDs, often prescribed for musculoskeletal pain and headache disorders, naproxen and low-dose ibuprofen are least likely to increase cardiovascular risks; whereas, diclofenac most significantly elevates those risks even at lower doses.

Writing online in PLoS Medicine, Patricia McGettigan, from Hull York Medical School, UK, and David Henry, of the Institute for Clinical Evaluative Sciences, Toronto, Canada, report a systematic review and meta-analysis of community-based controlled observational studies to compare cardiac risks of NSAIDs at typical daily doses [McGettigan and Henry 2011]. Their literature search, spanning 1985 through 2010 and examining 11 different NSAIDs [excluding aspirin), uncovered 30 case-control studies that included 184,946 cardiovascular events, and 21 cohort studies describing outcomes in more than 2.7 million NSAID-exposed individuals.

Of the more extensively studied NSAIDs (≥10 studies each), the highest overall cardiac risks were seen with rofecoxib and diclofenac, and the lowest with ibuprofen and naproxen. In a sub-set of studies, risk increases with ibuprofen were seen only with higher doses and naproxen was risk-neutral at all doses. Risks were elevated with even low doses of rofecoxib and diclofenac, and the risk increased in each case with higher doses. Of the less studied drugs (<10 studies each), etoricoxib, etodolac, and indomethacin carried relatively high risks.

The authors conclude that, from a clinical perspective, naproxen and low-dose ibuprofen have the most favorable cardiovascular risk profiles. However, this advantage should be weighed against the gastrointestinal risks of both drugs, and the antagonism of low-dose aspirin's beneficial effects that ibuprofen may incur. Additionally, the authors stress that patients at high risk of cardiovascular disease should be warned about not exceeding the maximum recommended dose (1,200 mg/day) for non-prescription ibuprofen.

While celecoxib was indistinguishable from naproxen in separate comparisons, the authors are reluctant to recommend celecoxib in high-risk patients. The data for etoricoxib (not marketed in the United States) are limited but this drug had a significantly higher relative risk than naproxen or ibuprofen, which raises concerns about its safety, particularly since etoricoxib analogues such as rofecoxib and lumiracoxib have been withdrawn from the market.

The authors expressed special concern about cardiac risks with diclofenac, particularly since it is available without prescription in several countries. The data show an elevated risk even with low doses of diclofenac, unlike most competitor drugs. The review also casts doubts on the safety of an older drug, indomethacin, which has a range of gastrointestinal and central nervous system effects that, combined with concerns about its cardiovascular risk, might question its continued clinical use, according to the authors.

COMMENTARY: The results of this systematic review by McGettigan and Henry are consistent with our earlier examination of randomized, controlled trials [see UPDATE here], which showed that all NSAIDs increase risks of cardiovascular adverse events to some degree. In that assessment, covering only 7 drugs and roughly 116,000 patients, naproxen was found to have the most favorable cardiac safety profile, while diclofenac and ibuprofen were of most significant concern.

This newer review included more patients and 4 additional NSAIDs not covered in the meta-analyses of randomized trial data: piroxicam, etodolac, meloxicam, and indomethacin. Except for indomethacin, data for the other agents were relatively sparse. Furthermore, all studies in this review were nonrandomized observational designs, which are a somewhat lower quality of evidence subject to some biases, such as in patient selection and data collection. A large amount of data were available from the 51 studies for analysis and, perhaps, over-analysis, since the authors examined the data in various ways to address questions that were not intended in the original designs.

NSAID RRs The graph at right shows the pooled relative-risk point estimates (RR, triangular dots) for each drug studied, along with their respective 95% confidence intervals (CIs, vertical lines). Some of the differences between RRs for a number of the NSAIDs were small, often with overlapping confidence intervals (CIs) suggesting a lack of significant difference in cardiac risk between them. At the same time, some of the CIs were relatively wide — eg, for etoricoxib, etodolac, piroxicam, and valdecoxib — and reflect the more limited data regarding these NSAIDs available for analysis and less precision in their RR point estimates.

Data collection for many of the studies relied on large databases and/or electronic health records, which, among other problems, do not record the use of over-the-counter NSAIDs or aspirin. And, unfortunately, this systematic review does not provide data on absolute risk increases with the drugs; that is, the incidence rates — such as, the number of adverse cardiovascular events per 1,000 patients taking the drugs — potentially resulting from specific NSAIDs.

Finally, in most of the analyses conducted by McGettigan and Henry, heterogeneity was significant. That is, outcome results of the collective studies for individual NSAIDs often varied considerably, leading to some inconsistencies even when the pooled RR point estimates and CIs seemed to be precise. Heterogeneity does not necessarily negate validity of meta-analysis results, but it does suggest that outcomes should be accepted circumspectly.

The authors affirm that, despite limitations of the data, the large sizes of the studies they reviewed, the presence of coherent relationships, and general agreement with the results of randomized trials provides some confidence in the results that they found. Therefore, they believe that the study is sufficiently robust to inform clinical decisions regarding the use of NSAIDs. Certainly, patient age, pre-existing cardiac risk factors, and dose and duration of NSAID therapy could play important roles, but the most general message seems to be that, along with the gastrointestinal risks of these agents, a potential for cardiovascular adverse events always should be taken into consideration when recommending or prescribing NSAIDs for patients with pain.

REFERENCE: McGettigan P, Henry D. Cardiovascular Risk with Non-Steroidal Anti-Inflammatory Drugs: Systematic Review of Population-Based Controlled Observational Studies. PLoS Med. 2011;8(9):e1001098 [full article here].

eNotificationsl Don’t Miss Out. Stay Up-to-Date on UPDATES!
Register [here] to receive a once-weekly e-Notification of new postings.