Nottingham, UK - A new British case-control study has found an increased risk of MI with two commonly used nonselective NSAIDs, ibuprofen and diclofenac, as well as the COX-2 inhibitor that was withdrawn last year, rofecoxib (Vioxx, Merck & Co). The researchers also found similar odds ratios for other NSAIDs, although the results did not reach as high a significance level.

"We think that enough concerns exist to warrant reconsideration of the cardiovascular safety of all NSAIDs," say Dr Julia Hippisley-Cox (University of Nottingham, UK) and colleagues in their paper in the June 11, 2005 issue of BMJ [1].

In an accompanying editorial, Dr Peter Juni (University of Berne, Switzerland) and others say observational studies such as this cannot test definitively whether there are small to moderate risks or benefits of a class of drugs and that meta-analyses or ultimately large-scale clinical trials are required [2].

"Our paper is new in that it adds information about the [MI] risk with more traditional NSAIDs and shows that this effect is not specific to COX-2 inhibitors," Hippisley-Cox told heartwire. In fact, the new findings virtually mirror those of a similar trial reported last month in the Archives of Internal Medicine [3], about which the lead author commented, "We need to be aware of the potential risk of all NSAIDs."

As a clinician I find it frightening that we have this massive jump from giving a drug to thousands of people for a few months in a clinical trial to giving it to millions of people for long periods of time.

Hippisley-Cox agrees with the editorial accompanying her study and believes that a first move in the right direction would be to conduct meta-analyses of all data—whether published or unpublished—on each drug individually to assess the risk, "and in some cases this may provide us with reassurance." However, if a meta-analyses, with its inherent biases, does not reassure, then a large-scale clinical trial will be required to definitively assess the risk of MI with each individual NSAID or COX-2 inhibitor, she says.

She admits that it is difficult to say who would actually conduct such trials, but she believes that the impetus for them should come from regulatory agencies. "As a clinician I find it frightening that we have this massive jump from giving a drug to thousands of people for a few months in a clinical trial to giving it to millions of people for long periods of time."

In their study, Hippisley-Cox et al identified 9218 patients aged 25 to 100 across England, Scotland, and Wales via the QRESEARCH database who suffered an MI for the first time over a four-year period. They tracked whether patients had been prescribed NSAIDs or a COX-2 inhibitor and compared them with 86 349 matched controls. The findings were adjusted for several risk factors.

For those prescribed NSAIDs or COX-2 inhibitors in the three months just before the MI, the risk increased compared with those who had not taken these drugs in the previous three years. For ibuprofen, the risk increased by almost a quarter (24%), and for diclofenac it rose by over a half (55%).

The most significant findings were for ibuprofen, diclofenac, and rofecoxib, say the authors. In terms of "numbers needed to harm" in the 65-and-over age group, for those taking diclofenac, one extra patient for every 521 patients was likely to suffer a first-time MI. For rofecoxib, the figure was one patient for every 695 patients; and for ibuprofen, one patient for every 1005 patients was at risk.

"Given the high prevalence of the use of these drugs in elderly people and the increased risk of myocardial infarction with age, the relatively large number of patients needed to harm could have considerable implications for public health," they comment.

They note that they found no evidence of a cardioprotective effect of naproxen, and in fact there was evidence of increased risk of MI with this drug, too. They also found no clinically important interactions between any NSAID and either aspirin or coronary heart disease.

In the editorial, Juni et al say that unbiased data on serious adverse events from clinical trials should be made available to independent researchers and the public and analyzed in a timely fashion. "Indeed, in the case of rofecoxib, cumulative meta-analysis showed that an increased risk of MI was evident from 2000 onward. Similar analyses are now required for the other COX-2 inhibitors."

They note that the US FDA and other licensing authorities are an important source of relevant data, but that, unfortunately, reports from such agencies "are not as useful as they could be." For example, only 16 out of at least 27 trials of celecoxib that were performed to 2002 were included in relevant reports, and in the case of valdecoxib "we found that many pages and paragraphs had been deleted because they contained 'trade secret and/or confidential information that is not disclosable.' "

"Surely, the protection of the public's health justifies full access to the safety data submitted by industry to the FDA and other licensing authorities and mandates transparent reporting harms in accordance with international guidelines," they conclude.