With the support of the National Research Programme (Musculoskeletal Health - Chronic Pain) , researchers at the University of Berne analysed the data in publicly available studies on the anti-inflammatory drug Vioxx.
The investigators came to the conclusion that the cardiovascular risks of the drug could have been demonstrated by the end of the year 2000 and that the risk is unrelated to dose and duration of use. The Lancet published these results in its on-line edition of 5th November 2004.
On 30th September, the pharmaceutical company Merck & Co. withdrew its anti-inflammatory drug Vioxx (rofecoxib), from the market throughout the world after a study had shown an increased risk of cardiovascular diseases linked to the use of this COX-2 inhibitor. The drug has been suspected for some time to increase the risk of heart attacks. This withdrawal raises the question as to whether studies could have revealed this risk earlier. A team led by Peter Jüni and Matthias Egger, of the Department of Social and Preventive Medicine at the University of Berne has now carried out a meta-analysis to examine the available study data as part of a National Research Programme project on Musculoskeletal Health - Chronic Pain (NRP 53). The results of this study were published in the on-line edition of 5th November of the renowned medical journal The Lancet.
The study is based on publicly available trial data from the American drug licensing authority (FDA) and on an extensive literature search. The meta-analysis included 18 randomised studies and 11 observational studies. Our analysis confirms the suspicion that the data proving an increased risk of cardiovascular disease linked to Vioxx were available by the end of 2000, says Peter Jüni. By the end of 2000, 52 myocardial infarctions had occurred among 20742 patients; 41 in patients assigned to Vioxx and 11 in patients not treated with Vioxx – a statistically significant difference.
In particular, the researchers at the University of Berne criticise the interpretation of the VIGOR study in 2000. This study compared the action and side effects of rofecoxib (Vioxx) with another analgesic, naproxen (e.g. Apranax or Aleve). Indisputable differences in the risk of a heart attack were apparent in this study. The investigators of the VIGOR study, however, attributed this difference to the supposedly protective properties of naproxen and not to any increase in the risk o a heart attack with rofecoxib. But we found no evidence to support this interpretation, says Matthias Egger. In the VIGOR study, an increased risk of heart attacks due to rofecoxib should have been considered.
In its press release in relation to the withdrawal of Vioxx, Merck stated that on the basis of the alternative possible treatments and the questions raised by these data, the company has come to the conclusion that a voluntary withdrawal from the market is best. Although this was announced in September 2004, the company could and should have made the same statement several years back, when the data that we analysed were already available, says Matthias Egger. However, at that time Merck issued a press release entitled: Merck again confirms the good cardiovascular safety of Vioxx.
The meta-analysis also questions other statements by Merck & Co. The product was withdrawn with the information that the risk only affected patients who had taken Vioxx for 18 months or longer. We could not confirm this, says Jüni. The data show that an increased risk of a heart attack is present within a few months of use, and this risk is independent of the dose taken.
It was also striking that studies in which the data on adverse events were evaluated by an independent committee showed a much clearer result with regard to the cardiovascular side effects. It could be, according to Egger, that without independent evaluation of the data, the assessment of adverse events is biased so that the risks of a drug appear smaller. We therefore recommend that all studies be carried out with independent external data evaluation.
The analysis of the Vioxx data also shows that the licensing authorities should critically review their procedures. New data and knowledge concerning a drug should be regularly included in the documentation and be continuously analysed. As the example of Vioxx shows, this is not always ensured at present, says Jüni. Only with continuous monitoring can we make sure that the public is protected against unnecessary side effects of medicines, adds Egger. The reasons why manufacturers and licensing authorities did not continuously evaluate and reassess new information and data in this case should be investigated by an independent panel of experts.
The study now published is part of a project being performed within the framework of the National Research Programme Musculoskeletal Health - Chronic Pain NRP 53. The project is systematically comparing the effects and side effects of various well-known and new drugs for arthritis pain. The results of this project should be available in mid-2006.