The Daily Telegraph reported that “powerful painkillers used by thousands of arthritis sufferers almost double the chance of a patient suffering a heart attack and stroke”. The newspaper reports that the Cox-2 inhibitor Vioxx was withdrawn in 2004 after a study showed that it doubled the risk of heart attacks. This most recent study followed the people who had stopped taking the drug for a longer period after the trial ended (a year after), and confirms the findings of the previous study.
The article says that there is concern that other Cox-2 inhibitors may have a similar effect.
As the newspaper reports, this study has confirmed the findings of a previous study, and refers to a drug that is no longer available (Vioxx, also called rofecoxib).
As the article reports, there is some concern that this increase in risk may also apply to other Cox-2 inhibitors, and doctors will continue to weigh these against the benefits of these drugs in individual patients.
Where did the story come from?
Professor John Baron and colleagues from Dartmouth Medical School, and other universities and medical departments in the US, UK, and Spain carried out this research. The study was funded by Merck Research Laboratories (Merck is the company that manufactured Vioxx). The study was published in the peer-reviewed medical journal, The Lancet.
What kind of scientific study was this?
This study provided longer-term follow-up of a double-blind randomised controlled trial (known as the APPROVe trial), which compared rofecoxib (Vioxx) with placebo in people at high risk of developing polyps in the large bowel. The APPROVe trial recruited people from 108 centres across the world between 2000 and 2001. It was stopped in 2004 because of concern that rofecoxib increased the risk of cardiovascular events, and the drug was withdrawn from the market. The most recent publication reports on the follow-up of people who took part in the APPROVe for at least one year after the trial ended.
The APPROVe trial recruited people aged 40 years and over who had one or more growths (adenomas) in their large bowel removed in the previous 12 weeks, with no growths (polyps) remaining after the surgery. The trial did not include anyone with uncontrolled high blood pressure; angina (chest pain) or chronic heart failure that showed symptoms with minimal physical activity; previous heart attack, or surgery for coronary artery disease in the past year; or a stroke or mini-stroke in the past two years. Participants were initially randomly allocated to receive 25mg rofecoxib daily, 50mg rofecoxib daily, or placebo (inactive pill that looked identical to the rofecoxib pills) for three years. However, the 50mg rofecoxib group (25 people) was stopped shortly after the trial began.
The participants were scheduled to have their vital signs assessed when the study began, and on 11 occasions during and after treatment completed, with a colonoscopy given at the end of treatment and one year later. Any adverse effects that occurred during treatment or within 14 days after completion of treatment were recorded. Once the cardiovascular risks became clear, monitoring for adverse effects was extended to one year after treatment finished.
The study was stopped in 2004. About one year after the study was terminated, the researchers tried to contact all participants by telephone to ask whether they had experienced heart attack, stroke, cancer, or other related events since they had last been assessed. For those people who had died or were otherwise unable to respond, the researchers asked someone else to complete the interview for them. A panel of reviewers then looked at the answers and identified those who had experienced any of the following: death from a cardiovascular, haemorrhagic or unknown cause; non-fatal heart attack; or non-fatal stroke. The researchers compared the proportion of people who had experienced these cardiovascular events in the rofecoxib and placebo groups over the follow-up period. They also looked at whether this risk changed substantially over time.
What were the results of the study?
The APPROVe trial included 2,587 participants, of whom 1,857 completed the three years of treatment (72%). The researchers obtained data about cardiovascular events for at least a year after completing treatment (average [median] of about 1.5 years) for 84% of the participants, and managed to obtain death data for 95% of participants.
It was found that 59 people from the rofecoxib group experienced cardiovascular events during follow-up, compared with 34 people in the placebo group. This represented a 79% increase in the risk of an event over this period (hazard ratio 1.79, 95% confidence intervals 1.17 to 2.73). When each type of event was analysed separately, there was about double the risk of having a stroke or heart attack, although only the increase in heart attack was statistically significant. There was an increase of 31% in the risk of death, but this increase did not reach statistical significance (hazard ratio 1.31, 95% CI 0.80 to 2.15).
When the cumulative incidence or absolute risk of having one of the specified cardiovascular events was examined (the individual’s chance of having an event by the end of the time period), the researchers found that 1.74% more people in the rofecoxib group had experienced an event than in the placebo group (95% confidence interval 0.47% to 3.01%).
What interpretations did the researchers draw from these results?
The researchers concluded that using rofecoxib was associated with an increased rate of cardiovascular events such as heart attack and stroke, and that this increase lasts for at least a year after stopping treatment.
What does the NHS Knowledge Service make of this study?
This further follow-up from a well-conducted trial confirms its earlier findings of increased cardiovascular risk with rofecoxib, and supports the decision to withdraw the drug from the market.
The authors note some limitations:
As there were relatively few events, this study did not have enough power to reliably look at risks in specific subgroups of people, or to compare the frequency of events over time.
Although the number of people who participated in the follow-up was relatively high, loss of some people may have affected results.
The researchers did not collect information about use of other painkillers such as aspirin or non-steroidal anti-inflammatory drugs (NSAIDs) after the end of the treatment period, and therefore could not look at whether this affected results.
The study looked at the effect of a relatively long period of rofecoxib use, the effect of a shorter period of use is not clear.
The researchers suggest it is likely that the relative differences in risk (for example, the 79% increase in risk of the specified cardiovascular events) can be generalised to different populations. However, different populations have different underlying levels of risk of these events and the differences in the cumulative risk of an event may not be generalised to these.
Other studies have suggested that other Cox-2 inhibitors may carry similar risks, and doctors will continue to weigh these up against the benefits of these drugs in individual patients.
Analysis by Bazian
Edited by NHS Website
Links to the science
Lancet 2008; Oct 14 [Early online publication]
Cochrane Database Syst Rev 2005, Issue 1
Cochrane Database Syst Rev 2005, Issue 1