The withdrawal of the drug Vioxx exposed flaws in a system designed to protect the public. But the battle over who is to blame ignores the fact that there is no such thing as a safe drug, says Simon Frantz.
When fireworks ushered in the beginning of 2004, few could have predicted how much they would have continued to blaze throughout a rocky year for the pharmaceutical industry.
The year began with accusations that drug companies had stifled the release of negative data about antidepressant use in children1. Then came the publication of several books accusing firms of manipulating clinical trials and using deceptive marketing2.
The industry's public image reached an all-time low. It came as no surprise to read reports that Michael Moore is making a film, provisionally titled Sicko, investigating the pharmaceutical industry.
And then along came rofecoxib. This blockbuster painkiller, sold as Vioxx by New Jersey-based Merck, was withdrawn from shelves across the world after it was discovered to increase patients' risk of heart attack and stroke.
Warnings and accusations flew back and forth (see "A painkiller's progress"). Merck was accused of knowingly selling a dangerous drug. Other members from the same class as rofecoxib, called COX2 inhibitors, were also charged with having similar risks. And the agency responsible for the approval and safety of drugs in the United States, the Food and Drug Administration (FDA), found itself under fire for not spotting the risks sooner.
So should we worry about the safety of our drugs? The fact is that even the safest drug can produce an adverse effect in some people, and the task of regulatory agencies is to strike the right balance between the risks and benefits for each drug in the context of the condition that is being treated.
Of course, the situation is rarely black and white. COX2 inhibitors were developed as safer alternatives to older painkillers for arthritis. Older drugs, such as aspirin and non-steroidal anti-inflammatory drugs (NSAIDs), can cause severe gastrointestinal bleeding problems and were estimated to lead to more than 16,000 deaths a year in the United States alone.
COX2 inhibitors may avoid gastrointestinal irritations, but trials assessing whether rofecoxib and celecoxib (sold as Celebrex) could also reduce polyps in patients with colon cancer revealed cardiovascular risks. But, while one trial with celecoxib detected heart risks, another found none. And early indications from the colon cancer trials have hinted that rofecoxib and celecoxib may have a therapeutic effect.
The regulatory agencies now have to decide whether the COX2 inhibitors stay on the market, and if so how to ensure they are given to the right patients, that is, those at a low risk of cardiovascular disease but with a history of gastrointestinal problems. They may also have to decide whether to set the risk/benefit scales at different levels for patients with colon cancer and those with arthritis.
But striking such difficult balances requires us to identify the risks associated with drugs in a rapid and reliable manner, and the current system is sadly inadequate. Trials conducted in order to get a drug approved for sale (known as Phase 1, 2 and 3 trials) are usually carried out in hundreds or thousands of patients and over a short period of months.
If problems occur at a low prevalence of around 1% or 2%, or after18 months, as was found with rofecoxib, then the probability of detecting them in such clinical trials is negligible. The risk becomes apparent only once the drug is taken by millions, that is, once the drug is already being prescribed.
Once a drug is on the market, the FDA can recommend that companies carry out long-term studies in large populations to assess safety (known as Phase 4 trials). But the agency has no authority to enforce them, and fewer than half of these studies are actually completed.
Such a process for tracking drug safety cannot persist in light of rofecoxib's withdrawal.
Graham and the editors of leading medical journals have already called for the FDA to follow Europe's lead in reassessing drugs after five years on the market. The EU regulatory body for drugs, the European Medicine Agency, has the power to enforce safety studies and is able to take stock of new evidence concerning benefits and risks.
Some in the United States also advocate the creation of a new regulator, responsible for drug safety alone.
Whatever the solution, it will require finance. At present, drug companies pay the FDA a fee when they submit new drugs for approval, to help provide the agency with the resources to make timely decisions.
An extension of this user-fee model to aid safety studies might be seen by patients and politicians as too close a link between the FDA and the industry that it regulates, although studies show that the agency has not been compromised by this source of funding4. The alternative would be for Congress to provide the resources for FDA to carry out the tasks that it mandates.
What is important is that whatever changes are made, and however they are funded, they do not encourage knee-jerk reactions that prevent valuable treatments from reaching patients. Risking the lives of patients by delaying approval for access to a drug may prove just as detrimental to public health as the adverse effects of an approved drug.
For example, in December, the FDA halted a study looking at Alzheimer's disease that reported a link between the NSAID naproxen (sold as Aleve) and heart attack and stroke. That has been deemed premature, even by the most vociferous critics of the COX2 inhibitors.
Critics say that the 3% of the 2,400 elderly patients who suffered heart attacks or strokes during the study represents a relatively small number for that age group, and the data have not been assessed by cardiologists.
Vioxx has exposed flaws in the current US drug regulatory system, and it is likely to mark a turning point for the industry.
But we must all be wary of overreactions that suffocate drug companies and discourage the development of much-needed drugs. No drug will ever be totally safe, and safer versions or alternative treatments are often required. At present, the companies are the only place where these drugs can be developed.
- Check, E. Nature 431, 122 (2004).
- Marris, E. firstname.lastname@example.org published online doi:10.1038/news041220-7 (2004).
- Snyderman, R. Nature Med. 10, 1292–1293 (2004).
- Marris, E. Nature 432, 537 (2004).
- Carpenter, D. et al. Health Aff. W3, 618–624 (2003).
- Mukherjee, D. et al. J. Am. Med. Assoc. 286, 954–959 (2001).
- Topol, E. J. N. Engl. J. Med. 351, 1707–1709 (2004).
- FitzGerald, G. A. N. Engl. J. Med. 351, 1709–1711 (2004).
- Editorial Lancet 364, 1287–1288 (2004).
- Fontanarosa, P. B. et al. JAMA 292, 2647–2650 (2004).