The past few months have added to confusion among both physicians and the public regarding the use of several highly effective medications that are taken by several million people, including many elderly patients who may be hypertensive or have heart disease. Conflicting results from major clinical trials have been highlighted on television and in newspapers and magazines. Physicians' telephones have been ringing off the hook. The Internet is swamped and numerous experts have surfaced with “I-told-you-so” comments. The antiarthritic medications and the cyclo-oxygenase-2 (COX2) inhibitors are under attack. Are patients being given appropriate advice? Physicians are concerned about prescribing them and patients have become increasingly worried about their possible adverse effects. The debate is escalating to a point where it may be based more on emotion than science.

All medications have some side effects. In general, the FDA has done a good job of protecting the public from drugs that may induce harmful adverse effects, despite the outcry of several consumer advocates. Recent criticism of this agency for its lack of surveillance of drugs has centered on post-approval evaluations. At present, medications are approved after efficacy and safety studies in several thousand people over a relatively short time period of time, typically 3-6 months. The COX2 issue has led to renewed demands that drugs should not be released unless safety data are available on many more patients over a longer time period. This is perhaps desirable but not practical.

The FDA process of drug testing in small numbers of people followed by studies in larger groups and finally in 12–24 week trials in several thousand people has largely been successful in weeding out many drugs for safety or efficacy reasons. In this process, however, there is an occasional adverse effect that does not become evident in these shortterm, relatively small studies.

A good example of this was a diuretic, ticrynafen, which was approved by the FDA many years ago. In addition to lowering blood pressure (BP), this agent also reduced uric acid levels—a possible benefit in the management of hypertensive patients and patients with heart failure. During the course of the pre-approval studies, there were a few cases of hepatitis reported, mostly from the West Coast. The cases were not considered to be related to the drug since there were many cases of hepatitis in this area at the time. The drug was approved after studies of approximately 3000-4000 people. Within a relatively short period of time, however, when several hundred thousand patients were given this agent, it became obvious that there was a relationship of the drug to the occurrence of hepatitis. This relationship had only became obvious when large numbers of people were exposed to its use. The drug manufacturer withdrew the drug from the market. The small percentage of cases of hepatitis noted during the initial clinical trials were originally believed to be a chance finding but they turned out to be of significance. Some might argue that the approval of this unique medication should have been delayed until many more thousands of people were studied?

There are numerous examples of FDA surveillance preventing the approval of a drug that lacked efficacy or had serious adverse effects. The FDA's responsible actions in the Laetrile or Thalidomide approval process, despite pressure from numerous consumer and commercial groups, are examples of how the agency has worked. While not always as effective as it should be in an ideal world, the FDA has, on balance, done a good job of protecting the public.

How does all this relate to hypertensive patients, to the withdrawal of Vioxx (rofecoxib; Merck & Co., White House Station, NJ) from the market and the pressures to remove Celebrex (celecoxib; Pfizer Inc., New York, NY) and Bextra (valdecoxib; Pfizer Inc., New York, NY) because of data linking these medications to heart disease and stroke risk? It might be helpful to review an example of what almost happened to a useful class of medications. In the 1960s and 70s there were numerous physicians, members of the press, and consumer groups clamoring for FDA action on thiazide diuretics because of their possible metabolic effects on heart disease risk. A headline in a major newspaper pronounced that “thiazides increase heart attack risk.” This was based on a recently completed trial. Fortunately, the FDA, upon reviewing the data, did not yield to the loud pronouncements; the safety and effectiveness of these agents have since been affirmed and reaffirmed over the years. Their use may have resulted in some chemical changes but benefit clearly has outweighed the risk. What a mistake it would have been to have placed warning labels on these medications or withdrawn them from the market.

But the world has changed over the past 20–30 years; pressures are greater to do something the minute a hint of a problem surfaces. Litigation potential lurks in every corner—ads on billboards, on the radio, on television, and in magazines and newspapers advising that “if you have ever been on a medication and you have ever had sleepless nights, anxiety, arthritis, chest pain, dizziness, blurred vision, or in any way felt unhappy, please call us for a free consultation; you may be entitled to monetary damages.” The recent COX2 publicity has spawned another new group of “liability chasers.” It is quite clear that not all suits against the pharmaceutical industry are frivolous or without merit, but people listening to what has happened since the Vioxx withdrawal must be disturbed—and if they are taking one of the other antiarthritis medications, must either be frightened or, at least, concerned. There is nothing that will increase cost or limit the search for new, effective medications more than the threat of billion-dollar settlements if an unforeseen side effect occurs after a drug has been released.

Were the cardiovascular (CV) effects of Vioxx clearly established, or only suspected, before the drug was approved? Following approval, this medication was heavily promoted with direct-to-consumer advertising which, by the way, should be disallowed, and which probably resulted in too many people taking this and other COX2 drugs instead of aspirin, Tylenol (acetaminophen; McNeil Consumer & Specialty Pharmaceuticals, Fort Washington, PA), or Advil (ibuprofen; Wyeth Consumer Healthcare, Madison, NJ) which, in many cases, may have been just as effective. A large number of people, however, did require the use of Vioxx or other medications in this class and did report good results. Many had been unable to tolerate some of the NSAIDs because of gastrointestinal problems, by history or as a possibility. Quality of life was improved in many people. The manufacturer had had a clue that the incidence of heart attacks or strokes might have been higher on Vioxx, based on some comparative preliminary data; but these were not definitive. When definitive data were uncovered in a long-term trial, the manufacturer withdrew the medication from the market. The “I-told-you-so” physicians from preeminent institutions surfaced—they knew this all along—the drug never should have been released. Conflicting opinions are now emerging about when research on a new drug should be stopped, or a drug withdrawn, from the market. Should this be after a few adverse reactions are reported or only after an outcome study shows a significant deleterious effect on outcome? If the former were true, most medications, including aspirin, probably would never have made it to market and diuretics would have been drugs of the past. Benefit to risk must be a major criterion for acceptance since there is a risk to someone from almost every medication.

Following the Vioxx story, a study reported that Bextra, another COX2 inhibitor, also increased strokes and CV disease. What the stories mentioned, but did not emphasize, was that these episodes had occurred in patients following bypass surgery. Many physicians reacted and stopped prescribing Bextra. Next came the Celebrex headlines. In one study, there was an increased incidence of CV disease. What the stories noted, but did not emphasize, was that the dosage that caused the trouble was 800 mg/d—doses rarely prescribed. At the usual dosage of 200 mg/d, there did not appear to be any adverse effects. But the message was out there—these drugs are dangerous, especially in patients at high risk for heart disease who took them for long periods of time.

The news media like to report in superlatives—“the risk of X compared to Y was double or triple.” These impress the reader but represent a relative, not an absolute, risk difference. In the case of the COX2 inhibitor, Celebrex for example, in a National Institutes of Health (NIH)-sponsored trial of more than 2000 patients over a period of 33 months, there was a relative 3.4-fold increase in cardiac events in patients on 800 mg/d compared with placebo. These are disturbing numbers, but the absolute increase in risk over this time period in these patients was represented by 23 events in the 800-mg group compared with seven in the placebo group—quite a difference in reporting (16 more events over 2-plus years in a large number of patients, rather than more than a 300% increase in risk). Now if you were one of the 16 people, it would not matter how many others were adversely affected. Here the risk/benefit ratio comes into play. If the COX2 inhibitors are more effective in some individuals than other pain relievers in reducing arthritis pain, perhaps the small risk is worth it. There is some evidence that smaller doses of 100 or 200 mg per day of Celebrex, the usual dose given in a patient with arthritis, do not pose a CV risk. It is true that there may be some elevation of BP with the COX2 inhibitors and some adverse effect on platelet aggregation, especially at higher doses, but BP changes may also be noted with the nonsteroidal anti-arthritis agents. The COX2 inhibitors probably suppress the vascular production of prostacyclin, a vasodilator and antithrombotic substance, without affecting the synthesis of thromboxane, which may promote thrombosis. The concurrent use of platelet-inhibiting aspirin with a COX2 inhibitor may negate at least some of this latter effect. In a hypertensive patient who needs one of these agents for pain relief, BPs should be monitored at home. The rise in pressure is not great, but a medication adjustment might be necessary.

The data about Vioxx do suggest that the use of the drug increases CV event risk, but the decision to continue to use it should have been made on the basis of its effectiveness and absolute, not relative, risk in people who are unresponsive or cannot tolerate other medications, not as a result of headlines and the pressure of litigation. There is a question raised that the hype about fewer gastrointestinal adverse effects with the COX2 inhibitors may be exaggerated; that Advil, Aleve (naproxen; Bayer, Pittsburgh, PA), and other NSAIDs are just as effective and less expensive. In many cases, this may be true but, as noted, there are patients who cannot tolerate these agents or in whom they do not work. There are some data to suggest that even naproxen may increase myocardial infarction and stroke. The available data are not definitive.

Additional information with further direct comparisons must be gathered. In the meantime, the rhetoric and hype should be toned down and the true risk of these agents evaluated by physicians and patients alike. If someone does not respond to Tylenol, aspirin, or one of the NSAIDs, the use of a COX2 inhibitor in recommended dosages should still be considered, especially if the patient has a history of gastrointestinal disorders. We must increasingly be vigilant about drug side effects and reactions, but should not let the headlines and threats of legal action prevent us from practicing good medicine. The FDA should be encouraged to increase its drug surveillance capabilities, but should not be forced to make decisions that are not based on carefully controlled scientific observations—benefit-to-risk still should be an overriding consideration.


  1. Top of page
  2. Addendum:

This editorial was written prior to the recommendations of the US Food and Drug Administration's advisory panel and publication of study results in the New England Journal of Medicine. The panel recommendations not to remove the COX2 inhibitors from the market appear to be reasonable.