As almost everyone knows, Merck & Company withdrew Vioxx from the market October 2004 after two studies showed that it increased the risk of heart attacks.Vioxx first became available in 1999. So it took five years to determine that it was dangerous to some patients.Was this long delay due to negligence by Merck or malfeasance at the US Food and Drug Administration (FDA), or was it something else? There are lawsuits beginning and there will be congressional hearings. You’ll hear a lot about this issue in the next several years. Here are some things to think about.
It is easy to detect a side effect if the side effect is common. Suppose a drug causes a side effect in one percent of people who take it, and that 1/10th of one percent of people who don’t take the drug also get the “side effect,” perhaps from another drug or maybe because the symptom is common in people generally. Believe it or not, it would take 3,200 people participating in a study to prove that the side effect was caused by the drug. That’s a lot of people.
Could we have detected an increase in heart attack rates caused by Vioxx in the NDB with 15,000 people completing questionnaires? Well, first we need to know the normal rate of heart attacks. In the NDB, about one percent of people report that they had a heart attack during one of our 6-month surveys, and about two percent of people used Vioxx each 6 months. That’s not a lot of heart attacks and not a lot of Vioxx use. When we analyzed NDB data, we found a tendency to an increase in heart attacks by Vioxx users. On average, we saw an increased risk of about 32 percent. But, because so few people were taking Vioxx in the NDB and the number of heart attacks was so small, we could not make a statistically valid conclusion. All we could have said is that the increase of risk could be as great as 223 percent but it also could be as low as –28 percent. That is, it was even possible that Vioxx might protect against heart attacks. What it all comes down to is that when an event is rare, like heart attacks, and the drug only causes it in a few people, it is very hard to prove the cause and effect link.
The end for Vioxx was the Kaiser-Permanente HMO study that found a 3-times increased risk for heart attacks among Vioxx users. The Kaiser study included 1.4 million people and 8,199 heart attacks. By contrast, among the 15,000 NDB participants there were only 560 heart attacks. By the way, we were able to detect that Vioxx caused hypertension and swelling (factors that might lead to heart attack) and we reported that in 2002. The NDB research paper was published this year.
Although it seems clear now that Vioxx caused heart attacks, it wasn’t clear when Vioxx was first released, although a few suspected a problem. It took many people taking the drug for a number of years before it became clear that the drug was at fault. In the flurry of lawsuits and investigations that will occur, remember that this knowledge came slowly.Was the drug company at fault, or maybe the FDA? Stay tuned, but remember that a problem like this is not always a sign of malfeasance. Perhaps researchers should change the way they identify serious side effects from clinical trials that have a small number of patients to large observational data banks, such as the NDB and Kaiser-Permanente.