Manipulating research findings

Drug manufacturers use a variety of tactics to get positive results from their clinical trials. For example, they can compare their new drug with an existing drug at an inadequate dose or choose patients who are most likely to get better on the treatment. If they reach a point where results are looking good, they can stop the trial early. Or, when they see that the results of clinical trials are unflattering, they can choose not to publish them.

As to the last point—not publishing unflattering data—Dr. Ben Goldacre, in the UK, describes a particular instance in his medical practice. He had researched the medical literature looking for a drug that would help his patient with depression. Because he’d had no luck with a variety of drugs, he began looking at the literature for reboxetine. “I’d read the trial data before I wrote the prescription, and found only well-designed, fair tests, with overwhelmingly positive results. It looked like a safe and effective treatment.” Later, through a long process of investigation, a group of researchers finally brought together all the trials that had ever been conducted on reboxetine. “When all this trial data was put together it produced a shocking picture,” he recounts. “Seven trials had been conducted comparing reboxetine against a placebo. Only one, conducted in 254 patients, had a neat positive result, and that one was published in an academic journal for doctors and researchers to read” It turns out that six more trials had been conducted on almost ten times as many patients. All of these trials showed that reboxetine was no better than a sugar pill. None of these six trials was published. “I had no idea they existed.” Reboxetine is still on the market.

On September 30, 2004, Merck pulled its arthritis drug, Vioxx, from the market, following increasingly alarming data that the drug increased the risk of heart attacks and strokes. This was five years after discovering, in 1999, that people who took Vioxx were 2.4 times more likely to experience cardiovascular problems than those taking naproxen. Neither the company’s marketing materials nor articles published in the New England Journal of Medicine indicated the magnitude of this problem. In other words, Merck had withheld this data and also misrepresented safety information. By the time the drug was pulled, an estimated two million people were taking it, and yearly sales amounted to $2.5 billion dollars. Merck was finally moved to withdraw the drug when, in a later trial to test the effect of Vioxx on polyps, researchers found that people taking Vioxx were developing almost twice as many heart attacks, strokes, and blood clots as people taking placebos. (These results were in fact similar to those in its 1999 study, the results of which were withheld.) One of the FDA’s scientists estimated that Vioxx had been associated with more than 27,000 heart attacks or deaths. (Vioxx, by the way, had not been shown to alleviate pain any better than the older medicines.)

These are just a couple of examples. Given the lack of reliable information, how can doctors (or you) trust the drug companies?

Next week: Passing off “me too” drugs as new.

For an introduction to this blog, see I Just Say No; for a list of blog topics, click the Topics tab.