A Top Cardiologist Says A Diet Drug Maker Misled Patients And Investors

, , ,

Update 7:15: Click here for Orexigen’s comments on this situation.

A top cardiologist is accusing a drug maker with whom he worked closely of making misleading statements about its weight loss medicine and refusing to release data that would undermine its previous claims.

The pill, Contrave, was prescribed 117,000 times in its first three months on the market, a better launch than any obesity drug in the past decade. Things seemed to get even better in March, when Contrave’s maker, Orexigen Therapeutics, released data that claimed that the drug not only helped patients lose weight but also prevented heart attacks, strokes and deaths caused by cardiovascular disease by a stunning 41%, a statistically significant amount.

But Steven Nissen, chair of cardiology at the Cleveland Clinic and the lead researcher on the 9,000-patient study, immediately protested that the result was premature and untrustworthy. So did John Jenkins, the head of the Office of New Drugs at the Food and Drug Administration. Now, via a Cleveland Clinic press release and an interview with Forbes, Nissen, well known for his role in raising alarms about the safety of drugs like Merck’s Vioxx and GlaxoSmithKline’s Avandia, has released the data to back up that contention.

The supposed benefit almost disappeared as time passed. As the number of heart attacks, strokes, and cardiovascular deaths increased from 94 to 192, the 41% difference between those getting Contrave and those getting placebo shrank to just 12% and was no longer statistically significant. More evidence it may be meaningless: patients given Contrave were more likely to have chest pain and more likely to die from causes other than heart disease. Many patients who supposedly had a benefit had actually stopped taking the drug.

“Patients were misled, investors were misled,” says Nissen. “It is so critically important that investors and other people understand why early data in a trial are unreliable or unstable.”

Worse, Orexigen has sought to block the release of the new data. On March 26, the study’s executive committee voted unanimously that the study should be stopped and the data publicized. But for six weeks Orexigen repeatedly refused to approve a press release, Nissen says. Today, Orexigen and its marketing partner, Takeda, issued a release saying only that the study was stopped. The Cleveland Clinic’s release, unauthorized by the companies and filled with data, followed minutes later.

Orexigen did not immediately return multiple phone calls and emails requesting comment.

Contrave is a combination of two old drugs: bupropion, the active ingredient in the antidepressant Wellbutrin, and naltrexone, which is used to treat alcohol addiction. In one study, 42% of patients treated with Contrave lost at least 5% of their body weight, compared with 17% of patients treated with a placebo. But for many patients, the drug does not work; the average weight loss is just 4% better than a placebo.

Because bupropion boosts blood pressure, the FDA wanted to be sure that Contrave did not increase the risk of heart attacks or strokes. So it required Orexigen to conduct the study Nissen was to run, called LIGHT. Nissen, and his group at the Cleveland Clinic were brought on in part because of his reputation for drug safety – if he proved the drug was safe, other doctors would be more likely to believe it.

The trial was designed to prove that Contrave caused no more than a 40% increase in heart attacks, strokes, and cardiovascular deaths (euphemistically called ‘cardiovascular events’ by doctors). It was expected to take until 2017.

But the FDA allowed Contrave onto the market sooner, using a regulatory mechanism that had originally been proposed by Nissen himself. Orexigen could peek at the data 25% of the way through and use that data to file with the FDA if it showed that Contrave was doing no worse than doubling the risk of a cardiovascular event. The deal, however, came with strings.

Nissen says that four Orexigen executives signed a 14-page document that said that the company would take extreme steps to make sure that only a carefully selected regulatory SWAT team would see the results. Specific areas of the company’s offices would be locked up to keep the results secret, and the team would not use Orexigen’s corporate email system to share them.

Something very different happened. According to FDA documents, more than 100 people saw the results, including top Orexigen officials, outside consultants, and lawyers. And, in a complete surprise, the results actually seemed to show that Contrave prevented cardiovascular events. The FDA chastised the company for allowing the data to leak, and said that the study could no longer be used to satisfy Orexigen’s regulatory requirement – the company would have to do a second study at an estimated cost of more than $100 million.

Instead of keeping the information quiet, Orexigen executives decided that they had a fiduciary duty to file for a patent based on the information. Nissen found out about the patent from Takeda executives, who had taken over the day-to-day management of the clinical trial from Orexigen. “I was livid,” he says. “I almost couldn’t believe it.”

When the patent was approved on March 3, Orexigen filed a document with the Securities and Exchange Commission that included the total number of cardiovascular events, and the preliminary estimate on the reduction in risk to patients. The company’s shares spiked 30%.

Even then, Nissen said, there were problems in the data that Orexigen was not disclosing to doctors, patients or investors. For one thing, it didn’t disclose that four more patients, a total of 29, had chest pain while on Contrave. But the more important issues had to do with the way clinical trials are analyzed.

Usually, researchers use a very tough standard, called intent-to-treat, for testing drug effectiveness. If a patient is randomly assigned to get a medicine, they are counted in the clinical trial even if they decide to stop taking it. Normally, this makes for a tougher test. Patients who stop taking a heart drug usually don’t get any benefit from it.

But in this case, probably because of the play of random chance, something different happened. In patients who stayed on Contrave or the placebo, there was a 30% reduction in the risk of heart attacks, strokes, or cardiovascular deaths. But in those who stopped taking the drug, the benefit was a much greater 53%. In other words, the results were better if patients were assigned to take Contrave, but then didn’t! The differences were dramatic. In the off-treatment group, there were 12 cardiovascular deaths in the placebo group and 1 in the Contrave group.

This isn’t surprising. When researchers want to take an early look at a clinical trial, the rules of statistical significance actually change. Stopping the study actually requires a far higher level of certainty the result is true, because the more often you look, the more likely it is that you’ll find the exact result you’re hoping for by chance.

Nissen and the FDA’s Jenkins warned that the result was probably spurious.  “Step back and think for a second,” Jenkins told me then.  “We required this study because we’re concerned that Contrave may cause adverse cardiovascular events because of its effect on blood pressure and heart rate. So the likelihood that that drug is going to have an early benefit is highly unlikely.”

The FDA initially worried that the results of the study had leaked to 100 people, and decided not to accept it. But now they had been released to the entire world. Nissen says he had hoped to continue to the end of the trial anyway, but that his executive committee faced another problem: patients were having fewer heart attacks and strokes than expected. In order for the trial to complete, doctors would have to enroll another 5,000 patients. It was futile.

The executive committee voted unanimously to recommend that the trial be stopped and that the data through to the halfway mark be released immediately. They told Orexigen, and were shown the more mature data. Not only did the cardiovascular benefit seem to vanish, but there was a difference in deaths from other causes: 26 deaths on Contrave versus 17 on placebo. The difference was not statistically significant.

Nissen started working with Takeda to craft a simple press release that would contain the more mature data. Each time Takeda would approve the release Orexigen would reject it. The process went on for six weeks, and even continued as Orexigen conducted an earnings call during which executives said that no decision about the study had been made.

Finally, he decided that the Cleveland Clinic would simply issue its own release, without the companies’ approval, after it was announced that the trial had stopped. This morning, that happened.

Also on Forbes:

Source: Forbes