The company contends, contrary to the authors, that it shared the findings of the 2006 study with the steering committee.
But in their article for NEJM, the authors focused mainly on the fact that Avandia had performed the best — that is, it was able to control blood sugar for the longest period.
As for those hints of cardiovascular risks that Nissen, the Cleveland Clinic cardiologist, had seen in the data? The authors pointed to no such trouble.
The risks of “cardiovascular events” associated with Avandia, the article said, were “similar” to those affecting patients who had taken metformin, one of the most prescribed diabetes drugs in the world.
The signal for doctors was clear: Don’t worry.
‘It didn’t look right’
But Nissen could not help but worry.
He had seen other data that suggested to him that Avandia could cause heart trouble. Another trial sponsored by the company, known as DREAM, had shown a slight trend, he thought, but the number of patients was too small to be considered statistically significant. Then, in the ADOPT trial results, he saw the same clues repeated, even if they were not remarked upon by authors of the article.
The trend in the data was suggestive, Nissen thought, though again not statistically significant. In the two groups of patients that had taken the commonly used drugs, there had been 14 and 20 serious heart attacks. The Avandia group had 24.
And there was another clue: The measures of bad cholesterol were notably higher in the Avandia group.
“The trend was in the wrong direction, and that’s what sent me off,” Nissen said. “It didn’t look right.”
To see whether his suspicions were warranted, Nissen, with colleague Kathy Wolski, set out to assemble the data from every trial of Avandia that they could find. The more data they had, the more likely they could accurately gauge the risks. The drugmaker refused Nissen’s requests for data, but because of litigation brought by Eliot Spitzer, then New York’s attorney general, the company had been forced to make some of it public. In all, he discovered the summaries of 42 trials — 35 of them unpublished. Most of them had been sponsored by Glaxo.
After analysis, the results were stark: Avandia raised the risks of heart attack by 43 percent and of death from heart problems by 64 percent.
Those findings would stand up. But the reach of the pharmaceutical companies to influence the science would create three more years of uncertainty.
Glaxo ready to respond
Nissen and Wolski submitted their findings to NEJM on May 2, 2007.
Normally, an article takes several months to get published, but Drazen put it on a fast track, publishing it on the NEJM Web site 19 days later, on May 21.
“This was a big surprise, and I wanted to get it out there,” Drazen said. “If it was right, thousands of people were having heart attacks because of this drug.”
Glaxo was surprisingly well prepared to respond.
How? What was not known until later is that the NEJM paper had been leaked to the company.
As part of the process of peer review, the paper had been sent to Steven M. Haffner, a Glaxo ally and a University of Texas professor who had helped conduct the ADOPT trial.
Without telling Drazen or Nissen, Haffner faxed a copy of the confidential unpublished paper to the company, according to documents released by the Senate.
More than 40 company executives would learn of its contents. They prepared a meticulous response to its publication that suggested that Nissen’s results were plain wrong.
“GSK strongly disagrees with the conclusions reached in the NEJM article, which are based on incomplete evidence and a methodology that the author admits has significant limitations,” the news release said.
But internally, scientists and statisticians at Glaxo largely agreed with Nissen’s calculations, the company e-mails released by the Senate show.
“To a great extent the numbers are the numbers, the Cleveland analysis is very similar to our own,” one of them reported via e-mail.
An ‘underpowered’ study
The company would also launch one other strategic counter to Nissen’s paper: They would publish the results of another, separate trial of Avandia that they were conducting, known as the RECORD trial.
One of the reasons that the Glaxo executives could be confident that the RECORD trial would show no danger is that the trial did not have enough patients enrolled to judge the drug’s heart-attack risks, as Glaxo scientists believed, according to the Senate report. It was, in the scientific jargon, “ underpowered.”
The Glaxo executives faced one big problem, however. The RECORD trial was two years from completion. Publishing the interim results of a trial is very unusual because it tips off patients and doctors in a way that could bias results.
Like the ADOPT trial, the RECORD trial was funded by Glaxo, which had in turn hired a steering committee of prestigious academics to lead it.
The researchers in the RECORD trial had many financial ties to the company, too. Of the eight authors of the RECORD trial report, one was a Glaxo employee. The other seven reported having received consulting fees or other support from Glaxo. One, Philip D. Home, reported donating such money to medical institutions.
While the academics were nominally in charge of the trial, it would be the company, not the academics, who would first decide to publish the interim results.
The day before the academics were to meet, Ronald L. Krall, Glaxo’s chief medical officer, told another employee in an e-mail, “We’ve decided we will disclose the results.”
If the steering committee objected, the executives were prepared to tell them that a “decision has been made — live with it,” according to an e-mail from Glaxo executive Trevor G. Gibbs.
When the academics were convened the next day, the group went along with the decision to publish interim results. They decided on their own, the steering committee’s chairman, Home, said via e-mail. He said they feared that Nissen’s warning could scare patients and doctors out of the trial, and they needed to reassure them.
“We had no choice but to publish,” he said. “The decision was inevitable if regrettable.”
In their first submission to NEJM, Home and his co-authors indicated that the RECORD trial results had undermined Nissen’s warning, according to a letter from the journal to the authors.
But NEJM’s peer reviewers noted that the data did not support that conclusion, and they demanded changes.
As result, when the article appeared in July 2007, it did not say anything definitive about Avandia and certain heart problems.
The paper said that the results of the RECORD trial were “inconclusive” as to whether the drug raised the risk of cardiovascular problems and that the data were “insufficient” to determine whether the drug raised the heart-attack risk.
Yet the language in the article, though equivocating, might still have helped Avandia sales by making the issue look like a muddle.
“What it did was it falsely reassured practitioners and patients that [Avandia] might be safe when in fact it wasn’t,” Nissen said. “They got three more years out of it.”
It was not until 2010 that Nissen was largely vindicated. An FDA reviewer indicated that the RECORD trial had been poorly designed and suggested that investigators had improperly missed heart problems suffered by Avandia patients.
In September 2010, the FDA announced major restrictions on the use of Avandia. On the same day, European regulators ordered it off the market.
In the wake of controversies arising around Vioxx, Avandia and Celebrex, many in the medical world have sought ways to ensure that drug research is free of commercial bias.
One of the leading proposals would be to compel drug companies to release all of the data from trials of drugs that are on the market.
Over the summer, the European Medicines Agency — the continent’s counterpart to the FDA — said it will move toward requiring the release of all such data. Glaxo, too, has said it is preparing for such a release, though other companies have yet to follow suit.
“Since 2004, we have posted summaries of all our clinical trial results on our Web site for the world to see,” Glaxo said in a statement. “All of these actions speak to the degree of commitment we have to be open with our research so there can be more understanding, and hopefully credibility, in what we are doing.”
Such transparency about industry-sponsored trials would not eliminate the ability of companies to avoid unflattering studies, or to hire like-minded researchers, or to design research that gives only positive views of their products.
But if such measures are carried out across the industry — and there is no sign at this point that they will be — independent researchers could analyze the data from trials and come to their own conclusions.
Many believe drug companies should feel obliged to share such information.
“If you have the privilege of selling a drug, in return should come the responsibility to share everything you know about the drug,” said Harlan Krumholz, a professor of medicine at Yale and a leading advocate of data access. “This is not about doing gotcha with industry. It’s about how to restore trust.”