TRIAL AND ERROR
Delays in Drug's Test
Fuel Wider Data Debate
By RON WINSLOW and SARAH RUBENSTEIN
March 24, 2008; Page A1
In January, Merck & Co. and Schering-Plough Corp. disclosed surprising news: A long-overdue study of their blockbuster cholesterol drug Vytorin found it was no better at fighting heart disease than a far-cheaper generic. Doctors and public officials questioned whether the companies had delayed the results for more than a year to protect billion of dollars in sales.
The firms said they had merely been trying to correct irregular data. But by late 2005, company officials overseeing the study already had spotted a threat to Vytorin. They noticed, they now say in interviews, that patients being studied were healthier than expected -- which could make it harder to demonstrate Vytorin's superiority at slowing disease.
The Vytorin study has fueled two broad controversies, both likely to be stoked this week as the findings are formally presented for the first time. Its handling has sharpened the debate over how much control sponsors should wield in clinical trials that influence doctors and regulators. And the study's results have led some skeptical doctors to question the value of cholesterol-lowering drugs that have become the front-line medical weapon against heart disease, the Western world's leading killer.
The management of the Vytorin trial, known as Enhance, was unusual in a number of ways. The companies have previously acknowledged that they kept the study under wraps for more than a year amid a prolonged effort to address their concerns about the data. They didn't provide for an independent steering committee to oversee scientific issues, as is typically -- though not always -- done.
The companies brought in a second lab to compete to produce more accurate results than the original research team. In another unusual step, they proposed changing the primary way the data would be analyzed after it had already been collected. Company officials began publicly minimizing the significance of the study as too narrow. Finally, the companies compiled the results themselves, without the participation of the outside academic scientist they had hired to lead the study, and announced the findings in a news release in January, rather than wait for the researcher's report.
Dutch researcher John Kastelein, the outside scientist who led the study, has called it "a trial from hell." In an interview in December, he said he wished he had stood up to the companies before they announced they planned to change how the data would be analyzed. They later abandoned that move amid a blitz of criticism. "This is an unusual case history of clinical research," says Jerry Avorn, a drug-policy researcher at Harvard Medical School not connected with the study.
Merck and Schering-Plough say they were acting with integrity. The companies say they didn't peek at the results or know Vytorin had failed in the study until very recently. "What we were trying to do was to improve...the precision and the accuracy of the data so that at the end, the results would be credible," says Enrico Veltri, Schering-Plough group vice president of global clinical development and a leader of the companies' joint venture. He says no independent steering panel was needed because the study was ostensibly simple, seeking only additional information about drugs that were already on the market. "There was nothing to steer," he says.
Sales Plunge
The ultimate disclosure of the findings sent sales and prescriptions of Vytorin plunging, along with the stock prices of the two companies, wiping out tens of billions of dollars in market capitalization. Thousands of patients inundated doctors' offices, many of them wanting to stop taking Vytorin or its sister drug, Zetia. Other patients were so unsettled that they wanted to stop taking another class of drugs including Zocor called statins, which are well-proven to reduce risk of death and heart attack.
The brouhaha came in the wake of scandals in which drug companies were accused of suppressing negative studies about antidepressants and other drugs, including Merck's own painkiller Vioxx. (Merck has denied withholding unflattering data.) A congressional probe was launched in December into the companies' handling of the Vytorin study, and New York authorities are among those examining whether insider stock trades before the announcement of the study's results were appropriate.
In Vytorin's case no safety risk emerged from the data, but the episode has renewed a growing concern among medical-journal editors and policy makers over "selective publication," in which positive trials are submitted quickly for dissemination while negative studies are delayed or shelved. The conduct of the trial helped "damage the public trust" in medical research, says Philip Greenland, dean for clinical research at Northwestern University's Feinberg School of Medicine, Chicago, and editor of the journal Archives of Internal Medicine, who wasn't associated with the Enhance study. Academic researchers should respond by avoiding studies where sponsors can "put [the data] in a drawer and trash it or have the endpoint changed," he says.
Amid the controversy, the Vytorin trial is getting unusual billing Sunday at a meeting of the American College of Cardiology, where Dr. Kastelein will present the findings for the first time in a scientific forum. The study is being featured at the professional society's opening plenary session in Chicago, and will be followed by an extraordinary roundtable of cardiologists to discuss the study's implications for treating patients.
Vytorin is a combination of Schering-Plough's Zetia and Merck's Zocor, two drugs that each block low-density lipoprotein or LDL cholesterol -- also known as "bad cholesterol" -- in a different way. Zocor, like other statins, blocks LDL production in the liver, while Zetia prevents its absorption from food in the gut. The companies formed a joint venture in 2000 to develop and market cholesterol drugs. One result has been to give Zocor, which went off patent in 2006 and lost most of its share of the market to generics, a profitable new life as a component of Vytorin.
Zetia, introduced in 2002, and Vytorin, in 2004, are both approved by the Food and Drug Administration to lower LDL cholesterol. But the companies haven't proved that Zetia contributes any heart-attack protection on its own or that Vytorin provides more than a statin alone, so many doctors still are reluctant to prescribe both drugs despite their popularity. The Enhance study was intended to provide a partial answer by determining whether Vytorin had more effect than Zocor alone on the wall thickness of neck arteries, a marker that other studies have shown correlates with risk of heart attack and stroke.
Enhance was the first large trial to use a new digital technology to record ultrasound images to measure tiny changes in the neck artery walls, looking for differences of tenths or even hundredths of a millimeter to weigh the effects. Half of the 720 patients in the study took a combination of Zetia and Zocor, while the control group took only Zocor. It was a randomized, double-blind trial -- the gold standard of study designs -- meaning patients were assigned to the drug groups randomly and neither they nor researchers knew which medicine they were taking during the trial. Patients were followed for two years, with measurements taken of their neck arteries when they entered the study and every six months.
In late 2005, Dr. Kastelein's team began sending complete measurements from the first group of patients to Schering-Plough's New Jersey headquarters. Company statisticians there, in a generally accepted practice, began routine checks to make sure the data were in order. The plan was to evaluate the last patient in April 2006, "unblind" the data that summer and send the results back to Dr. Kastelein for analysis. He expected to present the findings at a big cardiology meeting that fall.
But during the initial data checks, the companies got a jolt: The artery-wall measurements taken when the patients entered the study were near-normal -- even though they all had a genetic condition that causes unusually high cholesterol, says Schering-Plough's Dr. Veltri.
The likely reason: Most of the patients had already been taking high doses of cholesterol-lowering statins, which had become standard treatment before the study began. Researchers hadn't anticipated the impact. "It was obvious to anyone that this would be a higher hurdle," said Tom Musliner, a director of cardiovascular research at Merck and a member of the companies' clinical committee overseeing Enhance.
'Implausible' Data
The companies say they didn't conclude at the time that Vytorin had failed in the study. But they figured that artery-wall readings would have to be especially precise to show any statistically significant differences between the drugs. They also emphasize that early data checks turned up another problem they found even more troubling: missing or "implausible" data that the companies have previously cited as the reason for the long delay in reporting the findings. In some cases they were concerned about wide fluctuations in readings that were supposed to be precise. Other researchers say variation in readings is inevitable in most imaging studies, and that enrolling enough randomly assigned patients spreads any problems among both groups to avoid affecting the overall results.
The companies began sending what became a constant flow of inquiries to Dr. Kastelein asking him and his lab to check and clarify data and re-read images in a bid to resolve the problems. Early in 2006, the companies' committee proposed a different approach to reading the still-blinded data, and pitted Dr. Kastelein's lab against an outside research team to see whether one would be more accurate. There was no meaningful difference and Dr. Kastelein's team was kept on the case, spending the second half of 2006 re-reading most of the 40,000 images taken in the study, according to the companies. In January 2007, an independent consultant told the companies that the quality of the Enhance data was similar to what was found in other comparable trials.
But company officials still weren't satisfied. They say they kept exploring different ways to eliminate wayward readings and hone the study's precision. "It's very atypical for a trial to go through this sort of scrutiny," says Allen Taylor, chief of cardiology service at Walter Reed Army Medical Center, Washington, D.C., and an expert in imaging of neck arteries.
The companies defend the effort. "It wasn't that the study looked like it was totally inadequate," says Merck's Dr. Musliner. "The more you can reduce your variability, the greater your chances of showing the significance of smaller differences."
Dr. Kastelein, who was running his first large study, said in a December interview that he grew frustrated by the companies' demands and repeatedly tried to reassure the sponsor about the data and its readiness. He said then that he takes some responsibility for letting the delays go on so long, but that results would have been presented months ago if he had control of the data. He has since declined to speak publicly about Enhance while waiting to officially present the findings. He declined to comment about the companies' new assertion about the health of the study subjects' arteries.
By April 2007, Schering-Plough Chairman and Chief Executive Fred Hassan, who joined the company after the study was launched, was downplaying its significance in a conference call with analysts, saying it was conducted in "a very special population with very special doses." He reminded analysts about a much bigger study looking directly at heart attacks in high-risk patients that would be completed in 2011. He and Merck Chief Executive Richard Clark pointed again to that larger study early this year, and in an interview Mr. Hassan said the rationale for conducting the Enhance study wasn't clear to him. "There must have been some hypothesis they wanted to test at that time," he said.
Troubled Times
Meanwhile, sales of Vytorin and its sister drug, Zetia, continued to soar in 2007, garnering more than $5 billion combined and helping both companies recover from troubled times. Merck faced huge liability and a big sales loss after Vioxx was linked to heart attacks and pulled from the market. Schering-Plough's biggest drug, the allergy fighter Claritin, had plummeted after generic versions came out.
In November, in another unorthodox move, the companies convened an outside panel of imaging experts and announced that on its recommendation they planned to change the study's primary "endpoint" -- the main measurement of Vytorin's effectiveness. Such a change, akin to moving the finish line after the race is over, is taboo in clinical studies and would undercut credibility of the results. When Dr. Veltri called Dr. Kastelein, the researcher initially went along with the plan, and the companies announced it. "I was kind of overwhelmed," he said in an interview in December. "I didn't want to lose the trial."
Several days later, he called Dr. Veltri back to oppose it. Other scientists and the companies' own standing advisory committees were critical as well. The move came as study authors were increasingly being asked by major medical journals to attest that they, and not sponsors, controlled their data and analysis.
In mid-December, the companies dropped the plan themselves, but the damage was done. The same day, a congressional committee announced an inquiry. "We are concerned with the delay in releasing the results of the study...and the apparent manipulation of trial data," Michigan Democratic Reps. John Dingell and Bart Stupak wrote to the two companies. They have since widened their inquiry to include how the companies market the drugs and allegations of insider trading. Schering-Plough says an executive's trades in mid-2007 were approved under company rules.
On Dec. 31, 2007, the companies unblinded the Enhance study, revealing which drugs each patient took. Schering-Plough and Merck issued the results in a news release on Jan. 14.
Patients who took the Zetia-Zocor combination akin to Vytorin achieved LDL levels 17% lower on average than those who took Zocor alone. But there wasn't any significant difference in the patients' neck arteries, called carotid arteries, suggesting the additional cholesterol reduction didn't have a benefit.
That finding has led some doctors to question the benefits of lowering LDL, whether by Zetia or statins such as Zocor. Zetia faces more questions because more than a dozen studies have shown long-term use of statins significantly reduces heart attacks and deaths along with reducing LDL.
Some researchers believe the Enhance test was simply unreliable and that Vytorin's impact against cardiovascular disease won't be known until completion of the bigger study -- 12,500 patients -- comparing Vytorin against Zocor in preventing deaths and heart attacks in very high-risk patients. That's expected in 2011.
online.wsj.com |
