May 12th, 2015
Diet Drug Study Crashes and Burns in the Wake of Leaked Results
The ill-fated Light trial, which was supposed to examine the cardiovascular outcomes of the weight loss drug Contrave, a combination of naltrexone and bupropion marketed by Orexigen and Takeda, came to a spectacular halt today. The action was probably inevitable given the extreme controversy generated earlier this year when it became known that Orexigen had widely disseminated results from an early interim analysis of the study.
The news about the trial was announced in a press release from the companies and a press release from the Cleveland Clinic, the home institution of Steve Nissen, the trial’s chairman.
Contrave was approved by the FDA in 2014 partly on the basis of a pre-specified interim analysis of Light, which ruled out a significant doubling of cardiovascular risk for people taking the drug. (It is worth noting that the idea for approving drugs based on early results of outcomes trials was originally proposed by Nissen and Thomas Fleming, who was the chair of the data and safety monitoring committee of the trial.) The interim analysis was performed after the 9,000-patient trial had achieved 25% of its primary endpoint events, major adverse cardiovascular events. This analysis contained only 94 events: 59 in the placebo group and 35 in the Contrave group. The results were supposed to be held in strict confidence, but Orexigen shared them with over 100 people both inside and outside the company. When the FDA found out about the initial leak, it said that a second cardiovascular outcomes trial would have to be performed.
Then in March of this year, the interim results became public knowledge when Orexigen used the data to support a patent application. This caused a firestorm of criticism and controversy. In an interview, Nissen said that the trial’s executive committee unanimously agreed that the release of the 25% data “made it impossible to complete the Light trial as originally intended.” The committee further agreed that data from the 50% interim analysis should be released, “with the view that this information was important in that patients are receiving the drug on the mistaken belief of cardiovascular benefit.”
The Cleveland Clinic press release includes important additional information about the trial from the 50% interim analysis. Most importantly, the new results failed to confirm the earlier finding of cardiovascular benefit. The new analysis contains a total of 192 endpoint major adverse cardiovascular events: 102 in the placebo group compared with 90 in the Contrave group (HR=0.88, CI 0.66 – 1.17). This result reflects a shift in the outcomes: the early benefit in the first quarter of the trial was reversed in the second quarter, in which there were 43 events in the placebo group and 55 events in the Contrave group.
With some bitterness, Nissen notes that Orexigen was “perfectly willing to allow” the positive data from the 25% analysis to be made public but not the data from the 50% analysis. After 6 weeks of negotiations, the companies announced the termination of the trial but did not include the data from the 50% analysis. “We felt it was important to get this information out to the public,” said Nissen. The trial leaders therefore issued their own press release.
Nissen said that no one could be certain whether the change in direction represented a regression to the mean or an emerging signal of a possible long-term adverse effect associated with long-term use of the drug.
“These results show neither benefit nor harm for patients taking the drug, but are consistent with the requirement by the FDA that the Light Trial demonstrate an absence of a doubling of cardiovascular risk for patients taking the drug,” said Nissen in the Cleveland Clinic press release. “The inconsistency of effects on cardiovascular outcomes between the first 25 percent and the second 25 percent of the Light Study clearly illustrates the risks inherent in pre-judgment of clinical trial results based upon an interim analysis and demonstrate why interim results should remain confidential during any ongoing trial.”