LONDON, Sept 14 (Reuters) - The failure of AstraZeneca's (AZN) experimental drug NXY-059 for stroke last year shows worrying flaws in the way experimental drugs are tested first in animals, a leading neurologist said on Friday.
Better tests of the compound on animals would have revealed early on that the product did not work, saving AstraZeneca millions of dollars, according to Malcolm Macleod of Stirling Royal Infirmary.
AstraZeneca scrapped development of NXY-059 last October after the drug, which was licensed from Renovis (RNVS) and was designed to protect patients from permanent damage after a stroke, failed in a major clinical study.
…Macleod said AstraZeneca could have avoided this heartache if tests on animals had been properly randomised and blinded. In an analysis of the NXY-059 animal tests, presented at a meeting sponsored by the British Association for the Advancement of Science in York, he found the animal results were distorted when allocations to experimental groups were not made randomly and were not concealed from the researcher.
Studies which did not randomise animals concluded that NXY-059 improved outcome by more than 50 percent -- but those that did put the effect at only 20 percent.
"There is always an element of bias because scientists want to produce positive results," Macleod said in a telephone interview.
Typically, that background level of bias results in about a 10 percent over-statement of efficacy.
In the case of NXY-059, the bias was much larger and Macleod said this may have been due to the fact that the animal studies, although performed in academic laboratories, were funded by industry.
"That may lead to a motivation to try and make sure the company gets what you think they want -- although I think AstraZeneca would have much rather known that the studies failed," he said.
News 
Market Data 
Discover 
