BioWorld International Correspondent
LONDON - A widespread failure to control bias in animal tests is the cause of the unprecedented attrition rates that are seen when treatments for stroke are progressed into clinical trials, according to an international group of researchers.
Many of the issues raised by the finding apply to testing using animal models of disease in general.
The researchers have accumulated data on 700 drugs that have been tested in animal models of stroke. Of these, about 500 showed positive effects, but of almost 100 that subsequently advanced into clinical trials, only one, tissue plasminogen activator, was found to be effective in humans.
Most recently, NXY-059, being developed by AstraZeneca plc, showed no effect in a worldwide Phase III clinical trial in 3,200 patients, despite having improved outcomes by 44 percent in animal tests.
"So either the animal studies for NXY-059 were misleading, or the clinical trials weren't sensitive enough, or animal experiments aren't a good model of human disease," said Malcolm Macleod, consultant neurologist at Stirling Royal Infirmary in Scotland.
Macleod is leading an international research project, Collaborative Approach to Meta-analysis and Review of Animal Data in Experimental Stroke (CAMARDES), which aims to bring improvements in the design, conduct, reporting and use of animal data in the development of stroke drugs.
Speaking at the annual British Association for the Advancement of Science meeting at York University earlier this month, Macleod described how the animal data for NXY-059 was analyzed to assess if research bias had influenced the outcome.
It turned out that studies, which did not randomize animals, said NXY-059 improved outcome by more than 50 percent, while those that did, estimated the effect at 20 percent.
Similarly, studies, which did not conceal treatment allocation from the researcher inducing the stroke, indicated that NXY-059 improved outcome by more than 50 percent; those that did estimated the effect at 25 percent. And studies that did not blind the assessment of outcome said NXY-059 improved that outcome by 50 percent; those which did estimated the effect at less than 30 percent.
The researchers have reported similar findings for other putative stroke drugs, but Macleod said the impact of poor study quality was much more pronounced in the case of NXY-059 (which was known also as Cerovive).
Curiously, the 3,200 patient trial was at odds with an earlier 1,700 patient Phase III trial, which met its primary endpoint of a reduction of physical disability at 90 days post-treatment. AstraZeneca, of London, stopped development of NXY-059 when the results of the second Phase III trial were announced in November 2006.
"Another concern with the data for NXY-059 and for other interventions is that the number of animals used in individual experiments was generally too small to allow a precise estimate of outcome. In the long run, this means more animals have to be used as scientists attempt to replicate the imprecise results of other studies," Macleod noted.
Macleod estimated that over the past 20 years at least 250,000 animals have been used in the preclinical development of potential stroke drugs.
He argued that the CAMARADES findings showed there is significant scope for improvements in the design, conduct, analysis and reporting of animal experiments. By minimizing bias, such improvements would improve the amount of valid information gained from those animals used.