The VA Diabetes Trial analysis that showed no extra cardiovascular risk from rosiglitazone (Avandia) should not be taken as conclusive, cautioned independent investigators and statisticians.
"None of those analyses was randomized," said Daniel Seigel, Sc.D., a retired statistician formerly with the National Institutes of Health who served on the trial's data and safety monitoring committee "They were all akin to observational analyses."
"At none of our meetings did [study investigators] make strong claims that these were definitive analyses," he said.



