However, the NIH-sponsored studies were not powered to assess differences in infrequent but serious events such as death and stroke, which were found in previously-published large Medicare database analyses to be significantly more frequent with unlicensed Avastin.