In many treatment trials the aim is to prevent deaths from one type of disease. The cause of death evaluation is therefore of the utmost importance. A high autopsy rate is required, even as regards sudden deaths that cannot routinely be referred to as ischemic heart disease. Autopsies are also necessary for studying non-lethal and lethal side effects of the involved drug. Even in the best of cases the post-mortem rate is far from 100% and correctness of the cause of death evaluation is not ensured. If the treatment has prevented deaths from the cause studied, this gain must not be outweighed by increases of other causes of death, for the result to be considered reliably positive