Use of Mortality as an Endpoint in Noninferiority Trials May Lead to Ethically Problematic Conclusions