Mortality as an endpoint in studies in critically ill patients: a reappraisal of definitions

The recent literature was critically reviewed reporting heterogeneous mortality endpoints in studies involving critically ill patients. Both location- and duration-dependent definitions are used, more or less arbitrarily and sometimes with contradictory results. Location-dependent mortality refers to intensive care or in-hospital mortality whereas duration-dependent mortality concerns 7-180 day mortality after admission or intervention. There is no consensus on how mortality should be evaluated in the critically ill. It is argued that trialists should aim at uniformity in outcome evaluation of critical care, in order to allow comparison of studies, and that 28-day mortality should remain the primary endpoint for intervention studies.