Acute renal failure (ARF) is associated with morbidity and mortality in excess of 50% in the intensive care unit (ICU) setting. A variety of outcome measures have been described in published reports of ARF, however, the studies often do not distinguish between clinical outcomes and surrogate endpoints. Multiple factors can influence these outcomes, including variations in practice. It is important to be aware of the potential effects of these factors when clinical trials are planned and executed for ARF patients. For any intervention trial, knowledge of the natural history of the disease and process of care informs the design and conduct of the trial. Standardization of a definition for ARF and of the criteria for initiation, frequency, duration, and withdrawal of dialysis support would be of great benefit. This article provides a critical appraisal of outcomes research in ARF and describes an approach for selecting appropriate endpoints for future clinical research in ARF.