A cute kidney injury (AKI) is a common complication in hospitalized patients, with an incidence of 3 to 10% (1-4). In-hospital mortality rates that are associated with AKI remain high, in the range of 30 to 70% (5-8), despite significant improvements in dialytic technology as well as important advances in critical care, which have resulted in improved survival for other critical illnesses, including acute lung injury and sepsis (9). These improvements include continuous renal replacement therapies, which allow for continuous removal of solutes and fluid and may be tolerated better from a hemodynamic standpoint, and biocompatible dialysis membranes, which are associated with reduced complement and granulocyte activation.In general, indications for dialysis in the acute care setting have been extrapolated from those that are applied in chronic kidney disease, including volume overload that is refractory to diuretic therapy; electrolyte abnormalities (in particular hyperkalemia); uremic complications (pericarditis or pleuritis); severe acidosis (pH Ͻ 7.20); and selected toxic ingestions, such as methanol, ethylene glycol, and other water-soluble agents (10,11). However, the evidence base supporting specific dialysis practices in the acute care setting is limited. For example, several studies that were completed in the 1960s and 1970s compared "early" and "late" initiation of dialysis, using blood urea nitrogen (BUN) to define early and late (Table 1). These studies primarily were cohort studies that used historical controls, not randomized, clinical trials. The results of these investigations, along with extrapolation from the ESRD population, promoted recent practice patterns. Currently, many nephrologists often delay dialysis in the acute care setting until the patient has an impending complication of AKI, such as hyperkalemia leading to cardiac arrhythmias, acidosis resulting in hypotension, or oliguria leading to volume overload or hypoxemia, or until the BUN exceeds 100 mg/dl (18). Given differences in protein catabolism, diet, comorbid conditions, and severity of illness, it may be inappropriate to extrapolate any BUN or creatinine cutoff for dialysis initiation in the chronic setting to patients with AKI.Practice patterns regarding decisions to initiate dialysis in the acute care setting have not been well described and have not been subjected to randomized trials. Further investigation is needed to describe and provide evidence for these clinical practices. However, nephrologists and intensivists should recognize that recent evidence-based changes in the practice of critical care medicine have significant effects on the parameters that trigger the decision to initiate dialysis. For example, aggressive volume resuscitation for sepsis will exacerbate volume overload in an oliguric patient; as a result that patient may require dialysis earlier in the course of AKI to avoid complications of volume overload, including hypoxemia. In this article, we review recent changes in critical care practice and the potentia...