In the ICU setting it is often not possible to identify a single specific cause of acute renal failure (ARF). Rather, this is a complex event, with major involvement of ischaemia and direct nephrotoxicity. Recent data show that an acute kidney injury is not merely a consequence of a severe generalised illness, but may itself be a significant mediator of the illness. Via the effects of uraemia, immunosuppression, hypervolaemia or electrolyte imbalance acute-on-chronic RF may impact the outcome of the critically ill patient. Accordingly, it is of prime importance to avoid renal impairment, or to diagnose it in its earliest stages and to institute appropriate therapy as early as possible. With this in mind, the AKIN-staging system (Acute Kidney Injury Network) was introduced in 2007. An absolute increase in serum creatinine by only 0.3 mg/dl (26.4 mcmol/l) is classed as Stage I ARE Unfortunately, the increase in serum creatinine merely confirms an earlier renal injury, but the parameter also enters calculations of the glomerular filtration rate (GFR). The MDRD-calculation (Modification of Diet Renal Disease) is thus merely an approximate estimation of the GFR. More recent markers, such as the NGAL (Neutrophil Gelatinase-Associated Lipocalin) in serum and urine can serve as much earlier indicators of ischaemia-induced ARF. They enable a faster diagnosis of ARF than serum creatinine. Determinants of the GFR are the maintenance of mean arterial pressure and a balanced volume status. Any fluid deficit should always be corrected with crystalloids; colloids are justified only in the event of an acute intravascular volume deficit. The uncontrolled use of high doses of diuretics in intensive care is not recommended, as this strategy does not prevent the development of ARF. Nor is there consensus on the precise criteria for the initiation of renal replacement therapy. However, there is growing evidence in favour of early initiation and adequate intensity of renal replacement therapies. In summary: a significant increase in urea and serum creatinine indicating an incipient acute-on-chronic renal failure or ARF in the critically ill patient, progressive reduction of diuresis and an increasing demand for diuretics to maintain the fluid balance indicate a need for the timely commencement of renal replacement therapy, independent of the patient's absolute creatinine and urea values. In the ICU setting, renal replacement therapy should ideally be continuous or, if this is not possible, intermittent on a daily basis and as easy on the patient as possible. The choice between continuous and intermittent haemodialysis should be determined by the current level of experience of the ICU involved.