Résumé : Background. There is uncertainty on the effect of different dialysis modalities for the treatment of patients with acute kidney injury (AKI), admitted to the intensive care unit (ICU). This controlled clinical trial performed in the framework of the multicentre SHARF 4 study (Stuivenberg Hospital Acute Renal Failure) aimed to investigate the outcome in patients with AKI, stratified according to severity of disease and randomized to different treatment options. Methods. This was a multicentre prospective randomized controlled trial with stratification according to severity of disease expressed by the SHARF score. ICU patients were eligible for inclusion when serum creatinine was >2 mg/dL, and RRT was initiated. The selected patients were randomized to intermittent (IRRT) or continuous renal replacement therapy (CRRT). Results. A total of 316 AKI patients were randomly assigned to IRRT (n = 144) or CRRT (n = 172). The mean age was 66 (range 18-96); 59% were male. Intention-to-treat analysis revealed a mortality of 62.5% in IRRT compared to 58.1% in CRRT (P = 0.430). No difference between IRRT and CRRT could be observed in the duration of ICU stay or hospital stay. In survivors, renal recovery at hospital discharge was comparable between both groups. Multivariate analysis, including the SHARF score, APACHE II and SOFA scores for correction of disease severity, showed no difference in mortality between both treatment modalities. This result was confirmed in pre-specified subgroup analysis (elderly, patients with sepsis, heart failure, ventilation) and after exclusion of possible confounders (early mortality, delayed ICU admission). Conclusions. Modality of RRT, either CRRT or IRRT, had no impact on the outcome in ICU patients with AKI. Both modalities need to be considered as complementary in the treatment of AKI (Clinical Trial: SHARF 4, NCT00322933, http://ClinicalTrials.gov). © The Author [2008]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.