BACKGROUNDPostoperative acute kidney injury (AKI) is a cause of morbidity and mortality. Its diagnosis requires better markers than variations in diuresis or postoperative serum creatinine. OBJECTIVESThe aim of this study was to evaluate the accuracy of Doppler renal resistive index for early detection of AKI after hip or knee arthroplasty. DESIGNA prospective observational study. SETTINGA single-centre study in a university hospital. PATIENTSFifty men and women older than 65 years, requiring hip or knee replacement with at least two perioperative AKI risk factors, including diabetes, arteritis, chronic heart or renal dysfunction, and prescription of angiotensin-converting enzyme (ACE) inhibitors. Exclusion criteria were poor abdominal echogenicity, arrhythmia, respiratory failure or agitation. INTERVENTIONRenal resistive index was measured preoperatively and in the postanaesthesia care unit. RESULTSSixteen patients presented with AKI in the postoperative period. Resistive index was increased in this group in both the preoperative 0.72 (0.69 to 0.73) vs. 0.66 (0.58 to 0.71); P = 0.01 and postoperative periods 0.75 (0.71 to 0.75) vs. 0.67 (0.62 to 0.72); P = 0.0001. Resistive index evaluated by ROC curves and AUC to detect AKI was 0.862 95% confidence interval (95% CI) 0.735 to 0.943. The most accurate cut-off value was a postoperative resistive index of 0.705 (sensitivity = 94%, specificity = 71%, LR+ = 3.19 and LR– = 0.09). The grey area between 0.705 and 0.73, corresponding to the inconclusive zone, included 26% (13/50) of all the patients. CONCLUSIONPostoperative resistive index appears to be effective for early detection of AKI after major orthopaedic surgery. Resistive index can be measured in the postoperative care unit in patients at risk of AKI. TRIAL REGISTRATION NUMBER29-0512.