We performed a double-blind placebo-controlled trial to study whether early treatment with erythropoietin could prevent the development of acute kidney injury in patients in two general intensive care units. As a guide for choosing the patients for treatment we measured urinary levels of two biomarkers, the proximal tubular brush border enzymes gamma-glutamyl transpeptidase and alkaline phosphatase. Randomization to either placebo or two doses of erythropoietin was triggered by an increase in the biomarker concentration product to levels above 46.3, with a primary outcome of relative average plasma creatinine increase from baseline over 4 to 7 days. Of 529 patients, 162 were randomized within an average of 3.5 h of a positive sample. There was no difference in the incidence of erythropoietin-specific adverse events or in the primary outcome between the placebo and treatment groups. The triggering biomarker concentration product selected patients with more severe illness and at greater risk of acute kidney injury, dialysis, or death; however, the marker elevations were transient. Early intervention with high-dose erythropoietin was safe but did not alter the outcome. Although these two urine biomarkers facilitated our early intervention, their transient increase compromised effective triaging. Further, our study showed that a composite of these two biomarkers was insufficient for risk stratification in a patient population with a heterogeneous onset of injury.