We examined whether hemodynamic improvement after high-flow hemofiltration predicts survival in patients treated with standard continuous renal replacement therapy (CRRT). This was a prospective, observational cohort study of 169 patients, measuring the mean arterial pressure (MAP) and norepinephrine (NE) dosage before and 24 hours after CRRT. Responders were defined as having a 20% reduction in NE dosage or a 20% rise in MAP with no increase in NE, compared with nonresponders. Patients were considered to be unstable if they were receiving NE or their MAP was lower than 60 mm Hg before CRRT. Of the 169 patients, 68% were men; mean age was 53.8 years (52.7 to 54.9), with a mean Acute Physiology and Chronic Health Evaluation (APACHE) II at admission of 21.8 (21.2 to 22.3), of whom 114 were unstable at the start of CRRT. Overall mortality rate 15 days after the end of CRRT was 54.3% (57.7% in stable vs. 52.9% in unstable patients, p = NS). There were 99 responders and 70 nonresponders, the only differences being NE dosage (higher in responders, p < 0.01) and mortality rate (responders 30% vs. nonresponders 74.7%, p < 0.001). In unstable patients, mortality rate was 30% in responders versus 87% in nonresponders (p < 0.001) (72% sensitivity and 86% specificity for predicting death). Logistic regression analysis showed that the only variables associated with death were APACHE II at admission (OR, 1.06; 95% CI, 1.0 to 1.12), percent creatinine decrease (OR, 0.98; CI, 0.96 to 1.0), and lack of hemodynamic response to CRRT (OR, 7.04; CI, 3.3 to 15.02). Hemodynamic improvement after 24-hour CRRT is a strong predictor of survival.