Objective: to investigate the effects of continuous venovenous hemofiltration on the hemodynamics of critically ill patients with multiple organ dysfunction syndrome.Design: prospective, descriptive study of hemodynamic, respiratory and clinical data.Setting: tertiary university referral hospital. Multidisciplinary Intensive Care Unit (ICU).Patients: 55 consecutive patients with multiple organ dysfunction syndrome (DODS) requiring venovenous hemofiltration who had invasive hemodynamic monitoring for management of hypotension and/or hypoxemia.Interventions: Hemodynamic profile was recorded immediately before, at 3 hours, 6 hours and 12 hours alter the hemofiltration was started (mean of 3 set data each time).Measurements and main results: Fifty live patients with MODS who required hemodynamic monitoring and CVVH were analyzed. The mean age was 46 + 17 years (range 18-77 years). All patients were in marked positive fluid balance and were grossly edematous when CVVH was begun. All of them received inotropic support. Twenty of these 55 patients (36 %) recovered from their MODS, and survived to leave the intensive care unit. All patients were hemodynamically stable during hemofiltration. Mean arterial pressure rose from a mean of 79 +/- 12 to 92 +/- 15 mmHg (p < 0,001); PaO2/FiO(2) index from 106 +/- 43 to 145 +/- 52 (p < 0,001); tissue oxygen uptake index from 172 +/- 50 to 207 +/- 60 ml/min*m(2) (p < 0,01); and tissue oxygen extraction from 24 +/- 7 to 27 +/- 7 % (p < 0,01). There was a significant decrease in the mean pulmonary venous to arterial shunt fraction (37 +/- 10 to 30 +/- 9 %) (p < 0,01). Mean arterial pressure before starting treatment was reduced in the dying group, but rose during the treatment and remained stable. Systemic vascular resistance index was significantly reduced in the dying group before starling treatment, but also rose after hemofiltration began.Conclusions: continuous venovenous hemofiltration is associated with a significant improvement in several hemodynamic and respiratory variables in critically ill patients with MODS. We presume that the mechanism of action involves the clearance of vasoactive mediators. Further work is necessary to decide whether this technique can reduce the high mortality oi this condition.