OBJECTIVE To examine the validity of the low-dose "renal" dopamine regimen in critically ill patients by investigating the steady-state clearance of dopamine. METHODS A prospective clinical study. METHODS Teaching hospital intensive care unit. METHODS 48 haemodynamically stable patients receiving a dopamine infusion. METHODS Sampling of arterial blood and dopamine infusates. RESULTS Plasma and infusate dopamine levels were measured by liquid chromatography with electrochemical detection. Steady-state clearance was determined by dividing the actual infusion rate by the steady-state plasma concentration. Dopamine clearance for the whole group was 46.4 +/- 35.9 ml/kg per min (mean +/- SD), which is significantly lower than 70 +/- 15.2 ml/kg per min reported for elective surgical patients (p = 0.01). Twelve patients with renal dysfunction had significantly lower dopamine clearances (36 +/- 16.6 ml/kg per min) than the remaining 36 patients (61 +/- 38.5 ml/kg per min, p = 0.022). There was a very poor correlation between plasma dopamine level and infusion rate for the group as a whole (r = 0.47), and this worsened (r = 0.31) when only those patients on a "renal" dose of 2-5 microg/kg per min were considered (n = 30). CONCLUSIONS Plasma dopamine clearance is lower in critically ill patients and there is a large interindividual variation. It is therefore impossible to predict the plasma level from the infusion rate. Consequently, the concept of a selective renovascular low-dose dopamine infusion is invalid in critically ill patients.