BACKGROUND The pharmacokinetics of remifentanil, an opioid analgesic metabolized by non-specific esterases, and its principal metabolite, remifentanil acid (RA), which is excreted via the kidneys, were assessed as part of an open-label safety study in intensive care unit (ICU) patients with varying degrees of renal impairment. METHODS Forty adult ICU patients with normal/mildly impaired renal function (creatinine clearance [CL(cr)] 62.9 (sd) 14.5 ml min(-1); n=10) or moderate/severe renal impairment (CL(cr) 14.7 (15.7) ml min(-1); n=30) were included. Remifentanil was infused for up to 72 h, at a starting rate of 6-9 microg kg(-1) h(-1) titrated to achieve a target sedation level, with additional propofol (0.5 mg kg(-1) h(-1)) if required. Intensive arterial sampling was performed for up to 72 h after infusion. Pharmacokinetic parameters obtained by simultaneous modelling of remifentanil and RA data were statistically compared between the two groups. RESULTS Remifentanil pharmacokinetics were not significantly affected by renal status. RA clearance in the moderate/severe group was reduced to about 25% that of the normal/mild group (41 (29) vs 176 (49) ml kg(-1) h(-1), P<0.0001). Metabolic ratio, a predictor of the ratio of RA to remifentanil concentrations at steady state, was approximately eight-fold higher in the moderate/severe group relative to the normal/mild group (116 (110) vs 15 (4), P<0.0001). Maximum RA levels approached 700 ng ml(-1) in the moderate/severe group. CONCLUSIONS Although RA accumulates in patients with moderate/severe renal impairment, pharmacokinetic modelling predicts that RA concentrations during a 9 microg kg(-1) h(-1) remifentanil infusion for up to 15 days would not exceed those reported in the present study, for which no associated prolongation of mu-opioid effects was observed.