Long-term sedation in intensive care unit: A randomized comparison between inhaled sevoflurane and intravenous propofol or midazolam

Abstract

Purpose: To evaluate efficacy and adverse events related to inhaled sevoflurane for long-term sedation compared with standard intravenous (IV) sedation with propofol or midazolam. Methods: Randomized controlled trial. Sixty intensive care unit (ICU) patients expected to require more than 24 h sedation were randomly assigned to one of three groups: group S, inhaled sevoflurane; group P, IV propofol; group M, IV midazolam. All patients also received IV remifentanil for goal-directed sedation (Ramsay scale and pain score) until extubation or for a maximum of 96 h. Primary end points were wake-up times and extubation delay from termination of sedative administration. Proportion of time within Ramsay score 3-4, IV morphine consumption at 24 h post extubation, hallucination episodes after end of sedation, adverse events, inorganic fluoride plasma levels, and ambient sevoflurane concentrations were recorded. Results: Forty-seven patients were analyzed. Wake-up time and extubation delay were significantly (P < 0.01) shorter in group S (18.6 ± 11.8 and 33.6 ± 13.1 min) than in group P (91.3 ± 35.2 and 326.11 ± 360.2 min) or M (260.2 ± 150.2 and 599.6 ± 586.6 min). Proportion of time within desired interval of sedation score was comparable between groups. Morphine consumption during the 24 h following extubation was lower in group S than in groups P and M. Four hallucination episodes were reported in group P, five in group M, and none in group S (P = 0.04). No hepatic or renal adverse events were reported. Mean plasma fluoride value was 82 μmol l-1 (range 12-220 μmol l-1), and mean ambient sevoflurane concentration was 0.3 ± 0.1 ppm. Conclusions: Long-term inhaled sevoflurane sedation seems to be a safe and effective alternative to IV propofol or midazolam. It decreases wake-up and extubation times, and post extubation morphine consumption, and increases awakening quality. © 2011 jointly held by Springer and ESICM.