The usefulness of three clinical scales for monitoring sedation in intensive care units was compared. The Ramsay scale, Cook modification of the Glasgow scale, and combined Ramsay-Cook scale were evaluated. Thirty-five patients with continuous intravenous sedation were monitored using all three clinical scales. The scores were collected simultaneosly with a two-observer method carried out by the nurse in charge of the patient and a second observer (nurse or physician). The three scales were considered valid and reproductible, with "good" agreement for the Ramsay scale and "very good" agreement for the Cook modified coma scale and the combined Ramsay-Cook scale. In conclusion, since the Cook modified Glasgow scale was the most reproducible, it is preferred for use in research and daily practice. The Ramsay scale is more subjective, but may be useful for daily nursing activities.