RESUMO
OBJECTIVES: To study efficacy, systemic and cerebral haemodynamic, and cost of sedation with sevoflurane after midazolam failure. STUDY DESIGN: Prospective observational study in a mixed intensive care unit. PATIENTS AND METHODS: Mechanically ventiled patients in whom deep sedation failed (Ramsay score<5 despite midazolam 10mg/h and fentanyl 400µg/h) were enrolled. Sedation with sevoflurane and fentanyl (200µg/h) was performed during 48 hours. Sevoflurane was administered with a dedicated filter (AnaConDa™) and sevoflurane infusion rate was adjusted in order to achieve a Ramsay score ≥5. Ramsay score, mean arterial blood pressure, norepinephrine dose/24h, intracranial pressure and cerebral perfusion pressure in patients with brain injury were measured. Directs costs for sedation were calculated. An analysis of variance for repeated measures compared values between D0 (intravenous sedation), D1 and D2 (inhaled sedation). RESULTS: Twenty-five patients (age=51 [38-63], SAPS II=43 [33-49]) were enrolled. Ramsay score was 4 [4,5] at D0 and 6 [6] at D1 and D2 (P<0.05 vs D0). Mean arterial pressure was significantly lower at D1 (80 [73-86] mmHg) as compared to D0 (84 [77-92] mmHg) and D2 (84 [78-91] mmHg) (P<0,05). Norepinephrine consumption was lower at D2 as compared to D1 (P<0,05). Intracranial pressure was lower at D1 (9 [5-13] mmHg) and D2 (11 [7-15] mmHg) as compared to D0 (12 [7-17] mmHg) (P<0.05). PPC was stable at D1 and increased at D2. Directs costs were significantly increased with sevoflurane. CONCLUSION: Sevoflurane is an effective and safe alternative to midazolam in ICU patients associated with a moderate increase in costs.
