Volatile sedation in the intensive care unit: A systematic review and meta-analysis.

BACKGROUND: Volatile sedation in the intensive care unit (ICU) may reduce the number of adverse events and improve patient outcomes compared with intravenous (IV) sedation. We performed a systematic review and meta-analysis comparing the effects of volatile and IV sedation in adult ICU patients.

METHODS: We searched the PubMed, Embase, Cochrane Central Register, and Web of Science databases for all randomized trials comparing volatile sedation using an anesthetic-conserving device (ACD) with IV sedation in terms of awakening and extubation times, lengths of ICU and hospital stay, and pharmacologic end-organ effects.

RESULTS: Thirteen trials with a total of 1027 patients were included. Volatile sedation (sevoflurane or isoflurane) administered through an ACD shortened the awakening time [mean difference (MD), -80.0 minutes; 95% confidence intervals (95% CIs), -134.5 to -25.6; P = .004] and extubation time (MD, -196.0 minutes; 95% CIs, -305.2 to -86.8; P < .001) compared with IV sedation (midazolam or propofol). No differences in the lengths of ICU and hospital stay were noted between the 2 groups. In the analysis of cardiac effects of sedation from 5 studies, patients who received volatile sedation showed lower serum troponin levels 6 hours after ICU admission than patients who received IV sedation (P < .05). The effect size of troponin was largest between 12 and 24 hours after ICU admission (MD, -0.27 μg/L; 95% CIs, -0.44 to -0.09; P = .003).

CONCLUSION: Compared with IV sedation, volatile sedation administered through an ACD in the ICU shortened the awakening and extubation times. Considering the difference in serum troponin levels between both arms, volatile anesthetics might have a myocardial protective effect after cardiac surgery even at a subanesthetic dose. Because the included studies used small sample sizes with high heterogeneity, further large, high-quality prospective clinical trials are needed to confirm our findings.