Sedation Practices in Mechanically Ventilated Neurocritical Care Patients from 19 Countries: An International Cohort Study.

BACKGROUND

Our objective was to characterize the impact of common initial sedation practices on invasive mechanical ventilation (IMV) duration and in-hospital outcomes in patients with acute brain injury (ABI) and to elucidate variations in practices between high-income and middle-income countries.

METHODS

This was a post hoc analysis of a prospective observational data registry of neurocritically ill patients requiring IMV. The setting included 73 intensive care units (ICUs) in 18 countries, with a total of 1,450 patients with ABI requiring IMV. There were no interventions.

RESULTS

Patients were categorized into day 1 propofol, midazolam, propofol and midazolam, dexmedetomidine, or sodium thiopental. The primary outcome was duration of IMV. Secondary outcomes were ICU and hospital mortality, ICU length of stay, days to first extubation, extubation failure, and withdrawal of life-sustaining therapy. Multivariable analyses were adjusted for clinically preselected covariates. Of 1,450 included patients (median age 54 years, 66% male), 41.2% (n = 597) were started on propofol, 26.1% (n = 379) were started on midazolam, 19.9% were started on propofol and midazolam, 0.3% (n = 5) were started on sodium thiopental, 0.7% (n = 10) were started on dexmedetomidine, and 11.8% (n = 171) were treated without sedation. After adjustment, there was no significant difference in IMV duration between patients who received midazolam (aβ = 0.64, p = 0.43, 95% confidence interval [CI] - 0.96 to 2.24) or propofol and midazolam (aβ = 0.32, p = 0.46, 95% CI - 1.44 to 2.12) compared with patients who received propofol. Patients who were started on midazolam had an average length of ICU stay that was 2.78 days longer than patients started on propofol (p = 0.003, 95% CI 0.94-4.63). There were no differences in mortality, days to first extubation, extubation failure, or withdrawal of life-sustaining therapy. Patients from high-income countries (n = 1,125) were more likely to receive propofol on day 1 (45.7 vs. 25.5%), whereas patients from middle-income countries (n = 325) were more likely to receive midazolam (32.6 vs. 24.3%) (p < 0.001).

CONCLUSIONS

In an international registry of patients with ABI requiring IMV, IMV duration did not differ significantly relative to initial sedation strategy. However, patients started on midazolam had longer ICU stay.