Early deep sedation was associated with post-hospital one-year mortality in critically ill surgical patients: a propensity-matched retrospective cohort study.

Abstract

Objective: Sedation is a crucial issue in critical care, but the impact of early deep sedation on post-hospital mortality in critically ill surgical patients remains unclear.

Methods: We linked the 2015-2020 critical care database at Taichung Veterans General Hospital with the nationwide death registration in Taiwan. Log-rank test was used to estimate survival curves between patients with and without deep sedation, defined by the average Richmond Agitation-Sedation Scale (RASS) level within the first 3 days equal to or lower than - 3. A multivariable Cox proportional hazards regression model was used to determine hazard ratios (HR) and 95% confidence intervals (CI). Furthermore, we used propensity score-matching (PSM) analysis to validate the association.

Results: A total of 7,135 critically ill surgical patients were enrolled, and 13.7% of them experienced early deep sedation. Independent predictors for post-hospital one-year mortality included old age, male, more comorbidities, high acute physiology and chronic health evaluation (APACHE) II score, and low body mass index. We noted that receiving midazolam (aHR 1.368, 95% CI 1.052-1.780) or propofol (aHR 1.459, 95% CI 1.136-1.874) was associated with increased mortality compared with dexmedetomidine. Early deep sedation was independently associated with post-hospital mortality after adjusting for covariates (aHR 1.216, 95% CI 1.019-1.452), and the association remained robust in the PSM analysis (aHR 1.313, 95% CI 1.054-1.636).

Conclusion: We identified the association between early deep sedation and post-hospital mortality, a modifiable factor, in critically ill surgical patients. Further prospective studies are warranted to confirm our findings.