Abstract

Although sedation is a fundamental component of intensive care unit (ICU) management, the most effective approach remains ambiguous. The objective of this network meta-analysis (NMA) was to assess and contrast the efficacy of a variety of sedation protocols in adult ICU patients with respect to postoperative delirium (POD), ventilator-free days (VFDs), ICU length of stay (LOS), and mortality. The study included sixteen randomized controlled trials (RCTs) that involved 3,783 patients. Random-effects models were implemented to conduct Bayesian network meta-analyses. Dexmedetomidine was the most effective treatment for POD, substantially reducing delirium in comparison to midazolam and ranking first in 61.7% of simulations. In a distinct network, the most effective method for reducing POD was daily sedation interruption (DIS). However, isoflurane was the most effective in terms of VFDs, despite the absence of statistical significance. Lorazepam demonstrated the lowest performance (70.1% worst rank). In the second network, protocols that did not require sedation were the most effective in optimizing VFDs, with a top rank of 70.1%. Lorazepam was significantly associated with increased mortality, whereas mild sedation and no sedation were ranked most favorably in the alternative network model. Inhaled isoflurane and dexmedetomidine are among the most efficacious agents for reducing POD and increasing ventilator-free days, respectively. There is a correlation between reduced ICU LOS and mortality, and light sedation and no-sedation strategies. In contrast, lorazepam and midazolam were consistently linked to worse outcomes. These results advocated for the prioritization of mild, individualized sedation strategies in ICU care.

Key words: Sedation protocols, ICU, delirium, mortality, ventilator-free days, network meta-analysis