“Effects of different sedative agents in mechanically ventilated patients with sepsis: a GRADE-assessed systematic review and network meta-analysis”

Abstract Background The optimal sedative agent for mechanically ventilated adults with sepsis or septic shock remains uncertain. We evaluated the comparative effectiveness and safety of commonly used sedatives using network meta-analysis and GRADE. Methods PubMed, Embase, Scopus, Cochrane Library, and ClinicalTrials.gov were searched from inception to March 31, 2026. Randomized controlled trials enrolling mechanically ventilated adults with sepsis or septic shock were included. Random-effects network meta-analysis generated odds ratios (ORs) or mean differences (MDs) with 95% confidence intervals (CIs). P-scores were used for treatment ranking, and certainty of evidence was assessed using GRADE. Results Nine randomized controlled trials involving 2,557 participants were included. No sedative demonstrated a statistically significant advantage for short-term or long-term mortality. Likewise, no significant differences were observed for hospital length of stay or ventilator-free days. Propofol significantly reduced bradycardia compared with clonidine (OR 0.51, 95% CI 0.38–0.68), while dexmedetomidine did not differ from clonidine. Overall, ICU length-of-stay estimates showed some statistically significant pairwise differences, but all were supported by low-certainty evidence. Treatment rankings varied by outcome and were interpreted cautiously because several network nodes were informed by sparse data. Evidence certainty was generally low because of risk of bias and imprecision, although bradycardia comparisons were rated moderate certainty. Conclusion No sedative showed a clear survival benefit in mechanically ventilated patients with sepsis. Sedative choice should therefore remain individualized according to hemodynamic status, organ function, delirium risk, and clinical context. Larger, adequately powered randomized trials are needed. Clinical trial number Not applicable.