Effect of level of sedation on outcomes in critically ill adult patients: a systematic review of clinical trials with meta-analysis and trial sequential analysis

Background Sedation is routinely administered to critically ill patients to alleviate anxiety, discomfort, and patient- ventilator asynchrony. However, it must be balanced against risks such as delirium and prolonged intensive care stays. This study aimed to investigate the effects of different levels of sedation in critically ill adults. Methods Systematic review with meta-analysis and trial sequential analysis (TSA) of randomised clinical trials including critically ill adults admitted to the intensive care unit. CENTRAL, MEDLINE, Embase, LILACS, and Web of Science were searched from their inception to 13 June 2023. Risks of bias were assessed using the Cochrane risk of bias tool. Primary outcome was all-cause mortality. Aggregate data were synthesised with meta- analyses and TSA, and the certainty of the evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach. This study is registered with PROSPERO: CRD42023386960. Findings Fifteen trials randomising 4352 patients were included, of which 13 were assessed high risk of bias. Meta- analyses comparing lighter to deeper sedation showed no evidence of a difference in all-cause mortality (risk ratio (RR) 0.94, 95% confidence fi dence interval (CI) 0.83-1.06; - 1.06; p = 0.28; 15 trials; moderate certainty evidence), serious adverse events (RR 0.99, CI 0.92-1.06; - 1.06; p = 0.80; 15 trials; moderate certainty evidence), or delirium (RR 1.01, 95% CI 0.94-1.09; - 1.09; p = 0.78; 11 trials; moderate certainty evidence). TSA showed that when assessing mortality, a relative risk reduction of 16% or more between the compared interventions could be rejected. Interpretation The level of sedation has not been shown to affect the risks of death, delirium, and other serious adverse events in critically ill adult patients. While TSA suggests that additional trials are unlikely to significantly fi cantly change the conclusion of the meta-analyses, the certainty of evidence was moderate. This suggests a need for future high-quality studies with higher methodological rigor.