BACKGROUND: Intensive Care Unit (ICU) delirium is a multifactorial syndrome associated with prolonged hospitalization, increased morbidity and mortality, cognitive decline, and higher healthcare costs. Many randomised controlled trials of interventions to prevent or manage ICU delirium have been combined in systematic reviews. We aimed to collate and map the meta-analysed evidence for pharmacological interventions.
METHODS: Eligible reviews included RCTs of any pharmacological intervention designed to prevent or manage critically ill adults with, or at risk of, ICU delirium. We searched 8 databases from inception to 26 September 2023. Two reviewers independently screened search results and confirmed eligibility of full texts. We then mapped the effects for pharmacological interventions (single or combined drugs and sedation strategies) along with the certainty of the evidence for outcomes in the Del-CorS core outcome set, ICU and hospital length of stay.
RESULTS: Of 3,381 studies, we identified 56 relevant systematic reviews reporting our outcomes (17 included in mapping). Thirteen reviews with GRADE assessments were mapped for delirium outcomes (occurrence, duration or severity), six for ICU or hospital mortality, and 15 for ICU or hospital length of stay. The α2-adrenoceptor agonist drug class (primarily dexmedetomidine) had the largest evidence base and was probably favourable over placebo for preventing or reducing delirium occurrence (moderate to high-certainty evidence; 2 systematic reviews). The α2-adrenoceptor agonist drug class (primarily dexmedetomidine) was also probably favourable over placebo for reducing ICU and hospital length of stay (moderate-certainty evidence; one systematic review). The evidence was more variable for other pharmacological comparisons. For ICU mortality, there may have been little or no difference between dexmedetomidine and a non-dexmedetomidine comparator (low certainty), while the evidence was very uncertain on hospital mortality (one systematic review). No meta-analyses reported outcomes for cognition or emotional distress. Co-interventions, in particular non-pharmacological interventions, were often incompletely reported.
CONCLUSIONS: Mapped evidence suggests the α2-adrenoceptor agonist drug class (primarily dexmedetomidine), is most likely to be effective at managing delirium in the ICU. However, underlying conditions indicating or precluding intervention, and the impact of loss to follow-up, remain unclear. We found a lack of synthesised evidence for important core outcomes of cognition and emotional distress, and for deprescribing sedatives/analgesia as part of optimising sedation strategies.
SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13054-025-05615-0.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)