Closed-loop systems for automated hypnotic drug delivery during general anaesthesia: a systematic review and...

Closed-loop systems for automated hypnotic delivery integrate continuous bispectral index (BIS) monitoring with controller algorithms to adjust infusion rates in real time. Unlike earlier syntheses that focused on aggregate control metrics, we frame anaesthetic depth as an asymmetric risk problem by separating exposure to potentially inadequate anaesthesia from exposure to excessive depth of anaesthesia.

This PROSPERO-registered systematic review (CRD420251208039) searched PubMed, Embase, Cochrane CENTRAL, and Web of Science through November 2025. Randomised trials comparing BIS-guided closed-loop systems with clinician-directed manual control in adult noncardiac surgery were included. The primary outcome was the proportion of time within 10 BIS units of the target.

Seventeen trials encompassing 1898 patients were included. Closed-loop systems increased time within the target BIS range by 17.6% (95% confidence interval [CI] 11.8-23.5; P<0.001), corresponding to approximately an additional 20 min of optimal anaesthetic depth during a 2 h procedure. Bilateral safety was shown: time with BIS <40 was significantly reduced (mean difference [MD] -17.2%; 95% CI -26.8 to -7.5; P<0.001), whereas time with BIS >60 was unchanged (MD -0.8%; 95% CI -2.6 to 1.0; P=0.38). Controller performance metrics favoured automation (wobble, MDAPE, global score; all P<0.001). Time to extubation was shorter (MD -1.68 min; P<0.001), with no differences in propofol consumption (P=0.13) or vasopressor use (risk ratio 1.04; P=0.62).

BIS-guided closed-loop systems achieve a bilateral safety profile in anaesthetic depth control, reducing deep anaesthesia without increasing light anaesthesia, with modestly faster recovery and preserved haemodynamic stability. These findings support automated hypnotic control as a strategy to rebalance competing intraoperative risks while maintaining physiological stability.

PROSPERO CRD420251208039.