Autonomous closed-loop intravenous esmolol for systolic blood pressure regulation and beat-to-beat variability suppression with safety supervision: a large-animal validation study
J Curneen, J O'Dwyer, C Judge, D Graham, S A Cryan, E Roche, G DuffyAbstract
Background/Introduction
Acute hypertension management in high-acuity settings - including hypertensive emergencies and perioperative care - relies on manual drug titration, which is subject to operator delays, dose imprecision, and haemodynamic overshoot.1 Beyond threshold attainment, blood pressure variability (BPV) is increasingly recognised as an independent cardiovascular risk factor, with the contribution of short-term beat-to-beat fluctuations to haemodynamic stress in the acute setting requiring further exploration.2,3,4 Autonomous closed-loop control offers the potential to achieve both precise target attainment and BPV minimisation simultaneously, but preclinical validation with integrated safety supervision is lacking.
Purpose
To develop and validate a closed-loop intravenous esmolol infusion system for automated systolic blood pressure (SBP) regulation and BPV reduction in a large-animal model, incorporating real-time safety supervision and pharmacological rescue.
Methods
Following institutional ethical approval, five female Yorkshire swine underwent intravenous esmolol infusion across five experimental days. Open-loop dose-response characterisation (n=1) informed controller design. Closed-loop control employed a multi-layer feedback architecture with adaptive modulation to accommodate evolving haemodynamic conditions and variability dynamics. An independent supervisory safety layer provided autonomous override and pharmacological counter-regulation in the event of excessive blood pressure reduction. A predefined SBP reduction band was specified. Performance metrics included time-in-range (TIR) and average real variability (ARV).
Results
In closed-loop validation experiments, representative performance demonstrated TIR of 51.7% versus 0% during the untreated phase. During steady-state regulation, ARV was reduced by 31% (0.91 vs 1.32 mmHg), confirming simultaneous target attainment and beat-to-beat variability suppression. Diastolic blood pressure and heart rate decreased by 6.0% and 9.1% respectively, remaining within safe limits. In dedicated safety stress-testing, four autonomous SBP threshold violations triggered immediate pump suspension and drug rescue with successful controller re-engagement, without operator input.
Conclusions
A closed-loop intravenous esmolol system achieved autonomous SBP regulation within a predefined target band while simultaneously reducing beat-to-beat BPV, with preserved secondary haemodynamic stability and reliable automated safety responses in a large-animal model. These findings provide proof-of-concept for device-based autonomous hypertension and BPV control warranting translation to human studies.