Artificial intelligence-based predictive hemodynamic monitoring in conjunction with goal-directed therapy reduces duration, frequency, and severity of intraoperative hypotension in major maxillofacial and otolaryngological surgery-a prospective randomized controlled pilot trial.

BACKGROUND

Intraoperative hypotension (IOH) during non-cardiac surgery is associated with increased risk of postoperative complications, including acute kidney injury, myocardial injury, stroke, and mortality. Artificial intelligence-based predictive hemodynamic monitoring using the Hypotension Prediction Index (HPI), combined with goal-directed therapy (GDT), has been proposed to reduce IOH. However, its effectiveness in major maxillofacial and otolaryngologic surgery remains unclear. The purpose of the study was to assess whether HPI-guided management or classical GDT reduces IOH compared to standard care in patients undergoing major maxillofacial and otolaryngologic surgery.

METHODS

In this randomized controlled pilot trial at a university hospital, 75 patients were allocated to one of three groups: control (n = 25), HPI-guided GDT (n = 25), or classical GDT without HPI (n = 25). In the control group, the advanced hemodynamic monitoring was blinded to the anesthesiologist. IOH was defined as mean arterial pressure (MAP) < 65 mmHg for > 1 min. Primary endpoints were the number and total duration of IOH episodes. Secondary endpoints included the time-weighted average MAP < 65 mmHg (TWA65) and postoperative complications.

RESULTS

Seventy-four patients were analyzed. The HPI group showed significantly fewer IOH episodes (median 3.0 vs. 7.0; p = 0.02) and shorter IOH duration (7.0 min vs. 46.0 min; p < 0.01) compared to control. No significant difference was observed between the classical GDT and control groups. Secondary outcomes were comparable across all groups.

CONCLUSIONS

HPI-guided hemodynamic management significantly reduces the frequency and duration of IOH in major head and neck surgery. Larger studies are needed to evaluate effects on clinical outcomes.

TRIAL REGISTRATION: The trial was registered on clinicaltrials.gov (NCT04151264) on 14th October 2019.