Intraoperative hemodynamic patterns during hypotension prediction index-guided care: a high-resolution secondary analysis of the HYT trial.

BACKGROUND: Although the Hypotension Prediction Index (HPI) reduces intraoperative hypotension, improvements in patient-centered outcomes remain inconsistent. We performed a high-resolution signal analysis to describe the coexistence of arterial pressure, preload-related hemodynamic signals, and algorithm-derived hypotension risk during surgery. METHODS: We conducted a retrospective high-resolution signal analysis (20-second intervals) of 181 patients from the multicenter HYT trial. We quantified hypotension (mean arterial pressure [MAP] <65 mmHg), preload-dependent physiology (stroke volume variation [SVV] >13%), and their temporal relationship with HPI-derived risk signals. Patients were classified into four descriptive groups based on combinations of MAP and SVV burden. RESULTS: Overall hypotension exposure was low (median 1.3% of monitored time), whereas preload-dependent physiology was more frequent (approximately 10%) and often persisted during periods classified as low hypotension risk (HPI <80). SVV >13% preceded hypotensive events significantly earlier than HPI alerts (median 8.7 vs. 1.8 minutes; P=0.0119). Higher vasopressor burden was observed in patients with preserved MAP despite sustained SVV elevation. CONCLUSIONS: During HPI-guided care, preserved arterial pressure frequently coexists with sustained preload-dependent hemodynamic signals. These findings describe patterns of pressure and preload-related signal behavior during algorithm-guided hemodynamic management and should be interpreted as hypothesis-generating.