Hypotension prediction index: comparison between invasive and non-invasive pressure inputs

Area of Science:

• Anesthesiology
• Critical Care Medicine
• Cardiovascular Monitoring

Background:

• The Hypotension Prediction Index (HPI) aids in reducing intraoperative hypotension and perioperative complications.
• HPI can be derived from both invasive (Acumen™) and non-invasive (ClearSight™) blood pressure monitors.
• Limited concurrent comparisons exist between HPI values from these two sources.

Purpose of the Study:

• To compare HPI values and predictive performance using invasive and non-invasive blood pressure inputs.
• To enhance clinician confidence in non-invasive hemodynamic monitoring.
• To explore expanded applications and improved patient outcomes through reliable non-invasive monitoring.

Main Methods:

• Simultaneous hemodynamic data acquisition using two HemoSphere monitors with HPI software.
• Comparison of non-invasive (ClearSight™) HPI data against invasive (Acumen™) data.
• Statistical analyses included Bland-Altman, Spearman correlation, concordance, ROC analysis, and HPI alert agreement.

Main Results:

• 6,862 paired data points from 36 patients were analyzed.
• Strong correlation (r=0.76) and concordance (74%) between invasive and non-invasive HPI values.
• Similar predictive performance for hypotension (AUC invasive: 0.883, non-invasive: 0.860) and high alert agreement (86.3%).

Conclusions:

• HPI values derived from invasive and non-invasive pressure inputs are comparable.
• The predictive performance of HPI for hypotension is similar regardless of the pressure source.
• Findings support the clinical utility of non-invasive HPI monitoring.