Intraoperative hypotension (IOH) is a common and serious complication during surgery, closely associated with poor postoperative outcomes. Traditionally, anesthesiologists rely on real-time physiological parameters and alarms to monitor blood pressure, but the low alarm thresholds may lead to delayed interventions. The Hypotension Prediction Index (HPI) is a novel predictive tool that uses arterial waveform signals and advanced algorithms to forecast hypotensive events in advance. Recent observational studies have shown that HPI's accuracy in predicting hypotension is highly consistent with setting the physiological monitor's alarm threshold to 73 mmHg. This study will compare the effectiveness of HPI and a raised alarm threshold of 73 mmHg in preventing IOH. While HPI is promising with its AI-assisted approach to patient care, its high cost due to the advanced technology raises concerns. If its accuracy is comparable to simply raising the traditional monitor threshold, it may not lead to substantial changes in clinical practice.

Intraoperative hypotension (IOH) is a significant complication that affects surgical patients, potentially leading to adverse outcomes postoperatively. Standard practices involve relying on monitoring devices with low alarm thresholds for blood pressure, which may result in delayed interventions. The Hypotension Prediction Index (HPI) offers a predictive approach by analyzing arterial waveform signals and using complex algorithms to detect potential hypotensive episodes early. Recent observational studies have suggested that HPI's accuracy in predicting hypotension aligns closely with raising the physiological monitor alarm threshold to 73 mmHg. To further investigate this, this study will compare the effects of setting a traditional monitor alarm threshold at 73 mmHg with using HPI to prevent IOH.

In this study, patients will be randomly assigned to two groups. In the HPI group, interventions will be initiated when the HPI value exceeds 85. These interventions will follow a protocol that includes fluid administration, norepinephrine, and dobutamine to prevent hypotension. The control group will have their alarm threshold set at 73 mmHg. For these patients, interventions will be based on stroke volume variation (SVV) and clinical judgment, utilizing fluid and norepinephrine as needed. HPI is an attractive AI-based tool for medical care, but its high cost due to advanced technology raises questions. If its accuracy proves to be similar to simply raising the alarm threshold to 73 mmHg, it may not lead to meaningful changes in clinical practice. The study aims to compare the efficacy of these two methods in reducing the incidence of IOH.

Inclusion Criteria:

• A: Patients undergoing surgeries requiring general anesthesia lasting more than two hours, and requiring continuous arterial blood pressure monitoring via arterial catheter according to standard medical practice. This includes:

ASA Class II or higher. Estimated surgery duration of three hours or more. High cardiovascular risk, such as poorly controlled hypertension, diabetes, coronary artery disease, chronic kidney disease, or chronic emphysema.

• B: Patients aged 18 years or older.

Exclusion Criteria:

• ASA Class I: Patients with mild systemic disease.
• Pregnancy: Pregnant women.
• End-stage renal disease: Patients with eGFR below 30 ml/min/1.73 m².
• Cardiac shunt: Presence of intracardiac shunt.
• Severe arrhythmias: Including supraventricular tachycardia (heart rate >100 bpm), ventricular tachycardia, or ventricular fibrillation.
• Factors affecting SVV accuracy: Conditions such as atrial fibrillation (A-Fib) or thoracic surgery that can invalidate stroke volume variation (SVV) measurements.