Using machine learning to guide blood pressure treatment during surgery

Machine Learning-Assisted Intraoperative Hypotension Management: Developing Personalized Treatment Recommendations

This project will try a machine-learning tool to help tailor blood pressure treatments for adults having elective surgery under general anesthesia with invasive arterial monitoring.

Quick facts

• Study type: Observational
• Enrollment: 50 (estimated)
• Ages: 18 Years and up
• Sex: All
• Sponsor: Nevsehir Public Hospital (other gov)
• Locations: 1 site (Konya, Konya/Meram)
• Trial ID: NCT07396636 on ClinicalTrials.gov

What this trial studies

This is a prospective observational effort that collects continuous invasive arterial pressure waveforms and perioperative clinical data from adults undergoing elective surgeries of at least two hours under general anesthesia. Machine learning models will analyze waveform and clinical features to classify likely physiological causes of intraoperative hypotension (for example, low blood volume, vasodilation, or reduced heart contractility) and generate personalized decision-support information. Clinicians will continue routine care while the system's outputs are recorded to see how phenotype-based guidance relates to chosen treatments. The study links these physiologic phenotypes and treatment patterns to short-term outcomes such as organ injury and length of stay.

Who should consider this trial

Good fit: Adults aged 18 or older scheduled for elective surgical procedures under general anesthesia with planned continuous invasive arterial blood pressure monitoring and an expected surgery length of two hours or more who can give informed consent.

Not a fit: Patients having emergency surgery, those with sepsis or advanced cardiogenic shock, uncontrolled rhythm disturbances like atrial fibrillation, severe left ventricular dysfunction (EF <30%), or those who cannot maintain an arterial line or provide consent are unlikely to benefit from participation.

Why it matters

Potential benefit: If successful, the approach could help clinicians choose more targeted blood pressure treatments during surgery, potentially reducing organ injury and postoperative complications.

How similar studies have performed: Prior machine-learning tools have shown promise in predicting intraoperative hypotension in limited settings, but using ML to distinguish physiological causes and guide personalized treatment is relatively novel and not yet proven in large interventional trials.

Eligibility criteria

Inclusion Criteria:

* Adult patients aged 18 years or older
* Patients scheduled for elective surgical procedures
* Procedures performed under general anesthesia
* Availability of continuous invasive arterial blood pressure monitoring during the intraoperative period
* Planned surgical duration of at least 2 hours
* Ability to provide written informed consent for participation in the study

Exclusion Criteria:

* Emergency surgery
* Diagnosis of sepsis, septic shock, or advanced cardiogenic shock
* Cardiac rhythm disturbances that prevent reliable hemodynamic measurements (e.g., atrial fibrillation)
* Severe left ventricular dysfunction or advanced heart failure (ejection fraction \<30%)
* Inability to maintain intraoperative arterial cannulation due to technical or clinical reasons
* Inability to provide informed consent (e.g., cognitive impairment or refusal to participate)

Where this trial is running

Konya, Konya/Meram

• Konya City Hospital — Konya, Konya/Meram, Turkey (Türkiye) (RECRUITING)

Study contacts

• Study coordinator: Mehmet Akif Yazar, MD, PhD, Assoc. Prof.

Conditions: Intraoperative Hypotension, Perioperative Care, Hemodynamic Monitoring, Blood Pressure, Anesthesia, General, Intraoperative hypotension, Machine learning, Personalized treatment