Prediction of Hemodynamic Instability in Patients Undergoing Surgery

Intraoperative hypotension occurs often and is associated with adverse patient outcomes such as stroke, myocardial infarction and renal injury.

The aim of this study was to test the accuracy of a physiology-based machine-learning algorithm using continuous non-invasive measurement of the blood pressure waveform with the Nexfin® finger cuff during surgery.

Study Overview

• Status: Completed
• Conditions: Blood Pressure; Machine Learning; Hemodynamic Instability; Prediction Models
• Intervention / Treatment: Diagnostic test: Hypotension Probability Indicator

• Study Type: Observational
• Enrollment (Actual): 507

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

All adult patients undergoing surgery in the AMC were included in the study. Subjects were only excluded when technical problems or strong local vasoconstriction (i.e., cold fingers) prevented the Nexfin® non-invasive blood pressure finger cuff measurement. Subjects were not excluded for any other reason besides technical failure.

Description

Inclusion Criteria:

• all adult patients undergoing surgery

Exclusion Criteria:

• none

Study Plan

How is the study designed?

Design Details

• Observational Models: Cohort
• Time Perspectives: Prospective

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Sensitivity of the HPI algorithm	Sensitivity	three minutes prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	three minutes prior to the hypotensive event

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Predictive positive value of the HPI algorithm	Predictive positive value	one minute prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	two minutes prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	three minutes prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	four minutes prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	five minutes prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	ten minutes prior to the hypotensive event
Predictive positive value of the HPI algorithm	Predictive positive value	15 minutes prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	one minute prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	two minutes prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	four minutes prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	five minutes prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	ten minutes prior to the hypotensive event
Negative predictive value of the HPI algorithm	Negative predictive value	15 minutes prior to the hypotensive event
Time from HPI alarm to hypotensive event during surgery	Time from HPI alarm to hypotensive event, this can range from 0,1 min to a high number such as 30 or even 40 minutes.	From the onset of the HPI alarm to the hypotensive event during surgery, this is in minutes. (this can range from 0,1 min to a high number such as 30 or even 40 minutes)
Sensitivity of the HPI algorithm	Sensitivity	one minute prior to the hypotensive event
Sensitivity of the HPI algorithm	Sensitivity	two minutes prior to the hypotensive event
Sensitivity of the HPI algorithm	Sensitivity	five minutes prior to the hypotensive event
Sensitivity of the HPI algorithm	Sensitivity	ten minutes prior to the hypotensive event
Sensitivity of the HPI algorithm	Sensitivity	15 minutes prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	one minute prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	two minutes prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	five minutes prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	ten minutes prior to the hypotensive event
Specifity of the HPI algorithm	Specifity	15 minutes prior to the hypotensive event

Collaborators and Investigators

• Sponsor: Academisch Medisch Centrum - Universiteit van Amsterdam (AMC-UvA)

Publications and helpful links

General Publications

• Wijnberge M, van der Ster BJP, Geerts BF, de Beer F, Beurskens C, Emal D, Hollmann MW, Vlaar APJ, Veelo DP. Clinical performance of a machine-learning algorithm to predict intra-operative hypotension with noninvasive arterial pressure waveforms: A cohort study. Eur J Anaesthesiol. 2021 Jun 1;38(6):609-615. doi: 10.1097/EJA.0000000000001521.

Study record dates

Study Major Dates

• Study Start (Actual): April 1, 2015
• Primary Completion (Actual): December 1, 2016
• Study Completion (Actual): April 26, 2018

Study Registration Dates

• First Submitted: April 26, 2018
• First Submitted That Met QC Criteria: May 22, 2018
• First Posted (Actual): May 23, 2018

Study Record Updates

• Last Update Posted (Actual): May 23, 2018
• Last Update Submitted That Met QC Criteria: May 22, 2018
• Last Verified: May 1, 2018

More Information

Terms related to this study

• Other Study ID Numbers: W15_080

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: Undecided

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No