Hypotension in ICU Patients Receiving Vasopressor Therapy

Area of Science:

• Critical Care Medicine
• Biomedical Engineering
• Health Informatics

Background:

• Vasopressor infusion (VPI) is a critical intervention for managing hypotension in intensive care units (ICUs).
• Adherence to blood pressure (BP) targets during VPI is essential for patient outcomes.
• Predicting and preventing hypotension episodes remains a clinical challenge.

Purpose of the Study:

• To assess compliance with BP goals during VPI.
• To evaluate the efficacy of a statistical model for early hypotension prediction compared to a standard threshold alert.
• To analyze the characteristics of hypotension episodes during VPI.

Main Methods:

• Retrospective analysis of adult ICU patient data receiving VPI at submaximal dosages.
• Development of a logistic regression model using BP trend features to predict sustained hypotension (>15 min).
• Validation of the predictive model using prospective data.

Main Results:

• Hypotension episodes were frequent, occurring in over 47% of retrospective ICU stays.
• The logistic regression model achieved high detection rates (99.6% retrospective, 100% prospective) with a median advance forecast time (AFT) of 12 minutes (retrospective) and 22 minutes (prospective).
• The basic hypotension threshold alert had a 100% detection rate but a 0-minute AFT.

Conclusions:

• Hypotension is a common complication during vasopressor infusion in the ICU.
• A logistic regression model incorporating BP temporal trends offers significant advance warning of impending hypotension.
• This predictive model has the potential to improve clinical management and patient outcomes during VPI.