Blood pressure is the main determinant of organ perfusion. Hypotension is common in patients having surgery and in critically ill patients. The severity and duration of hypotension are associated with hypoperfusion and organ dysfunction. Hypotension is mostly treated reactively after low blood pressure values have already occurred. However, prediction of hypotension before it becomes clinically apparent would allow the clinician to treat hypotension pre-emptively, thereby reducing the severity and duration of hypotension. Hypotension can now be predicted minutes before it actually occurs from the blood pressure waveform using machine-learning algorithms that can be trained to detect subtle changes in cardiovascular dynamics preceding clinically apparent hypotension. However, analyzing the complex cardiovascular system is a challenge because cardiovascular physiology is highly interdependent, works within complicated networks, and is influenced by compensatory mechanisms. Improved hemodynamic data collection and integration will be a key to improve current models and develop new hypotension prediction models.
Keywords: artificial intelligence; blood pressure; cardiovascular dynamics; hemodynamic monitoring; hypotension prediction index; machine learning.
Copyright © 2019 Elsevier Ltd. All rights reserved.