November 29, 2025
Background: Intraoperative hypotension prediction has gained increasing attention because of its potential to reduce organ injury and the availability of large patient datasets and advanced machine learning techniques. Although prediction methods can reduce the duration of hypotension, they have not yet been convincingly shown to improve objective clinical outcomes and remain a subject of ongoing debate.
Review: This review summarizes the current state of intraoperative hypotension prediction and provides recommendations for future research. Existing prediction methods generally rely on the current mean arterial pressure as a key input variable and typically demonstrate good sensitivity and specificity but relatively low positive predictive value for forecasting short-term hypotensive events. The authors recommend improving the definition of acute hypotensive events and refining methods used to evaluate prediction models. They also propose extending prediction approaches to estimate reductions in blood flow and predict treatment responses. In addition, the review presents an initial risk-benefit analysis of implementing hypotension prediction methods in routine clinical practice.
Conclusions: Future research should focus on developing patient-specific hypotension prediction models and methods that not only predict hypotension but also identify the most appropriate treatment response. Such personalized approaches are considered the most promising strategy for improving perioperative outcomes.
©2026 OR. All Rights Reserved
