Prediction of lung mechanics throughout recruitment maneuvers in pressure-controlled ventilation. (December 2020)

• Record Type: Journal Article
• Authors: Morton, Sophie E.; Knopp, Jennifer L.; Tawhai, Merryn H.; Docherty, Paul; Heines, Serge J; Bergmans, Dennis C.; Möller, Knut; Chase, J. Geoffrey
• Is Part Of: Computer methods and programs in biomedicine. Volume 197(2020)
• Publication Date: 2020-12
• Subjects: Mechanical ventilation -- Pressure-controlled ventilation -- Recruitment maneuvers -- Respiratory mechanics -- Virtual patients -- Personalised care -- Prediction; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2020.105696
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: Clinical studies have shown excessive pressure and volume in mechanical ventilation leads to unintended lung injury. Development and validation of a predictive in-silico models is presented. This model predicts peak volumes accurately in commonly-used pressure controlled ventilation. Clinical use of virtual patients could improve clinician confidence and patient safety in delivering care. Abstract: Mechanical ventilation (MV) is a core therapy in the intensive care unit (ICU). Some patients rely on MV to support breathing. However, it is a difficult therapy to optimise, where inter- and intra- patient variability leads to significantly increased risk of lung damage. Excessive volume and/or pressure can cause volutrauma or barotrauma, resulting in increased length of time on ventilation, length of stay, cost and mortality. Virtual patient modelling has changed care in other areas of ICU medicine, enabling more personalized and optimal care, and have emerged for volume-controlled MV. This research extends this MV virtual patient model into the increasingly more commonly used pressure-controlled MV mode. The simulation methods are extended to use pressure, instead of both volume and flow, as the known input, increasing the output variables to be predicted (flow and its integral, volume). The model and methods are validated using data from N = 14 pressure-control ventilated patients during recruitment maneuvers, with n = 558 prediction tests over changes of PEEP ranging from 2 to 16 cmH2 O. Prediction errors for peak inspiratory volume for an increase of 16 cmH2 O were 80 [30 - 140] mL (15.9 [8.4 - 31.0]%), with RMS fitting errors of 0.05 [0.03 - 0.12] L. These results show very good prediction accuracy able to guide personalised MV care.