Automated characterization of patient-ventilator interaction using surface electromyography

Area of Science:

• Critical Care Medicine
• Biomedical Engineering
• Respiratory Physiology

Background:

• Characterizing patient-ventilator interaction in critically ill patients is crucial but time-consuming.
• Manual evaluation requires trained staff and can be subjective.

Purpose of the Study:

• To assess the feasibility of automating the quantification of patient-ventilator asynchrony using noninvasive surface electromyography (sEMG).
• To compare the accuracy and reliability of two algorithms for detecting inspiratory effort and classifying asynchronies.

Main Methods:

• Recorded sEMG from diaphragm and intercostal muscles, and esophageal pressure in mechanically ventilated ARDS patients.
• Utilized triangle and adaptive thresholding algorithms to automatically detect inspiratory effort and classify major asynchronies.
• Validated algorithmic results against manual annotations by two experts.

Main Results:

• Spontaneous breathing activity was detected in 22 out of 36 patients.
• Both algorithms demonstrated reliable detection performance with high sensitivity and positive predictive values against expert annotations.
• Automatic asynchrony index prediction showed reliable results, comparable to manual assessments.

Conclusions:

• Automating patient-ventilator asynchrony quantification using noninvasive sEMG is feasible in critically ill patients.
• This approach can facilitate more frequent diagnosis of asynchrony.
• Improved patient-ventilator interaction and clinical decision-making may be supported.