Abstract
BACKGROUND: Mechanical ventilation is essential for treating respiratory failure. However, ventilator over-assistance can lead to ventilator-induced diaphragm dysfunction (VIDD), and insufficient assistance can necessitate excessive effort, which leads to high stress and damage diaphragm function. Current methods monitoring respiratory effort face clinical implementation challenges due to invasiveness and complexity. This study introduces and validates a novel non-invasive real-time respiratory muscle pressure (N-Pmus) monitoring method. METHODS: (1) The bench study involved developing a non-invasive, real-time respiratory muscle pressure monitoring algorithm (N-Pmus) based on the equation of motion of the respiratory system and validated with efforts generated by the ASL simulator (ASL5000-Pmus) across 270 clinical scenarios. (2) A clinical validation was performed through a prospective observational study (n = 23), comparing N-Pmus with the Pmus derived from simultaneously monitored esophageal pressure (Pmus(eso)) to assess correlation and agreement. The association between N-Pmus and the established Pmus benchmarks was analyzed using linear mixed-effects models. Bias and agreement were evaluated through Bland-Altman analysis for repeated measures. RESULTS: (1) The bench study demonstrated that N-Pmus correlated well with ASL5000-Pmus, with marginal R²=0.993 and conditional R²=0.997. The bias was - 0.23 cmH₂O, with limits of agreement ranging from - 1.51 to 1.04 cmH₂O. (2) The clinical validation revealed strong N-Pmus/Pmus(eso) agreement with marginal R²=0.97 and conditional R²=0.971. The bias was - 0.2 cmH₂O, with limits of agreement ranging from - 2.22 to 1.83 cmH₂O. CLINICAL TRIAL NUMBER: Retrospectively registered with https://www.chictr.org.cn/ . REGISTRATION NUMBER: ChiCTR2300076940, registered 24 October 2023.