A non-invasive method to monitor respiratory muscle effort during mechanical ventilation.

Purpose: This study introduces a method to non-invasively and automatically quantify respiratory muscle effort (P_mus) during mechanical ventilation (MV). The methodology hinges on numerically solving the respiratory system's equation of motion, utilizing measurements of airway pressure (P_aw) and airflow (F_aw). To evaluate the technique's effectiveness, P_mus was correlated with expected physiological responses. In volume-control (VC) mode, where tidal volume (V_T) is pre-determined, P_mus is expected to be linked to P_aw fluctuations. In contrast, during pressure-control (PC) mode, where P_aw is held constant, P_mus should correlate with V_T variations.

Methods: The study utilized data from 250 patients on invasive MV. The data included detailed recordings of P_aw and F_aw, sampled at 31.25 Hz and saved in 131.1-second epochs, each covering 34 to 41 breaths. The algorithm identified 51,268 epochs containing breaths on either VC or PC mode exclusively. In these epochs, P_mus and its pressure-time product (P_musPTP) were computed and correlated with P_aw's pressure-time product (P_awPTP) and V_T, respectively.

Results: There was a strong correlation of P_musPTP with P_awPTP in VC mode (R² = 0.91 [0.76, 0.96]; n = 17,648 epochs) and with V_T in PC mode (R² = 0.88 [0.74, 0.94]; n = 33,620 epochs), confirming the hypothesis. As expected, negligible correlations were observed between P_musPTP and V_T in VC mode (R² = 0.03) and between P_musPTP and P_awPTP in PC mode (R² = 0.06).

Conclusion: The study supports the feasibility of assessing respiratory effort during MV non-invasively through airway signal analysis. Further research is warranted to validate this method and investigate its clinical applications.