Deidentified data comparing esophageal and model computed catheter data.

Purpose. Accurate assessment of respiratory muscle effort during mechanical ventilation is essential to prevent patient self-inflicted lung injury. Esophageal manometry is the clinical reference standard for estimating respiratory muscle pressure (Pmus) and its pressure–time product (PmusPTP), but its use is limited by invasiveness. We evaluated whether PmusPTP estimated non-invasively from airway pressure (Paw) and flow (Faw) signals shows close agreement with esophageal manometry.
Methods. Prospective, validation, observational study of mechanically ventilated adults who underwent simultaneous digital recording of Paw, Faw and esophageal pressure (Peso) after intubation. Patient-specific chest wall recoil pressure (Pcw) was determined during verified passive ventilation and a Pmus(t) function was calculated as Pcw(t) – Peso(t), referenced to end-expiration, for each recorded breath. We computed respiratory system mechanics non-invasively by numerical analysis of Paw and Faw signals and determined Pmus(t) with the classical one-compartment model of the respiratory system. Insufflation PmusPTP calculated by the model and by Peso was compared using linear regression and Bland–Altman agreement analysis.
Results. We analyzed 1586 individual breaths from ten patients. Model calculated and Peso-derived Pmus(t) waveforms closely matched across a wide range of respiratory efforts. Agreement between methods was robust at the individual breath and epoch levels, with R² of 0.85 and 0.92, respectively. Bias and limits of agreement were −0.05 ± 1.9 cmH₂O·s for individual breaths and −2.7 ± 14.6 cmH₂O·s·min⁻¹ for epoch-level data.
Conclusions. Estimates of PmusPTP based on airway signals agree with esophageal manometry and may enable a non-invasive approach to monitoring respiratory muscle effort during mechanical ventilation.