A30-15 Electrical Impedance Tomography Reveals a Consistent Physiologic Delay Between Mechanical and Aeration Signals During Continuous Positive Airway Pressure

Abstract Rationale Continuous positive airway pressure (CPAP) is widely used for sleep disordered breathing, acute hypoxemic respiratory failure, and perioperative lung protection. Bedside titration typically relies on global mechanical surrogates such as airflow, airway pressure, or chest wall motion, that do not reveal regional lung aeration. Electrical impedance tomography (EIT) enables noninvasive, radiation free imaging of aeration but its temporal relationship to mechanical signals remains unclear. We sought to quantify the delay (Δt) between mechanical end inspiration and EIT derived aeration onset in healthy adults receiving CPAP, and to determine whether subject characteristics influence this delay. Methods We retrospectively analyzed 79 de-identified subjects from the publicly available PhysioNet dataset “Respiratory Dataset from PEEP Study with Expiratory Occlusion” (University of Canterbury HREC 2023/04/LR-PS). Mechanical channels (airway pressure, differential flow, chest and abdominal motion) were sampled at 100 Hz; EIT frames (32 × 32 pixels) at 50 Hz using a Dräger PulmoVista 500 system. Data were synchronized via 5 V TTL hardware triggers and verified by cross-correlation of the first shared inspiratory peak. “Zero Time Main” was defined as the time of minimal flow following inspiration; “Zero Time EIT” as the first decline in global impedance after peak aeration. The difference Δt = EIT - Main represented the aeration lag. Linear, logistic, and penalized (LASSO) regression models assessed associations with age, sex, anthropometrics, asthma, and smoking/vaping status. Results EIT consistently lagged behind mechanical signals by 1.74 ± 0.91 s (paired t = 16.9, p < 1 × 101²27; Wilcoxon p < 1 × 10−114). Zero time correlations were strong (r = 0.85) (Fig. 1). Logistic regression dichotomizing Δt at its median yielded AUC = 0.77, indicating moderate discriminative performance. Neither sex, age, height, weight, asthma, nor smoking/vaping significantly predicted Δt (all p > 0.1). Both backward stepwise OLS and 10-fold LASSO cross-validation eliminated all predictors (final R² = 0.026). Conclusions EIT reveals a reproducible ∼1.7 s physiologic lag between mechanical and aeration events during CPAP that is independent of demographic or clinical factors. This aeration lag likely reflects continued alveolar filling after bulk flow cessation and provides a potential non-invasive biomarker of lung recruitment dynamics. Establishing this healthy reference may guide future studies using EIT to detect impaired regional ventilation or optimize CPAP and PEEP settings in patients with respiratory disease. This abstract is funded by: None