Iterative interpolative pressure reconstruction for improved respiratory mechanics estimation during asynchronous volume controlled ventilation

Abstract

Asynchronous events (AEs) during mechanical ventilation (MV) breathing support can lead to poor respiratory mechanics estimation, as the patient’s attempts to breath affects the measured airway pressure and flow. An algorithm that allows improved model-based estimation of respiratory system elastance, Ers during asynchronous volumecontrolled MV was developed. This method reconstructs a pseudo airway pressure waveform for each breath, that is similar to a breath that was unaffected by asynchronous efforts. The reconstructed waveforms can be used to estimate true respiratory system mechanics. To test the proposed algorithm, 10 retrospective airway pressure and flow datasets were obtained from 6 MV patients. Each dataset contains 475-500 breaths. Of the 9/10 datasets which contained AEs, 8 experienced a decrease in Ers mean absolute deviation (MAD) and the 5th-95th range (Range90) after pressure reconstruction. The median [maximum (max), minimum (min)] decrease in Range90 divided by median elastance, was 51.3% (67.4%,-16.7%). Additionally, the median elastance for reconstructed breaths in these datasets moved closer to the true, nonasynchronous, elastance value. The median elastance change was 48.7% closer towards the true value, with a maximum shift of 93.4%. The one dataset which did not experience an improvement was found to have a varying pressure amplitude indicative of external factors affecting the MV treatment, rather than a deficiency in the pressure reconstruction. The algorithm demonstrates the ability to consistently enhance elastance estimation in MV patients.