Human pressure-volume loops from the impedance catheter: Development and validation of semi-automated beat-by-beat analysis

<ns3:p> Introduction Pressure-volume (PV) loops offer a comprehensive evaluation of cardiac function. Impedance catheters enable the acquisition of synchronised intracardiac electrocardiogram (ECG), pressure, and volume data with high temporal resolution. However, current calibration methods are impractical and data interpretation is often inconsistent. Methods In the PREFER-CMR prospective, cohort study, 15 patients with suspected heart failure and preserved ejection fraction underwent same-day cardiac magnetic resonance (CMR) imaging and invasive impedance catheter studies. Signal processing algorithms were developed to semi-automatically determine PV-loop phases and calibrate impedance catheter volumes to CMR. Results of beat-by-beat and average loop analysis approaches were compared with reference methods and between each other. Results The second-order differential of the pressure-volume trace identified PV-loop phases on a beat-by-beat basis, but gradient smoothing prevented detection in average loops. Calibrated impedance catheter volumes, including left ventricular end diastolic (LVEDV) and end systolic (LVESV) volumes, correlated with CMR (r≥0.95, p<0.001) using both analysis methods. However, the average loop LVESV was overestimated by 8.1ml (p=0.031). For left ventricular end diastolic pressure, both beat-by-beat (r=0.73, p=0.002) and average loop (r=0.69, p=0.005) methods correlated with the fluid-filled manometer reference. Maximum pressure correlation was strong for both beat-by-beat (r=0.85, p<0.001) and average loop (r=0.80, p<0.001) methods, but was 10.1mmHg (p=0.040) lower in the average loop method. Between methods, significant correlations (r=0.73–0.99) were found across all pressures and volumes. Stroke work (r=0.94) and potential energy (r=0.96) significantly correlated (p<0.001) between methods, although Bland-Altman subgroup analysis suggested underestimation of stroke work in atrial fibrillation using the average loop method. Conclusions Impedance catheter volumes can be accurately calibrated using CMR. PV-loop phases can be robustly detected with a semi-automated algorithm. Both beat-by-beat and average loop approaches are viable for analysing multiple cardiac cycles, though beat-by-beat analysis may offer advantages for phase identification, pressure assessment, and in irregular rhythms. Trials registration ClinicalTrials.gov: NCT05114785. Registration date: 05/11/2021. <ns3:ext-link xmlns:ns4="http://www.w3.org/1999/xlink" ext-link-type="uri" ns4:href="https://clinicaltrials.gov/study/NCT05114785">https://clinicaltrials.gov/study/NCT05114785</ns3:ext-link> </ns3:p>