Intraoperative impedance and electrogram features can predict chronic capture threshold in active fixation ventricular leadless pacemakers

Abstract Background Leadless pacemakers (LP) with the capability to obtain electrical measurements such as impedance and electrograms (EGM) can provide early feedback on implant site selection, before committing to fixation and therefore minimizing the need to reposition the LP. Purpose The objective was to utilize intraoperative features of the electrogram (EGM) and paced impedance measurements to predict pacing capture thresholds (PCT) at the 3-month follow-up. Methods This is a retrospective study of a leadless pacemaker clinical trial (NCT#:05252702), including patients with complete sets of impedance measurements and intracardiac EGMs collected during the mapping phase and while in tether mode, and a capture threshold obtained at the 3-month follow-up. A computerized algorithm was developed to quantify features of the EGM signal: amplitudes of the R-wave, S-wave, and COI, the slope of the upstroke and downstroke, and the sharpness of the R-wave peak (calculated as average slope of points within 1 sample of the peak). Linear regression was performed to identify significant predictors of the chronic PCT. Binary logistic regression models were constructed by converting the 3-month PCT into a binary outcome using a cutoff of 1.5V and analyzed using receiver operating characteristic (ROC) curves. Results 88 patients were included. PCT at 3-months was 0.73±0.84 V. 8 patients had PCT >1.5V at 3 months. In univariate linear regression, impedance during mapping and tether, the sharpness of the R-wave during mapping, and the R-wave amplitude during tether were significant predictors of 3-month PCT (p=0.04, <0.01, 0.05, 0.03, respectively). Two logistic regression models were identified: 1) using only mapping variables (COI and impedance), 2) including both mapping COI and tether impedance. The mapping logistic regression model included COI (p=0.01) and impedance (p=0.1) during mapping and produced an area under the curve (AUC) of 0.88 with sensitivity and specificity of 100% and 70%, respectively. A logistic regression model including COI (mapping, p=0.04) and impedance (tether, p=0.03) produced an AUC of 0.92 with sensitivity and specificity of 100% and 81%, respectively. Test of the Χ2 statistic vs. constant model had p<0.01 in both models. Conclusion We developed a computerized prediction model using intraoperative EGM and impedance to predict 3-month PCT of a leadless pacemaker. This may be useful in enhancing procedural efficacy and efficiency.Linear Regression Results Binary Logistic Regression ROC Curves