Ambulatory AI-based Holter monitoring for delayed high-degree AV block after TAVI among high-risk patients

Background Permanent pacemaker implantation (PPMI) occurs in approximately 15% of patients within 30 days post-TAVI, with up to 30% of cases occurring after hospital discharge. Delayed high-grade atrioventricular block (DH-AVB) poses a significant clinical challenge as hospitalization durations shorten. While baseline risk factors including RBBB and new-onset conduction disturbances predict increased risk, their sensitivity for DH-AVB remains low. Current guidelines lack specific management strategies for high-risk patients. Purpose To evaluate the diagnostic utility of AI-based ambulatory electrocardiographic monitoring for early detection of DH-AVB in high-risk post-TAVI patients, identify clinical and procedural predictors of delayed conduction disorders and establish optimal surveillance strategies for preventing sudden cardiac complications. Methods This prospective, single-center observational study enrolled post-TAVI patients with high-risk conduction features: pre-procedural RBBB, post-procedural new LBBB (>24 hrs), PR interval >240ms (>24 hrs), QRS/PR prolongation ≥20ms, or transient peri-procedural high-degree AVB (HDAVB). Patients with pre-existing pacemakers or meeting criteria for PPMI were excluded. Each eligible patient underwent 14-day ambulatory electrocardiographic monitoring after hospital-discharge using a commercial AI-based Holter system with automated trigger detection for HDAVB, severe bradycardia, and conduction interval prolongation. Patients demonstrating development of HDAVB or high-risk features were recalled for urgent evaluation for PPMI. Results Over 9 months, 143 patients underwent TAVI (mean age 78.2±8.9 years, 49.6% female). Following exclusion of 15 patients with pre-existing pacemakers and 39 requiring immediate post-procedural PPMI (including 7 following an EPS), 34 high-risk patients were enrolled and monitored with AI-based Holter surveillance for 14 days post-discharge. Six patients (17.6%) required PPMI: five during the surveillance period due to HDAVB (n=1), conduction disorder progression (n=3), and symptomatic 2nd-degree AVB Mobitz I (n=1), and one patient at 30 days due to symptomatic sinus node dysfunction. No patients experienced sudden cardiac complications including syncope or cardiac arrest during or after Holter surveillance (median follow-up 160 days), and no late PPMI (>1 month) was required. Conclusion AI-guided ambulatory monitoring proved reliable and safe for post-TAVI surveillance, with no adverse events observed. This system provides reassurance for both patients and healthcare providers following hospital discharge. All patients requiring intervention were identified during the monitoring period, preventing potential sudden cardiac complications. Continued enrolment is planned to achieve adequate statistical power for identifying predictors of delayed PPMI. AI-based ambulatory monitoring may represent a valuable addition to routine post-TAVI electrical surveillance.