Abstract Background Cardiac resynchronization therapy (CRT) is considered the standard treatment for patients with heart failure and prolonged QRS duration. Recent studies suggest that atrioventricular delay (AVD) optimization can play a key role in QRS narrowing, thus improving CRT outcomes. For this purpose, device-based algorithms have been developed. In particular, the SyncAV algorithm continuously monitors the patient’s intrinsic AVD and dynamically adjusts the pacing AVD to allow fusion of the left ventricle pacing with intrinsic right ventricle depolarization. SyncAV optimization is usually performed in a single patient condition, i.e. at rest, and in a single operating scenario of the device, i.e. during atrial sensing. Purpose This study aims to compare QRS narrowing after AVD optimization using SyncAV in CRT-treated patients at rest, during moderate exertion, and during an increase in the atrial pacing rate. Methods This is a prospective, non-randomized, non-blinded, multicenter study. The study population consisted of patients in sinus rhythm, PR interval < 350 ms, and rest ventricular rate < 100 bpm. Within one month from the implant, patients were assessed for AVD optimization with SyncAV, selecting the offset (i.e. 10 to 40% AVD) that resulted in the shortest QRS duration (QRSd). Subsequently, patients performed moderate physical activity using a bicycle ergo-meter while a surface ECG was collected. QRSd was assessed once patients achieved a heart rate over 20 bpm to baseline heart rate, up to 100 bpm. Finally, QRSd narrowing was evaluated during a programmed increase in atrial pacing (i.e. >20 bpm of baseline atrial rate). Results Nine patients were treated with CRT and enrolled (age 67,3 ± 13,2 years; 66% male; LV ejection fraction 33,7±2,9 %; intrinsic QRSd 155,6 ± 12,7 ms) with intact atrioventricular conduction (PR interval 170,6±25,8 ms). The median SyncAV offset that provided the best QRSd reduction was 20 % (10 - 35%). Intrinsic QRSd was reduced to 132,7 ± 10,0 ms (-14,2±9,4%) with standard biventricular pacing and to 122,0 ± 11,2 ms (-21,2±8,5%) with SyncAV optimization. During exercise the mean heart rate was 88,1 bpm, the QRSd was reduced to 132,7 ± 10,0 ms (-14,2±9,4%) with standard biventricular pacing and to 124,9 ± 14,9 ms (-19,3±10,8 %) with SyncAV optimization. The mean heart rate during atrial pacing was 90 bpm with paced QRSd of 126,0 ±13,9 ms (-18,9±7,3% from baseline) with SyncAV optimization and paced PR interval of 198,2±22,7ms. Conclusion We evaluated the effectiveness of SyncAV for AVD optimization in out-of-clinic. This study demonstrates the possibility of narrowing the QRSd using SyncAV at several activity levels. Nevertheless, a deeper evaluation is needed to confirm our findings, investigating the SyncAV role in a larger population of patients.