Background
Sleep disordered breathing (SDB) may trigger nocturnal cardiac arrhythmias (NCA) in patients with heart failure with reduced ejection fraction (HFrEF). The NCA ancillary study of the ADVENT-HF trial will test whether, in HFrEF-patients with SDB, peak-flow-triggered adaptive servo-ventilation (ASVpf) reduces NCA. To this end, accurate scoring of NCA from polysomnography (PSG) is required.
Objective
To develop a method to detect NCA accurately from a single-lead electrocardiogram (ECG) recorded during PSG and assess inter-observer agreement for NCA detection.
Methods
Quality assurance of ECG analysis included training of the investigators, development of standardized technical quality, guideline-conforming semi-automated NCA-scoring via Holter-ECG software and implementation of an arrhythmia adjudication committee. To assess inter-observer agreement, the ECG was analysed by two independent investigators and compared for agreement on premature ventricular complexes (PVC) /h, premature atrial complexes/h (PAC) as well as for other NCA in 62 patients from two centers of the ADVENT-HF trial.
Results
The intraclass correlation coefficients for PVC/h and PAC/h were excellent: 0.99 (95%- confidence interval [CI]: 0.99–0.99) and 0.99 (95%-CI: 0.97–0.99), respectively. No clinically relevant difference in inter-observer classification of other NCA was found. The detection of non-sustained ventricular tachycardia (18% versus 19%) and atrial fibrillation (10% versus 11%) was similar between the two investigators. No sustained ventricular tachycardia was detected.
Conclusion
These findings indicate that our methods are very reliable for scoring NCAs and are adequate to apply for the entire PSG data set of the ADVENT-HF trial.