BACKGROUND: Heart rate control in atrial fibrillation (AF) is typically assessed by 24‑hour electrocardiography (ECG). There are scarce data on the use of 24‑hour ECG parameters to predict mortality in patients with AF. AIMS: We aimed to identify 24‑hour ECG parameters that predict mortality in patients with AF. METHODS: We enrolled 280 ambulatory patients (mean [SD] age, 72 [8.7] years; 57.9% men) with permanent or persistent AF. Data on mortality and pacemaker or defibrillator implantation during follow‑up were collected. Predictors of mortality were assessed using the Cox proportional hazards model and C statistic. RESULTS: Compared with survivors, 78 patients (28%) who died were older, more often had comorbidities, left bundle branch block (LBBB), reduced left ventricular ejection fraction, lower maximum heart rate, higher number of ventricular extrasystoles, and the longest R‑R interval below 2 seconds. Univariate analysis showed higher mortality in patients with the longest R‑R intervals below 2 seconds compared with those with R‑R intervals of 2 seconds or longer (P <0.001). Independent mortality predictors in the regression model included older age, renal failure, history of coronary intervention, chronic obstructive pulmonary disease, LBBB, and a high number (≥770) or absence of R‑R intervals of at least 2 seconds. The area under the curve for mortality prediction increased after including ECG parameters (0.748; 95% CI, 0.686-0.81; vs 0.688; 95% CI, 0.618-0.758; P = 0.02). CONCLUSIONS: A high number of R‑R intervals longer than 2 seconds or their absence on 24‑hour ECG may predict mortality in patients with AF.
BACKGROUND: Heart rate control in atrial fibrillation (AF) is typically assessed by 24‑hour electrocardiography (ECG). There are scarce data on the use of 24‑hour ECG parameters to predict mortality in patients with AF. AIMS: We aimed to identify 24‑hour ECG parameters that predict mortality in patients with AF. METHODS: We enrolled 280 ambulatory patients (mean [SD] age, 72 [8.7] years; 57.9% men) with permanent or persistent AF. Data on mortality and pacemaker or defibrillator implantation during follow‑up were collected. Predictors of mortality were assessed using the Cox proportional hazards model and C statistic. RESULTS: Compared with survivors, 78 patients (28%) who died were older, more often had comorbidities, left bundle branch block (LBBB), reduced left ventricular ejection fraction, lower maximum heart rate, higher number of ventricular extrasystoles, and the longest R‑R interval below 2 seconds. Univariate analysis showed higher mortality in patients with the longest R‑R intervals below 2 seconds compared with those with R‑R intervals of 2 seconds or longer (P <0.001). Independent mortality predictors in the regression model included older age, renal failure, history of coronary intervention, chronic obstructive pulmonary disease, LBBB, and a high number (≥770) or absence of R‑R intervals of at least 2 seconds. The area under the curve for mortality prediction increased after including ECG parameters (0.748; 95% CI, 0.686-0.81; vs 0.688; 95% CI, 0.618-0.758; P = 0.02). CONCLUSIONS: A high number of R‑R intervals longer than 2 seconds or their absence on 24‑hour ECG may predict mortality in patients with AF.