BackgroundWhile resting ECG parameters have been shown to independently predict adverse cardiovascular outcomes in various populations with and without cardiovascular disease, the prognostic value of these findings have not been specifically addressed in patients with atrial fibrillation.MethodsWe conducted a combined post hoc analysis on all 5436 patients with a history of non-permanent atrial fibrillation randomized to rhythm- versus rate-control therapy in AFFIRM and AF-CHF trials. A total of 7159 baseline ECGs (4848 in atrial fibrillation and 2311 in sinus rhythm) were analyzed to assess the predictive value of the PR interval, QRS duration, and QT intervals on hospitalizations and mortality. Univariate and stepwise multivariate Cox regression models were created for the following outcomes: cardiovascular mortality, all-cause mortality, arrhythmic death, time to first hospitalization, and time to first cardiovascular hospitalization. Separate analyses were performed in atrial fibrillation and in sinus rhythm.ResultsIncreased QRS duration (≥120 msec) was significantly and independently associated with all-cause mortality (HR 1.38; p<0.0001), cardiovascular death (HR 1.50; p<0.0001), arrhythmic death (HR 1.86; p=0.0001), any hospitalisation (HR 1.12; p=0.02), and cardiovascular hospitalisation (HR 1.24; p=0.0002). Increased PR interval (>200 msec) was significantly and independently associated with cardiovascular (HR 1.55; p=0.01), and arrhythmic mortality (HR 2.05; p=0.005). Prolonged QTc (440 msec) predicted cardiovascular hospitalisation (HR 1.23; p=0.04) and arrhythmic mortality (men; HR 1.42; p=0.03). Uncorrected QT interval and rate of change in QRS duration were not predictive of adverse outcomes.ConclusionSimple parameters on baseline ECG significantly and independently predict adverse cardiovascular outcomes in patients with a history of atrial fibrillation. BackgroundWhile resting ECG parameters have been shown to independently predict adverse cardiovascular outcomes in various populations with and without cardiovascular disease, the prognostic value of these findings have not been specifically addressed in patients with atrial fibrillation. While resting ECG parameters have been shown to independently predict adverse cardiovascular outcomes in various populations with and without cardiovascular disease, the prognostic value of these findings have not been specifically addressed in patients with atrial fibrillation. MethodsWe conducted a combined post hoc analysis on all 5436 patients with a history of non-permanent atrial fibrillation randomized to rhythm- versus rate-control therapy in AFFIRM and AF-CHF trials. A total of 7159 baseline ECGs (4848 in atrial fibrillation and 2311 in sinus rhythm) were analyzed to assess the predictive value of the PR interval, QRS duration, and QT intervals on hospitalizations and mortality. Univariate and stepwise multivariate Cox regression models were created for the following outcomes: cardiovascular mortality, all-cause mortality, arrhythmic death, time to first hospitalization, and time to first cardiovascular hospitalization. Separate analyses were performed in atrial fibrillation and in sinus rhythm. We conducted a combined post hoc analysis on all 5436 patients with a history of non-permanent atrial fibrillation randomized to rhythm- versus rate-control therapy in AFFIRM and AF-CHF trials. A total of 7159 baseline ECGs (4848 in atrial fibrillation and 2311 in sinus rhythm) were analyzed to assess the predictive value of the PR interval, QRS duration, and QT intervals on hospitalizations and mortality. Univariate and stepwise multivariate Cox regression models were created for the following outcomes: cardiovascular mortality, all-cause mortality, arrhythmic death, time to first hospitalization, and time to first cardiovascular hospitalization. Separate analyses were performed in atrial fibrillation and in sinus rhythm. ResultsIncreased QRS duration (≥120 msec) was significantly and independently associated with all-cause mortality (HR 1.38; p<0.0001), cardiovascular death (HR 1.50; p<0.0001), arrhythmic death (HR 1.86; p=0.0001), any hospitalisation (HR 1.12; p=0.02), and cardiovascular hospitalisation (HR 1.24; p=0.0002). Increased PR interval (>200 msec) was significantly and independently associated with cardiovascular (HR 1.55; p=0.01), and arrhythmic mortality (HR 2.05; p=0.005). Prolonged QTc (440 msec) predicted cardiovascular hospitalisation (HR 1.23; p=0.04) and arrhythmic mortality (men; HR 1.42; p=0.03). Uncorrected QT interval and rate of change in QRS duration were not predictive of adverse outcomes. Increased QRS duration (≥120 msec) was significantly and independently associated with all-cause mortality (HR 1.38; p<0.0001), cardiovascular death (HR 1.50; p<0.0001), arrhythmic death (HR 1.86; p=0.0001), any hospitalisation (HR 1.12; p=0.02), and cardiovascular hospitalisation (HR 1.24; p=0.0002). Increased PR interval (>200 msec) was significantly and independently associated with cardiovascular (HR 1.55; p=0.01), and arrhythmic mortality (HR 2.05; p=0.005). Prolonged QTc (440 msec) predicted cardiovascular hospitalisation (HR 1.23; p=0.04) and arrhythmic mortality (men; HR 1.42; p=0.03). Uncorrected QT interval and rate of change in QRS duration were not predictive of adverse outcomes. ConclusionSimple parameters on baseline ECG significantly and independently predict adverse cardiovascular outcomes in patients with a history of atrial fibrillation. Simple parameters on baseline ECG significantly and independently predict adverse cardiovascular outcomes in patients with a history of atrial fibrillation.