Early recognition of patients at high risk for atrial fibrillation may help to minimize potential health risks. The detection of susceptibility to develop atrial fibrillation is thus a real clinical challenge. Whereas many studies have used the signal-averaged P wave, the aim of this work is to determine whether electrocardiographic parameters resulting from the analysis of the P wave in ECG recorded during sinus rhythm could be markers for paroxysmal atrial fibrillation susceptibility. Our idea was to compare the ECG in sinus rhythm from two populations: healthy people and patients subject to paroxysmal atrial fibrillation. In addition to standard P wave parameters (P width, P-R interval,. . . ), the Euclidean distance between beat-to-beat P waves, which has been rarely addressed in this context, was studied on lead V1. Significant differences between the healthy and the paroxysmal atrial fibrillation groups were obtained for various parameters. Moreover, a classification of the two groups based on the joint analysis of P width and P-R interval was suggested. This proposed classification could lead to an effective identification of patients at risk to develop atrial fibrillation.