Utility of P-wave abnormalities for distinguishing embolic stroke from non-embolic stroke.

BACKGROUND: Our aim was to analyze the incidence of P-wave abnormalities in embolic and non-embolic strokes, and evaluate its clinical usefulness for predicting stroke etiology. METHODS: We included 376 consecutive patients hospitalized for acute ischemic stroke from January 2015 to September 2021. Among the patients in sinus rhythm at admission, 31 had ischemic stroke due to atrial fibrillation (AF)-related embolism, 59 had embolic stroke of unknown source (ESUS), and 143 had non-embolic stroke. P-wave abnormalities were defined as 1. P-wave axis abnormality (PWAA); 2. P-wave terminal force in V1 (PTFV1) ≤ -4000 µV*ms; 3. advanced inter-atrial block (A-IAB). RESULTS: The prevalence of each type of abnormality was consistently lower in patients with non-embolic stroke than in those with AF-related embolism (AF-related vs. ESUS vs. non-embolic; PWAA, 45% vs. 20% vs. 14%; PTFV1, 36% vs. 37% vs. 15%; and A-IAB, 55% vs. 31% vs. 13%, respectively). The identification of at least one type of P-wave abnormality improved the sensitivity compared to using a single abnormality parameter (sensitivity 72%, specificity 62%), while at least two types of abnormality had low sensitivity, but high specificity (sensitivity 29%, specificity 95%). Multivariate regression analysis revealed that identification of at least one type of P-wave abnormality was independently associated with embolic stroke (odds ratio 3.11, 95%CI 1.46-6.63). CONCLUSIONS: The incidence of each type of P-wave abnormality was significantly lower in patients with non-embolic stroke. A combination of PWAA, PTFV1, and A-IAB parameters could be useful for distinguishing embolic from non-embolic stroke.