Novel ECG-based scoring tool for prediction of takotsubo syndrome.

Background: Takotsubo syndrome (TS) usually involves ECG changes mimicking acute myocardial infarction (AMI). The differentiation of both disorders is crucial for selection of appropriate treatment. The aim of this study was to assess ECG parameters in patients with TS and AMI, and try to establish a scoring tool for TS prediction.

Methods: The study consisted of two study parts: evaluation and validation cohorts. Overall, the study included 82 patients with TS and 141 subjects with AMI. In addition to the major demographic characteristics and comorbidities, the following ECG parameters were analyzed: heart rate, QRS duration, QTc, QRS amplitudes in frontal and precordial leads, frequencies for ST-segment elevation, combined sign of positive ST-segment elevation in -aVR and absent in V, negative T-wave in lead I and positive in III, inverted or biphasic T-waves in V-V, T-wave inversions in frontal and precordial leads. All significant variables were identified in univariate regression analysis and further included for multivariate logistic regression analysis predicting TS.

Results: TS was frequently diagnosed in women and in elderly patients. Presence of ST-segment elevation, inverted/biphasic T-waves in V-V, QRS amplitudes in frontal and precordial leads were significantly different in evaluation group. By multivariate regression analysis sex, QRS amplitudes in frontal, inverted or biphasic T-waves in septal leads and QTc were identified as powerful variables to calculate TS probability. The diagnostic accuracy of the developed 6-points-TS-score was then evaluated in the validation group. Thus, no subject with a TS-score of ≥ 5 had AMI (specificity 99%, sensitivity > 92%).

Conclusion: The developed ECG-based TS-score model may be a useful complimentary tool for TS prediction in acute clinical setting.