Vecchione, Carmine

ABSTRACT Aims Takotsubo syndrome (TTS) is associated with a substantial rate of adverse events. We sought to design a machine learning (ML)-based model to predict the risk of in-hospital death and to perform a clustering of TTS patients to identify different risk profiles. Methods and results A ridge logistic regression-based ML model for predicting in-hospital death was developed on 3482 TTS patients from the International Takotsubo (InterTAK) Registry, randomly split in a train and an internal validation cohort (75% and 25% of the sample size, respectively) and evaluated in an external validation cohort (1037 patients). Thirty-one clinically relevant variables were included in the prediction model. Model performance represented the primary endpoint and was assessed according to area under the curve (AUC), sensitivity and specificity. As secondary endpoint, a K-medoids clustering algorithm was designed to stratify patients into phenotypic groups based on the 10 most relevant features emerging from the main model. The overall incidence of in-hospital death was 5.2%. The InterTAK-ML model showed an AUC of 0.89 (0.85–0.92), a sensitivity of 0.85 (0.78–0.95) and a specificity of 0.76 (0.74–0.79) in the internal validation cohort and an AUC of 0.82 (0.73–0.91), a sensitivity of 0.74 (0.61–0.87) and a specificity of 0.79 (0.77–0.81) in the external cohort for in-hospital death prediction. By exploiting the 10 variables showing the highest feature importance, TTS patients were clustered into six groups associated with different risks of in-hospital death (28.8% vs. 15.5% vs. 5.4% vs. 1.0.8% vs. 0.5%) which were consistent also in the external cohort. Conclusion A ML-based approach for the identification of TTS patients at risk of adverse short-term prognosis is feasible and effective. The InterTAK-ML model showed unprecedented discriminative capability for the prediction of in-hospital death.

Abstract Aims Acute pulmonary disorders are known physical triggers of takotsubo syndrome (TTS). This study aimed to investigate prevalence of acute pulmonary triggers in patients with TTS and their impact on outcomes. Methods and results Patients with TTS were enrolled from the International Takotsubo Registry and screened for triggering factors and comorbidities. Patients were categorized into three groups (acute pulmonary trigger, chronic lung disease, and no lung disease) to compare clinical characteristics and outcomes. Of the 1670 included patients with TTS, 123 (7%) were identified with an acute pulmonary trigger, and 194 (12%) had a known history of chronic lung disease. The incidence of cardiogenic shock was highest in patients with an acute pulmonary trigger compared with those with chronic lung disease or without lung disease (17% vs. 10% vs. 9%, P = 0.017). In-hospital mortality was also higher in patients with an acute pulmonary trigger than in the other two groups, although not significantly (5.7% vs. 1.5% vs. 4.2%, P = 0.13). Survival analysis demonstrated that patients with an acute pulmonary trigger had the worst long-term outcome (P = 0.002). The presence of an acute pulmonary trigger was independently associated with worse long-term mortality (hazard ratio 2.12, 95% confidence interval 1.33–3.38; P = 0.002). Conclusions The present study demonstrates that TTS is related to acute pulmonary triggers in 7% of all TTS patients, which accounts for 21% of patients with physical triggers. The presence of acute pulmonary trigger is associated with a severe in-hospital course and a worse long-term outcome.