Portal de Periódicos da CAPES

Prediction of short‐ and long‐term mortality in takotsubo syndrome: the InterTAK Prognostic Score

2019; Elsevier BV; Volume: 21; Issue: 11 Linguagem: Inglês

10.1002/ejhf.1561

1879-0844

Manfred Wischnewsky, Alessandro Candreva, Beatrice Bacchi, Victoria L. Cammann, Ken Kato, Konrad A. Szawan, Sebastiano Gili, Fabrizio D’Ascenzo, Wolfgang Dichtl, Rodolfo Citro, Eduardo Bossone, Michael Neuhaus, Jennifer Franke, Ioana Sorici‐Barb, Miłosz Jaguszewski, Michel Noutsias, Maike Knorr, Susanne Heiner, Christof Burgdorf, Behrouz Kherad, Carsten Tschöpe, Annahita Sarcon, Jerold S. Shinbane, Lawrence Rajan, Guido Michels, Roman Pfister, Alessandro Cuneo, Claudius Jacobshagen, Mahir Karakas, Wolfgang Köenig, Alexander Pott, Philippe Meyer, José David Arroja, Adrian Banning, Florim Cuculi, Richard Kobza, Thomas Fischer, Tuija Vasankari, Juhani Airaksinen, L. Christian Napp, Monika Budnik, Rafał Dworakowski, Philip MacCarthy, Christoph Kaiser, Stefan Osswald, Leonarda Galiuto, Christina Chan, Paul Bridgman, Daniel Beug, Clément Delmas, Olivier Lairez, Ibrahim El‐Battrawy, İbrahim Akın, Ekaterina Gilyarova, Alexandra Shilova, Mikhail Gilyarov, Martin Kozel, Petr Toušek, David E. Winchester, Jan Gałuszka, Christian Ukena, Gregor Poglajen, Pedro Carrilho‐Ferreira, Christian Hauck, Carla Paolini, Claudio Bilato, Abhiram Prasad, Charanjit S. Rihal, Kan Liu, P. Christian Schulze, Matteo Bianco, Lucas Jörg, Hans Rickli, Thanh H. Nguyen, Yoshio Kobayashi, Michael Böhm, Lars S. Maier, Fausto J. Pinto, Petr Widimský, Martin Borggrefe, Stephan B. Felix, Grzegorz Opolski, Ruediger C. Braun‐Dullaeus, Wolfgang Rottbauer, Gerd Hasenfuß, Burkert Pieske, Heribert Schunkert, Hölger Thiele, Johann Bauersachs, Hugo A. Katus, John D. Horowitz, Carlo Di Mario, Thomas Münzel, Filippo Crea, Jeroen J. Bax, Thomas F. Lüscher, Frank Ruschitzka, Jelena R. Ghadri, Christian Templin,

Cardiac Health and Mental Health

Recent evidence suggests comparable in-hospital and long-term outcomes between takotsubo syndrome (TTS) and acute coronary syndrome.1, 2 Medical scoring systems are practical tools for decision making and prognostic assessment. However, TTS-specific scoring systems for risk stratification have not yet been established. Recently, classification based on triggering conditions proved useful in predicting adverse outcomes in TTS (InterTAK Classification).1 Since clinical parameters other than triggering conditions can be associated with adverse outcomes in TTS, such as systolic blood pressure and heart rate,3 the present study aimed to establish a scoring system combining triggering factors with other important but easily- obtainable clinical parameters of daily clinical practice. Takotsubo syndrome patients were enrolled from the International Takotsubo (InterTAK) Registry, which is an all-comers TTS registry in collaboration with 26 centres worldwide.2 Univariate Cox regression analyses for all-cause mortality were performed with variables of interest based on updated knowledge and recent literature1-6: age (cut-off = 70 years), triggering factors (InterTAK Classification), sex, hypertension, diabetes mellitus, typical or atypical TTS type, heart rate on admission (cut-off = 94 bpm), systolic blood pressure on admission (cut-off = 119 mmHg), left ventricular ejection fraction (cut-off = 45%) and white blood cell count on admission. Patients with missing values for these parameters were excluded. All variables with significance on univariate analysis were entered into a Cox regression model for all-cause mortality with respect to factor interactions. The Cox regression model was internally validated using 1000 bootstrap samples. The bootstrap was used to develop a regression coefficient-based scoring system in the presence of competing risks. The regression (beta) coefficients of the significant risk factors were multiplied by 10 and rounded to the nearest integer to derive item points. These points were then incorporated into the InterTAK Prognostic Score, calculated by the summation of points associated with each risk factor: InterTAK Prognostic Score = γ1 × risk factor A1 + γ2 × risk factor A2 + … + γn × risk factor An, where γ1, γ2,…, γn denote the item points of the risk factor Ai. Optimal cut-off values were determined through decision tree analysis with exhaustive CHAID algorithm. Based on these cut-off values, patients were categorized into four groups: low, intermediate, high, and very high risk. Survival distributions and median survival times of each group were estimated using the Kaplan–Meier (KM) product–limit method and were compared using the log-rank test. Time-dependent receiver operating characteristic curves from censored survival data using KM or nearest neighbor estimation (NNE) method were used to assess score predictive performance in terms of area under the curve (AUC) for various time points. All tests were two-tailed and statistical significance was defined as P < 0.05. Statistical analyses were performed using R (version 3.5; R Foundation for Statistical Computing, Vienna, Austria) and SPSS (version 25; IBM Corp., Armonk, NY, USA). A total of 1160 patients (90.8% females; mean age 66.5 ± 13.0 years) were included in the present study. Overall, 80.6% of patients showed typical TTS type and mean left ventricular ejection fraction was 41.1% ± 11.6%. Mean systolic blood pressure and heart rate on admission were 130.2 ± 27.6 mmHg and 87.9 ± 21.9 bpm. The prevalence of diabetes mellitus was 13.4%. An emotional trigger (InterTAK Classification, Class I) was identified in 32.6% of TTS patients while 32.1% had preceding physical activities, medical conditions, or procedures (Class IIa) and 5.7% had preceding neurologic disorders (Class IIb). The remaining patients (29.7%) had no identifiable triggering factors (Class III). According to bootstrap results, points were assigned to each risk factor that was independently associated with all-cause mortality (Figure 1A). Based on the total points, patients were categorized into four risk groups: low risk ≤ 15 points (37.8%), intermediate risk 16–22 points (23.4%), high risk 23–28 points (28.4%), and very high risk ≥ 29 points (10.4%). The four risk groups showed significant differences in all-cause mortality when using the low-risk group as reference [intermediate risk: hazard ratio (HR) 3.15, 95% confidence interval (CI) 1.74–5.72, P < 0.001; high risk: HR 6.16, 95% CI 3.46–10.98, P < 0.001, and very high risk: HR 11.82, 95% CI 6.56–21.28, P < 0.001] (Figure 1B). The AUCs for KM or NNE showed that the score predictive performance is almost stable from 30 days post-admission (KM/NNE-AUC 0.74/0.74) up to 5 years (KM/NNE-AUC 0.78/0.77). Although well-known scores for other cardiac disorders, such as the CHA2DS2-VASc risk score,7 have been applied as proxy scores for risk evaluation in TTS, there has been no TTS-specific risk score thus far. The present study establishes a novel risk stratification score for TTS (InterTAK Prognostic Score) that only requires variables that are easily obtainable in the acute phase and could identify low to very high risk of all-cause mortality both at short- and long-term. Thus, the InterTAK Prognostic Score could serve as a useful clinical tool to stratify patients according to their risk. C.T. has been supported by the H.H. Sheikh Khalifa bin Hamad Al-Thani Research Programme and the Swiss Heart Foundation. L.S.M. was supported by EU HORIZON 2020 (SILICOFCM ID777204). The InterTAK Registry is supported by the Biss Davies Charitable Trust. Conflict of interest: none declared.