Abstract 20617: A Risk Score for Predicting Mortality in in 10,582 Patients Hospitalized With Takotsubo Cardiomyopathy
Background: Takotsubo cardiomyopathy (TTCM) is characterized by acute, reversible left ventricular dysfunction in the absence of epicardial coronary artery occlusion. Because of its rarity, little is known risk factors for mortality during initial presentation in patients with TTCM.
Objectives: To describe risk factors for in-hospital mortality and major adverse events (MAE) in TTCM.
Methods: Deidentified hospital data from California, New York, New Jersey, Vermont, New Hampshire, West Virginia, and Colorado between 2006 and 2012 were harmonized, and hospital records reporting the ICD-9 code 429.83 (takotsubo cardiomyopathy) were analyzed. Demographics, comorbidities, and in-hospital mortality were quantified. Characteristics independently associated with death were identified using multivariate logistic regression, and a risk score was calculated using positive multivariate predictors of in-hospital mortality. Performance in predicting death was compared with the full regression model using area under the curve (AUC) analysis.
Results: In total, 10,582 hospital records reporting TTCM were identified. Overall in-hospital mortality was 4.4%. Factors associated with increased mortality on multivariate analysis included age ≥ 60 (OR 1.76 [1.31-2.35], p<0.001), male gender (OR 1.91 [1.44-2.53], p<0.001), Asian race (OR 1.82 [1.13-2.91], p=0.01), atrial fibrillation (OR 1.68 [1.29-2.18], p<0.001), intracranial hemorrhage (OR 6.84 [4.31-10.9], p<0.001), stroke (OR 2.90 [1.88-4.47], p<0.001) and acute renal failure (OR 4.11 [3.22-5.23, p<0.001). A risk score calculated by counting positive predictors of mortality was significantly associated with mortality ranging from 42/2585 (1.6%) in patients with no risk factors to 3/8 (37.5%) in patients with ≥ 5 risk factors. Odds of mortality increased with each additional risk factor (OR 2.24 [2.04-2.46], p<0.001). The full multivariate regression model performed very well in predicting mortality (AUC 0.79), and the simplified risk score also performed well (AUC 0.70).
Conclusions: The presence of age ≥ 60, male gender, Asian race, atrial fibrillation, intracranial hemorrhage, stroke, and acute renal failure may identify TTCM patients at increased risk of in-hospital mortality.
Author Disclosures: D.P. Kao: None. J. Lindenfeld: None.
- © 2014 by American Heart Association, Inc.