ABSTRACT
BACKGROUND: Coronavirus disease COVID-19 produces a predominantly pulmonary affection, being cardiac involvement an important component of the multiorganic dysfunction. At the moment there are few reports about the behavior of echocardiographic images in the patients who have the severe forms of the disease. OBJECTIVE: Identify the echocardiographic prognostic markers for death within 60 days in patients hospitalized in intensive care. METHODS: A single-center prospective cohort was made with patients hospitalized in intensive care for COVID-19 confirmed via polymerase chain reaction who got an echocardiogram between May and October 2020. A Cox multivariate model was plotted reporting the HR and confidence intervals with their respective p values for clinical and echocardiographic variables. RESULTS: Out of the 326 patients included, 153 patients got an echocardiogram performed on average 6.8 days after admission. The average age was 60.7, 47 patients (30.7%) were females and 67 (44.7%) registered positive troponin. 91 patients (59.5%) died. The univariate analysis identified TAPSE, LVEF, pulmonary artery systolic pressure, acute cor pulmonale, right ventricle diastolic dysfunction, and right ventricular dilatation as variables associated with mortality. The multivariate model identified that the acute cor pulmonale with HR= 4.05 (CI 95% 1.09 - 15.02, p 0.037), the right ventricular dilatation with HR= 3.33 (CI 95% 1.29 - 8.61, p 0.013), and LVEF with HR= 0.94 (CI 95% 0.89 - 0.99, p 0.020) were associated with mortality within 60 days. CONCLUSIONS: In patients hospitalized in the intensive care unit for COVID-19, the LVEF, acute cor pulmonale and right ventricular dilatation are prognostic echocardiographic markers associated with death within 60 days.
Subject(s)
COVID-19 , Ventricular Dysfunction, Right , Critical Care , Echocardiography , Female , Humans , Prospective Studies , Ventricular Dysfunction, Right/complicationsABSTRACT
Resumen Introducción: Por tratarse de una tarea altamente compleja y de importancia clínica, el diagnóstico del síndrome coronario agudo se presta para su exploración por medio de modelado mediante sistemas inteligentes. Objetivo: desarrollar un sistema multiagente que ensamble las decisiones de varias redes neuronales para el diagnóstico del dolor torácico enfocado a los síndromes coronarios agudos. Metodología: estudio de pruebas diagnósticas en el que se entrenan un conjunto de redes neuronales con una precisión cercana al 70%, que luego son ensambladas mediante tres sistemas de votación para luego adicionar el resultado de redes especiales en poblaciones particulares y seleccionar la mejor configuración que hará parte de un sistema multiagente para el diagnóstico del dolor torácico. Resultados: Se generaron 84 redes con precisión promedio del 72% en pruebas; al ensamblarse aumentan dicha precisión hasta llegar a un máximo del 84% que tras la adición de los grupos especiales alcanza el 89%. Se escoge una conformación que brinda una sensibilidad del 96% con una especificidad del 77%, con valores predictivos positivo y negativo de 87 y 93% respectivamente para el diagnóstico de síndrome coronario agudo. Conclusiones: Es posible desarrollar una herramienta para el diagnóstico automático del síndrome coronario agudo a partir de un sistema multiagente que ensamble la disposición tomada por un conjunto de redes neuronales artificiales, cuyo rendimiento permite su consideración para su implementación dentro de un sistema de soporte a las decisiones clínicas.
Abstract Introduction: Because it is a highly complex task of a great clinical importance, the diagnosis of acute coronary syndromes allows for their analysis by means of intelligent system models. Motivation: To develop a multi-agent system that assembles the decisions of several neural networks for the diagnosis of chest pain with a focus on acute coronary syndromes. Methods: A study of diagnostic tests where a series of neural networks are trained with a precision close to 70%, and are later on assembled with three voting systems. Then the results of special networks on specific populations are added to select the best configuration that Will make part of a multi-agent system for diagnosing chest pain. Results: A total of 84 networks were generated, with an average precision of 72% during testing; once assembled this precision rises up to a maximum of 84%, which then reaches 89% when the special groups are included. A configuration that offers a sensitivity of 96% with a specificity of 77% and positive and negative predictive values of 87 and 93% respectively is chosen for the diagnosis of acute coronary syndrome. Conclusions: It is possible to develop a tool for the automatic diagnosis of acute coronary syndrome using a multi-agent system that assembles the dispositions taken by a set of artificial neural networks. Its performance allows taking it into consideration for implementing it within a clinical decision-making support system.
