Modelo de predicción para cuantificar la fracción de eyección recuperada del ventrículo izquierdo en taquimiocardiopatía

Abstract

Introduction and objectives

Tachycardiomyopathy is a frequent cause of reversible ventricular dysfunction, whose predictors of recovery are not properly identified. Their individual estimate would be useful to carry out the corresponding invasive procedures.

Methods

This is an observational, retrospective and unicentral study. Cases diagnosed with tachycardiomyopathy were collected between September 2015 and January 2023. The sample was split into 2 study groups: the first one was used for the construction of the linear regression model by performing a multivariate analysis based on evaluating all possible equations (sample 1, 70%); whereas the second one was intended for the validation of the model (sample 2, 30%).

Results

A total of 134 patients were gathered and left ventricular ejection fraction (LVEF) at recovery (the difference between final and original LVEF) was calculated in all of them. Within sample 1, setting parameters of 16,383 models were estimated. The model selected, based on Mallows’ Cp index, was the composed of the following variables: sex, arterial hypertension (HT), LVEF at diagnosis, achievement of rhythm control, and ablation. The selected model explains about half of the individual variability of the recovered LVEF (R² 0.493). Afterwards, individual LVEF at recovery predicted by the model was compared with the observed LVEF in sample 2. No overall significant differences were observed in the R² coefficient of determination.

Conclusions

Sex, arterial hypertension, LVEF at diagnosis, achievement of rhythm control, and ablation were considered as predictors of recovered LVEF in patients with tachycardiomyopathy. This has made possible a quantitative estimate when integrating them within the model.