Emergency Physician-performed Transesophageal Echocardiography in Simulated Cardiac Arrest

Abstract

Introduction

Transesophageal echocardiography (TEE) is a well-established method of evaluating cardiac pathology. It has many advantages over transthoracic echocardiography (TTE), including the ability to image the heart during active cardiopulmonary resuscitation. This prospective simulation study aims to evaluate the ability of emergency medicine (EM) residents to learn TEE image acquisition techniques and demonstrate those techniques to identify common pathologic causes of cardiac arrest.

Methods

This was a prospective educational cohort study with 40 EM residents from two participating academic medical centers who underwent an educational model and testing protocol. All participants were tested across six cases, including two normals, pericardial tamponade, acute myocardial infarction (MI), ventricular fibrillation (VF), and asystole presented in random order. Primary endpoints were correct identification of the cardiac pathology, if any, and time to sonographic diagnosis. Calculated endpoints included sensitivity, specificity, and positive and negative predictive values for emergency physician (EP)-performed TEE. We calculated a kappa statistic to determine the degree of inter-rater reliability.

Results

Forty EM residents completed both the educational module and testing protocol. This resulted in a total of 80 normal TEE studies and 160 pathologic TEE studies. Our calculations for the ability to diagnose life-threatening cardiac pathology by EPs in a high-fidelity TEE simulation resulted in a sensitivity of 98%, specificity of 99%, positive likelihood ratio of 78.0, and negative likelihood ratio of 0.025. The average time to diagnose each objective structured clinical examination case was as follows: normal A in 35 seconds, normal B in 31 seconds, asystole in 13 seconds, tamponade in 14 seconds, acute MI in 22 seconds, and VF in 12 seconds. Inter-rater reliability between participants was extremely high, resulting in a kappa coefficient across all cases of 0.95.

Conclusion

EM residents can rapidly perform TEE studies in a simulated cardiac arrest environment with a high degree of precision and accuracy. Performance of TEE studies on human patients in cardiac arrest is the next logical step to determine if our simulation data hold true in clinical practice.