Development and Implementation of a Multicenter Registry for Resuscitation-Focused Transesophageal Echocardiography

Abstract

Objectives:

To evaluate the clinical impact, safety, and clinical outcomes of focused transesophageal echocardiography (TEE) in the evaluation of critically ill patients in the emergency department (ED) and intensive care units (ICU).

Methods:

We established a prospective, multicenter, observational registry involving adult critically ill patients in whom focused TEE was performed for evaluation of out-of-hospital cardiac arrest (OHCA), in-hospital cardiac arrest (IHCA), evaluation of undifferentiated shock, hemodynamic monitoring, and/or procedural guidance in the ED, ICU, or operating room (OR) setting. The primary objective of the current investigation was to evaluate the clinical impact and safety of focused, point-of-care TEE in critically ill patients. Data elements included patient and procedure characteristics, laboratory values, timing of interventions, clinical outcomes, and TEE video images.

Results:

A total of 1,045 focused TEE studies were collected among 916 patients from 28 hospitals, including 585 (64%) intra-arrest and post-arrest OHCA and IHCA, 267 (29%) initial evaluation of undifferentiated shock, 101 (11%) procedural guidance, and 92 (10%) hemodynamic monitoring. TEE changed management in 85% of patients with undifferentiated shock, 71% of IHCA, and in 62% of cases of OHCA. There were no reported esophageal perforations or oropharyngeal injurie s, and other procedural complications were rare.

Conclusion:

A prospective, multicenter, and multidisciplinary TEE registry was successfully implemented, and demonstrated that focused TEE is safe and clinically impactful across multiple critical care applications. Further studies from this research network will accelerate the development of outcome-oriented research and knowledge translation on the use of TEE in emergency and critical care settings.

INTRODUCTION

Transesophageal echocardiography (TEE) has been incorporated over recent years as an additional modality of focused ultrasonography in the care of critically ill patients.^1–3 Once only utilized for comprehensive echocardiography applications such as intraoperative assessment, evaluation of suspected endocarditis, or screening for intra-cardiac thrombus prior to cardioversion, focused TEE has been increasingly adopted by clinicians to support emergency decision making in prehospital,⁴ Emergency Department (ED), ^5–9 Operating Room (OR),^10–12 Intensive Care Unit (ICU) settings. ^3,13–17 TEE is ideally suited for the evaluation of patients across these acute care environments, given superior image quality compared to transthoracic echocardiography, TEE probe proximity to cardiac structures, and its ability to provide continuous imaging during resuscitative interventions.^1,7

To date, observational, single-center studies have characterized the feasibility, safety, and clinical impact of focused TEE in various point-of-care applications including cardiac arrest resuscitation,^5–10,18 evaluation of shock and hemodynamic monitoring,^13,15,16,17 guidance of emergency endovascular procedures,^{5,13,15,16,19} and evaluation of blunt thoracic trauma.²⁰ In 2020, a multidisciplinary group of experts evaluated the use of this modality for resuscitation care and proposed a research agenda for this emerging field, highlighting the need for larger, multicenter studies evaluating TEE to provide evidence on safety, clinical impact, and outcomes, as well as the development of standardized and generalizable protocols for use.⁶ In response to these gaps, we developed and implemented the Resuscitative TEE Collaborative Registry (rTEECoRe), a novel multicenter and interdisciplinary registry aimed to evaluate the clinical impact, safety, and outcomes of focused TEE in the evaluation of critically ill patients across various acute care environments.

METHODS

Study design and patient population

We designed and implemented a prospective, multicenter, observational study involving critically ill adult patients in whom focused TEE was performed as part of routine care in the OR, ED, or ICU setting. Eligible cases were defined as any focused TEE studies performed by clinicians at the point-of-care, that were clinically indicated according to each institutional protocols at the participating units. TEE studies were performed by attendings, fellows, or residents from each participating unit. Their training and competency standards were determined by each institution’s credentialing process. Comprehensive or consultative TEE examinations (e.g. diagnostic studies performed by a cardiologist or comprehensive perioperative TEE studies) were not included in this study. The study design and reporting adhered to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for reporting observational studies.²¹

Study objectives and outcomes

The rTEECoRe cohort study was designed to study five applications of TEE in adult critically ill patients: (1) intra-arrest and post-arrest evaluation in both out-of-hospital cardiac arrest (OHCA) and (2) in-hospital cardiac arrest (IHCA); (3) initial evaluation of undifferentiated shock or acute hypotension; (4) hemodynamic monitoring in critically ill patients; and (5) guidance of emergency endovascular procedures. The primary objectives of this initial study of the rTEECoRe cohort were to evaluate the impact in management decisions and safety of focused, point-of-care TEE in critically ill patients across various acute care settings. The general clinical impact was defined as the proportion of cases where a management decision was attributed to data obtained from TEE. The clinical impact specifically during cardiac arrest resuscitation was defined as the proportion of cases where chest compression position during CPR was changed based on real-time feedback provided by TEE images. Safety was defined as the rate of confirmed or potential procedure-related complications. Secondary objectives of this study were to characterize the use of this imaging modality in critically ill patients with shock and cardiac arrest, including the evaluation of patient demographics, clinical indications, operator characteristics, TEE views utilized and findings, duration of studies, and patient outcomes. As a requirement for participation in the study, centers were required to enroll at least 90% of their eligible cases into the registry. Participating sites were asked to report the total number of eligible cases per month during the study period, and the total number of TEE studies that were entered into the registry.

Organizational structure of the registry

One academic medical center (University of Pennsylvania) provided operational and research infrastructure including the HIPAA-compliant shared data collection tool, built within a well-established clinical data management platform (REDCap, Vanderbilt University, TN; see below for more details) and a full-time research coordinator who managed onboarding of participating institutions, including execution of data use agreements (DUA), approval of institutional review board (IRB) protocols, and facilitation of data entry at each participating site. Participation in the registry was available to any hospital using focused TEE for one or more applications of interest. Information about the registry was disseminated through the American College of Emergency Physician’s Emergency Ultrasound and Critical Care Sections listservs, as well as other pertinent national and international emergency and critical care societies. Onboarding information to join the study was made publicly available through University of Pennsylvania’s Center for Resuscitation Science website (https://www.med.upenn.edu/resuscitation/rteecore).

An interdisciplinary scientific oversight committee (SOC) involving physicians with clinical and research experience in the field of TEE (FT, CGO, RA, KB, FMW, MN, JH, KA, AU, BSA), developed and refined the data collection instruments, and convened monthly to assess data completeness and quality. The SOC will be responsible in future work with the process of data requests and proposals for new studies using the registry dataset, coordinating registry reports, and ensuring high quality data standards.

Ethics and study registration

Since focused TEE represents a standard ultrasound modality in all participating centers and given the observational nature of the study, the protocol was considered of minimal risk to subjects. The study protocol was reviewed by the IRB at the primary site and determined to meet the criteria for exemption (authorized by 45 CFR 46.104, category 4). Using a template made available to all sites joining the study, each participating site submitted the study protocol to their local IRB, and obtained site-level approval to conduct the study. At all participating sites the study protocol was deemed to have no greater than minimal risk to subjects, and was approved to be conducted without the requirement of informed consent. This study was registered with ClinicalTrials.gov (NCT04972526).

Data capture instruments

We developed a shared online data abstraction platform using a well-established clinical trial database tool (REDCap). REDCap is a secure, HIPAA-compliant, web-based application designed to support data capture for research studies, providing an intuitive interface for validated data entry, audit trails for tracking data manipulation and export procedures, and automated export procedures to common statistical packages. Data capture instruments use branching logic that selectively prompts with specific questions and data elements depending on the indication of TEE and specific cohort of patients (e.g. OHCA vs IHCA, intra vs post-arrest evaluation, etc.).

A preliminary form captured data applicable to all indications, and additional forms captured data specific to the enrollment category. One preliminary form was completed for each patient entered in the study, but one or more forms capturing specific applications of TEE could be utilized for an individual patient (e.g. TEE was used for intra-arrest evaluation of IHCA and also for guidance of ECMO). Analysis of specific applications of TEE were performed using the total number of cases for any given application as the denominator. Abstracted data included patient demographics, clinical and TEE procedure details, operator findings, TEE views obtained, clinician image interpretation, and change in management (if any) dependent upon the information provided by TEE. In addition to data from the clinician’s findings on TEE, the registry protocol required the upload of a minimum of 3 and a maximum of 10 deidentified TEE video images, that were considered representative of the findings reported in each case enrolled. These video images were saved and stored according to each participating site’s local workflow, which at most participating sites includes the use of an image archiving software. Uploaded images were automatically stored as video files along with the rest of collected data associated with the corresponding record ID. For the current study, the analysis of diagnostic findings from TEE was analyzed only from data reported by the clinician performing the study and not from post hoc analysis of these images.

Standarization and training for data entry

Individualized and small group training sessions were conducted via teleconference for each onboarded site with the goal to train individuals that were abstracting hospital-level data into REDCap. These training sessions provided a step-by-step review of the study protocol, study definitions, and all data instruments in REDCap, including the process of saving, storing, deidentification, and uploading of TEE video images into the online database. These sessions included at minimum the site PI responsible for data entry, and usually a local research assistant or coordinator, as well other members of the clinical team that would be enrolling cases in each department. These training sessions were repeated upon request for sites that had included new clinicians or research coordinators, or had a change in their site PI. A digital binder of documents including study protocol, PDF copies of the online data instruments, and data entry manual, were made available to all participating sites.

Study definitions and outcomes

Change in management was defined as any intervention or decision making in the clinical management of the patient that based on the clinician’s judgement was attributed to the information provided by TEE. Potential procedure-related immediate complications were defined as including pharyngeal bleed, endotracheal tube dislodgement, or endotracheal cuff rupture following TEE probe insertion. Esophageal perforation, oropharyngeal injury, and gastrointestinal bleed (GIB), that were considered to be related to TEE, were evaluated at any point during the hospital stay. For cases of OHCA, data collection also included an analysis of pre-hospital care using standardized cardiac arrest elements following international guidelines and Utstein consensus reporting.^21,22 Study outcomes in arrest cohorts included return of spontaneous circulation (ROSC), survival to hospitalization and discharge, and neurological status at discharge including cerebral performance category (CPC) and modified Rankin score (mRS). A description of the registry architecture is provided in Figure 1. A complete list of the data elements, copies of the data capture instruments, and data entry manual, are provided in supplementary materials.

FIGURE 1.

Architecture of rTEECoRE.

rTEECoRe: Resuscitative TEE Collaborative Registry; OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest; TEE: Transesophageal echocardiography; CPR: Cardiopulmonary resuscitation; ED: Emergency department; ROSC: Return of spontaneous circulation; DNR: Do not resuscitate.

Standarization of echocardiographic evaluation in cardiac arrest

For cardiac arrest cases, an important parameter of interest in our work was the anatomical chest compression location on the chest wall, known as the area of maximal compression (AMC).⁶ Given that TEE is an operator-dependent diagnostic modality and that our study incorporated observational data, we standardized the assessment of AMC by establishing an a priori echocardiographic definition of AMC, specifically determined in the mid-esophageal long axis view (ME LAX). In ME LAX view, TEE shows the right ventricle (RV), left ventricle (LV), and left ventricular outflow tract (LVOT) in the same plane. We defined the AMC as the area of the heart with the greatest anteroposterior diameter shortening during the compression phase of CPR, as qualitatively estimated in ME LAX. The options provided in the study instrument include the following AMC locations; “AMC-LV,” AMC-LVOT, “AMC-Other location” (free text provided to describe), “AMC location not evaluated”, and “unable to determine AMC location”.

Quality assurance and data review

A full-time research coordinator (AK), the PI (FT) and SOC members performed weekly quality review of cases enrolled in the registry and followed up with site PIs regarding cases with inconsistencies or possible errors. When possible, missing or inconsistent data were confirmed or corroborated with data from the electronic medical record (EMR). Data entries with inconsistencies or errors that could not be verified and corrected were removed from the study database and marked as missing data.

Data analysis

Data were analyzed using standard statistical software (STATA 14.2, STATA Corp, College Station, TX). Descriptive statistics were used to analyze and describe demographics and baseline characteristics of study subjects, clinical findings and outcomes. We used interquartile range (IQR) to summarize continuous variables and counts and percentages for categorical variables. We calculated 95% confidence intervals (CIs) and statistical significance was assessed at the 0.05 level.

RESULTS

Participating sites

From January 2021 to December 2023, 28 hospitals participated in the registry (21 from the U.S. 6 from Canada, and 1 from Italy) with a median enrollment time of 22 months (IQR 17–24). Of these, 23 sites included TEE cases performed in the ED, 12 included ICU cases, 2 included OR cases, and 2 included cases performed in hospital wards. The median number of cases enrolled by site was 21 (IQR 7–41), with median of 1 (IQR 1–2) case / month per participating site during the study period. Twenty-five of the participating sites had between 1–9 attending physicians performing TEE, and 3 sites had between 10–20 at the time of onboarding in the study. Twenty-two sites were academic centers with residency and/or fellowship programs and 6 were community hospitals with affiliated residency programs. Figure 2 shows case enrollments and distribution by participating sites.

FIGURE 2.

Case enrollments and distribution by participating sites.

rTEECoRe: Resuscitative TEE Collaborative Registry; OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest.

Study enrollment and patient characteristics

Within the study period, a total of 1,045 focused TEE studies were performed among 916 patients across all indications, including 437 (47%) intra and post-arrest OHCA, 267 (29%) initial evaluation of undifferentiated shock, 148 (17%) intra and post-arrest IHCA, 101 (11%) procedural guidance, and 92 (10%) hemodynamic monitoring. Patients had a median age of 62 (IQR 48–72), were 68% male, with a median BMI 30 (IQR 25–36), and had a high prevalence of chronic medical conditions including HTN (41%) and CAD (22%). Cohort demographic characteristics are shown in Table 1.

TABLE 1.

Characteristics of patients enrolled in rTEECoRe across all indications of focused TEE.

	Total	Intra-arrest evaluation in OHCA	Post-arrest evaluation in OHCA	Intra-arrest evaluation in IHCA	Post-arrest evaluation in IHCA	Initial evaluation of undifferentiated shock	Hemodynamic monitoring	Procedural guidance
	N=1,045	N=278	N=159	N=88	N=60	N=267	N=92	N=101
Age (years)	62 (48-72)	63 (49-73)	64 (52-75)	66 (56-74)	58 (40-68)	59 (46-69)	60 (47-70)	54 (40-64)
Biological Sex
Male	68% (626)	69% (188)	65% (103)	60% (53)	78% (47)	74% (197)	71% (65)	72% (73)
Female	29% (266)	29% (81)	34% (54)	39% (34)	22% (13)	25% (68)	26% (24)	27% (27)
Race
Native American/Alaska Native	6% (54)	1% (4)	5% (8)	3% (3)	12% (7)	12% (31)	7% (6)	16% (16)
Native Hawaiian/Pacific Islander	1% (5)	0% (0)	1% (1)	1% (1)	0% (0)	1% (3)	0% (0)	0% (0)
Asian	2% (21)	2% (6)	2% (3)	3% (2)	3% (2)	2% (6)	0% (0)	4% (4)
Black or African American	8% (77)	11% (31)	8% (13)	10% (9)	5% (3)	6% (15)	8% (7)	9% (9)
Latinx	11% (102)	5% (15)	9% (15)	15% (13)	17% (10)	17% (46)	8% (7)	8% (8)
White	35% (325)	35% (97)	30% (48)	30% (26)	42% (25)	42% (111)	30% (28)	53% (54)
Unknown	1% (8)	0% (1)	1% (1)	1% (1)	2% (1)	1% (2)	2% (2)	0% (0)
Missing	35% (320)	44% (123)	43% (69)	36% (32)	20% (12)	19% (52)	45% (41)	10% (10)
BMI	30 (25–36)	28 (25–34)	26 (23–33)	30 (25–35)	30 (25–35)	31 (26–38)	28 (24–34)	33 (26–39)
Past Medical History
CAD	22% (202)	23% (65)	25% (40)	30% (26)	18% (11)	18% (49)	22% (20)	11% (11)
CHF	15% (137)	14% (39)	16% (25)	16% (14)	13% (8)	11% (30)	21% (19)	9% (9)
CKD	14% (127)	14% (40)	15% (24)	23% (20)	10% (6)	10% (27)	20% (18)	5% (5)
DM	28% (259)	24% (68)	33% (53)	48% (42)	25% (15)	27% (71)	27% (25)	26% (26)
pMI	6% (51)	8% (22)	9% (14)	8% (7)	3% (2)	2% (5)	5% (5)	1% (1)
VAD	1% (8)	1% (2)	1% (1)	1% (1)	0% (0)	0% (0)	3% (3)	2% (2)
ICD	2% (15)	2% (5)	1% (1)	3% (3)	0% (0)	0% (1)	4% (4)	1% (1)
HTN	41% (371)	37% (104)	50% (80)	55% (48)	37% (22)	35% (93)	46% (42)	25% (25)
COPD	11% (98)	8% (22)	11% (18)	16% (14)	8% (5)	12% (31)	12% (11)	6% (6)
pHTN	3% (32)	3% (8)	1% (2)	5% (4)	7% (4)	1% (3)	5% (5)	7% (7)

Data are presented as median (IQR) for continuous measures, and % (n) for categorical measures The % has been rounded to the nearest tenth. rTEECoRe: Resuscitative TEE Collaborative Registry; OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest; BMI: Body mass index; CAD: Coronary artery disease; CHF: Congestive heart failure; CKD: Chronic kidney disease; DM: Diabetes mellitus; pMI: Prior myocardial infarction; VAD; Implanted ventricular assist device; ICD; Internal cardiac defibrillator; HTN: Hypertension; COPD: Chronic obstructive pulmonary disease; pHTN: Pulmonary hypertension.

Among patients evaluated with intra arrest TEE in OHCA, 37% achieved ROSC, 21% survived to hospital admission, and 4% survived to hospital discharge. Of the 11 patients reported to survive to discharge, 5 had reported normal CPC, and mRS of 0–1 (no significant disability). Neurological outcomes was reported unknown / not available in the additional 6 patients. In the IHCA cohort evaluated intra-arrest, 43% had ROSC and 12% survived to hospital discharge. Of the 10 patients reported to survive to discharge, 3 had reported CPC of normal or moderate disability, 2 with moderate disability, and 3 with severe disability. Two were reported to have mRS of 0–1, and 5 with mRS 3–4 (moderate and severe disability). Neurological outcomes was reported unknown / not available in all remaining survivors.

Procedure characteristics

The majority of patients received TEE examinations in the ED (60%) and in the ICU (39%). These were performed by emergency physicians (EPs) in 57% of cases, with attending physicians being the operator in 70% of the studies. Probe insertion was performed without the assistance of laryngoscopy in 80% of cases, while direct laryngoscopy (DL) was used in 8%, and video laryngoscopy (VL) in 12% of cases. Eighty-three percent of cases reported successful probe insertion during the first attempt.

Impact of focused TEE in management of non-arrest cases

Focused TEE was reported to help support clinical decision making based on findings in 83% of patients in whom TEE was performed to evaluate undifferentiated shock and 47% of patients where TEE was used during hemodynamic monitoring. During procedural guidance, focused TEE was used in 68% of cases for cannulae placement in veno-arterial (VA) extracorporeal membrane oxygenation (ECMO), in 12% for veno-venous (VV) ECMO, in 5% for placement of intravenous pacemaker, and 3% and 1% for guidance of Impella intra-aortic ballon pumps respectively. Other endovascular procedures guided by TEE were reported in 13% of cases, which included its use assisting with pulmonary artery catheter (PAC), and central venous catheter (CVC) placement, and guidance of emergency pericardiocentesis.

TEE evaluation of cardiac arrest

Initial rhythms of OHCA were 36% asystole, 36% pulseless electrical activity (PEA), and 9% ventricular fibrillation (VF). For IHCA, initial rhythms were 14% asystole, 62% PEA, and 9% VF. Cardiopulmonary resuscitation (CPR) was performed using mechanical CPR devices in 36% OHCA and 25% IHCA cases. Focused TEE was reported to support clinical decision making based on findings in 71% of IHCA, and 62% of cases of OHCA. In OHCA, the most common intervention following TEE post-arrest included the decision to perform coronary angiography in 18%, the initiation of vasopressors for hemodynamic support in 11%, and the decision to initiate mechanical circulatory support in 6%. In IHCA, common interventions included the decision to initiate intravenous fluid administration in 19% and initiation of mechanical circulatory support 15%. Other interventions resulting from the findings of TEE were reported in 31% and 28% of patients in OHCA and IHCA respectively and included management changes such as the identification of wall-motion abnormalities suggesting an acute myocardial infarction, decision to not perform coronary angiography, and to withhold anticoagulation after finding of an acute aortic dissection. Table 2 summarizes changes in management based on TEE findings.

TABLE 2.

Change in management based on TEE findings in cardiac arrest.

	OHCA	IHCA
	N=81	N=93
Taken to cardiac catheterization lab	22% (18)	4% (4)
Vasopressors	16% (13)	13% (12)
Intravenous fluids	7% (6)	20% (19)
Mechanical circulatory support (i.e., ECMO)	6% (5)	16% (15)
Stop intravenous fluids	5% (4)	4% (4)
Thrombolytic agent (i.e., tPA)	2% (2)	9% (8)
Anticoagulation	2% (2)	0% (0)
Blood transfusion	1% (1)	2% (2)
Taken to the operating room	1% (1)	0% (0)
Other management change	36% (29)	31% (29%)

rTEECoRe: Resuscitative TEE Collaborative Registry; OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest; ECMO: Extracorporeal membrane oxygenation; tPA; Tissue plasminogen activator; OR: Operating room.

Pseudo PEA (an organized electrical rhythm with the presence of myocardial contractions observed on TEE), was reported in 16% of OHCA, and 27% of IHCA, and fine VF in 10% and 14% respectively. The initial AMC was determined to be over the left ventricle (LV) only in 28% (95% CI, 22.8 to 33.7%) OHCA cases, and 32% (95% CI, 22.2 to 42.6%) of IHCA cases. Following intra-arrest determination of AMC, chest compression location was modified under TEE guidance in 25% (95% CI, 19.2 to 30.5%) of OHCA patients and 25% (95% CI, 21.3 to 43.2%) of IHCA. In OHCA, the AMC was more frequently determined to be at the LV in patients resuscitated with manual CPR (39%; 95% CI, 30.1 to 48.9%) vs those with mechanical CPR (30%; 95% CI, 21.2 to 39.9%), or those who were resuscitated alternating both types of CPR (11%; 95% CI, 3.1 to 26%). The AMC in OHCA was changed under TEE guidance more often when CPR was being performed manually (43%; 95% CI, 26.2 to 46%) vs with mechanical device (39%; 95% CI, 9.1 to 25.8%). A similar, albeit not significant relationship was found in IHCA patients where AMC was found at the LV in 39% (95% CI, 25 to 54.6%) with manual CPR vs 36% (95% CI, 17.1 to 59.3%) undergoing mechanical CPR. Figure 3 shows characteristics of both cardiac arrest cohorts and main diagnostic findings.

FIGURE 3.

Characteristics and main diagnostic findings of TEE during cardiac arrest.

OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest; ManCPR: Manual cardiopulmonary resuscitation; MechCPR: Mechanical cardiopulmonary resuscitation; RV: Right ventricle; PEA: Pulseless electrical activity; VF: Ventricular fibrillation; AMC: Area of maximal compression; ME LAX: Mid esophageal long axis view.

Complications

Across all indications, there were 3 reported cases (0.4%) of pharyngeal bleed following probe insertion, and no cases of endotracheal tube dislodgement or cuff rupture. Among cardiac arrest cases, there were no reported esophageal perforations or oropharyngeal injuries. One case (0.1%) of upper gastrointestinal bleed (GIB) was captured as “other” complication in the cohort of evaluation of shock, and one (0.3%) was reported in a case of intra-arrest OHCA. Table 3 summarizes procedure and operator characteristics. The median duration of TEE examination was shortest in intra-arrest OHCA cases (13 min [IQR 8–20]), and longest in the evaluation of shock (16 min [IQR 14–29]). In 80% of examinations the TEE probe was inserted without assistance of laryngoscopy; insertion was successful in 83% of cases during the first insertion attempt. Table 4 summarizes additional procedure characteristics including the TEE views obtained during examinations.

TABLE 3.

Procedure and operator characteristics of focused TEE examinations.

	Total	Intra-arrest OHCA	Post-arrest OHCA	Intra-arrest evaluation in IHCA	Post-arrest evaluation in IHCA	Initial evaluation of undifferentiated shock	Hemodynamic monitoring	Procedural guidance
	N=1,045	N=278	N=159	N=88	N=60	N=267	N=92	N=101
Clinical unit
Emergency Department	59% (545)	98% (272)	97% (155)	68% (60)	53% (32)	14% (38)	22% (20)	64% (65)
Intensive Care Unit	39% (355)	1% (4)	2% (3)	26% (23)	43% (26)	85% (227)	76% (70)	32% (32)
Operating Room	0% (4)	0% (1)	0% (0)	0% (0)	2% (1)	0% (1)	1% (1)	1% (1)
Ward	0% (4)	0% (0)	0% (0)	3% (3)	2% (1)	0% (0)	0% (0)	1% (1)
Missing	1% (5)	0% (0)	1% (1)	0% (0)	0% (0)	0% (1)	1% (1)	1% (1)
Specialty of operator performing TEE
Emergency Medicine	57% (520)	85% (237)	90% (143)	69% (61)	53% (32)	16% (42)	42% (39)	47% (47)
Intensive Care	40% (363)	9% (26)	9% (14)	30% (26)	47% (28)	83% (221)	51% (47)	36% (36)
Cardiology	2% (18)	5% (13)	0% (0)	1% (1)	0% (0)	0% (0)	0% (0)	17% (17)
Anesthesiology	1% (10)	0% (1)	0% (0)	0% (0)	0% (0)	1% (4)	5% (5)	0% (0)
Missing	1% (5)	0% (1)	1% (1)	0% (0)	0% (0)	0% (1)	1% (1)	1% (1)
Level of operator performing TEE
Attending physician	70% (643)	74% (206)	67% (107)	81% (71)	72% (43)	74% (197)	50% (46)	65% (66)
Fellow	20% (186)	15% (41)	15% (24)	14% (12)	17% (10)	21% (56)	51% (42)	11% (11)
Resident	8% (75)	10% (28)	16% (25)	6% (5)	10% (6)	4% (12)	3% (3)	22% (22)
Other	1% (6)	0% (1)	1% (2)	0% (0)	2% (1)	1% (2)	0% (0)	1% (1)
Missing	1% (6)	0% (2)	1% (1)	0% (0)	0% (0)	0% (0)	1% (1)	1% (1)
Complications during probe insertion *
Pharyngeal bleed	0.3% (3)	0% (0)	0% (0)	1% (1)	2% (1)	1% (2)	0% (0)	1% (1)
Endotracheal tube dislodgement	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)
Endotracheal tube cuff rupture	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)
Other	0.2% (2)	0% (0)	0% (0)	0% (0)	0% (0)	1% (2)	0% (0)	0% (0)
Other complications **
Esophageal perforation	0% (0)	0% (0)	0% (0)	0% (0)	0% (0)	-	-	-
Oropharyngeal injury	0.1% (1)	0% (0)	0% (0)	0% (0)	2% (1)	-	-	-
Gastrointestinal bleed	0.3% (2)	0% (1)	1% (1)	0% (0)	0% (0)	-	-	-

Data are presented as median (IQR) for continuous measures, and % (n) for categorical measures. Given the relevance of precise numbers the % in complications is presented with one decimal. For all other variables the % has been rounded to the nearest tenth.

^*The total (first column) represents the unique number of reported complications. The total is not always equal to the sum of all seven indications (e.g., pharyngeal bleed) because some of the same complications are captured in more than one indication but correspond to the same event.

^**These complications were only captured in the cardiac arrest cohorts. rTEECoRe: Resuscitative TEE Collaborative Registry; OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest.

TABLE 4.

Additional procedure characteristics of focused TEE examinations.

	Total	Intra-arrest OHCA	Post-arrest OHCA	Intra-arrest evaluation in IHCA	Post-arrest evaluation in IHCA	Initial evaluation of undifferentiated shock	Hemodynamic monitoring	Procedural guidance
	N=1,045	N=278	N=159	N=88	N=60	N=267	N=92	N=101
Procedure duration (minutes) Technique used for probe insertion	15 (8-24)	10 (5-20)	12 (7-18)	15 (8-25)	14 (10-23)	20 (12-29)	22 (14-35)	11 (5-31)
No laryngoscopy used	79% (722)	69% (192)	89% (142)	73% (64)	82% (49)	83% (222)	90% (83)	77% (78)
Direct laryngoscopy used	7% (67)	13% (35)	4% (7)	12% (11)	3% (2)	4% (11)	2% (2)	17% (17)
Video laryngoscopy used Missing	12% (108)	15% (41)	5% (8)	15% (13)	15% (9)	13% (34)	7% (6)	5% (5)
Missing	2% (19)	4% (10)	01% (2)	0% (0)	0% (0)	0% (0)	1% (1)	1% (1)
Number of probe insertion attempts
1	81% (739)	80% (222)	84% (134)	90% (79)	75% (45)	79% (211)	78% (72)	84% (85)
2	12% (113)	10% (28)	11% (17)	7% (6)	12% (7)	18% (47)	13% (12)	10% (10)
3	3% (28)	4% (12)	1% (2)	2% (2)	7% (4)	2% (6)	5% (5)	4% (4)
4	1% (5)	0% (1)	1% (1)	0% (0)	2% (1)	1% (2)	2% (2)	0% (0)
5	0% (3)	0% (0)	0% (0)	1% (1)	5% (3)	0% (0)	0% (0)	1% (1)
8	0% (1)	0% (0)	0% (0)	0% (0)	0% (0)	0% (1)	0% (0)	0% (0)
Missing	3% (23)	5% (14)	1% (2)	1% (1)	0% (0)	0% (0)	1% (1)	1% (1)
TEE views obtained
ME 4C		79% (214)	93% (131)	88% (69)	75% (45)	94% (252)	95% (87)	38% (38)
ME LAX		80% (216)	89% (125)	90% (70)	72% (43)	91% (244)	92% (85)	38% (38)
ME BC		32% (87)	60% (84)	40% (31)	58% (35)	83% (222)	70% (64)	88% (89)
TG SAX		40% (107)	72% (101)	54% (42)	58% (35)	88% (234)	79% (73)	27% (27)
ME 2C		17% (46)	35% (50)	19% (15)	23% (14)	60% (161)	72% (66)	18% (18)
AV SAX		8% (21)	30% (43)	14% (11)	27% (16)	54% (145)	62% (57)	10% (10)
UE Asc. SAX; main PA view		10% (26)	19% (27)	9% (7)	25% (15)	51% (136)	45% (41)	8% (8)
UE Asc. LAX		5% (14)	15% (21)	4% (3)	8% (5)	27% (73)	26% (24)	5% (5)
ME RV I/O		6% (16)	28% (40)	12% (9)	42% (25)	64% (170)	65% (60)	16% (16)
ME DTA SAX		23% (63)	38% (54)	24% (19)	43% (26)	63% (168)	54% (50)	66% (67)
ME DTA LAX		18% (48)	24% (34)	15% (12)	27% (16)	48% (129)	35% (32)	45% (45)
TG LAX		1% (4)	7% (10)	5% (4)	12% (7)	45% (121)	28% (26)	5% (5)
dTG 5C		1% (3)	10% (14)	3% (2)	17% (10)	46% (122)	24% (22)	3% (3)
ME 5C		3% (9)	10% (14)	3% (2)	12% (7)	22% (60)	34% (31)	1% (1)

Data are presented as median (IQR) for continuous measures, and % (n) for categorical measures. The % has been rounded to the nearest tenth OHCA: Out-of-hospital cardiac arrest; IHCA: In-hospital cardiac arrest; ME 4C: Mid esophageal 4 chamber view; ME LAX: Mid esophageal long axis view; ME BC: Mid esophageal bicaval view; TG SAX: Transgastric short axis view; ME 2C: Mid esophageal 2 chamber view; AV SAX: Aortic valve short axis: UA Asc. SAX: Upper esophageal ascending aorta short axis view; UA Asc. LAX: Upper esophageal ascending aorta long axis view; ME RV I/O: Mid esophageal right ventricular inflow/outflow view; ME DTA SAX: Mid esophageal descending aorta short axis view; ME DTA LAX: Mid esophageal descending aorta long axis view; TG LAX: Trans gastric long axis view; dTG 5C: Deep transgastric 5 chamber view; ME 5C: Mid esophageal 5 chamber view.

LIMITATIONS

The data gathered in this multicenter registry relies on reporting from individual clinicians and the site PIs performing the studies and therefore are subject to recall and recording bias. This is particularly important in the case of clinician’s findings during the TEE exam including the reported change in management resulting from TEE, as well as procedure-related complications. Furthermore, our outcome measure of change in management is reported by the clinicians performing the studies, and therefore is subject to confirmation bias. This user bias along with the lack of a comparator (e.g., TTE, other diagnostic tool, or no diagnostic tool), represent an important limitation in the interpretation of clinical impact of the data produced by focused TEE, clinical care, and patient outcomes. To this end, multicenter research network will provide an ideal infrastructure to conduct trials that can assess causality and validate the impact of this intervention in specific applications such as cardiac arrest resuscitation care.

The lack of concealment, which is inherent to the nature of this diagnostic modality, is also a source of potential bias in the reporting of changes in management resulting from TEE. Another limitation inherent to this design is the potential for selection, and reporting bias both which could affect our outcomes of interest. Given the pragmatic nature of this study, patients who receive focused TEE for any of the applications of interest represent effectively a convenience sample. In most participating units, focused TEE is only performed when a physician trained in TEE is available, and therefore not all patients with a clinical indication for TEE in any of these applications had this diagnostic intervention. We have mitigated this by requiring all participating sites to maintain a minimum of 90% of eligible cases reported. The study team monitors this case capture by surveying each site monthly. Furthermore, it is possible that some of participating sites have similar (e.g. focused) TEE studies performed in other units potentially eligible for enrollment in this registry that we are not currently capturing.

The purpose of specific TEE examinations was in some cases purely diagnostic (i.e. to evaluate for cardiac thrombus, tamponade, or other reversible pathologies), or prognostic (i.e. characterization of myocardial activity), and in other cases was more interventional (i.e. to guide the quality of CPR, the implementation of ECMO, or other resuscitative interventions). Furthermore, some exams were also performed for other indications not relevant to the study (e.g. assessment for a cardiac source of embolism or evaluation of shunt in refractory hypoxia), potentially lengthening the exams beyond that required to address the specific clinical applications of interest in the study. Future studies will more carefully separate the various TEE goals and study them in more detail. As a pragmatic study, operator levels of experience varied from site to site, which may have led to variable quality of examinations. Furthrmore, besides the level (e.g. attending, fellow, or resident) of the clinician performing the study, our study did not capture more specific data regarding the level of training or experience of the individual clinicians performing the studies entered in the registry. Based on the practice standards at least in US and Canadian centers, all studies performed by residents or fellows were likely staffed and supervised directly by an attending physician.

Local decisions on TEE implementation may have led to some bias in the nature of included cases. Future work will attempt to better characterize site-specific use and the denominator of eligible patients.

DISCUSSION

In this initial report from the rTEECoRE research network, we present an analysis of prospectively collected data characterizing the use of resuscitative TEE in critically ill patients from a large international multicenter cohort. These data represent the most substantial evidence to date demonstrating that focused TEE is feasible, safe, and clinically impactful, across the applications of cardiac arrest resuscitation, evaluation of undifferentiated shock, hemodynamic monitoring, and procedural guidance. This work represents the first multicenter study of resuscitative TEE and extends prior single center studies characterizing the role of TEE in emergency and intensive care settings.^{6,7,8,9,15,16} The population described in this cohort of patients is representative of critically ill patients in whom TEE studies are often performed including a greater proportion of male patients (68%), and high prevalence of obesity (median BMI of 30), which represents an important limitation for transthoracic echocardiography.^14,15,16 Cases involving cardiac arrest (specifically OHCA) represent the majority of enrolled cases. This is likely explained by the predominance (59%) of EDs among the clinical units participating from the study.

Studies were performed primarily by attending physicians (69%), followed by fellows (22%), with no significant differences of this ratio across different applications of TEE. The fact that resuscitative TEE is still a relatively new point of care ultrasound modality that requires specific training may explain this finding. Operators were from EM in 59% of cases and ICU in 39% of cases. While training standards, certification, and credentialing vary across specialties (e.g. EM vs Critical Care Medicine) and across institutions, the training recommendations for emergency physicians is guided by national guidelines, which suggest a minimum of 4–6 hours of structured simulation training and a minimum of 10 live proctored examinations.²

While laryngoscopy was used infrequently during TEE probe insertion, this was accomplished during the first attempt in the vast majority (81%) of cases and reported complications were rare. This an interesting finding in light of recent evidence indicating that the use of video laryngoscope may improve success of probe placement.²² Future analysis will determine whether the use of laryngoscopy is associated with fewer attempts and fewer complications in our cohort.

Across all indications, TEE was found have changed management for 50–85% of patients by informing decision-making and interventions based on point-of-care findings. While our data captured the key interventions resulting from focused TEE, review of individual cases shows that TEE commonly leads to more than one intervention, highlighting the versatility of this modality which can provide diagnostic information (e.g., RV dilation), guide therapy (e.g., initiation of IV fluids or vasopressors), and assist with endovascular procedures. For instance, a patient in whom TEE was used to evaluate undifferentiated shock was confirmed to be in cardiogenic shock (diagnostic role), and subsequently placed on emergency VA-ECMO, and placement of cannulae was facilitated by TEE (procedural guidance role).

In cardiac arrest, transthoracic echocardiography (TTE) performed during resuscitation has been shown to help identify patients with poor prognosis, detect reversible pathology and guide ongoing resuscitation efforts.^25,26 A large multicenter study demonstrated that patients with reversible causes of cardiac arrest such as pericardial tamponade and patients with organized myocardial contractions (pseudo-PEA) carry the highest probability of survival compared to patients with no treatable pathology and cardiac standstill, respectively.²⁵ However, an important limitation of TTE in resuscitation is the technical difficulty in obtaining cardiac windows with a study reporting inadequate transthoracic images in up to 20% of patients²⁷ and some retrospective studies suggesting that this may result in longer CPR interruptions. ^28,29

In this context, and due to its retrocardiac position, TEE has been described as an ideal imaging modality during resuscitation.^6,10 Our data confirm that TEE can identify potentially treatable pathology and characterize myocardial activity, including the identification of patients with pseudo-PEA and fine VF. Based on exam findings, TEE can help direct provider teams to initiate post-arrest interventions such as the decision to undergo emergent cardiac catheterization, initiate vasopressor therapy, and utilize mechanical circulatory support.

A unique advantage of TEE, as compared to TTE, is that it allows real-time imaging of the heart during CPR. Over the past two decades, several studies have shown that due to significant anatomic variability, the standard CPR compression point at the inter-nipple line may not always provide effective chest compressions in cardiac arrest patients, suggesting that specific compression location on the chest may strongly influence CPR effectiveness. ^30–32 The proposed mechanistic principle underpinning this observation is that in many patients, due to the anatomic location of the LVOT and the ascending aorta, standard chest compression positioning may markedly impede forward blood flow.^6,33,34 Furthermore, animal research has shown that CPR performed directly over the LV rather than over the LVOT results in higher diastolic BP, coronary perfusion pressures, ETCO2, cerebral perfusion, and ROSC.^35–36 To this end, our data are consistent with previous single center studies, demonstrating that only a third of patients receiving standard, guideline-recommended CPR had echocardiographic evidence of compression of the LV during resuscitation, and that this real-time information provided by TEE led to a change in the chest compression location in a similar proportion of patients (27% in OHCA and 36% in IHCA). Interestingly, we found that patients receiving manual CPR were more likely to have adequate AMC over the LV. These patients were also more likely to have the AMC changed in those cases where AMC was not at the LV, compared to those with mechanical CPR, and those alternating between the two types of CPR. This could be potentially explained by the fact that current generation of mechanical CPR devices, are not designed to be easily moved or repositioned during resuscitation. Future dedicated analyses of this cardiac arrest cohort, controlling for relevant factors, will allow us to further understand the potential implications of different types of chest compressions. To date, several clinical trials have demonstrated no difference in survival or neurological outcomes between manual and mechanical CPR, however none of these studies have information regarding the anatomical effects of these different chest compression strategies as that which may be provided by intra-arrest TEE. ³⁷ Of note, while we have reported the outcomes data currently available for the cohorts of cardiac arrest, it was not the goal of this study to establish an outcomes benefit from TEE. The survival to hospital discharge rates of our cohort (4% OHCA; 12% IHCA), is consistent with the overall low survival rates reported in cardiac arrest patients nationally.^38,39 This both highlights the challenge of conducting outcome-oriented research to evaluate interventions in cardiac arrest, which require a large sample, but also the unique opportunity to improve survival if novel targets such as the AMC, can be identified.

Our cohort of cardiac arrest patients represents primarily those with non-shockable rhythms in whom CPR represents a crucial outcome-modifying therapy. Current resuscitation guidelines emphasize the need to monitor the quality of CPR, as high-quality CPR is strongly associated with survival.^39–40 Traditional CPR quality metrics such as chest compression depth and rate have supported the development of practical “one-size-fits-all” consensus recommendations for delivery, but they do not reflect individual physiologic needs or anatomic variation. A growing body of animal and human studies have indicated that physiologic parameters, such as ETCO₂ and diastolic blood pressure (DBP), provide valuable information about response to resuscitation and may improve survival ^41,42 This physiology-guided approach to CPR has been recommended by expert consensus and current resuscitation guidelines.^39,40 Our registry database, particularly given the availability of TEE images along with hemodynamic, physiologic and outcomes data, represents a unique opportunity to advance our understanding of CPR and to improve resuscitation outcomes. In doing so, TEE informed resuscitation could represent a paradigm change in cardiac arrest, shifting from population-based strategies to individualized, physiology-guided care. To this end, an ongoing project funded by the National Institutes of Health (K23 HL165150–01A1) will utilize the rTEECoRe network to evaluate the impact of TEE-guided chest compression location on hemodynamic endpoints and survival, and to characterize implementation determinants, barriers, and facilitators of TEE during resuscitation.⁴³

Our registry infrastructure will support several future studies, including retrospective data review of specific conditions and management decisions using TEE, as well as post hoc analyses of the bank of video images gathered from this registry. To that end, this large repository of TEE images represents a unique feature of this registry and an ideal resource to derive and validate a standardized protocol for the deployment of TEE-guided resuscitation. This work will be aided by the development of image analysis algorithms assisted by machine learning to improve standardization of TEE image interpretation. As interest in TEE continues to grow, it will be important to develop procedural competency evaluation methods and scalable training approaches. Our network investigators also envision the potential development of a federally funded research program in TEE, analogous to the Collaborative Pediatric Critical Care Research Network (CPCCRN).^44,45

In conclusion, a prospective, multicenter, and multidisciplinary registry evaluating the use of focused TEE across acute care settings was successfully implemented. Its initial study demonstrates that focused TEE is safe and clinically impactful in ED and ICU settings for the purpose of acute diagnostic evaluation during critical illness and resuscitative interventions such as CPR. In many cases TEE generated diagnostic information that led to specific management changes. Further work will be required to assess whether TEE use can improve outcomes in specific resuscitation conditions. Further studies from this research network will accelerate the development of outcome-oriented research and knowledge translation on the use of TEE in emergency and critical care settings.