Impact Of Transesophageal Echocardiography-guided Hemodynamic Optimization On Fast-track Recovery After Cardiac Surgery

Summary

This retrospective study evaluates the impact of intraoperative transesophageal echocardiography-guided hemodynamic optimization on fast-track recovery in cardiac surgery. By implementing a closed-loop protocol focused on volume phenotyping and cardiac function, the intervention was associated with shorter extubation time and reduced ICU length of stay compared with standard care.

Abstract

Perioperative volume misclassification and cardiac dysfunction often hinder fast-track recovery following cardiac surgery. This retrospective cohort study evaluated whether intraoperative transesophageal echocardiography (TEE)-guided hemodynamic optimization was associated with improved fast-track outcomes. Seventy-two adult patients undergoing cardiac surgery received either TEE-guided optimization (n = 34) or standard care (n = 38). The intervention involved goal-directed hemodynamic management based on TEE phenotyping of volume status and cardiac function at predefined intraoperative time points, including post-induction, prior to cardiopulmonary bypass, during weaning from bypass, and before chest closure. Clinical decisions were guided by predefined objective criteria, including left ventricular end-diastolic area and stroke volume responsiveness for fluid management, and ejection fraction and regional wall motion abnormalities for inotropic support. Inverse probability of treatment weighting was used to adjust for baseline imbalances. TEE-guided optimization was associated with shorter time to extubation and reduced intensive care unit length of stay compared with standard care. The intervention group also showed improved early postoperative perfusion profiles without an observed increase in major complications. These findings suggest that intraoperative TEE-guided hemodynamic optimization may support fast-track recovery after cardiac surgery, although larger prospective studies are needed to confirm these results.

Introduction

Cardiac surgery, particularly valve replacement or major vascular procedures involving extracorporeal circulation, induces complex pathophysiological changes¹. Surgical trauma, systemic inflammatory responses induced by extracorporeal circulation, and myocardial ischemia-reperfusion injury are primary drivers of intraoperative hemodynamic fluctuations. Driven by the principles of accelerated recovery after surgery (ERAS), the Fast Track Cardiac Surgery Protocol (FTCP) has become a critical pathway to improving patient outcomes^2,3. This protocol aims to minimize multi-organ stress through refined intraoperative management, thereby enabling early extubation and shortening the duration of intensive care unit (ICU) stay⁴. However, the successful implementation of the FTCP is highly dependent on intraoperative hemodynamic stability. Intraoperative occult hypotension or volume overload has been demonstrated to be closely associated with postoperative acute kidney injury, myocardial injury, and neurological complications⁵. Consequently, achieving precise intraoperative hemodynamic optimization remains a critical clinical challenge in the field of cardiac surgery.

In traditional clinical practice, anesthesiologists primarily rely on pressure parameters such as central venous pressure (CVP), mean arterial pressure (MAP), and urine output to assess patients' volume status. However, these static indicators poorly reflect volume responsiveness under complex physiological conditions⁶. CVP is susceptible to interference from multiple factors, including cardiac valve function, ventricular compliance, and positive pressure ventilation, often failing to accurately represent the true left ventricular preload⁷.

Reliance on this pressure-driven monitoring paradigm often leads to empirical treatment strategies. Indiscriminate volume expansion to maintain blood pressure may lead to interstitial edema and increased right ventricular functional load, impeding postoperative pulmonary recovery⁸. Conversely, the use of vasoactive drugs when volume targets are not met may temporarily stabilize blood pressure readings but result in inadequate microcirculatory perfusion. The inherent uncertainty of this monitoring approach often serves as a primary contributing factor to delayed postoperative recovery^9,10.

The clinical application of transesophageal echocardiography (TEE) provides a valuable paradigm for visualizing intraoperative hemodynamic management. Unlike pressure-based monitoring methods, this technique enables real-time acquisition of volumetric and dynamic parameters such as left ventricular end-diastolic area, stroke volume, and ejection fraction¹¹. Conventional tools such as pulmonary artery catheters and uncalibrated pulse contour analysis provide continuous monitoring, but they remain fundamentally pressure-derived and may not accurately reflect right-sided filling pressures or localized contractility deficits during acute physiological shifts¹¹. By directly observing cardiac sections, physicians can accurately assess ventricular filling status and myocardial contractility. This technology holds unique value in identifying occult cardiac dysfunction, allowing immediate detection of regional ventricular wall motion abnormalities or changes in the degree of valvular regurgitation¹². The shift from pressure monitoring to functional visualization may enable treatment decisions to be based on real-time anatomical evidence rather than surrogate estimations. Through repeated evaluation and correction at critical surgical junctures, clinicians can establish a closed-loop decision-making framework that integrates volume status assessment with cardiac function optimization, thereby supporting physiological matching¹³. This approach may be particularly relevant in higher-risk clinical scenarios, including patients with preexisting valvular lesions, occult right ventricular dysfunction, or borderline ejection fractions, who may be more susceptible to acute congestion and hypoperfusion during complex fluid shifts. Accordingly, the primary goal of the current study was to evaluate whether TEE-guided hemodynamic optimization was associated with improved fast-track recovery after cardiac surgery.

The current research focuses on whether refined intraoperative hemodynamic management can translate into clinical pathway-based rehabilitation benefits by improving organ perfusion and perioperative hemodynamic stability. This study aimed to evaluate the impact of volume and function optimization protocols guided by transesophageal echocardiography (TEE) on postoperative weaning time and early recovery quality in cardiac surgery patients. Given its retrospective design, this study is primarily hypothesis-generating. It seeks to provide clinical insight into whether visualization-guided management may support accelerated postoperative recovery pathways in cardiac surgery.

Protocol

The research protocol was reviewed and approved by the Institutional Review Board of the First Affiliated Hospital of China Medical University (AF-SOP-07-1.2-01). All investigative procedures were conducted in strict adherence to the ethical standards of the institutional research committee and the 1964 Helsinki Declaration. Given the retrospective study design and the use of anonymized patient data, the committee formally waived the requirement for informed consent.

Study design

This study employed a retrospective cohort design, utilizing a single-center cardiac surgery perioperative database and electronic medical record system to identify and include eligible patients based on clinical records from January 2021 to December 2023. Patients were retrospectively categorized into two groups based on their documented perioperative management strategies: the intraoperative TEE-guided optimization group (n = 34) and the standard hemodynamic management group (n = 38). Both groups received standard anesthesia monitoring and perioperative management, but the intervention group underwent standardized TEE evaluation at critical intraoperative time points and implemented targeted volume and cardiac function adjustments based on the evaluation results, establishing an evaluation-intervention-reevaluation closed-loop management process. The control group primarily relied on conventional monitoring indicators and clinical experience for fluid and medication decisions, without following the predefined TEE closed-loop optimization pathway. The study protocol, key time point settings for perioperative workflows, closed-loop decision-making logic, and postoperative rapid pathway transition strategies are illustrated in Figure 1.

To mitigate the impact of baseline differences and selection bias on outcome inference in retrospective studies, this study implemented classical confounding control strategies during statistical analysis to adjust for variations in preoperative risk factors, surgical complexity, and perioperative critical parameters between the two groups. The primary outcomes focused on core indicators of rapid pathway rehabilitation, while concurrently evaluating safety and efficacy endpoints, including ICU stay duration, hospitalization days, and major perioperative complications. This systematic evaluation aimed to assess the potential benefits of the TEE-guided closed-loop optimization strategy in real-world cardiac surgery populations and its clinical generalizability.

Participants

This study retrospectively included patients who underwent cardiac surgical procedures and completed perioperative management at our center within the study time window. To ensure subject homogeneity and comparability between intraoperative TEE evaluation and perioperative fast-track outcomes, the inclusion and exclusion criteria were as follows.

Inclusion criteria were as follows: (1) cardiac surgery patients aged ≥18 years, with surgical types including coronary artery bypass grafting (CABG), valvular surgery, or conventional cardiac surgical procedures such as CABG combined with valvular surgery; (2) completion of standardized general anesthesia and routine perioperative monitoring, with complete perioperative critical data, including at least surgical and anesthesia records, intraoperative fluid and vasoactive drug usage records, ICU records, and extubation-related information; (3) patients who underwent intraoperative TEE and obtained valid evaluation data, or those who underwent TEE but whose perioperative management pathway could be clearly classified as either the TEE-guided closed-loop optimization group or the standard management group; and (4) postoperatively, patients who were admitted to the ICU or underwent anesthesia recovery pathway and received rapid pathway rehabilitation management evaluation, with measurable primary outcomes and secondary outcomes.

Exclusion criteria were as follows: (1) patients with contraindications to TEE or conditions that prevented safe TEE placement or valid image acquisition, such as confirmed severe esophageal lesions (e.g., esophageal strictures, esophageal varices with high bleeding risk, esophageal tumors, or perforation history) or other anesthesia evaluations deemed unsuitable for TEE placement; (2) redo sternotomy, emergency or life-saving surgeries, preoperative cardiogenic shock, or cases requiring immediate extracorporeal life support (intraoperative or postoperative ECMO) or perioperative instability that precluded standardized comparable evaluation and management protocols; (3) patients with complex structural heart disease or anatomical abnormalities, severe arrhythmias, or other factors that impaired reliable interpretation of TEE key parameters or significantly interfered with volume state and cardiac function assessment; (4) missing or incomplete perioperative core data, particularly inability to obtain primary outcomes or critical exposure information, including implementation of TEE closed-loop optimization and key node records; and (5) death within 24 h postoperatively or discontinuation of treatment or transfer for non-study-related reasons that prevented complete outcome tracking.

TEE-guided optimization protocol

This study implemented a closed-loop optimization strategy for volume state and cardiac function guided by intraoperative transesophageal echocardiography (TEE) in the intervention group. The protocol involved a continuous decision-making chain of phenotype identification, cardiac function assessment, targeted intervention, and immediate reevaluation using standardized TEE planes at critical perioperative time points, thereby maintaining effective perfusion and promoting rapid postoperative recovery without unnecessary volume loading. TEE examinations were performed by qualified anesthesiologists using a multiplane adult probe connected to a high-end echocardiography system, in accordance with a standardized institutional imaging protocol. Baseline echocardiographic assessments were performed immediately after anesthesia induction, with subsequent structured evaluations at key time points, including before cardiopulmonary bypass (CPB), during pump removal or weaning from CPB, and before chest closure. Each evaluation followed uniform plane acquisition and measurement protocols to ensure comparability and traceability across different time points.

Regarding volume status assessment, the protocol focused on cardiac filling and volume tolerance phenotypes, integrating left ventricular filling status, right heart load characteristics, and perfusion-related ultrasonographic clues for stratified evaluation. Specifically, the protocol prioritized acquiring key cross-sectional views, including the mid-esophageal four-chamber view, the mid-esophageal two-chamber view, and the transgastric left ventricular short-axis view. Precise measurements of left ventricular end-diastolic area, stroke volume, and ejection fraction were recorded at each predefined milestone. Qualitative and semiquantitative assessments were completed with particular attention to left ventricular end-diastolic filling, left-right cardiac interaction indicated by ventricular septal morphology and motion, and left atrial pressure elevation trends consistent with volume overload. Left atrial pressure was indirectly estimated using the transmitral early diastolic velocity to the early tissue Doppler velocity ratio (E/e′ ratio). In cardiac function assessment, the protocol emphasized the synchronous identification of common clinical phenotypes such as left ventricular systolic dysfunction, right heart impairment, or transient myocardial suppression after pump removal. Reusable functional indicators were selected based on intraoperative circulatory status for trend analysis to support targeted and individualized subsequent interventions. The key TEE views and measurement points for acquisition and interpretation are systematically presented in Figure 2 to facilitate consistent evaluation across cases and time points.

For intervention and reevaluation, this study adopted a predefined algorithmic decision-making framework, wherein the intervention pathway relied on volume responsiveness, volume tolerance, and cardiac functional reserve. Positive volume responsiveness was defined as a stroke volume increase of >10% following a standardized 250 mL crystalloid fluid challenge. For patients with indications of volume responsiveness and volume tolerance, a low-dose fluid challenge was performed, followed by rapid reevaluation of left ventricular end-diastolic area, stroke volume, and perfusion-related ultrasound clues shortly after intervention to confirm whether effective preload improvement and cardiac output enhancement had been achieved. For patients with indications of poor volume tolerance or risk of congestion, empirical fluid replacement was avoided, and instead, vasoactive drugs were prioritized to regulate perfusion pressure or measures such as reducing afterload and adjusting ventilation parameters were employed to improve cardiac load conditions. If a phenotype of contractile insufficiency or right ventricular dysfunction was observed, or if the left ventricular ejection fraction fell below 45% or new regional wall motion abnormalities emerged relative to the post-induction baseline, positive inotropic support was initiated while maintaining adequate coronary perfusion pressure, and the directionality and effectiveness of the intervention were re-evaluated via TEE. Clinical interventions were documented within a structured evaluation-intervention-reevaluation framework. The closed-loop algorithm of phenotype identification, targeted intervention, and immediate reevaluation, and the key decision nodes and operational pathways under different phenotypes are visually summarized in Figure 3 to ensure clinical consistency and minimize bias caused by experiential differences.

Standard care description

The control group followed the center’s established perioperative anesthesia and hemodynamic management protocols for cardiac surgery, aiming to maintain circulatory stability, ensure organ perfusion, and minimize risks associated with hypoperfusion and volume overload at different surgical stages. All patients underwent invasive arterial pressure monitoring in addition to standard monitoring, with dynamic evaluation incorporating perfusion-related parameters such as heart rate, rhythm, urine output, blood gas analysis, and lactate levels. Perioperative management adhered to a clinical stratification approach, with comprehensive adjustments to volume management, vasoactive drugs, and positive inotropic support based on hemodynamic changes to meet the physiological demands during the induction phase, the transition period before and after cardiopulmonary bypass, and the perioperative period.

In volume management, the control group typically made fluid replacement decisions based on factors such as blood loss, urine output, arterial pressure levels, circulatory fluctuation amplitude, and changes in anesthesia depth. The primary fluid type for rehydration was crystalloids, supplemented with colloids and blood products when necessary to correct volume deficit and decreased hemoglobin levels. To address extracorporeal circulation-related blood dilution, altered volume distribution, and transient circulatory instability after pump removal, the control group primarily used a conventional indicator-driven fluid replacement and transfusion strategy, emphasizing gradual adjustment of volume load while maintaining acceptable hemodynamics. However, no structured interpretation or stratified management of volume responsiveness or volume tolerance phenotypes was performed.

In the management of vasoactive drugs and positive inotropic support, the control group primarily focused on maintaining target arterial pressure for hypotension or insufficient perfusion pressure. This was typically achieved through titration of pressor medications to correct anesthetic-related vasodilation, reduced vascular tone after cardiopulmonary bypass, or hypovolemia-induced hypotension. If clinical manifestations such as inadequate cardiac output or decreased myocardial contractility after pump removal were observed, positive inotropic agents were administered based on clinical judgment, combined with volume management and ventilation parameter adjustments for comprehensive treatment. For arrhythmias or unstable heart rates, pharmacological interventions or electrical cardioversion were performed according to conventional principles of extracorporeal anesthesia to restore acceptable hemodynamic status.

Regarding monitoring methods, the control group utilized transesophageal echocardiography only when clinically indicated, such as when valvular dysfunction, intracardiac thrombosis, or pump disengagement was suspected. However, the control group did not implement a predefined TEE-guided closed-loop optimization protocol, nor was there a requirement to complete standardized cross-sectional acquisition and recording at fixed key time points or to mandate immediate post-intervention TEE reevaluation and structured decision-making chains. Consequently, TEE utilization in the control group was more problem-oriented or empirically assisted decision-making rather than serving as a core tool for goal-directed volume and cardiac function optimization.

Regarding postoperative transition, the control group was managed according to the conventional cardiac monitoring and extubation protocol after ICU admission. Extubation and weaning assessments were typically based on comprehensive criteria including circulatory stability, oxygenation and ventilation status, recovery of consciousness, and levels of analgesia and sedation, with the ICU team determining the weaning rhythm based on clinical manifestations and routine indicators. The control group also adhered to the fundamental requirements of the fast-track approach, but its perioperative management lacked the structured physiological phenotypic support provided by intraoperative TEE closed-loop optimization, thus relying more on traditional indicators and clinical experience for pathway advancement and adjustment. Through the above description, the management strategies of the control group were clearly distinguished from those of the intervention group in terms of decision-making basis, standardization of assessment timing, and post-intervention reevaluation mechanisms, providing a foundation for subsequent comparison of the impact of the two strategies on fast-track outcomes.

Outcomes, definitions, and time windows

This study established outcomes based on core indicators of rapid pathway recovery in cardiac surgery. The primary outcomes focused on extubation-related metrics to reflect the achievement of rapid pathway criteria, specifically including extubation time or delayed extubation events. Delayed extubation was defined using a clinically common threshold, namely postoperative mechanical ventilation lasting more than 24 h. The observation period for the primary outcomes was typically set at the end of surgery or the time of ICU admission, with the endpoint being the completion of extubation or meeting the criteria for delayed extubation.

Secondary outcomes were used to evaluate the strategy’s efficacy and safety. Efficiency outcomes included ICU stay duration and hospitalization duration. Safety outcomes covered perioperative major complications, with key focuses on acute kidney injury (AKI), low cardiac output (LCO)-related events, pulmonary complications, reintubation, atrial fibrillation (AF), major bleeding-related events, and infections. AKI was assessed using the KDIGO staging criteria, pulmonary complications were classified based on clinical diagnosis and imaging support, and LCO-related events were determined by clinical records and supportive therapy needs. The observation window for complications was primarily the postoperative hospitalization period and was extended to short-term readmissions or re-ICU admissions after discharge if follow-up records were available in the medical record system. All outcomes were uniformly extracted and verified based on electronic medical records to ensure consistency and reproducibility of definitions.

Covariates, missing data, and quality control

To mitigate the influence of confounding factors in retrospective studies on outcome comparisons, this study predefined and extracted covariates closely associated with the fast-track outcomes, encompassing baseline patient risk factors, surgical complexity, and key perioperative intervention factors. Baseline covariates included age, gender, body mass index (BMI), baseline cardiac function, and major comorbidities. Perioperative covariates focused on surgical type, extracorporeal circulation (CPB) duration, aortic cross-clamping time, anesthesia and operative duration, intraoperative blood loss and transfusion, total fluid volume, as well as the use of vasoactive drugs and positive inotropic agents. These variables were used both to assess comparability between the two groups and to adjust for confounding in subsequent propensity score methods and multivariate models, thereby enhancing the robustness of causal inference.

Missing data were managed conservatively. First, completeness verification was conducted for key exposure information and primary outcomes. If primary outcomes or grouping information were missing, they were excluded from analysis. For minor missing values in covariates, complete case analysis was prioritized, and the impact of missing data on result stability was evaluated in sensitivity analyses. When key covariates had a high proportion of missing values, appropriate handling based on the missing mechanism was implemented, with explicit explanations in the results to avoid introducing systematic bias.

For quality control, this study evaluated the execution intensity of the TEE closed-loop optimization pathway using structured records, including whether TEE assessments were completed at key time points, whether post-intervention reevaluation records were available, and whether perioperative critical medications and fluid information were traceable. Data extraction was performed by trained researchers according to a predefined variable dictionary, with cross-verification conducted when necessary. Abnormal values and logically inconsistent records were reviewed against original medical records for confirmation. By standardizing variable definitions, unifying observation windows, and implementing stratified verification procedures, this study minimized information bias and enhanced the credibility and reproducibility of the results.

Statistical analysis

This study utilized standard statistical software for data processing and analysis. Continuous variables were first assessed for distribution characteristics. Data with approximately normal distribution were expressed as mean ± standard deviation, and intergroup comparisons were performed using independent samples tests. Data with skewed distribution were presented as median and quartiles, and intergroup comparisons were conducted using nonparametric tests. Categorical variables were expressed as frequency and percentage, and intergroup comparisons were performed using chi-square test or Fisher’s exact test. All tests were two-sided, with a P-value < 0.05 as the criterion for statistical significance.

Given that this study was designed as a retrospective cohort, inverse probability of treatment weighting (IPTW) was employed for primary analysis to control for confounding factors and mitigate the impact of baseline differences and selection bias on outcome inference. Propensity scores were estimated using binary logistic regression, with covariates selected based on clinical relevance and prior knowledge, covering demographic characteristics, underlying diseases, preoperative cardiac and renal function indices, as well as perioperative key parameters and surgical procedure categories reflecting operative complexity. Essential covariates included age, gender, EuroSCORE II, baseline left ventricular ejection fraction, and CPB duration. Subsequently, inverse probability weighting was applied for weighted analysis, and the weight distribution was validated to avoid instability caused by extreme weights. Before and after weighting, standardized mean differences (SMD) were used to assess covariate balance, with SMD values < 0.10 indicating good balance. The balanced diagnostic results are presented in Table 1.

The outcome analysis was conducted under a weighted framework. The primary outcomes were extubation-related metrics, specifically time to extubation and delayed extubation events, and effect estimates were reported with corresponding confidence intervals. For continuous outcomes such as ICU stay duration and hospitalization duration, appropriate weighted regression models were selected based on distribution characteristics for comparison. Perioperative complications and other categorical outcomes were also compared between groups and estimated for effects after weighting. To enhance the robustness of the results, sensitivity analyses were further performed, including repeated estimation of the primary outcome under different model settings and testing the impact of changes in key covariate selection on the direction and magnitude of effects, thereby verifying the consistency and reliability of the study conclusions.

Results

Baseline characteristics

This study enrolled 72 cardiac surgery patients: 34 in the TEE-guided closed-loop optimization group and 38 in the conventional management group. The two groups were generally comparable in most demographic characteristics and underlying comorbidities, with no statistically significant differences in gender distribution, body mass index (BMI), smoking status, or the proportions of hypertension, diabetes, chronic obstructive pulmonary disease (COPD), an...

Discussion

This retrospective two-group controlled study focused on the real-time assessment and intervention optimization of volume status and cardiac function via intraoperative transesophageal echocardiography (TEE)¹⁴. The results demonstrated an overall improvement in rapid pathway outcomes, with key manifestations including shorter mechanical ventilation duration and smoother perioperative transition¹⁵. These findings align with recent evidence supporting the rapid recovery conce...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Anesthesia Workstation	Dräger	Perseus A500	Used for general anesthesia delivery and ventilation
Hemodynamic Monitoring System	Edwards Lifesciences	EV1000	Continuous arterial pressure and hemodynamic monitoring
R software	R Foundation for Statistical Computing	Version 4.2.1	Open-source software utilized for propensity score estimation and weighted analysis
SPSS Statistics	IBM	Version 26.0	Utilized for demographic data processing and primary statistical analysis
TEE Adult Multiplane Probe	GE Healthcare / Philips Healthcare	Adult multiplane probe	Standard adult probe for perioperative transesophageal imaging
Transesophageal Echocardiography System	GE Healthcare / Philips Healthcare	EPIQ CVx / Vivid E95	Used for intraoperative cardiac imaging and hemodynamic assessment