Background

Detailed understanding of the association between intraoperative left atrial and left ventricular diastolic function and postoperative atrial fibrillation is lacking. In this post hoc analysis of the Posterior Left Pericardiotomy for the Prevention of Atrial Fibrillation after Cardiac Surgery (PALACS) trial, we aimed to evaluate the association of intraoperative left atrial and left ventricular diastolic function as assessed by transesophageal echocardiography (TEE) with postoperative atrial fibrillation.

Methods

PALACS patients with available intraoperative TEE data (n = 402 of 420; 95.7%) were included in this cohort study. We tested the hypotheses that preoperative left atrial size and function, left ventricular diastolic function, and their intraoperative changes were associated with postoperative atrial fibrillation. Normal left ventricular diastolic function was graded as 0 and with lateral e’ velocity 10 cm/s or greater. Diastolic dysfunction was defined as lateral e’ less than 10 cm/s using E/e’ cutoffs of grade 1, E/e’ 8 or less; grade, 2 E/e’ 9 to 12; and grade 3, E/e’ 13 or greater, along with two criteria based on mitral inflow and pulmonary wave flow velocities.

Results

A total of 230 of 402 patients (57.2%) had intraoperative diastolic dysfunction. Posterior pericardiotomy intervention was not significantly different between the two groups. A total of 99 of 402 patients (24.6%) developed postoperative atrial fibrillation. Patients who developed postoperative atrial fibrillation more frequently had abnormal left ventricular diastolic function compared to patients who did not develop postoperative atrial fibrillation (75.0% [n = 161 of 303] vs. 57.5% [n = 69 of 99]; P = 0.004). Of the left atrial size and function parameters, only delta left atrial area, defined as presternotomy minus post–chest closure measurement, was significantly different in the no postoperative atrial fibrillation versus postoperative atrial fibrillation groups on univariate analysis (–2.1 cm2 [interquartile range, –5.1 to 1.0] vs. 0.1 [interquartile range, –4.0 to 4.8]; P = 0.028). At multivariable analysis, baseline abnormal left ventricular diastolic function (odds ratio, 2.02; 95% CI, 1.15 to 3.63; P = 0.016) and pericardiotomy intervention (odds ratio, 0.46; 95% CI, 0.27 to 0.78, P = 0.004) were the only covariates independently associated with postoperative atrial fibrillation.

Conclusions

Baseline preoperative left ventricular diastolic dysfunction on TEE, not left atrial size or function, is independently associated with postoperative atrial fibrillation. Further studies are needed to test if interventions aimed at optimizing intraoperative left ventricular diastolic function during cardiac surgery may reduce the risk of postoperative atrial fibrillation.

Editor’s Perspective
What We Already Know about This Topic
  • New-onset postoperative atrial fibrillation occurs in approximately 20 to 30% of low-risk cardiac surgical patients. Patients with postoperative atrial fibrillation have longer in-hospital stays and increased postoperative morbidity and mortality.

  • Intraoperative transesophageal echocardiographic predictors of postoperative atrial fibrillation are poorly understood.

What This Article Tells Us That Is New
  • Patients who developed postoperative atrial fibrillation had more frequently abnormal baseline intraoperative transesophageal echocardiographic left ventricular diastolic function compared to patients who did not develop postoperative atrial fibrillation (75.0% vs. 57.5%; P = 0.004).

  • Baseline preoperative intraoperative left ventricular diastolic function is independently associated with greater postoperative atrial fibrillation after cardiac surgery (odds ratio, 2.02; 95% CI, 1.15 to 3.63; P = 0.016)

Postoperative atrial fibrillation, or new-onset atrial fibrillation occurring during hospitalization, is a common complication of cardiac surgery, occurring in approximately 20 to 30% of patients.1  Patients with postoperative atrial fibrillation have longer in-hospital stay and higher hospital costs, as well as higher morbidity from stroke and heart failure, increased risk of rehospitalization, and increased mortality.2,3  It is generally thought that postoperative atrial fibrillation is caused by the combination of acute triggers related to surgery acting on top of a predisposing atrial or ventricular substrate.4  However, the echocardiographic predictors of postoperative atrial fibrillation are poorly understood. While previous studies have examined the relationship between preoperative left atrial and left ventricular function assessed at preoperative transthoracic echocardiography and postoperative atrial fibrillation, a detailed understanding of the relationship between intraoperative left atrial and left ventricular function assessed by intraoperative transesophageal echocardiography (TEE), their intraoperative changes, and postoperative atrial fibrillation is lacking.5,6 

In this secondary analysis of a randomized trial evaluating the effect of posterior pericardiotomy on postoperative atrial fibrillation in cardiac surgical patients, we hypothesized that baseline left atrial size and function, left ventricular diastolic function, and their intraoperative changes would be associated with postoperative atrial fibrillation independently of posterior pericardiotomy.

Study Population

This is a post hoc cohort study of the Posterior Left Pericardiotomy for the Prevention of Atrial Fibrillation after Cardiac Surgery (PALACS) randomized controlled trial (NCT02875405), approved by the Weill Cornell Medicine Institutional Review Board (New York, New York; No. 1502015867). All study patients gave informed consent for participation, and data usage and the study protocol along with results of the main trial have been previously published.7,8  Data collected for the study will be made available by the corresponding author upon reasonable request after publication.

Patients were enrolled in the PALACS trial between September 18, 2017, and August 2, 2021, to evaluate whether performing posterior left pericardiotomy—an incision in the posterior pericardium that drains the pericardial sac into the left pleural cavity—was associated with a reduction in in-hospital postoperative atrial fibrillation in patients undergoing elective cardiac surgery without history of atrial fibrillation or any other arrhythmias (relative risk, 0.55; 95% CI, 0.39 to 0.78). In an explanatory analysis of the PALACS trial, we have previously shown that left atrial dimensions and function did not change significantly after posterior pericardiotomy and that reduction in posterolateral pericardial effusion was the most likely mechanism for the effect of posterior pericardiotomy in reducing postoperative atrial fibrillation.9  The randomization strategy used in the PALACS study employed stratification by congestive heart failure, hypertension, age, diabetes mellitus, previous stroke or transient ischemic attack or thromboembolism, vascular disease, age, sex category (CHA2DS2-VASc) score, which has been shown to predict risk of postoperative atrial fibrillation in cardiac surgical patients.7  Patients undergoing coronary artery bypass grafting (CABG), aortic valve repair or replacement, ascending aortic graft replacement, or a combination of these surgical interventions were included. Patients were randomized 1:1 to undergo either posterior left pericardiotomy or no intervention. All patients underwent general anesthesia with invasive monitoring including arterial line, central line, pulmonary artery catheter, and TEE. Fluid and vasopressor management were performed according to a predefined clinical protocol.

Image Acquisition and Hemodynamics

A comprehensive intraoperative TEE following a predefined protocol was performed on each patient. TEEs were acquired in accordance with European Society of Cardiology (Sophia Antipolis, France) and American Society of Echocardiography (Durham, North Carolina) guidelines by a dedicated group of board-certified cardiac anesthesiologists trained in the protocol using commercially available equipment (EPIQ 7, Philips Medical Systems, USA, and Vivid 7, GE Healthcare, USA, ultrasound systems).8 

Left atrial images were optimized before acquisition to ensure maximal left atrial areas in the midesophageal two and four-chamber views (Supplementary Figure 1, https://links.lww.com/ALN/D267). Left atrial length was measured as the distance from the mitral annulus to the posterior left atrial wall in both midesophageal four-chamber and two-chamber views in their maximal diameter. Left atrial area was obtained by tracing the endocardium in the left atrial–focused midesophageal four-chamber and two-chamber views, with the left atrial maximum area as the largest area during the cardiac cycle and left atrial minimum area as the smallest area during the cardiac cycle.6,10  The left atrial reservoir function was determined to be the percentage of change between the largest and smallest left atrial area divided by the largest left atrial area using the following formula6 :

LA maximum area LA minimum areaLA maximum area   x   100%

The left atrial (LA) area and length in the four-chamber and two-chamber views were used to quantify the left atrial volume using the biplane area/length method according to the equation

0.85 x   LA1 x LA2L1+L22

where LA1 and LA2 were the areas in the four-chamber and two-chamber views, respectively, and L1 and L2 signified left atrial length in the four-chamber and two-chamber views.11 

Parameters of diastolic function were measured according to established methods by echocardiographers blinded to the randomization assignment and outcomes12 : (1) spectral pulsed wave Doppler of mitral inflow at the tips of the mitral leaflets including the peak velocities during early filling (E), atrial contraction (A), the ratio of E/A, the deceleration time from peak E-wave were measured, and the duration of the mitral A-wave in the four-chamber view, using the most representative spectral Doppler trace greater than two to four cycles; (2) pulsed wave Doppler of pulmonary venous flow in the midesophageal two-chamber view with the sample volume placed within the proximal 1 cm of the left upper pulmonary vein, including the pulmonary systolic and diastolic flows, and atrial reversal duration (color Doppler was used to align the Doppler cursor parallel to the pulmonary venous flow); and (3) spectral tissue Doppler of early (e’) diastolic myocardial velocity at the lateral mitral annulus in the four-chamber view. If the lateral annulus was not well-visualized, the medial annulus was used for myocardial velocities (Supplementary Figure 2, https://links.lww.com/ALN/D267).

Diastolic parameters were integrated into a global diastolic function score by using grades 0 to 3.2,3,12,13  Normal left ventricular diastolic function was graded as 0 and defined as a lateral e’ velocity 10 cm/s or greater. Diastolic dysfunction was defined as lateral e’ less than 10 cm/s and graded following the algorithm reported in figure 1:

Fig. 1.

Algorithm for grading diastolic dysfunction. A, A-wave during the mitral inflow or pulmonary vein flow as appropriate; e’, early diastolic myocardial velocity; E, E-wave during the mitral inflow.

Fig. 1.

Algorithm for grading diastolic dysfunction. A, A-wave during the mitral inflow or pulmonary vein flow as appropriate; e’, early diastolic myocardial velocity; E, E-wave during the mitral inflow.

Close modal

Grade 1:

E/e’ 8 or less AND at least two of the following criteria:

  1. Mitral E/A less than 0.8

  2. Mitral deceleration time greater than 240 ms

  3. Pulmonary venous systolic wave velocity equal to or greater than diastolic wave

  4. Pulmonary venous A wave reversal duration shorter than mitral A wave duration

Grade 2:

E/e’ 9 to 12 AND at least two of the following criteria:

  1. Mitral E/A 0.8 to 1.5

  2. Mitral deceleration time 160 to 240 ms

  3. Pulmonary venous systolic wave velocity less than diastolic wave

  4. Pulmonary venous A wave reversal duration - mitral A wave duration greater than 30 ms

Grade 3:

E/e’ 13 or greater AND at least two of the following criteria:

  1. Mitral E/A greater than 1.5

  2. Mitral deceleration time less than 160 ms

  3. Pulmonary venous systolic wave velocity less than diastolic wave

  4. Pulmonary venous A reversal wave duration - mitral A wave duration 30 ms or greater

Cases of diastolic dysfunction (i.e., lateral e’ less than 10 cm/s) not meeting these criteria were graded as “indeterminate diastolic dysfunction” or adjudicated by a third reviewer.

All TEE and hemodynamic measurements were performed in standardized conditions and taken at two time points—(1) baseline intraoperative TEE (i.e., before sternotomy), and (2) postoperative TEE (i.e., after chest closure)—and were used to evaluate immediate changes in left atrial size and function and left ventricular diastolic function. Systemic blood pressure, central venous pressure, and pulmonary artery pressures were obtained from the electronic anesthesia record (CompuRecord, Philips Medical Systems, USA). Cardiac output and cardiac index were averaged from three similar values measured by thermodilution technique.

Postoperative Atrial Fibrillation Assessment

Postoperative atrial fibrillation was defined as the presence of an irregularly irregular rhythm without P waves lasting more than 30 s. Patient rhythm was continuously monitored via telemetry for the entire hospital stay, and all postoperative atrial fibrillation episodes were adjudicated independently by a committee consisting of two cardiologists and a cardiac surgeon blinded to patient clinical data.

Statistical Analysis

The Shapiro–Wilk test was used to assess whether continuous variables were normally distributed. Since all continuous variables were nonnormally distributed, they were reported as median and interquartile range and compared using the Mann–Whitney U test. Continuous variables were untransformed, and categorization of such variables was avoided to prevent information loss and spurious associations.14  Categorical variables were described as counts and proportions and compared using Pearson’s chi-square test. Variations between the pre- and postoperative time points (deltas) were analyzed by ANOVA. Baseline operative and echocardiographic variables were compared between patients with and without postoperative atrial fibrillation including age, sex, race, body mass index, diabetes, New York Heart Association class greater than 2, preoperative hematocrit, European System for Cardiac Operation Risk Evaluation (EuroSCORE) II, surgical procedure (CABG, aortic valve, vascular aortic), cardiopulmonary bypass and operative times, posterior pericardiotomy intervention, and intraoperative TEE measurements.

A multivariable logistic regression model was used to assess if postoperative atrial fibrillation was associated with specific characteristics and to adjust for confounders. To avoid overfitting, we built a parsimonious model that included CHA2DS2-VASc score greater than 3, left atrial volume, presence of posterior pericardiotomy, and any left ventricular diastolic dysfunction. Variables were selected based on clinical relevance and model performance (e.g., removing collinear variables). The inclusion of the CHA2DS2-VASc score, well-validated for predicting postoperative atrial fibrillation after cardiac surgery,15,16  combined the confounding effect of baseline characteristics in a single covariate. Left atrial volume and diastolic dysfunction were relevant to the research question, and the presence of posterior pericardiotomy was included to account for treatment allocation from the original trial. Results from the multivariable models are presented as odds ratios and corresponding 95% CIs. These analyses were established a priori without a predefined minimal clinically meaningful effect size.

Sensitivity Analyses.

Sensitivity analyses were established a priori and were performed to confirm the solidity of the main results and validate the baseline left atrial TEE measurements as well as the diastolic function grading algorithm. In detail,

  1. In patients (n = 118) with comprehensive, in-house preoperative transthoracic echocardiography examinations (within 5 [interquartile range, 2 to 17] days of surgery), Bland–Altman plots were used to test the agreement between left atrial length and area measurements at preoperative transthoracic echocardiography compared to preoperative TEE.17  The means of the paired differences between transthoracic echocardiography and TEE measurements were calculated as the bias and the 95% limits of agreement. Correlation coefficients and plots were also produced to show the relationship between transthoracic echocardiography and TEE measurements. Inter- and intraobserver reproducibility for baseline intraoperative left atrial length and area were of intraclass correlation coefficient,18  which measures the strength of agreement by comparing the variability in the ratings between and within raters. The closer the value to 1, the better the agreement. Reproducibility for left atrial measurements in 20 patients was measured with paired measurements by two raters.

  2. In the same cohort of patients, left ventricular diastolic function was evaluated using standard methods at preoperative transthoracic echocardiography12,19,20  (used as reference) and compared to diastolic function graded at baseline preoperative TEE.5  In addition, inter- and intrarater reproducibility for the evaluation of lateral e’ and mitral E-wave at baseline preoperative TEE were assessed by means of intraclass correlation coefficient18  in 20 patients by two raters.

As a post hoc sensitivity analysis, an interaction term between posterior pericardiotomy and diastolic dysfunction was added to the main multivariable model. Two-tailed P value less than 0.05 was considered statistically significant without adjustment for multiplicity. Absolute standardized differences were also provided to assess balance between groups. Values greater than 0.2 suggested the groups were imbalanced. The data analysis and statistical plan were written after the data were accessed but before any analysis was performed. All statistical analyses were performed using R Statistical Software (version 3.2.3, R Foundation for Statistical Computing, Austria).

Of the 420 patients enrolled in the PALACS trial, 402 (95.7%) had complete and analyzable intraoperative echocardiographic data. Baseline characteristics of the patients are summarized in table 1 and Supplementary Table 1 (https://links.lww.com/ALN/D267).

Table 1.

Demographic Characteristics of the Participants with and without Diastolic Dysfunction

Demographic Characteristics of the Participants with and without Diastolic Dysfunction
Demographic Characteristics of the Participants with and without Diastolic Dysfunction

A total of 230 of 402 patients (57.2%) had intraoperative diastolic dysfunction. Patients with diastolic dysfunction were older; more likely to be female; more likely to have hypertension, diabetes, or previous myocardial infarction; and more likely to use β-blockers. They also had a higher CHA2DS2-VASc score, EuroSCORE II, left atrial dimension, and need for aortic surgery. Posterior pericardiotomy intervention was not significantly different between the two groups (table 1).

A total of 99 of 402 patients (24.6%) developed postoperative atrial fibrillation. Patients with postoperative atrial fibrillation were older and had higher EuroSCORE II and CHA2DS2-VASc scores. Patients who developed postoperative atrial fibrillation spent a median time of 24.1 h in atrial fibrillation and had significantly longer in-hospital stay and intensive care stay but did not have significantly higher rate of postoperative complications including stroke, myocardial infarction, need for intra-aortic balloon pump, or re-intubation than patients without postoperative atrial fibrillation (Supplementary Table 2, https://links.lww.com/ALN/D267). The largest number of postoperative atrial fibrillation cases occurred on postoperative day 2 (41%), followed by postoperative day 3 (30%). The median postoperative atrial fibrillation time to onset was 50 h [interquartile range, 36 to 68].

Baseline intraoperative echocardiographic data are presented in table 2. Patients who developed postoperative atrial fibrillation had more frequently abnormal left ventricular diastolic function (75.0% in the postoperative atrial fibrillation group vs. 57.5% in the no postoperative atrial fibrillation group; P = 0.004; fig. 2; table 2); 16.2% (23 of 142) of patients with normal diastolic dysfunction, 31.7% (34 of 107) of patients with grade 1 dysfunction, 33.7% (26 of 77) with grade 2 dysfunction, 14.2% (2 of 14) with grade 3 dysfunction, and 21.8% (7 of 32) with undetermined grade dysfunction developed postoperative atrial fibrillation (Supplementary Figure 3 and Supplementary Table 3, https://links.lww.com/ALN/D267). When evaluating the severity of left ventricular diastolic dysfunction, the percentage of patients with grade 1 and 2 dysfunction was higher among patients who developed postoperative atrial fibrillation (grade 1, 40.0% [34/99] vs. 28.6% [73/303]; grade 2, 30.6% [26/99] vs. 20.0% [51/303]).

Table 2.

Baseline Preoperative Transesophageal Echocardiographic Data

Baseline Preoperative Transesophageal Echocardiographic Data
Baseline Preoperative Transesophageal Echocardiographic Data
Fig. 2.

Prevalence of left ventricular diastolic dysfunction among patients who develop and those who did not develop postoperative atrial fibrillation. Patients who developed postoperative atrial fibrillation more frequently had abnormal left ventricular diastolic function (75% in the postoperative atrial fibrillation group vs. 58% in the no postoperative atrial fibrillation group; P = 0.004).

Fig. 2.

Prevalence of left ventricular diastolic dysfunction among patients who develop and those who did not develop postoperative atrial fibrillation. Patients who developed postoperative atrial fibrillation more frequently had abnormal left ventricular diastolic function (75% in the postoperative atrial fibrillation group vs. 58% in the no postoperative atrial fibrillation group; P = 0.004).

Close modal

Patients who developed postoperative atrial fibrillation had significantly lower lateral and septal e’ values and longer pulmonary A-wave duration; however, the mitral A wave duration minus pulmonary atrial reversal wave duration time was not significantly different (table 2).

The hemodynamic variables at baseline and after surgical chest closure are summarized in table 3. There were no significant differences between patients with and without postoperative atrial fibrillation in terms of preoperative systolic and diastolic blood pressures, central venous pressure, pulmonary artery pressures, and cardiac output. After chest closure, there was a statistically significant but minimal clinical difference in central venous pressure. There was also a trend toward lower cardiac index in patients who developed postoperative atrial fibrillation.

Table 3.

Hemodynamic Characteristics

Hemodynamic Characteristics
Hemodynamic Characteristics

Univariate associations between postoperative changes in left atrial function and postoperative atrial fibrillation are shown in table 4. There was significant difference in the delta left atrial area (defined presternotomy minus post–chest closure measurement) between the two groups, with the group that developed postoperative atrial fibrillation having no change in left atrial area after bypass whereas the group that did not develop postoperative atrial fibrillation had a smaller left atrial area after bypass. The delta left atrial length demonstrated a similar trend, with patients who developed postoperative atrial fibrillation having a smaller decrease in left atrial length than those who did not develop postoperative atrial fibrillation. There was also a trend for a greater decrease in left atrial appendage velocities among patients who developed postoperative atrial fibrillation, suggesting that the relative postoperative change in atrial size and function may play a role in the etiology of postoperative atrial fibrillation.

Table 4.

Baseline versus Postoperative Transesophageal Echocardiographic Measurements

Baseline versus Postoperative Transesophageal Echocardiographic Measurements
Baseline versus Postoperative Transesophageal Echocardiographic Measurements

In multivariable analysis, baseline abnormal left ventricular diastolic function (odds ratio, 2.02; 95% CI, 1.15 to 3.63; P = 0.016) and pericardiotomy intervention (odds ratio, 0.46; 95% CI, 0.27 to 0.78; P = 0.004) were the only covariates independently associated with postoperative atrial fibrillation (table 5).

Table 5.

Multivariable Analysis for Predictors of Postoperative Atrial Fibrillation

Multivariable Analysis for Predictors of Postoperative Atrial Fibrillation
Multivariable Analysis for Predictors of Postoperative Atrial Fibrillation

Other Sensitivity Analyses

Left Atrial Indices.

The validity of TEE measurements was confirmed by the Bland–Altman plots, demonstrating moderate differences without bias between preoperative TEE and preoperative transthoracic echocardiography left atrial length and left atrial area measurements (Supplementary Table 4 and Supplementary Figure 4, https://links.lww.com/ALN/D267). Additional analyses showed strong correlations preoperative TEE and preoperative transthoracic echocardiography left atrial length and left atrial area (r = 0.77 and r = 0.83, respectively; Supplementary Figure 5, https://links.lww.com/ALN/D267). Intraclass correlation coefficient demonstrated high inter- and intraobserver reproducibility for left atrial length and area (Supplementary Tables 5 and 6, https://links.lww.com/ALN/D267).

Left Ventricular Diastolic Function.

Intraclass correlation coefficient demonstrated good consistency in the assessment of diastolic dysfunction by TEE and transthoracic echocardiography (intraclass correlation coefficient, 0.81; 95% CI, 0.70 to 0.86). Intraclass correlation coefficient also showed excellent inter- and intrarater reproducibility for both lateral e’ and mitral E assessments (Supplementary Tables 7 and 8, https://links.lww.com/ALN/D267). There was no interaction between pericardiotomy and diastolic dysfunction (P = 0.568).

In this secondary analysis of the PALACS trial of greater than 400 patients, we found that a baseline intraoperative left ventricular diastolic dysfunction was independently associated with postoperative atrial fibrillation after adjusting for patient characteristics, left atrial size, and pericardiotomy intervention. The original trial demonstrated that posterior pericardiotomy reduced the incidence of postoperative atrial fibrillation (relative risk, 0.55; 95% CI, 0.39 to 0.78).

When we assessed the impact of surgery on both left atrial function and left ventricular diastolic function, we found that intraoperative changes in left atrial indices were associated with postoperative atrial fibrillation. However, on multivariable analysis, only left ventricular diastolic dysfunction before surgery and pericardiotomy intervention were independently associated with postoperative atrial fibrillation. Interestingly, left atrial reservoir function and intraoperative changes therein were not associated with postoperative atrial fibrillation. This may suggest that, while immediate preoperative optimization of left ventricular diastolic function is important, intraoperative changes in left atrial size and function may not significantly affect the development of postoperative atrial fibrillation after cardiac surgery.

Postoperative atrial fibrillation, defined as new-onset atrial fibrillation occurring during the hospitalization after cardiac surgery, is thought to be a distinct entity from atrial fibrillation in nonsurgical settings, mediated both by pre-existing patient substrate and acute surgical changes including proarrhythmic substrate and fibrosis, and oxidative stress and inflammation after surgery.4  In addition, surgical manipulation, variations in cardiac function due to ischemia and inflammation, and fluctuations in fluid status leading to changes in filling pressure and in afterload in the perioperative period can significantly impact the risk of developing postoperative atrial fibrillation.

While the associations we found are consistent with previous studies, the strength of the associations and the odds ratios differ from previous reports. This may partly be related to differences in the sample size and comprehensiveness of the echocardiographic measures collected, variables selected for the multivariable model, patient characteristics, and baseline left atrial and left ventricular size and function.

Studies performed in clinical outpatient settings evaluated postoperative atrial fibrillation predictors before cardiac surgery and found that left ventricular diastolic dysfunction was a strong predictor of postoperative atrial fibrillation.5  Other intraoperative studies have evaluated predictors of postoperative atrial fibrillation such as left atrial size and function in limited cardiac surgical populations such as isolated coronary CABG patients.6,10,21  A small prospective study of 79 patients undergoing CABG found that prebypass left atrial appendage velocity and postbypass left upper pulmonary vein systolic/diastolic velocity ratio were significant predictors of postoperative atrial fibrillation,10  and another small study including 62 CABG patients found that after adjustment for age and cross-clamp time, there was no difference in transmitral velocities between patients who did and did not develop postoperative atrial fibrillation.21 

One larger study including 300 patients undergoing CABG evaluated intraoperative TEE parameters and found that postbypass parameters of atrial filling fraction and abnormal left ventricular relaxation (E duration greater than 270 ms) were independently associated with postoperative atrial fibrillation.6  Melduni et al. identified diastolic dysfunction on preoperative transthoracic echocardiography as an echocardiographic predictor of postoperative atrial fibrillation in a cardiac surgical population at high risk of postoperative atrial fibrillation that included patients with significant mitral regurgitation.5  The authors found a higher rate of postoperative atrial fibrillation in patients with diastolic dysfunction (38.5%) and found that the rate of postoperative atrial fibrillation increased exponentially with dysfunction severity (odds ratio, 5.12, 95% CI, 1.60 to 16.38, P = 0.006; odds ratio, 9.87, 95% CI, 3.23 to 30.13, P < 0.001; and odds ratio, 28.52, 95% CI, 7.11 to 114.44, P < 0.001, in grades 1, 2 and 3, respectively).5 

Methodologic differences in postoperative atrial fibrillation assessment may also explain the different results among studies. The negative study by Skubas et al.21  that did not find that preoperative TEE left atrial or mitral inflow indices were associated with postoperative atrial fibrillation used a less sensitive postoperative atrial fibrillation definition compared to the PALACS trial.21  In the mentioned study by Melduni et al.5  that identified diastolic dysfunction as the most significant predictor of postoperative atrial fibrillation, continuous telemetry during hospitalization or electrocardiogram within 30 days after surgery was used, similarly to our study.5 

While the current standard method for evaluation of left ventricular diastolic function is transthoracic echocardiography, recent recommendations state that diastolic function should be assessed during TEE when TEE is being performed for other purposes, which is true for all intraoperative TEE examinations.22  A recent study has also assessed the validity of left ventricular diastolic parameters measured by transthoracic echocardiography versus TEE under sedation, and found similar values and a statistically significant correlation.23  While other studies have chosen a more simplistic grading system, we chose an integrated diastolic grading strategy that was recommended and followed previously, and were able to successfully grade greater than 86% of the patients.5,13,24  Our study confirmed the validity of left ventricular diastolic function assessed at baseline TEE, lending credence to the importance of grading diastolic function intraoperatively.

While we found excellent correlation between left atrial indices on TEE and TEE, the moderate differences (0.5 cm for left atrial length and 3.8 cm2 for left atrial area) suggest that these values are not interchangeable. We speculate that this could be due to general anesthesia and positive pressure ventilation, resulting in changes in filling pressures, hemodynamics, and diastolic dysfunction. Previous studies have shown improvement of left ventricular diastolic function on intraoperative TEE compared to same-day awake preoperative transthoracic echocardiography in cardiac surgical patients.25  Another reason could be related to the inherent differences between transthoracic echocardiography versus TEE measurements. A prospective observational study of 121 patients undergoing simultaneous transthoracic echocardiography and TEE under sedation showed that TEE underestimated left atrial length by 9% using the short-axis view at the level of the aortic valve compared to the parasternal long-axis view.26  The mean differences in this study are similar with underestimation of 10% in left atrial length and 25% in left atrial area between preoperative transthoracic echocardiography and intraoperative TEE.

To the best of our knowledge, this is the largest study to date prospectively collecting and assessing intraoperative echocardiographic data in cardiac surgery patients using telemetry, i.e., the most sensitive method to detect even self-limiting episodes of postoperative atrial fibrillation. Our study builds on the existing literature by identifying intraoperative echocardiographic correlates and predictors of postoperative atrial fibrillation, also assessing the relative intraoperative changes of these echo parameters. These results fill in a gap on the associations and predictors of postoperative atrial fibrillation in an elective low-risk surgical population.

Limitations

There are several limitations to our study. This was a post hoc analysis of a randomized control trial where the intervention, posterior pericardiotomy, may have affected the association between diastolic dysfunction and postoperative atrial fibrillation. However, diastolic dysfunction was independently associated with postoperative atrial fibrillation, and the interaction between posterior pericardiotomy and diastolic dysfunction was not significant. We addressed known confounders in our multivariable analyses, but other unknown potential biases may remain due to the observational nature of the study. This was an exploratory study; formal sample size calculation was performed only for the parent trial. This was a single-center study that only included elective patients who did not have significant left atrial remodeling and had normal left ventricular ejection fraction, and excluded patients that had significant tricuspid or mitral valve disease, and therefore may not be generalizable to other populations, especially to those with significant left atrial remodeling, mitral disease, and a high risk of postoperative atrial fibrillation. We used a method of measuring intraoperative left atrial indices that has been previously published but is yet to be standardized in perioperative TEE guidelines. However, on sensitivity analysis, we found a strong correlation between left atrial measurements obtained at baseline intraoperative TEE compared to preoperative transthoracic echocardiography. We also found good concordance between preoperative transthoracic echocardiography and baseline intraoperative TEE in the evaluation of abnormal versus normal left ventricular diastolic function.

Conclusions

In conclusion, this is one of the largest and most comprehensive studies evaluating the association between left ventricular diastolic function and left atrial indices measured at intraoperative TEE and postoperative atrial fibrillation. Our results showed that baseline intraoperative left ventricular diastolic dysfunction graded on TEE, and not left atrial volume or function, was independently associated postoperative atrial fibrillation. This study also demonstrated the feasibility of an integrated diastolic grading system for intraoperative echocardiography. Further studies are needed to test if interventions aimed at optimizing baseline intraoperative left ventricular diastolic function during cardiac surgery may reduce the risk of postoperative atrial fibrillation.

Research Support

This work was supported by the National Institutes of Health (Bethesda, Maryland; K23 HL153836; Dr. Rong, Principal Investigator).

Competing Interests

Dr. Di Franco has consulted for Novo Nordisk (Bagsvaerd, Denmark) and Servier (Paris, France) and is an Advisory Board Member for Scharper (Milan, Italy). The other authors declare no competing interests.

Supplemental Tables 1 to 8, https://links.lww.com/ALN/D267

Supplemental Figures 1 to 5, https://links.lww.com/ALN/D267

1.
LaPar
DJ
,
Speir
AM
,
Crosby
IK
,
Fonner
E
,
Brown
M
,
Rich
JB
,
Quader
M
,
Kern
JA
,
Kron
IL
,
Ailawadi
G
;
Investigators for the Virginia Cardiac Surgery Quality Initiative
:
Postoperative atrial fibrillation significantly increases mortality, hospital readmission, and hospital costs.
Ann Thorac Surg
2014
;
98
:
527
33
;
discussion 533
2.
Gillinov
AM
,
Bagiella
E
,
Moskowitz
AJ
,
Raiten
JM
,
Groh
MA
,
Bowdish
ME
,
Ailawadi
G
,
Kirkwood
KA
,
Perrault
LP
,
Parides
MK
,
Smith
RL
,
Kern
JA
,
Dussault
G
,
Hackmann
AE
,
Jeffries
NO
,
Miller
MA
,
Taddei-Peters
WC
,
Rose
EA
,
Weisel
RD
,
Williams
DL
,
Mangusan
RF
,
Argenziano
M
,
Moquete
EG
,
O’Sullivan
KL
,
Pellerin
M
,
Shah
KJ
,
Gammie
JS
,
Mayer
ML
,
Voisine
P
,
Gelijns
AC
,
O’Gara
PT
,
Mack
MJ
for the
CTSN
:
Rate control versus rhythm control for atrial fibrillation after cardiac surgery.
N Engl J Med
2016
;
374
:
1911
21
3.
Eikelboom
R
,
Sanjanwala
R
,
Le
M-L
,
Yamashita
MH
,
Arora
RC
:
Postoperative atrial fibrillation after cardiac surgery: A systematic review and meta-analysis.
Ann Thorac Surg
2021
;
111
:
544
54
4.
Gaudino
M
,
Di Franco
A
,
Rong
LQ
,
Cao
D
,
Pivato
CA
,
Soletti
GJ
,
Chadow
D
,
Cancelli
G
,
Perezgrovas Olaria
R
,
Gillinov
M
,
DiMaio
JM
,
Girardi
LN
:
Pericardial effusion provoking atrial fibrillation after cardiac surgery: JACC review topic of the week.
J Am Coll Cardiol
2022
;
79
:
2529
39
5.
Melduni
RM
,
Suri
RM
,
Seward
JB
,
Bailey
KR
,
Ammash
NM
,
Oh
JK
,
Schaff
HV
,
Gersh
BJ
:
Diastolic dysfunction in patients undergoing cardiac surgery: A pathophysiological mechanism underlying the initiation of new-onset post-operative atrial fibrillation.
J Am Coll Cardiol
2011
;
58
:
953
61
6.
Leung
JM
,
Bellows
WH
,
Schiller
NB
:
Impairment of left atrial function predicts post-operative atrial fibrillation after coronary artery bypass graft surgery.
Eur Heart J
2004
;
25
:
1836
44
7.
Gaudino
M
,
Sanna
T
,
Ballman
KV
,
Robinson
NB
,
Hameed
I
,
Audisio
K
,
Rahouma
M
,
Di Franco
A
,
Soletti
GJ
,
Lau
C
,
Rong
LQ
,
Massetti
M
,
Gillinov
M
,
Ad
N
,
Voisine
P
,
DiMaio
JM
,
Chikwe
J
,
Fremes
SE
,
Crea
F
,
Puskas
JD
,
Girardi
L
;
PALACS Investigators
:
Posterior left pericardiotomy for the prevention of atrial fibrillation after cardiac surgery: An adaptive, single-centre, single-blind, randomised, controlled trial.
Lancet Lond Engl
2021
;
398
:
2075
83
8.
Abouarab
AA
,
Leonard
JR
,
Ohmes
LB
,
Lau
C
,
Rong
LQ
,
Ivascu
NS
,
Pryor
KO
,
Munjal
M
,
Crea
F
,
Massetti
M
,
Sanna
T
,
Girardi
LN
,
Gaudino
M
:
Posterior Left pericardiotomy for the prevention of postoperative Atrial fibrillation after Cardiac Surgery (PALACS): Study protocol for a randomized controlled trial.
Trials
2017
;
18
:
593
9.
Rong
LQ
,
Di Franco
A
,
Rahouma
M
,
Dimagli
A
,
Chan
J
,
Lopes
AJ
,
Kim
J
,
Sanna
T
,
Devereux
RB
,
Delgado
V
,
Weinsaft
JW
,
Crea
F
,
Alexander
JH
,
Gillinov
M
,
DiMaio
JM
,
Pryor
KO
,
Girardi
L
,
Gaudino
M
:
Postoperative pericardial effusion, pericardiotomy, and atrial fibrillation: An explanatory analysis of the PALACS trial.
Am Heart J
2023
;
260
:
113
23
10.
Shore-Lesserson
L
,
Moskowitz
D
,
Hametz
C
,
Andrews
D
,
Yamada
T
,
Vela-Cantos
F
,
Hossain
S
,
Bodian
C
,
Lessen
RJ
,
Konstadt
SN
:
Use of intraoperative transesophageal echocardiography to predict atrial fibrillation after coronary artery bypass grafting.
Anesthesiology
2001
;
95
:
652
8
11.
Jiamsripong
P
,
Honda
T
,
Reuss
CS
,
Hurst
RT
,
Chaliki
HP
,
Grill
DE
,
Schneck
SL
,
Tyler
R
,
Khandheria
BK
,
Lester
SJ
:
Three methods for evaluation of left atrial volume.
Eur J Echocardiogr
2008
;
9
:
351
5
12.
Nagueh
SF
,
Appleton
CP
,
Gillebert
TC
,
Marino
PN
,
Oh
JK
,
Smiseth
OA
,
Waggoner
AD
,
Flachskampf
FA
,
Pellikka
PA
,
Evangelista
A
:
Recommendations for the evaluation of left ventricular diastolic function by echocardiography.
J Am Soc Echocardiogr
2009
;
22
:
107
33
13.
Matyal
R
,
Skubas
NJ
,
Shernan
SK
,
Mahmood
F
:
Perioperative assessment of diastolic dysfunction.
Anesth Analg
2011
;
113
:
449
72
14.
Altman
DG
,
Royston
P
:
The cost of dichotomising continuous variables.
BMJ
2006
;
332
:
1080
15.
Chua
S-K
,
Shyu
K-G
,
Lu
M-J
,
Lien
L-M
,
Lin
C-H
,
Chao
H-H
,
Lo
H-M
:
Clinical utility of CHADS2 and CHA2DS2-VASc scoring systems for predicting postoperative atrial fibrillation after cardiac surgery.
J Thorac Cardiovasc Surg
2013
;
146
:
919
926.e1
16.
Chen
Y-L
,
Zeng
M
,
Liu
Y
,
Xu
Y
,
Bai
Y
,
Cao
L
,
Ling
Z
,
Fan
J
,
Yin
Y
:
CHA2DS2-VASc score for identifying patients at high risk of postoperative atrial fibrillation after cardiac surgery: A meta-analysis.
Ann Thorac Surg
2020
;
109
:
1210
6
17.
Altman
D
,
Bland
J
:
Measurements in medicine: The analysis of method comparison studies.
Statistician
1983
;
32
:
307
17
18.
Shrout
P
,
Fleiss
J
:
Intraclass correlation: Uses in assessing rater reliability.
Psychol Bull
1979
;
86
:
420
28
19.
Nishimura
RA
,
Tajik
AJ
:
Evaluation of diastolic filling of left ventricle in health and disease: Doppler echocardiography is the clinician’s Rosetta Stone.
J Am Coll Cardiol
1997
;
30
:
8
18
20.
Ommen
SR
,
Nishimura
RA
,
Appleton
CP
,
Miller
FA
,
Oh
JK
,
Redfield
MM
,
Tajik
AJ
:
Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: A comparative simultaneous Doppler-catheterization study.
Circulation
2000
;
102
:
1788
94
21.
Skubas
NJ
,
Barzilai
B
,
Hogue
CW
:
Atrial fibrillation after coronary artery bypass graft surgery is unrelated to cardiac abnormalities detected by transesophageal echocardiography.
Anesth Analg
2001
;
93
:
14
9
22.
Hahn
RT
,
Abraham
T
,
Adams
MS
,
Bruce
CJ
,
Glas
KE
,
Lang
RM
,
Reeves
ST
,
Shanewise
JS
,
Siu
SC
,
Stewart
W
,
Picard
MH
:
Guidelines for performing a comprehensive transesophageal echocardiographic examination: Recommendations from the American Society of Echocardiography and the Society of Cardiovascular Anesthesiologists.
J Am Soc Echocardiogr
2013
;
26
:
921
64
23.
Win
TT
,
Alomari
IB
,
Awad
K
,
Ratliff
MD
,
Qualls
CR
,
Roldan
CA
:
Transesophageal versus transthoracic echocardiography for assessment of left ventricular diastolic function.
J Integr Cardiol Open Access
2020
;
3
:
10.31487/j.jicoa.2020.01.05
.
24.
Swaminathan
M
,
Nicoara
A
,
Phillips-Bute
BG
,
Aeschlimann
N
,
Milano
CA
,
Mackensen
GB
,
Podgoreanu
MV
,
Velazquez
EJ
,
Stafford-Smith
M
,
Mathew
JP
;
Cardiothoracic Anesthesia Research Endeavors (CARE) Group
:
Utility of a simple algorithm to grade diastolic dysfunction and predict outcome after coronary artery bypass graft surgery.
Ann Thorac Surg
2011
;
91
:
1844
50
25.
Couture
P
,
Denault
AY
,
Shi
Y
,
Deschamps
A
,
Cossette
M
,
Pellerin
M
,
Tardif
J-C
:
Effects of anesthetic induction in patients with diastolic dysfunction.
Can J Anaesth
2009
;
56
:
357
65
26.
Block
M
,
Hourigan
L
,
Bellows
WH
,
Reeves
J
,
Romson
JL
,
Tran
M
,
Pastor
D
,
Schiller
NB
,
Leung
JM
:
Comparison of left atrial dimensions by transesophageal and transthoracic echocardiography.
J Am Soc Echocardiogr
2002
;
15
:
143
9