Secondary Logo

Journal Logo

Education: Original Clinical Research Report

Low Cardiac Output Syndrome After Adult Cardiac Surgery: Predictive Value of Peak Systolic Global Longitudinal Strain

Amabili, Philippe MD*; Benbouchta, Sabeha MD*; Roediger, Laurence MD, PhD*; Senard, Marc MD, PhD*; Hubert, Marie Bernard MD, PhD*; Donneau, Anne-Françoise MSci, PhD; Brichant, Jean-François MD, PhD*; Hans, Gregory A. MD, PhD*

Author Information
doi: 10.1213/ANE.0000000000002605



  • Question: Is the peak systolic global longitudinal strain (GLS) value measured during intraoperative transesophageal echo helpful to identify patients who develop postoperative low cardiac output syndrome (LCOS) after adult cardiac surgery?
  • Findings: Intraoperative GLS mildly improves the prediction of LCOS when added to previously described risk factors, including preoperative left ventricular ejection fraction.
  • Meaning: The prediction improvement of LCOS provided by intraoperative GLS is not clinically meaningful.

Low cardiac output syndrome (LCOS) is a recognized complication of cardiac surgery. It is associated with increased morbidity and mortality.1 Identification of patients at risk for postoperative LCOS is important to facilitate treatment, ie, timely introduction of circulatory support. It could also help avoid the deleterious effects of unnecessary administration of inotropes.2,3 Several risk factors for postoperative LCOS have been identified, including advanced age, prolonged bypass time, urgent surgery, and impaired left ventricular function.4–7

Left ventricular ejection fraction (LVEF) is the most widely used estimate of left ventricular systolic function. A decreased LVEF is an independent risk factor for both LCOS and mortality after cardiac surgery.8 The echocardiographic measurement of LVEF is, however, heavily dependent on image quality and operator experience.9

The peak systolic global longitudinal strain (GLS) is an alternative measure of LV systolic function that is based on the analysis of systolic myocardial deformation.10 Its measurement is more reproducible than 2D LVEF and remains accurate in patients with poor acoustic windows.10 GLS can also detect more subtle alterations in the contractility of the left ventricle.11 Recently, the GLS measured during preoperative transthoracic echocardiogram (TTE) was shown to predict mortality and the need for a postoperative inotropic support after cardiac surgery.12

GLS values obtained from TTE are only modestly correlated with those derived from transesophageal echocardiogram (TEE), which tends to overestimate GLS.13 Furthermore, general anesthesia and positive pressure ventilation affect loading conditions and therefore can impact GLS measurements.14 Our primary objective was to assess whether the GLS value measured using TEE images obtained before cardiopulmonary bypass (CPB) improves the prediction of postoperative LCOS over established risk factors. Our secondary objectives were to identify the cutoff value of GLS to best discriminate between patients who did and did not subsequently develop LCOS. The relationship between GLS and postoperative outcome was also investigated.


Study Design and Patients

Our ethics committee (Comite d’ethique hospitalo-facultaire universitaire de Liege, Chairperson Pr. V. Seutin, Ref:2016/85) approved this retrospective study and gave a waiver of consent. The manuscript adheres to the applicable Enhance the QUality and Transparancy Of health Research (EQUATOR) guidelines.

Eligible patients were adults undergoing elective or urgent (ie, patient not admitted for elective surgery but needing surgery before hospital discharge) on-pump cardiac surgery between January 2015 and June 2016 and in whom the images of the pre-CPB transesophageal echocardiography stored in our echo database met the following criteria: were acquired with a real-time 3D matrix array transducer (X7-2t, Philips Medical, Brussels, Belgium), had good quality electrocardiogram signal, and included the 3 midesophageal views of the left ventricle (4-chamber, 2-chamber, and long axis) at a minimum frame rate of 50 Hz. Exclusion criteria were emergency surgery (ie, to be performed before the next working day), preoperative use of inotropes or mechanical circulatory assistance, preoperative atrial fibrillation or ventricular pacing, and heart transplantation or ventricular assist device implantation.

Anesthesia and Hemodynamic Management

Target-controlled infusions of propofol and remifentanil were used to induce and maintain anesthesia in all patients. Doses were titrated to keep the entropy of the electroencephalogram between 40 and 60 (M-Entropy; Acertys-GE, Madison, WI). Full muscle relaxation was achieved with 1 mg·kg−1 of rocuronium before tracheal intubation. No additional bolus of muscle relaxant was given unless necessary (eg, patient coughing or moving despite appropriate anesthetic depth). Lungs were ventilated at a tidal volume of 8 mL·kg1 of ideal body weight with a positive end-expiratory pressure of 5 cmH2O. A continuous positive airway pressure of 5 cmH2O was applied during CPB. After skin closure, patients were transferred to the intensive care unit (ICU) sedated and ventilated. Propofol and remifentanil infusions were gradually discontinued once the criteria for extubation were met: normothermia, no excessive bleeding, hemodynamic stability, and acceptable gas exchanges.

All patients received a pulmonary artery catheter with continuous reading of the cardiac output and mixed venous oxygen saturation. Epicardial pacing was used for hemodynamic optimization whenever the heart rate was <80 beats per minute. Cardiac filling pressures, TEE, and visual inspection of the heart were used to guide fluid administration and assess whether inotropic support was needed during weaning from CPB. After chest closure, fluid responsiveness was assessed using pulse pressure variation, changes in the diameter of the inferior vena cava induced by mechanical ventilation, or passive leg raising maneuver. In the case of persistent cardiac index <2 L·minute1·m2 or mixed venous oxygen saturation <60% after correction of hypovolemia, the introduction or escalation of inotropes was considered. Before starting or escalating inotropes, an echocardiogram was usually performed to confirm systolic dysfunction of either ventricle and rule out tamponade. Dobutamine was the first-line inotrope and usually started at 3–5 µg·kg1·minute1. An intraaortic balloon pump was considered when the administration of dobutamine at a dose of 10 µg·kg1·minute1 failed to improve cardiac output. Veno-arterial extracorporeal membrane oxygenation was considered in case of cardiac failure refractory to the above-mentioned treatments. Epinephrine was only used as a rescue medication. Milrinone is rarely used in our department, and levosimendan was not available during the study period.

In the ICU, the dobutamine infusion was decreased by 0.5 µg·kg1·minute1 every 3 hours if the cardiac index remained >2 L·minute1·m2, the mixed venous oxygen saturation >60%, and the urinary output >0.5 mL·kg1·h1.

Outcome Measures and Variables

The primary outcome measure, LCOS, was arbitrarily defined as the need for an inotropic or a mechanical circulatory support for at least 24 hours postoperatively, or death from refractory heart failure within the first 24 postoperative hours. Patients who only required pure vasoconstrictors such as norepinephrine were not considered to have postoperative LCOS. Secondary outcome measures were time to complete weaning from inotropes, time to discharge from ICU and hospital, and 30-day mortality.

Baseline patient characteristics, operative data, and outcome measures were retrieved from our institutional electronic medical record. The preoperative LVEF was obtained from the preoperative TTE and categorized into normal (>50%), moderately decreased (31%–50%), or severely decreased (≤30%). These echocardiograms were usually performed during the month preceding surgery. As per local guidelines, the echocardiographers are requested to use the Simpson’s method to calculate LVEF whenever feasible.

Intraoperative TEEs were performed with an iE33 or an EPIQ 7 ultrasound system (Philips Medical, Brussels, Belgium). Pre-CPB images were acquired after induction of general anesthesia and before chest opening in mechanically ventilated and hemodynamically stable patients. A comprehensive examination was performed in all patients, and views were acquired according to the recommendations from the American Society of Echocardiography and the European Association of Cardiovascular Imaging.15 Studies meeting the above-mentioned criteria were transferred to a dedicated workstation (QLAB version 10.5 software, Philips Medical, Belgium) for off-line calculations by experienced echocardiographers (P.A., L.R., B.H., and G.H.). Two-dimensional speckle tracking with automated Cardiac Motion Quantification was used to calculate GLS. The peak of the R wave was used as a reference to identify end-diastole, and the time to aortic closure was measured using the midesophageal long-axis view of the aortic valve. The integrity of the tracking was visually checked, and GLS was computed only if at least 14 of the 17 segments were adequately tracked. Manual adjustments of the region of interest were made whenever necessary to optimize myocardial tracking. As recommended, changes in GLS are described using absolute values of the number. As a result, a decreased GLS means that GLS is becoming less negative and the LV function deteriorates.16

Statistical Analysis

Histograms, quartile plots, and the Shapiro-Wilk test were used to assess whether the distribution of continuous variables was Gaussian. Continuous variables were expressed as mean ± standard deviation or median (25th–75th percentile) according to their distribution. Categorical variables were summarized as count (%). The univariable association between established risk factors and LCOS was assessed using binary logistic regression. A first multivariable predictive model (model 1) of postoperative LCOS was created by entering all covariates significantly associated with LCOS at the univariable level (P < .05), except GLS, into a forward stepwise logistic regression. P value thresholds of .05 and .1 were used for entering and leaving the model, respectively. A second predictive model (model 2) including the GLS in addition to the covariates selected in model 1 was then created. The 2 models were compared using the likelihood-ratio test. The areas under the receiver operating characteristic (ROC) curve of the 2 models were also computed and compared using the algorithm suggested by DeLong et al.17 In addition, the clinical relevance of the change in predicted risk provided by the addition of GLS for an individual patient was estimated by computing the integrated discrimination index (IDI).

The optimal cutoff value of GLS associated with LCOS was determined using the Youden index (J= sensitivity + specificity − 1) of the ROC curve. As a sensitivity analysis, the GLS cutoff value that maximizes both sensitivity and specificity was also estimated.18

Adjusted and unadjusted relationships between GLS and secondary outcome measures were analyzed using Cox proportional-hazards regressions or logistic regressions as appropriate. Covariates retained in model 1 were used for adjustment in Cox regressions, and deaths competing with the outcome were censored at the maximum time to event for any patients. Given the small number of deaths, the analysis of 30-day mortality was only adjusted for the EuroSCORE II. The intraclass correlation coefficients (ICCs) ± 95% confidence interval (CI) were calculated on a subset of 30 randomly selected studies to assess intra- and interobserver variability of GLS measurements.

Considering a 20% baseline risk of postoperative LCOS, we estimated that including 164 patients would provide 80% power to detect an absolute increase of 20% in the risk of LCOS among patients with an impaired GLS at an α level of .05. As we had no information regarding the proportion of exams from which we would be able to calculate the GLS and the number of variables that would be retained in the multivariable analyses, all images acquired since January 2015 when we started using the 3D matrix array transducer (X7-2t) were reviewed. A P value ≤.05 was considered statistically significant. Statistical analyses were performed using Stata 13.1 (StataCorp LP, College Station, TX) and R version 3.2.4 (R Foundation for Statistical Computing).


Study Population

GLS was measured in 303 (33.4%) of the 907 patients who had cardiac surgery during the study period. Sixteen patients met exclusion criteria and 12 could not be included because of missing data on outcome or covariates. A total of 275 patients were thus retained for final analyses (Figure 1). Demographics, comorbidities, and intraoperative characteristics of these patients are summarized in Table 1 and in Supplemental Digital Content, Table,

Table 1.
Table 1.:
Demographics, Comorbidities, and Intraoperative Characteristics of the Study Population
Figure 1.
Figure 1.:
Flow chart of the study. ECG indicates electrocardiogram; GLS, peak systolic global longitudinal strain; LCOS, postoperative low cardiac output syndrome; TEE, transesophageal echocardiography.

Primary Objective

Among the 275 patients included, 59 (21.5%) patients had postoperative LCOS. The mean GLS value was significantly worse in the group of patients who developed postoperative LCOS (−15.2% ± 4.1% vs −18.5% ± 4.3%; P < .001;Figure 2).

Figure 2.
Figure 2.:
Relationships between global longitudinal strain (GLS) and low cardiac output syndrome (LCOS). Top: Scatter plot of peak systolic GLS measured during pre-CPB transesophageal echocardiogram (pre-CPB GLS) for patients with and without postoperative LCOS. Estimated difference in mean −3.3% (P < .001; 95% CI, −4.5% to −2.1%). Open circles, Individual measures. Black bar, median. Box, 25th percentile - 75th percentile. Adjacent lines, upper and lower adjacent values. Bottom: Receiver operating characteristic curve for the univariable association between peak systolic global longitudinal strain and postoperative low cardiac output syndrome. A GLS threshold value of −17% was identified as the optimal cutoff point using the Youden index. AUC indicates area under the curve; CI, confidence interval.

Univariable associations between established risk factors and postoperative LCOS are shown in Table 2. Patients suffering LCOS were more likely to be in New York Heart Association functional class III or IV, have a decreased LVEF (<50%), suffer chronic obstructive pulmonary disease, and have a history of peripheral artery disease. Mitral valve surgery, CPB time, combined surgical procedures, and urgent surgery were also significantly associated with postoperative LCOS.

Table 2.
Table 2.:
Risk Factors for Postoperative LCOS and Logistic Regression Models

Covariates, except for GLS, which were significantly associated with LCOS, were entered into a forward stepwise logistic regression to create a first predictive model (model 1). Predictors retained in this first model were the preoperative LVEF, a New York Heart Association functional class III or IV, the CPB duration, and mitral valve surgery (Table 2). A second model (model 2) was created by adding GLS to the variables retained in model 1. In this second model, GLS was found to be an independent predictor of postoperative LCOS (odds ratio, 1.1; 95% CI, 1.02–1.26; P = .02). The 2 models were compared using the likelihood-ratio test, and the model 2 fitted the data significantly better than model 1 (P = .02). However, as shown in Figure 3, the area under the ROC curve of the 2 models did not differ (difference in area under the ROC curve = 0.014; 95% CI, −0.0003 to 0.04; P = .15). Regarding the improvement in prediction of LCOS for an individual patient that resulted from adding GLS to the other predictors, the IDI was 0.02 (95% CI, 0.0003–0.04; P = .046). In other words, the difference in predicted risk between patients who experienced postoperative LCOS and those who did not increased by 2% when adding the GLS into the model.

Figure 3.
Figure 3.:
Receiver operating characteristic (ROC) curves of models 1 and 2. Model 2 includes the global longitudinal strain in addition to the predictors of model 1. The areas under the ROC curve (AUC) of the 2 models do not differ (P = .15). CI indicates confidence interval.

Secondary Objectives

Using the Youden index, the threshold GLS value of −17% (95% CI, −18.6% to −15.5%) was found to best distinguish patients who developed LCOS from those who did not suffer this complication. The sensitivity and specificity at this cutpoint were 0.69 (95% CI, 0.52–0.80) and 0.68 (95% CI, 0.47–0.79), respectively (Figure 2). A GLS threshold value of −17% was also obtained when attempting to maximize both sensitivity and specificity for the identification of LCOS in a sensitivity analysis.

Relationships between GLS and secondary postoperative outcome measures are summarized in Table 3. After adjusting for covariates included in model 1, a lower GLS value was significantly associated with a lower cumulative probability of weaning from inotropes postoperatively. No association was found between GLS and the other secondary outcome measures.

Table 3.
Table 3.:
Association of GLS With Secondary Postoperative Outcomes

Reproducibility Analyses

Reproducibility analyses were performed on a randomly selected subset of 30 patients. High ICCs were found for interobserver (ICC = 0.91; 95% CI, 0.81–0.95) and intraobserver (ICC = 0.96; 95% CI, 0.91–0.98) GLS measurements.


This study demonstrates that GLS measured from pre-CPB TEE images is an independent risk factor for postoperative LCOS. However, adding GLS to other risk factors only mildly improves the prediction of LCOS. Indeed, although the likelihood-ratio test indicates that the model including GLS fit the data better, the IDI suggests that the prediction improvement for an individual patient is of questionable clinical relevance.19 The fact that adding GLS into the model did not significantly improve the area under the ROC curve is a somewhat expected finding because a large independent association of the new predictor with the outcome is required to meaningfully improve the AUC.20 A GLS threshold value of −17% was found to best differentiate patients who developed postoperative LCOS from those who did not. Our study also shows that a lower GLS value is associated with a decreased cumulative probability of inotrope weaning postoperatively. Finally, after adjusting for confounders, we found no significant association between GLS and time to discharge from ICU and from hospital, and 30-day mortality.

GLS is used to assess the systolic LV myocardial deformation.21 It has several advantages over LVEF calculated according to the Simpson’s method, including a lower inter- and intraobserver variability.22 GLS better correlates with the LVEF obtained by magnetic resonance imaging10 and carries an independent prognostic value in a broad range of cardiac diseases.23 The results of the present study confirm the high intra- and interobserver reproducibility of TEE-derived 2D speckle tracking GLS previously reported.13

With respect to the prediction of postoperative LCOS, our results extend the findings from Ternacle et al12 to the GLS value measured using TEE in anesthetized and mechanically ventilated patients. Further comparison between the 2 studies is rendered impossible since Ternacle et al12 used different covariates for adjustment. In contrast to our findings, Ternacle et al12 also found GLS to be associated with 30-day mortality. One possible explanation for this discrepancy is that our study has a lower statistical power than the study of Ternacle et al.12 Indeed, considering a 6% mortality rate in patients with a normal GLS, a sample size of 712 patients would have been required to provide a 0.8 power to detect an absolute 6% difference in mortality rate at an α level of .05. Also, TTE- and TEE-derived strain are clearly not interchangeable,13 and the effects of anesthetics and loading conditions may further have affected the relationship between GLS and 30-day mortality.14

One strength of the present study is that caregivers, including cardiac anesthetists and intensivists, were “blinded” to the GLS because it was computed a posteriori using stored images. However, this study has limitations. Although the criteria used for the introduction and weaning of inotropes were well standardized and included objective measures derived from a pulmonary artery catheter, some degree of variability among caregivers may have influenced the outcome in this observational retrospective trial. Also, the retrospective design did not control for variables such as diastolic dysfunction of the left ventricle or right ventricular systolic function. Another limitation is that that we used the LVEF obtained from the preoperative TTE and not from the intraoperative TEE. However, it is unlikely that this affected the results. Indeed, TTE- and TEE-derived LVEF values are well correlated when calculated using the Simpson’s method, which was the case for most of the patients.24 In addition, this approach was consistent with the goal of our study, which was to assess whether measuring GLS during pre-CPB TEE would add clinically relevant information to the already available preoperative data that include a measure of LVEF by TTE. We used the QLab software to calculate GLS (QLab Software, Philips Medical, Belgium). Small but significant differences between vendors exist.25 Hence, we cannot rule out that using a different system would have led to slightly different cutoffs and results. To increase power and applicability, we included patients undergoing a broad range of cardiac surgical procedures. The resulting heterogeneity may have reduced the predictive ability of GLS. Full data were only obtained in 275 of 907 patients operated on during the study period, which limits the generalizability of these findings. More than half of the patients we excluded did not have intraoperative TEE. This finding is in agreement with current guidelines that do not support the routine use of TEE in low-risk coronary artery bypass graft surgeries.26 Our results could therefore not apply to this particular group of patients. In addition, the small sample size confers a limited statistical power and precludes interesting subgroup analyses. Finally, only 2 of the 4 TEE transducers available in the department allow measurement of the GLS. The 2 types of TEE transducers were used at random according to their availability, and this is unlikely to have resulted in a selection bias. Inadequate image quality and ECG artifacts are limitations inherent to the practice of intraoperative TEE. Specifically, acoustic shadowing caused by mitral valve calcifications, a poor definition of the lateral wall, or a low image resolution in the far field can hinder appropriate tracking of the myocardium and affect GLS values. We carefully adjusted the region of interest whenever necessary to optimize the tracking of the myocardium. The high intra- and interobserver reproducibility that we observed is somewhat reassuring. We believe that GLS could be measured in a higher proportion of patients in a prospective data collection because more attention would be paid to the quality of the images and the electrocardiogram signal. These limitations, however, underlie the fact that accurate measurement of GLS requires training, experience, and good image quality.

In conclusion, GLS obtained during pre-CPB TEE is an independent predictor of postoperative LCOS. However, adding GLS to the already established predictors of LCOS only mildly improves the prediction of this complication. The optimal threshold value of GLS associated with postoperative LCOS was −17%. This cutoff value gives 69% sensitivity and 68% specificity. Whether GLS has a greater prognostic value in specific surgical populations, ie, mitral valve surgery, warrants further investigations.


Name: Philippe Amabili, MD.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript.

Conflicts of Interest: P. Amabili received a travel grant from Medtronic.

Name: Sabeha Benbouchta, MD.

Contribution: This author helped conduct the study and analyze the data.

Conflicts of Interest: None.

Name: Laurence Roediger, MD, PhD.

Contribution: This author helped design the study, conduct the study, and analyze the data.

Conflicts of Interest: None.

Name: Marc Senard, MD, PhD.

Contribution: This author helped design the study, conduct the study, and write the manuscript.

Conflicts of Interest: None.

Name: Marie Bernard Hubert, MD, PhD.

Contribution: This author helped design the study and conduct the study.

Conflicts of Interest: None.

Name: Anne-Françoise Donneau, MSci, PhD.

Contribution: This author helped design the study, analyze the data, and write the manuscript.

Conflicts of Interest: None.

Name: Jean-François Brichant, MD, PhD.

Contribution: This author helped design the study and write the manuscript.

Conflicts of Interest: None.

Name: Gregory A. Hans, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Conflicts of Interest: G. A. Hans received a travel grant from Medtronic and honorarium from LivaNova.

This manuscript was handled by: Nikolaos J. Skubas, MD, DSc, FACC, FASE.


1. Rao V, Ivanov J, Weisel RD, Ikonomidis JS, Christakis GT, David TE. Predictors of low cardiac output syndrome after coronary artery bypass. J Thorac Cardiovasc Surg. 1996;112:38–51.
2. Nielsen DV, Algotsson L. Outcome of inotropic therapy: is less always more? Curr Opin Anaesthesiol. 2015;28:159–164.
3. Fellahi JL, Parienti JJ, Hanouz JL, Plaud B, Riou B, Ouattara A. Perioperative use of dobutamine in cardiac surgery and adverse cardiac outcome: propensity-adjusted analyses. Anesthesiology. 2008;108:979–987.
4. Ding W, Ji Q, Shi Y, Ma R. Predictors of low cardiac output syndrome after isolated coronary artery bypass grafting. Int Heart J. 2015;56:144–149.
5. Algarni KD, Maganti M, Yau TM. Predictors of low cardiac output syndrome after isolated coronary artery bypass surgery: trends over 20 years. Ann Thorac Surg. 2011;92:1678–1684.
6. Maganti M, Badiwala M, Sheikh A, et al. Predictors of low cardiac output syndrome after isolated mitral valve surgery. J Thorac Cardiovasc Surg. 2010;140:790–796.
7. Maganti MD, Rao V, Borger MA, Ivanov J, David TE. Predictors of low cardiac output syndrome after isolated aortic valve surgery. Circulation. 2005;112:I448–I452.
8. Topkara VK, Cheema FH, Kesavaramanujam S, et al. Coronary artery bypass grafting in patients with low ejection fraction. Circulation. 2005;112:I344–I350.
9. Kalogeropoulos AP, Georgiopoulou VV, Gheorghiade M, Butler J. Echocardiographic evaluation of left ventricular structure and function: new modalities and potential applications in clinical trials. J Card Fail. 2012;18:159–172.
10. Macron L, Lairez O, Nahum J, et al. Impact of acoustic window on accuracy of longitudinal global strain: a comparison study to cardiac magnetic resonance. Eur J Echocardiogr. 2011;12:394–399.
11. Smiseth OA, Torp H, Opdahl A, Haugaa KH, Urheim S. Myocardial strain imaging: how useful is it in clinical decision making? Eur Heart J. 2016;37:1196–1207.
12. Ternacle J, Berry M, Alonso E, et al. Incremental value of global longitudinal strain for predicting early outcome after cardiac surgery. Eur Heart J Cardiovasc Imaging. 2013;14:77–84.
13. Marcucci CE, Samad Z, Rivera J, et al. A comparative evaluation of transesophageal and transthoracic echocardiography for measurement of left ventricular systolic strain using speckle tracking. J Cardiothorac Vasc Anesth. 2012;26:17–25.
14. Burns AT, La Gerche A, D’hooge J, MacIsaac AI, Prior DL. Left ventricular strain and strain rate: characterization of the effect of load in human subjects. Eur J Echocardiogr. 2010;11:283–289.
15. Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2015;16:233–270.
16. Voigt JU, Pedrizzetti G, Lysyansky P, et al. Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. Eur Heart J Cardiovasc Imaging. 2015;16:1–11.
17. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837–845.
18. Gallop RJ, Crits-Christoph P, Muenz LR, Tu XM. Determination and interpretation of the optimal operating point for ROC curves derived through generalized linear models. Underst Stat. 2003;2:219–242.
19. Kerr KF, Wang Z, Janes H, McClelland RL, Psaty BM, Pepe MS. Net reclassification indices for evaluating risk prediction instruments: a critical review. Epidemiology. 2014;25:114–121.
20. Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med. 2008;27:157–172.
21. Chandra S, Skali H, Blankstein R. Novel techniques for assessment of left ventricular systolic function. Heart Fail Rev. 2011;16:327–337.
22. Barbier P, Mirea O, Cefalù C, Maltagliati A, Savioli G, Guglielmo M. Reliability and feasibility of longitudinal AFI global and segmental strain compared with 2D left ventricular volumes and ejection fraction: intra- and inter-operator, test-retest, and inter-cycle reproducibility. Eur Heart J Cardiovasc Imaging. 2015;16:642–652.
23. Kalam K, Otahal P, Marwick TH. Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction. Heart. 2014;100:1673–1680.
24. Stoddard MF, Dillon S, Peters G, Kupersmith J. Left ventricular ejection fraction is increased during transesophageal echocardiography in patients with impaired ventricular function. Am Heart J. 1992;123:1005–1010.
25. Farsalinos KE, Daraban AM, Ünlü S, Thomas JD, Badano LP, Voigt JU. Head-to-head comparison of global longitudinal strain measurements among nine different vendors: the EACVI/ASE Inter-Vendor Comparison Study. J Am Soc Echocardiogr. 2015;28:1171.e2–1181. e2.
26. Hahn RT, Abraham T, Adams MS, et al.; American Society of Echocardiography; Society of Cardiovascular Anesthesiologists. Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American Society of Echocardiography and the Society of Cardiovascular Anesthesiologists. Anesth Analg. 2014;118:21–68.

Supplemental Digital Content

Copyright © 2017 International Anesthesia Research Society