Strain as an Outcome Predictor

Myocardial deformation analysis by strain has emerged in the past decade as a valuable tool to evaluate myocardial contractility. A large number of publications have shown the utility of strain to evaluate myocardial function outside of the operating room and demonstrated its value for risk stratification in both surgical and nonsurgical patients. However, few studies have examined strain measured from intraoperative transesophageal echocardiography (TEE) as a predictor for perioperative outcomes.1,2 As the field of myocardial deformation imaging develops, major considerations that can impact the application of strain measures in the perioperative setting are clinical applications, technique (TEE versus transthoracic echocardiography [TTE]), as well as 2-dimensional (2D) vs 3-dimensional (3D) analysis algorithms. In this issue of Anesthesia & Analgesia, Sonny et al3 and Amabili et al4 provide evidence that TEE-measured global longitudinal strain (GLS), assessed by speckle tracking, is a predictor of adverse outcomes for patients undergoing cardiac surgery. While both studies are similar in scope, significant differences in their patient populations, techniques, and findings warrant discussion here. Strain is a measurement expressing myocardial systolic function. The epicardial and endocardial myocardial fibers in the left ventricle (LV) are longitudinally oriented, and their systolic shortening is expressed as negative deformation or strain. With speckle tracking, strain is measured from the motion of the grey pixels in a 2D echocardiographic image of the LV. The average strain from all myocardial segments, obtained from the 3 midesophageal LV views, is termed GLS and is a measure of global LV systolic function. Strain measurements are performed in a semiautomated fashion from vendor-specific software so that normal (and abnormal) strain values are not necessarily interchangeable between echocardiographic systems. The studies by Sonny et al3 and Amabili et al4 presented in this issue both investigated the relationship between postoperative outcomes and preoperative peak systolic GLS measured from TEE—after anesthetic induction and before cardiopulmonary bypass. Sonny et al3 demonstrated that GLS (and global strain rate) were independent predictors of prolonged hospitalization, while LV ejection fraction (LVEF) was not. Amabili et al4 similarly proved the independent predictive ability of GLS for postoperative low cardiac output syndrome but found that GLS did not add benefit to the already established risk factors, including LVEF. The differences in findings may be due to different patient populations, intraoperative management, and strain analysis techniques. Major differences between the studies included: (1) cardiac surgical procedures performed (aortic valve replacement for aortic stenosis versus mixed coronary artery bypass grafting, mitral valve, and major aortic surgeries), (2) anesthetic management (volatile inhalational versus total intravenous anesthesia), and (3) choice of inotropic support for cardiovascular management (epinephrine versus dobutamine). In addition, while both Sonny et al3 and Amabili et al4 investigated GLS measured via 2D speckle tracking, the 2 studies used different ultrasound systems and software for semiautomatic speckle-tracking analysis. Differences in echocardiography platforms and software should be kept in mind when interpreting and comparing strain results between studies because absolute values for strain can vary by vendor.5 The results of Sonny et al3 and Amabili et al4 demonstrate a weaker relationship between preoperative GLS and postoperative outcomes as compared to previously published studies. For example, Ternacle et al6 demonstrated that GLS was an independent predictor of postoperative mortality and added incremental value over LVEF, as compared to Amabili et al,4 who found no added value of strain in addition to LVEF. These conflicting conclusions may be explained by differences in sample size, imaging modality, and echocardiographic technique used. Both Sonny et al3 and Amabili et al4 have relatively small sample sizes, whereas the Ternacle et al6 results are supported by a larger study (n = 425) with far greater power. In addition, Ternacle et al6 used TTE, not TEE-measured strain. Preoperative TTE versus intraoperative TEE imaging are associated with images acquired at different time points in the perioperative period.7 TTE-based studies, including Ternacle et al,6 most often acquire preoperative images when patients are awake and spontaneously ventilating before general anesthesia. In contrast, TEE-based preoperative strain measurements, including Sonny et al3 and Amabili et al,4 are acquired under general anesthesia with mechanical ventilation. Therefore, TEE versus TTE-acquired preoperative strain reflect measures LV systolic function taken at different times and under different loading conditions, and as such the results and the relationships to LVEF and perioperative outcomes may vary. When discussing myocardial deformation measures and the clinical correlations to outcomes, it is also important to discuss the differences in 2D vs 3D echocardiography. Myocardial contraction and relaxation is a complex 3D motion involving left ventricular shorting, wall thickening, and rotation.8 Global strain can be measured with 2D echocardiography in the longitudinal, circumferential, and radial axes.2 However, in the aforementioned 3 studies (Ternacle et al,6 Sonny et al,3 and Amabili et al4), only 2D strain in the longitudinal axis (GLS) was measured. The 2D echocardiographic imaging planes are narrow, and speckle tracking may be lost between frames. The 3D echocardiography overcomes this limitation because the entire left ventricular myocardium is captured in a single imaging sector, so that speckle-tracking software can more accurately measure myocardial deformation.9 The 3D myocardial deformation imaging enables the measurement of longitudinal and circumferential and radial global strain. In addition, a composite strain measurement, global area strain, which incorporates changes in both longitudinal and circumferential strains, can be measured.9 A recently published study in patients undergoing cardiac surgery demonstrated that decreased preoperative 3D strain global area strain was an independent predictor of acute and long-term outcomes after both coronary artery bypass grafting and valve surgeries.10 Currently, 3D speckle-tracking strain imaging is only available for TTE. However, with the rapid advancement in technology, it is expected to be available for real-time, quick, easy, and semiautomated TEE modality in the near future. Myocardial strain imaging has been shown to be a reproducible and accurate measure of ventricular function with robust data for clinical utility. However, its intraoperative use appears limited. The anesthesiologists’ hesitation to embrace this new technology may be due to scarce anesthesia literature demonstrating the clinical value of intraoperative strain. It is important to note that there are differences in strain measures of ventricular function analyzed from TTE and TEE,7 as well as different imaging and software platforms. It is therefore important for studies to be performed using intraoperative TEE strain in the perioperative period. As technology continues to improve the reliability of strain measurements across vendors and make semiautomated analysis easier to use, the utility of this new tool will continue to grow among anesthesiologists and intensivists. Future studies will hopefully aim at demonstrating even greater utility of intraoperative strain. DISCLOSURES Name: Kimberly Howard-Quijano, MD, MS. Contribution: This author helped conceptualize, write, and edit the manuscript. Name: Kan Zhang, MD. Contribution: This author helped conceptualize, write, and edit the manuscript. Name: Nikolaos J. Skubas, MD, DSc, FACC, FASE. Contribution: This author helped conceptualize, write, and edit the manuscript. This manuscript was handled by: Roman M. Sniecinski, MD. Acting EIC on final acceptance: Thomas R. Vetter, MD, MPH.