Change in Left Ventricular Peak Global Longitudinal Strain in Patients Undergoing Uneventful On-Pump Coronary Artery Bypass Grafting Surgery

INTRODUCTION

Systolic function of the left ventricle (LV) is used mainly to predict perioperative mortality and morbidity in patients undergoing coronary artery bypass grafting (CABG) surgery.^[1] Trans-esophageal echocardiography (TEE) plays the main role in assessing LV function intra-operatively in patients undergoing CABG surgery.^[2] While LV ejection fraction (EF), measured by Simpson’s biplane method, is the most widely used parameter and is considered a cornerstone for measurement of LV systolic function as well as for prediction of outcome, it has various limitations such as inter and intra-observer variability, load dependency, and tethering.^[3,4] Two-dimensional (2D) speckle tracking imaging (STE) enables measurement of myocardial deformation, which provides high quality, precise, and objective information regarding regional and global myocardial function in real time. Global longitudinal strain (GLS) by STE can identify subtle LV dysfunction, has significantly less inter and intra-operator variability than EF, and is more precise method for evaluation of LV function with/without echocardiographic training.^[4,5] Systolic LV function assessed using GLS is particularly important as it can reveal systolic abnormalities despite normal LVEF, which are associated with less favorable outcome.^[6] LVEF is predominantly related to LV circumferential shortening, whereas GLS depicts LV longitudinal shortening. As the vulnerable subendocardial myofibers are responsible for the longitudinal shortening of LV, GLS is able to detect subtle changes in myocardial function.^[3,6]

During cardiac surgery, the LV function is measured intraoperatively by TEE. However, in the immediate post-CABG period, the LV function is influenced by general anesthesia, positive pressure ventilation, myocardial protection technique used, the effect of myocardial revascularization, and the effect of inotropic and vasopressor agent used.^[7] However, the magnitude of the combined effect of the above-mentioned confounding factors in the immediate post-operative period on LV function is not known. Moreover, although the change of peak LV GLS over the perioperative period of CABG surgery has been studied;^[8] to the best of our knowledge, literature is scarce specifically regarding the change in peak LV GLS from immediate postoperative period to immediate pre-hospital discharge period (Postoperative day 7–10), i.e., the change of peak LV GLS over the early postoperative period in patients undergoing CABG surgery. As the effects of anesthesia, positive pressure ventilation, cardioplegia, inotropes, and vasopressors wean off by the time of hospital discharge (postoperative day 7–10), finding the difference between peak LV GLS measured in the immediate post-operative period and that measured just before hospital discharge (postoperative day 7–10) can estimate the combined effect of all the intra-operative confounders on peak LV GLS.

Therefore, we intended to perform a study evaluating the change in the LV function, measured by peak LV GLS, from the immediate post-operative period to the immediate pre-hospital discharge period (7–10 days after surgery) in patients undergoing uneventful, on-pump CABG surgery and having uneventful early post-operative (during hospital stay) course.

MATERIAL AND METHODS

RESULTS

Among eighty-eight (88) patients approached for the study, the study could be successfully completed as well as analyzed in fifty-one (51) patients [Figure 1]. The demographic parameters, comorbidities, pre-operative and pre-discharge medications, surgical and CPB characteristics, baseline (pre-induction) echocardiographic and hemodynamic parameters are summarized in Table 1.

Close

Figure 1:

A schematic flow diagram summarizing the selection of the study population LVEF: Left ventricular ejection fraction, MR: Mitral regurgitation, CABG: Coronary artery bypass grafting, POD: Post-operative day, TTE: Transthoracic echocardiography.

Export to PPT

Pre-induction Peak LV GLS (Mean [SD]) and LVEF (Median [IQR]) was 14.5 (4.9)% and 58 (55–64)%, respectively [Table 1]. There were statistically significant differences in Peak LV GLS, C.I, LVEF, HR, and MAP between the immediate post-operative period and pre-hospital discharge period [Table 2 and Figure 2]. However, There was no statistically significant difference of change in Peak LV GLS from Immediate post-operative period to pre-hospital discharge period between patients with Normal (n = 17) and Reduced (n = 34) baseline (pre induction) Peak LV GLS (Mean [SD] difference −1.8 (3.9)% vs. −4.1 (4.8)%, P = 0.079, Welch’s t-test). There was statistically significant, but weak association of the baseline peak LV GLS with the change in Peak LV GLS from immediate postoperative period to the pre-hospital discharge period (R² = 0.156, P = 0.003) [Figure 3]. However, post-operative VIS was not found to be significantly associated with the change in Peak LV GLS from the immediate post-operative period to pre-hospital discharge period (R² = 0.003, P = 0.696) [Figure 4].

Close

Figure 2:

Boxplot showing comparison between immediate post-operative and pre-hospital discharge peak LV GLS (The horizontal bars represent the median values; the upper and the lower margins of the boxes represent the 1^st and the 3^rd quartiles, respectively; the top and the bottom whiskers represent the minimum and the maximum values, respectively. X represents Mean values). (*P < 0.05 in comparison to Immediate Post-operative Peak LV GLS). LV GLS: Left ventricular global longitudinal strain.

Export to PPT

Close

Figure 3:

Scatterplot with superimposed linear regression line showing association of change in Peak LV GLS from immediate post-operative period to pre-hospital discharge period with the baseline (pre-induction) Peak LV GLS (absolute value). LV GLS: Left ventricular global longitudinal strain.

Export to PPT

Close

Figure 4:

Scatterplot with superimposed linear regression line showing association of change in Peak LV GLS from immediate post-operative period to Pre-Hospital discharge period with VIS at immediate post-operative period. LV GLS: Left ventricular global longitudinal strain, VIS: Vasoactive-inotropic score.

Export to PPT

Based on the VIS score in the immediate post-operative period, the patients were divided into three equal-sized subgroups according to VIS-tertiles – Low (Median 0.09 [IQR 0.00–1.33]), Intermediate (Median 4.62 [IQR 3.43–5.13]), and High (Median 12.8 [7.7–17.5]).

Although VIS score was significantly different between the VIS tertiles (P-value 0.000 with Friedman’s test), change of Peak LV GLS from Immediate Postoperative to Pre-Hospital Discharge period, Peak LV GLS in the immediate postoperative period, Change in C.I from Immediate Postoperative to Pre Hospital Discharge period, C.I in the immediate postoperative period were not significantly different between the subgroups of patients according to these VIS-tertiles [Table 3].

The inter- and intra-observer agreements of the echocardiographic measurements at both the Immediate Postoperative period and the Pre Hospital Discharge period showed that except LVEF, all other parameters had acceptable (ICC>0.8) inter-observer as well as intra-observer agreement at both the timepoints.

DISCUSSION

In our patients, although LVEF improved statistically significantly before hospital discharge in comparison to the immediate post-operative period (56.4 (53–62)% to 58.9 (56–64.1)%, P = 0.011), the peak LV GLS showed a statistically significant decrease over the same time period (−12.8 [3.8]% to −10.2 [2.4]%, P = 0.000). However, LVEF values in both immediate post-operative, as well as pre-discharge period were > 50% and the difference in LVEF between these two time points was only 2.5%. Hence, the difference in LVEF between these time points was clinically insignificant.

Myocardial strain analysis is a more sensitive marker of global as well as regional LV dysfunction than LVEF. Reproducibility and inter-observer agreement of Peak LV GLS are better than those of LVEF calculated by Simpson’s method.^[13] Two-thirds of our patients had impaired pre-operative Peak GLS despite having normal pre-operative LVEF. This is consistent with existing literature that a significant number of CAD patients scheduled for CABG have impaired Peak GLS despite having normal LVEF.^[14,15] Lack of improvement in GLS after coronary revascularization is associated with negative LV remodeling.^[16] While a few studies have reported improvement in LV GLS after CABG,^[17] a few have reported decrement, too.^[14] However, most of the studies have assessed GLS after 1 to 24 months of CABG.

Difference of GLS between pre-operative baseline state and postoperative follow-up is mainly dependent on reverse remodeling of myocardium due to revascularization, which may continue to evolve over months.^[18] Perioperative period, on the other hand, is more dynamic. Changes in Peak LV GLS within the span of the perioperative period of CABG surgery can be affected by a multitude of perioperative factors, like effect of anesthetic agents, altered loading conditions, positive pressure ventilation, ischemia-reperfusion injury, CPB, cardioplegia, vasopressors, and/or inotropes.^[8] Perioperative echocardiographic assessment of cardiac function is a quintessential part of cardiac anesthesiology and intraoperative assessment of LV function is predominantly dependent on TEE.^[19]

The peak LV GLS measured intraoperatively by TEE immediately after CABG, which is used to quantify post-CPB left ventricular function, is expected to change at the time of the patient’s discharge, when the perioperative factors influencing LV function are absent. This magnitude of change in peak LV GLS in patients receiving a standardized perioperative management and having an uneventful perioperative course describes the course of peak LV GLS in the immediate postoperative period. Although a few studies have investigated the change of peak LV GLS in the perioperative period, the change in LV function over the early post-operative period has not been addressed adequately. Our study fills this particular knowledge gap in literature. This knowledge for making prediction regarding pre-discharge LV function of a person undergoing CABG and having uncomplicated early post-operative course, as pre-discharge TTE is often not feasible in a busy tertiary care hospital due to the enormous workload on the cardiology department and also due to poor transthoracic imaging window in a significant number of patients in the early post-operative period.

A multitude of reasons can be attributed to the significant deterioration of peak LV GLS from the immediate postoperative period after CABG (measured just before shifting from Operating Room [OR] to Intensive Care Unit [ICU]) to that just prior to the hospital discharge, despite having uneventful intraoperative and post-operative course. The intraoperative measurement of peak LV GLS was done with TEE after completion of CABG and sternal closure, just before shifting to ICU. The patients were under anesthesia with a median BIS of 42, were on positive pressure ventilation, and were on vasoactive-inotropic support with a median VIS of 4.8. Moreover, they were shortly weaned off from CPB after CABG, therefore, still under the immediate effects of myocardial revascularization, ischemia-reperfusion effect of aortic clamping and unclamping, systemic effect of CPB as well as residual effect of cardioplegia. On the other hand, pre-discharge peak LV GLS was assessed using TTE, patients were awake and spontaneously breathing, not under any vasoactive-inotropic support and a median of 9 days had passed since they had undergone surgery, therefore, the immediate effects of CPB, ischemia-reperfusion injury and cardioplegia had likely weaned off by that time although these changes can continue to evolve over time.^[20]

Labus et al.^[8] found significant decrease of 2D LV GLS from pre-operative baseline to pre-hospital discharge period after on-pump CABG with the uneventful perioperative course, although LVEF remained unchanged. We also found a similar decrease in 2D LV GLS from pre-induction baseline to the pre-hospital discharge period, with no clinically significant difference in LVEF. Existing studies also suggest that 2D-LV GLS deteriorates after on-pump cardiac surgery as well as after off-pump CABG.^[14,18,21]

Our study adds further to the literature by gaining insight into the change of peak LV GLS from immediate postoperative period after CABG to the pre-hospital discharge period, thus quantifying the change of LV function over the dynamic early post-operative period. As we chose patients who had uneventful surgery with uneventful intraoperative as well as postoperative period, the change in peak LV GLS over this time period arguably represents the usual change of LV function that occurs over this early post-operative period with gradual attenuation of the intraoperative factors that were present during the intraoperative assessment of LV function immediately after CABG.

Labus et al.^[8] did not find any significant difference between TTE-derived peak LV GLS obtained in awake, spontaneously breathing state and TEE-derived peak LV GLS obtained in anesthetized state with positive pressure ventilation in a similar population of patients like ours. Although TTE-based studies like study by Dalla et al.^[7] have found significant deterioration of peak LV GLS upon induction of anesthesia and onset of positive pressure ventilation, their study population was devoid of any known myocardial disease. Cinotti et al.,^[22] in their study involving non-cardiac population had found improvement in peak LV GLS after cessation of general anesthesia and extubation. Based on these, it can be argued that the deterioration of the peak LV GLS from immediate post-operative period after CABG to the pre-hospital discharge period was not contributed significantly by the transition from anesthetized state with positive pressure ventilation to awake, spontaneously breathing state.

Peak LV GLS is also affected by inotropic and vasopressor therapy through their effect on myocardium and the loading conditions.^[23] Unlike the study of Labus et al.^[8] where peak LV GLS was measured intraoperatively by TEE once before and once after sternotomy before institution of CPB, without any influence of vasopressor or inotropic support, peak LV GLS was measured in our study intraoperatively after CABG was completed and after weaning from CPB, a period where vasoactive or inotropic support is almost unavoidable. LV strain has been found to increase dose dependently with the use of inotropic agents.^[23] We did not find any significant linear relation between the magnitude of the decrement of peak LV GLS from the immediate postoperative period after CABG to the pre-hospital discharge period and the VIS at the immediate post-operative period after CABG. The finding was further supplemented by comparing the decrement of peak LV GLS among the patients receiving low, intermediate and high vasoactive-inotropic support at the immediate postoperative period after CABG, where no significant difference was found among the patients in low, intermediate, and high VIS tertiles. This lack of association can possibly be explained by the fact that vasopressors and inotropes are used in the immediate post-operative period after cardiac surgery to counteract the effects of revascularization, CPB, aortic cross clamp (AXC), and cardioplegia on the myocardium and vasculature to achieve and maintain adequate cardiac output as well as systemic perfusion pressure.^[24] The effects of the intraoperative factors on the myocardium and vasculature vary from patient to patient leading to different amounts of dysfunction of myocardium and vasculature on different patients. The more the dysfunction, the more is the requirement of the vasoactive-inotropic drugs to maintain adequate cardiac output and/or systemic perfusion pressure. Therefore, the measurements made at the immediate postoperative period after CABG represent a balanced state of counteraction between at one side the factors that depress LV function and the factors that enhance LV function at the other side. Due to the same reason, the magnitude of change of peak LV GLS as well as C.I did not have an association with the VIS score at the immediate post-operative period. Although based on the findings of our study, we cannot definitely say that the deterioration of peak LV GLS was not contributed by the presence and absence, respectively, of vasoactive-inotropic support, we can say that the magnitude of the decrement in peak LV GLS was not associated with the level of vasoactive-inotropic support present at the immediate post-operative period after CABG.

Impaired peak LV GLS in the pre-operative period is an independent predictor of poor prognosis after cardiac surgery,^[15] over and above LVEF.^[25] In our study, the deterioration of peak LV GLS from the immediate postoperative period to pre-hospital discharge period was not significantly different between patients having normal and impaired pre-operative peak LV GLS. The association between pre-operative peak LV GLS and the decrement in peak GLS was also found to be weak. However, as the sample size of our study was not calculated to look at this association specifically, a definitive conclusion regarding the same cannot be drawn from our findings.

Peak LV GLS is dependent on loading conditions on heart at the time of measurement.^[26] Peak LV GLS increases with increase in preload and decreases with decrease in preload.^[27] There was no significant difference in LVEDA between the two time points. CVP estimated by IVC diameter and variability at the time of hospital discharge was higher than the directly measured CVP at the immediate post-operative period after CABG. Therefore, the preloads at the two time-points were arguably similar. Although comparison of pulmonary capillary wedge pressure measured through a pulmonary arterial catheter between the two time points would have been the gold standard, it is not feasible for a patient to have pulmonary arterial catheter at the time of hospital discharge. Therefore, based on our findings, it is to a reasonable extent unlikely that the decrement of peak LV GLS was related to preload.

The afterload, estimated by SVRI, was significantly higher at the time of hospital discharge than at the immediate postoperative period after CABG. This could potentially be a significant contributor toward the decrement of peak LV GLS, as increase in afterload is associated with decrease in peak LV GLS. Lower afterload immediately after CABG in comparison to that at the time of hospital discharge can possibly be explained by presence of general anesthesia-induced dilation of systemic resistance vessels,^[28] reduced transmural pressure due to positive pressure ventilation,^[29] use of inodilator agent (milrinone), and vasoplegia after CPB at the intraoperative period immediately after CABG.^[30] Although no significant linear relation was found between the change in peak LV GLS as well as C.I and the change in SVRI, the lack of linear relation can be explained by the complex interplay and inter-dependency of the loading conditions and myocardial contractility in the loop-type human circulatory system.

There were certain strengths of our study, the intraoperative as well as the post-operative management of our patients was essentially uniform, according to a pre-standardized protocol. All the surgeries were performed by a single surgeon, using essentially the same surgical technique. The techniques of myocardial preservation were also similar among the patients. The CPB and AXC durations, with a mean of 138 min and 97 min, respectively, were not prolonged. The pre-operative risk status of the patients, with a median (IQR) Euroscore II of 0.87 (0.78–0.99), was low. Only the patients with an uneventful, uncomplicated intraoperative as well as postoperative course till hospital discharge were included. Hence, findings of our study essentially represent the usual course of change of LV function from that at the immediate postoperative period, to that at the time of hospital discharge, in low-risk patients with normal pre-operative LV function (by LVEF) undergoing uneventful, multivessel CABG with uneventful early post-operative course. Thus, our findings shed light into the change of LV function over the early post-operative course till the time of discharge from the hospital after CABG a scarcely investigated domain till now. The findings of our study can form the base upon which course of change of LV function in this early post-operative period in high-risk patients, patients with pre-operative impaired LVEF, and patients having not-so-benign intraoperative and/or early post-operative course after CABG can be studied.

The association between the change of peak LV GLS over this early post-operative period with the likely perioperative factors that can affect peak LV GLS values was also looked upon in our study, although probably underpowered for the same. With the help of our findings, adequately powered studies can be designed in the future to understand the contribution of different perioperative factors in affecting the change of LV function over this scarcely investigated early post-operative period after CABG.

CONCLUSION

Low-risk patients with normal LVEF undergoing multivessel CABG surgery with uneventful intraoperative and early postoperative (during hospital stay) course have significant decline in peak LV GLS from the immediate intraoperative period (immediately before shifting from OR to ICU) to the time just before their discharge from hospital, despite having no clinically significant difference in LVEF between the two time points. Increase in afterload explains in part this decline of peak LV GLS over this time period.

Further studies with adequate power are required to find whether the decrement of peak LV GLS over the early postoperative period bears any prognostic significance, which can, in turn, direct the need for routine measurement of peak LV GLS in the perioperative and early post-operative period in patients undergoing CABG.