Clinical significance of regional constructive and wasted work in patients receiving cardiac resynchronization therapy

Background Previous studies have shown that global constructive work (CW) and wasted work (WW) predict response to cardiac resynchronization therapy (CRT). This study evaluated the predictive value of regional CW and WW for reverse remodeling and clinical outcomes after CRT. Methods We performed a prospective study involving 134 CRT candidates with left bundle branch block and left ventricular ejection fraction ≤35%. Global and regional CW and WW were calculated using pressure-strain loop analysis. CRT response was defined by reverse remodeling as a reduction of ≥15% in left ventricular end-systolic volume after six months. Results At six-month follow-up, 92 (69%) patients responded to CRT. Of the regional CW and WW measures, lateral wall (LW) CW and septal WW were most strongly and significantly correlated with reverse remodeling. At multivariate analysis, LW CW and septal WW were both independent determinants of reverse remodeling. When LW CW and septal WW were included in the model, global CW and WW were not independently associated with reverse remodeling. LW CW and septal WW predicted reverse remodeling with an area under the curve (AUC) of 0.783 (95% CI: 0.700–0.866) and 0.737 (95% CI: 0.644–0.831), respectively. Using both variables increased the AUC to 0.832 (95% CI: 0.755–0.908). Both LW CW ≤878 mmHg% (HR 2.01; 95% CI: 1.07–3.79) and septal WW ≤181 mmHg% (HR 2.60; 95% CI: 1.38–4.90) were significant predictors of combined death and HF hospitalization at two-year follow-up. Conclusion LW CW and septal WW before CRT are important determinants of reverse remodeling and clinical outcomes.


Introduction
Cardiac resynchronization therapy (CRT) reduces morbidity and mortality in patients with heart failure (HF) and a wide QRS complex (1).However, a significant portion of patients who receive CRT do not respond favorably to the therapy (2).Several echocardiographic measures have been suggested to predict CRT response by analyzing the timing of mechanical events (3)(4)(5).Although these time-delay parameters initially showed promise, randomized controlled trials have shown that these parameters are not reliable predictors of CRT response (6,7).One possible explanation for these findings is that the mechanical dyssynchrony caused by primary electric dyssynchrony is the modifiable substrate for CRT (8).The problem with conventional time-delay indices of mechanical dyssynchrony is that they can also be caused by regional contractile disparities such myocardial ischemia, infarction, or scar, which are less likely to be amenable to CRT (8,9).As an alternative, visual assessments of apical rocking, septal flash, and left bundle branch block (LBBB) contraction pattern are used to assess left ventricular (LV) mechanical dyssynchrony, potentially overcoming the limitations of previously suggested parameters (10,11).
In a healthy heart, all LV segments contract synchronously and myocardial energy is used efficiently to eject blood into the aorta.However, when there is a delay in electrical conduction, segments that activate early and late contract at different times, leading to the wastage of myocardial energy in stretching opposing walls.Several studies have shown that non-invasive estimates of global constructive work (CW) or wasted work (WW) using pressure-strain loops predict reverse remodeling or mortality after CRT better than dyssynchrony indices do (12)(13)(14)(15).The combined assessment of myocardial CW and WW involves evaluating the contractile reserve and wasted energy caused by LV dyssynchrony, providing a comprehensive approach to evaluate the mechanisms underlying the CRT response.However, the prognostic value of regional CW and WW in CRT candidates has rarely been defined (16).In addition, a recent study showed that the combination of work difference between the septum and lateral wall (LW) with septal viability can be used to predict CRT response (17).The study employed cardiac magnetic resonance imaging with late gadolinium enhancement to evaluate septal viability.The current study aims to assess the efficacy of combining regional CW and WW in predicting reverse remodeling and clinical outcomes of patients undergoing CRT.

Study population
This was a prospective single-center study.We assessed patients with HF and LBBB who were undergoing CRT.We excluded patients who had atrial fibrillation, severe heart valve disease, or poor apical acoustic window.All patients were receiving optimized medical therapy at the time of CRT.An ischemic etiology was defined as a history of myocardial infarction, coronary revascularization, or angiographic evidence of multi-vessel disease or single-vessel disease with >75% stenosis of the left main or proximal left anterior descending artery.The study was approved by the institutional review board and complied with the Declaration of Helsinki.All patients provided written informed consent to participate in the study.

Conventional echocardiographic analysis
All patients underwent transthoracic echocardiography using a commercially available ultrasound probe and device (M5S probe, Vivid E9, GE Healthcare, Horten, Norway) before and six months after CRT.Two dimensional and pulsed wave Doppler data were stored and analyzed offline.LV volumes and function were obtained using the modified Simpson's rule.

Speckle tracking analysis
The study used digital loops of two-dimensional LV images for offline speckle-tracking analysis with a commercially available software (EchoPAC, GE Vingmed Ultrasound, Horten, Norway).The gain settings and sector width were adjusted to optimize the image quality with frame rates of 50-90 Hz.Twodimensional LV images were obtained at the apical fourchamber, two-chamber, and long-axis views for speckle-tracking strain analysis.To analyze LV longitudinal strain, the endocardial border was traced on an end-systolic frame, and the width of the region of interest was adjusted to include most of the LV myocardium.The software automatically tracked myocardial motion and generated six curves of segmental longitudinal strain for each apical view.Global longitudinal strain was computed as the average of peak systolic longitudinal strain of all LV segments.

Myocardial work assessment
The study utilized a vendor-specific software (EchoPAC version 202, GE Vingmed Ultrasound) to assess global and regional myocardial work.The peak LV pressure was assumed to be equivalent to the brachial systolic blood pressure, measured before the echocardiographic study.The software produced a previously validated noninvasive LV pressure curve that was adjusted based on the timing of ejection and isovolumic phases (18).These phases were defined by the timing of aortic valve and mitral valve opening and closing using spectral Doppler tracings.LV strain measured by speckle-tracking analysis and LV pressure curve were synchronized by aligning cardiac cycle phases and peak LV pressure.We quantified myocardial work by computing the rate of regional shortening via strain curve differentiation and multiplying this value by estimates of instantaneous LV pressure.Myocardial CW measurements quantified the amount of work performed during systolic shortening and the negative work performed while lengthening during isovolumic relaxation.Myocardial WW measurements quantified the amount of negative work performed while lengthening in systole and work performed while shortening in isovolumic relaxation.We computed regional CW and WW values for six regional walls (the inferior, posterior, lateral, anterior, anteroseptal, and septal walls) as the averages of the values for the basal-and mid-LV segments.We calculated the global values of CW and WW as the mean values for all LV walls.

Alternative approaches
Two experienced observers assessed the existence of septal flash, apical rocking, and LBBB contraction pattern before CRT.Septal flash was characterized as the thickening and thinning of the septum during the isovolumic contraction period, while apical rocking was described as the movement of the LV apical myocardium vertical to the LV long axis (10).The LBBB contraction pattern was recognized by analyzing longitudinal strain curves in the apical four-chamber view using three criteria (11).These criteria included: (1) early shortening of at least one basal-or mid-LV segment in the septum and early lengthening in at least one basal-or mid-LV segment in the LW; (2) early septal peak shortening occurring within the initial 70% of the ejection period; and (3) LW reaching peak shortening after aortic valve closure (11).The work difference between the LW and septum was calculated in the apical four-chamber view as the absolute difference between net myocardial work in the LW and septum (17).

Endpoints
The study's primary objective was to assess LV reverse remodeling, which was defined as a reduction in LV end-systolic volume (ESV) of ≥15% after six months of CRT.The secondary endpoint was the composite of all-cause death or hospitalization due to HF during a two-year follow-up.

Statistical analysis
Continuous data are presented as mean ± standard deviation and categorical data as number and percentage.Comparisons among continuous variables were examined using the Student's t-test.Comparisons among categorical data were performed using the chi-squared test.We evaluate the predictive performance of global and regional CW and WW for reverse remodeling by calculating receiver-operating characteristic cures and areas under the curve (AUCs).To identify CRT responders, we selected an optimal cut-off value that maximized the Youden index (sensitivity + specificity − 1).Pearson's correlation analysis was conducted to examine the association between values of CW and WW and the decrease in LV ESV following CRT.To assess the predictive value of variables for reverse remodeling, we employed logistic regression analysis.Variables that had a univariate p value of <0.05 were included in a multivariate model.We utilized a series of nested models by incorporating CW (global or lateral) and WW (global or septal) parameters.The incremental predictive ability of each model was assessed by comparing chi-square values at each stage.To determine the cumulative probabilities of all-cause death or HF hospitalization after CRT, we employed the Kaplan-Meier method, and between-group comparisons of cumulative event rates were calculated using the log-rank test.We evaluate the inter-and intra-observer agreement for CW and WW in 20 randomly chosen patients.A p-value of <0.05 was considered statistically significant.

Results
Table 1 summarizes the baseline characteristics of the 134 patients included in the study, with an average age of 69.0 ± 11.9 years, 54.5% of whom were male, and 37.3% had ischemic etiology.Five patients died before the six-month follow-up and were classified as non-responders.Of the remaining 129 patients, 92 achieved the primary endpoint of a reduction in LV ESV of ≥15%, resulting in a response rate of 69%.Responders exhibited a higher prevalence of non-ischemic etiology, less dilated LV, and a more preserved LV ejection fraction and global longitudinal strain than non-responders.Prior to CRT, there were significant differences in regional CW between responders and nonresponders in the posterior, lateral, anterior, and anteroseptal walls.There were also significant differences in regional WW in the anteroseptal and septal walls.Figure 1 displays the segmental values of myocardial work, CW, and WW in a responder (Panel A) and non-responder (Panel B) before CRT and after six months.Prior to CRT, the responder had marked differences in myocardial work, CW, and WW between regional walls, with large septal WW, which was converted to large CW with CRT.On the other hand, the non-responder shows smaller variations in myocardial work, CW, and WW before CRT.After CRT, there was only a modest improvement of septal function with noticeable WW in the posterior wall (LV pacing site).Supplementary Table S1 shows the effects of CRT on LV function and myocardial work.At follow-up, responders showed a significant improvement in LV ejection fraction and global longitudinal strain, whereas non-responders did not experience any changes in these parameters.Responders also exhibited significant improvements in septal WW, global CW, global WW, and work difference after six months, whereas non-responders did not show any significant changes in LW CW, septal WW, and global WW at follow-up.

Predictive characteristics for reverse remodeling after CRT
Based on the binary definition of reverse remodeling, posterior wall CW, LW CW, anterior wall CW, anteroseptal WW, and septal WW had an AUC greater than that under the line of no information (Table 2).Of the regional CW values, the LW CW varied the most between responders and non-responders [AUC: 0.783, 95% confidence interval (CI) 0.700-0.866,cut-off value 878 mmHg%, sensitivity 72%, specificity 74%].Of the regional WW values, septal WW varied the most between responders and non-responders (AUC: 0.737, 95% CI: 0.644-0.831,cut-off value: 181 mmHg%, sensitivity 88%, specificity 55%).Combining LW CW and septal WW increased the AUC to 0.832 (95% CI: 0.755-0.908).Figure 2 displays the response rates of patients whose regional CW and WW values met (true-positive rate) or did not meet (false-negative rate) the cut-off values.The results show that LW CW was superior to the other regional CW measures and global CW, with a true-positive rate of 86% and a false-negative rate of 46%.Septal WW was superior to the other regional WW measures and global WW, with a true-positive rate of 81% and a false-negative rate of 32%.The AUCs for global CW and WW were 0.732 (95% CI: 0.639-0.825)and 0.692 (95% CI: 0.589-0.796),respectively.Combining the global CW and WW increased the AUC to 0.759 (95% CI: 0.669-0.850).

Variables associated with reverse remodeling
Multivariate analysis, using the significant variables from the univariate analysis (Supplementary Table S2) revealed that non-ischemic etiology and LV end-diastolic volume were independently associated with reverse remodeling, and they were thus included in the baseline model (χ 2 = 26.7,Table 3).We then added the CW (lateral and global) and/or WW (septal and global) parameters to the model.The LW CW [odds ratio (OR) 1.26, 95% CI: 1.10-1.44 per 100-mmHg% increase] and septal WW (OR 1.33, 95% CI: 1.07-1.66per 100-mmHg% increase) were both independently associated with reverse remodeling.Model power improved when LW CW (χ 2 difference: 24.4,p < 0.001) and septal WW (χ 2 difference 17.2, p < 0.001) were added to the model.In contrast, global CW and WW were not independently associated with reverse remodeling when LW CW or septal WW was included in the model.The addition of LW CW >878 mmHg% (OR 4.09; 95% CI: 1.44-11.62)and septal The bull-eye plots and bar charts showing the values of myocardial work, constructive work (CW) and wasted work (WW) before and 6 months after cardiac resynchronization therapy (CRT).In (Panel A), which represents a responder, high values of CW were observed in the lateral wall, while predominant WW was seen in the septum before CRT.Following CRT, there was a significant improvement in septal function, and the septal WW was converted to substantial CW.On the other hand, (Panel B), which represents a non-responder, showed lower values of lateral wall CW and septal WW compared to the responder.After CRT, there was only a moderate improvement in septal function, and noticeable WW was observed in the posterior wall.Response rate at 6 months after CRT by determining whether the cut-off value for each CW and WW parameter was met in all patients.The dashed line indicates the response rate observed when ignoring the parameter.The number of patients who met or did not meet the cut-off criterion for each parameter is shown inside each bar.3).There were 66 patients (49%) with both LW CW >878 mmHg% and septal WW >181 mmHg%.Of this group, 29% (n = 19) showed ischemic cardiomyopathy, which was a significantly smaller proportion than was observed in the other group (p = 0.044).This presence of both LW CW >878 mmHg% and septal WW >181 mmHg% was associated with a high response rate (91%).There were 23 patients (17%) with both LW CW ≤878 mmHg% and septal WW ≤181 mmHg%.Their response rate was only 21%.The response rate in the 45 patients (34%) who had either LW CW >878 mmHg% or septal WW >181 mmHg% was 60%.

Event-free survival
Figure 4 displays the Kaplan-Meier curves dichotomized according to LW CW ≤878 mmHg% (log-rank p = 0.024) and septal WW ≤181 mmHg% (log-rank p = 0.001).Both CW ≤878 mmHg% (HR 2.01; 95% CI: 1.07-3.79,p = 0.031) and septal WW ≤181 mmHg% (HR 2.60; 95% CI: 1.38-4.90;p = 0.003) were significant predictors of combined all-cause death and hospitalization due to HF at two-year follow-up.Figure 5 displays the Kaplan-Meier curves stratified by the combined LW CW and septal WW parameters.Patients categorized in the "both" group, characterized by both LW CW >878 mmHg% and septal WW >181 mmHg%, demonstrated the most favorable  "either" group, with only one parameter meeting the criteria, were positioned between the "both" and "neither" groups in terms of their outcomes.

Discussion
This study extends prior researches on myocardial work and presents the novel finding that the assessment of regional CW and WW via non-invasive pressure-strain loops can offer valuable prognostic insights for individuals who were being considered for CRT.Prior to CRT, LW CW and septal WW were significantly correlated with the reductions in LV ESV after CRT and independently predicted reverse remodeling and clinical outcomes after CRT.Global CW and WW were similarly correlated with the extent of reverse remodeling; however, they did not independently predict reverse remodeling when LW CW and septal WW were taken into account.The latter two measures were useful for predicting CRT response among both ischemic and non-ischemic patients.
The rationale for using LW CW and septal WW to predict CRT outcomes is that electrical conduction delay in the failing heart provokes discoordinate contraction between the early-activated septum and the late-activated LW.In patients with HF and LBBB, the ventricular septum contracts early during the isovolumic contraction phase, and during ejection, the out-of-phase septal relaxation counteracts LV free wall contraction.Regional CW quantifies the work performed during systolic shortening and negative work while lengthening during isovolumic relaxation, and reflects the contractile reserve.Regional WW computes the amount of negative work performed while lengthening during systole and work performed while shortening during isovolumic relaxation, and reflects energy waste caused by mechanical dyssynchrony.CRT can recruit myocardial work that is internally wasted by discoordinate contraction, and assessing the LW CW and septal WW facilitates identification of the contractile reserve and recruitable substrate that are amenable to CRT.
Previous studies have shown the prognostic value of global CW and WW in CRT candidates (12)(13)(14)(15).In a study of 97 patients undergoing CRT, global CW was associated with CRT response and was significantly correlated with the reductions in LV ESV after CRT (12).Despite higher values of LW CW and septal   (12).One possible reason for this discrepancy is that global measures of CW and WW, derived from the average of all segments, may lose significant information that is embedded in the nonhomogeneous distribution of regional CW and WW in CRT candidates.Our results agree with the results of two other studies that separately showed the prognostic value of LW CW or septal WW in patients undergoing CRT (16,19).In a brief report on 168 CRT candidates, LW CW rather than septal WW was independently associated with CRT response, and a LW CW >881 mmHg% was associated with a 2.2-fold increase in CRT response odds (16).In a small study of 21 patients receiving CRT, septal WW rather than global WW was the only myocardial work factor that predicted LV ESV reductions after CRT (19).However, the definition of septal WW, negative work in percentage of positive work, differs from ours (19).
In the present study, we found that septal WW was less related to reverse remodeling after CRT in patients with ischemic cardiomyopathy.Distinguishing between systolic lengthening of the septum due to transmural scar and septal systolic stretching resulting from LBBB, presents a challenge.Consequently, the similarity in septal WW between patients with myocardial scar and patients with electrical conduction delay may account for the weaker association of septal WW with reverse remodeling after CRT in patients with ischemic cardiomyopathy.The considerable variability in the extent of septal WW among patients with LBBB likely reflects, at least in part, this mixed etiology of systolic lengthening.
The use in clinical practice of myocardial work assessment derived from non-invasive pressure-strain loops for prognostic and clinical decision-making purposes is increasing (17,(20)(21)(22)(23)(24)(25)(26).The reliability of non-invasive measures of myocardial work in comparison to invasive measures has been validated in experimental evaluations and computer simulations (18,27).In the present study, the combined approach of LW CW and septal WW offers a clinically feasible and relatively simple method for identifying CRT responders.Both parameters were measured from the basal-and mid-segments of the LW and septum in the apical four-chamber view, which can be obtained for all patients.Evaluating LW CW and septal WW incorporates the assessment of contractile reserve and energy waste, which are key factors determining the response to CRT.

Limitations
This study has some limitations.Firstly, it is a single-center study, which may limit the generalization of its findings to clinical practice.Secondly, the lack of a validation cohort to examine the results further limits its generalizability.Thirdly, to assess myocardial work, a vendor-specific module (EchoPAC, GE) that combines LV strain data with a non-invasive LV pressure curve is required.Lastly, the study did not evaluate septal viability, and it is unclear whether it provides additional valve over septal WW and LW CW.Further study may be needed to address this issue.

Conclusion
This study revealed that LW CW and septal WW before CRT, assessed based on pressure-strain loops, predicted reverse remodeling and clinical outcomes after CRT.These two measurements reliably identified potential CRT responders in both ischemic and non-ischemic patients, and may better identifying CRT responders than the work difference between septum and LW.

FIGURE 3
FIGURE 3Predicting reverse remodeling after CRT.Model χ 2 values are presented for a series of nested models.The baseline model included non-ischemic etiology and left ventricular end-diastolic volume (LVEDV).

FIGURE 4
FIGURE 4 Association of lateral wall (LW) CW and septal WW with event-free survival.The Kaplan-Meier curves were stratified based on the cut-off values for LW CW (878 mmHg%, Panel A) and septal WW (181 mmHg%, Panel B).

TABLE 1
Baseline characteristics of the entire population and based on CRT response.

TABLE 2
Predictive characteristics of regional constructive work and wasted work prior to cardiac resynchronization therapy.

TABLE 3
Variables associated with CRT response in the baseline model and after addition of constructive work and wasted work parameters.CI, confidence interval; CW, constructive work; LVEDV, left ventricular end-diastolic volume; LW, lateral wall; OR, odds ratio; WW, wasted work.

TABLE 5
Multivariate logistic regressiona analysis with LV reverse remodeling as dependent variable.

TABLE 4
Comparison of the area under the curves for predicting reverse remodeling after CRT.