Association between left ventricular geometry and global myocardial work in patients with heart failure with preserved ejection fraction: assessment using strain-pressure loop

Concentric LV remodeling and hypertrophy are common structural abnormalities in patients with heart failure with preserved ejection fraction (HFpEF) and tend to be accompanied by impaired LV function. Assessment of global myocardial work (GMW) using strain-pressure loop may provide more comprehensive assessment of LV myocardial function, overcoming the limitations of the conventional parameters. We investigated the value of GMW in patients with HFpEF and assessed the relationship of GMW with concentric remodeling and hypertrophy. Consecutive patients with HFpEF (n = 107) and sex-matched healthy controls (n = 32) were prospectively enrolled. Clinical and conventional echocardiography variables were obtained. Further analyses of offline data were performed to obtain GMW indices including global work index (GWI), global constructive work (GCW), global waste work (GWW), and global work efficiency (GWE). Association of concentric remodeling and hypertrophy with GMW was analyzed by univariate and multivariate analysis. HFpEF patients showed lower GWE (94% vs 96%, P < 0.001) and higher GWW (114 mmHg% vs 78 mmHg%, P = 0.003) than control group, while GWI (2111 mmHg% vs 2146 mmHg%, P = 0.877) and GCW (2369 mmHg% vs 2469 mmHg%, P = 0.733) were comparable in the two groups. HFpEF patients with relative wall thickness (RWT) > 0.42 had reduced GWE (94% vs 95%, P = 0.034) compared to HFpEF patients with RWT ≤ 0.42, while GWI, GCW, and GWW were comparable between these two subgroups. Multivariate analysis showed an independent association of RWT with GWI, GCW, and GWE, respectively. Impaired global myocardial work was detected in patients with HFpEF. Impaired LV GMW may be associated with increased RWT.


3
Concentric remodeling and hypertrophy are the most common left ventricular (LV) structural abnormalities in patients with HFpEF and these changes are associated with adverse prognosis [5][6][7]. Pathological LV remodeling and hypertrophy induced by abnormal hemodynamic stress further lead to LV dysfunction [8].
Although left ventricular ejection fraction (LVEF) is the conventionally used index for assessment of LV function, it largely reflects the pump function rather than the mechanical function [9]. Recently, global longitudinal strain (GLS) has been shown to be an effective parameter for evaluating the mechanical function [10]. Studies have demonstrated deteriorated strain in patients with normal LVEF, particularly in patients with HFpEF [11,12]. However, GLS is pressure-load dependent [13], and may not be the best parameter to evaluate the LV myocardial function in patients with increased afterload (such as patients with HFpEF). Therefore, development of a new method for assessment of LV myocardial function that is not dependent on afterload is a key imperative [14].
Global myocardial work (GMW) has been introduced as a novel non-invasive method for myocardial function assessment [15]. GMW is derived from the non-invasive pressure strain loop, and this method has been validated against invasive methods for myocardial function assessment. GMW is a useful method for estimation of LV myocardial function as it integrates both pressure load and LV longitudinal strain; this is particularly valuable in patients with HFpEF, in whom hypertension is one of the most common comorbidity [16,17]. Moreover, global myocardial work efficiency (GWE) derived from pressure strain loop method has been shown to be a sensitive marker of subclinical myocardial damage in HFpEF patients [18].
Recently, the Heart Failure Association (HFA) of the European Society of Cardiology (ESC) suggested taking increased relative wall thickness (RWT) into account when diagnosing HFpEF in clinical practice [19], while the guidelines for heart failure diagnosis take account of concentric remodeling and hypertrophy as an objective evidence of HFpEF [20]. Previous studies had showed that heart failure patients with increased RWT may show worse LV myocardial function [21,22]. However, the association of concentric remodeling and hypertrophy with non-invasive global myocardial work is not well characterized in the contemporary literature. Thus the aim of this study was to explore the relationship between impaired global myocardial work with concentric LV remodeling and hypertrophy.

HFpEF group
According to the latest ESC guidelines [20], we enrolled 168 consecutive patients with signs or symptoms of HF and LVEF ≥ 50% at the Beijing Chaoyang Hospital, Capital Medical University, Beijing, China. The recruitment period lasted from March 2021 to March 2022. Patients were further divided into two subgroups based on the median value of GWE.
The exclusion criteria were: congenital heart disease, severe mitral annular calcification, severe mitral regurgitation, severe aortic valve disease, prosthetic heart valve or prosthetic ring, severe chronic obstructive pulmonary disease, persistent atrial fibrillation, and poor acoustic window.

Control group
Thirty-two sex-matched asymptomatic subjects who were referred to our hospital for screening were included as the control group. Subjects were excluded if they had: (1) hypertension, (2) type 2 diabetes, (3) coronary artery disease, (4) congenital heart disease, (5) more than trivial valve regurgitation or valve stenosis of any degree, (6) any previous cardiac or vascular surgery or interventional procedure, (7) any kind of cardiac therapy.
This prospective study was approved by the ethics committee of Beijing chaoyang Hospital Affiliated to capital medical University (No: 2021-11-26-17).

Clinical variables
At the time of transthoracic echocardiography (TTE), data pertaining to the following demographic and clinical variables were collected: age, sex, systolic and diastolic blood pressure, heart rate, height, and weight. Medical history of hypertension, diabetes, coronary artery disease, and smoking were obtained simultaneously. Plasma B-type natriuretic peptide and serum creatinine levels were determined within 7 days of the TTE. The estimated glomerular filtration rate (eGFR) was measured using the modified MDRD Study equation for Chinese (the Chinese equation) [23].

Conventional echocardiography
Two-dimensional and Doppler echocardiography were performed using standard equipment (Vivid E95, General Electric Medical Systems, Milwaukee, WI, USA). At least 3 consecutive cardiac cycles of standard echocardiographic views were acquired and stored digitally for subsequent analysis by trained physicians. LVEF, left atrial volume index (LAVi), and stoke volume (SV) were measured through tracing the apical four-and two-chamber views based on the modified biplane Simpson's method. Left atrial enlargement was defined as LAVi ≥ 34 mL/m 2 . Left ventricular dimensions variables, including left ventricular end-diastolic diameter (LVEDd), interventricular septum thickness at end diastole (IVSd), and posterior wall thickness at end diastole (PWTd), were measured according to the recommended criteria, and LV mass index (LVMi) ⟨LVMi = {0.8 * 1.04 * � (IVSd + LVEDd + PWTd) 3 − LVEDd 3 � + 0.6 � ∕BSA⟩ and relative wall thickness (RWT) were calculated accordingly. Mitral inflow peak (E), peak (A), and issue Doppler mitral annular early diastolic velocity (e') were measured for calculation of E/A and E/e'. All indices were measured in accordance with the latest chamber quantitation guidelines of the American Society of Echocardiography [24].

Speckle tracking imaging
Global longitudinal myocardial strain was assessed using a semiautomated 2D speckle tracking software (EchoPac 204, General Electric Medical Systems) in the apical four chamber (A4C) view, apical two chamber (A2C) view, and apical three chamber (A3C) view with typical temporal resolution of 60-90 frames/s. After manual tracing of the endocardial border and selecting the appropriate wall thickness, the software automatically identified six segments in each view and tracked the motion of acoustic markers. Segments that failed to track satisfactorily were readjusted manually, and, if this was ineffective, these were excluded from further analysis. Timing of the aortic and mitral valve opening and closure were obtained using single-gated pulsed-wave Doppler traces. GLS was derived as the average of the 17 segmental strain values and is described as an absolute value for simplicity.

Global myocardial work evaluation
Global myocardial work was calculated as recommended [25]. In brief, GMW was measured using the combination of LV GLS and the subject's brachial cuff blood pressure in place of left ventricular pressure. LV GLS was measured as above taking R-wave in ECG as the onset. After 5-10 min rest and choosing an appropriate cuff, blood pressure was measured in the supine position immediately before the echocardiographic study. All echocardiography and blood pressure data were stored and then analyzed offline. After strain measuring and valvular event timing setting, we could get a pressure strain loop (PSL). The software further calculated the following four parameters:

Intra-and inter-observer variability
Twenty-five patients were randomly selected and independently remeasured by another two researchers. Intraobserver variability was assessed at two different time points more than 4 weeks apart, while inter-observer variability was assessed by comparing the measurements by two different researchers on the same patient data. The reproducibility and variability were analyzed using Bland-Altman plots and intraclass correlation coefficient (ICC).

Statistical analysis
Normality of distribution of continuous variables was assessed using the Kolmogorov-Smirnov test. Continuous variables are reported as mean ± standard deviation or median (interquartile range), as appropriate. Categorical variables are expressed as frequency (percentage). Betweengroup differences with respect to continuous variables were assessed using the Mann-Whitney U test or Kruskal-Walli's test. Between-group differences with respect to categorical variables were assessed using the Chi-squared test or Fisher exact test, and a covariance analysis was performed when necessary. Spearman's rank correlation coefficient was used to assess the correlation between different variables, and then linear multivariate regression or logistic multivariate analysis was performed to evaluate the independent effect of covariates. Patients with HFpEF were further divided into two subgroups with different RWT according to the latest guidelines for cardiac chamber quantification [24]. P values < 0.05 were considered indicative of statistical significance. All statistical analyses were performed using IBM SPSS Statistics version 25 (IBM, Armonk, New York), Graphpad Prism version 8.0 and Origin Pro 2021.

Patient characteristics
Between January 2021 and February 2022, a total of 168 patients were clinically diagnosed as HFpEF and all of them underwent TTE. After exclusion, the final study population consisted of 107 patients (mean age: 65 ± 12 years; 50% male) and 32 sex-matched subjects (mean age: 60 ± 8 years; 47% male). The study flow chart is shown in Fig. 1.
Representative examples of significantly different LV myocardial work characteristics in two patients with different RWT are illustrated in Fig. 2. The patient with higher RWT (0.60) showed a smaller area of PSL, more impaired GWE, higher GWW, and slightly reduced GCW than the other patient with lower RWT (0.38).

Correlation analysis
GWI, GCW, and GWE showed a positive correlation with LVEF (r = 0.33, 0.32, and 0.30, respectively; P < 0.05 for all) and GLS (r = 0.61, 0.58, and 0.58, respectively; P < 0.05 for all); this indicated a good correlation of GWI, GCW, and GWE with the conventional indices for myocardial function assessment. To varying degrees, all parameters of left ventricular myocardial function showed a correlation with LVMI and RWT. GWE showed a negative correlation with both LVMi (r = − 0.44, P < 0.05) and RWT (r = − 0.31, P < 0.05). GWW showed a positive correlation with LVMi (r = 0.37, P < 0.05) and GWI, while GCW showed a weak negative correlation with RWT (r = − 0.26 and − 0.24, respectively; P < 0.05 for all). As for left ventricular filling pressure, only GWW and GWE showed a negative and positive correlation with E/A, respectively (r = − 0.34 and 0.27, respectively; P < 0.05 for both) (Supplementary Fig. 1.)

Multivariate analysis
We performed univariate and multivariate analysis of the potentially factors relevant to LV myocardial work. RWT showed an independent association with GWI as well as GCW after adjusting for sex, age, and LVEF (all P < 0.05, Tables 2 and 3). As for GWW, only SBP remained in the multivariate model after including all variables that showed significant association in univariate analysis (P < 0.05, Supplemental Table 1). Among all the echocardiography and structural variables, only RWT showed an independent association with GWE after adjusting for age, male sex, DBP, LVEF, ESV, and LVMi (P < 0.05, Table 4).

Intra-and inter-observer variability
There was good intra-and inter-observer consistency in the assessment of LV myocardial work parameters (Fig. 3).

Discussion
The present study evaluated the association between RWT and GMW in patients with HFpEF. Our main findings are summarized as follows: (i) compared with control subjects, patients with HFpEF showed impaired LV myocardial function, though LVEF was not significantly different. (ii) Patients with higher RWT had worse LV performance. (iii) RWT showed an independent association with GWE as well as GCW and GWI after adjusting for conventional clinical and echocardiography variables.
Several previous studies have demonstrated impaired LV myocardial function among HFpEF patients with reduced GWE and increased GWW, though their LVEF was normal [18,[26][27][28]. Our findings are consistent with those of these studies. We further divided the patients with HFpEF into two subgroups based on the median value of GWE in our cohort, as there is no standard reference range for GWE; the results showed that patients with higher GWE showed better LV myocardial function. Besides, GMW indices were related to LVEF. These results are consistent with those of previous studies [28,29]. As showed in Zhu's study [29], GMW indices were strongly correlated to patients with reduced LVEF while only mildly correlated to patients with normal LVEF. In our study, LVEF showed a relatively low correlation between LVEF and GMW as all the patients were with preserved EF. The discrepancy in correlation between two subgroups with different EF might due to GMW and EF reflect different aspects of LV function. LVEF represented left ventricular pump function while GMW indices are parameters about LV myocardial mechanical function.  Though LVEF and GLS are useful indices for the assessment of LV myocardial function [30], there are several limitations of these conventional parameters. According to Boe's study [31], GLS could be decreased when the load pressure increased. This indicates that GLS may be not reliable for assessment of myocardial function in HFpEF patients with different load condition. GMW, derived from PSL, appeared to be a relatively more appropriate and direct method for LV mechanical function assessment while LVEF and GLS are indirect methods. With consideration of GLS and pressure load, GMW indices provide an overall evaluation of LV myocardial function [26]. Besides, several studies have demonstrated the excellent correlation between PSL area and invasive measurements [16,31]. According to Laplace law, the left ventricle wall stress would be reduced in patients with increased wall thickness, which means that the pressure would be overestimated when we use blood pressure to replace left ventricle wall stress [32][33][34] and this would lead to overestimation of GWI, GCW, and GWW. However, GWE may not be greatly affected by the overestimated LV wall stress because GWE is the ratio of GCW and GCW plus GWW. It also worth noting that GWE has been proved to be associated with maximal Watts in stress test [18]. This indicates the subtle alteration of myocardial mechanics at rest may reflect the exercise capacity in patients with HFpEF. Thus, GWE could be a valuable and sensitive parameter to stratificate HFpEF patients.
Patients with HFpEF typically have several comorbidities that could induce cardiac remodelling and increase in inefficient cardiomyocytes, which further deteriorates the LV myocardial function. The presence of Coronary artery disease might be a vital factor of GWW [29,35] and CAD was detected in 28% patients with HFpEF in our study. The proportion of CAD, however, were comparable between subgroups by GWE and RWT. Heart failure patients with hypertrophy characterized by increased RWT tend to progress to systolic dysfunction on long-term follow up [36][37][38]. In our study, RWT showed an independent association with both GCW and GWE. Patients with increased RWT tend to have more severe concentric hypertrophy and inhomogeneity in cardiomyocytes [39]. The relationship between geometry and LV myocardial function has been demonstrated by Tan YT, et al. previously [40,41], i.e. LVH and inhomogeneity leads to incoordinate contraction and distorts the normal twisting motion. These dyssynchronous patterns could reduce GCW. Then as a result of delayed untwisting, the loss of suction further induced abnormal relaxation. As GWW mailly generated by the abnormal active relaxation, patients with cententric remodelling and hypertrophy would have increased GWW, which further reduced GWE.

Limitations
This was a single-center study with small sample size. Thus, our findings require further validation in a larger cohort. Although HFpEF patients were carefully identified, we cannot exclude the possibility that some patients may have been missed because we did not perform any invasive tests for the diagnosis of HFpEF. However, the diagnosis of HFpEF in our study was based on clinical assessment according to the latest guidelines.

Conclusion
In this study, patients with HFpEF showed impaired LV myocardial function when evaluated by global myocardial work. Impaired LV myocardial function may be associated with increased RWT.
Author contributions CQZ and LXZ conceived and designed the research; LMM, QYY, DXY, ZM and ZWW collected data and GWI global work index, GCW global constructive work, GWW global waste work, GWE Global work efficiency, obs observer, ICC intraclass correlation coefficient