Author + information
- Received November 18, 2010
- Revision received December 22, 2010
- Accepted January 3, 2011
- Published online April 1, 2011.
- Alex Pui-Wai Lee, MBChB⁎,
- Qing Zhang, PhD⁎,†,
- Gabriel Yip, MD⁎,
- Fang Fang, PhD⁎,
- Yu-Jia Liang, MM⁎,†,
- Jun-Min Xie, MM⁎,
- Yat-Yin Lam, MBChB⁎ and
- Cheuk-Man Yu, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Cheuk-Man Yu, Division of Cardiology, Department of Medicine and Therapeutics, Prince of Wales Hospital, Li Ka Shing Institute of Health Sciences, The Chinese University of Hong Kong, Shatin, Hong Kong 852, China
Objectives The aim of this study was to evaluate the role of left ventricular (LV) mechanical dyssynchrony in heart failure with preserved ejection fraction (HFPEF) complicating acute coronary syndrome (ACS).
Background In systolic heart failure, LV mechanical dyssynchrony worsens cardiac function and cardiac resynchronization therapy improves clinical outcome. The role of LV mechanical dyssynchrony in HFPEF complicating ACS is unknown.
Methods One hundred two patients presenting with ACS (ejection fraction ≥50%) and 104 healthy controls were studied using tissue Doppler imaging: group 1 (n = 55) had HFPEF on presentation and group 2 (n = 47) had no clinical HFPEF. The SD of time to peak systolic myocardial velocity and the SD of early diastolic (Te-SD) myocardial velocity of 12 LV segments were obtained for evaluation of dyssynchrony. Longitudinal mean myocardial ejection systolic velocity (mean Sm) and mean early diastolic velocity (mean Em) were measured.
Results Te-SD was greater in group 1 (33 ± 13 ms) than group 2 (21 ± 9 ms) (p < 0.001), and diastolic mechanical dyssynchrony was evident in 35% of patients in group 1 but in only 9% in group 2 (p < 0.001). Conversely, the SD of time to peak systolic myocardial velocity was similar in the 2 ACS groups (34 ± 16 ms vs. 32 ± 18 ms; p = NS), showing a similar prevalence of systolic mechanical dyssynchrony (47% vs. 43%; p = NS). Worsening of the diastolic dysfunction grade was associated with a parallel increase in Te-SD (grades 0, 1, 2, and 3: 16 ± 3 ms, 21 ± 5 ms, 28 ± 9 ms, and 41 ± 17 ms, respectively; p < 0.001). Te-SD correlated negatively with mean Em (r = −0.56, p < 0.001) and positively with peak mitral inflow velocity of the early rapid-filling wave/Em (r = 0.69, p < 0.001); mean myocardial ejection systolic velocity correlated significantly with mean Em (r = 0.56, p < 0.001), SD of time to peak systolic myocardial velocity (r = −0.42, p < 0.001) and Te-SD (r = −0.23, p = 0.001). Multivariate analysis identified peak mitral inflow velocity of the early rapid-filling wave/Em as the only variable independently associated with HFPEF (odds ratio: 1.48, p = 0.001). When peak mitral inflow velocity of the early rapid-filling wave/Em was excluded from the model, Te-SD (odds ratio: 1.13, p < 0.001) and mean Em (odds ratio: 0.37, p < 0.001) became independently associated with HFPEF.
Conclusions LV diastolic mechanical dyssynchrony may impair diastolic function and contribute to the pathophysiology of HFPEF, complicating ACS.
There is growing recognition that heart failure (HF) with preserved ejection fraction (HFPEF) is common and is associated with significant morbidity and mortality (1). Coronary artery disease, in addition to hypertension, diabetes, and aging, is a frequent risk factor for HFPEF (2). In the setting of acute coronary syndrome (ACS), HFPEF often complicates the acute clinical course. Although left ventricular (LV) performance can be impaired as a result of ischemia-induced systolic dysfunction, infarct-related increase in myocardial stiffness, and/or ischemic mitral regurgitation, the pathogenic mechanisms of acute ischemic HFPEF are not completely understood.
Mechanical dyssynchrony in LV contraction has been established as an important pathogenic mechanism in systolic HF. In the setting of acute ischemic HFPEF, acute myocardial ischemia leads to delayed onset and slower rate of contraction and relaxation in regional myocardial segments and thus may generate LV mechanical dyssynchrony (3,4), which may in turn compromises LV systolic and diastolic performance and leads to clinical HF. In the present study, we sought to elucidate the impact of LV mechanical dyssynchrony on ventricular function and its relationship with the occurrence of acute HFPEF in patients presenting with ACS.
One hundred two consecutive patients (age 64 ± 10 years, 76 men) who were admitted to a tertiary care hospital for ACS and had a normal ejection fraction (EF) (≥50%) were prospectively studied. All subjects presented to the emergency department with acute chest pain >30 min and had ischemic changes on the electrocardiogram (ST-segment depression or elevation ≥0.1 mV and/or T-wave inversion on at least 2 contiguous leads) and/or elevated serum cardiac biomarker (troponin T) within 6 h of presentation. Patients with a depressed EF (<50%), atrial fibrillation, pacemaker implantation, more than a mild degree of valvular dysfunction, a prosthetic valve, pericardial constriction, and myocardial rupture were excluded from the study. The diagnosis of HF was made clinically by the attending physicians during acute hospitalization and independently verified by a cardiologist based on documented symptoms of HF (acute onset or worsening of dyspnea), signs of fluid retention (elevated jugular venous pressure and dependent edema), in addition to radiological evidence of pulmonary vascular congestion. There were 55 patients who had HFPEF during hospitalization (group 1) and 47 patients in whom HFPEF did not develop during the hospital stay (group 2). Patients underwent cardiac catheterization and invasive revascularization if clinically indicated. One hundred four age- and sex-matched healthy subjects were studied as controls. They had no history of cardiovascular or systemic diseases, had normal findings on a physical examination and electrocardiogram, and had an echocardiogram showing no evidence of structural heart disease. The study was approved by the ethics committee of the institution, and written informed consent was obtained from all subjects.
Echocardiography (Vivid 7, Vingmed-General Electric, Horten, Norway) was performed within 72 h after hospital admission and before any coronary revascularization procedures were performed. Two-dimensional and Doppler echocardiography was performed in standard parasternal, apical, and subcostal views. Tissue Doppler imaging (TDI) was performed in apical 4-chamber, 2-chamber, and long-axis views for evaluation of LV longitudinal function. Color-coded TDI optimized for pulse repetition frequency, color saturation, and sector size and depth were obtained to maximize the frame rate to 100 Hz or higher. At least 3 consecutive beats in sinus rhythm were stored, and the images were analyzed offline using customized software (EchoPac-PC, version 7.0.0, Vingmed-General Electric). All measurements were averaged over at least 3 consecutive cardiac cycles. The echocardiographers who obtained the images and the investigators who performed the offline analysis were blinded to the clinical information of the subjects.
Evaluation of LV volumes and systolic and diastolic function
LV end-diastolic volume, end-systolic volume, and EF were assessed using the modified Simpson method in the apical 4- and 2-chamber views. Regional wall-motion abnormality was evaluated, and a wall motion score index (WMSI) was determined according to the recommendations by the American Society of Echocardiography (5). Longitudinal LV systolic function was assessed by averaging the peak myocardial systolic velocities at the 6 basal segments (Sm) obtained by offline TDI analysis with the sample volumes placed just above the mitral annulus (6).
To assess diastolic function, peak mitral inflow velocity of the early rapid-filling wave (E), peak velocity of the late filling wave due to atrial contraction (A), and deceleration time of early mitral inflow velocity (DT) of the early filling were recorded by using Doppler echocardiography. The longitudinal LV diastolic function was assessed by averaging the myocardial early diastolic velocities at the 6 basal segments (mean Em) at offline TDI analysis. Diastolic dysfunction was graded with reference to a classification scheme previously described (7). Normal diastolic function was defined as E/A = 0.9 to 1.5, DT = 160 to 240 ms, and E/Em <10; grade 1 (abnormal relaxation) if E/A <0.9 and DT >240 ms; grade 2 (pseudo-normal) if E/A = 0.9 to 1.5, DT = 160 to 240 ms, plus either E/Em ≥10 or E/A reversal on Valsalva maneuver; and grade 3 (restrictive filling) if E/A >2, DT <160 ms and E/Em ≥15. The diastolic function in 10 patients did not fall into any category and was classified as undetermined.
Evaluation of LV mechanical dyssynchrony
To assess LV mechanical dyssynchrony in both systole and diastole, myocardial velocity curves obtained from color-coded TDI were reconstituted offline using the 12-segment (6 basal, 6 mid) model that consisted of the anterior, inferior, anteroseptal, inferoseptal, anterolateral, and inferolateral segments at both basal and mid-levels of LV (8). The basal segments were sampled just above the mitral annulus, and the mid-segments were sampled at the mid-level of LV. The time to peak myocardial systolic velocity during the LV ejection period (Ts) and the time to peak myocardial early diastolic velocity (Te) during the early LV filling period were measured for each segment with reference to the onset of QRS complex. Continuous-wave Doppler imaging of the aortic and mitral flow was used to determine the timing of aortic and mitral valve opening/closure, respectively. Markers of valve opening and closing events would appear on the electrocardiographic recordings during offline TDI analysis to assist in accurate measurement of Ts and Te. The SD of Ts (Ts-SD) and of Te (Te-SD) of the 12 LV segments were calculated to measure systolic and diastolic mechanical dyssynchrony, respectively. The interobserver and intraobserver variability for Ts-SD and Te-SD was evaluated in 60 patients and was 4.7% and 3.2%, respectively. Using the upper 2 SDs of normal controls as a cutoff, systolic mechanical dyssynchrony was defined as Ts-SD >33 ms and diastolic mechanical dyssynchrony as Te-SD >34 ms, as previously reported (6,8). Post-systolic shortening, defined as myocardial contraction occurring after aortic valve closure (positive velocity greater than the peak ejection velocity), was distinguished from myocardial early diastolic velocity by their different timing and opposite directions. Post-systolic shortening, albeit considered to be a marker of ischemia, may not be a marker of mechanical dyssynchrony in ischemic cardiomyopathy (9). Therefore, in the present study, post-systolic shortening was not included in the assessment of LV mechanical dyssynchrony.
Data were analyzed using statistical software (SPSS for Windows, version 13.0, SPSS Inc., Chicago, Illinois). Results were presented as mean ± SD or number and percentage of patients. Comparisons among patient groups and among various grades of diastolic dysfunction were performed using 1-way analysis of variance with Scheffé test or Pearson chi-square test as appropriate. Pearson coefficient was used for correlation analysis. Univariate analysis was performed for all clinical and echocardiographic variables including age, sex, type of ACS, hypertension, diabetes, systolic and diastolic blood pressures, heart rate, QRS duration, LV end-diastolic volume, LV end-systolic volume, LV ejection fraction, mean Em, mean Sm, E/Em, E/A, DT, WMSI, Ts-SD, and Te-SD. Variables with p < 0.1 on univariate analysis were tested in the multivariate logistic regression with the forward stepwise method to identify independent associations with HFPEF. A value of p < 0.05 was considered significant.
The clinical and demographic characteristics of group 1 and group 2 ACS patients are shown in Table 1. Patients in group 1 were significantly older (p = 0.03) and had a higher prevalence of anterior myocardial infarction (p = 0.02) than those in group 2. The vast majority (96%) of patients in both groups had a QRS duration <120 ms. Other clinical characteristics and medications before admission were similar in the 2 groups.
LV volumes and systolic and diastolic function
Both groups of ACS patients had normal EF and similar WMSI (Table 2). Yet there were subtle abnormalities in systolic parameters in both ACS groups compared with controls; namely, ACS patients had a greater end-systolic volume (p < 0.05), lower EF (p < 0.001), higher WMSI (p < 0.001), and lower mean Sm (p < 0.001) than normal controls (Table 2). Of note, the extent of systolic abnormality was similar in group 1 and group 2 leading to the absence of intergroup differences. On the other hand, diastolic dysfunction was more severe in group 1, which had a higher prevalence of pseudo-normal and restrictive filling patterns (p < 0.001) and, in particular, a lower mean Em and a higher E/Em ratio by TDI (all p values <0.001 vs. group 2 or controls).
Systolic and diastolic mechanical dyssynchrony
Ts-SD was increased in both ACS groups compared with controls (p < 0.001), but it did not differentiate group 1 from group 2 (Table 2). The prevalence of systolic mechanical dyssynchrony (i.e., Ts-SD >33 ms) was similar in group 1 and group 2 (47% vs. 43%; p = NS) (Fig. 1). In contrast, Te-SD was significantly increased in group 1 (p < 0.001 vs. the other 2 groups) (Table 2). Diastolic mechanical dyssynchrony (i.e., Te-SD >34 ms) was evident in 19 of 55 patients (35%) in group 1 but in only 4 of 47 patients (9%) in group 2 (p < 0.001) (Fig. 1).
Combining the 2 ACS groups, comparisons of echocardiographic variables among various diastolic function grades are shown in Table 3. As expected, mean Em decreased, whereas E/Em increased progressively with escalating severity of diastolic dysfunction (p < 0.001 for all comparisons). Although conventional systolic parameters including EF and WMSI showed no significant differences among the 4 grades, mean Sm was relatively preserved in patients with grade 0 (normal) diastolic function (p < 0.001 vs. other grades). Nevertheless, there was no intergroup difference (p = NS) in mean Sm from grade 1 to grade 3 diastolic dysfunction.
Interestingly, Te-SD increased progressively from grade 1 to grade 3 diastolic dysfunction (p < 0.001) (Fig. 2). This was in contrast to Ts-SD, which showed no significant differences across diastolic function grades. There was a significant correlation between Te-SD and E/Em (r = 0.69, p < 0.001) (Fig. 3) and, inversely, with mean Em (r = −0.56, p < 0.001), but not between Ts-SD and E/Em (r = 0.08, p = NS) or mean Em (r = 0.08, p = NS). On the other hand, mean Sm correlated significantly with mean Em (r = 0.56, p < 0.001), Ts-SD (r = −0.42, p < 0.001), and Te-SD (r = −0.23, p = 0.001). Ts-SD and Te-SD correlated significantly with each other, albeit modestly (r = 0.16, p = 0.019). Examples of dyssynchrony TDI curves and mitral Doppler inflow patterns in controls and group 1 and group 2 ACS patients are shown in Figure 4.
Fifty-seven patients (30 in group 1 and 27 in group 2; p = NS) underwent cardiac catheterization and revascularization as clinically determined, and all of them had >70% luminal stenosis in at least 1 major epicardial coronary artery. To examine the relationship between coronary anatomy and mechanical dyssynchrony, a subset of patients with single-vessel disease were studied (n = 35). The distribution of the most delayed segments on echocardiography with regard to the coronary artery anatomy on angiography is shown in Figure 5. Interestingly, the segments with the most delayed contraction or relaxation (i.e., segments with the latest Ts or Te) in a patient did not always correspond anatomically to the coronary vascular territory on angiography. The most delayed contraction or relaxation was observed in the remote segments with normal coronary arteries in 25.7% and 28.5% of cases, respectively (Fig. 5).
Associating factors of HFPEF in ACS on multivariate analysis
Several clinical and echocardiographic variables were associated with HFPEF on univariate analysis including age, type of ACS (anterior ST-segment elevation myocardial infarction), E/A, DT, mean Em, E/Em, and Te-SD (p < 0.1) (Table 4). E/Em was the only variable independently associated with HFPEF on multivariate analysis (odds ratio: 1.48; 95% confidence interval: 1.17 to 1.88; p = 0.001). If E/Em was excluded from the model, Te-SD (odds ratio: 1.13; 95% confidence interval: 1.10 to 1.20; p < 0.001) and mean Em (odds ratio: 0.37; 95% confidence interval: 0.25 to 0.55; p < 0.001) were independently associated with HFPEF.
The present study demonstrated that patients with ACS complicated by acute HFPEF had significantly increased temporal dispersion in the regional timing of myocardial relaxation detectable by TDI. The results of this study show that diastolic mechanical dyssynchrony is closely linked to diastolic dysfunction grade and noninvasive estimates of filling pressure (E/Em), suggesting that LV diastolic dyssynchrony may be a contributing factor for acute ischemic HFPEF. Interestingly, such an association was not apparent for systolic mechanical dyssynchrony.
Role of LV mechanical dyssynchrony in the pathogenesis of HFPEF complicating ACS
HF is a frequent complication of ACS and is associated with poor prognosis (10,11). Many patients with ACS complicated by acute HF on presentation have relatively preserved LV systolic dysfunction. In the Global Registry of Acute Coronary Events, for instance, only 48.7% of the HF patients had depressed EF (11). However, the pathogenic mechanisms of HFPEF in these patients are not completely understood.
The presumed pathophysiological abnormality leading to HFPEF is diastolic dysfunction, a common finding in patients with coronary artery disease (12). Normal diastolic function is characterized by rapid decrease in LV pressure during the isovolumic and the early ventricular filling phases, as well as high compliance of LV walls during the late diastolic filling phase (13). The active process of myocardial relaxation that generates the rapid LV pressure decrease is normally homogeneous in all regional segments. Increase in the temporal heterogeneity in this process with some fibers lengthening later than the others may jeopardize normal ventricular filling (14). Further prolongation of early myocardial relaxation could delay diastolic LV minimum pressure well into late diastole and could therefore contribute to the elevation of LV filling pressure (15). We previously demonstrated that LV systolic and diastolic mechanical dyssynchrony were prevalent in patients with chronic coronary artery disease (16) and HFPEF (17,18). Wang et al. (19) observed that LV diastolic mechanical dyssynchrony had inverse correlations with the time constant of relaxation and pulmonary capillary wedge pressure in chronic hypertensive HFPEF. In the present study, parallel increase in diastolic mechanical dyssynchrony with increasing severity of diastolic dysfunction supported this hypothesis. E/Em was apparently the single most important associating factor of HFPEF on multivariate analysis, which is probably not surprising because any mechanisms that lead to acute HFPEF should eventually increase the LV filling pressure with resultant pulmonary edema. However, Te-SD and mean Em became independently associated with HFPEF after excluding E/Em from the regression model. These results, together with the close correlation between E/Em and Te-SD, support the hypothesis that diastolic mechanical dyssynchrony may contribute to HFPEF through elevation of the LV filling pressure.
As myocardial ischemia is often a regional phenomenon, it is possible that regional delay in relaxation leads to diastolic mechanical dyssynchrony during ACS. Among patients with chronic coronary artery disease and preserved LV systolic function, diastolic mechanical dyssynchrony has been shown to predict exercise-induced ischemia with resultant impairment of early ventricular filling (20). In the present study, the location of coronary stenosis and the site of regional mechanical delay were, in general, concordant with each other. However, a dissociation of coronary stenosis and mechanical delay did occur in some patients. Several possible explanations exist. First, myocardial perfusion is modified by physiological factors such as vasomotor tone and microvascular resistance, which are not well appreciated on coronary angiography. Microvascular ischemia may be present in diabetic and hypertensive patients, who were prevalent in our study population. On the other hand, nonischemic factors such as electrical delay may lead to mechanical dyssynchrony. Although a wide QRS complex of >120 ms was uncommon, electrical delay that was not manifested on surface electrocardiography could not be entirely excluded in our study population.
In this study, we observed that the mean Sm correlated significantly with mean Em, Ts-SD, and Te-SD. Longitudinal motion of the LV base toward LV apex during systole is largely a result of torsional deformation of the spirally arranged LV myofibers. The potential energy stored by LV torsion during the systolic phase is reinstituted during LV untwisting and aids in diastolic suction. An abnormal activation sequence of the ventricles (e.g., due to right ventricular pacing, regional ischemia, distortion of myofiber architecture, electrical conduction delay) can result in dyssynchronous LV contraction, with reduced torsion and longitudinal shortening (21). Recent studies pointed out that LV dyssynchrony was inversely related to LV torsion in advanced HF patients with prolonged QRS duration (22,23), underscoring that LV torsion (and associated longitudinal function) may be a parameter that reflects the extent of LV dyssynchrony. The relationships of mean Sm with mean Em, Ts-SD, and Te-SD in our study lent further support that systolic and diastolic mechanical events of the LV are closely coupled, and mechanistic links may exist between LV longitudinal systolic function and LV dyssynchrony in patients with ACS. Intriguingly, although systolic mechanical dyssynchrony was evident in ACS patients, it did not appear to be associated with an acute occurrence of HFPEF. We postulate that systolic mechanical dyssynchrony may be common in the setting of ACS but per se might not be enough to cause acute HFPEF when the global systolic function is relatively preserved and diastolic dysfunction is mild.
First, although this study demonstrated a strong association between LV mechanical dyssynchrony and HFPEF in the setting of ACS, a causal relationship cannot be directly assumed because of the cross-sectional design of the study. Second, there is a possibility of undetected transient systolic dysfunction, or mitral regurgitation, because echocardiography was performed within 72 h but not immediately after the onset of acute symptoms. Third, the findings of this study may not be extrapolated to patients with atrial fibrillation who were excluded owing to technical difficulty in evaluating dyssynchrony in this group of patients. Finally, clinical outcomes of ACS patients with or without HFPEF were not assessed due to the relatively small sample size and short follow-up period in this study. However, the clinical significance of HFPEF in ACS was established (9,10), and the present study aimed to provide insight into the pathophysiological mechanisms underlying this disorder.
The present study demonstrated that LV diastolic mechanical dyssynchrony is associated with global diastolic dysfunction and may be a predisposing factor of HFPEF in acute ischemic patients. This finding may provide a new therapeutic target for ischemic HFPEF. Whether LV diastolic mechanical dyssynchrony can be reversed or modified by medication, revascularization, or even cardiac resynchronization therapy warrants further investigations.
This study was supported by the Research Grant Council of Hong Kong (grant reference number: 477907). The authors have reported that they have no relationships to disclose.
- Abbreviations and acronyms
- peak velocity of the late filling wave due to atrial contraction
- acute coronary syndrome
- deceleration time
- peak mitral inflow velocity of the early rapid-filling wave
- ejection fraction
- mean early diastolic velocity of 6 basal left ventricular myocardial segments
- heart failure
- heart failure with preserved ejection fraction
- left ventricular
- peak systolic velocity of 6 basal left ventricular myocardial segments
- tissue Doppler imaging
- time to peak myocardial early diastolic velocity
- SD of the time to peak early diastolic velocity of 12 myocardial segments
- time to peak myocardial systolic velocity during the LV ejection period
- SD of the time to peak systolic velocity of 12 myocardial segments
- wall motion score index
- Received November 18, 2010.
- Revision received December 22, 2010.
- Accepted January 3, 2011.
- American College of Cardiology Foundation
- Abraham T.P.,
- Belohlavek M.,
- Thomson H.L.,
- et al.
- Pislaru C.,
- Belohlavek M.,
- Bae R.Y.,
- Abraham T.P.,
- Greenleaf J.F.,
- Seward J.B.
- Cerqueira M.D.,
- Weissman N.J.,
- Dilsizian V.,
- et al.
- Yu C.M.,
- Sanderson J.E.,
- Marwick T.H.,
- Oh J.K.
- Lester S.J.,
- Tajik A.J.,
- Nishimura R.A.,
- Oh J.K.,
- Khandheria B.K.,
- Seward J.B.
- Yu C.M.,
- Lin H.,
- Zhang Q.,
- Sanderson J.E.
- Yu C.M.,
- Fung J.W.,
- Zhang Q.,
- et al.
- Steg P.G.,
- Dabbous O.H.,
- Feldman L.J.,
- et al.
- Bonow R.O.,
- Bacharach S.L.,
- Green M.V.,
- et al.
- Brutsaert D.L.,
- Sys S.U.
- Yu C.M.,
- Zhang Q.,
- Yip G.W.K.,
- et al.
- Lee A.P.,
- Song J.K.,
- Yip G.W.,
- et al.
- Wang J.,
- Kurrelmeyer K.M.,
- Torre-Amione G.,
- Nagueh S.F.
- Delgado V.,
- Tops L.F.,
- Trines S.A.,
- et al.
- Bertini M.,
- Marsan N.A.,
- Delgado V.,
- et al.