Mechanical dispersion is associated with poor outcome in heart failure with a severely depressed left ventricular function and bundle branch blocks

Abstract

Objectives: Bundle branch blocks (BBB)-related mechanical dyssynchrony and dispersion may improve patient selection for device therapy, but their effect on the natural history of this patient population is unknown.

Methods: A total of 155 patients with LVEF ≤ 35% and BBB, not treated with device therapy, were included. Mechanical dyssynchrony was defined as the presence of either septal flash or apical rocking. Contraction duration was assessed as time interval from the electrocardiographic R-(Q-)wave to peak longitudinal strain in each of 17 left ventricular segments. Mechanical dispersion was defined as either the standard deviation of all time intervals (dispersion_SD) or as the difference between the longest and shortest time intervals (dispersion_delta). Patients were followed for cardiac mortality during a median period of 33 months.

Results: Mechanical dyssynchrony was not associated with survival. More pronounced mechanical dispersion_delta was found in patients with dyssynchrony than in those without. In the multivariate regression analysis, patients’ functional class, diabetes mellitus and dispersion_delta were independently associated with mortality.

Conclusions: Mechanical dispersion, but not dyssynchrony, was independently associated with mortality and it may be useful for risk stratification of patients with heart failure (HF) and BBB.

Key Messages

• Mechanical dispersion, measured by strain echocardiography, is associated with poor outcome in heart failure with a severely depressed left ventricular function and bundle branch blocks.

• Mechanical dispersion may be useful for risk stratification of patients with heart failure and bundle branch blocks.

Keywords:

• Mechanical dyssynchrony
• mechanical dispersion
• bundle branch block
• heart failure
• survival

Introduction

Patients with heart failure (HF) and a severely reduced left ventricular ejection fraction (LVEF ≤ 35%) often present with electrical dyssynchrony, defined as prolonged QRS duration ≥120 ms [1,2]. In this patient population, left bundle branch block (LBBB) is more common than right bundle branch block (RBBB) [3], but the mere presence of BBB is an independent predictor of high post-discharge mortality [4–6]. An unfavourable prognosis of HF patients with severely reduced LVEF and BBB who are refractory to optimal medical treatment can potentially be altered by cardiac resynchronization therapy (CRT), an implantable cardioverter-defibrillator (ICD) or both [7]. By inducing left ventricular (LV) reverse remodelling, CRT improves symptoms and reduce morbidity and mortality, while ICD prolongs life by preventing sudden cardiac death [7]. A severely reduced LVEF is the current guideline criterion for both ICD and CRT device implantation, whereas a QRS prolongation is additionally required for the latter [7]. However, these criteria appear to be suboptimal. Despite years of technological advances and experience with device therapy in HF, the non-response rate to CRT is still high, while the recent study reported that prophylactic ICD implantation in patients with non-ischemic HF was not associated with survival benefit [8,9]. In order to better define a mechanical substrate and thus improve patient selection for both CRT and ICD implantation, echocardiographic definitions of mechanical dyssynchrony and dispersion have been proposed [10–13]. Septal flash (SF) and apical rocking (ApRock) are mechanical signatures of LBBB that can be visually assessed on routine echocardiograms and their correction by CRT was associated with beneficial volumetric response and improved survival in previous studies [10–13]. However, their impact on the natural history of CRT candidates has not been previously investigated.

Left ventricular mechanical dispersion, measured as regional heterogeneity of myocardial contraction by speckle tracking echocardiography (STE) was predictive of ventricular arrhythmia in post-myocardial infarction (MI) patients who underwent an ICD implantation [14,15]. Unfortunately, patients with BBB, who may account for approximately one third of prophylactic ICD implantations, were excluded from original dispersion studies [14,15] and, therefore, the association of mechanical dispersion and survival in HF patients with BBB could not have been tested. It also seems unclear whether the dispersion of myocardial contraction in patients with BBB only reflects mechanical dyssynchrony or poses an additional risk for unfavourable outcome. In this retrospective study, we investigated the impact of mechanical dyssynchrony and dispersion on the natural history of HF in patients with severely reduced LVEF and BBB.

Methods

This was a retrospective, observational study. All patients consecutively admitted to the Department of Cardiology, University Clinical Hospital Centre Zemun, Belgrade, Serbia, between January 2011 and December 2014 with the diagnosis of worsening HF and severely reduced LVEF (≤35%) were considered for inclusion in the study. Patients’ flow through the study is shown in supplementary Figure S1. The study was approved by the Institutional Ethics Committee. After excluding patients without BBB, as well those with pacemakers, acute coronary syndrome, severe non-operated valvular lesions, significant comorbidities with a life expectancy less than 6 months and those who died during hospitalization, a total of 155 patients (133 with LBBB and 22 with RBBB) were enrolled in the study. Of note, all enrolled patients were fulfilling guideline criteria and were considered for CRT and ICD implantation, but eventually they did not undergo device implantation due to resource constrains. In brief, Serbia has one the highest rates of cardiovascular disease in Europe, but after several decades of political instability and low economic growth, the health care system was unable, until recently, to meet the demands for all effective but costly therapies, such as CRT/CRT-D. No other preselection criteria for device therapy were applied. The diagnosis of left and right BBB was made using conventional criteria, including QRS duration >120 ms and: QS or rS in lead V1, monophasic R wave with no Q wave in lead V6 and I (for LBBB) and RSR’ pattern in V1-2 with wide, slurred S wave in the leads I, aVL and V5-6 (for RBBB) [16]. The ECGs were reviewed by an investigator unaware of any other patient data. Demographic, clinical and laboratory data were generated as a part of routine patient care and were collected from medical records. An ischemic origin of HF was proven by coronary angiography or by a documented history of MI.

Patients were followed for cardiac mortality during a median period of 33 months [interquartile range: 13–44 months]. Data on cardiac mortality were acquired from medical records, by interview with the patients’ cardiologists or relatives and/or from national death registry. Seven patients were lost to follow-up.

Echocardiographic data

According to the hospital protocol, all patients underwent a comprehensive transthoracic echocardiography during hospitalization using commercially available scanner (Vivid 7 Pro, GE Vingmed Ultrasound, Horton, Norway). Three cardiac cycles (five in patients with atrial fibrillation) from each standard echocardiographic view were recorded and the exam was digitally stored in full DICOM format for off-line analysis using EchoPac workstations (version BT12; GE Healthcare, Milwaukee, WI). The LVEF was calculated using the modified biplane Simpson method. All conventional echocardiographic measurements, mechanical dyssynchrony assessment and 2D-STE analysis were performed by the investigator blinded to all other patient data.

Visual assessment of mechanical dyssynchrony

Mechanical dyssynchrony was visually assessed by evaluating ApRock and SF and was defined as a presence of SF and/or ApRock. This method for assessing dyssynchrony has been described and evaluated in the previous studies [11,13,18]. In brief, in the apical 4-chamber view, the presence of SF was identified visually as a short inward motion of the septum in early systole (within the QRS width), whereas ApRock is characterized by a short septal motion of the apex in early systole and a subsequent long motion to the lateral wall during ejection phase (Figure 1, Video 1).

Figure 1. Visual assessment of mechanical dyssynchrony. In patients with left bundle branch block, an early activation of ventricular septum (lightning bolt symbol in the left panel) results in a premature septal contraction (within QRS width) which can be visually appreciated as a short, inward motion (septal flash, yellow arrow in the middle panel). This early septal contraction also causes the apex to move septally. A delayed activation of the posterolateral wall moves the apex laterally while stretching the septum, which can be visually appreciated as apical rocking (Video 1). Modified from Stankovic et al. [11].

Two-dimensional speckle tracking analysis

Longitudinal strain was obtained from all apical views at >40 frames/s by manually tracing endocardial surface in each view, with a region of interest adjusted to include the entire myocardial thickness. After careful visual inspection of the adequacy of tracking, the region of interest was readjusted in all cases of unsatisfactory tracking. Patients with two or more segments with poor tracking even after manual correction were excluded from further analysis. Peak systolic longitudinal strain values were recorded for each segment and global longitudinal strain (GLS) was calculated as the average of all segmental values, using a 17-segment LV model.

Myocardial contraction duration was measured as the time from the first deflection of the QRS complex (onset of R- or Q-wave) to maximum myocardial shortening in each segment.

Mechanical dispersion (dispersion_SD) was defined as the standard deviation of time to peak negative strain intervals from the 17 LV segments [14,15,17]. Segments with only positive strain (dyskinetic segments) were not included in time measurements [15]. As an alternative measure of mechanical dispersion, the difference between the longest and shortest time interval from onset Q/onset R-wave to the maximum myocardial shortening (dispersion_delta) was calculated for all patients [14,15,17]. Independently and blindly to mechanical dispersion and visual dyssynchrony assessment, longitudinal strain curves in the apical 4-chamber view were also inspected for the presence of “classical” LBBB-induced strain pattern. It was defined as an early contraction of at least one basal or midventricular segment in the septal wall (with early peak contraction not exceeding 70% of the ejection phase) and early stretching in at least one basal or midventricular segment in the opposing lateral wall [18] (Figure 2(A,B)).

Figure 2. Dispersion of segmental longitudinal strain curves in the apical 4-chamber view with respect to the presence of mechanical dyssynchrony and myocardial scar. Each color-coded curve corresponds to one of six segments of septal and lateral walls, as indicated in the bottom of the figure. (A) Mechanical dyssynchrony in a patient with non-ischemic heart failure – classical pattern (Video 2A); early contraction of septal segments stretches lateral wall segments while the opposite can be seen in the late systole (“mirroring curves”); (B) Mechanical dyssynchrony in a patient with scarred mid (cyan) and apical (green) septal segments (Video 2B); basal septal segment shows typical dyssynchronous contraction pattern while the remaining two scarred segments are dyskinetic (positive curves) resulting in pronounced dispersion; (C) Dispersion of strain curves in a patient with no dyssynchrony and apical scar (Video 2C); (D) Dispersion of strain curves in a patient with no dyssynchrony and septal scar (Video 2D). AVC indicates aortic valve closure. Yellow arrows indicate positive strain curves of dyskinetic segments.

Statistical analysis

Depending on the distribution of the data, continuous data were expressed as mean with 95% confidence interval (95%CI) and compared between groups using unpaired t-test. Categorical data were summarized by proportions with 95%CI and compared using a Fisher’s exact test. For comparison of continuous variables among three groups, analysis of variance was performed, with post hoc pairwise comparisons between the groups, whereby adjustments to the significance level were made using the Tukey–Kramer method. Correlation between different measures of dispersion (dispersion_SD and dispersion_delta) and QRS width was described by Spearman correlation coefficient. Survival rates were assessed with Kaplan–Meier analysis, while differences in survival were compared between-groups by a log-rank test. Univariable and multivariable Cox regression analyses were performed to determine parameters associated with cardiac mortality. The multivariable regression model included variables with univariate p value<.05. The intra- and interobserver variability for QRS complex duration was tested in 30 randomly selected patients using interclass correlation coefficients (ICCs). All statistical tests were two-tailed, and a p value <.05 was considered significant. Statistical analysis was performed using commercially available software PASW Statistics 18 version 18 (SPSS, Inc., Chicago, IL).

Results

From 840 patients with HF and severely reduced LVEF initially considered for the study, BBB were present in 203 patients (24%). Characteristics of the study population are summarized in Table 1. Among patients with BBB, LBBB was more frequent than RBBB. Patients with RBBB were younger, but more often had atrial fibrillation and diabetes mellitus than those with LBBB. Echocardiographic examinations were incomplete or irretrievable in 19 patients. Echocardiographic data of the remaining 136 patients are shown in Table 2. Our experience with inter-and intraobserver variability for assessing mechanical dyssynchrony and dispersion has been reported previously [11,17,18]. ICCs for QRS width measurements were: 0.91 [95% confidence interval (CI) 0.80–0.96] for intraobserver variability and 0.83 [95% CI 0.67–0.91] for interobserver variability.

Table 1. Characteristics of the study population.

	All patients (n = 155)	Patients with LBBB (n = 133)	Patients with RBBB (n = 22)	Mean difference (95% CI)
Age, years	68 [66–70]	69 [67–70]	63 [56–67]	−6.69 [−11.29, −2.09]
QRS width, ms	155 [151–159]	155 [151–159]	154 [143–164]	−1.63 [−12.98, 9.71]
				Risk ratio [95% CI]
Atrial fibrillation	0.25 [0.19–0.33]	0.22 [0.16–0.30]	0.45 [0.27–0.65]	0.48 [0.28–0.85]
Female sex	0.35 [0.28–0.43]	0.36 [0.28–0.45]	0.27 [0.13–0.48]	0.88 [0.66–1.17]
Ischemic aetiology	0.38 [0.31–0.46]	0.38 [0.30–0.46]	0.41 [0.23–0.61]	0.92 [0.53–1.59]
NYHA functional class
II	0.33 [0.26–0.41]	0.35 [0.28–0.44]	0.18 [0.08–0.39]
III	0.59 [0.51–0.66]	0.57 [0.49–0.65]	0.68 [0.47–0.84]
IV	0.08 [0.05–0.14]	0.08 [0.04–0.13]	0.14 [0.05–0.33]
Hypertension	0.89 [0.83–0.93]	0.90 [0.84–0.94]	0.82 [0.61–0.93]	1.10 [0.90–1.35]
Diabetes mellitus	0.38 [0.31–0.46]	0.34 [0.26–0.42]	0.64 [0.43–0.80]	0.53 [0.36–0.79]
Impaired kidney function^a	0.32 [0.25–0.41]	0.32 [0.24–0.40]	0.36 [0.20–0.57]	0.87 [0.47–1.59]
COPD	0.14 [0.10–0.21]	0.15 [0.10–0.22]	0.10 [0.03–0.28]	1.65 [0.42–6.59]
ACEi/ARB	0.97 [0.94–0.99]	0.97 [0.93–0.99]	1.00 [0.85–1.00]	0.99 [0.92–1.06]
Beta-blocker	0.85 [0.78–0.89]	0.83 [0.76–0.89]	0.91 [0.72–0.97]	0.92 [0.79–1.07]
Aldosterone antagonists	0.83 [0.76–0.88]	0.82 [0.75–0.88]	0.86 [0.67–0.95]	0.95 [0.79–1.14]
Digoxin	0.22 [0.16–0.29]	0.22 [0.16–0.30]	0.23 [0.10–0.43]	0.96 [0.42–2.21]
Amiodarone	0.10 [0.06–0.16]	0.10 [0.06–0.16]	0.14 [0.05–0.33]	0.72 [0.22–2.31]

Data are summarized by mean with 95% confidence interval (95% CI) or by proportion with 95% CI.

^a Defined as an estimated glomerular filtration rate <60 mL/min/1.73 m² by the Modification of Diet in Renal Disease (MDRD) study equation.

ACEi/ARB: angiotensin-converting-enzyme inhibitor/angiotensin II receptor blocker; COPD: chronic obstructive pulmonary disease; LBBB: left bundle branch block; RBBB: right bundle branch block; NYHA: New York Heart Association.

Table 2. Echocardiographic data.

	All patients	Patients with LBBB	Patients with RBBB	t Value mean difference [95% CI]
Left atrium, mm (n = 152)^a	45 [44–47]	45 [44–47]	46 [43–48]	t = 0.130 0.21 (−3.04, 3.47)
Septum, mm (n = 152)	11 [10–11]	11 [10–11]	11 [10–12]	t = 0.729 0.39 (−0.67, 1.45)
Posterior wall, mm (n = 152)	11 [10–11]	11 [10–11]	11 [9–11]	t = −1.11 −0.60 (−1.68, 0.48)
LVEDD, mm (n = 152)	62 [60–64]	62 [61–64]	60 [56–64]	t= −1.08 −2.47 (−6.99, 2.06)
LVEF, % (n = 152)	20 [19–22]	20 [19–21]	23 [19–26]	t = 1.49 2.67 (−0.87, 6.19)
LVEDV, ml (n = 152)	188 [175–201]	188 [174–203]	187 [152–222]	t = −0.91 −1.68 (−38.06, 34.70)
LVESV, ml (n = 152)	146 [135–158]	146 [134–159]	146 [114–177]	t = 0.64 −1.06 (−34.01, 31.99)
GLS, % (n = 88)	−7.2 [−7.8, −6.5]	−7.2 [−7.9, −6.5]	−6.8 [−9.1, −4.7]	t = 0.312 0.33 (−1.79, 2.46)
DispersionSD, ms (n = 88)	97 [90–104]	100 [93–107]^b	73 [59–87]	t = −2.64 −27.44 (−48.0, −6.80)
Dispersiondelta, ms (n = 88)	292 [267–314]	299 [276–322]	242 [167–317]	t = −1.69 −56.62 (−123.31, 10.06)
				Risk ratio [95% CI]
Mechanical dyssynchrony	0.60 [0.52–0.68]	0.66 [0.58–0.74]^b	0.18 [0.07–0.41]	3.77 [1.34–10.59]

Data are summarized by mean with 95% confidence interval (95% CI) or by proportion with 95% CI.

^a Numbers in brackets indicate the number of patients in whom the respective analysis was possible;

^b Denotes p < .05 vs. right bundle branch block from the independent samples t-test or Fisher exact test.

GLS: global longitudinal strain; LV EDD: left ventricular end-diastolic diameter; LVEF: left ventricular ejection fraction; LBBB: left bundle branch block; RBBB: right bundle branch block.

Mechanical dyssynchrony

Mechanical dyssynchrony could be visually assessed in all patients with available echo data (100% feasibility). It was seen in 82 (60%) of patients – more often in patients with LBBB than in those with RBBB. It was also more prevalent in patients with non-ischemic origin of HF and among women. Patients with mechanical dyssynchrony had lower LVEF than those without, while GLS was similar between the groups (Table 3). The classical strain pattern was observed in all patients with visual mechanical dyssynchrony, and in none of those without visual dyssynchrony.

Table 3. Mechanical dyssynchrony and dispersion in relation to clinical and echocardiographic data.

	Mechanical dyssynchrony (n = 82)	No mechanical dyssynchrony (n = 54)	t Value mean difference [95% CI]
Age, years	68 [66–80]	68 [65–71]	t = 0.296 0.53 (−2.99, 4.05)
QRS width, ms	158 [153–161]	150 [143–157]	t = −1.91 −7.89 (−16.04, 0.27)
LVEF, %	19 [17–20]^a	23 [20–25]	t = 2.96 4.04 (1.34, 6.74)
GLS, %	−6.7 [−7.57, −5.82]	−7.8 [−7.79,−6.63]	t = −1.49 −1.02 (−2.37, 0.33)
DispersionSD, ms	107 [99–116]^a	86 [76–96]	t = −3.28 −21.3 (−34.2, −8.4)
Dispersiondelta, ms	320 [289–350]^a	262 [231–293]	t = −2.66 −57.8 (−100.8, −14.8)
			Risk ratio [95% CI]
Ischemic aetiology	0.29 [0.21–0.40]^a	0.54 [0.41–0.66]	0.55 [0.36–0.83]
Female sex	0.39 [0.29–0.50]^a	0.20 [0.12–0.33]	0.77 [0.62–0.95]
LBBB	0.96 [0.90–0.99]^a	0.74 [0.61–0.84]	1.30 [1.11–1.53]

Data are summarized by mean with 95% confidence interval (95% CI) or by proportion with 95% CI.

^a Denotes p < .05 from the t-test for independent samples or Fisher exact test;

GLS: global longitudinal strain; LBBB: left bundle branch block; LVEF: left ventricular ejection fraction.

Mechanical dispersion

Speckle tracking-derived measures of mechanical dispersion (dispersion_SD and dispersion_delta) were assessed in 88 patients. Of 136 patients with echocardiographic data, 39 patients with atrial fibrillation were excluded from mechanical dispersion analysis. From the remaining 97 patients, nine patients were further excluded due to low image quality or frame rates precluding reliable speckle tracking analysis (91% feasibility).

There was a strong positive correlation between dispersion_SD and dispersion_delta (Spearman’s rho = 0.84, p < .001). However, dispersion_SD correlated to QRS width (Spearman’s rho = 0.25, p = .018), while dispersion_delta did not (Spearman’s rho = 0.40, p = .709).

Mean values of both mechanical dispersion parameters were greater in patients with mechanical dyssynchrony than in those without (Table 3). Patients with LBBB had greater mechanical dispersion expressed as dispersion_SD than those with RBBB. In addition, patients with a history of MI had greater mechanical dispersion expressed as dispersion_delta (332 ms, 95% CI 283–380 ms vs. 277 ms, 95% CI 253–302 ms, mean difference −54.14, 95% CI −103.12 to −5.16, p = .038) than those without, while no such difference was observed for dispersion_SD (104 ms, 95% CI 89–120 ms vs. 94 ms, 95% CI 87–102 ms, mean difference −9.87, 95% CI −25.05 to 5.31, p = .199). Further, patients with no mechanical dyssynchrony and no previous MI had a lower mechanical dispersion_delta (235 ms, 95% CI 203–267 ms) than both, patients with either dyssynchrony or previous MI (309 ms, 95% CI 280–337 ms) and patients with mechanical dyssynchrony and previous MI (367 ms, 95% CI 280–453 ms; F (2, 85) = 8.02, eta-squared = 0.158, p < .001 for overall comparison from ANOVA) (Figure 3). Typical examples of dispersion of segmental longitudinal strain curves with respect to the presence of mechanical dyssynchrony and myocardial scar are shown in Figure 2 (for LBBB) and Figure 4 (for RBBB). A comparison of clinical, electroctrocardiographic and echocardiographic parameters according to gender is shown in Table 4.

Figure 3. The extent of mechanical dispersion with regard to mechanical dyssynchrony and a history of myocardial infarction (MI). Patients without mechanical dyssynchrony and previous MI had a lower mechanical dispersion_delta than both, patients with either dyssynchrony or MI and patients with mechanical dyssynchrony and previous MI. Error bars indicate one standard deviation, while asterisk indicates significant differences between groups.

Figure 4. Dispersion of segmental longitudinal strain curves in the apical 4-chamber view in patients with right bundle branch block. (A) In a patient with non-ischemic cardiomyopathy, despite visually synchronous left ventricular contraction, there is a significant dispersion of segmental longitudinal strain curves’ peaks. (B) In a patient with ischemic cardiomyopathy, left ventricular contraction appears synchronous on both visual inspection and strain imaging. Lower values of peak systolic longitudinal strain of the lateral wall segments are concordant with lateral wall akinesis. AVC indicates aortic valve closure.

Table 4. A comparison of clinical, electroctrocardiographic and echocardiographic parameters according to gender.

	Women	Men	t Value mean difference [95%CI]
Age, years (n = 54/101)^a	72 [70–75]^b	65 [63–67]	t = 4.22 6.99 (3.72, 10.27)
QRS width, ms (n = 54/101)	156 [149–163]	155 [150–159]	t = 0.35 1.44 (−6.66, 9.53)
GLS, % (n = 28/60)	−7.8 [−9.2, −6.6]	−6.8 [−7.6, −6.0]	t= −1.40 −1.03 (−2.50, 0.43)
LVEF, % (n = 52/100)	22 [20–24]	19 [18–21]	t = 1.94 2.56 (−0.05, 5.17)
NYHA class (n = 54/101)	2.7 [2.5–2.9]	2.8 [2.7–2.9]	t= −1.06 −0.12 (−0.31, 0.09)
DispersionSD, ms (n = 28/60)	105 [90–120]	93 [86–101]	t = 1.62 11.76 (−2.68, 26.19)
Dispersiondelta, ms (n = 28/60)	307 [262–353]	285 [260–310]	t = 0.93 22.47 (−25.41, 70.36)
			Risk ratio [95%CI]
Ischemic aetiology (n = 54/101)	0.41 [0.29–0.54]	0.37 [0.28–0.46]	0.90 [0.60–1.36]
LBBB (n = 54/101)	0.89 [0.78–0.95]	0.84 [0.76–0.90]	0.95 [0.83–1.07]
Mechanical dyssynchrony (n = 43/93)	0.74 [0.60–0.85]^b	0.54 [0.44–0.64]	0.72 [0.56–0.93]

Data are summarized by mean with 95% confidence interval (95%CI) or by proportion with 95%CI.

^a Numbers in brackets indicate the number of patients in whom the respective analysis was possible.

^b Denotes p < .05 vs. men from the independent samples t-test of Fisher exact test;

GLS: global longitudinal strain; LBBB: left bundle branch block; LVEF: left ventricular ejection fraction.

Clinical and echocardiographic parameters and cardiac mortality

During a median period of 33 months, 73 patients (43%) died. Table 5 shows univariate and multivariate Cox regression analyses to identify parameters associated with cardiac mortality. Age, NYHA class, impaired kidney function, diabetes mellitus, COPD, GLS and dispersion_delta were associated with cardiac mortality in the univariate analysis. Mechanical dyssynchrony and dispersion_SD were not associated with outcome. In the multivariate analysis, only NYHA class, diabetes mellitus and dispersion_delta were independently associated with mortality. Of note, when regression analysis was confined to patients with LBBB, dispersion_delta, along with NYHA class and impaired kidney function, remained independently associated with mortality (Table S1). A significant separation of fully adjusted survival curves for two cutoffs of dispersion_delta, provisionally set at the first (221 ms) and third quartile (347 ms) of the data set is shown in Figure 5.

Figure 5. Fully adjusted survival curves based on the Cox model shown in Table 5. More pronounced mechanical dispersion was associated with worse overall survival. Cutoff values of 221 ms (A) and 347 ms (B) represent the first and third quartiles (25th and 75th percentiles) of dispersion_delta data set.

Table 5. Univariate and multivariate regression analyses to identify parameters associated with all-cause mortality.

	Univariate		Multivariate
	HR [95% CI]	p Value	HR [95% CI]	p Value
Female vs. male	1.035 [0.634–1.690]	.890	–	–
Ischemic aetiology (yes/no)	1.450 [0.906–2.322]	.122	–	–
NYHA class (per 1 class)	2.142 [1.468–3.128]	<.001	2.121 [1.139–3.948]	.018
GLS (per 1%)	1.163 [1.048–1.290]	.004	1.102 [0.947–1.281]	.209
LVEF (per 1%)	0.973 [0.942–1.005]	.096	–	–
COPD (yes/no)	2.095 [1.195–3.674]	.010	1.420 [0.607–3.324]	.419
Impaired kidney function (yes/no)	2.141 [1.333–3.438]	.002	1.489 [0.676–3.280]	.323
Diabetes mellitus (yes/no)	1.757 [1.098–2.810]	.019	2.152 [1.028–4.502]	.042
Age (per 1 year)	1.035 [1.011–1.061]	.005	1.017 [0.974–1.063]	.438
ACEi/ARB (yes/no)	0.591 [0.145–2.420]	.465	–	–
Beta-blocker (yes/no)	0.625 [0.343–1.147]	.130	–	–
Amiodarone (yes/no)	1.082 [0.518–2.262]	.833	–	–
Mechanical dyssynchrony (yes/no)	1.462 [0.866–2.469]	.155	–	–
DispersionSD (per 1 ms)	1.003 [0.994–1.013]	.496	–	–
Dispersiondelta (per 1 ms)	1.005 [1.002–1.008]	.002	1.004 [1.000–1.008]	.043
LBBB vs. RBBB	1.332 [0.660–2.686]	.423	–	–
QRS width (per 1 ms)	0.997 [0.987–1.007]	.533	–	–
Haemoglobin (per 1 g/L)	0.989 [0.974–1.004]	.167	–	–
Hypertension (yes/no)	2.108 [0.767–5.778]	.148	–	–
Atrial fibrillation (yes/no)	1.266 [0.752–2.131]	.374	–	–

ACEi/ARB: angiotensin converting enzyme inhibitors/angiotensin receptor blockers; CI: confidence interval; COPD: chronic obstructive lung disease; HR: hazard ratio; GLS: global longitudinal strain; LVEF: left ventricular ejection fraction; LBBB: left bundle branch block; NYHA: New York Heart Association; RBBB: right bundle branch block.

Discussion

In this study, we investigated the effects of mechanical dyssynchrony and dispersion on the natural history of HF patients with severely reduced HF and BBB. Mechanical dispersion, expressed as the difference between the longest and shortest time interval from ECG onset Q/onset R-wave to the maximum myocardial shortening, was associated with mortality, while the presence of mechanical dyssynchrony was not. Further, mechanical dispersion remained independently associated with mortality after being adjusted for LVEF, GLS and clinical parameters known to adversely affect survival of patients with HF.

Expectedly, mechanical dyssynchrony was more frequently seen in patients with LBBB, non-ischemic cardiomyopathy and among women, and at the rate comparable to that of the PREDICT-CRT study which used the same definition of dyssynchrony [11]. Although SF and ApRock represent mechanical consequences of dyssynchronous contraction induced by LBBB, these dyssynchrony markers were also seen in 18% of patients with RBBB confirming previous notions that LBBB may masquerade as RBBB [20,21].

Our observation that the presence of SF or ApRock is not associated with unfavourable prognosis concur with results of the PREDICT-CRT study demonstrating that patients with no mechanical dyssynchrony had similar long-term survival as those with uncorrected dyssynchrony despite CRT [11]. Therefore, it appears that SF and ApRock per se are not the heralds of a poorer survival, but rather a potential mechanical substrate for favourable response to CRT.

The differences and similarities between mechanical dyssynchrony and dispersion

In this study, patients with mechanical dyssynchrony showed a greater extent of mechanical dispersion than those without. This can be expected since SF and ApRock are visual surrogate markers of temporal inhomogeneities within the left ventricle in the presence of LBBB. As shown in Figure 2(A), a premature septal contraction in early systole stretches the opposite posterolateral segments while the delayed contraction of the posterolateral wall stretches the septum during the ejection phase. This classical pattern of dispersion (“mirroring strain curves”) represents a typical opposing wall mechanics induced by “true” LBBB, and it was observed in all patients with visual mechanical dyssynchrony [19,22]. It is correctable by CRT and has been linked to beneficial volumetric response and improved long-term survival after CRT [11–13,19,22–24]. On the other hand, in patients with myocardial scar and no visual dyssynchrony (Figure 2(C,D)), there is dissociation between dispersion and dyssynchrony – mechanical dispersion describes pure functional inhomogeneities (i.e. passive motion due to scar). This pattern of dispersion is not correctable by CRT, but it may represent a substrate for ventricular arrhythmias [14,15,25]. Finally, when myocardial scar and dyssynchrony coincide, mechanical dispersion most likely reflects both active motion due to LBBB and passive motion due to scar [26] (Figure 2(B,C)). In line with this, in this study, patients with no mechanical dyssynchrony and no previous MI showed a lower mechanical dispersion than patients with either dyssynchrony or previous MI, as well as patients with both dyssynchrony and previous MI (Figure 3).

Our data are in line with previous studies demonstrating that LV contraction patterns are related to different myocardial electrical activation patterns [20,27,28]. In a study by Risum et al. [27], there was a close association between LBBB defined by strict criteria proposed by Strauss et al. [29] and LBBB contraction pattern by speckle-tracking strain. In addition, Leeters et al. [28] reported that wall motion abnormalities between inferior and anterior LV walls in patients with RBBB and left anterior fascicular block are similar to abnormalities found between septal and lateral LV walls in patients with LBBB. However, it should be noted that prognostic implications of QRS morphology can also be gender-specific, as shown by Zusterzeel et al. [30]. Among patients with LBBB, women had a 21% lower mortality risk than men while there were no sex differences in non-LBBB group [30]. In this study, there was a higher prevalence of mechanical dyssynchrony among women, and it might have translated into more favourable survival, if patients had been treated by CRT. However, since patients in this study did not receive CRT, it was not surprising that the gender was not associated with outcome.

Range versus standard deviation for defining dispersion in bundle branch block

To the best of our knowledge, the effect of mechanical dispersion on the natural history of HF patients with low LVEF and BBB has not so far been investigated, mainly because device therapy has become their standard treatment. The relationship between mechanical dispersion and arrhythmic events has been studied in patients with ICDs, but since the presence of LBBB-induced mechanical dyssynchrony may interfere with the extent of mechanical dispersion, patients with LBBB were largely excluded from those trials. Of note, baseline dispersion_SD (which was referred to as mechanical dyssynchrony) did not predict ventricular arrhythmias or death in mild HF patients with LBBB treated with device therapy [31]; however, CRT-induced improvement of dispersion_SD was associated with significant reduction of arrhythmias in two trials involving patients with LBBB [31,32]. Methods for quantifying mechanical dispersion used in the present and previous studies were based on two absolute measures of dispersion: range (dispersion_delta) and standard deviation (dispersion_SD). In previous trials excluding patients with LBBB, both measures of mechanical dispersion showed similar accuracy of predicting arrhythmic death or coronary artery disease [14,15,17]. Although a strong positive correlation between the two measures of dispersion was also observed in this study, only dispersion_delta was independently associated with overall survival even in the presence of BBB. Given the fact that mechanical dispersion in patients with BBB may represent a combination of temporal and functional heterogeneities, it is possible that the range (dispersion_delta), defined as the difference of two extreme values, may represent a more robust measure of mechanical dispersion than standard deviation (dispersion_SD) which is less affected by extreme values. In line with this, in the current study, dispersion_SD correlated to QRS width, while dispersion_delta did not. On the other hand, albeit statistically significant, a correlation between dispersion_SD and QRS width is clinically meaningless. This further support the hypothesis that echocardiographic measures of dispersion convey the information that is independent of and possibly superior to QRS duration.

Clinical implications

While mechanical dispersion assessed by STE might be a promising tool for improving patient selection for ICD implantation, it has not been fully validated in patients with BBB. Our data suggest that mechanical dispersion_delta may be useful risk stratifying HF patients with BBB. Of note, although the terms mechanical dyssynchrony and dispersion are often used interchangeably, our data concur with previous reports [33] implying that mechanical dyssynchrony appears to be a pattern of mechanical discoordination specific to LBBB carrying different prognostic implications.

Study limitations

Our study is limited by its retrospective, observational and single centre design. A large proportion of patients had to be excluded from dispersion analysis due to atrial fibrillation which is frequently encountered in patients with advanced HF. Further, to ensure high quality data, patients with insufficient image quality and low frame rates were excluded resulting in a relatively small sample size. In addition, we refrained from proposing a cutoff for mechanical dispersion and tested only two provisional cutoff values for illustrative purposes. Since all patients in this study were fulfilling criteria for CRT-P or CRT-D implantation, proposing a cutoff value associated with natural history would be of limited clinical value. Finally, despite being promising in the research arena, the strain-based echocardiographic parameters of mechanical dispersion and dyssynchrony are yet to be validated in a large-scale randomized study, since the previous prospective studies failed to show acceptable accuracy and reproducibility [34,35].

Conclusions

Mechanical dispersion is associated with survival of patients with HF and BBB, independently of left ventricular ejection fraction and GLS. Mechanical dyssynchrony was not associated with survival, although more pronounced mechanical dispersion was observed in patients with dyssynchrony.

Disclosure statement

The authors declare that there is no conflict of interest.

Additional information

Funding

This research received no grant from any funding agency in the public, commercial or not-for-profit sectors. Biljana Putnikovic and Aleksandar N. Neskovic received funding from the Ministry of Education, Science and Technological Development of the Republic of Serbia (grant No 175099).