Late Systolic Myocardial Loading is Associated with Left Atrial Dysfunction in Hypertension

Abstract

Background

Late systolic load has been shown to cause diastolic dysfunction in animal models. Whereas the systolic loading sequence of the ventricular myocardium likely affects its coupling with the left atrium (LA), this issue has not been investigated in humans. We aimed to assess the relationship between the myocardial loading sequence and LA function in human hypertension.

Methods and Results

We studied 260 subjects with hypertension and 19 normotensive age- and sex-matched controls. Time-resolved central pressure and LV geometry were measured with carotid tonometry and cardiac MRI, respectively, for computation of time-resolved ejection-phase myocardial wall stress (MWS). The ratio of late/early ejection-phase MWS time-integrals was computed as an index of late systolic myocardial load. Atrial mechanics were measured with cine-SSFP MRI using feature-tracking algorithms. Compared to normotensive controls, hypertensive participants demonstrated increased late/early ejection-phase MWS and reduced LA function. Greater levels of late/early ejection-phase MWS were associated with reduced LA conduit, reservoir and booster pump LA function. In models that included early and late ejection-phase MWS as independent correlates of LA function, late-systolic MWS was associated with lower, whereas early-systolic MWS was associated with greater LA function, indicating an effect of the relative loading sequence (late vs. early MWS) on LA function. These relationships persisted after adjustment for multiple potential confounders.

Conclusions

A myocardial loading sequence characterized by prominent late systolic MWS was independently associated with atrial dysfunction. In the context of available experimental data, our findings support the deleterious effects on late systolic loading on ventricular-atrial coupling.

Afterload is recognized as an important determinant of myocardial function. Animal studies have demonstrated that within physiologic ranges, increased afterload results in differential responses in relaxation depending on its timing. Increases in early systolic load resulted in unchanged or slightly enhanced relaxation, whereas increases in late systolic load resulted in a slow rate of diastolic ventricular pressure fall.^1,2 More recently, greater late-systolic central pressure^3,4 was reported to be associated with impaired early diastolic relaxation in humans. These studies implicate the loading sequence as a potential mechanistic determinant of diastolic function.

Afterload ultimately affects the time-varying mechanical load (stress) on the myocardium, which is determined by complex interactions between myocardial contractile elements, instantaneous LV geometry and the time-varying hydraulic load imposed by the arterial tree.⁵ All key determinants of myocardial stress (wall thickness, cavity size and LV pressure, which in turn depends on arterial properties for any given flow delivered against the input impedance of the systemic circulation) exhibit marked variations during systole. Therefore, although aortic (and LV) pressure during systole is related to developed myocardial wall stress (MWS), LV geometric changes that occur during systole determine a profound but highly variable change in the relationship between LV/aortic pressure and wall stress.⁵ Consequently, time-varying MWS estimations need to account for both time-varying pressure and time-varying geometry and are poorly represented by any single time-point measurement. Recently, a relationship between time-varying MWS and abnormal diastolic relaxation has been demonstrated, suggesting that the systolic myocardial loading sequence is a key determinant of systolic-diastolic coupling.^6–8 However, its impact on left atrial (LA) structure and function is unknown.

LA dysfunction contributes to a number of complications in hypertension. LA function is important to regulate and promote LV filling during diastole, contributing to cardiac output adaptations to changes in loading conditions, inotropic stimulation and heart rate. With early/mild hypertension, conduit function of LA decreases, whereas booster function increases.⁹ However, with progression of hypertensive heart disease, booster function also declines and is associated with heart failure^10,11 and a poor prognosis.¹² Given the clinical relevance of LA dysfunction in hypertension, a better understanding of its determinants is highly desirable.

In this study, we aimed to assess whether the myocardial loading sequence (i.e., early versus late ejection-phase MWS) is associated with reduced LA function, as assessed by MRI-based measurements of atrial deformation and phasic volumes, among adults with hypertension.

Methods

Study population

We enrolled a convenience clinical sample of 260 subjects with a history of hypertension referred for a cardiac MRI at the Corporal Michael J. Crescenz VA Medical Center. We also studied a group of age- and gender-matched control subjects without a history of hypertension, diabetes mellitus, cardiomyopathy, or vasoactive medication use (n=19). The protocol was approved by the Philadelphia VA Medical Center Institutional Review Board, and written informed consent was obtained from all participants.

Key exclusion criteria were as follows: (1) Claustrophobia; (2) Presence of metallic objects or implanted medical devices in body; (3) Atrial fibrillation, flutter or significant arrhythmia at the time of enrollment, which may compromise the study measurements; (4) Other conditions that would make the study measurements less accurate or unreliable (i.e., inability to perform an adequate breath hold for cardiac MRI acquisitions); (5) Any degree of aortic stenosis, which increases LV pressure relative to carotid pressure, and therefore confounds assessment of LV MWS.

CMR Imaging Protocol

Participants underwent a cardiac-MRI examination to assess LV structure and function, using a 1.5 Tesla (T) whole body MRI scanner (Avanto or Espree, Siemens, Malvern, PA) equipped with a phase-array cardiac coil. LV volumes and ejection fraction were determined using balanced steady-state free-precession (SSFP) cine imaging. Typical parameters were as follows: TR=30.6 ms; TE=1.3 ms; Phases=30; Slice thickness=8 mm; Matrix size=192×192; Parallel image (IPAT) factor=2. LV short-axis stack cine images were analyzed along with at least one long-axis view to assess time-resolved LV endocardial and epicardial contours in each cardiac phase, using CMR42 software (Circle CVI, Calgary, AB, Canada). LV wall volume was computed as the difference between epicardial and endocardial volumes at each cardiac phase. LV mass (LVM) was computed as the difference between epicardial and endocardial volumes, multiplied by myocardial density. LVM was normalized for body height in meters raised to the allometric power of 1.7.¹³ We measured proximal aortic flow using velocity-encoded (phase-contrast) imaging in a short-axis proximal aortic plane prescribed at or below the level of the right pulmonary artery. We acquired aortic flow during free-breathing using the following parameters: TR=10 ms; TE=3.2 ms; Flip angle=30; FOV=340×340; matrix size=256×256; Slice thickness=8 mm; gating=retrospective; VENC=130 cm/sec (prescribed ad hoc to avoid aliasing); number of phases maximized according to heart rate. Aortic through-plane contrast images were processed with the freely available software Segment.¹⁴

LA longitudinal strain and volumetric analysis

We used feature-tracking techniques for the measurement of LA phasic strain, as previously described.¹⁵ LA analyses were performed using cvi42 image analysis software (Circle Cardiovascular Imaging Inc., Calgary, Canada). LA endocardial borders were manually traced in apical 2- and 4-chamber views using LV end-diastole as the point of reference. An automated tracking algorithm was applied, and the tracking of all atrial segments was confirmed. Manual adjustments were performed as needed to optimize wall tracking. An example of atrial wall tracking is shown in Figure 1 and in online Videos 1 and 2.

Figure 1. Representative example of measures of atrial deformation derived from atrial tissue tracking from cine SSFP-MRI images.

Panel A shows the diastatic LA phase, in which reference points are prescribed. Panel B shows the tracking of atrial tissue, shown at a different phase of the cardiac cycle (see online Video 1). Panels C and D show representative plots of strain and strain rate, respectively, along with measures of reservoir (R), conduit (C) and booster pump (P) function derived from strain and strain rate curves. Note that atrial diastasis was used as the reference length for all strain/strain rate measurements. SR-E = early diastolic (conduit) atrial strain rate. SR-A=late diastolic (booster pump) atrial strain rate.

Values of segmental deformation were exported and further processed using custom software programmed in Python (Python Software Foundation, Wilmington, Delaware, USA). We computed longitudinal atrial strain, defined as the change of atrial myocardial length throughout the atrial cycles (L₁) compared to its resting (or reference) length (L₀) in a relaxed state at diastasis (end of atrial diastole). Strain was computed as (L₁-L₀)/L₀. We computed strain (deformation) and strain rate (SR) relative to the diastatic LA length (rather than end-diastolic length), since diastasis represents the length for LA tissue at the end of atrial diastole (reference phase). As shown in Figure 1, strain and SR were calculated to assess reservoir (total longitudinal strain), conduit (positive longitudinal strain and early-diastolic SR) and booster (negative longitudinal strain and late-diastolic SR) LA function. Maximum (LA_MAX), minimum (LA_MIN), and diastatic (LA_DIAS) LA volumes were also measured. LA expansion index, passive LA emptying fraction and active LA emptying fraction were calculated as volumetric measures of reservoir, conduit and booster phases, respectively. LA expansion index was calculated as (LA_MAX-LA_MIN)/LA_MIN, passive LA emptying fraction as (LA_MAX-LA_DIAS)/LA_MAX, and active LA emptying fraction as (LA_DIAS-LA_MIN)/LA_MAX.

Carotid tonometry

Carotid artery applanation tonometry was performed immediately before or after the cardiac MRI using a SphygmoCor-Px device (AtCor Medical), which incorporates a high-fidelity Millar applanation tonometer (Millar Instruments). Carotid pressure waveforms were calibrated according to brachial mean and diastolic pressure (measured with a Hewlett-Packard 78352c device; Hewlett-Packard, Palo Alto, CA), given that diastolic and mean blood pressures do not vary substantially along the arterial tree. The carotid waveform was used because it is a direct surrogate of the aortic pressure waveform; this approach therefore does not require the use of a generalized transfer function. We used the systolic portion of the carotid pressure waveform to assess LV pressure during ejection (Figure 2).

Figure 2.

Example of carotid pressure waveform obtained via arterial tonometry (top panels) and time-resolved LV geometry assessed with cardiac MRI (bottom panels), which is used to compute time-resolved ejection-phase MWS (right-sided panel). “Early” ejection-phase wall stress was quantified as the stress-time integral (area under the curve) during the first half of ejection (blue area), whereas “late” ejection-phase myocardial wall stress was quantified as the stress-time integral during the second half of ejection (red area). Note that the time-axis in the MWS plot corresponds only to the ejection phase, whereas the carotid pressure waveform corresponds to an entire cardiac cycle.

Computation of ejection-phase MWS

The computation of ejection-phase MWS is schematized in Figure 2. We applied the formula developed by Arts et al,^16,17 which is applicable to axisymmetric ventricles for computation of average LV MWS. This method does not neglect radially-directed forces or forces generated within the wall that oppose fiber shortening, which vary significantly with cavity and wall thickness and can interfere with direct comparisons of myocardial fiber stress at different times during ejection. The formula is based on LV cavity volume (V_LV), LV wall volume (V_W) and pressure:

$$\mathit{Fiber}\sigma =\frac{P}{\frac{1}{3}ln(1+\frac{V_W}{V_{LV}})}$$

where P=pressure, ln=natural logarithm, V_W=wall volume and V_LV=ventricular cavity volume (computed at each time point).

We computed time-varying LV cavity volume based on the integration of the proximal aortic systolic flow waveform. The time-integral of the systolic flow waveform represents the cumulative ejected volume at each time point. The systolic flow time-integral was therefore calibrated to stroke volume (computed as end-diastolic minus end-systolic LV cavity volume measured with SSFP MRI), and in turn was used to compute the cumulative ejected volume from the onset of ejection to each time point during ejection. The latter value was subtracted from end-diastolic LV cavity volume to compute the time-resolved LV cavity volume during ejection. LV wall volume was assumed to be constant throughout ejection, as previously demonstrated.¹⁸

To assess early vs. late systolic MWS, we computed the time-integrals of MWS in the first and second half of ejection, respectively, as previously described (Figure 2).⁶

In 5 repeated measures performed on the same day, coefficients of variation were as follows: LA expansion index: 10.9%; total longitudinal strain 11.4%; LA passive EF: 12.5%; conduit longitudinal strain: 9.7%; early diastolic longitudinal strain rate 15.6%; LA active EF: 5.9%; late diastolic longitudinal strain 25.5%; late diastolic longitudinal strain rate: 33.7. All measures of MWS (Peak MWS, Time to peak MWS, End-systolic Stress, Early ejection-phase MWS, Late ejection-phase MWS and the ratio of late/early ejection-phase MWS demonstrated CVs<5%, whereas the time to peak fiber stress demonstrated a CV of 7.3%.

Statistical Methods

Continuous variables are presented as mean±SD and compared using t-tests of analysis of variance, as appropriate. Categorical variables are presented as frequencies and percentages and compared with the chi-square or Fisher’s exact text, as appropriate. Pearson correlations coefficients were computed to examine the relationship various measures of LA structure and function. Linear regression was performed to assess the relationship between early systolic MWS, late systolic MWS (both included as independent variables), and various parameters of LA function (included as dependent variables in individual regression models) in hypertensive participants. For each parameters of LA function, we built unadjusted models (referred to as “Model 1”), models that adjusted for age, ethnicity and sex (“Model 2”) and models that additionally adjusted for left atrial volume, body mass index, systolic blood pressure, diastolic blood pressure, diabetes mellitus, coronary artery disease and heart failure status (“Model 3”). For easier comparisons of the magnitude of the relationships of early and late systolic wall stress in various models, we present standardized regression coefficients (β). For a more intuitive representation of the relationship between the loading sequence and LA function, we computed the ratio of late/early MWS-time integrals. Hypertensive subjects were then stratified according to tertiles of the eary/late MWS-time-integral ratio, and indices of LA function were compared between the tertiles using one-way analysis of variance (ANOVA), with post-hoc pairwise comparisons with Bonferroni correction. Statistical significance was defined as a 2-tailed p value<0.05. All probability values presented are 2-tailed. Statistical analyses were performed using SPSS software (SPSS v24 for Mac; IBM SPSS version 24, Chicago IL).

Results

Baseline characteristics of study participants are presented in Table 1. As expected, compared to the control group, hypertensive subjects demonstrated greater systolic, diastolic and mean arterial pressure and body mass index (Table 1). Hypertensives subjects also exhibited greater LV mass, LV mass index, prevalence of LV hypertrophy¹⁹, and LA volume (but not LA volume index). There were no significant differences in peak MWS, time to MWS, end-systolic MWS, or the absolute values of early (1st half or ejection) or late (2nd half of ejection) MWS-time-integrals. However, the ratio of late/early ejection-phased MWS-time integrals was significantly higher in the hypertensive group (Table 2), indicating higher late systolic myocardial load relative to early systolic myocardial load in hypertensive subjects. Relative to normotensive controls, hypertensive subjects demonstrated significantly reduced conduit and reservoir LA function (lower LA expansion index, LA passive ejection fraction, conduit longitudinal strain, and less negative early diastolic longitudinal strain rate), without significant differences in measures of LA booster pump function (Table 2). Correlations between LA volume and various measures of reservoir, conduit and booster pump function are described in the supplemental section (Supplemental Table 1).

Table 1.

Baseline Characteristics of Study Participants

Characteristics	Hypertensives (n=260)	Non-hypertensive controls (N=19)	P value
Age, years	62.7±9.6	59.6±7.7	0.65
Male gender	240 (92.3)	17 (89.5)	0.45
Ethnicity
Caucasian	114 (43.8)	10 (52.6)	<0.001
African-American	139 (53.5)	6 (31.6)
Other	7 (2.69)	3 (15.8)
Body mass index, kg/m2	30.7±6.2	27±5.2	0.014
Body surface area, m2	2.18±0.27	1.98±0.21	0.002
Systolic blood pressure, mmHg	144.4±19.6	129.3±14.8	0.001
Diastolic blood pressure, mmHg	83.5±12.1	76.8±9.4	0.019
Mean blood pressure, mmHg	113.3±16.5	103±13	0.008
Diabetes Mellitus	135 (51.9)	*
Insulin Use	59 (22.7)	*
eGFR, mL/min/1.73 m2	80.9±28.5	92.5±18.1	0.082
Beta Blocker Use	165 (63.5)	*
ACE inhibitor use	153 (58.8)	*
ARB use	40 (15.4)	*
Calcium channel blocker use	80 (30.8)	*
Thiazide use	79 (30.4)	*
Statin use	193 (74.2)	9 (47.4)	<0.001
Spironolactone use	14 (5.4)	*

Values are mean±SD or counts (percentages). eGFR=estimated glomerular filtration rate; ACE=Angiotensin convertase enzyme; ARB=Angiotensin receptor blocker.

^*Excluded by design.

Table 2.

Left ventricular structure, MWS and left atrial structure and functional variables in hypertensive vs. non-hypertensive control subjects

Characteristics	Hypertensives (n=260)	Non-hypertensive controls (N=19)	P value
LV end-diastolic volume, mL	174.2±62.1	148.2±33.2	0.07
LV ejection fraction, %	52±14.3	57/3±11.7	0.12
LV mass, g	167.7±50.1	125.7±29	<0.001
LV mass index to height, g/m1.7	63.5±19	49.5±11.7	0.002
LV mass index to BSA, g/m2	76.9±22.2	64.0±17.5	0.017
LA volume, ml	77.7±37.9	66.3±15	0.011
Left atrial volume index, ml/m2	35.8±18.7	34±9.4	0.69
LV hypertrophy*	77 (29.7)	1 (5.3)	0.022
Peak MWS, kdynes/cm2	724.9±163.5	708.5±124.1	0.67
Time to peak MWS, ms	88.7±29.9	79.1±18.8	0.17
End-systolic MWS, kdynes/cm2	431.2±136.9	382.4±63.4	0.13
Early ejection-phase MWS (kdynes·cm-2·s)	106.5±28.2	107.9±207.9	0.82
Late ejection-phase MWS (kdynes·cm-2·s)	87.9±28.8	80.5±14. 6	0.27
Ratio of late/early ejection-phase MWS	0.82±0.1	0.75±0.1	0.008
LA expansion index	1.12±0.45	1.38±0.69	0.019
Reservoir (total) LA longitudinal strain (%)	23.8±10.5	28.2±14	0.09
LA passive emptying fraction	0.22±0.09	0.26±0.09	0.028
Conduit LA longitudinal Strain (%)	10.8±7.1	14.5±8	0.035
Early diastolic longitudinal LA Strain Rate (%/s)	−82.4±47.4	−116.9±70	0.004
LA active emptying fraction	0.37±0.12	0.38±0.15	0.56
Late diastolic longitudinal LA strain amplitude (%)	−13±6.1	−13.7±7.3	0.63
Late diastolic longitudinal strain rate (%/s)	−138.7±63.7	−148.8±81.2	0.52

Values are mean±SD or counts (percentages). BSA=body surface area; MWS=myocardial wall stress.

^*LV mass index >85 (males) or >81 (females) g/m² of BSA¹⁹

Correlation of early/late systolic MWS with LA function

Table 3 and Figure 3 show differences in measures of reservoir, conduit and booster pump LA function between subjects stratified according to tertiles of the ratio of late/early MWS-time-integrals (<0.77, 0.77 to 85, and >0.85). Representative examples of pressure, flow, time-resolved LV wall/cavity ratio, and ejection-phase MWS among subjects in the lowest, mid- and highest tertile are shown in supplemental Figures 1–3, respectively.

Table 3.

Left atrial structure and functional variables in hypertensive subjects in the three tertiles of late/early ejection-phase MWS-time-integrals

Characteristics	Tertile 1 (<0.77)	Tertile 2 0.77 to .85	Tertile 3 (>0.85)	P value
LA volume, ml	77.8 (67.8 to 87.8)	75.1 (67.8 to 82.4)	80.2 (71.5 to 88.9)	0.72
LA volume index, ml/m2	36.2 (30.7 to 41.6)	33.5 (30.3 to 36.7)	37.7 (33.8 to 41.5)	0.39
LA expansion index	1.26 (1.17 to 1.36)	1.19 (1.09 to 1.29)	0.9 (0.82 to 0.99)	<0.001 †‡
Reservoir (total) LA longitudinal strain (%)	27.7 (25 to 30.4)	24.3 (22.2 to 26.5)	19.6 (17.8 to 21.4)	<0.001 †‡
LA passive emptying Fraction	0.25 (0.23 to 0.27)	0.21 (0.19 to 0.23)	0.19 (0.17 to 0.21)	<0.001 *†
Conduit LA longitudinal Strain (%)	13.7 (11.8 to 15.6)	10.9 (9.3 to 12.4)	8 (7.1 to 8.9)	<0.001 *†‡
Early diastolic longitudinal LA Strain Rate (%/s)	−103 (−116 to −90)	−80 (−89 to −70)	−65 (−72 to −57)	<0.001 *†
LA active emptying Fraction	0.39 (0.36 to 0.41)	0.39 (0.37 to 0.42)	0.32 (0.29 to 0.35)	<0.001†‡
Late diastolic longitudinal LA Strain amplitude (%)	−14 (−15.5 to −12.5)	−13.5 (−14.8 to −12.2)	−11.6 (−12.8 to −10.3)	0.027 †
Late diastolic longitudinal Strain rate (%/s)	−150 (−164 to −135)	−146 (−161 to −131)	−121 (−134 to −108)	0.08†‡

Pairwise comparisons with Bonferroni correction:

^*1 vs 2;

^†1 vs. 3;

^‡2 vs 3.

Figure 3.

Measures of LA reservoir, conduit and booster pump function in hypertensive subjects in the lower, middle, and higher tertile of the late/early MWS-time integral ratio. Higher levels of this ratio indicate greater late-systolic load relative to early systolic load.

Whereas LA volume was not significantly different between the tertiles, there were significant differences in all measures of LA function. LA conduit function was progressively reduced among subjects in the mid- and highest tertile, compared to subjects in the lowest tertile. In contrast, indices of LA booster pump function were reduced only in the highest tertile of late/early MWS. Reservoir function (which is influenced by both conduit and booster pump function) exhibited an intermediate trend, with progressive reductions from the lowest to the highest tertile, which were more clear (and statistically significant) in the highest tertile.

Correlation of early and late systolic MWS with LA function in multivariable models

The results of regression models in which early and late systolic MWS are examined as independent predictors of LA reservoir, conduit and booster pump function are presented in Supplemental Tables 2–4, respectively. In these models, early systolic MWS was consistently related to higher LA reservoir, conduit and booster pump function whereas late systolic MWS was consistently related to reduced LA reservoir, conduit and booster pump function. These relationships remained significant and only modestly attenuated after adjustment for age, sex, and race (model 2) or after further adjustment for LA volume, body mass index, systolic and diastolic blood pressure, presence of diabetes mellitus, coronary artery disease, use of angiotensin convertase enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, spironolactone, and diuretics.

Interestingly, the values of the standardized regression coefficients for both early and late systolic MWS were lower in models that assessed parameters of booster pump LA function (Supplemental Table 4) than the values obtained in models that assessed parameters of LA conduit or reservoir function (Supplemental Tables 2–3). Nevertheless, these relationships remained significant and only modestly attenuated after adjustment for demographic factors (model 2) and after further adjustment for LA volume and multiple potential confounders (Model 3).

Discussion

In this study, we assessed the relationship between the myocardial loading sequence (early vs. late MWS) and LA dysfunction in adults with hypertension. LA function was assessed from classic phasic LA volumes as well as novel measurements of LA myocardial deformation (strain). We report, for the first time, that a myocardial loading sequence characterized by prominent late systolic wall stress (relative to early systolic wall stress) is independently associated with reduced reservoir, conduit and booster pump LA function. In the context of available experimental data, our findings support the deleterious effects of late systolic LV loading on systolic-diastolic coupling.

LA structure and function represent highly clinically-relevant phenotypes in hypertensive patients. The prognostic importance of phasic LA function, independent of LA size, is increasingly recognized.^9–12,20 Furthermore, there is a differential clinical course of reservoir and conduit function vs. booster pump function in hypertension. Eshoo et al⁹ reported that, among mild untreated hypertensive subjects, reservoir and conduit LA function was decreased, whereas booster function was increased,. As hypertensive heart disease progresses, however, booster pump function also decreases,^10,11 which is associated with an increased risk of cardiovascular morbidity and mortality in population-based and hypertensive cohorts.^20,21 Our findings indicate that even average levels of late systolic load (late/early MWS time-integral ratios of ~0.77–0.85) are associated with reduced conduit LA function, compared to subjects with low levels of late/early load, with further reduction seen at higher levels of late/early MWS (ratios >0.85). In contrast, only subjects in the highest tertile of late/early MWS (late/early MWS time-integral ratios >0.85) demonstrated reduced LA booster pump function. Since reservoir function is influenced by both conduit and booster pump function, the observed trend in indices of reservoir function was progressive from the lowest to the highest tertile, but more clear in the highest tertile.

In multivariable analyses, we found that, in models that adjust for late systolic load, early systolic load is associated with increased LA function, indicating that it is not only the absolute load, but the myocardial loading sequence, that determines LA dysfunction in hypertension. These findings advance our understanding of systolic-diastolic coupling in the intact heart. Causal links between the systolic loading pattern and abnormal diastolic function has been shown in experimental settings.^22,23 Interestingly, the pattern of systolic load,^1,24,25 rather than the absolute afterload itself,²⁶ appears to be the most important determinant of diastolic function. The intact ejecting LV has been shown to respond to late systolic load with delayed relaxation in animal^1,24 and human²⁵ experiments. Similarly, late systolic MWS has been shown to be associated with impaired diastolic relaxation in a large middle-aged community-based sample⁶ and in subjects with hypertension.⁷

MWS represents the time-varying mechanical load on the myocardium, which is determined by complex interactions between myocardial contractile elements, instantaneous LV geometry and the time-varying hydraulic load imposed by the arterial tree.^5,7,8 As such, MWS integrates the influence of arterial load and LV structure and function on myocardial load. The mechanism by which late systolic load acutely impacts diastolic function remains incompletely understood. Although absolute MWS tends to be lower in late systole, the myocardium may be particularly vulnerable to even small MWS increases during this period. This may be because of intrinsic differences in cellular processes between early and late ejection. Loading during active cross-bridge formation (early systolic load) increases the number of interacting cross-bridges (cooperative activity),²⁷ a physiologic mechanism that allows adequate matching of the number of cross-bridges with systolic load. However, when increased load occurs after the onset of myocardial relaxation, the number of interacting cross-bridges can no longer adapt, which results in a mismatch between the number of cross-bridges and load, and an increased stress imposed on individual cross-bridges.²⁷ Of note, the timing of transition from myocardial fiber contraction to relaxation differs from the timing of transition from ventricular systole to diastole, normally occurs early during the ejection phase,^27–29 and is related to the descending limb of the myocyte cytoplasmic calcium transient.^27,28 In addition to the acute effect of late systolic load on diastolic function demonstrated in various studies, late systolic load has been shown to induce more LV remodeling and fibrosis than early systolic load.³⁰ LV fibrosis may increase diastolic passive stiffness of the LV, imposing excessive load on the atrial myocardium, ultimately contributing to LA dysfunction.³¹

Interestingly, in our study, the myocardial loading sequence was more strongly associated with parameters of conduit and reservoir LA function, although the loading sequence was also associated with reduced LA booster pump LA function, with lower booster pump function observed in subjects in the highest tertile of MWS. A potential explanation for a weaker association between the loading sequence and booster pump function (relative to conduit and reservoir function) relates to the effect of atrial contraction on LV early systolic load. It has been shown that booster pump function imparts kinetic energy to LV inflow in late diastole, which is preferentially preserved by blood that passes directly from inflow to outflow in a single cardiac cycle (referred to as “direct” flow).³² During early LV contraction, this direct flow has the greatest amount of kinetic energy, the shortest distance to the LV outflow tract and is moving in a favorable direction relative to the LV outflow tract, effectively unloading the LV during early systole.³² Therefore, booster pump function reduces early systolic LV load, which may induce preferential shortening over fiber stress generation in early systole. This mechanism (which is specific to booster pump function) may confound/attenuate the relationship between worse LA function and a loading pattern characterized by lower early and greater late systolic MWS.

Our study raises the possibility that manipulating the loading sequence may lead to increased LA function in hypertension. Arterial wave reflections selectively increase mid-to-late systolic LV load and MWS.¹⁷ However, the effect of wave reflections on MWS is influenced by the LV contraction pattern.^5–8 Normally, brisk force development and fiber shortening occur in early systole, which is followed by a dynamic reconfiguration of LV geometry that results in a mid-systolic reduction in MWS relative to LV pressure, thus protecting the cardiomyocytes against excessive load in late systole (a period of increased vulnerability). However, in the presence of contractile abnormalities that compromise early systolic ejection, the dynamic geometric reconfiguration of the LV that favors a reduced MWS relative to pressure is less efficient,^5–8 increasing late systolic MWS and facilitating the ill effects of wave reflections. A reduction in late systolic MWS may therefore be accomplished by: (1) interventions that reduce the hemodynamic load induced by wave reflections (such as inorganic nitrates or other nitric oxide donors),^33,34 and/or; (2) interventions that alter the contraction pattern, which could be accomplished by manipulating myocardial contractility, chronic LV remodeling and/or by enhancing shortening-deactivation, which has recently been proposed as a mechanism linking late systolic MWS with impaired relaxation.^6–8 Our observational study was not designed to assess whether manipulating the loading sequence enhances LA function, which should be the focus of future studies.

Our study should be interpreted in the context of its strengths and limitations. Our relatively large sample, the application of methods suitable for time-resolved MWS estimations, and use of volume-independent measures of LA function (derived from semi-automated measures of LA myocardial longitudinal deformation), are strengths of our study. Similarly, our assessment of LV wall and cavity volumes for MWS computations was based on SSFP cine-MRI; unlike 2D echocardiography (which we utilized in previous studies), this method does not rely on geometric assumptions. Furthermore, we used novel tissue-tracking algorithms to assess LA strain (deformation), a more direct measure of LA myocardial function. We found that the loading sequence was associated with LA function in all three domains (reservoir, conduit and booster); these associations were consistently demonstrated with various measures of LA function in each domain, and were robust to adjustment for potential confounders. The high consistency and robustness of our results adds confidence to our findings. Our study also has several limitations. Our observational study can only demonstrate associations, and does not prove causality. Although the administration of a vasoactive drug such as nitroglycerin, could be introduced to induce changes the loading sequence in an experimental design,⁷ effects of the drug on preload or the atrial myocardium would confound assessments of the cause-effect relationship between changes in the loading sequence and atrial function. Residual confounding cannot be excluded; therefore, unmeasured mechanisms (such as neurohormonal factors or genetic polymorphisms) may lead to non-causal associations between time-resolved MWS patterns and LA function. We utilized a convenience sample to recruit our study participants at VA Medical Center. Our control group was small, which could have led to type II error in some comparisons with the hypertensive group. Therefore, our study population was composed predominantly of males and it may not be adequate to extrapolate these findings to women, younger, or community-based samples of hypertensive adults.

In summary, we assessed the relationship between the myocardial loading sequence (early vs. late MWS) and LA dysfunction in adults with hypertension. We found that a LV myocardial loading sequence characterized by prominent late systolic MWS (relative to early systolic MWS) is independently associated with reduced reservoir, conduit and booster pump LA function in hypertensive adults. We found that that conduit function progressively decreased from the lowest to the highest tertile of late/early ejection-phase MWS, whereas booster pump function was selectively reduced in the highest tertile of late/early ejection-phased MWS. Time-varying MWS represents an integrated index of myocardial-ventricular-arterial coupling.⁵ In the context of available experimental data from previous studies, our findings suggest that time-varying systolic MWS influences atrial function and is thus important for ventricular-arterial coupling. Future mechanistic studies regarding the role of the loading sequence on systolic-diastolic coupling may enhance our understanding of the pathophysiology of abnormal myocardial function and may aid in identifying potential targets or interventions to improve LV diastolic relaxation and systemic adaptations to adverse patterns of systolic load.

Clinical Perspective.

Myocardial wall stress (the load that cardiomyocytes “experience”) during ejection cannot be derived solely from blood pressure. The systolic loading sequence underlies ventricular-arterial coupling in health and disease and likely underlies the development of maladaptive LV hypertrophy and progression to heart failure. We assess myocardial wall stress during ejection using a combination of arterial tonometry and cardiac MRI in patients with hypertension and in normotensive controls. We also measured left atrial function with cardiac MRI. Our study demonstrates that early-systolic myocardial wall stress correlates positively, whereas late-systolic wall stress correlates negatively with LA function. A loading sequence characterized by a prominent late systolic load (relative to early systolic load) is associated with impaired reservoir, conduit and booster pump LA function. Our study supports the role of the systolic loading sequence as a determinant of atrial function in hypertension. The systolic loading sequence represents a suitable therapeutic target to improve diastolic function and atrial function in hypertension, which should be tested in future trials.