|Home | About | Journals | Submit | Contact Us | Français|
1) Compare left ventricular (LV) systolic stiffness and contractility in normal subjects, hypertensives without heart failure, and patients with heart failure and preserved ejection fraction (HFpEF); and 2) Determine whether LV systolic stiffness or myocardial contractility are associated with mortality in HFpEF.
Arterial load is increased in hypertension and is matched by increased end-systolic LV stiffness (ventricular-arterial coupling). Increased end-systolic LV stiffness may be mediated by enhanced myocardial contractility or processes which increase passive myocardial stiffness.
Healthy controls (n=617), hypertensives (No HF, n=719) and patients with HFpEF (n=244, 96% hypertensive) underwent echo-Doppler characterization of arterial (Ea) and LV end-systolic (Ees) stiffness (elastance), ventricular-arterial coupling (Ea/Ees ratio), chamber-level and myocardial contractility (stress-corrected midwall shortening).
Ea and Ees were similarly elevated in hypertensives with or without HFpEF compared with controls, but ventricular-arterial coupling was similar across groups. In hypertensives, elevated Ees was associated with enhanced chamber-level and myocardial contractility, while in HFpEF, chamber and myocardial contractility were depressed compared with both hypertensives and controls. Group differences persisted after adjusting for geometry. In HFpEF, impaired myocardial contractility (but not Ees) was associated with increased age-adjusted mortality.
While arterial load is elevated and matched by increased LV systolic stiffness in hypertension with or without HFpEF, the mechanisms of systolic LV stiffening differ substantially. These data suggest that myocardial contractility increases to match arterial load in asymptomatic hypertensive heart disease, but that progression to HFpEF may be mediated by processes which simultaneously impair myocardial contractility and increase passive myocardial stiffness.
Half of patients with heart failure (HF) have a preserved ejection fraction (HFpEF)(1-4). HFpEF predominantly afflicts elderly, hypertensive patients(1,3,4). Vascular stiffness increases with age, promoting systolic hypertension and increased effective arterial elastance (Ea)(5,6). Left ventricular (LV) end-systolic stiffness (elastance, Ees) increases in tandem(5-7), such that the relationship between ventricular and arterial elastance (ventricular-arterial coupling) remains relatively constant(8-10).
End-systolic LV elastance is a measure of contractility, but it is also influenced by chamber geometry and passive myocardial stiffening(8,11). Ees is elevated in HFpEF(7,10), yet many(12-14), though not all(15) prior studies have reported that various systolic function indices are mildly depressed in HFpEF. Indeed, it is well recognized that impairments in myocardial contractility may coexist with preserved EF in hypertensives with concentric remodeling(16-20). This phenomenon allows preservation of endocardial motion despite reduced shortening of individual myofibers, such that the EF remains normal(19,21).
We sought to compare and contrast chamber and myocardial contractility, LV endsystolic stiffness and ventricular-arterial coupling in three groups of patients: healthy controls without cardiovascular disease, hypertensives without HF and patients with HFpEF, using multiple load-independent measures of chamber and myocardial contractility. All participants were drawn from a large scale, non-selected, population-based sample. To account for differences in ventricular geometry, contractile indices were contrasted within each pattern of chronic chamber remodeling. To determine the clinical significance of these findings, relationships between contractility or LV systolic stiffness and mortality were examined.
The unique aspects of the Rochester Epidemiology Project for population-based research have been described(2,3). The study was approved by the Mayo Institutional Review Board. A random sample (n=2042) of the Olmsted County, Minnesota population aged ≥45 years underwent echocardiography and record review. From this cohort, two control groups were identified(2,3,10): healthy, non-obese controls without cardiovascular disease or diabetes, and hypertensive controls without HF. From the same community, consecutive patients with HFpEF and no significant valvular disease were identified using the Framingham criteria(2,3). Vital status through March 2008 was determined from the Mayo Clinic registration database and the Rochester Epidemiology Project death database(2,3). Mortality data was ascertained from medical records, death certificates for Olmsted County residents, obituaries and notices of death in the local newspapers. Data on all Minnesota deaths were obtained from the State of Minnesota annually. Some clinical characteristics and ventricular function parameters from subjects in this study have previously been published(1-3,10), but most of the systolic indices and their associations with outcomes have not.
Comprehensive echocardiographic assessment was performed by registered diagnostic cardiac sonographers using standardized instruments and techniques, with studies interpreted in a blinded fashion(10). Ventricular dimensions, wall thickness, chamber volumes, and stroke volume were determined in triplicate from 2-dimensional, M-mode echocardiography, and Doppler spectra using standard methods(10). Sex-specific definitions for ventricular hypertrophy and geometry patterns based on LV mass index and relative wall thickness (normal, concentric remodeling, concentric hypertrophy, and eccentric hypertrophy) were used(10). Left ventricular end-diastolic pressure was estimated from echo-Doppler and tissue-Doppler(10). Brachial blood pressure (BP) was determined by sphygmomanometry. End-systolic pressure was determined from the product of 0.9*systolic BP(8). Effective arterial elastance (Ea=end-systolic pressure/stroke volume) and circumferential end-systolic wall stress (cESS) were determined as measures of ventricular afterload(8,19).
Endocardial fractional shortening (eFS) was determined from two-dimensional systolic and diastolic dimensions. LV end-systolic elastance (Ees) was determined by the single beat technique(22). Ventricular-arterial interaction was quantified by the coupling ratio (Ea/Ees). To account for both afterload and preload, two additional load-independent measures of chamber contractility were examined: (1) wall-stress-corrected eFS (sc-eFS), determined by expressing observed eFS as a percentage of that predicted for any given wall stress, based upon the regression equation derived in the healthy controls(18); (2) preload recruitable stroke work (PRSW), determined using a validated single-beat technique(23).
Measures of chamber-level contractility do not necessarily reflect myocardial contractility(16-19,21), because motion at the endocardial surface is greater than predicted by sarcomere shortening alone, due the phenomenon of cross-fiber shortening. Shortening of muscle fibers oriented in orthogonal directions at the inner and outer surfaces of the heart causes marked thickening in the radial axis(21). This effect is enhanced in the setting of concentric remodeling, allowing individual reductions in myofiber contraction to achieve the same net displacement of endocardium, preserving endocardial-based parameters such as EF(16,18,19,21). To assess myocardial contractility, circumferential midwall fractional shortening (mFS) was assessed using the two-shell method of Shimizu(16-19). To minimize afterload dependence, stress-corrected mFS (sc-mFS) was determined as a percentage of that predicted for any given wall stress using the regression equation derived from the healthy control population(18).
Categorical variables were compared by the Chi-square test, and continuous variables were compared using one-way ANOVA with Bonferroni correction. Regression analysis was used to adjust for age, sex, body size, chamber size and morphology, or the presence of other diseases, where the dependent variable was the normally distributed continuous (linear least-squares regression) or categorical (logistic regression) outcome variable of interest. Any interaction between these variables was also evaluated and accounted for as appropriate. The Kaplan-Meier method tested for differences in survival between groups by the log-rank test. Cox proportional-hazards regression was used to adjust for the effect of differences in baseline characteristics on survival.
Of 2042 randomly selected community residents, 617 met the criteria for the healthy control group and 719 subjects met the criteria for the hypertension without HF group. A total of 244 patients constituted the HFpEF group. Nearly all HFpEF patients had a history of hypertension, and were older, more obese, and had higher prevalence of coronary artery disease and diabetes than hypertensives or controls (Table 1).
Hypertensives and HFpEF displayed increased afterload (Ea and cESS) compared to controls (Table 2). While Ea was similarly increased in hypertensives and HFpEF, cESS was higher in HFpEF. As previously reported in this population(10), Ees was similarly increased in hypertensives and HFpEF compared with controls (Table 2). Overall, Ees was strongly correlated with Ea (Figure 1A), and both this relationship and the mean coupling ratios (Ea/Ees) were similar in all three groups (Table 2), indicating preserved ventricular-arterial coupling.
As compared to controls, EF was similar but eFS and mFS were higher in the hypertensive group. In contrast, in HFpEF, EF, eFS and mFS were reduced as compared to hypertensives or controls (Table 2). Similarly, load-independent measures of chamber contractility (PRSW and sc-eFS) were higher in hypertensives as compared to controls and lower in HFpEF compared with hypertensives or controls (Table 2, Figure 1). Adjusting for wall stress (cESS), mFS was higher in hypertensives compared to controls (Figure 2A) and lower in HFpEF compared to controls (Figure 2B) and hypertensives (Figure 2C). HFpEF displayed lower scmFS than hypertensives or controls (Table 2, Figure 1), even after adjusting for age, gender, body size, renal function, beta-blocker use, history of coronary disease and diabetes (p<0.0001). The cumulative distribution of sc-mFS was shifted rightward from controls in hypertension and leftward in HFpEF (Figure 2D), indicating that myocardial contractility was systematically enhanced in hypertensives and impaired in HFpEF.
As previously described in this cohort(3,10), relative wall thickness increased from controls to hypertension to HFpEF (Table 1). The prevalence of left ventricular hypertrophy was similarly increased in hypertensives and HFpEF. The distribution of geometry among hypertensives and HFpEF patients was different from that of control patients (Figure 3A), and tended to be different in HFpEF versus hypertensives (p=0.045).
In healthy controls, Ees (Figure 3 B) and most contractile indices (Figure 3 C) were systematically elevated or reduced as a function of chamber geometry alone. To adjust for confounding effects of chamber remodeling between the groups, we compared Ees and each contractile index within each geometry pattern. Figure 4 shows that, regardless of geometry, Ees was consistently elevated in hypertensives and HFpEF as compared to controls, while PRSW, sceFS and sc-mFS were each consistently higher in hypertensives as compared to controls and lower in HFpEF as compared to both hypertensives and controls.
Median follow-up was 3.1 years (mean 3.1 ± 0.6 years) in the HFpEF group. Mortality at 3 years was 36.4% in HFpEF, 3.1% in hypertension, and 0.8% in controls. In the HFpEF group, survival decreased with greater impairment of myocardial contractility (Figure 5). After adjusting for age, sc-mFS below the median was associated with a 33% increase in mortality (p=0.013). Impaired sc-mFS remained a significant predictor of mortality after adjusting for age, body mass index, coronary disease, hypertension and diabetes mellitus (p=0.01). In contrast, EF, Ees, Ea, and geometry pattern were not associated with age-adjusted mortality (p>0.05).
This is the largest population-based study to date examining left ventricular systolic properties in HFpEF, exploring the mechanisms underlying ventricular-arterial coupling in hypertensive heart disease according to the presence or absence of HFpEF. Among hypertensives, increases in end-systolic LV stiffness were associated with increased chamber and myocardial contractility. In contrast, similar increases in HFpEF patients were associated with impaired contractility, suggesting that elevated Ees may be related to passive myocardial stiffening to a greater extent in this group. Disparities in contractile function were not due to differences in chamber geometry, and impaired myocardial contractility in HFpEF was associated with increased mortality. We speculate that over time, patients with hypertensive heart disease who develop HFpEF acquire structural or functional perturbations which impair myocardial contractility, and that these perturbations contribute to the transition to and progression of overt HF, despite preserved EF.
The interaction of the heart with the arterial system (ventricular-arterial coupling) is a key determinant of cardiovascular performance(8,9). Ea is a lumped parameter reflecting total arterial afterload, incorporating mean and pulsatile components. Ees is determined invasively from the slope and intercept of the end-systolic pressure-volume relationship, but may also be measured noninvasively(22), allowing Ees to be determined in larger patient populations. Ventricular-arterial coupling is expressed by the Ea/Ees ratio(8).
While EF is the most commonly utilized measure of systolic function in clinical practice, it is potently influenced by loading conditions and chamber remodeling(24,25). EF is more accurately conceptualized as a measure of ventricular-arterial coupling. Under normal circumstances, the Ea/Ees ratio varies from 0.5-1.0, a range where cardiac work and efficiency are optimized(8,9). While normal ventricular-arterial coupling ratios (and EF) were observed in each patient group, there were dramatic differences in the ways in which coupling was maintained—enhanced contractility in hypertensives without HF but impaired contractility in HFpEF.
It is well recognized that changes in contractile performance alter Ees(22,24) but Ees is also influenced by chamber geometry, and by factors which alter the passive stiffness of the myocardium(8,11). With aging, increases in arterial stiffness are associated with tandem increases in both systolic and diastolic LV stiffness(5,6,8). Indeed, in this study population, we have previously reported that diastolic ventricular stiffness is elevated in both hypertensives and HFpEF compared with healthy controls, but is highest in HFpEF(10). Taken together with the current findings of impaired contractility despite elevated Ees in HFpEF, we speculate that the processes which contribute to diastolic stiffening in HFpEF influence systolic stiffness as well.
Seminal reports from the 1980's and 90's demonstrated that abnormal myocardial contractility may coexist with a normal EF, because concentric geometric chamber remodeling preserves the extent of endocardial motion relative to the diastolic cavity size(16-19,21). A number of studies have reported abnormalities in regional systolic function in HFpEF, particularly shortening in the longitudinal axis(12,26-29). However, the significance of these findings has been questioned(30), because systolic velocities vary inversely with afterload(31), typically elevated in HFpEF patients(8,10), and because longitudinal shortening does not fully reflect chamber-level contractility(30). Examining load-independent parameters of chamber and myocardial contractility in a large, population-based study, we show that patients with HFpEF indeed do display systolic dysfunction compared with both hypertensives without HF and healthy controls.
Two important but smaller-sized studies also found that roughly a third of HFpEF patients fell below the 95% prediction bands for the relationship between mFS and cESS observed in healthy controls(14,15). More importantly, over 90% of HFpEF patients fell below the mean regression line describing healthy controls. This is consistent with the systematic shift in the distribution of myocardial contractility in HFpEF observed in the current study. However, prior studies did not compare HFpEF with hypertensive controls, HFpEF subjects were highly selected, there were no adjustments for differences in LV geometry, and the impact of impaired myocardial contractility on survival was not examined.
Ees increases with decreasing LV size, yet even after adjusting for differences in geometry, Ees remained significantly elevated in HFpEF, while each additional load-independent index of chamber-level and myocardial contractility was impaired. This “disconnect” between Ees and other measures of contractility has been observed in animal models of pressure overload HF, where increased Ees coexists with impaired chamber, myocardial and myocyte contractility, fibrosis, diastolic dysfunction and impaired beta adrenergic signaling(32). The association of resting contractile dysfunction with increased mortality in HFpEF, viewed in light of these animal studies and recent studies demonstrating abnormal contractile reserve with stress in HFpEF(33-35) indicates that impaired contractility, however mild at rest, may not simply be an innocuous bystander in HFpEF, but rather may reflect processes which mediate progression to overt HF.
Hypertension is a dominant risk factor for HFpEF(1,4), and many of the cardiovascular features in HFpEF are also seen in asymptomatic hypertensives(13,33). As such, comparisons between these two groups provide valuable mechanistic insight into what specifically distinguishes the HFpEF phenotype. While elevated Ees in HFpEF coexisted with impaired contractility, elevated Ees was associated with enhanced contractility in hypertensives. Earlier studies have reported reduced myocardial contractility in hypertension, and that the presence of impaired myocardial contractility is associated with higher rates of cardiovascular events(36). However, the patient groups populating these earlier referral-based studies were often pre-selected for the presence of hypertrophy, and had more extreme levels of concentric remodeling(13,16,18,19). We also found that sc-mFS was categorically impaired in the presence of concentric relative to normal geometry—regardless of patient group (Figures (Figures33 and and4).4). However, within each geometry pattern, contractility was consistently enhanced in hypertensives without HF. The hypertensive patients in the current study were younger than HFpEF, and it may be that most of the hypertensive controls in this population-based study were at an “earlier” stage of disease, where enhanced contractility may be observed as has been reported in human (37,38) and animal studies(39). Alternatively, hypertensives in this more contemporary, population-based study may have been more optimally treated(40), since the extent of concentric remodeling was not as extreme as seen in earlier referral-based studies. If hypertensive heart disease and HFpEF exist in a continuum as has been suggested(10), it may be that the processes leading to concomitant loss of contractile hyper-function and passive stiffening play a role in the transition from hypertension to clinically-evident HFpEF. Alternatively, the myocardial response to pressure overload may differ in patients predisposed to develop HFpEF. These concepts merit future study in chronic longitudinal studies.
The methods employed to assess systolic function are validated against gold standard techniques(10,22,23), but echo-Doppler data inherently have greater variability compared with invasive measurements. While sampling bias was minimal in this population-based study, our study cohort was almost exclusively white, and these results may not be applicable to other ethnic groups. Cause of death data is not available from this population, and we are unable to determine how impaired contractility might be related to mode of death. These data are observational in nature and therefore cannot prove causality or temporal progression.
While ejection fraction and ventricular-arterial coupling are similarly “normal” in hypertensives with or without HFpEF, the mechanisms whereby LV systolic elastance increases to match arterial load differs according to the presence of HF. Patients with hypertension without HF display enhanced Ees and contractility, while Ees in HFpEF is elevated despite impaired contractility, at both the chamber and myocardial levels. These differences are independent of geometry, and myocardial contractile dysfunction is associated with increased mortality in HFpEF, emphasizing its clinical importance. Therapies targeting processes which mediate concomitant contractile dysfunction and passive stiffening may prove useful in the treatment of patients with HFpEF.
Support: NIH HL 63281, HL 72435, HL 55502, and the Marie Ingalls Career Development Award in Cardiovascular Research
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest: None
The authors have no disclosures.