|Home | About | Journals | Submit | Contact Us | Français|
While several studies have examined the cardiac functional determinants of exercise capacity, few have investigated the effects of structural remodeling. The current study evaluated the association between cardiac geometry and exercise capacity. Subjects with ejection fraction ≥ 50% and no valvular disease, myocardial ischemia or arrhythmias were identified from a large prospective exercise echocardiography database. Left ventricular mass index and relative wall thickness were used to classify geometry into normal, concentric remodeling, eccentric hypertrophy and concentric hypertrophy. All subjects underwent symptom-limited treadmill exercise according to standard Bruce protocol. Maximal exercise tolerance was measured in metabolic equivalents. Of 366 (60±14 years; 57% male) subjects, 166(45%) had normal geometry, 106(29%) had concentric remodeling, 40(11%) had eccentric hypertrophy and 54(15%) had concentric hypertrophy. Geometry was related to exercise capacity: in descending order, the maximum achieved metabolic equivalents was 9.9±2.8 in normal, 8.9±2.6 in concentric remodeling, 8.6±3.1 in eccentric hypertrophy and 8.0±2.7 in concentric hypertrophy (all p<0.02 vs normal). Left ventricular mass index and relative wall thickness were negatively correlated with exercise tolerance in metabolic equivalents (r= -0.14; p=0.009 and r= -0.21; p<0.001, respectively). Augmentation of heart rate and ejection fraction with exercise were blunted in concentric hypertrophy compared to normal, even after adjusting for medications. In conclusion, the pattern of ventricular remodeling is related to exercise capacity among low-risk adults. Subjects with concentric hypertrophy display the greatest limitation and this is related to reduced systolic and chronotropic reserve. Reverse remodeling strategies may prevent or treat functional decline in patients with structural heart disease.
A decline in exercise capacity is one of the most notable effects of aging and an almost ubiquitous symptom of cardiovascular diseases. To better understand how age and disease-related changes can affect exercise capacity, several recent studies have examined cardiac functional correlates of reduced exercise capacity with aging 1, 2 or hypertension 3-5. In contrast, few have studied the effects of left ventricular structural remodeling and geometry on exercise performance 6, 7.
Echocardiography is a well-established tool to characterize cardiac remodeling in hypertension 8 and systolic heart failure 9. Based on the non-invasive measurements of left ventricular (LV) mass and relative wall thickness by echocardiography, the spectrum of LV geometric patterns can be classified into four patterns 10: normal (NL), concentric remodeling (CR), concentric hypertrophy (CH) and eccentric hypertrophy (EH). While the prognostic implications of LV remodeling are widely recognized in terms of cardiovascular events 11, the influence of LV geometry on exercise capacity remains unclear. A greater appreciation of how adverse LV remodeling may contribute to exercise limitation is needed, because any such abnormalities may represent novel therapeutic targets in the prevention or treatment of exertional symptoms in patients with structural heart disease 12, 13.
Accordingly, the aims of this study were 1) to investigate how LV geometry and its components (LV mass and relative wall thickness) are related to exercise capacity; and 2) to determine the functional correlates of LV remodeling in low risk adults undergoing exercise echocardiography. We hypothesized that abnormal LV geometry would be related to reduced exercise capacity. Specifically, we hypothesized that the pattern of geometry with both abnormal components of relative wall thickness and LV mass (i.e. CH) would be associated with the greatest reduction in exercise capacity compared to normal geometry, while geometry patterns with only one abnormal component of relative wall thickness or LV mass (i.e., CR or EH, respectively) would be associated with intermediate exercise capacity. We further hypothesized that there would be differences in blood pressure, heart rate, or LV diastolic or systolic responses to exercise among LV geometry groups, even in the absence of myocardial ischemia or overt systolic failure.
Subjects were identified from a prospective database of consecutive patients referred for exercise echocardiography at Mayo Clinic, Rochester, MN during the year 2006. Patients were included if they had a comprehensive resting transthoracic echocardiogram (to allow characterization of cardiac geometry) within 90 days (median 1 day) of exercise echocardiography. Patients were excluded if they had reduced ejection fraction (<50%), moderate or severe valvular heart disease, evidence of exercise-induced myocardial ischemia, or atrial fibrillation/ flutter at the time of exercise. This study was approved by the Institutional Review Board.
Clinical and anthropometric variables were recorded at the time of the baseline echocardiogram. Medication use and medical history were abstracted from the record and entered into a prospectively maintained database by specially trained nurses 2. Coronary artery disease was defined as previous coronary revascularization or history of myocardial infarction.
LV mass index (LVMI) was determined using the American Society of Echocardiography -recommended formula based on modeling the ventricle as a prolate ellipse of revolution 10. Relative wall thickness (RWT) was calculated by the formula (2*PWTd)/LVIDd where PWTd = posterior wall thickness and LVIDd = LV internal diameter at end-diastole10. Cardiac geometry was classified as: 1. NL: LVMI ≤ 95 g/m2 in women or ≤ 115 g/m2 in men and RWT ≤ 0.42; 2. CR: LVMI ≤ 95 g/m2 in women or ≤ 115 g/m2 in men and RWT > 0.42; 3. EH: LVMI > 95 g/m2 in women or > 115 g/m2 in men and RWT ≤ 0.42; or 4. CH: LVMI > 95 g/m2 in women or > 115 g/m2 in men and RWT > 0.42 10. Left atrial volume was measured offline using the biplane area–length method 10. All volume measurements were indexed (I) to body surface area.
Doppler echocardiography was performed before starting exercise and immediately after symptom-limited treadmill exercise according to standard Bruce protocol, with maximal exercise tolerance measured in metabolic equivalents (METs). One metabolic equivalent of task (MET) is the energy expended by an average individual at rest, defined by convention as a whole-body oxygen consumption of 3.5 ml oxygen per kilogram of body weight per minute. Because oxygen consumption was not directly measured, exercise capacity (in METs) was estimated on the basis of standardized increments in the speed and grade of the treadmill 14. Patients were not allowed to grip the handle bars during exercise. Blood pressure was measured at the end of each stage of exercise using an aneroid gauge sphygmomanometer with the cuff placed over the upper arm and auscultation of the brachial artery. Echocardiographic images were uniformly acquired and analyzed according to standard recommendations 15. Ejection fraction was measured using the modified method of Quinones et al 16 or by visual estimation 17. The reliability of these methods and close agreement between subjective and volumetric assessments of ejection fraction in our institution have been previously published 18. In the apical 4-chamber view, transmitral early (E) inflow velocity was recorded in the pulsed-wave Doppler mode, and the early diastolic velocity within the septal mitral annulus (e′) was measured by Doppler tissue imaging using the pulsed-wave Doppler mode, placing a 5-mm sample volume at the septal region of the mitral annulus. Transmitral E and septal e′ were defined as the first occurring peak velocities in diastole after the isovolumic relaxation signals.
Baseline characteristics were compared across the four geometry groups using Pearson's Chi-square test (categorical variables) or one-way ANOVA (continuous variables) with comparison of each abnormal geometry group to the normal geometry group using Dunnett's test. During exercise echocardiography, rest versus stress parameters were analyzed using the paired t-test for within-group comparisons, and using one-way ANOVA with Dunnett's test for comparisons among groups. The univariate associations of LVMI and RWT with achieved METs was assessed by Pearson's (parametric) or Spearman's (non-parametric) correlation coefficient. For multivariate analyses, multiple linear regression was used to adjust for age and sex in group comparisons, where the dependent variable was achieved METs (log transformed to satisfy normality assumptions) and factors entered into the model included age, sex and geometry group. All analyses were 2-sided, and significance was judged at P<0.05.
Of a total of 366 (age 60±14 years; 57% male) patients, 166 (45%) had NL geometry, 106 (29%) had CR, 40 (11%) had EH, and 54 (15%) had CH (Table 1). Test indications (p=0.13) and reasons for stopping exercise (p=0.08) were similar among geometry groups. Across geometry groups, patients varied in age, systolic blood pressure and history of hypertension, with the CH group being the oldest, having the greatest proportion with a history of hypertension, and having the highest systolic blood pressure at the time of baseline echocardiography. Patients with any abnormal geometry were more likely to be treated with angiotensin system blockers than those with normal geometry, with a similar trend for beta-blockers. LV volumes differed across groups, with CR having the smallest volume. Resting heart rate was highest in CR; cardiac index was similar across groups. Left atrial volume and E/e’ ratio increased across the groups from NL to CH.
In the entire sample, LV geometry was related to exercise capacity (ANOVA p<0.0001; p for trend <0.0001) (Figure 1A). In descending order, the maximum achieved METs was 9.9±2.8 in NL, 8.9±2.6 in CR (p=0.008 vs NL), 8.6±3.1 in EH (p=0.016 vs NL), and 8.0±2.7 in CH (p<0.001 vs NL). Results were similar after excluding the 13 subjects who stopped exercising due to electrocardiographic changes: 9.9±2.8 METs in NL, 8.9±2.7 in CR, 8.7±3.2 in EH, and 8.0±2.7 in CH (p<0.001). In multivariable analyses adjusting for age and sex, geometry group remained independently associated with METs (p=0.04). After adjusting for relevant baseline clinical differences (as listed in Table 1, i.e., age, systolic blood pressure, heart rate, hypertension, smoking status, and usage of beta-blockers and angiotensin-converting-enzyme inhibitors and angiotensin receptor blockers), the association between geometry and METs remained statistically significant (p=0.037).
The individual components of geometry, LVMI and RWT, were negatively correlated with METs (r= -0.14; p=0.009 and r= -0.21; p<0.001, respectively) in the overall sample (Figure 1B & 1C). Among women, there was no significant relationship between LVMI and METs (p=0.11), while RWT remained significantly associated with METs (r= -0.17; p=0.031). Among men, both LVMI and RWT were significantly associated with METs (r= -0.26 and -0.23, respectively; p≤0.001 for each). These associations between the individual components of geometry and achieved METs were no longer statistically significant after adjusting for age in sex-stratified analyses or after adjusting for age and sex in the entire sample.
Systolic blood pressure, heart rate and ejection fraction each increased from rest to exercise overall (p<0.001) and within each geometry group (all p<0.01) (Table 2). However, the magnitude of observed changes with exercise differed among geometry groups (Figure 2). The increase in systolic blood pressure with exercise tended to be smaller in CH compared to NL (p=0.06), despite the higher peak blood pressures reached in CH. The heart rate response to exercise was significantly attenuated in CH compared to NL, even after adjusting for beta-blockade (p=0.001) or excluding all subjects on beta-blockers (p=0.015). Despite similarly increased LV mass in EH and CH, exercise heart rate response was preserved in EH (p=0.7 compared to NL; p=0.04 compared to CH). Augmentation of ejection fraction with exercise was blunted in CH, even after accounting for medications (beta-blockers, calcium channel blockers, angiotensin-converting-enzyme inhibitors and angiotensin receptor blockers) (p=0.004). The increase in echo-estimated filling pressures (E/e’ ratio) with exercise was similar among geometry groups (p=0.40), as was the increase in the individual velocities of early mitral inflow (E) (p=0.50) or early mitral annular motion (e') (p=0.93).
This is the largest study to date examining the relationships between left ventricular geometry and exercise capacity. In low-risk adults free of valvular heart disease, myocardial ischemia or arrhythmias, LV geometry was related to exercise tolerance. Compared to normal geometry, those with CH had the worst exercise capacity, while those with CR or EH had intermediate levels of exercise intolerance. This relationship remained significant even after adjusting for age and sex. Intriguingly, poor exercise performance in CH was related to decreased heart rate and LV systolic (ejection fraction) reserve responses with exercise, even after accounting for medications. These findings support the notion that ventricular remodeling plays a role in the pathogenesis of functional decline in patients with structural heart disease.
The importance of LV geometry has been recognized for decades 19, 20. LV remodeling is known to play a central role in the pathophysiology of cardiovascular disease 8. According to the classic paradigm, a stimulus for remodeling, typically pressure overload, causes myocytes to increase in width, thereby normalizing the pressure-induced increase in wall stress by Laplace law 19. In some patients, the predominant stimulus to LV hypertrophy is volume overload, with resultant EH 21. The adaptive increase in myocardial mass, however, may be associated with maladaptive alterations and LV dysfunction. Indeed, the prognostic impact of LV geometric patterns on cardiovascular events has been extensively described 22, 23. Recent studies have raised serious questions with the notion that hypertrophic remodeling may ever be “adaptive” 24, 25, leading some investigators to suggest that hypertrophic remodeling in itself represents an important therapeutic target 26.
While abnormal geometry is clearly associated with increased cardiovascular outcomes, surprisingly little data has been published regarding the impact of LV geometry on physiology, exercise capacity and reserve function. Two previous studies examined the association between LV geometry and exercise capacity 6, 7. Tomiyama et al. 6 studied a male population with hypertension (N=192; mean age 42-45 years) and found that among the 4 geometric groups, CH was associated with reduced treadmill exercise time compared to normal geometry. Further, men who developed CH during 3 years of follow-up demonstrated a reduction in exercise time. Pierson et al.7 studied 89 hypertensive adults (44% male, mean age 46-48 years) and similarly found shorter exercise treadmill time in CH compared with any other geometric pattern. Importantly, peak exercise oxygen consumption was also measured in the latter study and shown to be lower in CH than in CR or normal geometry.
Examining the separate components of RWT and LV mass in multivariate analysis adjusting for age and sex, Pierson et al. demonstrated an independent effect of LV mass indexed to height, but not RWT, on peak exercise oxygen consumption. A sex-stratified analysis was not available in this study, but was performed in a study by Gharavi et al.27 where LV mass predicted maximal VO2 in hypertensive men but not women - an interesting finding also observed in the current study. While RWT was not examined in the prior study, we further found that despite a lack of association with LV mass in women, increasing RWT was still associated with decreasing exercise capacity in women, suggesting that LV “shape relative to size” may be a more important determinant of exercise performance in women than “absolute LV size”. However, none of these associations between individual components of LV geometry and exercise capacity remained statistically significant after accounting for age in men and women in our study, and further studies are warranted.
In aggregate, our data are consistent with the prior and extend the findings to older subjects in the largest study of its kind to date. The large sample size in this study provided statistical power to detect even modest (r<0.2) associations of potential physiological significance. Similarly, modest associations of known clinical importance include the correlation between severity or duration of systemic hypertension and degree of left ventricular hypertrophy28. The large number of subjects in the current study also allowed adjustment for important covariates (age, sex, medications) to yield clinically meaningful results. Clinical characteristics and distribution of LV geometry patterns in this sample closely resemble those observed in population-based studies from the Olmsted County, MN general community 29 of subjects typically at risk of heart failure with preserved ejection fraction (HFpEF) 5, 29, 30.
The current data suggest that impaired systolic and chronotropic reserves may contribute to reduced exercise capacity in adults with CH. Similar findings have been observed in other populations with cardiovascular disease. Patients with HFpEF often display CH, and hypertensive heart disease is considered a precursor condition to HFpEF 5, 29, 30. In prior studies comparing hypertensive controls to patients with HFpEF, blunted increases in ejection fraction with peak exercise 5, 31 , as well as attenuated chronotropic responses 5, 32, were demonstrated in HFpEF. While a direct association with LV geometry was not investigated in these prior studies, HFpEF patients 5 had smaller LV volumes and increased LV mass compared to hypertensive controls with predominantly CH. More recent studies from this group 33 have again shown that contractile reserve is attenuated in HFpEF compared with hypertensives, even during submaximal effort. These findings lead to the speculation that “occult” systolic dysfunction may be present at rest in CH, despite preservation of ejection fraction and chamber contractility, and that this systolic dysfunction becomes apparent with exercise stress. Indeed, it is known that the increase in relative wall thickness in CR allows normal circumferential shortening at the endocardium and preservation of ejection fraction, even when midwall myocardial shortening is depressed 34, and that midwall LV shortening is inversely related to LV mass index under resting conditions 35.
The current results suggest that hypertrophic remodeling in itself may contribute to exercise disability by impairing systolic and chronotropic reserve function. The observation across studies of a progressive impairment from normal geometry to CH to HFpEF suggests that upstream therapies targeting hypertrophic remodeling may prevent the progression of asymptomatic hypertension (Stage B) to symptomatic heart failure (Stage C). These data lend support to the use of anti-remodeling strategies to preserve or improve exercise tolerance in future trials. Intriguingly, aerobic exercise itself has been shown to have a beneficial effect on ventricular remodeling. This “physiological” remodeling is manifested as eccentric hypertrophy in athletes and serves as an adaptive response to enhance exercise performance. Attenuation of LV concentricity by exercise training, without affecting LV mass, has been elegantly demonstrated in animal models of hypertension36, 37. In this regard, it is of interest that chronotropic response was relatively preserved in EH compared to CH despite similar use of beta-blockers in both groups in the current study. Further, while EF responses to exercise were similar, an increase of 6% in EH represents a larger absolute increase in stroke volume, because of the larger end-diastolic volume in EH. We speculate that both these mechanisms may potentially contribute to better exercise tolerance in EH. Importantly, we cannot discern “pathologic” from “physiologic” eccentric remodeling in this study, and the EH group likely included both types of patients. It may be that any benefits from physiologic EH were offset by pathologic EH. This deserves further study.
Invasive measurements and expiratory gas analysis were not available in this study. However, echocardiography is a well-established clinical tool to assess LV function, and METs provide a reasonable measure of exercise capacity that has been validated and widely applied in prior studies 2, 38-41.While it would be preferable to hold all patients at peak workload to obtain all measurements, this was not feasible due to limiting symptoms of dyspnea and fatigue. All groups were examined with the same protocol (imaging achieved immediately following peak), and there was no systematic disparity among geometry groups in the way heart rate and blood pressure decayed during early recovery (p>0.05 for all group comparisons with Dunnett's test). Ejection fraction varies inversely with afterload, and there may have been differences in the change in LV loading during exercise among the geometry groups. Nonetheless, the change in systolic blood pressure during exercise was lowest in CH and EH patients, arguing against afterload mismatch as an explanation for the impaired EF response in these patients. The large standard deviations of E/e’ measurements during exercise may have limited our inability to detect a statistically significant difference. The notable increase in left atrial size across geometry groups suggests an increasing diastolic burden across groups; the contribution of left atrial function to exercise capacity deserves further study.
Among low risk patients referred for exercise echocardiography, abnormal ventricular geometry was associated with impaired exercise performance. Patients with CH displayed the most impaired exercise capacity, in association with reduced systolic and chronotropic reserve. These results support the notion that ventricular remodeling influences cardiovascular performance with exercise, and may serve as a novel therapeutic target to prevent or treat functional decline in patients with structural heart disease.
SOURCES OF FUNDING
This work was supported by a grant from the Mayo Foundation.
CONFLICT OF INTEREST/DISCLOSURE