Author + information
- Received September 29, 2003
- Revision received February 12, 2004
- Accepted April 13, 2004
- Published online August 4, 2004.
- Samer S. Najjar, MD⁎,⁎ (, )
- Steven P. Schulman, MD†,
- Gary Gerstenblith, MD, FACC†,
- Jerome L. Fleg, MD, FACC⁎,
- David A. Kass, MD†,
- Frances O’Connor, MPH⁎,
- Lewis C. Becker, MD† and
- Edward G. Lakatta, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Samer S. Najjar, Laboratory of Cardiovascular Science, National Institute on Aging, National Institutes of Health, 5600 Nathan Shock Drive, Baltimore, Maryland 21224.
Objectives The goal of this study was to examine the age-associated differences in ventricular-vascular coupling, defined by the ratio of arterial elastance (EaI) to left ventricular systolic elastance (ELVI), and its components, at rest and during exercise.
Background Ejection fraction (EF) increases during exercise, but the EF reserve decreases with aging. Ejection fraction is inversely related to EaI/ELVI, an index of the interaction between arterial and ventricular properties, which is an important determinant of cardiac performance. Thus, age differences in EaI/ELVI during exercise, due to age differences in EaI, ELVI, or both, may help to explain the age deficit in EF reserve.
Methods We noninvasively characterized EaI/ELVI = end-systolic volume index (ESVI)/stroke volume index (SVI) and its two determinants EaI = end-systolic pressure/SVI, and ELVI = end-systolic pressure/ESVI, at rest and during exercise in 239 healthy men and women (age range, 21 to 87 years). Blood pressures were assessed with cuff sphygomanometry, and cardiac volumes with gated blood pool scintingraphy.
Results Resting EaI/ELVI was not age related in men or women. In both sexes, EaI/ELVI decreased during exercise and declined to a lesser extent in older subjects. There were gender differences in the components of EaI/ELVI during exercise: EaI was greater in older versus young women (p = 0.01) but was unaffected by age in men. Left ventricular systolic elastance increased to a greater extent in young versus older subjects (p = 0.0001 for men, p = 0.07 for women).
Conclusions Age-associated differences in EaI/ELVI occur in both genders during exercise. Sub-optimal ventricular-vascular coupling helps to explain the age-associated blunting of maximal exercise EF, and its underlying mechanisms appear to differ between men and women.
Although ejection fraction (EF) is clinically considered an important index of left ventricular (LV) systolic function, it is not exclusively governed by LV properties. Rather, EF is determined by the interaction of arterial and ventricular properties (1–3). Effective arterial elastance (EaI) is a steady-state arterial parameter that characterizes the functional properties of the arterial system by incorporating peripheral vascular resistance, total lumped vascular compliance, characteristic impedance, and systolic and diastolic time intervals (1). Effective arterial elastance shares common units with elastance measures of ventricular function (ELVI), and their ratio (EaI/ELVI), an index of arterial-ventricular coupling (1), is mathematically inversely related to EF (4).
In healthy subjects free of cardiovascular disease, EF increases by up to 30% during exercise (5). In order for the EF to increase during exercise, the coupling ratio EaI /ELVI must decline. Furthermore, studies have shown that with aging, the increase in EF during exercise (i.e., EF reserve) becomes blunted (5,6), suggesting age-associated differences in the shift of the coupling ratio and its components during exercise. Therefore, we sought to noninvasively measure EaI/ELVI and to characterize its arterial and ventricular determinants in healthy normotensive men and women at rest and during graded aerobic exercise: 1) to describe the age-associated changes in these parameters; 2) to compare how they differ between women and men; and 3) to investigate the functional changes that underlie the compromised EF reserve with aging during such (dynamic) exercise.
The study population consisted of 239 healthy, community dwelling participants, 103 women and 136 men, from the Baltimore Longitudinal Study of Aging. All were normotensive (systolic blood pressure [SBP] ≤140 mm Hg and diastolic blood pressure [DBP] ≤90 mm Hg) and free of cardiovascular disease as determined by detailed history and physical examination and normal resting and treadmill exercise electrocardiograms. Men over age 40 and women over age 50 years also had a negative exercise thallium scan. No subject took any cardiac or antihypertensive medications. All subjects gave informed consent to participate, and the study was approved by the institutional review board.
Exercise protocol and blood pressure
All subjects underwent gated blood pool scintigraphy in supine and upright positions at rest and during graded upright maximal cycle exercise, with determination of oxygen consumption from expired gas analysis. Upright graded exercise on an electronically braked cycle ergometer started at a 25-W workload and was increased by 25 W every 3 min, with pedal speed maintained at 60 rpm until exhaustion. Rest and exercise SBP and DBP were determined by cuff sphygmomanometry. Pulse pressure (PP) was calculated as the difference between SBP and DBP and mean arterial pressure as (2 · DBP + SBP)/3. Additional blood pressure (BP) determinations included end-systolic pressure (ESP) approximated by (2 · SBP + DBP)/3. This noninvasive assessment of ESP accurately predicts LV pressure-volume loop measurements of ESP (7).
Gated blood pool scans
Upright rest and exercise gated blood pool scans were obtained in a 40° left anterior oblique position to best define the ventricular septum after in vivo labeling of red blood cells with 25 to 30 mCi of technetium-99m, as previously described (8). Calculation of EF and cardiac volumes was performed by validated count-based methods described in detail elsewhere (8). All gated blood pool scan-derived volumes were normalized to body surface area, yielding their respective indexes: end-diastolic volume index, end-systolic volume index (ESVI), stroke volume index (SVI), and cardiac index.
The following indexes were calculated: 1) arterial-ventricular coupling index (EaI/ELVI) = ESVI/SVI; 2) EaI = ESP/SVI; and 3) LV systolic elastance index (ELVI) = ESP/ESVI. Of note, EaI/ELVI is mathematically related to EF according to the formula EaI/ELVI = (1/EF) − 1. The noninvasive values of EaI/ELVI are not dependent on BP measurements and can be regarded as relatively accurate. On the other hand, central pressures (including ESP) cannot be accurately measured during exercise; thus, the values of EaI and ELVI should be viewed as approximations.
Because the cardiac response to exercise differs in men and women (5), the data were analyzed separately by gender. Linear regression analyses were performed to determine the relationships between age and EF, EaI/ELVI and its components, EaI, and ELVI, ESVI, SVI, and ESP at rest in the sitting position, at 50%, and at maximal effort. The age relationships of these variables were also evaluated at an absolute work rate of 100 W in men and 75 W in women. The interaction of gender with age as a continuous variable in the entire cohort was determined by analysis of covariance (ANCOVA). If the interaction was not significant, then gender differences in EF, EaI/ELVI, EaI, ELVI, ESVI, SVI, and ESP at the various workloads were determined by ANCOVA. Additionally, to better illustrate age differences, hemodynamic parameters in 38 women and 39 men who were <40 years of age were compared with those from 26 women and 55 men >60 years by analysis of variance. To eliminate the influence of resting differences, we also calculated reserve function, defined as the change in a hemodynamic parameter from rest to individual work rates of 50% of maximum and to maximal exercise. Comparisons of the reserve parameters between young and older subjects were made by unpaired ttests. Associations between peak oxygen consumption (Vo2max) and the elastance parameters were examined using Pearson’s correlation coefficient. To evaluate the independent association between Vo2max and EaI/ELVI, EaI and ELVI, multiple regression analyses were performed for each of these elastance parameters, with age, Vo2max, and age × Vo2max as independent variables in these models. Data are expressed as the mean and SD. A two-tailed p < 0.05 was required for statistical significance.
The subjects’ ages ranged from 21 to 87 years, with a mean of 52.8 ± 16.2 years for the 136 men and 47.9 ± 15.7 years for the 110 women (43 postmenopausal, of whom 17 were on estrogen replacement therapy). Maximal cycle work rate ranged from 50 to 275 W (mean, 149.2 W) in men and 50 to 175 W (mean, 107.3 W) in women. Peak Vo2was 34.1 ± 9.0 ml/kg/min in men (range, 12.1 to 61.5 ml/kg/min, n = 96) and 27.7 ± 5.2 ml/kg/min in women (range, 15.1 to 37.7 ml/kg/min, n = 84). The Vo2max did not differ between the postmenopausal women who were (n = 12) and those who were not (n = 19) on estrogen replacement therapy (p = NS). The age regressions for resting and exercise ESP for men and women are shown in Table 1.In this normotensive cohort, resting ESP significantly increased with age in women but not in men. Similarly, at maximal exercise, ESP significantly increased with age in women but not in men. By ANCOVA, there were significant gender differences in maximal exercise measurements of ESP.
Resting EF was not age-related in either men or women (Table 1). During exercise, at submaximal work rates of 100 W and 50% of maximal workload, and at maximal workload, EF decreased with increasing age in men. During exercise, at 50% of maximal workload, and at maximal workload, EF decreased with increasing age in women. There were gender differences in EF only in the resting state, with women having higher values.
Figure 1Ashows the changes in EF from supine to the sitting position, to 50% maximal workload, and to maximal exercise. To illustrate the effects of age and gender on these workloads, women and men <40 years and >60 years of age are shown. Exercise EF increased to a greater extent in younger than in older men and women. Significant age differences were also evident in EF reserve (i.e., the change) from rest to maximal exercise.
Resting EaI/ELVI was not age-related in either men or women (Table 1). During exercise, EaI/ELVI varied directly with age in men at submaximal work rates of 100 W and 50% of maximal workload, and at maximal workload. In women, EaI/ELVI was significantly increased with age only at maximal workload.
There was a significant gender difference in EaI/ELVI in the resting state, with men having higher values. There was a significant age-gender interaction for EaI/ELVI at maximal exercise.
Figure 1B shows the rest and exercise EaI/ELVI in younger and older subjects. In both genders, EaI/ELVI decreased during exercise. However, exercise EaI/ELVI declined to a greater extent in younger than in older men and women. Similarly, significant age differences in EaI/ELVI reserve were evident in both sexes, with younger men and women having a greater reserve than their older counterparts.
Next, we separately evaluated the age-associated changes in the two components of the EaI/ELVI ratio (i.e., EaI and ELVI) at rest and during cycle exercise.
Arterial elastance index
At seated rest, there was no relationship between age and EaI in men or women (Table 1). During exercise, there were no age relationships for EaI in men at any workload. In women, EaI increased with increasing age at submaximal work rates of 75 W and 50% of maximal workload as well as maximal workload. Significant gender differences in EaI were noted both at rest and maximal effort.
Figure 2Aillustrates rest and exercise EaI in women and men <40 years and >60 years of age. Effective arterial elastance index reserve was also evaluated to take into account differences in resting EaI; EaI reserve from rest (in the sitting position) to each work rate in older women and men was similar to that in their younger counterparts. These data suggest that the exercise EaI differences between younger and older women may be related to greater resting EaI in the latter group.
LV elastance index
At seated rest, ELVI is negatively associated with age in men, but not in women (Table 1). During exercise, ELVI declined at 50% of maximal and at maximal workloads with increasing age in men. At maximal workload, ELVI trended (p = 0.07) to decline with increasing age in women. No gender differences were noted for ELVI at rest. However, a significant age-gender interaction was observed at maximal exercise.
Figure 2B illustrates rest and exercise ELVI in younger and older subjects; ELVI reserve was significantly greater in younger than in older men, and in younger than in older women. Whereas absolute measures of the exercise ELVI differed only between younger and older men, ELVI reserve was decreased in both older men and women.
To determine whether there were any differences in elastance responses by maximal exercise capacity, we evaluated the association of Vo2max with EaI/ELVI, EaI, and ELVI at rest and at maximal exercise. In women, there were no significant associations between Vo2max and these three elastance measures, either at rest or at maximal exercise. In men, in the resting state, Vo2max was associated with EaI (r = 0.3, p = 0.01), but not with ELVI (p = NS) or EaI/ELVI (p = NS). At maximal exercise, Vo2max was not associated with EaI (p = NS), but was weakly associated with ELVI (r = 0.2, p = 0.04) and inversely associated with EaI/ELVI (r = −0.3, p = 0.004). However, none of these associations remained significant in multiple regression models that included age, Vo2max, and age × Vo2max interaction as independent variables.
The interaction of ventricular and arterial properties, or coupling, is an important and largely under-appreciated determinant of cardiac performance (1–3,9). Because cardiac reserve during maximal exertion decreases with increasing age in healthy humans (5), evaluation of rest and exercise ventricular-vascular coupling in a large, healthy cohort illuminates reasons for this decline. This study is the first: 1) to evaluate the coupling index and to dissect out its elastance components in an attempt to glean insights into the mechanisms that underlie the coupling and its changes with exercise; and 2) to investigate the influence of gender differences on the coupling index and its elastance components.
Ventricular-vascular coupling at rest
We found no age difference in the resting ventricular-vascular coupling index, similar to prior invasive studies (2,4). Resting EaI/ELVI in this study averaged 0.5 to 0.6, which is similar to both noninvasive and invasive coupling measurements in prior studies and a value at which the work efficiency of the heart is maximal (6,10).
We also found no age associations with resting EaI in our cohort. However, previous invasive studies had reported that resting EaI increases with age (2,4). This discrepancy may be related to the careful screening of our healthy subjects for the absence of cardiovascular disease, and for the absence of hypertension as defined by the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure-VI criteria (11). Indeed, when subjects with SBP up to 160 mm Hg are included in the analysis, women did display an age-associated increase in resting EaI (data not shown).
Ventricular-vascular coupling during exercise
Prior studies both in animal models (12) and in humans (13) have shown that EaI/ELVI decreases from rest to peak exercise, similar to our findings in this study. Thus, although in the resting state arterial-ventricular coupling is maintained in a range that maximizes the efficiency of the heart, when the system is stressed, energetic efficiency is sacrificed in favor of cardiac efficacy, manifest by a decrease in the coupling index (i.e., a greater relative increase in ventricular contractility than arterial load).
In this study, EaI/ELVI decreased less in older versus younger persons, suggesting that aging is associated with less reserve capacity, or the inability to attain maximal efficacy, manifest by a lesser reduction in the coupling index.
At rest, we found significant gender differences in the age regressions of EaI/ELVI and EaI, but not ELVI. Saba and colleagues (14) previously reported gender differences in arterial load, and they observed in 71 normotensive subjects that EaI was significantly higher in women than in men. Prior population studies also demonstrate that PP, an index of arterial stiffness, increases with age to a greater extent in women than in men (15).
Men and women displayed mechanistic differences in their predisposition for the age-associated suboptimal efficacy during exercise. Indeed, we found that exercise ELVI is higher in younger than in older men and tended to be higher in younger than in older women, with a significant age-gender interaction. In contrast, we observed that exercise EaI is higher in older than in younger women during all stages of exercise, with no age differences in men, and with a significant gender difference at maximal exercise. Interestingly, the marked gender differences in EaI and ELVI noted in the young individuals at peak exercise (Figs. 2A and 2B) were attenuated in older subjects.
Therapies that improve the coupling of ventricular and vascular elastances are likely to improve cardiac function and exercise tolerance in healthy subjects. This concept is supported by two recent studies in healthy older subjects: Administration of the direct vasodilator sodium nitroprusside caused reductions in reflected waves (manifest as a reduction in the augmentation of the carotid PP) as well as reductions in preload, resulting in reduced cardiac volumes and higher EF at rest and during maximal exercise as compared with placebo therapy (16). Administration of intravenous verapamil reduced noninvasive indexes of arterial and ventricular systolic stiffness and improved exercise tolerance and oxygen consumption before reaching anaerobic threshold (17).
Although measurement of the arterial-ventricular coupling index relies only on evaluation of volumes and is independent of BP, a potential limitation of this study was the use of noninvasive measures to separately characterize the individual components of this index (i.e., effective arterial and LV elastances). The estimation of ELVI by ESP/ESVI assumes V0= zero. Because our subjects were healthy, we likely overestimated actual end-systolic elastance. Furthermore, while acute changes in LV elastance assessed in the catheterization laboratory often reflect changes in contractility, long-term age-associated changes in ventricular elastance estimated in this cohort probably reflect, in addition, increased wall thickness and myocardial fibrosis (6,9). Nonetheless, this approach has been used in prior noninvasive evaluations (4), and importantly, the values of EaI/ELVI in this study are similar to those obtained invasively (3,4). Secondly, brachial artery measurements of BP do not completely describe central aortic pressures, although we should emphasize that the arterial to ventricular coupling ratio is not affected by central pressures because the pressure terms in the numerator and the denominator cancel out. Nevertheless, prior studies suggest that ESP estimated from cuff BPs closely approximate central pressures (7,18). Nussbacher et al. (16) reported that brachial cuff BP measurement overestimate estimated central ESP values by <5% in a group of older, healthy subjects. Furthermore, while central pressures can be derived in the resting state, brachial pressure measurements are the only feasible method to assess ventricular and vascular elastances during exercise in a large, healthy cohort. A final limitation in this study is the comparison of exercise parameters across age groups and the greater maximal workload achieved in younger subjects, resulting in higher arterial and ventricular indexes by virtue of the greater workload. However, even at submaximal workloads, these relationships of age and gender are still evident.
In conclusion, in healthy men and women, noninvasive evaluation of EaI/ELVI allows characterization of age and gender differences in EF at rest as well as the decline in exercise EF reserve with increasing age in both genders. In women, this is likely due to an age-associated increase in exercise arterial elastance without an appropriate rise in ventricular elastance. In men, this is probably due to an age-related decline in exercise LV elastance index, as arterial elastance does not change with age. These arterial-ventricular mismatches in older men and women provide potential mechanisms for the age-associated blunting of maximal exercise EF. Therapies aimed at improving ventricular-vascular coupling in the elderly may improve exercise performance.
Supported, in part, by NO-1 AG-8-2109, National Institute on Aging, and the Intramural Research Program, National Institute on Aging. Drs. Najjar and Schulman contributed equally to this research.
- Abbreviations and Acronyms
- blood pressure
- diastolic blood pressure
- effective arterial elastance index
- ejection fraction
- left ventricular systolic elastance index
- end-systolic pressure
- end-systolic volume index
- left ventricle/ventricular
- pulse pressure
- systolic blood pressure
- stroke volume index
- Received September 29, 2003.
- Revision received February 12, 2004.
- Accepted April 13, 2004.
- American College of Cardiology Foundation
- Sunagawa K.,
- Maughan W.L.,
- Bukhoff D.,
- Sagawa K.
- Chen C.-H.,
- Nakayama M.,
- Nevo E.,
- et al.
- Fleg J.L.,
- O’Connor F.,
- Gerstenblith G.,
- et al.
- Lakatta E.G.
- Kelly R.P.,
- Ting C.T.,
- Liu C.P.,
- et al.
- Schulman S.P.,
- Lakatta E.G.,
- Fleg J.L.,
- et al.
- Ishihara H.,
- Yokota M.,
- Sobue T.,
- Saito H.
- Little W.C.,
- Cheng C.P.
- Franklin S.S.,
- Gustin W. IV.,
- Wong N.D.,
- et al.
- Nussbacher A.,
- Gerstenblith G.,
- O’Connor F.,
- et al.
- Chen C.-H.,
- Nakayama M.,
- Talbot M.,
- et al.
- Kroeker E.J.,
- Wood E.H.