Left Ventricular Filling in Arterial Hypertension
Influence of Obesity and Hemodynamic and Structural Confounders
We assessed the relations of left ventricular filling to load and geometry by Doppler echocardiography in 80 normotensive subjects (40 normal-weight [36±12 years, 24 women] and 40 obese [35±13 years, 24 women]) and 61 hypertensive subjects without silent coronary heart disease (29 normal-weight [43±13 years, 15 women] and 32 obese [42±13 years, 19 women]) and comparable left ventricular midwall performance. Left ventricular mass divided by height to the 2.7 power was higher in all groups than in normotensive normal-weight subjects (all P<.0001) and in hypertensive than normotensive obese subjects (P<.001). After controlling for age, sex, blood pressure, and heart rate, isovolumic relaxation time was prolonged in hypertensive subjects and normotensive obese subjects compared with normotensive normal-weight subjects (all P<.0001). Body mass index, left ventricular dimension and mass, and circumferential end-systolic stress did not influence these differences. In pooled groups, prolonged isovolumic relaxation time was predicted by high mean blood pressure (β=0.52, P<.001), low end-systolic stress (β=−0.33, P<.001), increased left ventricular mass (β=0.24, P<.004), and high body mass index (β=0.14, P<.05, multiple R=.72, SEE=16.5 milliseconds, P<.0001). Between-group differences in peak early transmitral flow velocity, the deceleration time of early filling velocity, and the ratio of early to late left ventricular filling disappeared after controlling for left ventricular mass. Thus, (1) isovolumic relaxation time is prolonged in both arterial hypertension and obesity; (2) the presence of obesity does not significantly increase isovolumic relaxation time in hypertension; and (3) abnormalities of left ventricular filling in arterial hypertension are offset after controlling for left ventricular mass.
Left ventricular (LV) diastolic dysfunction has often been described in arterial hypertension1 2 3 4 5 6 7 8 in either the presence2 3 4 or the absence5 6 of LV hypertrophy. The reported abnormalities have been attributed to abnormalities of cardiac load,6 7 demographic characteristics,8 or increased blood pressure (BP) by itself.3 8 Obesity is among potential confounders that may determine LV diastolic dysfunction in arterial hypertension.9 10 Impairment of LV relaxation, which have been shown to be independent of cardiac load and geometry, have been recently described in obese normotensive individuals.11 12 Because the prevalence of overweight to frank obesity in hypertensive populations is very high,13 obesity might well be an important confounder for the evaluation of LV diastolic dysfunction in hypertension. Accordingly, we designed this study to assess the effect of obesity on LV relaxation and filling in arterial hypertension.
Sixty-one consecutive hypertensive individuals were studied after a preliminary examination in the Outpatient Clinic of the Hypertension Unit of the Department of Clinical and Experimental Medicine (DCEM) of the Federico II University Hospital in Naples. The individuals came to the unit either because of bad BP control (systolic BP higher than 160 mm Hg and diastolic BP higher than 95) in several clinic measurements or as newly diagnosed hypertensive individuals (n=19). Individuals under medication discontinued treatment for at least 3 weeks after giving formal informed consent, in accord with the rules of the Institutional Ethics Committee. Among hypertensive individuals, 29 (15 women) were of normal weight and 32 (19 women) overweight to obese. Eighty normotensive subjects served as controls. Forty of them (24 women) were overweight to obese; they enrolled after informed consent in the Outpatient Clinic of the Nutrition Unit of the DCEM. These subjects came to the unit with the purpose of losing weight for aesthetic reasons and had no signs or symptoms of disease, including diabetes. The remaining 40 normotensive subjects were volunteers involved in a screening program of the department staff.12 Overweight was defined as a body mass index higher than 27.8 kg/m2 in men and 27.3 kg/m2 in women.14 Coronary artery disease was excluded in all participants on the basis of negative family and clinical histories and the absence of symptoms and cardiovascular risk factors (smoking, hypercholesterolemia). In addition, both hypertensive and normotensive subjects older than 40 years underwent a bicycle exercise test to exclude silent ischemia. Further characteristics of the study group are reported below in “Results.”
Doppler echocardiography was performed in a dimly light room with subjects in the partial left decubitus position using a commercially available machine (SIM7000/Challenge, ESAOTE) equipped with a 2.5-MHz annular-array transducer. M-mode tracings were recorded from the parasternal short-axis view as previously described in detail,12 15 coded or recoded when already used for different studies, and read by two observers using a digitizer pointer device and a graphic tablet (Summasketch III, Summagraphic) connected to a personal computer who were blinded to the knowledge of the subjects' sex, body weight, and BP. LV chamber dimensions and septal and posterior wall thicknesses were measured according to the American Society of Echocardiography.16 The Penn convention was used for calculation of LV mass.17 LV mass was normalized for height to the 2.7 power and LV diastolic dimension for the first power of height.18 LV systolic function was evaluated as midwall shortening both as an absolute value and as a percentage of the value predicted from the corresponding circumferential end-systolic wall stress.2 19 20 LV relaxation and filling were evaluated by pulsed-wave Doppler interrogation of the LV inflow tract from the apical four-chamber view, with the sample volume placed first at the mitral annulus and thereafter at the tips of the mitral valve. After a stable signal of the transmitral flow velocity was obtained, the Doppler cursor was moved toward the LV outflow tract in the apical five-chamber view for recording of both mitral and aortic signals, including the closing click of the aortic valve and the opening click of the mitral valve.12 21 Doppler signals were recorded at high speed (80 to 120 mm/s) on videotape with the subjects in held expiration and were measured by two observers using electronic pointer devices. The average of five beats was used for analysis.
Isovolumic relaxation time was calculated as the time from the closure click of the aortic valve to the opening click of the mitral valve. When either the closing or opening click was not identified (in eight hypertensive and six normotensive subjects), the time from the end of the aortic flow to the onset of mitral flow from the continuous wave interrogation of the LV inflow-outflow tract was used. Peak early transmitral flow velocity (E), peak late transmitral flow velocity (A), and the deceleration time of E velocity were measured at the tips of mitral leaflets at the maximum amplitude of E velocity. Deceleration time was measured as the time from peak E velocity to the time when the E wave descent intercepted the zero line. Atrial filling fraction was calculated at the mitral annulus as the ratio of the integral of the A wave to the total diastolic integral of the transmitral flow.
Variables are expressed as mean±SD. One-factor ANOVA was used for comparison of parameters of LV filling between hypertensive and normotensive subjects. The step-down multiple-stage F post hoc test was used for multiple comparisons. Potential confounders (age, sex, systolic and diastolic BPs, and heart rate) were identified by a least-squares linear correlation matrix and were taken into account by forcing covariates into the ANOVA models before between-groups comparison, according to a hierarchical design. In a second step, LV mass and body mass index were sequentially added to the previous confounders to control statistically significant differences for the effects of LV geometry and body size. LV end-diastolic diameter as a rough index of preload and circumferential end-systolic stress as a measure of myocardial afterload were also used as covariates. Multiple regression analysis was performed to identify independent predictors of prolonged isovolumic relaxation time by a stepwise procedure. In all the procedures, sex was used as a dummy variable by assigning 1 to men and 2 to women. The null hypothesis was rejected at a two-tailed value of P≤.05.
Table 1⇓ shows characteristics of the study groups. Hypertensive subjects were older than normotensive subjects (P<.0001). There was no difference in sex distribution. Differences in body size and BP were detected by design; however, diastolic BP was statistically higher in obese than in normal-weight normotensive control subjects (P<.01).
LV Geometry and Systolic Function
Obese subjects had higher LV dimensions than normal-weight individuals in either the presence or absence of arterial hypertension (Table 2⇓, all P<.001). Left atrial size was larger in both normal-weight and obese hypertensive subjects as well as in obese normotensive subjects than in normal-weight normotensive individuals (all P<.0001). LV mass was higher in all three experimental groups than in the normal-weight normotensive control group (all P<.0001), with the highest values found when hypertension and obesity coexisted (P<.005 versus obese normotensive subjects). Relative wall thickness was equally higher in normal-weight or obese hypertensive subjects than in normotensive individuals (both P<.001). End-systolic stress was similar among groups, as well as midwall shortening either as an absolute value or a percentage of predicted from circumferential end-systolic stress.
LV Diastolic Phase
Table 3⇓ shows the differences in Doppler measurements among the four groups: isovolumic relaxation time was similarly prolonged in normal-weight and obese hypertensive subjects and, to a less extent, in obese compared with normal-weight normotensive subjects (all P<.0001). This prolongation was also significantly greater in both hypertensive groups than in normotensive obese individuals (both P<.001). Peak E velocity was lower and the deceleration time of E velocity longer in both hypertensive groups and in obese than normal-weight normotensive individuals (all P<.0001). However, no statistical difference in peak E velocity and the deceleration time of E velocity was found between the normotensive obese group and both hypertensive groups. The E velocity time integral did not show any statistical difference among the groups.
Peak A velocity was greater in hypertensive than normotensive subjects. The A velocity time integral tended to be higher in hypertensive subjects but achieved a statistically significant difference only in hypertensive obese subjects compared with both normotensive groups. The E/A velocity ratio was lower in both hypertensive groups (P<.001) than in normotensive normal-weight control subjects because of a combination of low peak E velocity and high peak A velocity. On the other hand, reduction in peak E velocity was responsible for the lower E/A velocity ratio in normotensive obese subjects compared with normotensive normal-weight subjects (P<.05). No statistical difference in atrial filling fraction was found among groups, although both hypertensive groups and obese normotensive subjects exhibited average values lower than in normal-weight normotensive individuals.
Univariate Correlation Matrix
A univariate correlation matrix between Doppler LV diastolic measurements and potential demographic and clinical confounders is shown in Table 4⇓ for the entire study population. On the basis of these correlations, potential confounders were identified and removed in the between-group comparisons.
Isovolumic Relaxation Time
After controlling for confounders identified in Table 4⇑ (ie, age and BP), isovolumic relaxation time remained prolonged in hypertensive normal-weight subjects (adjusted mean=92 milliseconds), hypertensive obese subjects (adjusted mean=94 milliseconds), and normotensive obese subjects (adjusted mean=89 milliseconds) compared with normotensive normal-weight subjects (adjusted mean=65 milliseconds; all P<.00001), but differences between both groups of hypertensive subjects and obese normal-weight subjects disappeared.
Effect of obesity on isovolumic relaxation
In pooled groups, isovolumic relaxation time was positively related to body mass index (r=.29, P<.0008) also after forcing age and systolic and diastolic BPs in a multiple regression model (partial r=.18, P<.05). After the addition of body mass index to previous confounders in the ANCOVA, isovolumic relaxation time remained significantly prolonged in all groups (adjusted means: 91 milliseconds in hypertensive normal-weight, 93 in hypertensive obese, and 92 in normotensive obese subjects) compared with normotensive normal-weight subjects (64 milliseconds; all P<.0001).
Effect of LV geometry on isovolumic relaxation
Isovolumic relaxation time was also related to LV mass index in pooled groups (r=.57, P<.0001). When separate groups were examined, this relation was closer in normotensive subjects (80 subjects, including normal-weight and obese individuals [r=.49, P<.0001]) than in hypertensive subjects (61 subjects, both normal-weight and obese [r=.30, P<.02]). In normotensive subjects, isovolumic relaxation time increased by 10 milliseconds each 10 g/m2.7 of increase in LV mass index, whereas in hypertensive subjects, this increase was 4 milliseconds each 10 g/m2.7 (slope difference significant at P<.02). The regression line became flat (r=.15, P=NS) in the subgroup of subjects with LV mass index values greater than 51 g/m2.7 (the value that has been recognized as a marker of future cardiovascular events in a follow-up study)22 independently of the presence of hypertension and/or obesity, whereas the relation below that LV mass value was closer (r=.57, P<.0001; isovolumic relaxation time increased by 15 milliseconds each 10 g/m2.7 of LV mass index). At comparable levels of LV mass index (<51 g/m2.7), isovolumic relaxation time was similarly prolonged in both hypertensive groups and in normotensive obese subjects (adjusted means: normotensive obese, 88 milliseconds; hypertensive normal-weight, 94; hypertensive obese, 91 milliseconds; all P<.0001 versus normotensive normal-weight, 69 milliseconds). Thus, LV mass only partially influences isovolumic relaxation in hypertension and obesity.
Effect of LV load on isovolumic relaxation
In pooled groups, isovolumic relaxation time was directly related to LV end-diastolic dimension index (a rough index of preload) (r=.28, P<.002). However, when LV end-diastolic dimension was added as a covariate to age and BP, isovolumic relaxation time remained prolonged in normotensive obese subjects (adjusted mean=89 milliseconds) and in hypertensive subjects who were either obese (adjusted mean=96 milliseconds) or normal-weight (adjusted mean=95 milliseconds) than in normotensive normal-weight individuals (adjusted mean=65 milliseconds; all P<.0001).
Isovolumic relaxation time was also weakly negatively related to circumferential end-systolic stress (r=−.19, P<.03), but again, between-group differences were not influenced by removing the effect of wall stress (adjusted means: 84 milliseconds in normotensive obese, 97 in hypertensive normal-weight, and 102 in hypertensive obese subjects; all P<.0001 versus 61 milliseconds in normotensive normal-weight control subjects).
Multiple regression analysis showed that the main predictor of prolonged isovolumic relaxation time was high mean BP (slope=0.76, β=0.52, P<.001), with an additional independent contribution by low circumferential end-systolic stress (slope=−0.25, β=−0.33, P<.001), high LV mass (slope=0.09, β=0.24, P<.004), and increased body mass index (slope=0.43, β=0.14, P<.05) (multiple R=.72, SEE=16.5 milliseconds, P<.0001).
LV Early Filling
After controlling for demographic and clinical confounders identified by the correlation matrix shown in Table 4⇑, peak E velocity remained significantly lower only in obese normotensive subjects than in normotensive normal-weight control subjects (adjusted means: 61 versus 71 milliseconds; P<.02), whereas in hypertensive subjects, the values were comparable to those in normotensive subjects (adjusted means: 66 milliseconds in hypertensive normal-weight and 68 in hypertensive obese subjects, respectively). The deceleration time of E velocity remained prolonged in all three experimental groups (adjusted means: 172 milliseconds in normal-weight hypertensive, 183 in obese hypertensive, and 177 in normotensive obese subjects) compared with the value in normotensive normal-weight individuals (adjusted mean=146 milliseconds; all P<.0001), without significant differences among the obese normotensive group and both hypertensive groups.
Effect of obesity on early filling
In pooled groups, peak E velocity was negatively related to body mass index (r=−.23, P<.01), even after forcing confounders in a multiple regression model (partial r=−.22, P<.01). The deceleration time of E velocity was directly related to body mass index (r=.25, P<.003); after controlling for age, BP, and heart rate, this relation remained significant (partial r=.25, P<.003).
Effect of LV geometry on early filling
After controlling for age, sex, and BP, LV mass index was negatively related to peak E velocity (partial r=−.39, P<.0001) and positively to the deceleration time of E velocity (partial r=.46, P<.0001). After addition of LV mass to age, sex, and BP in ANCOVA, differences were largely offset in both peak E velocity (adjusted means: 69 cm/s in normotensive normal-weight, 62 in normotensive obese, 66 in hypertensive normal-weight, and 69 in hypertensive obese subjects; all P=NS) and the deceleration time of E velocity (adjusted means: 155 milliseconds in normotensive normal-weight, 170 in normotensive obese, 165 in hypertensive normal-weight, and 176 in hypertensive obese subjects; P=NS).
Effect of LV load on early filling
The addition of LV end-diastolic diameter and wall stress to the ANOVA model did not influence between-group differences detected in univariate analysis.
LV Late Filling
After controlling for age, sex, BP, and heart rate (Table 4⇑), peak A velocity was not any more significantly different among the four groups (adjusted means: 54 cm/s in normotensive normal-weight, 60 in normotensive obese, 62 in hypertensive normal-weight, and 64 in hypertensive obese subjects; all P=NS). After adjusting for age and BP (Table 4⇑), the E/A velocity ratio remained lower only in normotensive obese subjects than in normotensive normal-weight individuals (adjusted means: 1.06 versus 1.24; P<.04), whereas differences between normotensive subjects and both normal-weight and obese hypertensive subjects were offset (adjusted means: 1.15 and 1.14, respectively).
Effect of obesity on late filling
No significant independent influence of body mass index was detected for any of the late passive filling variables.
Effect of LV geometry on late filling
All differences among groups in E/A velocity ratio disappeared after addition of LV mass to previous covariates (adjusted means: 1.2 in normotensive normal-weight, 1.1 in normotensive obese, 1.2 in hypertensive normal-weight, and 1.2 in hypertensive obese subjects; all P=NS).
Effect of LV load on late filling
No significant between-group differences in LV passive filling variables were detected after addition of either LV end-diastolic diameter or circumferential end-systolic stress to confounders reported in Table 4⇑.
The prevalence of obesity is very high in hypertensive populations.13 Because primary impairment of LV isovolumic relaxation has been reported in obese normotensive subjects independently of LV geometry and load,11 12 obesity may be considered a potential confounder when LV diastolic phase is evaluated in arterial hypertension.
Diastolic filling depends on the extent and velocity of relaxation23 as well as on the passive properties of the myocardium.24 During relaxation, active Ca2+ extrusion from the cytosol to the sarcoplasmic reticulum allows the detachment of the actomyosin crossbridges generating myocyte lengthening. This active process is markedly influenced by loading conditions25 26 27 28 and determines the initial value of the atrioventricular pressure gradient and consequently the magnitude of the peak velocity of rapid filling.21
Passive properties of the myocardium are represented by LV pressure-volume relations. Determinants of the passive pressure-volume relation include LV volume and geometry, intrinsic myocardial stiffness, and extrinsic factors such as pericardial function, intrathoracic pressure, and left to right ventricular interaction.24 These determinants may be affected by age and sex.29 As expected, in our study population, age, sex, BP, and heart rate influenced all diastolic variables, contributing variably to explain their variances by 4% to 35%. Thus, to best understand the significance of diastolic abnormalities in hypertensive individuals, Doppler-derived LV filling parameters need to be evaluated independently of confounders including age, sex, obesity, loading conditions, and LV geometry.
Isovolumic Relaxation Time
Isovolumic relaxation time was prolonged similarly in hypertensive individuals and normotensive obese subjects independently of their body weight even after all potential demographic, geometric, and hemodynamic confounders were removed. In particular, isovolumic relaxation time was positively related to LV mass index, but this relation became flat as LV mass increased over 51 g/m2.7. The “leveling off” of isovolumic relaxation at the higher values of LV mass is likely due to the increase in left mean atrial pressure that occurs in hypertensive individuals in the body's attempt to balance the increased chamber stiffness secondary to the higher values of LV mass, thereby avoiding further prolongation of isovolumic relaxation.30 Factors other than LV mass index also influenced the prolonged isovolumic relaxation: in the multiple regression model, LV mass index was only one of the predictors of prolonged isovolumic relaxation time in pooled groups.
In this study, the LV end-diastolic dimension index was used as a surrogate of preload, and isovolumic relaxation time was shown to be independent of it. This finding is in agreement with previous studies demonstrating that isovolumic relaxation is relatively independent of preload.26 28
The influence of afterload on LV relaxation is different whether forces are applied during the early or late contraction period: Load applied during the early ejection period prolongs relaxation (contraction load),23 25 30 31 32 33 34 whereas load applied in late contraction or during either isovolumic relaxation or the rapid filling phase determines relaxation to be faster (relaxation load).23 25 27 30 31 32 33 34 35 Thus, BP may well represent a contraction load, whereas end-systolic stress may be considered a relaxation load. Accordingly, a negative relation between circumferential end-systolic stress and isovolumic relaxation time was expected and found.
Prolonged isovolumic relaxation time in hypertensive individuals could be also due to a lower atrial pressure, which reduces the atrioventricular pressure gradient and consequently the relaxation filling load23 25 31 ; however, this mechanism is unlikely in hypertensive individuals as well as in normotensive obese subjects. Thus, prolonged isovolumic relaxation in both hypertensive and obese subjects appears to be substantially independent of LV load. Because the nonuniformity of relaxation and load23 was unlikely, as silent myocardial ischemia was excluded, the prolonged isovolumic relaxation found in hypertensive groups as well as in normotensive obese individuals might be due to the primary impairment of the intrinsic process of inactivation, only in part related to increased LV mass, BP, and body mass index. The extent of this abnormality in arterial hypertension is only slightly greater than in obesity, without any clear-cut evidence of a summative effect. In ANOVA, as well as in multivariate analysis, the effect of arterial hypertension appears to largely incorporate that of obesity, suggesting the possibility of a similar mechanism. Compared with previous studies demonstrating prolonged isovolumic relaxation time as a marker of diastolic dysfunction,9 11 the present study shows that this abnormality is common to hypertension and obesity and is independent of demographic, hemodynamic, and structural confounders.
Early LV Filling
After controlling for demographic and clinical confounders, including BP, E velocity was lower only in normotensive obese subjects than in normal-weight control subjects, whereas it was normal in hypertension. The low peak E velocity found in normotensive obese individuals depends on the prolonged relaxation that determines a low initial value of the left atrioventricular pressure gradient at the mitral valve opening.12 In hypertensive obese as well as in hypertensive normal-weight subjects, peak E velocity was higher than in normotensive obese individuals, suggesting that elevated mean atrial pressure may offset the consequences of the prolonged isovolumic relaxation.
The prolonged deceleration time of E velocity was relatively independent of age, BP, and heart rate in all three experimental groups compared with the normotensive normal-weight control group. During this period, LV filling is influenced by both the ongoing process of active relaxation and LV chamber stiffness. Slower relaxation prolongs the deceleration time of E velocity by reducing the rate of increase in LV pressure during early filling. Changes in structural properties of the myocardium may either prolong relaxation21 23 24 30 or increase LV chamber stiffness; the magnitude of LV hypertrophy, concentric remodeling of the LV chamber, and either increase or remodeling of the collagen network may shift the diastolic pressure-volume relation upward.24 36 37 Thus, the deceleration time of E velocity may be either prolonged or shortened, moving from the pattern of prolonged relaxation to that of increased chamber stiffness21 through intermediate patterns of pseudonormalization.38 In the present study, consistent with the observed pattern of altered relaxation, the deceleration time of E velocity was prolonged in both hypertensive and obese subjects compared with normotensive normal-weight control subjects, and the between-group difference was largely offset after controlling for LV mass.
Late Passive LV Filling
In this study, late passive filling depended more on BP levels than on obesity or other demographic confounders: Peak A velocity was indeed higher and the E/A velocity ratio was lower in both hypertensive groups than in normotensive subjects. As expected, after controlling for confounders, including BP, differences in peak A velocity between hypertensive and normotensive subjects were eliminated, indicating that high BP contributes to increased late LV filling velocity.
The LV active relaxation phase is impaired in arterial hypertension; prolongation of isovolumic relaxation time is relatively independent of age, sex, BP, body size, and LV geometry and load and therefore might be related to a primary alteration of the process of deactivation. Obesity does not increase significantly the prolonged isovolumic relaxation time found in individuals with arterial hypertension. Alterations of passive properties of the myocardium that can be identified by Doppler echocardiographic measurements are largely due to changes in LV geometry and load occurring in arterial hypertension independent of the presence of obesity. Demographic, hemodynamic, and structural confounders should be taken into account when LV diastolic phase is studied by Doppler echocardiography in arterial hypertension.
This work was supported in part by grant 18/01/5495 MURST/60%, 1994, and by CNR-FATMA-Project No. 201.12.74/2.
- Received February 29, 1996.
- Revision received April 10, 1996.
- Revision received May 31, 1996.
Shimuzu G, Zile MR, Blaustein AS, Gaasch WH. Left ventricular chamber filling and midwall fiber lengthening in patients with left ventricular hypertrophy: overestimation of fiber velocities by conventional midwall measurements. Circulation. 1985;71:266-272.
Fouad FM. Left ventricular diastolic function in hypertensive patients. Circulation. 1987;75(suppl I):I-48-I-55.
Papademetriou V, Gottdiener JS, Fletcher RD, Freis ED. Echocardiographic assessment by computer-assisted analysis of diastolic left ventricular function and hypertrophy in borderline or mild systemic hypertension. Am J Cardiol. 1985,56:546-550.
National Institutes of Health. Consensus development panel on the health implication of obesity. Ann Intern Med. 1985;103:1073-1077.
de Simone G, Di Lorenzo L, Costantino G, Moccia D, Buonissimo S, de Divitiis O. Supernormal contractility in primary hypertension without left ventricular hypertrophy. Hypertension. 1988;11:457-463.
Sahn DH, DeMaria A, Kisslo J, Weyman A. The Committee on M-Mode Standardization of the American Society of Echocardiography: recommendations regarding quantitation in M-mode echocardiography. Results of a survey of echocardiographic measurements. Circulation. 1978;58:1072-1083.
Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man: anatomic validation of the method. Circulation. 1977;55:613-618.
Gaasch WS, Zile MR, Hosino PK, Apstein CS, Blaustein AS. Stress-shortening relations and myocardial blood flow in compensated and failing canine hearts with pressure-overload hypertrophy. Circulation. 1988;79:872-873.
Thomas JD, Weyman AE. Echocardiographic Doppler evaluation of left ventricular diastolic function. Circulation. 1991;84:977-990.
Gilbert JC, Glantz SA. Determinants of left ventricular filling and of the diastolic pressure-volume relation. Circ Res. 1989;64:827-852.
Su JB, Hittinger L, Laplace M, Crozatier B. Loading determinants of isovolumic pressure fall in closed-chest dogs. Am J Physiol. 1991;260:H690-H697.
Gaasch WH, Carroll JD, Blaustein AS, Bing OHL. Myocardial relaxation: effects of preload on the time course of isovolumic relaxation. Circulation. 1986;73:1037-1041.
Zile MR, Gaasch WH, Wiegner AW, Robinson KG, Bing OH. Mechanical determinants of the rate of isotonic lengthening in rat left ventricular myocardium. Circ Res. 1987;60:815-823.
Starling MR, Montgomery DG, Mancini GBJ, Walsh RA. Load independence of the rate of isovolumic relaxation in men. Circulation. 1987;76:1274-1281.
Thomas JD, Flachskampf FA, Chen C, Guerrero JL, Picard MH, Levine RA, Weyman AE. Isovolumic relaxation time varies predictably with its time constant and aortic and left atrial pressures: implication for the noninvasive evaluation of ventricular relaxation. Am Heart J. 1992;124:1305-1313.
Ariel Y, Gaasch WH, Bogun DK, McMahon TA. Load-dependent relaxation with late-systolic volume steps: servo-pump studies in the intact canine heart. Circulation. 1987;75:1287-1294.
Gillebert TC, Law WYW. Influence of systolic pressure profile on rate of left ventricular pressure fall. Am J Physiol. 1991;261:H805-H813.
Zile MR, Gaasch WH. Load dependent left ventricular relaxation in conscious dogs. Am J Physiol. 1991;261:H691-H699.
Zatko FJ, Martin P, Bahler PC. Time course of systolic loading is an important determinant of ventricular relaxation. Am J Physiol. 1987;21:H461-H466.
Hori M, Kitakaze M, Ishida Y, Fukunami M, Kitabatake A, Inoue M, Kamada T, Yue DT. Delayed end ejection increases isovolumic ventricular relaxation rate in isolated perfused canine hearts. Circ Res. 1991;68:300-308.