Normalization of Diastolic Dysfunction in Aortic Stenosis Late After Valve Replacement
Background The remodeling of the left ventricle in patients with aortic stenosis after aortic valve replacement (AVR) is a complex process involving structural and functional changes.
Methods and Results Twenty-two patients were included in the present analysis. Twelve patients with severe aortic stenosis were studied before surgery, early (22±8 months) and late (81±22 months) after AVR using left ventricular biplane angiograms, high-fidelity pressure measurements, and endomyocardial biopsies. Ten healthy subjects were used as controls. Left ventricular systolic function was assessed from biplane ejection fraction, and diastolic function from the time constant of relaxation, the peak filling rate, and the myocardial stiffness constant. Left ventricular structure was evaluated from interstitial fibrosis, fibrous content, and muscle fiber diameter. Left ventricular muscle mass was significantly increased before surgery in patients with aortic stenosis and remained increased early after surgery, although there was a 35% decrease. Late after AVR, muscle mass decreased significantly but remained slightly (P=NS) elevated. Left ventricular ejection fraction increased slightly after AVR. Left ventricular relaxation was significantly prolonged before surgery and returned toward normal early and late after AVR. Peak filling rates remained unchanged before and after surgery. Myocardial stiffness constant was increased before surgery in patients with aortic stenosis compared with controls and increased even further early after AVR but was normalized late after surgery. Muscle fiber diameter was elevated in patients with aortic stenosis before and after surgery compared with controls; however, it decreased significantly early and late after AVR with respect to preoperative data but remained hypertrophied even late after surgery. Interstitial fibrosis and fibrous contents were larger before surgery than in control subjects and increased even more early but decreased significantly late after AVR.
Conclusions Diastolic stiffness increases in aortic stenosis early after AVR parallel to the increase in interstitial fibrosis, whereas relaxation rate decreases with a reduction in left ventricular muscle mass. Late after AVR, both diastolic stiffness and relaxation are normalized due to the regression of both muscular and nonmuscular tissue. Thus, reversal of diastolic dysfunction in aortic stenosis takes years and is accompanied by a slow regression of interstitial fibrosis.
Left ventricular (LV) hypertrophy is an adaptive process that compensates for pressure overload associated with aortic stenosis.1 2 3 4 5 This process is accompanied by a remodeling of the LV that involves the muscular and nonmuscular compartments of the ventricle. As a result of this remodeling process, muscle fiber hypertrophy and abnormalities of the collagen network occur that are responsible for changes in systolic and diastolic functions.1 5 6 Although it has been previously shown that aortic valve replacement may lead to immediate hemodynamic improvement and to prolongation of survival,7 8 it has been reported that regression of myocardial hypertrophy after relief of the hemodynamic burden is a process that may continue for decades after valve replacement.9 However, abnormal exercise hemodynamics may persist late after valve replacement despite a normal systolic response, suggesting impaired diastolic function in these patients.10
The purpose of the present study was to evaluate in patients with aortic stenosis whether remodeling of the LV after aortic valve replacement is accompanied by similar improvements in myocardial structure and passive elastic properties as it has been reported for the improvement in systolic function and whether these changes are dependent on the duration and time interval after aortic valve replacement.
Twelve patients with pure aortic stenosis (mean age, 44±13; age range, 25 to 65 years) were studied before surgery, early (22±8 months) and late (81±24 months) after successful aortic valve replacement. Ten subjects with normal LV function (mean age, 48±8 years; age range, 33 to 58 years) served as control subjects.
Informed consent was obtained from all patients. Cardiovascular medications were withheld for 24 hours before the procedure. Premedication consisted of 10 mg chlordiazepoxide orally 1 hour before catheterization. Right and left heart catheterizations were carried out in all patients. Biplane left ventriculography was performed in the right anterior oblique (30 degree) and left anterior oblique (60 degree) projections (Cardioscope, Siemens-Albis) at a filming rate of 50 frames per second. LV pressure was measured simultaneously with ventriculography with a Millar 7F micromanometer catheter (Millar Instruments) introduced transseptally into the LV via an 11.5F Brockenbrough guiding catheter. In control subjects, a Millar 8F pigtail micromanometer was introduced retrogradely into the LV. Central aortic pressure was measured through a fluid-filled 8F pigtail catheter in patients with aortic stenosis. All pressures were recorded at a paper speed of 250 mm/s (VR-12, Electronics for Medicine) with a standard lead of the ECG, the first derivative of LV high-fidelity pressure (dP/dt), and the time markers corresponding with the digital numbers on the angiographic images.1 2 5 6 Cardiac index was measured by the Fick method.
LV angiographic silhouettes were drawn manually from an adequately opacified sinus beat, excluding extrasystolic and postextrasystolic beats. LV volumes were determined on a frame-by-frame basis using the area–length method.11 Dimensional and volumetric data were filtered using the moving average technique.1 2 5 6 The LV pressure tracing was digitized for one cardiac cycle with an electronic digitizer (Numonics Corp) interfaced to a computer (PDP-11/34, DEC). Pressure and volume data were analyzed every 20 milliseconds for one cardiac cycle. End diastole was defined as the time point of the rapid upstroke of dP/dt and end systole at the time of incisural pressure in the aortic pressure curve. LV mass was determined according to the method of Rackley et al.12 Circumferential wall stress was calculated from a simplified version of Mirsky’s thick wall model.13 Mean systolic circumferential wall stress was defined as the mean wall stress occurring during the systolic ejection period.
The ratio of the LV end-diastolic volume to mass was calculated and was used as a parameter for assessing chamber geometry of the LV.
Assessment of LV Function
Systolic function was determined from biplane LV ejection fraction, LV end-diastolic pressure, and mean systolic wall stress.
Diastolic function was evaluated from isovolumic relaxation, peak diastolic filling rates, and passive elastic properties. LV relaxation was determined from the time constant of isovolumic pressure decline, which was calculated as the negative reciprocal of the slope of the linear relation between LV pressure and negative dP/dt.2 The isovolumetric relaxation period was defined as the time interval beginning immediately after maximal negative dP/dt and ending when pressure had decreased to 5 mm Hg above LV end-diastolic pressure.2 14 From this time interval, usually 7 to 14 points were available for calculation of the time constant of isovolumic pressure decline in the individual patient.
Peak diastolic filling rate was defined as the largest value of diastolic inflow (mL/m2 per second) during the first half (early) and the second half (late peak filling rate) of diastole. The filling phase was considered to begin 20 milliseconds before the first frame showing the entry of unopacified blood into the LV and to end at end diastole.2 14 Instantaneous diastolic filling rates were calculated every 20 milliseconds after mitral valve opening. To minimize error due to random noise, raw data were filtered using the fifth-grade moving average.14 Diastolic filling rate (FR) was calculated from the following equation: FR(t)=[V(t+20)−V(t−20)]/0.04, where t is time and V is volume.
Diastolic passive elastic properties were determined during the period from minimum ventricular pressure to end diastole.15 LV myocardial properties were evaluated from the diastolic stress–strain relation using an elastic model with shifting asymptote: S=a · ebE+c or dS/dF=b(S–c), where S is LV circumferential wall stress (in kdyne/cm2), a is elastic constant (in kdyne/cm2), b is constant of myocardial stiffness, F is diastolic strain (LaGrangian definition) c is asymptote of the stress–strain relation (in kdyne/cm2), and dS/dF is instantaneous myocardial stiffness (in kdyne/cm2). The three constants a, b, and c were determined by an iteration procedure.1 2 The constant of myocardial stiffness is mathematically represented by the slope of the stress–strain curve, and the tangent to this function is defined as the operative instantaneous myocardial stiffness dS/de.1 2
LV endomyocardial biopsies were performed with the King’s College bioptome (Olympus), which was introduced into the LV through the 11.5F Brockenbrough catheter (USCI).1 2 5 6 In each patient, two to four biopsy samples were obtained from the anterolateral wall of the LV. Immediately after biopsy, specimens were fixed in glutaraldehyde, cut in semithin sections, and evaluated by light microscopy.
Assessment of Cellular Hypertrophy and Interstitial Nonmuscular Tissue
(1) Muscle fiber diameter is the average fiber diameter of at least 100 measurements determined at the level of the nucleus from several randomly chosen cross sections with the use of a mechanical-optical pen (Kontron).
(2) Interstitial nonmuscular tissue (IF) is determined with the point-counting system excluding areas with arterioles and perivascular tissue as previously described.1 5 6 We used the term “interstitial fibrosis” (as did others) for this tissue because fibrous tissue is the predominant component of the interstitial space.1 5 6
Statistical comparisons were carried out with a one-way ANOVA with data from the control group and patients with aortic stenosis before and early as well as late after surgery. If the analysis showed a significant difference, the Scheffé procedure was applied. A one-way ANOVA for repeated measures was used for comparing preoperative, early postoperative, and late postoperative data of patients with aortic stenosis. In all tables, data are given as mean±SD.
Patient characteristics are presented in Table 1⇓. Mean age was similar in the two groups. In patients with aortic stenosis, the aortic valve area was 0.4 cm2/m2 and systolic pressure gradient was 76 mm Hg. Concomitant aortic regurgitation was mild. Cardiac index was, however, significantly reduced in preoperative patients with aortic stenosis compared with control subjects.
After valve replacement, the aortic valve area was 1.8±0.4 cm2 at the early follow-up and 1.7±0.4 cm2 at the late follow-up. The respective systolic pressure gradients were 10±4 and 11±4 mm Hg.
Heart rate was comparable in the different groups before, early, and late after valve replacement compared with the control subjects (Table 2⇓). LV peak systolic and end-diastolic pressures were significantly increased before surgery in patients with aortic stenosis but were normalized early and late after valve replacement. End-diastolic and end-systolic volumes were significantly larger before surgery than in control subjects but were no longer different than those of the control subjects after successful valve replacement. Ejection fraction was slightly reduced in preoperative patients with aortic stenosis and increased early as well as late after surgery. LV wall thickness was significantly increased before surgery and showed a significant regression after valve replacement, although it remained significantly thicker than in control subjects even years after surgery. LV muscle mass was significantly elevated in patients with aortic stenosis and remained increased early after surgery, although there was a 32% decrease in mass. Late after valve replacement, there was a further decrease by 15%. LV end-diastolic volume–to–mass ratio was significantly reduced before surgery as well as early after surgery compared with control subjects but was nearly normalized late after valve replacement.
Diastolic Function Data
LV relaxation was significantly prolonged before surgery and returned toward control levels early and late after valve replacement (Table 3⇓). Early and late peak filling rates were not significantly different in control subjects and in patients with aortic stenosis before and after surgery. Myocardial stiffness constant was doubled in patients with aortic stenosis compared with control subjects and further increased early after valve replacement but was normalized late after surgery (Fig 1⇓).
Muscle fiber diameter was increased in patients with aortic stenosis before and after surgery compared with control subjects; however, it decreased significantly early and late after valve replacement with respect to preoperative data but remained hypertrophied even late after surgery (Table 4⇓). Interstitial fibrosis was larger before surgery than in control subjects and increased even further early after the operation, whereas it decreased significantly late after valve replacement. Fibrous content was significantly higher before as well as early after valve replacement compared with control subjects; however, it decreased significantly late after valve replacement with respect to the preoperative and early postoperative data.
LV relaxation was closely correlated to LV muscle mass (n=34, r=.76, P<.001) and muscle fiber diameter (n=34, r=.57, P<.001) but not to interstitial fibrosis, fibrous content, or fibrous content–to–end-diastolic volume ratio. The constant of myocardial stiffness was correlated to interstitial fibrosis (n=34, r=.76, P<.001), fibrous content (n=34, r=.70, P<.001), and fibrous content–to–end-diastolic volume ratio (n=34, r=.79, P<.001). This index, however, was not correlated to LV mass or muscle fiber diameter.
There has been a dramatic change in the natural history of aortic stenosis after successful valve replacement.16 It has been clearly shown, even in patients with reduced systolic function, that substantial improvement in LV systolic function results from the relief of the hemodynamic burden.17 In children18 and in elderly women,19 a normal end-diastolic volume and a supernormal ejection fraction have been reported that were not, however, found in other studies5 6 or in the present study. The plot of the ejection fraction–afterload relation (Fig 2⇓) clearly shows that only three preoperative patients with aortic stenosis were below the normal shortening–afterload relation. Nine of 12 patients showed a normal systolic ejection fraction, with a mean value of 60%. Despite normal or near-normal LV ejection performance, abnormalities in LV diastolic function have been reported in patients before1 2 3 4 5 and early after5 aortic valve surgery. Therefore, the present study was carried out to determine the effect of aortic valve replacement on LV diastolic function and the effect of regression of LV hypertrophy on myocardial structure.
One of the major findings of the present study was that diastolic stiffness deteriorates early after valve replacement parallel to the relative increase in interstitial fibrosis, whereas relaxation rate improves with the reduction in LV muscle mass. These changes have been explained by a remodeling of the myocardium with a decrease in muscle mass but a “relative” increase in fibrous tissue due to the reduction in the muscular tissue. Late after aortic valve replacement, both diastolic stiffness and relaxation are normalized due to the regression of both muscular and nonmuscular, predominantly collagen, tissue. Thus, reversal of diastolic dysfunction in aortic stenosis takes years and is accompanied by a slow regression of interstitial fibrosis, whereas reversal of systolic dysfunction occurs more rapidly and is associated with the regression of muscular hypertrophy.
LV Structure and Function Early After Valve Replacement
Early after valve replacement (mean, 22±8 months), regression of LV macroscopic (–32%) and microscopic (–12%) hypertrophy has been described as well as a relative increase in interstitial fibrosis due to the regression of the muscular tissue.5 The relative nature of this increase in interstitial fibrosis is supported by the fact that the total amount of fibrous tissue of the LV remains unchanged.5 6 Similar results were observed in a previous study from our group.5 6 These changes in the microscopic structure of the myocardium are associated with alterations in LV diastolic function; the time constant of relaxation is prolonged and the constant of myocardial stiffness is increased in the presence of severe pressure overload of the LV due to aortic stenosis.
It has been demonstrated that relaxation may vary according to the degree of LV hypertrophy,1 14 whereas it is not directly influenced by the nonmuscular collagen tissue.1 In fact, the improvement in relaxation is directly dependent on the regression of LV hypertrophy. Two years after aortic valve replacement, the LV shows still some residual hypertrophy, although LV mass decreased by 35% (Table 2⇑). Because of the incomplete regression of hypertrophy 2 years after valve replacement, relaxation is not completely normalized, although it is improved significantly (Table 3⇑). Murakami and coworkers14 have found a nonsignificant improvement in relaxation early after valve replacement in patients with aortic stenosis. However, their study group included patients with a lower degree of LV hypertrophy and, therefore, a milder impairment of relaxation before aortic valve replacement than our present group and, thus, probably the improvement in relaxation was less after aortic valve replacement because of the lesser regression of LV hypertrophy.
Passive elastic properties are influenced mainly by the nonmuscular components of the myocardium1 20 21 and, at least in part, by the LV geometry.5 22 Thus, in the present study, the increase in myocardial stiffness early after valve replacement is directly related to the augmentation of interstitial nonmuscular tissue, whereas the decrease in stiffness late after surgery is associated with a decrease in interstitial fibrosis. When interstitial fibrosis or the total amount of fibrous tissue divided by the end-diastolic volume is plotted against the constant of myocardial stiffness, a curvilinear or exponential relation is observed (FC/EDVI). Early after valve replacement, an increase in the ratio between fibrous content and end-diastolic volume was observed because there was a reduction in the end-diastolic volume but no change in fibrous content (Table 4⇑). These changes are associated with an increase in the constant of myocardial stiffness. Similar results concerning the early postoperative changes in myocardial stiffness and structure have been reported by Hess and coworkers5 using echocardiography instead of cineangiography to measure LV diastolic function. These authors concluded that structural changes of the LV myocardium with a relative increase in interstitial fibrosis are probably responsible for the increase in myocardial stiffness early after aortic valve replacement. An increase in stiffness became evident in the presence of a fibrous content–to–end-diastolic volume ratio of >0.20 g/mL. In the present study, all patients with aortic stenosis had a fibrous content–to–end-diastolic volume ratio of >0.20 g/mL early after valve replacement, thus explaining the increase in myocardial stiffness after valve surgery.
LV Structure and Function Late After Valve Replacement
It has been reported that regression of angiographic LV mass is a slow process that may continue over 1 decade after aortic valve replacement.9 Late after aortic valve replacement, Krayenbuehl et al6 could not show any further change in muscle fiber diameter after the early postoperative period but could show a significant reduction in the nonmuscular interstitial tissue and LV fibrous content. They also noted a reduction in the fibrous content–to–end-diastolic volume ratio late after operation. These authors speculated that in patients with aortic stenosis “indirect” evidence existed for improvement of before surgery increased LV diastolic myocardial stiffness late after aortic valve replacement.
In the present study, a significant decrease in interstitial fibrosis, fibrous content, and the fibrous content–to–end-diastolic volume ratio was found late after aortic valve replacement (81±22 months). Parallel to the reduction in fibrous tissue, a normalization of the passive elastic properties was found late after valve replacement, which was not the case early after operation. In contrast, the time constant of relaxation was directly related to the regression of LV hypertrophy and was normalized late after aortic valve replacement parallel to the reduction of LV mass.
The rate of regression of LV hypertrophy is different with regard to the muscular and nonmuscular compartments of the LV. This behavior of the different structures of the myocardium influences systolic and diastolic function differently. After a mean time of 81 months following valve replacement, a 43% reduction of global muscular hypertrophy was observed, which occurred mainly (two thirds) within the first 2 years. At the cellular level, the regression of muscle fiber hypertrophy was similar with regard to the time course (two thirds within the first 2 years) but is quite different with regard to the extent of regression of fiber hypertrophy (15%). In contrast, the regression of the nonmuscular tissue is a different process that occurs in two steps: a 69% increase early after aortic valve replacement and a 52% decrease within the following 4 to 5 years after the early phase of valve replacement.
Despite significant reductions in muscular and nonmuscular tissue, 6 to 7 years after valve replacement the process of regression of LV hypertrophy remains incomplete, probably due to the persistence of a certain degree of pressure overload induced by the prosthetic valve.
The possibility of influencing, with pharmacological intervention (eg, angiotensin-converting enzyme or aldosterone inhibitors), the process of remodeling increases the importance of understanding the structure–function interplay in patients with LV hypertrophy—not only in the early but also in the late follow-up of these patients.23 24
One limitation of the present study is the small number of patients included in the present analysis. This could have some influence on the statistical power of variables, such as preoperative LV ejection fraction. However, the study protocol is complicated, including repeated recatheterizations over several years or even decades. The same patient had to undergo three different catheterizations, one or two of which were carried out on an ambulatory basis. All patients gave informed and written consent to the study, and no complications occurred in any of the patients. The family physician of each patient was contacted and informed about the nature of the procedure.
The number of biopsy specimens obtained in a given patient per cardiac catheterization was two to four in the present study. Baandrup and coworkers25 concluded that at least five biopsy samples are necessary to establish correlations between ventricular structure and function in patients with cardiomyopathies. However, in patients with aortic stenosis, morphological changes in the myocardium are uniformly distributed over the entire ventricle,26 so four biopsy samples should adequately reflect myocardial collagen and myocyte size in this selected cohort of patients.1 5 6 Although the relation between function and structure (cause and effect) is not known, our correlations suggest that relaxation is influenced by the extent of hypertrophy,1 whereas passive elastic properties are correlated to the changes in the nonmuscular tissue.1
This work was supported in part by the Swiss National Science Foundation, Berne, Switzerland.
This report is dedicated to Prof Hans-Peter Krayenbuehl, who died in July 1993.
- Received September 8, 1994.
- Revision received November 2, 1994.
- Accepted November 21, 1994.
- Copyright © 1995 by American Heart Association
Hess OM, Villari B, Krayenbuehl HP. Diastolic dysfunction in aortic stenosis. Circulation. 1993;87(suppl IV):IV-73-IV-77.
Hess OM, Ritter M, Schneider J, Grimm J, Turina M, Krayenbuehl HP. Diastolic stiffness and myocardial structure in aortic valve disease before and after valve replacement. Circulation. 1984;69:855-865.
Krayenbuehl HP, Hess OM, Monrad SE, Schneider J, Mall G, Turina M. Left ventricular myocardial structure in aortic valve disease before, intermediate and late after aortic valve replacement. Circulation. 1989;79:744-755.
Kennedy JW, Doces J, Stewart DK. Left ventricular function before and following aortic valve replacement. Circulation. 1977;56:944-950.
Monrad ES, Hess OM, Murakami T, Nonogi H, Corin WJ, Krayenbuehl HP. Time course of regression of left ventricular hypertrophy after aortic valve replacement. Circulation. 1988;77:1345-1355.
Monrad ES, Hess OM, Murakami T, Nonogi H, Corin WJ, Krayenbuehl HP. Abnormal exercise hemodynamics in patients with normal systolic function late after valve replacement. Circulation. 1988;77:613-624.
Rackley CE, Dodge HT, Coble YD, Hay RE. A method for determining left ventricular mass in man. Circulation. 1964;29:666-671.
Mirsky I. Left ventricular stresses in the intact human heart. Biophys J. 1969;9:189-208.
Murakami T, Hess OM, Gage JE, Grimm J, Krayenbuehl HP. Diastolic filling dynamics in patients with aortic stenosis. Circulation. 1986;73:1162-1174.
Gaasch WH, Battle WE, Oboler AA, Banas JS, Levine HJ. Left ventricular stress and compliance in man: with special reference to normalized ventricular function curves. Circulation. 1972;45:746-762.
Smith N, McAnulty JH, Rahimtoola SH. Severe aortic stenosis with impaired left ventricular function and clinical heart failure: results of valve replacement. Circulation. 1978;58:255-264.
Borow KA, Colan SD, Neumann A. Altered left ventricular mechanics in patients with valvular aortic stenosis and coarctation of the aorta: effects on systolic performance and late outcome. Circulation. 1985;72:515-522.
Carroll JD, Carroll EP, Feldman T, Ward DM, Lang RM, McGaughey D, Karp RB. Sex-associated differences in left ventricular function in aortic stenosis of the elderly. Circulation. 1992;86:1099-1107.
Jalil JE, Doering CW, Janicki JS, Pick R, Shroff SG, Weber KT. Fibrillar collagen and myocardial stiffness in the intact hypertrophied rat left ventricle. Circ Res. 1989;64:1041-1050.
Holubarsch C, Holubarsch T. Passive elastic properties in different models and stages of hypertrophy: a study comparing mechanical, chemical and morphometric parameters. Perspect Cardiovasc Dis. 1983;7:323-336.
Carroll JD, Lang RM, Neumann DAL, Borow KM, Rajfer SI. The differential effects of positive inotropic and vasodilator therapy on diastolic properties in patients with congestive cardiomyopathy. Circulation. 1986;74:815-825.
Brilla CG, Janicki JS, Weber KT. Cardioreparative effects of lisinopril in rats with genetic hypertension and left ventricular hypertrophy. Circulation. 1991;83:1771-1779.
Baandrup U, Florio RA, Olsen EJC. Do endomyocardial biopsies represent the morphology of the rest of the myocardium? Eur J Cardiol. 1982;3:171-178.
Schoen FJ, Lawrie GM, Titus JL. Left ventricular cellular hypertrophy in pressure- and volume-overload valvular heart disease. Hum Pathol. 1984;15:171-178.