Pressure and Volume Loading of the Right Ventricle Have Opposite Effects on Left Ventricular Ejection Fraction
Background Left ventricular ejection fraction has been reported to be depressed in patients with right ventricular volume overload (RVVO) due to Ebstein’s anomaly and uncomplicated atrial septal defect, whereas it is usually preserved in right ventricular pressure overload (RVPO) due to congenital pulmonic stenosis. In the present study, we examined the hypothesis that the differential timing of active displacement of the ventricular septum into the left ventricle in RVPO (end systole) and RVVO (end diastole) results in opposite effects of RVPO and RVVO on left ventricular ejection fraction.
Methods and Results Ten patients with severe tricuspid regurgitation after tricuspid valve resection for endocarditis and 10 patients with primary pulmonary hypertension were studied as models of isolated RVVO and RVPO, respectively. Left ventricular ejection fraction, end-diastolic volume, and regional systolic shortening were measured with the use of echocardiographic techniques in these 20 patients and 10 healthy control subjects. In RVPO, despite marked underfilling of the left ventricle relative to the healthy control subjects (end-diastolic volume, 48±26 versus 77±20 mL; P<.02), left ventricular ejection fraction was similar to that of the control subjects (56±5% versus 60±4%; P=.07) and only 1 of 10 RVPO patients had an ejection fraction of less than 50%. In contrast, in RVVO the left ventricle was volume replete (end-diastolic volume, 84±26 versus 77±20 mL; P=NS), but left ventricular ejection fraction was significantly depressed (51±4% versus 60±4%, P<.001) compared with the control subjects, and 4 of 10 RVVO patients had an ejection fraction of less than 50%. Analysis of systolic fractional shortening along two perpendicular short-axis diameters and the mutually orthogonal long axis demonstrated isolated augmentation of fractional shortening in the ventricular septal–to–posterolateral free wall dimension in RVPO (47.4±13.7% versus 34.2±13.1%, P<.05) and isolated depression of fractional shortening along that same dimension in RVVO (13.7±11.8% versus 34.2±13.1%, P<.001) compared with the control subjects.
Conclusions End-systolic leftward ventricular septal shift in RVPO results in isolated augmentation of systolic shortening in the septal–to–free wall dimension, whereas end-diastolic leftward ventricular septal shift in RVVO results in isolated reduction in systolic shortening in the septal–to–free wall dimension. As a result, despite relative underfilling of the left ventricle in RVPO, resting left ventricular ejection fraction is preserved, whereas ejection fraction is depressed for the volume-replete left ventricle of patients with RVVO.
Congenital and acquired diseases of the pulmonary circulation and right heart impose abnormal pressure and volume loads on the right ventricle. Left ventricular filling1 2 3 and left ventricular systolic performance may be secondarily influenced due to the interdependence between the two sides of the heart created by their arrangement as pumps in series, as well as through direct mechanical interactions via the shared ventricular septum and the circumferential myocardial fibers that encircle both chambers.4 Dexter5 first proposed that volume loading of the right ventricle in patients with uncomplicated atrial septal defect might result in impaired filling of the otherwise normal left ventricle by a “reverse Bernheim’s syndrome.” Although some investigators have demonstrated relative reductions in resting left ventricular ejection fraction in patients with atrial septal defect,6 others have found resting left ventricular ejection fraction to be relatively preserved.7 8 Recently, studies in patients with severe tricuspid regurgitation due to Ebstein’s anomaly of the tricuspid valve have demonstrated significant depression in resting left ventricular ejection fraction in patients with right ventricular volume overload (RVVO).9 10 11 12
In patients with right ventricular systolic hypertension due to congenital pulmonic stenosis, left ventricular ejection fraction has been reported to be normal at rest,13 although some investigators have shown that there is a relative decline in left ventricular ejection fraction with progressively severe degrees of right ventricular systolic hypertension.14 These findings are in general agreement with studies demonstrating relative preservation of resting left ventricular ejection fraction in patients with pulmonary arterial hypertension.15 16
Conclusions from the foregoing studies regarding the impact of right ventricular loading on left ventricular ejection fraction are limited by two considerations: (1) the congenital malformations may directly or indirectly be associated with intrinsic abnormalities of left ventricular structure, development, or both, and (2) in many of these patients, right ventricular pressure and volume overload coexist to varying degrees, rendering it difficult to identify the independent effects of isolated right ventricular pressure or volume overload. In the present study, the impact of right ventricular pressure and volume loading on left ventricular ejection fraction was reexamined in two unique populations of patients with relatively pure pressure or volume overload and intrinsically (structurally and developmentally) normal left ventricles. As a model for right ventricular pressure overload (RVPO), patients were selected with severe primary pulmonary hypertension and no more than mild tricuspid regurgitation. This provided the opportunity to study the effect of isolated right ventricular systolic hypertension on the left ventricle of patients, in whom intrinsic left ventricular disease was rigorously excluded. For comparison, patients with severe tricuspid regurgitation due to resection of the tricuspid valve for endocarditis limited to that valve were examined to evaluate the effects of pure RVVO on left ventricular ejection fraction. These two patient populations provided the unique opportunity to compare and contrast the effects of pressure and volume loading of the right ventricle on left ventricular ejection fraction.
Ten patients (18 to 48 years old; mean age, 33±10 years) with unexplained pulmonary hypertension and no more than mild tricuspid regurgitation (as assessed by Doppler echocardiography) underwent a comprehensive evaluation that excluded secondary causes of pulmonary hypertension17 and demonstrated no evidence of primary myocardial, valvular, or coronary artery disease. Each patient underwent simultaneous right heart catheterization and Doppler echocardiographic examination. Mean pulmonary artery pressure was 53±15 mm Hg, mean pulmonary capillary wedge pressure was 8±3 mm Hg, and mean right atrial pressure was 5±5 mm Hg. These patients with primary pulmonary hypertension formed the study group for isolated RVPO.
Ten patients (25 to 45 years old; mean age, 33±8 years) with severe tricuspid regurgitation (as assessed by Doppler echocardiography) due to tricuspid valve excision for isolated involvement of that valve by infectious endocarditis were compared with the patients with RVPO and have previously been described in detail2 ; one patient has been excluded because the echocardiographic study was insufficient to complete all the measurements required for the present study. Surgical exploration and pathological examination demonstrated that all tricuspid valve vegetations were excised and that the underlying valvular tissue was intrinsically normal. There was no evidence in these patients for prior pulmonary, myocardial, valvular, or coronary disease by clinical or echocardiographic examination. With the sternal angle assumed to be 5 cm above the center of the right atrium, mean right atrial pressure was estimated from the vertical height of the meniscus of the internal jugular venous pulsations as 14±4 mm Hg. Invasive hemodynamic monitoring of right heart pressures was not considered to be indicated in these patients in view of their clinical stability. No patient was judged to have significant right ventricular systolic hypertension because the peak systolic velocity of tricuspid regurgitation as assessed by continuous wave Doppler ultrasound was less than or equal to 2 m/s (right ventricle to right atrium peak systolic pressure differential, ≤16 mm Hg), and therefore peak right ventricular systolic pressure was ≤30±4 mm Hg for each patient. These patients with primary tricuspid regurgitation formed the study group for isolated RVVO.
Ten healthy subjects (25 to 47 years old; mean age, 33±8 years) served as a control population for comparison with the patients with RVPO and RVVO. These healthy individuals were free of clinical or Doppler echocardiographic evidence of myocardial or valvular disease. The study had the prior approval of the institutional review board.
Doppler Echocardiographic Examination
Examinations were performed while subjects were in sinus rhythm, breathing quietly in the left lateral recumbent position. Conventional two-dimensional echocardiographic equipment was used to obtain standard (1) parasternal short-axis images of the left ventricle at the papillary muscle and mitral leaflet levels and (2) apical four-chamber images through the crux of the heart. Pulsed Doppler spectral recordings of transmitral flow velocities were obtained with the transducer placed at the left ventricular apex and the Doppler cursor aligned with the left ventricular long axis. The pulsed Doppler sample volume was positioned at the midpoint of the mitral annulus, and minor adjustments in transducer position were made to obtain optimal pulsed Doppler spectra.1 2 3
Calculation of Left Ventricular Volumes and Ejection Fraction
A previously validated18 implementation of the modified Simpson’s rule was used to compute left ventricular end-diastolic and end-systolic volumes, from which left ventricular ejection fraction was derived. The volumetric model uses parasternal short-axis images at the mitral valve and papillary muscle levels to compute the volumes of a stack comprised of a basal cylinder, a truncated cone, and an apical cone.18 The short-axis cavity areas of the left ventricle at mitral valve (Am) and papillary muscle (Ap) levels were planimeterized off-line with an ImageVue Workstation (Nova MicroSonics) along the innermost boundary between the endocardium and blood pool, excluding the papillary muscles. The long axis of the left ventricular cavity (L) was measured from the apical four-chamber view as the distance from the left ventricular apical endocardium to the midpoint of the mitral annulus. Volume (V) at end diastole and end systole was computed as18 : (Am)L/3+(Am+Ap)L/6+(Ap)L/9. Measurements from three representative cardiac cycles were averaged to provide the final measurement for a given subject. As previously reported,19 the average intraobserver error was 5% and the average interobserver error was 12% for these measurements in our laboratory.
Measurement of Peak Transmitral Flow Velocities
Peak transmitral flow velocities of early diastolic filling (VE) and atrial systolic filling (VA) were measured from the midpoint of the envelope of the pulsed Doppler spectrum representing the instantaneous time-varying modal flow velocity.1 2 3 The ratio VA/VE was computed to characterize the relative magnitude of atrial systolic and early diastolic left ventricular filling velocities. Measurements were performed off-line on spectra digitized with the ImageVue Workstation.
Measurement of Left Ventricular Eccentricity and Systolic Fractional Shortening
Eccentricity of the short-axis left ventricular cavity profile (Figure⇓) was assessed from parasternal short-axis images at the papillary muscle level.2 3 The length of the short-axis diameter from the left ventricular septal endocardium to the endocardium of the posterolateral free wall was defined as D1. The length of the orthogonal short-axis diameter between the endocardial surfaces of the anterior and inferior left ventricular free walls was defined as D2. Left ventricular eccentricity at end systole or end diastole was defined as the ratio D2/D1. In control subjects for whom the left ventricular short-axis cavity retains a circular profile throughout the cardiac cycle, D2/D1 remains near unity. As the ventricular septum flattens and is displaced leftward toward the center of the left ventricle due to abnormal right ventricular loading, D1 decreases relative to D2, and D2/D1 becomes progressively more than unity.
Systolic fractional shortening was computed for the two short-axis dimensions D1 and D2 as well as for the left ventricular long axis, L, to provide a measure of regional systolic shortening along these three mutually perpendicular axes of the left ventricle (Figure⇑):
All measurements were performed off-line from digitized two-dimensional echocardiographic images on the ImageVue Workstation.
Data from the three study groups were initially evaluated with a one-way ANOVA. Assuming the critical F statistic was exceeded, pairwise comparisons were performed on the group mean values with a Student’s unpaired t statistic assuming unequal variances. Two-tailed P<.05 values after Bonferroni correction were considered statistically significant. All values are given as mean±1 SD.
Left Ventricular Ejection Fraction
Control subjects had left ventricular ejection fractions ranging from 53% to 66% with a group mean of 60±4%. The left ventricular ejection fraction in patients with RVPO due to primary pulmonary hypertension (56±5%) tended to be lower than the left ventricular ejection fraction in control subjects, but this difference had borderline statistical significance (P=.07). Left ventricular ejection fraction in RVPO ranged from 46% to 66%, and only 1 patient had a left ventricular ejection fraction of less than 50%. In contrast, patients with RVVO, due to tricuspid valve resection, had significantly depressed left ventricular ejection fractions (51±4%) compared with either control subjects (P<.001) or patients with RVPO (P<.05). Left ventricular ejection fraction in RVVO ranged from 46% to 57%. Four of the 10 patients with RVVO had a left ventricular ejection fraction of less than 50%, and 6 of the 10 patients had a left ventricular ejection fraction of less than the lowest left ventricular ejection fraction in the control group.
Left Ventricular End-Diastolic Volume and Pattern of Left Ventricular Filling
In patients with RVPO, the left ventricle was underfilled relative to measurements of left ventricular end-diastolic volume in control subjects (48±26 versus 77±20 mL, P<.02), and left ventricular transmitral filling was decreased during early diastole and shifted to late diastole (VA/VE=1.65±0.98) (Table 1⇓). In comparison, in RVVO left ventricular end-diastolic volume was comparable to that of control subjects (84±26 versus 77±20 mL, P=NS) but late diastolic filling at the time of atrial systole was relatively diminished (VA/VE=0.56±0.20).
Ventricular Interdependence Via the Ventricular Septum
Systolic right ventricular loading in patients with RVPO resulted in shift of the ventricular septum leftward and toward the center of the left ventricle at end systole (left ventricular eccentricity, D2/D1=1.66±0.50) with restoration to a more normal position at end diastole (D2/D1=1.34±0.35) (Figure⇑; Table 2⇓). In contrast, diastolic right ventricular loading in patients with RVVO resulted in shift of the ventricular septum leftward and toward the center of the left ventricle at end diastole (D2/D1=1.34±0.14) with restoration to more normal geometry at end systole (D2/D1=1.08±0.09).
In control subjects, systolic fractional shortening was symmetrical about both orthogonal short axes (Figure⇑: D1 [34.2±13.1%] versus D2 [36.1±11.8%], P=NS). In control subjects, base-to-apex systolic fractional shortening was 16.5±3.2%. In contrast, in patients with RVPO the ventricular septal–to–posterolateral free wall axis, D1, exhibited increased systolic fractional shortening compared with control subjects (47.4±13.7% versus 34.2±13.1%, P<.05). On the other hand, systolic fractional shortening along D2 (35.5±17.6% versus 36.1±11.8%, P=NS) and L (16.1±4.1% versus 16.5±3.2%, P=NS) was not significantly different from control values. In patients with RVVO, the ventricular septal–to–posterolateral free wall axis, D1, demonstrated decreased systolic fractional shortening compared with control subjects (13.7±11.8% versus 34.2±13.1%, P<.002). The systolic fractional shortening along D1 was significantly lower in patients with RVVO than in patients with RVPO (13.7±11.8% versus 47.4±13.7%, P<.001). For patients with RVVO, systolic fractional shortening along D2 (31.6±4.3% versus 36.1±11.8%, P=NS) and L (16.1±3.1% versus 16.5±3.2%, P=NS) was not significantly different from measurements in control subjects.
The results of the present study demonstrate that resting left ventricular ejection fraction is significantly depressed in patients with RVVO compared with in patients with RVPO. Relative to measurements in control subjects, patients with RVPO have preserved left ventricular ejection fraction, whereas patients with RVVO have significantly depressed left ventricular ejection fraction. Compared with control subjects, the absolute depression of left ventricular ejection fraction (9 ejection fraction points) in patients with RVVO is not large; nevertheless, 6 of the 10 RVVO patients had an ejection fraction of less than the lowest ejection fraction in control subjects, and 4 of the 10 RVVO patients had an ejection fraction of less than 50%. In contrast, all RVPO patients except 1 had a left ventricular ejection fraction of more than 50%.
What factors may explain these divergent responses of left ventricular ejection fraction to right ventricular pressure and volume loading? The answer does not appear to lie in the relative degree of overall left ventricular filling. The present study demonstrated that the left ventricle was relatively underfilled in RVPO and volume replete in RVVO and that the left ventricular end-diastolic volume of patients with RVPO was only 57% of the left ventricular end-diastolic volume of patients with RVVO. Thus, in RVPO impairment in early diastolic left ventricular filling with compensatory increases in atrial systolic filling (VA/VE=1.65) does not achieve adequate overall left ventricular filling. In contrast, in RVVO impairment of late diastolic filling (VA/VE=0.56) does not appear to compromise overall left ventricular filling. These observations lead to the unanticipated finding that the underfilled left ventricle found in RVPO has relatively preserved ejection fraction, whereas the volume-replete left ventricle found in RVVO exhibits depressed ejection fraction. The diastolic pressure–volume relations of these patients differed in one important respect: the pericardium was intact in the patients with primary pulmonary hypertension (RVPO), whereas it was left open after surgical excision of the tricuspid valve (RVVO). The absence of pericardial restraint in the patients who underwent tricuspid valvulectomy may explain why the mean left ventricular end-diastolic volume (84±26 mL) was slightly larger than in control subjects (77±20 mL). If the pericardium had been closed surgically at the end of the operation, full distention of the left ventricle might have been compromised, as the volume-overloaded right ventricle competed with the left ventricle for filling within the limited confines of the pericardial space. Under these circumstances, curtailed left ventricular filling might have resulted in further reduction in left ventricular systolic performance.
In the present study, we examined regional systolic shortening and the timing of ventricular septal deformation as potential mechanisms for the divergent responses in left ventricular ejection fraction to right ventricular pressure and volume overload. The analysis of left ventricular eccentricity confirmed previous observations1 2 3 20 that flattening of the ventricular septum resulting from leftward displacement of the septum toward the center of the left ventricle is most marked at end systole in RVPO and at end diastole in RVVO. As a consequence, in RVPO ventricular septal shift actively augments short-axis systolic shortening in the ventricular septal–to–posterolateral free wall dimension, after which the ventricular septum returns to a more normal position (moving rightward away from the center of the left ventricle) at end diastole. Quite the opposite occurs in RVVO, in which the ventricular septum is abnormally displaced leftward toward the center of the left ventricular cavity at end diastole, opposing the normal forces of left ventricular distention. The restoration of normal ventricular septal curvature at end systole results in a component of ventricular septal motion, which opposes the inward motion of the ventricular septum toward the center of the left ventricle during systolic contraction. As a result, the net shortening along the ventricular septum–to–posterolateral free wall short axis in RVVO is depressed relative to shortening along this axis in control subjects.
To evaluate whether the observed differences in ventricular septal–to–posterolateral free wall systolic shortening can explain the significant depression in global left ventricular ejection fraction in RVVO compared with RVPO, it was necessary to compare these findings with the measured systolic shortening in the two other mutually orthogonal axes of the left ventricle. Control subjects demonstrated symmetrical systolic fractional shortening in the short-axis plane. Base-to-apex systolic shortening along the long axis of the left ventricle was less than short-axis fractional shortening in these control subjects. In RVVO, the alterations in systolic fractional shortening were confined to the decrease in septal–to–free wall systolic fractional shortening, whereas systolic fractional shortening along the orthogonal short axis and along the long axis of the left ventricle was not significantly different from that in control subjects. Similarly, in RVPO the alterations in systolic fractional shortening were confined to the increase in septal–to–free wall systolic fractional shortening, whereas systolic fractional shortening along the orthogonal short axis and along the long axis of the left ventricle was within the normal range. These findings strongly suggest that the alterations in ventricular septal–to–posterolateral free wall systolic fractional shortening are mechanistically related to the observed differences in resting left ventricular ejection fraction in RVPO and RVVO. The absence of compensatory alterations in systolic fractional shortening in the other two orthogonal dimensions indicates that septal distortion due to abnormal right ventricular loading does not simply induce a rearrangement in overall left ventricular contraction pattern. Instead, systolic shortening is maintained in the normal range in all dimensions except the septal–free wall dimension. The regional nature of this impairment of systolic function also argues strongly against a systemic factor (eg, loading alteration, neurohumoral interaction, autonomic influence, etc) being the cause for depression of left ventricular ejection fraction in RVVO and preservation of left ventricular ejection fraction in RVPO. If a systemic factor were responsible for these changes in left ventricular ejection fraction, we would have anticipated more uniform changes in systolic fractional shortening in all three of the mutually orthogonal axes. Thus, the underfilled left ventricle in RVPO maintains near-normal resting left ventricular ejection fraction because of enhanced septal–free wall shortening resulting from abnormal end-systolic leftward ventricular septal displacement. In contrast, despite a volume-replete left ventricle, resting ejection fraction is significantly depressed in RVVO as a consequence of impaired septal–free wall systolic shortening due to end-diastolic leftward ventricular septal displacement.
Relation to Previous Studies
A preliminary communication from Handa et al21 identified enhanced ventricular septal–to–left ventricular free wall shortening resulting from abnormal leftward end-systolic ventricular septal shift in patients with primary pulmonary hypertension. Badke22 examined the effects of chronic pulmonary artery constriction on left ventricular systolic shortening in conscious dogs. This study demonstrated that in the chronic state, left ventricular systolic contraction remained normal despite the anticipation that decreases in left ventricular end-diastolic volume might result in reduced systolic fractional shortening. It was concluded that active end-systolic ventricular septal displacement toward the center of the left ventricle in this canine model of RVPO was responsible for the maintenance of the septal contribution to left ventricular contraction and the preservation of global left ventricular ejection. The timing of maximal ventricular septal distortion in RVVO is end diastolic such that the restoring forces, which return the ventricular septum to a nearly normal curvature by end systole, run counter to the normal inward motion of the ventricular septum in systole. Bove and Santamore4 have postulated that such a mechanism might account for the small but significant depression in left ventricular systolic function seen in patients with RVVO.
Although our patients represent the two extremes of abnormal right ventricular pressure and volume loading in individuals with intrinsically normal left ventricles, we cannot entirely exclude potentially confounding variables. A small amount of tricuspid regurgitation was common among the patients with primary pulmonary hypertension, but the patients were selected because they had no more than mild tricuspid regurgitation by Doppler echocardiographic criteria, minimizing the degree of concurrent RVVO. The presence of low mean right atrial pressures in these patients (5 mm Hg) confirms the fact that tricuspid regurgitation was mild and that the vastly predominating perturbation of the right ventricle was pressure overload. The patients who underwent tricuspid valve resection for isolated tricuspid valve endocarditis were relatively young (average age, 33 years) and thus at low risk for occult left ventricular disease. These patients did not undergo preoperative left ventriculography or coronary arteriography but were free of suspected heart disease before the onset of endocarditis as assessed by clinical criteria and Doppler echocardiography. Within these limitations, these two groups of patients represent examples of relatively isolated RVPO or RVVO in individuals with intrinsically normal left ventricles.
The present study demonstrates in patients with intrinsically normal left ventricles (from a structural and developmental standpoint) that RVPO and RVVO have opposite effects on resting left ventricular ejection fraction. The timing of ventricular septal flattening and leftward shift appears to be critical to its impact on ventricular septal–to–left ventricular free wall systolic fractional shortening. This dynamic regional distortion in left ventricular geometry and contraction pattern does not result in compensatory changes in the orthogonal left ventricular dimensions. Instead, the depression of septal–free wall shortening in RVVO appears to have a direct impact on decreasing left ventricular ejection fraction, whereas the augmentation of septal–free wall shortening in RVPO helps to maintain normal left ventricular ejection fraction. These observations help explain the apparent paradox that the underfilled left ventricle in RVPO maintains a near-normal ejection fraction, whereas the volume-replete left ventricle in RVVO exhibits a small but significant depression in resting left ventricular ejection fraction compared with control subjects.
- Received November 3, 1994.
- Revision received January 23, 1995.
- Accepted February 8, 1995.
- Copyright © 1995 by American Heart Association
Dexter L. Atrial septal defect. Br Heart J. 1956;18:209-225.
Bonow RO, Borer JS, Rosing DR, Bacharach SL, Green MV, Kent KM. Left ventricular functional reserve in adult patients with atrial septal defect: pre- and postoperative studies. Circulation. 1981;63:1315-1322.
Benson LN, Child JS, Schwaiger M, Perloff JK, Schelbert HR. Left ventricular geometry and function in adults with Ebstein’s anomaly of the tricuspid valve. Circulation. 1987;75:353-359.
Nootens M, Kaufmann E, Wolfkiel C, Rich S. Right and left ventricular systolic interaction at rest and exercise in pulmonary hypertension. J Am Coll Cardiol. 1994;23(special issue):418A.
Rich S, Dantzker DR, Agres SM, Bergofsky EH, Brundage BH, Detre KM, Fishman AP, Goldring RM, Groves BM, Koerner SK, Levy PC, Reid LM, Vreim CE, Williams GW. Primary pulmonary hypertension: a national prospective study. Ann Intern Med. 1987;107:216-223.
Folland ED, Parisi AF, Moynihan PF, Jones DR, Feldman CL, Tow DE. Assessment of left ventricular ejection fraction and volumes by real-time, two-dimensional echocardiography: a comparison of cineangiographic and radionuclide techniques. Circulation. 1979;60:760-766.
Lin SS, Reynertson SI, Louie EK, Levitsky S. Right ventricular volume overload results in depression of left ventricular ejection fraction: implications for the surgical management of tricuspid valve disease. Circulation. 1994;90(suppl II):II-209-II-213.
Handa S, Miyamori R, Ohnishi S, Akaishi M, Ogawa S, Yamazaki H, Nakamura Y, Yokoyama T. Left ventricular geometry and contraction in patients with primary pulmonary hypertension. Circulation. 1985;72(suppl III):III-411.
Badke FR. Left ventricular dimensions and function during right ventricular pressure overload. Am J Physiol. 1982;242(Heart Circ Physiol. 11):H611-H618.