Right Ventricular Diastolic Dysfunction in Heart Failure
Background Left ventricular (LV) diastolic dysfunction is common in heart failure and is an important predictor of prognosis and mortality. Less attention has been paid to right ventricular (RV) diastolic function. In this study, we compared RV diastolic function in a large cohort of patients with heart failure (HF) with two groups: patients with pulmonary hypertension and normal LV function (the PHT group) and normal subjects.
Methods and Results Transtricuspid and pulmonary artery flow were assessed by two-dimensional Doppler echocardiography at maximum inspiration and expiration in 185 subjects: 114 symptomatic HF patients (ejection fraction <0.5), 31 PHT patients (pulmonary artery systolic pressure >40 mm Hg), and 40 normal subjects. A subset was matched for age and heart rate. The results showed a high prevalence of RV diastolic abnormalities: HF patients had lower tricuspid E-A ratios, lower peak E-wave velocity, and prolonged RV isovolumic relaxation time (all P<.0001). Tricuspid E-wave deceleration time was significantly shorter only in those who had an LV restrictive filling pattern. The PHT group had similar findings. Compared with a normal range, more than half of the patients had lower tricuspid E-A ratios (HF, 55%; PHT, 69%), and 61% of HF and 58% of PHT patients had a prolonged RV isovolumic relaxation time. In the PHT group, RV diastolic parameters (E-wave deceleration time, E-A ratio, and isovolumic relaxation time) correlated significantly with pulmonary artery systolic pressure (P<.05). In the HF group, however, only tricuspid E-wave deceleration time correlated significantly with pulmonary artery systolic pressure, and HF patients with normal pulmonary artery systolic pressures had significantly lower tricuspid E-A ratios and prolonged RV isovolumic relaxation times compared with normal subjects. A close correlation existed between individual RV and LV diastolic parameters, suggesting that LV diastolic dysfunction may directly affect RV function, but there was no relation between LV size or systolic function and RV diastolic dysfunction.
Conclusions RV diastolic function is frequently abnormal in HF patients, and this is not related to elevated pulmonary artery systolic pressure alone, although high pulmonary artery pressure by itself also is associated with impaired RV diastolic function. Assessment of the role of right ventricular diastolic function in determining the symptoms and prognosis of heart failure is warranted.
Left ventricular diastolic dysfunction is common in patients with congestive HF and appears to be an important predictor of prognosis and mortality.1 2 3 By use of combined Doppler transmitral flow velocities and measurement of the isovolumic relaxation time, three patterns of LV diastolic dysfunction have been described: ARP, RFP, and a pseudonormal pattern.4 5 Recent data suggest that the RFP is the best clinical predictor of cardiac death in a group of patients with congestive HF and dilated cardiomyopathy.1 2 3 However, there is little information about RV diastolic dysfunction. Many cardiac diseases affect both the left and the right ventricles, and LV failure may secondarily impair RV diastolic performance through elevation of the pulmonary artery pressure or ventricular interdependence.6 Therefore, the aim of this study was to assess the feasibility of measuring RV diastolic function by Doppler echocardiography in a group of patients with congestive HF and relating the findings to PASP and LV diastolic function. A second group of patients with PHT and normal LV function was studied to ascertain the separate effect of increased PASP.
One hundred and eighty-five subjects (mean age, 58.69±1.17 years; 58% men) were recruited for the study. They formed three groups: patients with typical congestive heart failure (the HF group), those with pulmonary hypertension with normal LV function (the PHT group), and normal control subjects.
One hundred fourteen consecutive patients with symptomatic HF (mean age, 62.7±1.23 years; 66% men) were studied. The cause of HF was idiopathic dilated cardiomyopathy in 42 patients (36.8%), ischemic heart disease in 59 (51.8%), hypertensive heart disease in 8 (7.0%), and aortic valve disease (aortic regurgitation in 3 and mixed stenosis and regurgitation in 2) with severe LV dysfunction in 5 (4.4%). Idiopathic dilated cardiomyopathy was diagnosed if there was no clear cause and ejection fraction was <50%. Coronary angiography and endomyocardial biopsy were not considered mandatory and were performed in 60% and 43% of patients, respectively. Patients with ischemic heart disease had either a history of myocardial infarction or severe coronary artery disease on arteriogram, ejection fractions <50%, and LV enlargement on the echocardiogram (end-diastolic dimension >5.6 cm on long-axis view by M-mode echocardiography).
Thirty-one consecutive patients (mean age, 55±3.35 years; 48% men) with PHT were studied. PASP was >40 mm Hg by continuous-wave Doppler echocardiography with normal LV systolic function (LV ejection fraction >55%). The cause of PHT was chronic obstructive airways disease in 19 patients (61.3%), valvular heart disease in 6 (aortic regurgitation in 3, mitral regurgitation in 3; 19.3%), congenital heart disease in 3 (atrial septal defects in 2, ventricular septal defect in 1; 9.7%), systemic lupus erythematosus in 2 (6.5%), and idiopathic PHT in 1 (3.2%).
Normal Control Subjects
Forty normal subjects (mean age, 50.2±2.66 years; 45% men) free of heart or lung disease were selected as control subjects.
Doppler Echocardiographic Examination
Two-dimensional echocardiography with continuous- and pulse-wave Doppler studies was performed. Standard M-mode measurements of LV systolic function were performed. Pulse-wave Doppler echocardiography was performed for both the left and right ventricles. In the left ventricle, mitral in-flow velocities in the apical four-chamber view with the sampling window placed at the mitral annulus (to standardize measurements) were recorded. Diastolic parameters were measured for at least three beats. These parameters included peak early mitral valve filling velocity (E wave), peak atrial filling velocity (A wave), E-A ratio, and DT. IVRT was measured by moving the sampling window to a position between the anterior mitral leaflet and LV outflow tract. The LV diastolic mitral flow pattern was divided into a normal pattern, ARP, and RFP as previously described. ARP was characterized by prolonged DT (>240 ms), reversed E-A ratio, and prolonged IVRT (>100 ms); RFP was characterized by a short DT (<140 ms), large E-A ratio (>2), and short IVRT (<70 ms).2 4
RV diastolic function was assessed with the parasternal short-axis view at the level of the tricuspid valve. The sampling window was placed at the tricuspid annulus. RV diastolic parameters corresponding to those of the left ventricle were measured. The RV-IVRT was defined as the time interval between the closure of the pulmonary valve and opening of the tricuspid valve. This was estimated by subtracting the time interval between the peak of the R wave on the ECG and the onset of the tricuspid valve opening from the interval between the peak of the R wave and the end of the pulmonary systolic flow profile. Because the RV diastolic filling is affected by respiration, measurement of beats was timed with respiration. At least three beats from the end inspiration and three beats from the end expiration were recorded, and their values were averaged. PASP was estimated by continuous-wave Doppler echocardiograms recorded in the apical four-chamber view as the peak systolic pressure gradient across the tricuspid valve (peak regurgitation velocity) plus the estimated right atrial pressure (10 mm Hg).
The echocardiographic data between different diagnostic groups and subgroups and between ARP and RFP were compared by use of an unpaired t test or ANOVA as appropriate. The correlation of PASP with RV diastolic parameters was examined by multiple linear regression. The relationship between LV and RV diastolic parameters was examined by correlation analysis. All data are expressed as mean±SEM. A value of P<.05 was considered statistically significant.
HF and PHT Groups Versus Normal Control Subjects
There was a significant difference in the RV diastolic parameters between the two patient groups and control subjects. Patients with HF had lower tricuspid E-A ratios (1.00±0.03 versus 1.26±0.04, P<.0001), lower tricuspid peak E-wave velocity (43.73±0.97 versus 51.53±1.73 cm/s, P=.0001), and longer RV-IVRT (97.31±4.37 versus 54.61±3.55 ms, P<.0001). Similar results were observed in the group with PHT with reversal of the E-A ratio (P<.0001), lower tricuspid peak E-wave velocity (P=.0006), higher tricuspid peak A-wave velocity (P=.018), and longer RV-IVRT (P<.0001), but tricuspid DT was nonsignificantly shorter (P=.06; Table 1⇓). A normal range of each individual RV diastolic parameter was obtained by use of the mean±SD of normal control subjects: tricuspid peak E-wave velocity, 52±10 cm/s; tricuspid peak A-wave velocity, 42±10 cm/s; tricuspid E-A ratio, 1.26±0.27; tricuspid DT, 187±44 ms; and RV-IVRT, 55±21 ms. There was a high incidence of RV diastolic dysfunction in patients with either HF or PHT (Table 2⇓). More than half of the patients had lower E-A ratios (HF, 55.4%; PHT, 69%), whereas 36.0% of the HF and 26.7% of the PHT patients had shortened tricuspid DTs.
Heart rate was significantly higher in the PHT (79.8±3.0 bpm) and HF (82.1±1.6 bpm) groups than in normal control subjects (70.0±1.7 bpm, both P<.01). However, when groups of normal subjects and HF patients (28 from each group) matched for heart rate (mean, 70.5±1.9 and 70.5±1.9 bpm, respectively; P=.996) and age (mean, 54.0±3.1 and 56.2±2.8 years, respectively; P=.60) were compared, the HF patients still had significantly lower tricuspid peak E-wave velocity (44.1±1.8 versus 52.2±1.9 cm/s, P=.003), lower tricuspid E-A ratio (1.06±0.07 versus 1.25±0.04, P=.022), and longer RV-IVRT (99.7±7.4 versus 55.5±4.1 ms, P<.0001), but no significant difference was seen in PASP (30.4±2.6 versus 27.7±1.6 mm Hg, P=.39). In a similar comparison of patients with PHT and normal control subjects (15 in each group) matched for heart rate (mean, 71.9±2.6 and 73.0±2.9 bpm, respectively; P=.77) and age (mean, 50.0±4.7 and 55.4±4.7 years, respectively; P=.43), tricuspid peak E-wave velocity and E-A ratio were lower in the PHT group (E-wave velocity, 43.2±3.6 versus 52.8±2.5 cm/s, P=.035; E-A ratio, 0.85±0.04 versus 1.18±0.08, P=.001), and RV-IVRT was longer (90.1±10.7 versus 53.7±5.0 ms, P=.003). DT was not significantly different.
Correlation of RV Diastolic Parameters With PASP by Multiple Linear Regression
When the individual RV diastolic parameters of all patients were correlated with PASP by multiple regression (Table 3⇓), there was a significant negative correlation of PASP with DT (r=− .453, P<.0001) and tricuspid E-A ratio (r=−.341, P<.0001; Figs 1⇓ and 2⇓). However, among the HF patients, only the tricuspid DT correlated significantly with PASP (r=−.476, P<.0001). On the other hand, in the PHT group, PASP correlated with not only DT (r=− 0.593, P=.0082) but also tricuspid E-A ratio (r=−.628, P=.0023) and RV-IVRT (r=.123, P=.0360).
HF Group: Normal Versus High PASPs
Patients with HF who had normal PASP also had significant differences in most RV diastolic parameters compared with normal control subjects (Table 4⇓). In addition, there were differences in the RV diastolic parameters between HF patients with normal and high PASPs (Table 4⇓). In particular, the DT in HF patients with normal PASP was similar to that of normal subjects but was significantly reduced in HF patients with high PASP. There also was a significantly greater prolongation of the RV-IVRT in HF patients with high PASP, indicating that the elevated PASP per se affects predominantly the DT and the RV-IVRT.
Dilated Cardiomyopathy Versus Ischemic Heart Disease With Anterior Myocardial Infarction Patients
Forty patients with idiopathic dilated cardiomyopathies were compared with 41 patients with HF caused by a previous large anterior myocardial infarction (Table 5⇓). There was no difference in either RV or LV diastolic parameters in these two groups.
Relation of RV Diastolic Function to LV Diastolic Function, LV Size, and LV Systolic Function in the HF Group
There was a significant correlation between individual RV and LV diastolic parameters: E-wave velocity, r=.216, P=.003; A-wave velocity, r=.289, P<.001; E-A ratio, r=.214, P=.007; DT, r=.519, P<.001; and IVRT, r=.230, P=.004. When the HF patients were classified according to the different patterns of LV diastolic dysfunction, 34 had ARP, 69 had RFP, and 11 had a normal pattern (possibly pseudonormal but pulmonary vein flow velocities were not measured). Compared with control subjects, there were significant differences in most of the RV diastolic parameters in both LV-ARP and LV-RFP subgroups (Table 1⇑). When the patients with normal PASPs in these two groups were selected for comparison, differences were still apparent in tricuspid peak A-wave velocity (LV-ARP versus LV-RFP, 45.11±1.71 versus 39.77±1.56 cm/s, P=.025), tricuspid E-A ratio (0.95±0.05 versus 1.14±0.06, P=.012), and DT (227.98±11.11 versus 165.56±7.17 ms, P<.0001; Table 6⇓). In addition, to ascertain the effect of LV size on RV diastolic function, multiple regression analysis was done between LV end-diastolic dimension (the dependent variable) and RV-IVRT, tricuspid peak A-wave velocity, peak E-wave velocity, E-A ratio, and DT; there were no significant relations (P=.53, P=.59, P=.55, P=.79, and P=.86, respectively). Similarly, LV end-diastolic dimension was similar (P=.98) in those patients with normal (6.4±0.23 cm) or abnormal (6.4±0.11) RV diastolic function. Similarly, there was no significant association between RV diastolic parameters (either individual values or patterns) and LV fractional shortening or LV ejection fraction. Thus, RV diastolic dysfunction does not correlate with either LV size or LV systolic function.
This is the first published study that has systematically assessed RV diastolic function in HF patients and compared it with that in a group of patients with PHT (and normal LV function) and normal control subjects. In the past, a reluctance to use transtricuspid flow velocities for the assessment of RV diastolic function may have been due in part to the well-recognized observation that tricuspid flow depends significantly on respiration.7 Uiterwaal et al8 showed that the maximum velocities of the tricuspid E and A waves were significantly higher during inspiration than during expiration when the flow was measured on the atrial side. When measured on the ventricular side of the tricuspid valve, respiration was found to affect only the maximum velocity of the E wave. The tricuspid E-A ratio was not greatly affected by respiration, and the authors concluded that information about the respiratory cycle is not relevant for this particular measurement. It is not clear if measurements should be taken at end inspiration or end expiration. In a recent study of RV diastolic performance after complete repair of the tetralogy of Fallot, Cullen et al9 measured Doppler tricuspid flow at end inspiration and end expiration, and although there were small differences, they were not significant for peak E- and peak A-wave velocities. Furthermore, the effect of restrictive physiology was apparent in measurements recorded at both end inspiration and end expiration.9 Thus, average values would suffice. We found that by averaging values measured at end expiration and end inspiration, useful, clinically relevant information can be derived that separates those patients with clearly abnormal RV diastolic function from normal subjects. In contrast to transmitral flow on which age has an important effect, Pye et al10 found that there was no significant correlation between any tricuspid flow parameters and age, although two smaller studies found a weak correlation.11 12 Berman et al12 also found that tricuspid late velocity and atrial filling fraction were each modestly inversely related to the RR interval. Our HF and PHT patient groups both had higher resting heart rates than normal subjects, but a subset matched for heart rate (and age) had similar statistically significant abnormalities, confirming that neither heart rate nor age can account entirely for the differences in RV diastolic function.
Few studies have been done on RV function in HF, and its role has probably been underestimated. Polak et al13 found that patients with an RV ejection fraction <35% at rest had a significantly higher 2-year mortality. Recently, Di Salvo et al14 in the largest study to date concluded that RV ejection fraction >0.35 was a more potent predictor of survival in advanced HF than Vo2. In that study, LV ejection fraction at rest was not predictive of overall or event-free survival in any univariate or multivariate analysis.14 However, the only study in the literature that has assessed RV diastolic function in HF patients is a report by Riggs15 of 6 children with dilated cardiomyopathy, all of whom had abnormal RV filling. In our study, we confirmed that RV diastolic dysfunction is common, with prolonged IVRT occurring in nearly 60% of patients and reversed tricuspid E-A ratio occurring in 55% of patients with HF. Because LV diastolic function, particularly the restrictive filling pattern, has been shown to provide important prognostic information,1 2 3 it is possible that RV diastolic dysfunction is equally important. Follow-up of this group of patients will determine whether additional information provided by assessment of RV diastolic function will increase the accuracy of predictions of outcome and prognosis.
We have confirmed that whatever the cause, RV diastolic function is impaired in PHT patients. In a group of patients with chronic obstructive lung disease, Marangoni et al16 similarly found a good correlation between PASP and indexes of diastolic function. Obviously, one mechanism for impaired RV diastolic function in HF patients is PHT secondary to increased left atrial pressure. Both the tricuspid E-A ratio and DT correlated significantly with PASP for all subjects. In the present study, the HF and PHT groups had similar abnormalities of RV diastolic function. The DT tended to be shorter in the PHT group (P=NS), and this was the only parameter that correlated with PASP in the HF group. In the group of patients with HF and normal PASP, however, abnormalities of RV diastolic function were still common. It would appear, therefore, that there is a separate mechanism independent of PASP. Our study does not show precisely what this may be. However, we were unable to demonstrate any significant difference between patients with idiopathic dilated cardiomyopathy (who may have impaired RV function as part of the disease process) and patients with HF caused by large anterior myocardial infarctions in whom RV systolic function is normally preserved,17 although RV disease with myocyte loss and collagen accumulation is found in those patients with severe three-vessel disease and ischemic cardiomyopathy.18 Thus, it would appear that there may be other factors involved apart from the absence or presence of RV disease. It is clear that the anatomic and functional integrity of each ventricle is important determinant of the functional characteristics of the other ventricle.6 Although the effects of PHT and RV dilation on LV function have been studied,19 less attention has been paid to the effect of LV diastolic dysfunction on the right ventricle. An increased distention of either ventricle during diastole has been shown to alter the compliance and geometry of the opposite ventricle by either a direct mechanical effect (displacement of the septum) or some other indirect process.20 21 22 The right and left ventricles share common muscle bundles, septum, and pericardium.23 24 The mechanisms of ventricular interaction are unknown but may relate to restriction of ventricular filling by the pericardium,25 although most work has assessed only the effect of RV volume expansion on LV function26 rather than vice versa. Furthermore, the influence of volume changes in one ventricle on the filling dynamics of the other has not been studied. We found some correlation between individual RV diastolic parameters with those of the LV, suggesting that disordered mechanics of filling in one ventricle can directly affect the other, but this is unlikely to be the entire explanation.
In summary, we have demonstrated that RV diastolic dysfunction is a common feature in patients with HF and PHT with normal LV function. Second, we have shown that patients with HF with normal PASPs also have impaired RV diastolic dysfunction. This may be due in part to the disease process, such as in patients with dilated or ischemic cardiomyopathy, but also may be caused indirectly by coexistent LV diastolic dysfunction resulting from ventricular interdependence. It is possible that RV diastolic dysfunction may be equally important in determining symptoms and prognosis. Follow-up of this group of patients will help to answer this question.
Selected Abbreviations and Acronyms
|ARP||=||abnormal relaxation pattern|
|bpm||=||beats per minute|
|DT||=||deceleration time of the tricuspid or mitral E wave|
|IVRT||=||isovolumic relaxation time|
|PASP||=||pulmonary artery systolic pressure|
|RFP||=||restrictive filling pattern|
- Received September 18, 1995.
- Revision received November 2, 1995.
- Accepted November 5, 1995.
- Copyright © 1996 by American Heart Association
Pinamonti B, Di Lenarda A, Sinagra G, Camerini F, for the Heart Muscle Disease Study Group. Restrictive left ventricular filling pattern in dilated cardiomyopathy assessed by Doppler echocardiography: clinical, echocardiographic and hemodynamic correlations and prognostic implications. J Am Coll Cardiol. 1993;22:808-815.
Rihal CS, Nishimura RA, Hatle LK, Bailey KR, Tajik AJ. Systolic and diastolic dysfunction in patients with clinical diagnosis of dilated cardiomyopathy: relation to symptoms and prognosis. Circulation. 1994;90:2772-2779.
Uiterwaal C, Van Dam I, De Boo T, Van Keulen P, Folgering H, Hopman J, Daniels O. The effect of respiration on diastolic blood flow velocities in the human heart. Eur Heart J. 1989;10:108-112.
Cullen S, Shore D, Redington A. Characterization of right ventricular diastolic performance after complete repair of tetralogy of Fallot: restrictive physiology predicts slow postoperative recovery. Circulation. 1995;91:1782-1789.
Marangoni S, Scalvini S, Schena M, Vitacca M, Quadri A, Levi G. Right ventricular diastolic function in chronic obstructive lung disease. Eur Respir J. 1992;5:438-443.
Tobinick E, Schelbert HR, Henning H, LeWinter M, Taylor A, Ashburn W, Karliner JS. Right ventricular ejection fraction in patients with acute anterior and inferior myocardial infarction assessed by radionucleotide angiography. Circulation. 1978;57:1078-1084.
Beltrami CA, Finato N, Rocco M, Feruglio GA, Puricelli C, Cigola E, Quaini F, Sonnenblick EH, Olivelti G, Anversa P. Structural basis of end-stage failure in ischemic cardiomyopathy in humans. Circulation. 1994;89:151-163.
Stojnic BB, Brecker SJ, Xiao HB, Helmy SM, Mbaissouroum M, Gibson DG. Left ventricular filling characteristics in pulmonary hypertension: a new mode of ventricular interaction. Br Heart J. 1992;68:16-20.
Laks MM, Garner D, Swan HJC. Volumes and compliance measured simultaneously in the right and left ventricles of the dog. Circ Res. 1967;20:565-569.
Santamore WP, Lynch PR, Meiner GM, Heckman J, Bove AA. Myocardial interaction between the ventricles. J Appl Physiol. 1976;41:362-368.
Taylor RR, Covell JW, Sonnenblick EH, Ross J Jr. Dependence of ventricular distensibility on filling of the opposite ventricle. Am J Physiol. 1967;213:711-718.
Bernard D, Alpert JS. Right ventricular function in health and in disease. Curr Probl Cardiol. 1987;13:423-449.
Janicki JS. Influence of the pericardium and ventricular interdependence on left ventricular diastolic and systolic function in patients with heart failure. Circulation. 1990;81(suppl III):III-15-III-20.