Utility of Preload Alteration in Assessment of Left Ventricular Filling Pressure by Doppler Echocardiography: A Simultaneous Catheterization and Doppler Echocardiographic Study
Author + information
- David G. Hurrell, MDA,1,1,
- Rick A. Nishimura, MD, FACCA,*,
- Duane M. Ilstrup, MSA and
- Christopher P. Appleton, MD, FACCB
- ↵*Dr. Rick A. Nishimura, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Abstract
Objectives. The aim of this study was to demonstrate the usefulness of preload alterations in assessing left ventricular filling pressures with transmitral Doppler velocity curves.
Background. Doppler mitral inflow velocities, used to estimate left ventricular filling pressures noninvasively, are limited in predicting left ventricular filling pressures, especially in patients with normal systolic function and a “pseudonormal” mitral filling pattern.
Methods. Forty-nine patients were studied in the cardiac catheterization laboratory with simultaneous Doppler echocardiography using high fidelity catheters to compare left ventricular diastolic filling pressures (pre–A wave left ventricular pressure) and Doppler mitral inflow at baseline and during reduction of preload during the strain phase of the Valsalva maneuver (n = 27) or sublingual nitroglycerin (n = 36), or both (n = 14). Doppler measurements consisted of E (initial peak velocity), A (velocity at atrial contraction), deceleration time (time from E velocity to deceleration of flow extrapolated to baseline) and absolute A wave velocity (A′ [peak A wave velocity minus velocity at onset of atrial contraction]).
Results. In patients with high pre–A wave pressure (≥15 mm Hg), there was a greater change in the E/A′ ratio during the Valsalva maneuver than in patients with a normal pre–A wave pressure (−1.22 ± 1.1 vs. −0.35 ± 0.17; p = 0.02). A similar change was seen when comparing the change in the E/A′ ratio after administration of nitroglycerin in patients with a high versus a normal pre–A wave pressure (0.81 ± 0.49 vs. 0.18 ± 0.17; p < 0.001). These differences were present in patients with a normal E/A ratio at baseline.
Conclusions. Alterations in preload during assessment of Doppler echocardiographic indexes may be useful in noninvasively assessing left ventricular filling pressures.
Knowledge of left ventricular filling pressures is important for the diagnosis, prognosis and treatment of patients with cardiac disease. Pulsed wave Doppler transmitral flow velocities have been proposed as a noninvasive method for examining diastolic filling of the left ventricle, and thus have been used in an attempt to determine left ventricular filling pressures. In patients with a decreased ejection fraction, the ratio of early transmitral flow velocity to atrial flow velocity (E/A) and deceleration time accurately predict left ventricular filling pressures [1–7]. As left ventricular filling pressure increases, deceleration time decreases and the E/A ratio increases. However, in patients with normal or borderline systolic function, transmitral Doppler indexes are of less value [4]. Analysis of pulmonary venous flow velocity curves from pulsed wave Doppler echocardiography may provide information in addition to that of mitral inflow indexes, but these curves may be difficult to acquire in all patients.
Early stages of diastolic dysfunction are seen in the transmitral Doppler indexes by a decline in E wave velocity and prolongation of deceleration time because of impaired or slowed left ventricular relaxation. With disease progression, left ventricular compliance decreases, which results in increased left atrial pressure, increased E wave velocity and decreased deceleration time, thus mimicking a normal filling pattern. This “pseudonormal” filling pattern with increased left ventricular filling pressure is difficult to distinguish from normal transmitral indexes and normal left ventricular filling pressures. Alterations in loading conditions change these transmitral filling patterns in a predictable manner [8–22]. Thus, it is hypothesized that changes in the mitral flow velocities caused by reductions in preload may be useful in distinguishing patients with a pseudonormal pattern from those with normal mitral filling pattern and in assessing left ventricular filling pressures.
1 Methods
The total study group consisted of 49 patients (28 men and 21 women; mean [±SD] age 67 ± 11 years) who underwent simultaneous left ventricular cardiac catheterization with high-fidelity micromanometer catheters and transmitral Doppler echocardiographic assessment. All patients were hemodynamically stable and in normal sinus rhythm. Diagnoses included coronary artery disease in 36 patients and idiopathic dilated cardiomyopathy in four patients; the other nine patients had angiographically normal coronary arteries and normal left ventricular function. The mean ejection fraction was 58 ± 17%. This study was approved by the Institutional Review Board of the Mayo Foundation, and informed consent was obtained from all participants.
1.1 Cardiac catheterization technique.
Cardiac catheterization was performed with the patient in the fasting state. High fidelity micromanometer catheters (Millar Instruments) were used to obtain left-sided heart pressure waveforms and balanced to fluid-filled systems, as described previously [23, 24]. All recordings were performed before the administration of contrast dye. Measured pressure recordings included left ventricular systolic pressure, left ventricular end-diastolic pressure and left ventricular pre–A wave pressure, which is defined as the pressure at the onset of atrial contraction. The pre–A wave pressure was measured to provide an estimate of mean left atrial pressure (Fig. 1) [5].
Transmitral Doppler echocardiography performed with simultaneous cardiac catheterization using a high fidelity micromanometer catheter. Pulsed wave Doppler trace:E = peak E wave velocity; A = A wave velocity; E at A = velocity at atrial contraction; DT = mitral deceleration time; and A′ = absolute A wave velocity. Left ventricular pressure trace(LV): Pre-A = left ventricular diastolic pressure before atrial contraction.
1.2 Doppler analysis.
Doppler echocardiography was performed during cardiac catheterization, with the patient in the supine position. The Doppler signals were recorded on a hard copy strip chart at 100 mm/s, with a simultaneous left ventricular pressure recording. To record mitral flow velocity, a pulsed wave Doppler sample volume (1 to 2 mm) was placed between the mitral leaflet tips and aligned parallel with the inflow, as determined with color flow Doppler echocardiography from the apical window. Continuous wave Doppler echocardiography was performed during the Valsalva maneuver using a visually directed transducer (2.5 MHz) in three patients in whom adequate pulsed wave spectral analysis could not be obtained. The Doppler variables that were measured included the peak E wave velocity and the peak A wave velocity (Fig. 1). Deceleration time was taken as the time from peak E wave velocity to an extrapolation of the deceleration of the flow to baseline. From these variables, the absolute A wave (A′) velocity was calculated as the peak A wave velocity minus the E wave velocity at onset of atrial contraction to account for the effects of heart rate response [25]. The E/A′ ratio was calculated from these data.
1.3 Preload reduction.
Before insertion of the catheters, the patient was taught the Valsalva maneuver (straining against a closed glottis). After the catheters were inserted, baseline pressure waveforms from the high fidelity catheters as well as simultaneously recorded Doppler indexes were measured at end-expiration. The patient was then told to begin the Valsalva maneuver and was instructed to release it at 12 s. During this time, simultaneous left ventricular pressures and mitral inflow Doppler recordings were obtained. An adequate Valsalva maneuver was defined as a 10% reduction in maximal E wave velocity from baseline [14]. Three separate attempts were performed, and the best recordings were used for analysis. Doppler echocardiographic data could not be obtained in 16 patients because of an inadequate Valsalva maneuver in one (as measured by Doppler criteria), Doppler recordings of poor quality in 10 and a clinical condition in five (i.e., presumed acute myocardial infarction). Twenty-seven patients were included in the analysis.
After performing the Valsalva maneuver, 40 patients received 800 μg of nitroglycerin sublingually. We excluded patients who were receiving nitroglycerin intravenously at the time of arrival in the cardiac catheterization laboratory [9]or who had a left ventricular systolic blood pressure <110 mm Hg [4]. After 5 min of observation, the left ventricular pressure and Doppler recordings were repeated. An adequate response was defined as a 15% decrease in the pre–A wave left ventricular pressure. Four patients did not meet these criteria and were excluded from the analysis. Measurements for both cardiac catheterization and Doppler echocardiography were repeated, as defined earlier.
1.4 Statistics.
Data are expressed as mean value ± SD. A priori, it was decided to evaluate continuous variables separately before versus after an intervention for the patients with normal and abnormal pre–A wave pressures, with normal pressures being defined as pre–A wave pressure <15 mm Hg [26]. Thus, comparison of continuous variables before versus after an intervention was made by a repeated measures design, using a paired ttest for those with a gaussian distribution and the Wilcoxon signed-rank test for those with a nongaussian distribution. To determine the difference in the change in E/A ratio and E/A′ ratio induced by an intervention between patients with normal and abnormal pre–A wave pressures, an unpaired two-sample ttest was used for a gaussian distribution and the rank-sum test was used for a nongaussian distribution. Statistical significance was defined as p < 0.05. Correlation between like variables after the administration of nitroglycerin or during the Valsalva maneuver was determined using linear regression analysis.
2 Results
Forty-nine patients were studied in the cardiac catheterization laboratory with simultaneous Doppler echocardiography using one or both methods to reduce preload. Baseline data were obtained and repeated during the Valsalva maneuver (n = 27) or after the administration of nitroglycerin (n = 36). Thus, 13 patients underwent the Valsalva maneuver only, 22 had administration of nitroglycerin only and 14 had both.
2.1 Baseline data.
Baseline data for all 49 study patients are given in Table 1. The baseline E/A ratio and deceleration time versus the pre–A wave pressures are shown in Fig. 2, with E/A cutoff points >2.0 and deceleration time <150 ms, as described previously [3, 6, 27]. All patients who had either a baseline E/A ratio ≥2.0 (n = 3) or deceleration time ≤150 ms (n = 11) had a pre–A wave pressure ≥15 mm Hg. Although these indexes were 100% specific for a pre–A wave pressure ≥15 mm Hg, they were only 10% and 38% sensitive, respectively. There were 23 patients who had a pre–A wave pressure ≥15 mm Hg with either an E/A ratio <2.0 or a deceleration time >150 ms—that is, a “pseudonormal” pattern.
Left, Distribution of left ventricular diastolic pressure before atrial contraction (Pre a) by baseline mitral flow velocity E/A ratio. Dashed linesseparate patients with pre–A wave pressure >15 mm Hg and E/A ratio >2.0. Right,Distribution of left ventricular diastolic pressure before atrial contraction by baseline mitral deceleration time (DT). Dashed linesseparate patients with pre–A wave pressure >15 mm Hg and DT <150 ms. Open circles= ejection fraction <50%; solid circles= ejection fraction ≥50%.
Baseline Doppler Echocardiographic and Cardiac Catheterization Data for All 49 Study Patients
2.2 Valsalva maneuver.
The hemodynamic catheterization and Doppler echocardiographic data for the patients who performed the Valsalva maneuver are given in Table 2. The overall relation between the decrease in E/A′ ratio and pre–A wave pressure is shown in Fig. 3(left). With the higher pre–A wave pressures, there was a larger negative difference in the E/A′ ratio between baseline and during the Valsalva maneuver. Patients with normal and abnormal systolic function are represented by solid and open circles, respectively, in Fig. 3.
Left, Distribution of left ventricular diastolic pressure before atrial contraction (Pre a) by change in mitral flow velocity E/A′ ratio from baseline to peak strain on the Valsalva maneuver (ΔE/A′). Dashed linesseparate patients with pre–A wave pressure >15 mm Hg. Open circlesidentify those patients with normal systolic function. Right,Distribution of left ventricular pressure before atrial contraction by change in E/A′ ratio from baseline to after administration of nitroglycerin [Δ (E/A′)]. Dashed linesseparate patients with pre–A wave pressure >15 mm Hg. Open circles= ejection fraction <50%; solid circles= ejection fraction ≥50%.
Doppler Echocardiographic and Cardiac Catheterization Data at Baseline and Doppler Echocardiographic Data at Peak Strain for the 27 Patients Who Performed the Valsalva Maneuver
Patients were classified into two groups on the basis of their pre–A wave left ventricular pressure: group I consisted of 10 patients with a normal pre–A wave pressure <15 mm Hg, and group II consisted of 17 patients with an increased pre–A wave pressure ≥15 mm Hg. Representative Doppler echocardiographic recordings from a patient in each group before and during the Valsalva maneuver are shown in Fig. 4. At the peak strain period of the Valsalva maneuver, the maximal E wave velocity decreased in both groups. The A′ wave velocity decreased in group I (p = 0.004) but was unchanged in group II (p = 0.40). There was a significant difference in the change in E/A′ between group I and group II patients (−0.35 ± 0.17 vs. −1.22 ± 0.11; p = 0.02), as shown in Fig. 5(top left). Also shown in Fig. 5(top right) is the difference in the change in E/A between group I and group II patients (−0.34 ± 0.08 vs. −0.74 ± 0.48; p = 0.01). The difference in the change in E/A′ between group I and group II was present in the subgroup of patients with a baseline E/A ratio >1 or <2, as depicted by the solid circles in Fig. 5(top). In these patients with E/A ratios between 1 and 2, the difference between group I and group II remains significant (−0.31 ± 0.21 vs. −1.15 ± 0.94; p = 0.05).
Transmitral Doppler recordings in two patients with simultaneous left ventricular pressure trace at baseline (left)and at peak strain on Valsalva maneuver (right). Top,Patient with normal left ventricular filling pressures. During the Valsalva maneuver, E, A and A′ velocities decrease equally (ΔE/A′ = 0.15), whereas the mitral deceleration time increases. Bottom,Patient with increased left ventricular filling pressures. During the Valsalva maneuver, E wave velocity decreases, whereas A′ velocity remains nearly unchanged, resulting in a decreased E/A′ ratio (ΔE/A′ = 0.50). Mitral deceleration time increases. Pre-A = left ventricular diastolic pressure before atrial contraction.
Top, Scatterplots of the changes in the E/A′ and E/A ratios in group I versus group II during the Valsalva maneuver. Top left,Change in E/A′ measurements. Top right,Change in E/A measurements. Bottom,Scatterplots of the changes in the E/A′ and E/A ratios in group I versus group II during administration of nitroglycerin versus baseline. Bottom left,Change in E/A′ ratios. Bottom right,Change in E/A ratios. Gp I = pre–A wave pressures <15 mm Hg (group I); Gp II = pre–A wave pressures ≥15 mm Hg (group II); open circles= E/A ratio >1 and <2; solid circles= E/A ratio ≤1 or ≥2.
2.3 Nitroglycerin.
Table 3displays similar hemodynamic data for the patients who received nitroglycerin sublingually. The overall relation between the decrease in E/A′ ratio and pre–A wave pressure is shown in Fig. 3(right). With higher pre–A wave pressures, there is a larger negative difference in the E/A ratio between baseline and nitroglycerin administration. Patients with normal and abnormal systolic function are represented by solid and open circles, respectively, in Fig. 3.
Doppler Echocardiographic and Cardiac Catheterization Data at Baseline and After Nitroglycerin Administration
Patients were classified into those with a pre–A wave pressure <15 mm Hg (group IA; n = 15) and those with a pre–A wave pressure ≥15 mm Hg (group IIA; n = 21). Fig. 6shows representative recordings from a patient in each group and their response to the nitroglycerin. The maximal E wave velocity decreased in both groups (p < 0.001). The A′ wave velocity decreased in group I (p < 0.05) but was unchanged in group II (p = 0.23). There was a significant difference in the change in E/A′ ratio between group I and group II (−0.18 ± 0.17 vs. −0.81 ± 0.49; p < 0.001), as shown in Fig. 5(bottom left). Also shown in Fig. 5(bottom right) is the difference in the change in E/A between group I and group II (−0.18 ± 0.12 vs. −0.45 ± 0.23; p < 0.001). The difference in the change in the E/A′ ratio between group I and group II remained significant in the group of patients with a baseline E/A ratio >1 and <2 (−0.13 ± 0.19 vs. −0.84 ± 0.51; p = 0.008). These patients are depicted by the closed circles in Fig. 5(bottom).
Transmitral pulsed wave Doppler recordings from two patients with simultaneous left ventricular (LV) pressure trace at baseline (left)and after administration of nitroglycerin (NTG) sublingually (right). Top,Patient with normal left ventricular filling pressures. After nitroglycerin, E, A and A′ wave velocities decrease equally, whereas mitral deceleration time increases (change in E/A′ = 0.20). Bottom,Patient with increased left ventricular filling pressures. After nitroglycerin, E and A wave velocities decrease, whereas A′ wave velocity remains unchanged, resulting in a decreased E/A ratio (change in E/A′ = 1.50). Mitral deceleration time increases. Pre–A = left ventricular diastolic pressure before atrial contraction.
2.4 Correlation of techniques.
Fourteen patients performed the Valsalva maneuver and received nitroglycerin with adequate test responses, as defined earlier. The correlation for the E/A′ ratio difference between the interventions was excellent (r = 0.88, p = 0.03), consistent with a similar response to two different interventions aimed at reducing preload.
3 Discussion
It has been hypothesized [1, 28, 29], and subsequently proven in an experimental model [30]and in patients with amyloidosis [2], that progression of transmitral flow velocity patterns occur in the course of cardiac disease. In normal, young to middle-aged patients, the E/A ratio is slightly >1.0 and deceleration time is ∼200 ± 40 ms. With the onset of diastolic dysfunction and impaired left ventricular relaxation, E wave velocity decreases and deceleration time decreases, with no or a minimal increase in mean left atrial pressure. With disease progression, the E wave velocity increases because of higher left atrial pressure on mitral valve opening and deceleration time decreases in parallel with a decline in ventricular compliance. This is termed “pseudonormalization” of the transmitral flow pattern, because the mitral flow velocity curve simulates the normal transmitral flow pattern. An additional increase in left atrial pressure and a decline in compliance lead to an increase in E wave velocity and a decrease in deceleration time. This is known as a “restrictive transmitral filling pattern” [1].
By using standard two-dimensional echocardiography and the Doppler variables of transmitral flow (i.e., E, A and deceleration time), left ventricular filling pressures can be predicted accurately in selected patients [1–7]. These predictions are based on the hypothesis that with higher filling pressures in an abnormal ventricle, E wave velocity will be greater and deceleration time will be shorter. As has been recognized in patients with dilated cardiomyopathy or other disease states with decreased systolic function, there is an inverse relation between deceleration time and left ventricular filling pressure [1, 3, 6, 7]. However, in patients in whom a normal transmitral pattern may exist (i.e., those without left ventricular systolic dysfunction), the transmitral E wave velocity and deceleration time appear to be less accurate in estimating left ventricular filling pressures, because the pseudonormal pattern cannot be differentiated from a normal pattern on mitral flow velocity curves alone [4]. It is these patients in whom assessment of diastolic filling may be most important, because one-third of patients with congestive heart failure have normal systolic function [31].
Several methods have been proposed for differentiating the normal pattern from the pseudonormal pattern to overcome these limitations. Analysis of pulsed wave Doppler pulmonary vein A wave duration, color M-mode flow propagation velocity and two-dimensional echocardiographic determination of left atrial size have all been identified as potential markers of pseudonormal filling [26, 32–36]. Because mitral velocity curves change in a characteristic manner with the loading conditions of the left ventricle, it was hypothesized that changes in preload may also be able to differentiate the normal pattern from the pseudonormal one.
Previous research has defined the effects of altered preload on the indexes of transmitral Doppler echocardiography [8–20]. In normal subjects, a reduction in preload, produced by either the Valsalva maneuver or nitroglycerin, results in a decrease in both the E and A wave velocities and prolongation of deceleration time. This hypovolemic response produces little change in the E/A ratio. In contrast, diseased hearts with a pseudonormal pattern demonstrate a similar decline in the E wave velocity but less or no decline in the A wave velocity. This results in a decreased E/A ratio, with the emergence of an impaired relaxation pattern. Isolated impairment of relaxation causes a decrease in early filling of the left ventricle, with a compensatory increase in filling and atrial contraction. Patients with markedly reduced compliance and severely restrictive filling may or may not have the ability to alter their filling pattern in response to a reduction in preload, depending on left ventricular operating chamber compliance. Preliminary data suggest that failure to revert to a pseudonormal pattern in patients with restrictive filling with preload reduction is associated with a worse prognosis [37]. These hypothesized changes in mitral filling patterns with alterations in preload are depicted in Fig. 7, in which one-way arrows represent dynamic changes in the mitral filling patterns that do not occur with reducing preload and two-way arrows depict the filling patterns that are changeable.
Depiction of left ventricular filling patterns assuming a constant heart rate of 70 beats/min. In addition to the four basic patterns of transmitral Doppler flow, a fifth pattern shows the effects of preload reduction on a normal filling pattern. Here, the E and A wave velocities decrease to a similar degree, with no change in the E/A ratio. Also depicted is a severely restrictive pattern thought to represent a marked reduction in left ventricular compliance, with a tall E wave, low A wave velocity and short deceleration time. One-way arrows= dynamic changes in mitral filling patterns that do not occur with reducing preload; two-way arrows= filling patterns that are changeable.
In the present study, both the Valsalva maneuver and the administration of nitroglycerin were used to reduce preload and to assess the dynamic response of mitral inflow indexes to changes in load. The velocity at atrial contraction (E at A) was measured to obtain A′, because both maneuvers may cause an increase in both heart rate and peak A wave velocity. Measurement of the A′ wave velocity minimizes the confounding effects of alterations in heart rate [25]. The utility of using the E/A′ versus E/A ratio is shown in Fig. 5, in which there is a greater separation of patients with normal versus high filling pressures when using the E/A′ ratio.
Overall, the mitral E wave velocity decreased and the deceleration time increased in response to a reduction in preload. The A′ wave velocity also decreased in those patients with lower filling pressures; however, it remained the same or increased in patients with significantly increased pre–A wave pressures. There was a significant difference in the change in the E/A′ ratio during intervention between patients with low and high left ventricular filling pressures. This finding was present regardless of the baseline E/A ratio or left ventricular systolic function.
3.1 Study limitations.
The number of patients in the study was relatively small, and most of the patients were referred to the cardiac catheterization laboratory with a primary diagnosis of coronary artery disease. Despite this limitation, the range of baseline hemodynamic data and ejection fractions was wide. Mean left atrial pressure was not measured directly; however, Yamamoto et al. [5]demonstrated a good correlation between pre–A wave pressure and mean left atrial pressure in patients with heart disease.
Not all patients had both interventions, owing to clinical judgment, inadequate Doppler recordings during the Valsalva maneuver or failure to respond adequately to preexisting administration of nitroglycerin. The responses to the Valsalva maneuver were not fully standardized, and the magnitude of the intrathoracic pressure generated was not assessed quantitatively, but could be with sonospirometry. The hemodynamic response to nitroglycerin varied because there was no useful way for controlling this response. Finally, all Doppler echocardiographic recordings were obtained simultaneously with catheterization data with the patient in the supine position, which limited optimal positioning of the patient. This may have decreased the ability to obtain transmitral recordings (other investigators have had a higher success rate in obtaining transmitral flow velocity curves during the Valsalva maneuver [14]).
3.2 Clinical implications of the study.
The results of the present study are consistent with previous data indicating that left ventricular diastolic pressures are increased in patients who have a restrictive transmitral filling pattern and no further assessment needs. However, as shown herein, there is a high prevalence of a pseudonormal pattern in patients in whom further evaluation is necessary to determine left ventricular filling pressures. Alterations in mitral flow velocity variables produced by the Valsalva maneuver or administration of nitroglycerin help predict filling pressure in patients with all other types of left ventricular filling patterns.
It must be emphasized that no “cutoff” values were provided based on the results of the study. Despite the significant difference in the change in the E/A′ ratio during preload reduction between patients with low and high left ventricular filling pressures, overlap was present. Small changes in the change in the E/A′ ratio during preload reduction could make a significant difference in the sensitivity and specificity of the single cutoff value. Other investigators have used the criterion of a baseline E/A ratio >1.0 becoming <1.0 during the Valsalva maneuver [14, 38]. However, this criterion could not be applied to our patients because 90% of patients with a baseline E/A ratio >1.0 had a decrease in E/A ratio to <1.0 during preload reduction. Thus, this study should be viewed as a preliminary investigation to support future studies to determine the ultimate clinical utility of this technique.
Footnotes
↵1 Present address: Minneapolis Heart Institute Foundation, Minneapolis, Minnesota.
- Received July 12, 1996.
- Revision received March 21, 1997.
- Accepted April 24, 1997.
- The American College of Cardiology
