Author + information
- Received April 15, 2009
- Revision received May 14, 2009
- Accepted May 26, 2009
- Published online October 1, 2009.
- Kenya Kusunose, MD,
- Hirotsugu Yamada, MD, PhD⁎ (, )
- Susumu Nishio, RMS,
- Noriko Tomita, MD, PhD,
- Toshiyuki Niki, MD,
- Koji Yamaguchi, MD, PhD,
- Kunihiko Koshiba, MD, PhD,
- Shusuke Yagi, MD, PhD,
- Yoshio Taketani, MD,
- Takashi Iwase, MD, PhD,
- Takeshi Soeki, MD, PhD,
- Tetsuzo Wakatsuki, MD, PhD,
- Masashi Akaike, MD, PhD and
- Masataka Sata, MD, PhD
Reprint requests and correspondence:
Dr. Hirotsugu Yamada, Department of Cardiovascular Medicine, Institute of Health Biosciences, University of Tokushima Graduate School, 2-50-1 Kuramoto, Tokushima 770-8503, Japan
Objectives We evaluated the usefulness of the ratio of the early diastolic transmitral flow velocity (E) to the mitral annular velocity (e′) calculated from simultaneously recorded E and e′ in atrial fibrillation (AF).
Background The ratio of the E to the e′ (E/e′) has been reported as a useful index even in AF patients. However, E and e′ were measured during different beats in the previous studies.
Methods Fifty-six AF patients with preserved systolic function (mean age 66 ± 11 years) underwent routine echocardiographic study. The E/e′ was calculated from the E and e′ simultaneously recorded by the dual Doppler echocardiography. A single-beat E/e′ was calculated from simultaneously recorded E and e′ when the preceding RR interval/pre-preceding RR interval = 1. Brain natriuretic peptide (BNP) levels were also examined. Twenty-one patients underwent simultaneous pulmonary artery catheterization.
Results The single-beat lateral E/e′ correlated with pulmonary capillary wedge pressure (PCWP) (r = 0.74, p < 0.001). The single-beat lateral E/e′ of ≥11 could predict elevated PCWP (≥15 mm Hg) with a sensitivity of 90% and a specificity of 90%. The single-beat lateral E/e′ also correlated well with the log BNP concentration. The single-beat lateral E/e′ of ≥9.2 predicted a plasma BNP level of ≥200 pg/ml with 88% sensitivity and 84% specificity.
Conclusions The single-beat lateral E/e′ correlated with plasma BNP level and PCWP in AF patients with preserved systolic function. In addition, the single-beat lateral E/e′ (≥11) was a good predictor of elevated PCWP (≥15 mm Hg). Dual Doppler echocardiography offers an advantage of providing the single-beat lateral E/e′ correctly even in AF patients, for the evaluation of left ventricular diastolic function.
Atrial fibrillation (AF) is the most frequent form of arrhythmia, and the number of patients with AF has increased with the rapid aging of society (1). AF is an important risk factor for heart failure (2), wherein left ventricular (LV) diastolic function is more closely related to the symptoms, exercise tolerance, and prognosis of patients compared with systolic function (3,4). In the clinical setting, LV diastolic function is mainly evaluated by the transmitral flow velocity pattern using Doppler echocardiography. Evaluation of LV diastolic function is also needed in patients with AF which is extremely challenging because of the lack of atrial systolic transmitral flow wave and the irregularity of Doppler parameters caused by irregular R-R intervals (5–7).
As an alternative method for evaluating LV diastolic function, mitral annular velocity waveforms, which can be recorded using tissue Doppler echocardiography, have been used. The ratio of early diastolic transmitral flow velocity (E) to early diastolic mitral annular velocity (e′), E/e′, has been widely used to estimate the LV filling pressure in patients with various forms of heart disease and to predict their prognosis (8,9). Several studies have also reported the clinical usefulness of the E/e′ ratio in patients with AF (10,11). However, in these studies, E and e′ were separately measured during different beats, and the assessment was not theoretically accurate. Therefore, a recently developed novel dual Doppler method facilitates the recording of Doppler waveforms at 2 different points by dividing the transmitted pulses of a pulse Doppler echocardiography into 2 and alternately transmitting/receiving them. As such, both transmitral flow and mitral annular velocity waveforms can be recorded during the same beat and E/e′ can be instantly calculated.
If the E/e′, which can be noninvasively obtained by Doppler echocardiography, reflects the plasma BNP level and the filling pressures, it may be useful for assessing the condition of chronic AF, selecting therapeutic strategies, and evaluating the treatment response (12–19). The present study was undertaken to evaluate the role of E/e′, when E and e′ were calculated at the same time using the dual Doppler method, in the estimation of the plasma BNP level and filling pressures in chronic AF patients with preserved systolic function. Furthermore, we examined which site of the mitral annulus should be used and which beat should be selected for the calculation of E/e′ in AF.
The study subjects consisted of 56 patients (40 male, 16 female; mean age 66 ± 11 years) with chronic AF. Twenty-one patients (mean age 72 ± 6 years) had simultaneous right heart catheterization with echocardiographic examination. Patients were excluded if they had paroxysmal AF, severe valvular heart disease, congenital heart disease, LV systolic dysfunction (LV ejection fraction <50%), renal dysfunction (serum creatinine >2.0 mg/dl), or LV regional wall motion abnormality at the basal lateral or septal segment. Clinical data were obtained by a complete review of each patient's medical record, history taking, physical examination, and transthoracic echocardiography. Conventional medical therapies such as angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, diuretics, digoxin, and vasodilators were continued throughout the study. The institutional review board of the University of Tokushima approved the protocol of this study, and written informed consent was obtained from all subjects.
All patients underwent pre-cordial M-mode, 2-dimensional, and Doppler echocardiography while in the left lateral position. We used an ultrasound machine EUB-7500 (Hitachi Medical Corporation, Kashiwa, Japan). All images were stored digitally for playback and analysis. Left ventricular end-diastolic and end-systolic dimensions were measured from the M-mode echocardiogram. The LV mass index was estimated from the formula of Devereux et al. (20). Pulmonary artery systolic pressure was estimated by measurement of tricuspid regurgitation velocity (v) and estimate right atrial pressure based on size and collapsibility of inferior vena cava with the formula: pulmonary artery systolic pressure = 4v2 + estimate right atrial pressure (21). The LV ejection fraction (LVEF) and the left atrial (LA) volume were calculated by the Simpson method using 2-dimensional images. The LA volume was indexed to body surface area. Transmitral flow and mitral annular motion velocities were simultaneously recorded over 30 s by the newly developed dual Doppler echocardiography in the apical 4-chamber view. Transmitral flow velocity was recorded at the tip of the mitral leaflet, and mitral annular velocities were recorded at the mitral annulus of the LV lateral wall and the interventricular septal sides (Fig. 1). The ratio E/e′ of peak E to peak e′ was calculated by 3 different methods: 1) mean E/e′ was the average of instantaneous E/e′ during the simultaneous recording of E and e′ in 30 s; 2) single-beat E/e′ was calculated from simultaneously recorded E and e′ when the preceding RR interval (RRp)/pre-preceding RR interval (RRpp) = 1; and 3) conventional E/e′ was calculated from separately measured E and e′ in randomly picked up 3 cardiac cycles, respectively. The coefficient of variation of Doppler measurements was the ratio of the standard deviation to the mean values. The deceleration time of E-wave (DT) was measured. All Doppler measurements and calculations were performed without knowledge of the plasma BNP concentrations and pulmonary capillary wedge pressure (PCWP). No patients were excluded because of suboptimal images.
Measurement of BNP concentration
Blood samples were taken at the end of echocardiography. A 2-ml blood sample was drawn from an antecubital vein after 10 min of supine rest, placed in a tube containing ethylene diamine tetraacetic acid, and analyzed within 2 h. Plasma BNP concentrations were measured by chemiluminescence enzyme immunoassay.
Mean right atrial pressure, pulmonary artery pressure, and PCWP were measured with a pulmonary artery catheter. An investigator unaware of the echocardiographic data acquired the pressure measurements. Each parameter was averaged in 12 cycles at end-expiratory apnea. Fluid-filled transducers were balanced before the study with the zero level at the midaxillary line.
Comparisons of values within the same individuals were assessed using a paired t test. The diagnostic abilities of the E/e′ for separating plasma BNP concentrations and PCWP were determined by the receiver-operating characteristic (ROC) curve. Linear regression analysis was used to evaluate the correlations between echocardiographic variables and the log BNP concentration or PCWP. Potential determinants of the BNP level were identified by univariate regression analysis, and all identified predictors were then entered in a stepwise manner into a multivariate regression model. Bland-Altman analysis was used as a means of assessing systematic differences between the single-beat lateral and conventional lateral E/e′. Values were considered significantly different at p < 0.05. Statistical analysis was performed primarily using a statistical software package (MedCalc Software, Maria-kerke, Belgium).
Ten randomly selected studies were measured for reanalysis of the single-beat lateral E/e′ by 1 observer at 2 separate times, and the other 10 studies were chosen at random for reanalysis of E/e′ by 2 observers for determination of intraobserver and interobserver variabilities, respectively.
Baseline clinical characteristics
The mean age of the patients was 66 ± 11 years (range 36 to 85 years). Thirty-eight percent of the patients had hypertensive heart disease; 34% had lone AF. The mean heart rate of the patients was 74 ± 10 beats/min. The heart rate was fairly controlled in most of the patients. The LVEF was preserved. The mean LA size of the patients was greater than that for the normal control subjects (Table 1). The characteristics of 21 patients who underwent pulmonary artery catheterization are shown in Table 2. The E, conventional lateral E/e′, and single-beat lateral E/e′ were significantly greater, and the DT was significantly smaller in the higher PCWP group (≥15 mm Hg) than in the lower PCWP group (<15 mm Hg). However LV mass index, LA volume index, and LVEF did not differ between the 2 groups.
Mitral annular motion in septal and lateral sides
There was a good positive linear relationship between the mean lateral E/e′ and the mean septal E/e′ (r = 0.95, p < 0.001) (Fig. 2A). In all patients, the mean lateral e′ was higher than the mean septal e′ (9.2 ± 2.6 cm/s vs. 8.1 ± 2.8 cm/s, p < 0.05), and the mean lateral E/e′ was lower than the mean septal E/e′ (8.9 ± 3.1 vs. 9.9 ± 3.6, p < 0.05).
Beat-to-beat variability and single-beat measurement
The values of E, e′, and E/e′ did not correlate well with the preceding RR interval (RRp)/pre-preceding RR interval (RRpp) in a particular patient (Fig. 3). The mean coefficient of variation of e′ and E/e′ was smaller than E (e′: 12.2 ± 6.1%, E/e′: 14.0 ± 5.2%, E: 18.7 ± 5.4%). There was a good positive linear relationship between the mean lateral E/e′ and the single-beat lateral E/e′ at RRp/RRpp = 1 (r = 0.97, p < 0.001) (Fig. 2B). Intraobserver variability in E/e′ measurement was 4.9 ± 8.2%, and interobserver variability was 6.6 ± 8.7%, respectively.
Relationship between BNP and echocardiographic parameters
There was a good positive linear relationship between log BNP and the mean lateral E/e′ (r = 0.87, p < 0.001) (Fig. 4A) and the mean septal E/e′ (r = 0.84, p < 0.001). However, the LA volume index was not significantly related to the log BNP (Fig. 4B). The log BNP also correlated with age, LV mass index, E, DT, and pulmonary artery systolic pressure. To investigate the independent predictors of the log BNP, a stepwise multilinear regression analysis was used with several echocardiographic variables, and of these, the single-beat lateral E/e′ (when RRp/RRpp = 1) was the best predictor of the log BNP (Table 3). The ROC curves for diagnosing plasma BNP concentration of ≥200 pg/ml are shown in Figure 5. The single-beat lateral E/e′ value of ≥9.2 predicted a plasma BNP level of ≥200 pg/ml with 88% sensitivity and 84% specificity according to ROC analysis. The area under the curve was 0.92 for the single-beat lateral E/e′ (Fig. 5). The area under the curve for the single-beat lateral E/e′ to predict plasma BNP concentration of ≥200 pg/ml was significantly larger than that for the conventional lateral E/e′ (0.92 vs. 0.84, p < 0.05). The correlation coefficient between the single-beat lateral E/e′ and the conventional lateral E/e′ was r = 0.85. The mean difference ±1.96 SD was 0.5 ± 1.8 by Bland-Altman analysis (Fig. 6).
Relationship between PCWP and echocardiographic parameters
The PCWP correlated with BNP, conventional lateral E/e′, and single-beat lateral E/e′ (Table 4). The single-beat lateral E/e′ correlated well with PCWP (r = 0.74, p < 0.001) (Fig. 7A). The single-beat lateral E/e′ of ≥11 could predict elevated PCWP (≥15 mm Hg) with a sensitivity of 90% and a specificity of 90%. The conventional lateral E/e′ correlated with PCWP (r = 0.57, p < 0.01) (Fig. 7B). The conventional lateral E/e′ of ≥10 could predict elevated PCWP (≥15 mm Hg) with a sensitivity of 80% and a specificity of 64%. The plasma BNP level of ≥260 pg/ml could predict elevated PCWP (≥15 mm Hg) with a sensitivity of 80% and a specificity of 72%.
We calculated per-beat E/e′ using the newly developed dual Doppler method for the simultaneous recording of the transmitral flow and the mitral annular velocity waveforms in AF patients with preserved systolic function. The mean lateral E/e′ correlated with the plasma BNP level. In addition, the mean lateral E/e′ over 30 s correlated with the single-beat lateral E/e′ when RRp/RRpp = 1, and the single-beat lateral E/e′ also correlated with the plasma BNP level and PCWP.
Assessment of diastolic function in AF patients
It is known that the Frank-Starling mechanism and the interval–force relationship are involved in changes in LV systolic function between irregular heartbeats in AF patients. Few studies have reported interbeat changes of diastolic function. Tabata et al. (22) reported that LV relaxation depended on the systolic LV pressure during the preceding heartbeat in an experiment using dogs that underwent thoracotomy under anesthesia. In AF patients whose atrial systolic waves are absent, it is difficult to evaluate diastolic function based on transmitral flow velocity waveforms. An increase in the E/e′ reflects an elevation of the LV filling pressure, and this parameter may be useful for evaluating LV diastolic function in AF patients. Sohn et al. (11) indicated a correlation between the E/e′ and the LV filling pressure in AF patients. Okura et al. (10) reported that the E/e′ was a prognostic factor in patients with nonvalvular AF. However, in these studies, the E and the e′ were determined during different heartbeats. Therefore, the E/e′ was calculated based on 2 values measured under different loading conditions in AF patients in whom the R-R intervals were irregular. Theoretically, this method is not accurate, which is a limitation of these studies. In the present study, we simultaneously recorded the transmitral flow and mitral annular velocity waveforms using dual Doppler echocardiography to overcome this limitation.
Dual Doppler echocardiography
We first used dual Doppler echocardiography on the evaluation of LV performance in patients with AF. This method enabled simultaneous recording of 2 pulsed Doppler waveforms of flow–flow, flow–tissue, and tissue–tissue velocities. The simultaneous single-beat lateral E/e′ was more sensitive and specific for detecting elevated BNP and PCWP than the conventional lateral E/e′ (Figs. 5 and 7), indicating incremental benefit of this novel technique. Further beneficial application of this method will be comparisons of timing between flow and tissue velocities.
Mitral annular motion in septal and lateral sides
In previous studies examining the E/e′, e′ values have been measured on the lateral or septal sides of the mitral annulus (23). However, few studies have examined the appropriate site for taking e′ measurements to evaluate the LV diastolic function. Hadano et al. (24) recommended that e′ should be measured on the lateral side in patients after thoracotomy and not on the septal side. Usually, lateral e′ was higher than the septal e′. Therefore, changes are more sensitively detected on the lateral side. However, the swinging motion of the heart markedly influences mitral annular motion on the lateral side. On the other hand, mitral annular motion on the ventricular septal side is parallel to the Doppler beam direction, and the influence of cardiac movement may be less marked. In our AF patients, the lateral e′ values were greater than the septal e′ values. Furthermore, the E/e′ correlated with the plasma BNP level regardless of the site of e′ measurement. However, the correlation coefficient calculated was higher on the lateral side than the septal side. The subjects in this study did not include patients with regional wall motion abnormalities on the lateral side nor those with marked mitral annular calcification. However, in such patients, different results should be expected (25).
The relationship between E/e′ and the plasma BNP level
During sinus rhythm, the plasma BNP level correlates with the end-diastolic LV pressure (26) and reflects the prognosis of heart failure (12–15). Therefore, this parameter is often used for assessing heart failure based on various heart diseases. Several studies reported that the plasma BNP level correlated with the LV diastolic function rather than LVEF (27,28). We confirmed the clinical usefulness of the E/e′ by examining its relationship with the plasma BNP level. The mean E/e′ value in chronic AF patients correlated with the plasma BNP level, which is consistent with the results of previous studies (29). The E/e′ measured by this procedure reflected the plasma BNP level, suggesting that this parameter is useful for evaluating heart failure and predicting prognosis, even in patients with AF.
Single-beat lateral E/e′
In this study, we simultaneously recorded both the transmitral flow and mitral annular velocity waveforms for 30 s and performed statistical analysis using the mean E/e′ calculated per heartbeat. However, it is unfeasible in clinical practice, unless automatic measurement becomes possible. Therefore, we conducted a similar analysis using the E/e′ when the RRp/RRpp = 1, which is reported to reflect the average LV function in AF patients. The single-beat lateral E/e′ value correlated with the mean lateral E/e′ value over 30 s, the plasma BNP level, and PCWP. However, when the R-R interval is shorter, the error may be large.
The detection and assessment of diastolic dysfunction in AF patients with preserved LV function is clinically relevant. The results of this study showed that the plasma BNP level and PCWP could be estimated based on the single-beat lateral E/e′, even in AF patients with preserved systolic function. Therefore, this parameter may facilitate the evaluation of LV diastolic function in the presence of AF, which is difficult to assess by conventional Doppler echocardiography, suggesting its clinical usefulness. Our simultaneous E/e′ should be useful for evaluating response to the treatment of AF patients with heart failure and predicting their prognosis. Furthermore, dual Doppler echocardiography enables us to instantaneously calculate the single-beat lateral E/e′, making measurement in the presence of AF possible and allowing the accurate evaluation of E/e′ changes under different loading conditions or after drug loading.
The main limitation of this study is the small number of patients who underwent pulmonary artery catheterization. It was difficult for us to perform catheterization on these subjects because they did not require invasive studies for the clinical purpose. Furthermore, this study included a relatively homogeneous and selective population of AF patients with preserved LVEF and controlled heart rate. In addition, the BNP level may be influenced by right ventricular function. However, none of our subjects had right ventricular wall motion abnormality and hypertrophy. Pulmonary vein velocities may have additional information (30); however, we did not evaluate the velocities in this study because the clinical usefulness of the flow velocities in AF is still unclear.
The single-beat lateral E/e′ correlated with plasma BNP level and PCWP in AF patients with preserved systolic function. In addition, the single-beat lateral E/e′ (≥11) was a predictor of elevated PCWP (≥15 mm Hg), and the sensitivity and specificity of the index were better than those of the plasma BNP level (≥260 pg/ml). The dual Doppler echocardiography has an advantage because it provides the single-beat lateral E/e′ correctly even in AF patients for the evaluation of LV diastolic function in AF patients.
- Abbreviations and Acronyms
- atrial fibrillation
- brain natriuretic peptide
- deceleration time of E wave
- early diastolic transmitral flow velocity
- early diastolic mitral annular velocity
- left atrial/atrium
- left ventricle/ventricular
- left ventricular ejection fraction
- pulmonary capillary wedge pressure
- preceding RR interval
- pre-preceding RR interval
- Received April 15, 2009.
- Revision received May 14, 2009.
- Accepted May 26, 2009.
- American College of Cardiology Foundation
- Lloyd-Jones D.M.,
- Wang T.J.,
- Leip E.P.,
- et al.
- Zile M.R.,
- Lewinter M.M.
- Nakamura Y.,
- Konishi T.,
- Nonogi H.,
- Sakurai T.,
- Sasayama S.,
- Kawai C.
- Nagueh S.F.,
- Mikati I.,
- Kopelen H.A.,
- Middleton K.J.,
- Quinones M.A.,
- Zoghbi W.A.
- Ommen S.R.,
- Nishimura R.A.,
- Appleton C.P.,
- et al.
- Okura H.,
- Takada Y.,
- Kubo T.,
- et al.
- Wei C.M.,
- Heublein D.M.,
- Perrella M.A.,
- et al.
- Yoshimura M.,
- Yasue H.,
- Tanaka H.,
- et al.
- Berger R.,
- Huelsman M.,
- Strecker K.,
- et al.
- Rienstra M.,
- Van Gelder I.C.,
- Van den Berg M.P.,
- Boomsma F.,
- Van Veldhuisen D.J.
- Devereux R.B.,
- Reichek N.
- Ceyhan C.,
- Unal S.,
- Yenisey C.,
- Tekten T.,
- Ceyhan F.B.