Tissue Doppler in Spectral and Color ModeMethodological ConsiderationsNon-invasive Estimation of Left Ventricular Filling Pressures using the E/Em index

Abstract: Aims: Tissue Doppler (TD) in spectral mode (spectral TD) and color TD are the two modalities available today in tissue velocity echocardiography (TVE). Measurements of left ventricular (LV) myocardial velocities and displacement may yield different results depending on the employed sonographic modality and the subjective adjustments during data analysis. The ratio of transmitral early diastolic velocity (E) to early diastolic myocardial velocity (Em) has been suggested as a non-invasive estimate of LV filling pressures. The present studies aimed to evaluate the compatibility between the two TD modalities and the effect of temporal filtering and offline gain on velocity and displacement measurements obtained using these two methods. The validity of E/Em in identifying elevated LV filling pressures after acute reductions in hemodynamic LV loading was assessed.Methods and Results: In 57 patients, longitudinal myocardial velocity profiles were acquired from the basal LV segments, using spectral and color TD. Peak systolic (Sm) and early diastolic (Em) myocardial velocities and the myocardial displacement during the ejection phase were measured offline. Spectral TD measurements were performed using three different gain settings (0%-, 50%- 100% offline gain). Color TD analysis were performed before and after the application of temporal filtering at 30, 50 and 70 ms filter width. The E/Em ratio was calculated at the different settings. The correlation between spectral- and color TD measurements was good. Changes in offline gain and filter width resulted in significant alterations on spectral and color TD derived measurements, respectively. Sm and Em were significantly higher (p < 0.001) whereas the E/Em was significantly lower (p < 0.001) for measurements performed with spectral TD compared to color TD and the discrepancy among the measurements increased at increasing filter width and gain level. In Study II the results of spectral- and color TD derived displacement were compared to anatomic M-mode. Spectral TD at different gain settings significantly overestimated M-mode derived displacement measurements, whereas the concordance between color TD and M-mode was considerably better.In Study III, LV longitudinal systolic myocardial velocities and displacement during ejection period were quantified at the basal septal and lateral wall in 24 healthy individuals using spectral TD, color TD and M-mode. Mean spectral TD systolic velocity and displacement values were obtained from the outer and inner borders of the spectral velocity signal. The estimated mean spectral TD systolic myocardial velocities were highly concordant with corresponding color TD measurements (mean difference 0.1 ± 0.6 cm/s, septal wall; 0.1 ± 1.0 cm/s, lateral wall). Similarly, myocardial displacement obtained by integration over time of mean spectral TD velocities was in good agreement with color TD (mean difference 0.2 ± 0.7 mm, septal wall; 0.02 ± 0.86 mm, lateral wall) as well as with the corresponding M-mode measurements.In Study IV, simultaneous LV catheterization and echocardiographic examination was performed in 68 consecutive patients referred for coronary angiography. Doppler signals of the transmitral flow and spectral TD signals at the level of mitral annulus were obtained and the E/Em was then calculated. All examinations were repeated after nitroglycerin (NTG) infusion. At baseline, the predictive ability of E/Em in identifying elevated LVDP was modest at best (area under curve [AUC] ± SE = 0.71 ± 0.08, p < 0.01). The index was more strongly associated with LVDP in patients with reduced ejection fraction (EF) < 55% (r = 0.68; p < 0.01) compared to patients with EF ? 55%(r = 0.4; p < 0.05). Following NTG administration, the diagnostic ability of E/Em in detecting elevated LVDP was significantly reduced (area under curve [AUC] ± SE =0.6 ± 0.08, p > 0.05). Changes in LVDP were not reliably tracked using E/Em (p > 0.05).Conclusions: Spectral TD yields significantly higher myocardial velocity and displacement values and lower E/Em compared to color TD modality. Increasing gain and temporal smoothing significantly alters the obtained velocity and displacement information and yield greater disparity between measurements derived using the two TD methods. Spectral TD significantly overestimates myocardial displacement obtained using M-mode. On the other hand, the agreement between color TD and anatomic M-mode is considerably better. Measurements based on mean spectral TD velocities were highly concordant with those provided by color TD and M-mode. However, unfavorable limits of agreement discourage the interchangeable use of these modalities. The diagnostic ability of E/Em was poor and declined significantly following acute reduction in LV hemodynamic loading. Changes in LVDP were not predicted by alterations in E/Em. Our findings imply that E/Em might not be sufficiently robust to be employed as a single non-invasive estimate for tailoring medical therapy in patients with elevated LVDP.