Incremental Value of Ultrasonic Tissue Characterization (Backscatter) in the Evaluation of Left Ventricular Myocardial Structure and Mechanics in Essential Arterial Hypertension
Background— Ultrasonic backscatter parameters were analyzed in hypertensive patients and divided into groups according to both severity of left ventricular hypertrophy (LVH) (group A: no LVH [n=52]; B: mild to moderate LVH [n=55]; and C: severe LVH [n=10]) and left ventricular geometry (normal geometry [n=44]; concentric remodeling [n=8]; concentric hypertrophy [n=25]; and eccentric hypertrophy [n=40]).
Methods and Results— We studied 117 male, essential hypertensive patients and 19 normotensive, age-matched (40±5 years), healthy subjects who served as controls. Ambulatory and office blood pressure measurements were taken and 2-dimensional Doppler echocardiography and ultrasonic myocardial integrated backscatter (IBS) were performed. A group from the hypertensive study population (n=16) was observed after a period of pharmacological antihypertensive treatment to determine the behavior of backscatter parameters in relation to eventual regression of left ventricular mass (LVM). The cyclic variation index (CVIs) of the backscatter signal at the septum level was grouped according to each LVM level and was 29.4±9.3 (controls), 15±11 (group A), 9.5±10 (group B), and −1.5±8.6 (group C) (P<0.001). CVI septum values grouped according to left ventricular geometry were 15±11 (normal geometry), 12±7 (concentric remodeling), 7±11 (concentric hypertrophy), and 7.8±11 (eccentric hypertrophy) (P<0.01). Follow-up data demonstrate a significant reduction of LVM after therapy, as well as a significant increase in CVIs toward normal values.
Conclusions— Hypertensive patients with higher LVM had the worst prognosis; in fact, those patients had the most significant CVI alterations. Regression of LVM subsequent to chronic pharmacological therapy induces a normalization of ultrasonic backscatter parameters. Ultrasonic tissue characterization (backscatter) analysis could allow early identification of patients at risk of developing complications of hypertensive cardiopathy.
Received July 19, 2002; revision received September 17, 2002; accepted September 23, 2002.
Experimental1,2 and autoptical3 data support the hypothesis that the pressure-volume overload that characterizes essential arterial hypertension, by itself or via interaction of complex humoral factors such as the renin-angiotensin-aldosterone system, could cause an increase in intramyocardial fibrosis with alteration of collagen/myocytic content ratio.4 Systemic arterial hypertension represents a major cause of cardiovascular morbidity and mortality; the predictive value of echocardiographic5 evidence of left ventricular hypertrophy (LVH) for subsequent cardiovascular morbidity and mortality is well established. Abnormalities of left ventricular diastolic function have been found in patients affected by arterial hypertension and LVH.6 In the present study, we evaluated 1) the capability of integrated backscatter (IBS) to detect alterations in myocardial structure in both static and dynamic modality; 2) the incremental diagnostic and prognostic value of IBS in comparison with conventional echocardiography; and 3) the potential parallel effects of a chronic antihypertensive therapy on left ventricular mass (LVM) and tissue characterization parameters.
Exclusion criteria were malignant hypertension, heart failure, cardiomyopathy or valvular heart disease, diabetes and/or obesity, coronary artery disease, and renal and connective tissue disease. Furthermore, after repeated casual blood pressure measurements and 24-hour blood pressure monitoring, patients were selected on the basis of their LVM as shown by echocardiography. On the basis of these criteria, 117 male, never-treated, essential hypertensive patients with good hemodynamic compensation and absent to severe LVH were recruited. This group of patients was compared with a group of 19 carefully age- and sex-matched normotensive subjects. The study was approved by the Ethical Committee of the University of Pisa.
Conventional echocardiographic studies were performed with a digital Philips Sonos 5500 echograph (S4 fusion imaging probe, 2 to 4 MHz) in a fundamental imaging mode. Left ventricular diameter and septum and posterior wall thickness were measured according to the procedures of the American Society of Echocardiography.7 LVM was calculated with Devereux’s formula (Penn convention) and normalized for body surface area (LVMbs) and height2,7 (LVMh). Relative wall thickness was also measured according to standard formula. Midwall fractional shortening of the left ventricle was calculated according to Shimuzu’s model.8 Meridional end-systolic stress was calculated using the standard formula. After pulsed Doppler, transmitral flow velocity parameters were evaluated, including peak E, peak A, E/A ratio, mitral acceleration time, mitral deceleration time, and isovolumic relaxation time. All heart rates were corrected by Bazett’s formula.
Hypertensive patients were distributed into 3 balanced subgroups. Group A comprised patients with LVMbs values within the normal ranges of our laboratory (LVMbs ≤124 g/m2) (n=52), group B comprised patients with LVMbs values between ≥125 and ≤174 g/m2 (n=55), and group C comprised patients with LVM values >175 g/m2 (n=10) (Table 2). On the basis of the relationship between relative wall thickness and LVM,9 the total patient population (n=117) was then divided into 4 different groups (Table 3). The groups consisted of hypertensive patients with normal relative wall thickness and LVM (normal geometry; n=44; 37%); patients with concentric remodeling (n=8; 7%); patients with concentric hypertrophy (n=25; 21%); and patients with eccentric hypertrophy (n=40; 35%). A group of 16 hypertensive subjects selected from the hypertensive study population, all of whom had a high degree of LVH, underwent a new echocardiographic examination after 1 year of chronic homogeneous pharmacological therapy (angiotensin-converting enzyme inhibitors plus Ca2+ antagonist). All patients have achieved good blood pressure level control during this period.
A commercially available Acoustic Densitometry software package (Philips) was used on a Sonos 5500 (Philips) for the quantitative analysis of integrated backscatter. A detailed methodology for IBS analysis has been described previously.10,11
Continuous variables were expressed as mean±SD. Intra-group differences were evaluated using an unpaired Student’s t test. Upper and lower 95% confidence limits for each variable were calculated with the use of a 2-tailed Student’s t test distribution using the formulas mean ± (2.042×SD) and mean − (2.042×SD), respectively. Relations between IBS and 2-dimensional echocardiographic measurements were expressed in terms of linear multiple regression analysis. A paired t test was applied to compare the same patients before and after therapy. A probability value <0.05 was considered significant.
Age (40.3±5.6 versus 38.7±6.1 years) and body mass index (27.5±9.2 versus 26.6±5.7 kg/m2) were similar in hypertensive and normotensive patients. Average daytime ambulatory blood pressure was 145±7/93±4 versus 118±6/76±5 mm Hg (P<0.0001) for hypertensive and normotensive patients, respectively, and casual blood pressure averaged 156±13/99±6 versus 123±14/79±6 mm Hg (P<0.0001). Average heart rate was comparable in the 2 groups (73.5±11 in hypertensive versus 69.3±13 in normotensive subjects). Septum and posterior wall thickness were greater in hypertensive subjects, whereas the end-diastolic diameters were comparable between groups. LVMbs was significantly higher in the hypertensive group (Table 1). Left ventricular end-diastolic diameter and relative fractional shortening overlapped in the 2 groups (Table 1) and were in the normal range. Midwall fractional shortening was significantly lower in hypertensive patients, whereas meridional end-systolic stress was significantly higher (Table 2). The E/A ratio was significantly higher in normotensive subjects (Table 1). The A velocity time integral was significantly higher in hypertensive patients than in controls; as a result, the E/A velocity time integral ratio was significantly lower in hypertensive patients than in controls (Table 1).
In the subgroup with no LVH, the cyclic variation indexes (CVIs) of both the septum and posterior wall at the medium and proximal levels were significantly lower than in controls. The E/A ratio, however, did not differentiate between hypertensive subjects and controls (Figure 1). In presence of a moderate degree of LVH, the CVIs for both the septum and posterior wall were significantly lower in comparison with both hypertensive patients with no LVH and controls. An increase in LVMbs produced a decrease in CVI values in the group of hypertensive subjects with severe LVH (Table 2).
The diastolic values of IBS at both the septum and posterior wall levels, indexed for pericardial reflection, showed a significant increase only in patients with the severe form of LVH (Table 2, Figure 2). Midwall fractional shortening was significantly lower in patients with severe LVH or in those with concentric and eccentric LVH (Table 3). Meridional end-systolic stress showed a significant increase in patients with severe LVH, including subjects with concentric and eccentric LVH (Table 2 and Table 3⇑). When considering the sensitivity of the 2 tests in discriminating hypertensive patients from control subjects through individual analysis, we found that the E/A ratio was only able to discriminate 34% (40/117) of hypertensive patients from controls, whereas individual analysis for CVI at both the septum and posterior wall levels was able to discriminate 70% (80/117; P<0.01) of hypertensive patients from controls (Figure 1).
LVM was significantly lower after 1 year of treatment because of a significant reduction of septum and parietal thickness. Fractional shortening remained unchanged after 1 year, but midwall fractional shortening significantly increased. Left ventricular diastolic function demonstrated by transmitral flow analysis showed a slight but significant improvement. Importantly, after 1 year, we observed a trend toward normalization in IBS parameters, in particular in the CVIs at all sampled levels (Table 4, Figure 3).
Relationship Between the Quantitative Backscatter Analysis Data, the Echocardiographic Parameters, and Blood Pressure
CVIs at both medium and proximal levels of the septum and posterior wall were unrelated to the left ventricular fractional shortening and diastolic functional parameters. Systolic arterial pressure values were closely linked to both CVIs (mid-septum: r=−0.44, P<0.003; mid-posterior wall: r=−0.58, P<0.005). CVIs at both the septum and posterior wall levels showed an inverse, significant correlation with LVMbs (septum: r=−0.50, P<0.005; posterior wall: r=−0.53, P<0.004) and with meridional end-systolic stress (septum: r=−0.51, P<0.005; posterior wall: r=−0.52, P<0.004). Furthermore, CVIs at both the septum and posterior wall levels showed a significant correlation with midwall fractional shortening (septum: r=0.46, P<0.005; posterior wall: r=0.53, P<0.001) (Table 5). A stepwise multivariate regression analysis has shown a significant relationship (multiple r=0.79; r2=0.63; P<0.02) between the CVI at the mid-septum level after 1 year of therapy subtracted from its basal value (dependent variable) and the values of systolic arterial pressure (P<0.05), LVM (P<0.05), meridional end-systolic stress (P<0.05), and midwall fractional shortening (P<0.06) after 1 year year subtracted from their basal values (independent variables).
Left ventricular hypertrophy has been divided arbitrarily into 3 phases: an adaptive phase and a compensatory phase, in both of which relief of increased load is associated with reversal of contractile dysfunction, and a pathological phase, in which contractile function is abnormal and removal of excessive load is not accompanied by return to normal contractile function. A recent autopsy study in human hearts12 showed that collagen volume fraction increases in parallel with the severity of LVH. The most important pathological findings were LVH (cardiac myocyte hypertrophy) and interstitial, perivascular, and replacement fibrosis (myocardial apoptosis), the interaction of which play an important role in the determination of left ventricular systolic-diastolic performance.
End-diastolic IBS values indexed for IBS value of pericardium at both the septum and posterior wall levels are significantly higher than controls only in severe form of LVH, which suggests the presence of inappropriate hypertrophy with disproportionate connective tissue growth.
CVI, which is the expression of the intrinsic myocardial contractility, is also altered when the LVM is in a normal range or when there are initial alteration of left ventricular geometry (geometric remodeling) present. This parameter shows a progressive alteration with the increase of LVM and with concentric or eccentric LVH. The fact that there is no correlation with fractional shortening but there is a correlation with meridional end-systolic stress and with midwall fractional shortening is an expression of the intrinsic functional correlates of CVI with both afterload parameters and with indexes of midwall function. CVI reduction observed in hypertensive patients with a “normal” fractional shortening could be considered an “early,” independent index of abnormal intrinsic contractility.
Methodological Considerations about Ultrasonic Backscatter Analysis
We acknowledge that the CVI formula, which involves dividing a quantity expressed in a logarithm domain by another logarithm, could present some mathematical problems, but its intuitive approach has induced us to utilize it. The measurements of cyclic variation IBS based on the difference between average peak and average nadir value are more robust,13 and it is likely that this approach will improve our results when applied in our laboratory. Several authors have demonstrated the potential inaccuracy of a peak-to-peak measurements, which are applied in the present study, and of the magnitude of cyclic variation, not taking in account the dependence of the “apparent” magnitude on the time delay of cyclic variation of myocardial backscatter.14 In our echocardiographic images, the examined structures (septum and posterior wall) were nearly perpendicular to the insonifying ultrasonic beam, and thus we obtained a good signal-to-noise ratio and minimized problems associated with tissue anisotropy.15–17
Collagen and Acoustic Myocardial Properties
Different structural components of the myocardium can influence its acoustic properties under different physiological and pathological conditions. Collagen is the primary determinant of both scattering and attenuation of myocardial tissue; a linear relationship was found between IBS and hydroxyproline content in autopsied human hearts, with fibrotic changes associated with remote myocardial infarction.18 Furthermore, a significant direct correlation was found between collagen content analyzed by myocardial biopsy and regional echo amplitude.19 Myocardial scattering intensity depends directly on myocyte cellular size; the microstructural arrangement of myocardial cells embedded in a collagen matrix may provide a sufficient local acoustic impedance mismatch to account for the scattering from normal myocardium.11
Alterations of Ultrasonic Backscatter Parameters in Essential Hypertension
Despite the lack of histopathological (endobyopsy) data, which is not ethically acceptable in this type of subject, some hypotheses could be made to explain the alterations in the acoustic properties of the myocardium in hypertension. The increase of the myocardial collagenic network that is realized in hypertension (interstitial, perivascular, and replacement fibrosis) could determine, in systole, an increase in scattering, thereby causing a reduction of its normal cyclic variation. Moreover, the pressure-volume overload in hypertension, which causes a stimulus on the myocardium, could determine a change in the orientation, structure, or geometry of both the muscle fibers and the collagen network, thus influencing the acoustic properties of the tissue.
Comparison of Ultrasonic Backscatter Parameters With Transmitral Doppler Echocardiography
The analysis of the transmitral flow velocity is largely used in the evaluation of global diastolic function. In the pressure-volume overload of hypertensive patients, an alteration of the passive end-diastolic phase (increased stiffness) was observed. The E/A ratio is inversely related to LVM, and it is lower in the concentric hypertrophy group, thereby selecting the patients with the worst cardiovascular prognosis. Our study, which confirms a previous videodensitometric observation, demonstrated that CVI is better able to differentiate between normal and hypertensive patients than is the E/A ratio.20
Myocardial Midwall Mechanics and Ultrasonic Backscatter Parameters
The significant relationships between midwall fractional shortening, meridional end-systolic stress, and CVI lends credence to the hypothesis that the CVI data could be considered as an index of intrinsic myocardial contractility that is relatively insensitive to the afterload conditions. In this respect, we note that concentric LVH shows a higher impairment of myocardial intrinsic function, both with a lower CVI at septum and posterior wall levels and with a lower midwall fractional shortening, in comparison with other “geometric subgroups.”
What Really Adds the Ultrasonic Tissue Characterization to the Analysis of Hypertensive Heart?
After 1 year of appropriate and homogeneous antihypertensive therapy, we observe a significant reduction in LVM (mainly an expression of myocytic compartment) and a decrease of IBS at septum level (mainly an expression of myocardial collagen volume resetting), whereas CVI shows a significant increase in comparison with basal values (expression of intrinsic contractility improvement). These preliminary follow-up data clearly show that ultrasonic tissue characterization is able to detect both the changes of myocardial collagen volume and the reversal of intrinsic contractile dysfunction under the influence of effective antihypertensive therapy. Interestingly, all patients received angiotensin-converting enzyme inhibitors, which partially block the production of angiotensin, a key regulatory factor in collagen synthesis in the extracellular matrix. An indirect confirmation of the link between backscatter parameters and cardiac fibrosis (serum type I procollagen level) in hypertension is documented by a recent article.21
Ultrasonic backscatter analysis could allow physicians to assess if a patient is still in an adaptive or compensatory phase before irreversible damage (pathological phase) occurs. More extensive, randomized backscatter studies are needed to evaluate the effect of different pharmacological treatments (Ca2+ antagonist, β-blockers, or angiotensin-converting enzyme inhibitors) on CVI and IBS indexed values.
The authors thank Dr Roberto Farina, Philips Medical System (Milano, Italy), and EMAC (Genova, Italy) for their technological support.
Jalil J,E, Doering CW, Janicki JS, et al. Fibrillar collagen and myocardial stiffness in the intact hypertrophied rat left ventricle. Circ Res. 1989; 64: 1041–1050.
Sahn DJ, DeMaria A, Kisslo J, et al. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation. 1978; 58: 1072–1083.
Shimizu G, Hirota Y, Kawamura K, et al. Left ventricular midwall mechanics in systemic arterial hypertension: myocardial function is depressed in pressure-overload hypertrophy. Circulation. 1991; 83: 1676–1684.
Recchia D, Miller JG, Wickline SA. Quantification of ultrasonic anisotropy in normal myocardium with lateral gain compensation of two-dimensional integrated backscatter images. Ultrasound Med Biol. 1993; 497–505.
Holland MR, Wilkenshoff UM, Finch-Johnston AE, et al. Effects of myocardial fiber orientation in echocardiography: quantitative measurements and computer simulation of the regional dependence of backscatter ultrasound in parasternal short-axis view. J Am Soc Echocardiogr. 1998; 11: 929–937.
Hoyt RH, Collins SL, Skorton DJ, et al. Assessment of fibrosis in infarcted human hearts by analysis of ultrasonic backscatter. Circulation. 1985; 71: 740–744.
Lythall DA, Bishop J, Greenbaum RA, et al. Relationship between myocardial collagen and echo amplitude in non fibrotic hearts. Eur Heart J. 1993; 14: 344–350.
Di Bello V, RPedrinelli, DGiorgi, et al. Ultrasonic myocardial texture versus doppler analysis in hypertensive heart: a preliminary study. Hypertension. 1999; 33: 66–73.
Maceira AM; Barba J, Varo et al. Ultrasonic backscatter and serum marker of cardiac fibrosis in hypertensives. Hypertension. 2002; 39: 923–928.