Glinical Investigator
C1in Investig (1992) 70: S 79-S 86
© Springer-Verlag 1992
Left ventricular hypertrophy regression during antihypertensive treatment H. Eichstaedt, O. Danne, R.J. Schroeder, and D. Kreuz Departments of Cardiology and Radiology, University Hospital Rudolf Virchow, Free University of Berlin, and Division of Cardiology, Health Science Center at Brooklyn, State University of New York
Summary. For more than 20 years hypertrophy regression has been in the focus of hypertension research. Many studies in animals have shown impressive reduction of left ventricular hypertrophy after medical treatment of hypertension. The most important result seems to be that hypertrophy can be almost completely reversed in young animals, whereas in older animals regression of left ventricular hypertrophy appears to be less complete. Hypertrophy regression in man seems much more difficult to prove. The direct correlation between left ventricular muscle mass and ECG changes has been disappointing in many studies. Echocardiography is able to show a comparatively good impression of left ventricular muscle mass and therefore can also demonstrate regression of left ventricular hypertrophy within its methodological limits. There is no doubt that today magnetic resonance imaging has by far the best imaging quality of all the clinical methods and is able to demonstrate both hypertrophy and its regression with incomparable accuracy. In the present clinical study hypertrophy regression has been demonstrated after 6 months of treatment with Carvedilol. Key words: Hypertension Left ventricular hypert r o p h y - Hypertrophy regression Magnetic resonance - Carvedilol
Observations in animal studies Different anti-hypertensive drugs produce different degrees of regression of left ventricular hypertroAbbreviations." LVH=left ventricular hypertrophy; SHR= spontaneously hypertensive rats; Part=mean arterial pressure; WKY= Wistar-Kyoto rats; ACE= angiotensin converting enzyme; VCG = vector cardiography; MRI = magnetic resonance imaging; aU=arbitrary units; RNV=radionuclide ventriculography; ISA = intrinsic activity
phy (LVH), as Sen et al. [45] first showed in spontaneously hypertensive rats (SHR). For example, with alpha-methyldopa there was substantial lowering of LV mass, but the same lowering of blood pressure produced by minoxidil did not cause regression of LVH. In addition, propranolol produced substantial reversal of LVH in doses that did not lower blood pressure. By contrast, Motz et al. [37] found the vasodilator agent hydralazine did produce regression of LVH. However, there was further reduction in LV mass when metoprolol was added to hydralazine, with only slight additional lowering of blood pressure. In this study, there was substantial reduction in LV mass/volume ratio with therapy, indicating structural remodelling similar to the findings in humans. In SHR, lowering of mean arterial pressure Part is associated with substantial reduction in LV mass, though the values reached remain usually somewhat above those of Wistar-Kyoto rats (WKY) [37, 49]. This was also observed by Fletcher [20] after normalizing BP following reversal of renal cellophane wrap hypertension in rabbits, by removing the fibrous capsule encasing the kidney: in untreated renal hypertensive rabbits, LV mass/BW was 2 g/kg, compared with a mean value of 1.3 g/kg in normal rabbits. Six weeks after the kidneys were unwrapped, LV mass/BW was 1.5 g/kg, indicating substantial regression of LVH. Development of LVH parallels the rise of blood pressure from 4-20 weeks in SHR [32]. Korner et al. investigated three groups of SHR with the ACE inhibitor enalapril in a dose that normalized BP. Treatment was given from (a) 4-10 weeks; (b) 4-14 weeks, which corresponds to the major phase of development of hypertension; and (c) 1420 weeks, when adult levels of blood pressure have been reached [32]. In all three groups, LVH had been reversed completely by the end of treatment, with LV mass/BW ratios the same as those of WKY. When treatment was stopped, BP remained
$80 stable and at significantly lower values than in untreated SHR. In SHR treated from 4-14 weeks, BP rose by only a small amount by the end of the study, when the rats were 3-40 weeks of age, and this was associated with a corresponding slight redevelopment of LVH. In the other two groups, 15-20 weeks after treatment, BP had stabilized about halfway between the levels obtained in untreated SHR and WKY, with corresponding redevelopment of a modest degree of LVH, which was somewhat greater than in SHR treated from 414 weeks. During antihypertensive therapy, LVH can be completely reversed in young animals, as was shown by the above-mentioned studies, and its subsequent redevelopment is in proportion to the elevation in BP. In older animals, regression of LVH, though substantial, appears to be less complete. This may relate to the development of irreversible structural changes, such as the deposition of collagen, owing to damage associated with poor perfusion of the hypertrophied muscle fibres. Methods for measurement in humans
Electrocardiographic methods Until recently, diagnosis of left ventricular hypertrophy was confirmed mainly by electrocardiographic (ECG) criteria of LVH. These various criteria of ECG-LVH were based upon increased electromotive precardial voltage forces plus repolarization changes on the standard electrocardiogram. There are more than 30 different criteria that have been developed by various investigators to relate electrocardiographic measurements to more direct evidence of increased anatomic ventricular mass. In general, these electrocardiographic criteria of LVH are rather consistently insensitive, yet they are fairly specific for detecting increased left ventricular mass [6]. These electrocardiographic criteria by definition provide only a categorical variable for the continuous variable of left ventricular mass; and therefore, the direct correlation of the actual LVH mass with the ECG changes has been disappointing [42]. It follows that the clinical premorbid diagnosis of LVH has generally referred to the ECG pattern rather than the actual mass of the left ventricle. The relative risk for the patient to develop an adverse cardiovascular event or death is markedly increased when LVH shows on the ECG [23]. In the prospectively conducted Framingham study, development of ECG-LVH was documented in individuals who had had previously normal ECGs.
Thus, in this study 35 percent of all men who manifested ECG-LVH died within five years of the appearance of this adverse marker. Furthermore, men and women between 35 and 64 years of age had a 10- to 19-fold increased risk of cardiovascular mortality when ECG-LVH was present as compared to age- and gender-matched individuals without these ECG abnormalities [30]. Moreover, the risk of sudden death is particularly increased in patients with ECG-LVH [29], and left ventricular ectopy appears to be more prevalent on ambulatory monitoring in individuals with ECG-LVH [35, 36]. Prognostic implications of ECG-LVH are also demonstrable in other forms of adverse cardiovascular events. Risk of later development of angina pectoris, myocardial infarction, stroke, symptomatic peripheral vascular disease, and overt congestive heart failure are all significantly increased in patients with ECG-LVH [30]. Indeed, the greatest risk of developing overt congestive heart failure could be predicted from the presence of ECGLVH. In the Framingham study, asymptomatic men and women with ECG-LVH had more than a 17-fold chance of ultimately developing congestive heart failure than age- and gender-matched cohorts without evidence of ECG-LVH [30]. In vector-cardiography similar criteria for quantification have been developed and show a comparable value for diagnosis and prognosis of arterial hypertension.
Echocardiography Anatomical evidence of LVH can certainly be present without being detected by the ECG. So it has become apparent that the echocardiogram provides an excellent noninvasive means for more accurately assessing the presence of increased left ventricular mass. In expert hands, premortem echocardiographic measurements of left ventricular mass has correlated very closely with actual anatomic measurements made from patients with a wide range of ventricular weights who died as a result of a wide variety of cardiovascular disorders [7]. In contrast, correlations against ECG and vector cardiography (VCG) are relatively low. Echocardiographic technique provides a reliable measure of left ventricular mass, but since mass clearly is a continuous variable, it follows that for this noninvasive diagnostic measurement to be applied to general populations it must address more directly the question of the true prevalence of increased left ventricular mass. Levy and coworkers performed echocardiographic evalua-
$81 tions on almost 5,000 participants of the Framingham Heart Study. This important study demonstrated that left ventricular hypertrophy (by echocardiographic criteria) was not uncommon in the general population, being present in 16 percent of the men and 19 percent of the women evaluated [33]. Moreover, an important age effect was demonstrated as well as a significant association between blood pressure levels and the presence of echocardiographic evidence of left ventricular hypertrophy. Another finding of this populationbased study was the influence of exogenous obesity as an independent risk for the occurrence of echocardiographic left ventricular hypertrophy. The importance of both age and the level of systolic arterial pressure on the prevalence of echocardiographically demonstrable LVH support the concept that both height of arterial pressure, and duration of hypertension are important determinants of actual LVH. We also investigated different types and locations of left ventricular hypertrophy in patients with hypertension compared to aortic stenosis [10]. Magnetic resonance studies
Magnetic resonance imaging (MRI) has emerged as a method for visualization of the heart and has been available in our institution for cardiac imaging since the end of 1983. Because MRI displays the blood pool as having very low or no signal intensity with the spin-echo technique, the cardiac chambers are visualized without contrast medium. This method is excellent for determining the cardiac axis, layer depth, and time of triggering in the cardiac cycle [5, 40]. Therefore, MRI is a very suitable method for determination of myocardial parameters before and after drug treatment [12]. Until now there have been only very few studies on hypertrophy regression using magnetic resonance [13-17]. The main reason seems to be that while normally the treating cardiologist is the same person who also investigates echocardiography, other specialists typically interpret magnetic resonance images. We have the rare opportunity to work as cardiologists with MR. In the last eight years we have undertaken numerous regression studies under medications with diuretics, alpha-blockers, calcium antagonists, beta-blockers, clonidine, and various ACE inhibitors. In the following we will, as an example, describe the study with the new nonselective betaadrenoceptor antagonist Carvedilol, which has been shown to have additional alpha-l-blocking properties.
The study population comprised 27 outpatients (21 men and 6 women) aged f~=53.9 years with arterial hypertension. Their diastolic blood pressures were measured noninvasively and ranged from 95 to 115 mmHg. Ten untreated control subjects were also included in our study protocol. Patients with myocardial failure controlled by drugs, dysrhythmias, coronary heart disease, stenosis of the renal artery, or endocrine hypertension were excluded from the study. The study protocol consisted of a 2-week washout phase followed by a treatment phase of 25 mg Carvedilol per day over a period of 6 months. Prior to Carvedilol treatment, blood pressure, plasma calcium, potassium, magnesium, sodium, and creatinine were measured and the baseline values for ventricular parameters were determined by electrocardiography, echocardiography, radionuclide ventriculography, and MRI. After 2 weeks of treatment, blood pressure was measured, blood samples taken, and LVH measurements obtained. Five patients interrupted the study in this first phase. Final measurements of blood pressure, MRI scans, radionuclide ventriculograms, and electrocardiograms were performed after 3 and 6 months of Carvedilol treatment for the remaining patients. Serum calcium, potassium, magnesium, sodium, and creatinine were measured again at this time. Blood pressure was measured noninvasively using a mercury sphygmomanometer in both patients and normal controls. Electrocardiograms at rest were also recorded with quantification of the left precordial R waves. Measurements of left ventricular wall thickness were made in three planes based on a special angulation algorithm and the mean of three readings calculated for each. Mmode echocardiograms were performed using standard techniques with a Diasonics DRF-400 (digital radio frequency). Measurements of interventricular septal thickness had to amount to more than 15 mm in end-systole and 12 mm in end-diastole. Baseline values for posterior wall thickness were allowed to be 1 mm less. The left ventricular (LV) end-diastolic diameter had to be < 50 mm and the LV end-systolic diameter had to be 30 mm in keeping with concentric hypertrophy. The magnetic resonance tomograph (Siemens Magnetom) used a superconducting magnet operating at a field strength of 0.5 T (Table 1). The Sirecust 404 ECG system with ultrathin copper leads assured exact end-systolic and end-diastolic triggering. The repetition rate, or interval between sets of radiofrequency pulses, was determined by each subject's heart rate and ranged from 400 to
$82 Table 1. Specification of technical parameters of the M R imaging system at the University Hospital Berlin-Charlottenburg Diameter of the magnet Magnetic field strength
50 cm 0.5T
High-frequency field strength
8 to 10 kW
Duration of excitation impulse
0.5 to 1.0 s
Spin-echo seqeunce
35 and 70 ms
Trigger delay time
360 ms
Slice thickness
10 mm
Imaging matrix
256 x 256 matrix points
Resolution between 2 lines (735 matrix lines and 31-cm diagonal screen)
C1in Investig (1992) 70: S 79-S 86
© Springer-Verlag 1992
Left ventricular hypertrophy regression during antihypertensive treatment H. Eichstaedt, O. Danne, R.J. Schroeder, and D. Kreuz Departments of Cardiology and Radiology, University Hospital Rudolf Virchow, Free University of Berlin, and Division of Cardiology, Health Science Center at Brooklyn, State University of New York
Summary. For more than 20 years hypertrophy regression has been in the focus of hypertension research. Many studies in animals have shown impressive reduction of left ventricular hypertrophy after medical treatment of hypertension. The most important result seems to be that hypertrophy can be almost completely reversed in young animals, whereas in older animals regression of left ventricular hypertrophy appears to be less complete. Hypertrophy regression in man seems much more difficult to prove. The direct correlation between left ventricular muscle mass and ECG changes has been disappointing in many studies. Echocardiography is able to show a comparatively good impression of left ventricular muscle mass and therefore can also demonstrate regression of left ventricular hypertrophy within its methodological limits. There is no doubt that today magnetic resonance imaging has by far the best imaging quality of all the clinical methods and is able to demonstrate both hypertrophy and its regression with incomparable accuracy. In the present clinical study hypertrophy regression has been demonstrated after 6 months of treatment with Carvedilol. Key words: Hypertension Left ventricular hypert r o p h y - Hypertrophy regression Magnetic resonance - Carvedilol
Observations in animal studies Different anti-hypertensive drugs produce different degrees of regression of left ventricular hypertroAbbreviations." LVH=left ventricular hypertrophy; SHR= spontaneously hypertensive rats; Part=mean arterial pressure; WKY= Wistar-Kyoto rats; ACE= angiotensin converting enzyme; VCG = vector cardiography; MRI = magnetic resonance imaging; aU=arbitrary units; RNV=radionuclide ventriculography; ISA = intrinsic activity
phy (LVH), as Sen et al. [45] first showed in spontaneously hypertensive rats (SHR). For example, with alpha-methyldopa there was substantial lowering of LV mass, but the same lowering of blood pressure produced by minoxidil did not cause regression of LVH. In addition, propranolol produced substantial reversal of LVH in doses that did not lower blood pressure. By contrast, Motz et al. [37] found the vasodilator agent hydralazine did produce regression of LVH. However, there was further reduction in LV mass when metoprolol was added to hydralazine, with only slight additional lowering of blood pressure. In this study, there was substantial reduction in LV mass/volume ratio with therapy, indicating structural remodelling similar to the findings in humans. In SHR, lowering of mean arterial pressure Part is associated with substantial reduction in LV mass, though the values reached remain usually somewhat above those of Wistar-Kyoto rats (WKY) [37, 49]. This was also observed by Fletcher [20] after normalizing BP following reversal of renal cellophane wrap hypertension in rabbits, by removing the fibrous capsule encasing the kidney: in untreated renal hypertensive rabbits, LV mass/BW was 2 g/kg, compared with a mean value of 1.3 g/kg in normal rabbits. Six weeks after the kidneys were unwrapped, LV mass/BW was 1.5 g/kg, indicating substantial regression of LVH. Development of LVH parallels the rise of blood pressure from 4-20 weeks in SHR [32]. Korner et al. investigated three groups of SHR with the ACE inhibitor enalapril in a dose that normalized BP. Treatment was given from (a) 4-10 weeks; (b) 4-14 weeks, which corresponds to the major phase of development of hypertension; and (c) 1420 weeks, when adult levels of blood pressure have been reached [32]. In all three groups, LVH had been reversed completely by the end of treatment, with LV mass/BW ratios the same as those of WKY. When treatment was stopped, BP remained
$80 stable and at significantly lower values than in untreated SHR. In SHR treated from 4-14 weeks, BP rose by only a small amount by the end of the study, when the rats were 3-40 weeks of age, and this was associated with a corresponding slight redevelopment of LVH. In the other two groups, 15-20 weeks after treatment, BP had stabilized about halfway between the levels obtained in untreated SHR and WKY, with corresponding redevelopment of a modest degree of LVH, which was somewhat greater than in SHR treated from 414 weeks. During antihypertensive therapy, LVH can be completely reversed in young animals, as was shown by the above-mentioned studies, and its subsequent redevelopment is in proportion to the elevation in BP. In older animals, regression of LVH, though substantial, appears to be less complete. This may relate to the development of irreversible structural changes, such as the deposition of collagen, owing to damage associated with poor perfusion of the hypertrophied muscle fibres. Methods for measurement in humans
Electrocardiographic methods Until recently, diagnosis of left ventricular hypertrophy was confirmed mainly by electrocardiographic (ECG) criteria of LVH. These various criteria of ECG-LVH were based upon increased electromotive precardial voltage forces plus repolarization changes on the standard electrocardiogram. There are more than 30 different criteria that have been developed by various investigators to relate electrocardiographic measurements to more direct evidence of increased anatomic ventricular mass. In general, these electrocardiographic criteria of LVH are rather consistently insensitive, yet they are fairly specific for detecting increased left ventricular mass [6]. These electrocardiographic criteria by definition provide only a categorical variable for the continuous variable of left ventricular mass; and therefore, the direct correlation of the actual LVH mass with the ECG changes has been disappointing [42]. It follows that the clinical premorbid diagnosis of LVH has generally referred to the ECG pattern rather than the actual mass of the left ventricle. The relative risk for the patient to develop an adverse cardiovascular event or death is markedly increased when LVH shows on the ECG [23]. In the prospectively conducted Framingham study, development of ECG-LVH was documented in individuals who had had previously normal ECGs.
Thus, in this study 35 percent of all men who manifested ECG-LVH died within five years of the appearance of this adverse marker. Furthermore, men and women between 35 and 64 years of age had a 10- to 19-fold increased risk of cardiovascular mortality when ECG-LVH was present as compared to age- and gender-matched individuals without these ECG abnormalities [30]. Moreover, the risk of sudden death is particularly increased in patients with ECG-LVH [29], and left ventricular ectopy appears to be more prevalent on ambulatory monitoring in individuals with ECG-LVH [35, 36]. Prognostic implications of ECG-LVH are also demonstrable in other forms of adverse cardiovascular events. Risk of later development of angina pectoris, myocardial infarction, stroke, symptomatic peripheral vascular disease, and overt congestive heart failure are all significantly increased in patients with ECG-LVH [30]. Indeed, the greatest risk of developing overt congestive heart failure could be predicted from the presence of ECGLVH. In the Framingham study, asymptomatic men and women with ECG-LVH had more than a 17-fold chance of ultimately developing congestive heart failure than age- and gender-matched cohorts without evidence of ECG-LVH [30]. In vector-cardiography similar criteria for quantification have been developed and show a comparable value for diagnosis and prognosis of arterial hypertension.
Echocardiography Anatomical evidence of LVH can certainly be present without being detected by the ECG. So it has become apparent that the echocardiogram provides an excellent noninvasive means for more accurately assessing the presence of increased left ventricular mass. In expert hands, premortem echocardiographic measurements of left ventricular mass has correlated very closely with actual anatomic measurements made from patients with a wide range of ventricular weights who died as a result of a wide variety of cardiovascular disorders [7]. In contrast, correlations against ECG and vector cardiography (VCG) are relatively low. Echocardiographic technique provides a reliable measure of left ventricular mass, but since mass clearly is a continuous variable, it follows that for this noninvasive diagnostic measurement to be applied to general populations it must address more directly the question of the true prevalence of increased left ventricular mass. Levy and coworkers performed echocardiographic evalua-
$81 tions on almost 5,000 participants of the Framingham Heart Study. This important study demonstrated that left ventricular hypertrophy (by echocardiographic criteria) was not uncommon in the general population, being present in 16 percent of the men and 19 percent of the women evaluated [33]. Moreover, an important age effect was demonstrated as well as a significant association between blood pressure levels and the presence of echocardiographic evidence of left ventricular hypertrophy. Another finding of this populationbased study was the influence of exogenous obesity as an independent risk for the occurrence of echocardiographic left ventricular hypertrophy. The importance of both age and the level of systolic arterial pressure on the prevalence of echocardiographically demonstrable LVH support the concept that both height of arterial pressure, and duration of hypertension are important determinants of actual LVH. We also investigated different types and locations of left ventricular hypertrophy in patients with hypertension compared to aortic stenosis [10]. Magnetic resonance studies
Magnetic resonance imaging (MRI) has emerged as a method for visualization of the heart and has been available in our institution for cardiac imaging since the end of 1983. Because MRI displays the blood pool as having very low or no signal intensity with the spin-echo technique, the cardiac chambers are visualized without contrast medium. This method is excellent for determining the cardiac axis, layer depth, and time of triggering in the cardiac cycle [5, 40]. Therefore, MRI is a very suitable method for determination of myocardial parameters before and after drug treatment [12]. Until now there have been only very few studies on hypertrophy regression using magnetic resonance [13-17]. The main reason seems to be that while normally the treating cardiologist is the same person who also investigates echocardiography, other specialists typically interpret magnetic resonance images. We have the rare opportunity to work as cardiologists with MR. In the last eight years we have undertaken numerous regression studies under medications with diuretics, alpha-blockers, calcium antagonists, beta-blockers, clonidine, and various ACE inhibitors. In the following we will, as an example, describe the study with the new nonselective betaadrenoceptor antagonist Carvedilol, which has been shown to have additional alpha-l-blocking properties.
The study population comprised 27 outpatients (21 men and 6 women) aged f~=53.9 years with arterial hypertension. Their diastolic blood pressures were measured noninvasively and ranged from 95 to 115 mmHg. Ten untreated control subjects were also included in our study protocol. Patients with myocardial failure controlled by drugs, dysrhythmias, coronary heart disease, stenosis of the renal artery, or endocrine hypertension were excluded from the study. The study protocol consisted of a 2-week washout phase followed by a treatment phase of 25 mg Carvedilol per day over a period of 6 months. Prior to Carvedilol treatment, blood pressure, plasma calcium, potassium, magnesium, sodium, and creatinine were measured and the baseline values for ventricular parameters were determined by electrocardiography, echocardiography, radionuclide ventriculography, and MRI. After 2 weeks of treatment, blood pressure was measured, blood samples taken, and LVH measurements obtained. Five patients interrupted the study in this first phase. Final measurements of blood pressure, MRI scans, radionuclide ventriculograms, and electrocardiograms were performed after 3 and 6 months of Carvedilol treatment for the remaining patients. Serum calcium, potassium, magnesium, sodium, and creatinine were measured again at this time. Blood pressure was measured noninvasively using a mercury sphygmomanometer in both patients and normal controls. Electrocardiograms at rest were also recorded with quantification of the left precordial R waves. Measurements of left ventricular wall thickness were made in three planes based on a special angulation algorithm and the mean of three readings calculated for each. Mmode echocardiograms were performed using standard techniques with a Diasonics DRF-400 (digital radio frequency). Measurements of interventricular septal thickness had to amount to more than 15 mm in end-systole and 12 mm in end-diastole. Baseline values for posterior wall thickness were allowed to be 1 mm less. The left ventricular (LV) end-diastolic diameter had to be < 50 mm and the LV end-systolic diameter had to be 30 mm in keeping with concentric hypertrophy. The magnetic resonance tomograph (Siemens Magnetom) used a superconducting magnet operating at a field strength of 0.5 T (Table 1). The Sirecust 404 ECG system with ultrathin copper leads assured exact end-systolic and end-diastolic triggering. The repetition rate, or interval between sets of radiofrequency pulses, was determined by each subject's heart rate and ranged from 400 to
$82 Table 1. Specification of technical parameters of the M R imaging system at the University Hospital Berlin-Charlottenburg Diameter of the magnet Magnetic field strength
50 cm 0.5T
High-frequency field strength
8 to 10 kW
Duration of excitation impulse
0.5 to 1.0 s
Spin-echo seqeunce
35 and 70 ms
Trigger delay time
360 ms
Slice thickness
10 mm
Imaging matrix
256 x 256 matrix points
Resolution between 2 lines (735 matrix lines and 31-cm diagonal screen)
