Left Ventricular Hypertrophy and QT Dispersion in Hypertension
The interlead variation in QT length on a standard electrocardiograph reflects regional repolarization differences in the heart. To investigate the association between this interlead variation (QT dispersion) and left ventricular hypertrophy, we subjected 100 untreated subjects to 12-lead electrocardiography and echocardiography. Additionally, 24 previously untreated subjects underwent a 6-month treatment study with ramipril and felodipine. In the cross-sectional part of the study, QT dispersion corrected for heart rate (QTc dispersion) was significantly correlated with left ventricular mass index (r=.30, P<.01), systolic pressure (r=.30, P<.01), the ratio of peak flow velocity of the early filling wave to peak flow velocity of the atrial wave (E/A ratio) (r=−.22, P=.02), isovolumic relaxation time (r=.31, P<.01), and age (r=.21, P<.04). In the treatment part of the study, lead-adjusted QTc dispersion decreased from 24 to 19 milliseconds after treatment, and after a subsequent 2 weeks of drug washout remained at 19 milliseconds (P<.01). The changes in left ventricular mass index at these stages were 144, 121, and 124 g/m2 (P<.01). Systolic pressure decreased from 175 to 144 mm Hg and increased again to 164 mm Hg after drug washout (P<.01).The E/A ratio (0.97, 1.02, and 1.02; P=.69) and isovolumic relaxation time (111, 112, and 112; P=.97) remained unchanged through the three assessment points. In conclusion, QT dispersion is increased in association with an increased left ventricular mass index in hypertensive individuals. Antihypertensive therapy with ramipril and felodipine reduced both parameters. If an increased QT dispersion is a predictor of sudden death in this group of individuals, then the importance of its reduction is evident.
In hypertensive individuals, LVH predicts an increased morbidity and mortality. This is chiefly through an increase in cardiovascular events. These events include sudden death, which is up to 10 times more prevalent in individuals with LVH than those without.1
A prolonged QT segment on the standard ECG has been found to be associated with an increase in sudden death in various patient groups2 3 and also in apparently healthy individuals.4 More recently, the interlead variations in QT segment length (dispersion) have attracted much interest. An increased QT dispersion may be a more important parameter than QT segment length because it reflects regional repolarization differences in the heart, which may predispose to reentry arrhythmias.
In this study, we investigated the relationship between QT dispersion and LVH. Our hypothesis was that individuals with LVH would have an increased QT dispersion and that antihypertensive therapy causing LVH regression would result in a reduction in QT dispersion.
For the first part of the study, we studied 100 previously untreated subjects. Hypertensive subjects were recruited from the Peart-Rose Clinic at our institution. These consisted of 52 subjects with essential hypertension (BP >160 mm Hg systolic or 90 diastolic) and 21 subjects with borderline hypertension (BP 140 to 159 mm Hg systolic or 85 to 89 diastolic). Additionally, we studied 27 normotensive members of the medical school staff and their acquaintances.
All subjects had normal systolic function determined by two-dimensional echocardiography and no clinical or Doppler evidence of valvular stenosis or regurgitation. Subjects were excluded if they had a history of ischemic heart disease, peripheral vascular disease, congestive cardiac failure, diabetes mellitus, or alcohol abuse. Secondary causes of hypertension were excluded by standard clinical investigation. Each subject underwent echocardiography and a standard 12-lead ECG examination.
The second part of the study involved 24 previously untreated hypertensive subjects who satisfied the same exclusion criteria and underwent antihypertensive treatment. Each subject entered a 4-week placebo run-in phase during which they visited the clinic three times. Baseline BP was taken to be the value obtained at the third visit just before antihypertensive therapy was begun. During this run-in phase, each subject underwent a baseline echocardiographic study and a standard 12-lead ECG. After the third visit, subjects began taking 5 mg ramipril once daily. If BP remained uncontrolled (above 140/90 mm Hg) after 2 weeks, the dose was increased to 10 mg once daily. If BP remained uncontrolled, felodipine was added, initially at a dose of 5 mg once daily and titrated up to 10 mg once daily if necessary. Five subjects achieved BP control on ramipril alone and 19 required the addition of felodipine. Thereafter, subjects were reviewed monthly with BP checks in the clinic. Tablets were counted at each visit to monitor compliance. After 6 months of BP control, echocardiography and 12-lead ECGs were repeated both before and after a 2-week drug washout phase.
The study was approved by the St Mary's Hospital Ethics Committee, and all subjects gave informed consent.
BP was measured at each clinic visit with an automated monitor (Sentron, CR Bard Inc) and suitably sized cuffs after the subject had been seated for 5 minutes. The third of three sitting measurements made 2 minutes apart was taken as the BP.
Two-dimensional echocardiographic studies were performed with a phased-array sector scanner (3.5-MHz transducer, Pass II, General Electric Inc) using a standard examination protocol.5 LV septal wall thickness, posterior wall thickness, and cavity size were measured from the LV short-axis view by two-dimensionally guided M-mode echocardiography. Particular attention was paid to obtaining a precise cross-sectional “on-axis” image of the left ventricle at the papillary muscle tip level. The papillary muscles were then bisected by the M-mode beam, and simultaneous two-dimensional and M-mode images were obtained. Measurements of LV septal, posterior wall, and cavity dimensions were made at end diastole with the use of the Penn convention.6 Three consecutive cardiac cycles were measured and average values obtained.
LV mass was determined by an area×length method that has been validated in humans.7 For this calculation, two echocardiographic views are required: a parasternal short-axis view of the left ventricle at the papillary muscle tip level for assessment of the area of the myocardium, and an apical four-chamber view that maximizes the distance from the mitral valve annulus to the LV apex for determination of the length of the ventricle. LV mass is then calculated from the formula LV Mass=1.04[5/6(A1×L1)−5/6(A2×L2)], where A1 and A2 represent the epicardial and endocardial areas, respectively, measured by planimetry, and L1 and L2 represent the length of the left ventricle from the mitral annulus to epicardial and endocardial borders, respectively. LVMI was determined by dividing LV mass by body surface area.
LVH involves myocyte hypertrophy and an increase in collagen interstitial matrix. It is possible that this latter component may be an important factor in causing abnormalities in diastolic function. Therefore, we assessed diastolic functional parameters in addition to cardiac structural parameters because they might provide additional insight into the mechanisms of any QT segment abnormalities.
Doppler echocardiography is an accepted method for assessment of LV diastolic function.8 The parameters E/A ratio and IVRT were measured as follows: Pulsed Doppler examination of transmitral flow for assessment of LV filling was done with reference to the two-dimensional echocardiographic image from the apical four-chamber view. The sample volume was located between the mitral annulus and the tips of the mitral leaflets, and the position was adjusted so that the sample volume was maintained at an angle as near parallel to transmitral flow as possible by means of an audible signal and spectral velocity display. When the maximal transmitral velocity for the early filling wave was detected, the velocity profile was recorded with the subject in passive end expiration. The peak flow velocities of the early and atrial waves were measured from three consecutive cardiac cycles displaying the highest measurable velocity profiles, and an average value was used for subsequent analysis.
Additionally, patients underwent 24 hour ambulatory BP monitoring at baseline with Spacelabs 90207 machines. Measurements were made every half hour through the day (6 AM to 12 PM and hourly at night. Mean BP was calculated from the readings over the whole 24 hour period. Ambulatory monitoring was deemed acceptable if more than 90% of the readings were recorded.
Continuous-wave Doppler was used for examination of an area between the mitral and aortic valves on the apical four-chamber view so that both mitral and aortic flow patterns were visualized. The velocity profiles were recorded with the subject in passive end expiration. IVRT was taken to be represented by the time interval between the end of the aortic velocity envelope and the onset of the early filling wave. Measurements were made from three consecutive cycles, and the results were averaged.
All two-dimensional and Doppler echocardiographic measurements and analyses were carried out by one observer (J.M.).
All 12-lead ECGs were performed at 25 mm/s with standard lead positions. QT intervals were measured on all possible leads by a single observer blinded to all clinical details using a digitizer. QT intervals were taken to be from the onset of the QRS complex to the end of the T wave, which was defined as its return to the T-P baseline. If U waves were present, the QT interval was measured to the nadir of the curve between the T and U waves. QT intervals were corrected with Bazett's formula9 to compensate for its known dependence on heart rate: QTc=QT/(RR)1/2.
QTc dispersion was determined as the difference between the maximal and minimal QTc interval in different leads. QTc dispersion was adjusted to correct for the known influence of the number of measurable leads by dividing by the square root of this number.10 No subject had fewer than nine measurable leads.
All descriptive data are expressed as mean±SE. Simple regression analysis was used for assessment of correlations between variables. Multiple regression models were constructed by forward variable selection. Repeated measures ANOVA was used for assessment of the significance of longitudinal changes in measured variables.
The subject group consisted of 67 men and 33 women. Fifty were Caucasian, 46 were Afro-Caribbean, and 4 were Asian. Hemodynamic, echocardiographic, and ECG data are presented in Table 1⇓. Mean maximal QTc was 436±3 milliseconds, and mean lead-adjusted QTc dispersion was 20±0.8 milliseconds.
Maximal QTc was significantly correlated with LVMI (r=.41, P<.01), systolic BP (r=.43, P<.01), E/A ratio (r=−.33, P<.01), IVRT (r=.35, P<.01), and age (r=.30, P<.01) (Table 2⇓). In a multiple regression model containing LVMI, systolic BP, 24-hour systolic BP, E/A ratio, and age, maximal QTc was independently related only to 24-hour systolic BP (F=22.3, R2=.19, P<.001).
QTc dispersion was similarly significantly correlated with LVMI (r=.30, P<.01), systolic BP (r=.30, P<.01), E/A ratio (r=−.22, P=.02), IVRT (r=.31, P<.01), and age (r=.21, P<.04) (Table 3⇓). The relationship between QTc dispersion and LVMI is presented in the Figure⇓. In a multiple regression model containing LVMI, systolic BP, 24-hour systolic BP, E/A ratio, and age, QTc dispersion was independently related only to systolic BP (F=10.0, P<.01, R2=.08).
When subjects were divided into those with and those without LVH, maximal QTc was significantly greater in the group with than without LVH (447±5.6 versus 428±2.9 milliseconds, P=.003). Additionally, lead-adjusted QTc dispersion was significantly greater in the group with LVH (21.7±1.4 versus 18.2±0.8 milliseconds, P=.003).
Lead-adjusted QTc dispersion decreased from 24 to 19 milliseconds after treatment (P<.01) and remained unchanged after drug washout (19 milliseconds). Maximal QTc decreased from 460 to 435 milliseconds after treatment (P<.01) and was 442 milliseconds after drug washout. The changes in LVMI at these stages were 144, 121 (P<.01 for the difference), and 124 g/m2. Systolic BP decreased from 175 to 144 mm Hg (P<.01) and increased again to 164 mm Hg after drug washout (P<.01). E/A ratio (0.97, 1.02, and 1.02) and IVRT (111, 112, and 112) remained unchanged through the three assessment points (Table⇑s 4 and 5).⇓⇓
The change in maximal QTc was not significantly correlated with the change in LVMI (r=.29, P=.20), systolic BP (r=.32, P=.16), or heart rate (r=.35, P=.12). Similarly, the change in QTc dispersion was not significantly correlated with the change in LVMI (r=.20, P=.35), systolic BP (r=.22, P=.32), or heart rate (r=.23, P=.32).
QT Dispersion and LVH
The main findings of the cross-sectional part of the study were that QTc dispersion and maximal QTc length were associated with LVMI, BP, E/A ratio, IVRT, and age. The relationships with LVMI are reinforced by the findings of a greater QTc dispersion and maximal QTc length in subjects with LVH compared with those without LVH. The treatment arm of the study allowed a further analysis of the most important factors determining the QT parameters. Through the course of the study, E/A ratio and IVRT, both indexes of LV diastolic function, remained unchanged, whereas BP and LVMI decreased along with QTc dispersion and maximal QTc length. After a drug washout phase, BP increased and QTc dispersion, maximal QTc length, and LVMI remained unchanged. This strongly suggests that LVMI, rather than acute BP levels, is an important factor in determining QTc dispersion and maximal QTc length. Two factors probably account for the fact that changes in QTc dispersion and maximal QTc length were not significantly related to changes in LVMI: First, variables that have only modest baseline correlations were being compared, and second, all of the parameters had a degree of measurement variability.
There is evidence that an increased QT dispersion on the ECG represents regional differences in myocardial recovery of excitability11 and this may lead to a more arrhythmogenic substrate. An increased QT dispersion is associated with an increased risk of arrhythmias in individuals with long QT syndrome,12 who are at high risk of sudden death. More recent work has shown that an increased QT dispersion predicts sudden death in congestive heart failure.13 Additionally, in individuals with hypertrophic cardiomyopathy, a greater QT dispersion was present in those who had had episodes of ventricular tachycardia or ventricular fibrillation than in those who had not.14 There are also data concerning QT dispersion during myocardial infarction. This is greatest just after infarction and decreases with time and successful thrombolysis. It may be highest in those individuals who develop ventricular fibrillation.15 Additionally, an increased QT dispersion 4 weeks after myocardial infarction may be associated with subsequent mortality.16 Thus, there is increasing evidence of the potential importance of this parameter in predicting dangerous ventricular arrhythmias and sudden death in a wide variety of cardiac diseases.
Hypertension affects approximately 20% of the adult population, and a large proportion of these individuals develop LVH, which significantly increases their morbidity and mortality and in particular their risk of sudden death.1 Hypertensive individuals, especially those with LVH, have been noted to have an increase in ventricular arrhythmias,17 but the significance of this is not known. There is little evidence to show that those individuals with frequent or complex arrhythmias are prone to developing sustained ventricular tachycardia and/or die suddenly. Studies using programmed ventricular stimulation have failed to stimulate these potentially life-threatening arrhythmias in hypertensive individuals with LVH and normal ventricular function, even in those with very frequent ventricular ectopic heart beats and nonsustained runs of ventricular tachycardia.18 19 20 Additionally, the Framingham study has failed to convincingly demonstrate that an increase in ventricular ectopic heart beats confers additional risk.21 QT dispersion may be a better predictor of risk of sustained ventricular arrhythmias and sudden death than isolated ventricular ectopic heart beats. Large prospective studies are required to address this issue.
Potential Mechanisms of an Increased QT Dispersion in LVH
At a microscopic level, LVH is characterized by both myocyte hypertrophy and an increase in collagen interstitial matrix. Myocyte hypertrophy may cause a lengthening of action potential duration, and an increase in interstitial fibrosis may be associated with a reduced action potential amplitude and membrane potential, shortened action potential duration, or electrical quiescence.22 Either of these features could result in an increased QT dispersion if the changes in different parts of the ventricle are nonhomogeneous and it is apparent how such changes may result in reentry circuits and ventricular arrhythmias.
Reduction of QT Dispersion With Antihypertensive Therapy
The treatment part of the present study suggests that antihypertensive therapy with an angiotensin-converting enzyme inhibitor and calcium antagonist can reduce QT dispersion. This improvement would seem to be mediated by a reduction in LVMI rather than a direct antiarrhythmic effect of the drugs, since QT dispersion did not increase again after drug washout. It is not possible from this study to comment on whether a differential effect on QT dispersion would be produced by the various classes of antihypertensive drugs. However, some evidence suggests that angiotensin-converting enzyme inhibitors may be particularly good at reversing changes in the collagen interstitial matrix of the left ventricle.23 If this fibrosis is the important mediator in the increase in QT dispersion in LVH, simply reducing LV mass by reversing the myocyte hypertrophy may not be sufficient to reduce QT dispersion. It is unclear whether different antihypertensive agents have differential effects of regression of myocyte hypertrophy and interstitial fibrosis, so these results cannot be extrapolated to indicate that regression of LVH with other antihypertensive agents will produce similar results.
One limitation of the present study is the possibility of some subjects having ischemic heart disease, which might influence QT dispersion; although no subjects had a suggestive history and all had normal ECGs and no regional wall motion abnormalities on echocardiography, silent ischemia cannot be totally excluded. Second, all of the ECGs were recorded with a standard 12-lead machine. Thus, the leads were not all recorded simultaneously, and this affects both QT dispersion and RR interval measurement. We calculated QTc dispersion rather than focusing on QT dispersion to compensate for the varying RR intervals among the leads.
QT dispersion is increased in association with an increased LVMI in hypertensive individuals. However, antihypertensive therapy with ramipril and felodipine reduces both parameters. If an increased QT dispersion is a predictor of sudden death in this group of individuals, then the importance of its reduction is evident.
Selected Abbreviations and Acronyms
|E/A ratio||=||ratio of peak flow velocity of the early filling wave to peak flow velocity of the atrial wave|
|ECG||=||electrocardiograph, electrocardiographic, electrocardiogram|
|IVRT||=||isovolumic relaxation time|
|LVH||=||left ventricular hypertrophy|
|LVMI||=||left ventricular mass index|
We would like to thank Hoechst AG and the Coronary Flow Trust for their financial support. We would also like to thank the staff of the Peart-Rose Clinic and the Cardiology Department at St Mary's Hospital for their help throughout the study.
- Received January 24, 1996.
- Revision received February 28, 1996.
- Revision received May 21, 1996.
Kannel WB. Left ventricular hypertrophy as a risk factor in arterial hypertension. Eur Heart J. 1992;13(suppl D):82-88.
Schouten EG, Dekker JM, Meppelink P, Kok FJ, Vandenbroucke JP, Pool J. QT-interval prolongation predicts cardiovascular mortality in an apparently healthy population. Circulation.. 1991;84:1516-1523.
Foale RA, Nihoyannopoulos P, McKenna W, Kleinebenne A, Nadazdin A, Rowland E, Smith G. Echocardiographic measurement of the normal adult right ventricle. Br Heart J.. 1986;56:33-44.
Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man: anatomic validation of the method. Circulation.. 1977;55:613-618.
Reichek N, Helak J, Plappert T, Sutton M, Weber K. Anatomic validation of left ventricular mass estimates from clinical two dimensional echocardiography: initial results. Circulation.. 1983;67:348-352.
Bazett HC. An analysis of time relation of electrocardiograms. Heart.. 1920;7:353-357.
Day CP, McComb JM, Matthews J, Campbell RWF. Reduction in QT dispersion by sotalol following myocardial infarction. Eur Heart J.. 1991;12:423-427.
Higham PD, Hilton CJ, Aitcheson DA, Furniss SS, Bourke JP, Campbell RWF. QT dispersion does reflect regional variation in ventricular recovery. Circulation. 1992;86(suppl I):I-392. Abstract.
Day CP, McComb JM, Campbell RWF. QT dispersion: an indication of arrhythmia risk in patients with long QT intervals. Br Heart J.. 1990;63:342-344.
Higham PD, Campbell RWF. QT dispersion. Br Heart J.. 1994;71:508-510.
Messerli FH, Grodzicki T. Hypertension, left ventricular hypertrophy, ventricular arrhythmias and sudden death. Eur Heart J. 1992;13(suppl D):66-69.
Bethge C, Motz A, Hehn V, Strauer BE. Ventricular arrhythmias in hypertensive heart disease with and without heart failure. J Cardiovasc Pharmacol. 1987;10(suppl 6):S119-S228.
Coste P, Clementy J, Besse P, Bricaud H. Left ventricular hypertrophy and ventricular dysrhythmic risk in hypertensive patients: evaluation by programmed electrical stimulation. J Hypertens. 1988;6(suppl 4):S116-S118.
Pye MP, Cobbe SM. Mechanisms of ventricular arrhythmias in cardiac failure and hypertrophy. Cardiovasc Res.. 1992;26:740-750.
Weber KT, Brilla CG, Janicki JS. Myocardial fibrosis: functional significance and regulatory factors. Cardiovasc Res.. 1993;27:341-348.