AJH
1990;
3:512-517
Clinical Studies of Drug Reversal of Hypertensive Left Ventricular Hypertrophy Philip R. Liebson
From the Section of Cardiology, Rush-Presbyterian-St. Luke's M e d ical C e n t e r , R u s h M e d i c a l C o l l e g e , C h i c a g o , I l l i n o i s . T h i s a r t i c l e is b a s e d o n d a t a p r e s e n t e d a t t h e A m e r i c a n S o c i e t y o f Hypertension A n n u a l Meeting, M a y 1 0 to 1 2 , 1 9 8 9 . Address c o r r e s p o n d e n c e a n d reprint requests to Dr. Philip R L i e b son, Section of Cardiology, Rush-Presbyterian-St. Luke's Medical C e n t e r , 1 6 5 3 W . C o n g r e s s P a r k w a y , C h i c a g o , Illinois 6 0 6 1 2 .
© 1990 by the American
Journal
of Hypertension,
Inc.
cates that correlation of reduction of LV mass with decrease in blood pressure has been weak (r = 0.255, Ρ < .025). Drug monotherapy studies have shown that use of calcium channel blockers and angiotensin converting enzyme inhibitors predominantly decrease LV mass. /^-Blockers reduce LV mass in approximately 50% of studies with this class of agent, and diuretics and periph eral vasodilators decrease LV mass in far fewer than 50% of studies with these drug classes. De creased LV mass has not been found to produce ad verse effects on LV systolic function, but most study participants were included without evidence of underlying heart disease at entry. Improvement in LV diastolic filling has been found with or without substantial decrease in LV mass. Unfortu nately, clinical trials assessing LV mass regression have rarely included the patients most vulnerable to possible adverse effects, that is those with established cardiac disease. The demonstration of a possible benefit or adverse effects of LVH in hypertension on morbidity and mortality must await the results of future clinical trials of large numbers of patients with prolonged serial followup. Am J Hypertens 1990;3:512-517
Left ventricular hypertrophy, diuretics, calcium channel blockers, methyldopa, /^-blockers, angiotensin converting enzyme inhibitors, vasodilators. KEY WORDS:
T
he rationale for attempting to decrease left ven tricular (LV) mass in patients with left ventricu lar hypertrophy (LVH) is the evidence that in creased mass is an independent risk factor for congestive heart failure, coronary artery disease and sudden death. This has been amply demonstrated in the
0895-7061/'90/'$3.50
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
Because left ventricular hypertrophy (LVH) is a strong risk factor for cardiac morbidity and mortal ity considerable interest has been generated con cerning the possible effects of modifying LVH in hypertensive patients. In animal models dispropor tionate regression of left ventricular (LV) mass in relation to blood pressure may produce adverse effects on coronary blood flow reserve or LV systolic function. Clinical studies of modulation of LVH with antihypertensive drug therapy have been of two types: (1) large scale community comparison trials in which the prevalence of LVH and its regression were evaluated serially by electrocardiography (ECG) with stepped care drug therapy and (2) relatively short term (less than 2 years) studies with echocardiographic (ECHO) determination of LVH. The development of ana tomically validated measurement of LV muscle mass by ECHO has demonstrated ECHO to be much more sensitive than ECG to assessment of LVH. Community comparison trials of stepped care therapy such as the Multiple Risk Factor Interven tion Trial (MRFIT) and Hypertension Detection and Followup Study (HDFP) have not had the statistical power to demonstrate significant effects on risk reduction despite evidence for significant decrease in ECG-LVH despite participant numbers in the range of 8,000 to 10,000 followed serially for 5 to 6 years. Analysis of over 60 studies of hypertensive treatment using ECHO for assessment of LVH mass changes, averaging 10 to 15 subjects/study, indi-
AJH-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
Framingham study using electrocardiographic (ECG) criteria of LVH. Increased risk for cardiovascular events has been recently demonstrated using more sensitive echocardiographic criteria for LVH. 1
2
513
tion of antihypertensive therapy with an average of 13 patients per study. In addition to these studies, the large scale HDFP and MRFIT stepped care community com parison trials provided serial data of LVH and its regres sion using ECG assessment. The validation of echo cardiographic calculation of LV mass with necropsy LV weight was first accomplished in 19 7 7 and echocardio graphic criteria for LVH based upon population studies was first reported in 1986. ECG, which was initially used to assess the presence of LVH and risk in the Fram ingham study, has been shown to be much less sensitive than echocardiography to the presence of LVH. Clinical studies of changes in the LV mass using echo evaluation have shown considerable modification in LV mass over a brief period of time, not only in relation to blood pressure changes. Detraining of runners with physiologic LVH caused a mean decrease of LV mass of 42 g / m in five weeks. Leisure activity has been corre lated with LV mass. Discontinuation of hypertensive agents has been shown to produce significant increase in LV mass on the order of 10 g / m m Hg increase over five weeks. Finally, two studies of salt restriction and weight reduction have shown regression of LVH in hy pertensive patients over 6 weeks and 21 weeks respec tively. Most of these studies have involved 15 or fewer participants. The echocardiographic (ECHO) approach to assessment of LV mass in current clinical trials in volves M-mode evaluation of LV wall thickness and dimension at end diastole with the use of a cube formula to derive LV mass. If the calculated LV mass is corrected for body surface area, LV mass index is expressed in g / m which excludes changes in LV mass due to obesity. More recently, assessment of LV mass has been ex pressed in g / m (height), which accounts for effects of body weight on LV mass. The 66 clinical trials of monotherapy and multiple drug therapy in which ECHO LV mass was assessed serially are summarized in Table 1. These 66 trials in volved a total of 1,009 participants (approximately 15.3 participants/study). In only three of these studies was LV mass a criterion for entrance into the study. Despite that, many studies have demonstrated mean baseline values for LV mass in the established range for LVH (approximately 130 to 135 g / m for men and 100 to 110 g / m for women). Of the 66 trials, 46 were studies of monotherapy, allowing evaluation of effects of individ ual classes of drug interventions on regression of LV mass. Monotherapy trials have been subdivided into those involving diuretics, calcium channel blockers, central sympatholytic agents, /^-blockers, angiotensin convert ing enzyme (ACE) inhibitors and peripheral vasodila tors. Most monotherapy studies, over 9 0 % , evaluated participants serially for one year or less. The longest trial, involving the most participants as well, was a five year multitherapy trial of 68 patients. This is the only 4,5
PATHOPHYSIOLOGIC C O N S I D E R A T I O N S
6
3
3
CLINICAL TRIALS
7
8
9
2
9
9
9
2
8
General Considerations Clinical studies of reversal of LVH in hypertension have focused primarily on a dem onstrated decrease of LV mass with antihypertensive drugs, with no substantive data on the effects of LVH regression on modifying risks for cardiovascular events. The reason for this is the large group of participants needed in a clinical trial of mild to moderate hyperten sion to demonstrate such a significant reduction in risk, unless participants were selected with LVH as an entry criterion. Even the Hypertension Detection and Followup Study (HDFP) and Multiple Risk Factor Intervention Trial (MRFIT), involving thousands of participants as sessed for over five years, could not provide enough power to show significant risk reduction in relation to reduction in LVH by ECG criteria. 4,5
Echo Assessment of Changes in LV Mass We have reviewed over 60 published clinical studies of the effects of drug or lifestyle changes in modulation of LVH in hypertension using echocardiographic criteria for LV mass published between 1977 and 1989. These studies were usually short-term trials of up to one year in dura
2
2
10
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
The evidence that LVH is a strong risk factor for cardio vascular events suggests the attractive hypothesis that reversal may beneficially modify risk. Experimental and clinical studies, however, suggest that pathophysiologic consequences of a reduction in LV mass disproportion ate to that of blood pressure reduction may lead to ad verse effects. These may include: (1) effects on LV sys tolic wall stress leading to increased myocardial oxygen consumption and decreased systolic performance; and (2) decrease in subendocardial flow reserve. The first case may arise if LV mass decreases out of proportion to blood pressure change, the second if blood pressure de creases with Utile reduction in LV mass. Animal studies have suggested that regression of LV mass with decrease in hypertensive systolic pressure levels will produce a decrease in peak LV output if the systemic pressure is again raised to those levels without a concomitant in crease in LV mass. Despite these concerns about ad verse effects of LV function, few data in clinical studies of regression of LVH have demonstrated adverse effects on systolic performance. A recent clinical study showed that LV function improved in a group of patients with antihypertensive treatment and regression of LVH after antihypertensive agents were discontinued for four weeks.
514
AJH -JUNE
LIEBSON
co
ON
r-H CO
CD ^
on
(Ν
ν©
^
CM
1990-VOL.
3, NO.
6, PART 1
i n i-H
vO
s
ix ιη co
> g c ο
CM
0
vO
tx
i-H
CO Mn
CM
^
CO
co i n
1-H i-H
00 CM
VO
in
On
^
co
vO
Ο CO
1^1 vO
in 1
1
Λ
CM
m
IX
co
CM
in co
ι-·
vol ο co
in
oo
r-l
CO
Q
οι
ο
^Ι|χ·
τ* Ο t-H t X
00
Ο
υ
ON vO
m
00
CM IX CM
in CO
vO CM
in
ON m
«3
vO
Ο CM
CD
o c ο
3 ο Η
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
r-i
A]H-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
11
9
9
9
of the results among studies. These differences include age and sex distribution, previous treatment with anti hypertensive agents, modes of assessment of hyper trophy by echocardiographic techniques, degree of hy pertrophy present at the beginning of the study, reproducibility of results, duration of therapy, compari son with a control group, and patient compliance. Very few of the studies have attempted to account for weight changes and exercise profiles. These are potentially confounding factors in studies of LV regression. Evalua tion of 37 studies between 1977 and 1986 show that in only 9 was diet evaluated and in only 6 were there attempts at weight control, usually by salt restriction. In only 1 study was weight reduction used as an indepen dent intervention. In only 1 study as well was salt re striction evaluated as an interventional modality. Of the 37 studies, 10 provided information on weight changes in serial evaluations, but these were not accounted for in evaluation of changes in LV mass. Another consideration in the evaluation of these short-term trials is the effect of previous hypertensive therapy in relation to baseline studies of LV mass. Of the 47 ECHO studies in which information on the presence or absence of antihypertensive therapy was provided, only 10 involved participants who were not on previous medications. There was usually a 1 to 8 week drug washout period for those participants on previous medi cations before baseline studies were accomplished, an average of 2 weeks. It is theoretically possible that LV remodeling by previous antihypertensive agents might have influenced the characteristics of regression for pa tients entering these clinical trials, especially since drug effects may have lasted well beyond the washout pe riod. Another point concerning methodology is the lack of a placebo group in most of these studies. Only 3 of over 60 studies evaluated had a placebo control group which was evaluated serially with the experimental group. In the majority of studies, changes in LV mass were usually assessed by paired serial data analysis of the experimen tal group. The lack of an untreated placebo group can be explained perhaps by ethical concerns about not using an antihypertensive agent in established hypertension. Some studies used normotensive controls to compare information on baseline data, but this information does not provide insight into serial changes. In large scale community control studies, such as the HDFP and MRFIT trials, the comparison group received usual com munity physician antihypertensive management. 9
9
9
Community Comparison Trials with E C G Assess ment of Changes in LVH Although the MRFIT and HDFP trials were not primarily studies of regression of LVH, the large number of participants in each study allowed for some assessment of changes in its develop-
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
study reported beyond a two year follow up of echocardiographic regression of LV mass. Table 1 demonstrates that calcium channel blockers and ACE inhibitors are usually associated with regres sion of hypertrophy. Peripheral vasodilators and di uretics do not produce regression of LV mass in most studies. /?-Blockers show a balance of regression or no change in LV mass. For the purpose of evaluation, we arbitrarily based successful reduction of LV mass as 1 0 % or greater regression from baseline values. The results reflect somewhat those of previous animal studies which demonstrated that sympatholytic agents tended to cause regression of LV mass, while diuretics and va sodilators were less likely to do so. The lack of a strong diuretic or vasodilator effect on LV mass regression is presumably due to reflex catecholamine and angioten sin stimulation, which could maintain LV mass despite the lowering of blood pressure. It is possible that diuretic treatment is relatively ineffective in decreasing LV mass, though this has not yet been substantiated. In 2 of 6 monotherapy studies of diuretics, a decrease in LV mass was accomplished (Table 1). Moreover, in 5 of 8 trials of bitherapy in which a diuretic was one of these agents, reduction of LV mass was demonstrated. Of the 3 stud ies in which no reduction occurred, 2 involved a combi nation with a peripheral vasodilator. Monotherapy studies of pure vasodilators are minimal. Only one study has evaluated a direct peripheral vasodilator (hy dralazine) and this involved only 8 participants studied for one year. Effects of /^-blocker therapy have varied consider ably, which may be expected because of heterogeneity of actions of subclasses. It might be expected that agents with intrinsic sympathomimetic activity (ISA) would be associated with less reduction of LV mass than those without this property. A group of studies of pindolol, a /^-blocker with ISA, did indeed show an increase in LV mass despite adequate reduction of blood pressure. On the other hand, one study of acebutolol, another agent with ISA, showed adequate reduction of wall mass. It is possible to demonstrate reduction of LV mass with little change in blood pressure using low doses of methyldopa. This may have some bearing on the po tential to decrease subendocardial coronary flow re serve. In 62 studies comparing decrease in blood pressure with decrease in LV mass or mass index, a weak but significant correlation was found (r = 0.255), (P < .025). This weak correlation is not surprising since factors other than blood pressure may be responsible for modulating LV mass and LV geometry, such as activity status, change in weight, and neurohumoral effects, all of which may modify LV load or wall stress indepen dent of blood pressure. Differences in cohorts and methodologic factors may partially account for the diversity
515
516
AJH-JUNE
LIEBSON
5
12
3, NO. 6,
PARTI
tion with other agents. Finally, despite a large number of participants in each of these trials, the number with baseline LVH was insufficient to determine effect of regression of hypertrophy on prognosis. CHANGES IN LEFT VENTRICULAR FUNCTION Systolic and diastolic performance of the left ventricle have also been evaluated in echocardiographic studies as part of studies of regression of hypertrophy. In most of the studies reported, absence of heart disease was a criterion for participation, since segmental wall motion abnormalities of the left ventricle could affect echocar diographic measurements and calculations. Thus, baseline ECHO criteria for systolic performance such as fractional shortening, ejection fraction and velocity of circumferential fiber shortening (Vcf) were almost in variably normal. No study suggested a significant im pairment of LV systolic performance associated with regression of LVH, but participant numbers were invar iably low to begin with and baseline performance was within normal levels. These studies did not clearly dis sociate the effects of blood pressure reduction from those of regression of LVH. To assess adequately the functional consequences of a reduction in LV mass, one must consider not only variations in arterial pressure levels, but also any concomitant change in LV wall thickness and diameter, since LV stress is determined by all three variables. In the few studies which evaluated LV stress changes associated with regression of LV mass, wall stress did not change or decreased. Noninvasive techniques have been used to measure diastolic function of the heart in hypertension. These have included radionuclide studies of LV filling or Doppler echo techniques. A slow LV filling rate has been demonstrated early in hypertension even before changes in LV mass have become evident. Moreover, diastolic function appears to be present even with nor mal systolic performance. The few studies of reversal of LVH with serial evaluation of LV filling have shown differing effects. For example, one study of the calcium channel blocker demonstrated no decrease in abnormal filling despite a 1 0 % decrease in LV mass and blood pressure. Another study using a ^-blocker demon strated parallel decrease in LV mass and improvement in LV filling rate. The prognostic significance of abnor malities of LV filling and consequences of improvement by antihypertensive agents have not yet been evaluated. 9
9
9
13
14
ONGOING TRIALS OF LVH REGRESSION IN MILD HYPERTENSION Two large ongoing trials of mild hypertension will pro vide further elucidation of effects of monotherapy of hypertension on echocardiographic assessment of re gression of LV mass.
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
ment or its regression using ECG criteria. Both studies used a stepped care antihypertensive drug approach, with a diuretic as the first agent. Community physician care served as the comparison. Most of the comparison group also received a diuretic as the first agent. In the HDFP study, 535 of the 10,940 participants (5%) had LVH at baseline. In the MRFIT trial, 8,012 participants had hypertension and of these, 189 had ECG-LVH at baseline (2.4%). Criteria for ECG diagnosis of LVH were Minnesota Code categories, including R wave am plitude and ST segment and Τ wave abnormalities. In both the MRFIT and HDFP studies, hypertensives were treated using a stepped care multitherapy approach in the intervention group, and usual community physician approach in the control group. Those on stepped care in each study received a diuretic as the first agent. A large minority of participants in the special intervention (SI) and usual care (UC) groups in MRFIT (48%) received diuretic monotherapy. Regression of ECG-LVH was assessed by decrease in R wave amplitude and/or reversion of ST segment and Τ wave abnormalities. In the HDFP study over 5 years of serial evaluation, 5 0 % of the ECGs with LVH in the stepped care group and 4 7 % in the referred care group demonstrated a return to normal. In a 6 year follow up in the MRFIT study, the prevalence of ECG-LVH de creased in 7 5 % of the SI group and in 6 0 % of the UC groups by the sixth year. Most of the decrease occurred in the first year. Reduction of systolic pressure was in dependently associated with decrease in LVH. In both studies, ECG-LVH developed in some partici pants during the study. In the HDFP trial, 1.4% of par ticipants developed LVH from totally normal ECGs at baseline and an additional 4.7% developed increased R wave amplitude alone. In the MRFIT trial, 4.2% of the SI group and 5.4% of the UC group without ECG-LVH at baseline developed LVH. Characteristics which were independently associated with development of LVH despite treatment included: (1) higher baseline blood pressures, (2) black race, and (3) lesser reduction of blood pressure. These findings suggest that LVH can develop despite antihypertensive intervention, al though specific information on drug therapy in those who developed LVH has not been published. The use of ECG for the assessment of LVH and its regression represents a "tip of the iceberg" approach, considering its lack of diagnostic sensitivity compared to echocardiography. Nonetheless, such findings in these large scale studies suggest not only a considerable re gression hypertrophy over a long period, but evidence that a small segment of patients develop hypertrophy despite therapeutic intervention. These findings also suggest that diuretics may have some effect in decreas ing LVH because of the large percentage of drug treated patients receiving a diuretic, either alone or in combina
1990-VOL
AJH-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
TABLE 2. PARTICIPANTS IN TOMHS STUDY OF MILD HYPERTENSION
REFERENCES Kannel WB, Gordon T, Offutt D: Left ventricular hyper trophy by electrocardiogram: prevalence, incidence and mortality in the Framingham study. Ann Int Med 1969;71:89-105.
136 132 131 134 135 234 902
2.
Devereux RB: Importance of left ventricular mass as a predictor of cardiovascular morbidity in hypertension. Am J Hypertens 1 9 8 9 ; 2 : 6 5 0 - 6 5 4 .
3.
Messerli FH, Oren S, Grossman E: Left ventricular hypertrophy and antihypertensive therapy. Drugs 1988;35(suppl 5 ) : 2 7 - 3 3 .
4.
Hypertension Detection and Followup Program Cooper ative Group: Five year findings of the hypertension de tection and followup program. Prevention and reversal of left ventricular hypertrophy with antihypertensive drug therapy. Hypertension 1 9 8 5 ; 7 : 1 0 5 - 1 1 2 .
The Veterans Administration Cooperative Study of Monotherapy of Hypertension is evaluating partici pants with diastolic pressures in the 95 to 109 mm Hg range. Over 900 participants will be entered into six drug treatment groups (diuretic, peripheral α-blocker, central α-blocker, calcium channel blocker, ^-blocker and ACE inhibitor) as well as a placebo control group for two years with nutritional intervention (Gottdiener, J., personal communication). The Treatment of Mild Hypertension Study (TOMHS) has entered 902 participants with blood pres sure range of 90 to 99 mm Hg in five drug treatment groups (diuretic, peripheral α-blocker, calcium channel blocker, /J-blocker and ACE inhibitor) in addition to a placebo group. Over 125 participants in each drug intervention group and over 200 in the placebo group in TOMHS have received baseline ECHO evaluations for wall mass (Table 2). This study will provide serial evalu ations for at least two years beyond the baseline period. The placebo group can be justified since each drug ther apy group as well as the placebo group also receives nutritional intervention to decrease weight when appro priate and modify salt and alcohol intake. This study will carefully assess weight changes, salt intake on the basis of diet recall and urine sodium evaluation, and activity on the basis of a leisure activity scale. Thus, several of the other factors influencing LV mass changes other than blood pressure can be evaluated. In conclusion, drug studies of the reversal of LVH in hypertension have demonstrated regression of LVH as sociated in some instances with improvement of dia stolic filling of the left ventricle. The extent of regression appears to be related to drug class. There is little evi dence for impairment of LV systolic function after re gression of LVH. On the other hand, clinical trials have had few participants with significant LVH followed for along enough time to demonstrate possible beneficial or adverse effects of regression of LV mass. In order to demonstrate such effects, future studies should focus on participants with established LVH at baseline.
5.
McMahon A, Collins G, Rautahaharju P, et al: Electro cardiographic left ventricular hypertrophy and effects of antihypertensive drug therapy in hypertensive partici pants in the Multiple Risk Factor Intervention Trial. Am J Cardiol 1 9 8 9 ; 6 3 : 2 0 2 - 2 1 0 .
6.
Devereux RB, Reichek N: Echocardiographic determina tion of left ventricular mass in man. Anatomic validation of the method. Circulation 1 9 7 7 ; 5 5 : 6 1 3 - 6 1 8 .
7.
Hammond IW, Devereux RB, Alderman MH, et al: The prevalence and correlates of echocardiographic left ven tricular hypertrophy among employed patients with uncomplicated hypertension. J Am Coll Cardiol 1986;7:639-650.
8.
Levy D, Anderson KM, Savage DD, et al: Echocardiographically detected left ventricular hypertrophy: preva lence and risk factors. Ann Int Med 1 9 8 8 ; 1 0 8 : 7 - 1 3 .
9.
Liebson PR, Savage DD: Echocardiography in hyperten sion: a review. II. Echocardiographic studies of the ef fects of antihypertensive agents on left ventricular wall mass and function. Echocardiography 1 9 8 7 ; 4 : 2 1 5 - 2 4 9 .
10.
Schlant RC, Felner JM, Blumenstein Β A, et al: Echocar diographic documentation of regression of left ventricu lar hypertrophy in patients treated with essential hyper tension. Eur Heart J 1982;suppl A:171 - 1 7 5 .
11.
Leenen FHH, Smith DL, Farkas RM, et al: Vasodilators and regression of left ventricular hypertrophy. Am J Med 1987;72:969-978.
12.
Grimm RH Jr: The drug treatment of mild hypertension in the Multiple Risk Factor Intervention Trial. A review. Drugs 2986;31(suppl 1 ) : 1 3 - 2 1 .
13.
Szlachic J, Tubau JF, Vollmer C, Massie BM: Effects of diltiazem on left ventricular mass and diastolic filling in mild to moderate hypertension. Am J Cardiol 1989;63:198-201.
14.
White WB, Schulman P, Karimeddin MK, Smith VE: Re gression of left ventricular mass is accompanied by im provement in rapid left ventricular filling following anti hypertensive therapy with metroprolol. Am Heart J 1989;117:145-150.
15.
Stamler J, Prineas RJ, Neaton JD, et al: Background and design of the New US Trial on Diet and Drug Treatment of "Mild Hypertension" (TOMHS). Am J Cardiol 1987;59:51G-60G.
15
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
1.
Number at Baseline Diuretics (chlorthalidine) j?-Blocker (acebutolol) Calcium channel blocker (amlodipine) Peripheral alpha α-blocker (doxazosin) ACE inhibitor (captopril) Placebo Total
517
1990;
3:512-517
Clinical Studies of Drug Reversal of Hypertensive Left Ventricular Hypertrophy Philip R. Liebson
From the Section of Cardiology, Rush-Presbyterian-St. Luke's M e d ical C e n t e r , R u s h M e d i c a l C o l l e g e , C h i c a g o , I l l i n o i s . T h i s a r t i c l e is b a s e d o n d a t a p r e s e n t e d a t t h e A m e r i c a n S o c i e t y o f Hypertension A n n u a l Meeting, M a y 1 0 to 1 2 , 1 9 8 9 . Address c o r r e s p o n d e n c e a n d reprint requests to Dr. Philip R L i e b son, Section of Cardiology, Rush-Presbyterian-St. Luke's Medical C e n t e r , 1 6 5 3 W . C o n g r e s s P a r k w a y , C h i c a g o , Illinois 6 0 6 1 2 .
© 1990 by the American
Journal
of Hypertension,
Inc.
cates that correlation of reduction of LV mass with decrease in blood pressure has been weak (r = 0.255, Ρ < .025). Drug monotherapy studies have shown that use of calcium channel blockers and angiotensin converting enzyme inhibitors predominantly decrease LV mass. /^-Blockers reduce LV mass in approximately 50% of studies with this class of agent, and diuretics and periph eral vasodilators decrease LV mass in far fewer than 50% of studies with these drug classes. De creased LV mass has not been found to produce ad verse effects on LV systolic function, but most study participants were included without evidence of underlying heart disease at entry. Improvement in LV diastolic filling has been found with or without substantial decrease in LV mass. Unfortu nately, clinical trials assessing LV mass regression have rarely included the patients most vulnerable to possible adverse effects, that is those with established cardiac disease. The demonstration of a possible benefit or adverse effects of LVH in hypertension on morbidity and mortality must await the results of future clinical trials of large numbers of patients with prolonged serial followup. Am J Hypertens 1990;3:512-517
Left ventricular hypertrophy, diuretics, calcium channel blockers, methyldopa, /^-blockers, angiotensin converting enzyme inhibitors, vasodilators. KEY WORDS:
T
he rationale for attempting to decrease left ven tricular (LV) mass in patients with left ventricu lar hypertrophy (LVH) is the evidence that in creased mass is an independent risk factor for congestive heart failure, coronary artery disease and sudden death. This has been amply demonstrated in the
0895-7061/'90/'$3.50
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
Because left ventricular hypertrophy (LVH) is a strong risk factor for cardiac morbidity and mortal ity considerable interest has been generated con cerning the possible effects of modifying LVH in hypertensive patients. In animal models dispropor tionate regression of left ventricular (LV) mass in relation to blood pressure may produce adverse effects on coronary blood flow reserve or LV systolic function. Clinical studies of modulation of LVH with antihypertensive drug therapy have been of two types: (1) large scale community comparison trials in which the prevalence of LVH and its regression were evaluated serially by electrocardiography (ECG) with stepped care drug therapy and (2) relatively short term (less than 2 years) studies with echocardiographic (ECHO) determination of LVH. The development of ana tomically validated measurement of LV muscle mass by ECHO has demonstrated ECHO to be much more sensitive than ECG to assessment of LVH. Community comparison trials of stepped care therapy such as the Multiple Risk Factor Interven tion Trial (MRFIT) and Hypertension Detection and Followup Study (HDFP) have not had the statistical power to demonstrate significant effects on risk reduction despite evidence for significant decrease in ECG-LVH despite participant numbers in the range of 8,000 to 10,000 followed serially for 5 to 6 years. Analysis of over 60 studies of hypertensive treatment using ECHO for assessment of LVH mass changes, averaging 10 to 15 subjects/study, indi-
AJH-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
Framingham study using electrocardiographic (ECG) criteria of LVH. Increased risk for cardiovascular events has been recently demonstrated using more sensitive echocardiographic criteria for LVH. 1
2
513
tion of antihypertensive therapy with an average of 13 patients per study. In addition to these studies, the large scale HDFP and MRFIT stepped care community com parison trials provided serial data of LVH and its regres sion using ECG assessment. The validation of echo cardiographic calculation of LV mass with necropsy LV weight was first accomplished in 19 7 7 and echocardio graphic criteria for LVH based upon population studies was first reported in 1986. ECG, which was initially used to assess the presence of LVH and risk in the Fram ingham study, has been shown to be much less sensitive than echocardiography to the presence of LVH. Clinical studies of changes in the LV mass using echo evaluation have shown considerable modification in LV mass over a brief period of time, not only in relation to blood pressure changes. Detraining of runners with physiologic LVH caused a mean decrease of LV mass of 42 g / m in five weeks. Leisure activity has been corre lated with LV mass. Discontinuation of hypertensive agents has been shown to produce significant increase in LV mass on the order of 10 g / m m Hg increase over five weeks. Finally, two studies of salt restriction and weight reduction have shown regression of LVH in hy pertensive patients over 6 weeks and 21 weeks respec tively. Most of these studies have involved 15 or fewer participants. The echocardiographic (ECHO) approach to assessment of LV mass in current clinical trials in volves M-mode evaluation of LV wall thickness and dimension at end diastole with the use of a cube formula to derive LV mass. If the calculated LV mass is corrected for body surface area, LV mass index is expressed in g / m which excludes changes in LV mass due to obesity. More recently, assessment of LV mass has been ex pressed in g / m (height), which accounts for effects of body weight on LV mass. The 66 clinical trials of monotherapy and multiple drug therapy in which ECHO LV mass was assessed serially are summarized in Table 1. These 66 trials in volved a total of 1,009 participants (approximately 15.3 participants/study). In only three of these studies was LV mass a criterion for entrance into the study. Despite that, many studies have demonstrated mean baseline values for LV mass in the established range for LVH (approximately 130 to 135 g / m for men and 100 to 110 g / m for women). Of the 66 trials, 46 were studies of monotherapy, allowing evaluation of effects of individ ual classes of drug interventions on regression of LV mass. Monotherapy trials have been subdivided into those involving diuretics, calcium channel blockers, central sympatholytic agents, /^-blockers, angiotensin convert ing enzyme (ACE) inhibitors and peripheral vasodila tors. Most monotherapy studies, over 9 0 % , evaluated participants serially for one year or less. The longest trial, involving the most participants as well, was a five year multitherapy trial of 68 patients. This is the only 4,5
PATHOPHYSIOLOGIC C O N S I D E R A T I O N S
6
3
3
CLINICAL TRIALS
7
8
9
2
9
9
9
2
8
General Considerations Clinical studies of reversal of LVH in hypertension have focused primarily on a dem onstrated decrease of LV mass with antihypertensive drugs, with no substantive data on the effects of LVH regression on modifying risks for cardiovascular events. The reason for this is the large group of participants needed in a clinical trial of mild to moderate hyperten sion to demonstrate such a significant reduction in risk, unless participants were selected with LVH as an entry criterion. Even the Hypertension Detection and Followup Study (HDFP) and Multiple Risk Factor Intervention Trial (MRFIT), involving thousands of participants as sessed for over five years, could not provide enough power to show significant risk reduction in relation to reduction in LVH by ECG criteria. 4,5
Echo Assessment of Changes in LV Mass We have reviewed over 60 published clinical studies of the effects of drug or lifestyle changes in modulation of LVH in hypertension using echocardiographic criteria for LV mass published between 1977 and 1989. These studies were usually short-term trials of up to one year in dura
2
2
10
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
The evidence that LVH is a strong risk factor for cardio vascular events suggests the attractive hypothesis that reversal may beneficially modify risk. Experimental and clinical studies, however, suggest that pathophysiologic consequences of a reduction in LV mass disproportion ate to that of blood pressure reduction may lead to ad verse effects. These may include: (1) effects on LV sys tolic wall stress leading to increased myocardial oxygen consumption and decreased systolic performance; and (2) decrease in subendocardial flow reserve. The first case may arise if LV mass decreases out of proportion to blood pressure change, the second if blood pressure de creases with Utile reduction in LV mass. Animal studies have suggested that regression of LV mass with decrease in hypertensive systolic pressure levels will produce a decrease in peak LV output if the systemic pressure is again raised to those levels without a concomitant in crease in LV mass. Despite these concerns about ad verse effects of LV function, few data in clinical studies of regression of LVH have demonstrated adverse effects on systolic performance. A recent clinical study showed that LV function improved in a group of patients with antihypertensive treatment and regression of LVH after antihypertensive agents were discontinued for four weeks.
514
AJH -JUNE
LIEBSON
co
ON
r-H CO
CD ^
on
(Ν
ν©
^
CM
1990-VOL.
3, NO.
6, PART 1
i n i-H
vO
s
ix ιη co
> g c ο
CM
0
vO
tx
i-H
CO Mn
CM
^
CO
co i n
1-H i-H
00 CM
VO
in
On
^
co
vO
Ο CO
1^1 vO
in 1
1
Λ
CM
m
IX
co
CM
in co
ι-·
vol ο co
in
oo
r-l
CO
Q
οι
ο
^Ι|χ·
τ* Ο t-H t X
00
Ο
υ
ON vO
m
00
CM IX CM
in CO
vO CM
in
ON m
«3
vO
Ο CM
CD
o c ο
3 ο Η
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
r-i
A]H-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
11
9
9
9
of the results among studies. These differences include age and sex distribution, previous treatment with anti hypertensive agents, modes of assessment of hyper trophy by echocardiographic techniques, degree of hy pertrophy present at the beginning of the study, reproducibility of results, duration of therapy, compari son with a control group, and patient compliance. Very few of the studies have attempted to account for weight changes and exercise profiles. These are potentially confounding factors in studies of LV regression. Evalua tion of 37 studies between 1977 and 1986 show that in only 9 was diet evaluated and in only 6 were there attempts at weight control, usually by salt restriction. In only 1 study was weight reduction used as an indepen dent intervention. In only 1 study as well was salt re striction evaluated as an interventional modality. Of the 37 studies, 10 provided information on weight changes in serial evaluations, but these were not accounted for in evaluation of changes in LV mass. Another consideration in the evaluation of these short-term trials is the effect of previous hypertensive therapy in relation to baseline studies of LV mass. Of the 47 ECHO studies in which information on the presence or absence of antihypertensive therapy was provided, only 10 involved participants who were not on previous medications. There was usually a 1 to 8 week drug washout period for those participants on previous medi cations before baseline studies were accomplished, an average of 2 weeks. It is theoretically possible that LV remodeling by previous antihypertensive agents might have influenced the characteristics of regression for pa tients entering these clinical trials, especially since drug effects may have lasted well beyond the washout pe riod. Another point concerning methodology is the lack of a placebo group in most of these studies. Only 3 of over 60 studies evaluated had a placebo control group which was evaluated serially with the experimental group. In the majority of studies, changes in LV mass were usually assessed by paired serial data analysis of the experimen tal group. The lack of an untreated placebo group can be explained perhaps by ethical concerns about not using an antihypertensive agent in established hypertension. Some studies used normotensive controls to compare information on baseline data, but this information does not provide insight into serial changes. In large scale community control studies, such as the HDFP and MRFIT trials, the comparison group received usual com munity physician antihypertensive management. 9
9
9
Community Comparison Trials with E C G Assess ment of Changes in LVH Although the MRFIT and HDFP trials were not primarily studies of regression of LVH, the large number of participants in each study allowed for some assessment of changes in its develop-
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
study reported beyond a two year follow up of echocardiographic regression of LV mass. Table 1 demonstrates that calcium channel blockers and ACE inhibitors are usually associated with regres sion of hypertrophy. Peripheral vasodilators and di uretics do not produce regression of LV mass in most studies. /?-Blockers show a balance of regression or no change in LV mass. For the purpose of evaluation, we arbitrarily based successful reduction of LV mass as 1 0 % or greater regression from baseline values. The results reflect somewhat those of previous animal studies which demonstrated that sympatholytic agents tended to cause regression of LV mass, while diuretics and va sodilators were less likely to do so. The lack of a strong diuretic or vasodilator effect on LV mass regression is presumably due to reflex catecholamine and angioten sin stimulation, which could maintain LV mass despite the lowering of blood pressure. It is possible that diuretic treatment is relatively ineffective in decreasing LV mass, though this has not yet been substantiated. In 2 of 6 monotherapy studies of diuretics, a decrease in LV mass was accomplished (Table 1). Moreover, in 5 of 8 trials of bitherapy in which a diuretic was one of these agents, reduction of LV mass was demonstrated. Of the 3 stud ies in which no reduction occurred, 2 involved a combi nation with a peripheral vasodilator. Monotherapy studies of pure vasodilators are minimal. Only one study has evaluated a direct peripheral vasodilator (hy dralazine) and this involved only 8 participants studied for one year. Effects of /^-blocker therapy have varied consider ably, which may be expected because of heterogeneity of actions of subclasses. It might be expected that agents with intrinsic sympathomimetic activity (ISA) would be associated with less reduction of LV mass than those without this property. A group of studies of pindolol, a /^-blocker with ISA, did indeed show an increase in LV mass despite adequate reduction of blood pressure. On the other hand, one study of acebutolol, another agent with ISA, showed adequate reduction of wall mass. It is possible to demonstrate reduction of LV mass with little change in blood pressure using low doses of methyldopa. This may have some bearing on the po tential to decrease subendocardial coronary flow re serve. In 62 studies comparing decrease in blood pressure with decrease in LV mass or mass index, a weak but significant correlation was found (r = 0.255), (P < .025). This weak correlation is not surprising since factors other than blood pressure may be responsible for modulating LV mass and LV geometry, such as activity status, change in weight, and neurohumoral effects, all of which may modify LV load or wall stress indepen dent of blood pressure. Differences in cohorts and methodologic factors may partially account for the diversity
515
516
AJH-JUNE
LIEBSON
5
12
3, NO. 6,
PARTI
tion with other agents. Finally, despite a large number of participants in each of these trials, the number with baseline LVH was insufficient to determine effect of regression of hypertrophy on prognosis. CHANGES IN LEFT VENTRICULAR FUNCTION Systolic and diastolic performance of the left ventricle have also been evaluated in echocardiographic studies as part of studies of regression of hypertrophy. In most of the studies reported, absence of heart disease was a criterion for participation, since segmental wall motion abnormalities of the left ventricle could affect echocar diographic measurements and calculations. Thus, baseline ECHO criteria for systolic performance such as fractional shortening, ejection fraction and velocity of circumferential fiber shortening (Vcf) were almost in variably normal. No study suggested a significant im pairment of LV systolic performance associated with regression of LVH, but participant numbers were invar iably low to begin with and baseline performance was within normal levels. These studies did not clearly dis sociate the effects of blood pressure reduction from those of regression of LVH. To assess adequately the functional consequences of a reduction in LV mass, one must consider not only variations in arterial pressure levels, but also any concomitant change in LV wall thickness and diameter, since LV stress is determined by all three variables. In the few studies which evaluated LV stress changes associated with regression of LV mass, wall stress did not change or decreased. Noninvasive techniques have been used to measure diastolic function of the heart in hypertension. These have included radionuclide studies of LV filling or Doppler echo techniques. A slow LV filling rate has been demonstrated early in hypertension even before changes in LV mass have become evident. Moreover, diastolic function appears to be present even with nor mal systolic performance. The few studies of reversal of LVH with serial evaluation of LV filling have shown differing effects. For example, one study of the calcium channel blocker demonstrated no decrease in abnormal filling despite a 1 0 % decrease in LV mass and blood pressure. Another study using a ^-blocker demon strated parallel decrease in LV mass and improvement in LV filling rate. The prognostic significance of abnor malities of LV filling and consequences of improvement by antihypertensive agents have not yet been evaluated. 9
9
9
13
14
ONGOING TRIALS OF LVH REGRESSION IN MILD HYPERTENSION Two large ongoing trials of mild hypertension will pro vide further elucidation of effects of monotherapy of hypertension on echocardiographic assessment of re gression of LV mass.
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
ment or its regression using ECG criteria. Both studies used a stepped care antihypertensive drug approach, with a diuretic as the first agent. Community physician care served as the comparison. Most of the comparison group also received a diuretic as the first agent. In the HDFP study, 535 of the 10,940 participants (5%) had LVH at baseline. In the MRFIT trial, 8,012 participants had hypertension and of these, 189 had ECG-LVH at baseline (2.4%). Criteria for ECG diagnosis of LVH were Minnesota Code categories, including R wave am plitude and ST segment and Τ wave abnormalities. In both the MRFIT and HDFP studies, hypertensives were treated using a stepped care multitherapy approach in the intervention group, and usual community physician approach in the control group. Those on stepped care in each study received a diuretic as the first agent. A large minority of participants in the special intervention (SI) and usual care (UC) groups in MRFIT (48%) received diuretic monotherapy. Regression of ECG-LVH was assessed by decrease in R wave amplitude and/or reversion of ST segment and Τ wave abnormalities. In the HDFP study over 5 years of serial evaluation, 5 0 % of the ECGs with LVH in the stepped care group and 4 7 % in the referred care group demonstrated a return to normal. In a 6 year follow up in the MRFIT study, the prevalence of ECG-LVH de creased in 7 5 % of the SI group and in 6 0 % of the UC groups by the sixth year. Most of the decrease occurred in the first year. Reduction of systolic pressure was in dependently associated with decrease in LVH. In both studies, ECG-LVH developed in some partici pants during the study. In the HDFP trial, 1.4% of par ticipants developed LVH from totally normal ECGs at baseline and an additional 4.7% developed increased R wave amplitude alone. In the MRFIT trial, 4.2% of the SI group and 5.4% of the UC group without ECG-LVH at baseline developed LVH. Characteristics which were independently associated with development of LVH despite treatment included: (1) higher baseline blood pressures, (2) black race, and (3) lesser reduction of blood pressure. These findings suggest that LVH can develop despite antihypertensive intervention, al though specific information on drug therapy in those who developed LVH has not been published. The use of ECG for the assessment of LVH and its regression represents a "tip of the iceberg" approach, considering its lack of diagnostic sensitivity compared to echocardiography. Nonetheless, such findings in these large scale studies suggest not only a considerable re gression hypertrophy over a long period, but evidence that a small segment of patients develop hypertrophy despite therapeutic intervention. These findings also suggest that diuretics may have some effect in decreas ing LVH because of the large percentage of drug treated patients receiving a diuretic, either alone or in combina
1990-VOL
AJH-JUNE
1990-VOL
3, NO. 6, PART
1
DRUG REVERSAL OF HYPERTENSIVE LVH
TABLE 2. PARTICIPANTS IN TOMHS STUDY OF MILD HYPERTENSION
REFERENCES Kannel WB, Gordon T, Offutt D: Left ventricular hyper trophy by electrocardiogram: prevalence, incidence and mortality in the Framingham study. Ann Int Med 1969;71:89-105.
136 132 131 134 135 234 902
2.
Devereux RB: Importance of left ventricular mass as a predictor of cardiovascular morbidity in hypertension. Am J Hypertens 1 9 8 9 ; 2 : 6 5 0 - 6 5 4 .
3.
Messerli FH, Oren S, Grossman E: Left ventricular hypertrophy and antihypertensive therapy. Drugs 1988;35(suppl 5 ) : 2 7 - 3 3 .
4.
Hypertension Detection and Followup Program Cooper ative Group: Five year findings of the hypertension de tection and followup program. Prevention and reversal of left ventricular hypertrophy with antihypertensive drug therapy. Hypertension 1 9 8 5 ; 7 : 1 0 5 - 1 1 2 .
The Veterans Administration Cooperative Study of Monotherapy of Hypertension is evaluating partici pants with diastolic pressures in the 95 to 109 mm Hg range. Over 900 participants will be entered into six drug treatment groups (diuretic, peripheral α-blocker, central α-blocker, calcium channel blocker, ^-blocker and ACE inhibitor) as well as a placebo control group for two years with nutritional intervention (Gottdiener, J., personal communication). The Treatment of Mild Hypertension Study (TOMHS) has entered 902 participants with blood pres sure range of 90 to 99 mm Hg in five drug treatment groups (diuretic, peripheral α-blocker, calcium channel blocker, /J-blocker and ACE inhibitor) in addition to a placebo group. Over 125 participants in each drug intervention group and over 200 in the placebo group in TOMHS have received baseline ECHO evaluations for wall mass (Table 2). This study will provide serial evalu ations for at least two years beyond the baseline period. The placebo group can be justified since each drug ther apy group as well as the placebo group also receives nutritional intervention to decrease weight when appro priate and modify salt and alcohol intake. This study will carefully assess weight changes, salt intake on the basis of diet recall and urine sodium evaluation, and activity on the basis of a leisure activity scale. Thus, several of the other factors influencing LV mass changes other than blood pressure can be evaluated. In conclusion, drug studies of the reversal of LVH in hypertension have demonstrated regression of LVH as sociated in some instances with improvement of dia stolic filling of the left ventricle. The extent of regression appears to be related to drug class. There is little evi dence for impairment of LV systolic function after re gression of LVH. On the other hand, clinical trials have had few participants with significant LVH followed for along enough time to demonstrate possible beneficial or adverse effects of regression of LV mass. In order to demonstrate such effects, future studies should focus on participants with established LVH at baseline.
5.
McMahon A, Collins G, Rautahaharju P, et al: Electro cardiographic left ventricular hypertrophy and effects of antihypertensive drug therapy in hypertensive partici pants in the Multiple Risk Factor Intervention Trial. Am J Cardiol 1 9 8 9 ; 6 3 : 2 0 2 - 2 1 0 .
6.
Devereux RB, Reichek N: Echocardiographic determina tion of left ventricular mass in man. Anatomic validation of the method. Circulation 1 9 7 7 ; 5 5 : 6 1 3 - 6 1 8 .
7.
Hammond IW, Devereux RB, Alderman MH, et al: The prevalence and correlates of echocardiographic left ven tricular hypertrophy among employed patients with uncomplicated hypertension. J Am Coll Cardiol 1986;7:639-650.
8.
Levy D, Anderson KM, Savage DD, et al: Echocardiographically detected left ventricular hypertrophy: preva lence and risk factors. Ann Int Med 1 9 8 8 ; 1 0 8 : 7 - 1 3 .
9.
Liebson PR, Savage DD: Echocardiography in hyperten sion: a review. II. Echocardiographic studies of the ef fects of antihypertensive agents on left ventricular wall mass and function. Echocardiography 1 9 8 7 ; 4 : 2 1 5 - 2 4 9 .
10.
Schlant RC, Felner JM, Blumenstein Β A, et al: Echocar diographic documentation of regression of left ventricu lar hypertrophy in patients treated with essential hyper tension. Eur Heart J 1982;suppl A:171 - 1 7 5 .
11.
Leenen FHH, Smith DL, Farkas RM, et al: Vasodilators and regression of left ventricular hypertrophy. Am J Med 1987;72:969-978.
12.
Grimm RH Jr: The drug treatment of mild hypertension in the Multiple Risk Factor Intervention Trial. A review. Drugs 2986;31(suppl 1 ) : 1 3 - 2 1 .
13.
Szlachic J, Tubau JF, Vollmer C, Massie BM: Effects of diltiazem on left ventricular mass and diastolic filling in mild to moderate hypertension. Am J Cardiol 1989;63:198-201.
14.
White WB, Schulman P, Karimeddin MK, Smith VE: Re gression of left ventricular mass is accompanied by im provement in rapid left ventricular filling following anti hypertensive therapy with metroprolol. Am Heart J 1989;117:145-150.
15.
Stamler J, Prineas RJ, Neaton JD, et al: Background and design of the New US Trial on Diet and Drug Treatment of "Mild Hypertension" (TOMHS). Am J Cardiol 1987;59:51G-60G.
15
Downloaded from http://ajh.oxfordjournals.org/ by guest on August 11, 2015
1.
Number at Baseline Diuretics (chlorthalidine) j?-Blocker (acebutolol) Calcium channel blocker (amlodipine) Peripheral alpha α-blocker (doxazosin) ACE inhibitor (captopril) Placebo Total
517
