From the Division of Cardiology, Veterans General Hospital, Taipei,
Taiwan (C.-H.C., S.-J.L., T.-L.H., S.-J.H., M.-S.C.); the Division of
Cardiology, Veterans General Hospital, Taichung, Taiwan and National Yang Ming
University, Taiwan (C.-T.T.); the Institute of Public Health, National
Yang-Ming University, Taiwan (P.C.); Gerontology Research Center, National
Institute on Aging, Baltimore, Md (F.O., H.S., E.L.); and the Division of
Cardiology, Johns Hopkins University School of Medicine, Baltimore, Md
(F.C.P.Y.).
Correspondence to Frank C.P. Yin, MD, PhD, Department of Biomedical Engineering, Campus Box 1097, One Brookings Dr, Washington University, St Louis, MO 63130. E-mail yin{at}biomed.wustl.edu
Methods and Results—The influence of anthropometry, cardiac size
and contractility, arterial structure and
function, as well as indices of lifestyle, physical activity, and
dietary salt intake on LVM (by two-dimensionally guided M-mode
echocardiography) was analyzed in 1315
Chinese subjects who were either normotensive or had untreated
hypertension. Effects of many cardiac and arterial factors
were assessed. In univariate analysis, almost all
measured noncardiovascular, cardiac, and
arterial variables were significantly correlated with
LVM. In multivariate linear regression
analyses, when age, sex, body habitus, fasting serum C-peptide
level, dietary salt, physical activity, and lifestyle were accounted
for, the optimum multivariate linear regression main
effects model had an adjusted model r2 of
0.740, with 98% of the model variance accounted for by the 5
independent determinants of LVM: stroke volume (49.6%),
systolic blood pressure (30.7%), contractility
(14.7%), body mass index (1.8%), and aortic root diameter (1.6%).
Other proposed arterial indices were significant
independent determinants of LVM only when blood pressure was removed
from the model and, even then, these indices not only resulted in less
powerful prediction but also accounted for only a very small percentage
of the total variance of LVM.
Conclusions—In a large population, we (1) confirmed that age,
body habitus, and some indexes of arterial structure and
function are independent determinants of LVM; (2) found aortic diameter
to be an independent structural determinant of LVM; (3) demonstrated
that the effects of the derived measures of arterial
function were small and provided no better predictive power than blood
pressure alone; and (4) showed that when the best measures of cardiac
and vascular load were included, the single most potent predictor was
an index of left ventricular size.
This study comprehensively examines the independent and interactive
impact of cardiac factors as well as arterial structure and
function on LVM while accounting for well-known
noncardiovascular modulators of LVM such as age, sex,
body size and habitus, salt intake, physical activity, and
lifestyle2 7 15 16 17 18 19 20 21 22 23 24 25 26 27 in a large number of racially
homogeneous individuals. We examined noninvasive indexes of
arterial structure and function such as aortic and carotid
artery diameters, carotid artery wall thickness, blood pressure,
arterial pulse wave velocity, carotid artery elastic
modulus, arterial compliance, peripheral
resistance and carotid augmentation index as well as cardiac functional
indexes of size, contractility, heart rate, stroke
volume, and stroke work to determine which of these factors were
independent determinants of LVM.
LV Parameters
Vascular Structural and Functional Parameters
Wall properties of the aorta and carotid arteries were indexed by the
trunk PWV, E, and carotid distensibility. PWV was calculated from
sequential nondirectional Doppler (Parks model 802) flow velocity
and a simultaneous ECG at the right carotid artery and
femoral artery.6 32 The elastic modulus was
calculated as
E=(4a2b
Noncardiovascular Modulators of LVM
The effect of insulin status was indexed by fasting serum
C-peptide39 by radioimmunoassay (GP
serum M1221, Novo). The coefficients of variation were 6.9%
(intra-assay) and 8.0% (interassay), with a minimal detection level of
0.1 nmol/L. Blood hematocrit was measured. Finally, the effect of
dietary salt intake was indexed by the sodium content in an overnight
(8 to 10 hours) urinary sodium specimen.
Statistical Analysis
Once the final main effects model was obtained, the relative importance
of each independent variable was determined by performing forward
stepwise multiple regression analysis by calculating the ratios
of individual partial r2 to the full model
r2 . Expansion of the optimal main effects
model to include each possible pairwise interaction was also evaluated.
Finally, because many of the derived vascular functional
parameters as well as contractility
contained one of the blood pressure parameters and/or LV
chamber volume, additional regression analyses were performed
with either the BP variables, LVEDV, or ESSV removed and the
derived parameters substituted in their place. Significance
in any model was accepted at the P<0.05 level.
Table 4
Substituting the other BP variables for SBP lowered the adjusted
variance, so SBP was used in all models. Although most of the derived
vascular variables were significantly correlated with LVM in the
univariate analyses, none were significant
independent predictors of LVM.
We repeated the multiple regression procedure to see whether
determinants of LV wall thickness (a direct rather than a derived
measure of LV hypertrophy) or relative wall thickness (wall
thickness/end-diastolic diameter) were similar to those for
LVM. The optimal models each contained the same 9 predictors as for LVM
from the M-mode and the same top 5 predictors (accounting for 93% of
the total model variance), that is, SBP (37%), ESSV (33%), SV (12%),
BMI (6%), and aortic diameter (5%). The adjusted model variances for
wall thickness and relative wall thickness were, however, much lower
than for LVM (0.473 and 0.477, respectively).
As shown in the remaining columns in Table 4
There might be, however, concern as to the generalizability of our
results. A recent echocardiographic study of the
determinants of LVM in an occidental population of patients on
hemodialysis found M-mode–derived LVM to depend on LVEDV and
ESSV.40 Although this study did not examine
vascular parameters, the determinants of LVM they found
were two of the strongest predictors we observed. The surprising lack
of dependence on BP in that population was attributed to the high
proportion of patients with myocardial dysfunction. Additionally, we
found another 199 separate individuals (90 men and 109 women) in the
same age range as our study population (mean age 59 years) with mean
SBP and DBP of 152 and 91 mm Hg, respectively, who were being
treated for hypertension. Because this population was not
homogeneous in terms of duration of hypertension or
duration and type of treatment, we did not include this group in our
main study. Nevertheless, they were examined in the same manner. With
LVM assessed by the M-mode formula, the overall model variance was 0.55
and the top predictors were SV (45%), ESSV (31%), SBP (10%), and BMI
(7%). Aortic diameter was not a significant predictor. With the use of
the 2-D formula for LVM, the overall model variance was 0.52 and the
top predictors were SV (69%), SBP (19%), and ESSV (6%), with BMI and
aortic diameter not significant predictors. Thus even in these two
widely differing populations, the most important independent predictors
of LV mass were essentially the same as in our main population. These
results suggest that our findings may be generalizable.
As for noncardiovascular modulators of LVM, our
results confirmed the independent relation between body mass and LVM
previously reported.2 18 23 24 In all our models
BMI was consistently among the 5 strongest predictors. There
was also a small independent effect of BSA as reported by
others.16 21 22 The BSA effect was only about
one-fifth that of BMI. Previous studies showing a relation between BSA
and LVM, however, did not account for the effects of either SV or
LVEDV. We also found an independent age-related effect on LVM as
previously reported.7 16 41 However, the age
effect was very minor, accounting for only
An increase in the vascular load-because it affects the internal and
external LV loads-must cause the LV to remodel. Despite many studies
demonstrating a relation between one or more vascular
parameters and LVM, it is still not clear which aspect(s)
of cardiac load many of these indices affect. These
parameters can be categorized as affecting BP,
arterial size, arterial wall property, and
integrated load. SBP is an integrated but nonspecific external load on
the left ventricle during ejection because it is affected by wall
stiffness, wave reflections, and transmission speed and is itself
affected by LV ejection. Despite its nonspecificity, its integrative
power conveys sufficient information about the arterial
system such that none of the more specific vascular indexes is a
significant independent predictor of LVM when SBP is in the model.
DBP, on the other hand, affects the left ventricle before ejection, and
by determining the LV pressure required to open the aortic valve,
affects LV internal workload. Yet DBP is not an independent determinant
in the presence of SBP and by itself is a weaker determinant than SBP.
This is not too surprising because DBP does not directly affect
ejection. MBP is, likewise, only a steady load component and is not as
strong a predictor as SBP. One might have thought that PP would be as
strong a predictor as SBP because it manifests the pulsatile load.
This, however, was not the case. PP appeared to index contractile
function to some extent because it could replace ESSV in the model
(albeit with lower power and still required SBP). By itself, PP led to
a slightly weaker model than SBP. Thus SBP is the strongest single BP
determinant of LVM, although the other pressure terms carry much of the
same information.
One surprising and novel finding is the importance of the aortic
diameter. In all the models we found aortic root diameter to be among
the 5 strongest predictors and as potent as BMI. This loading probably
arises from the inertia of the blood in the ascending aorta on
ejection. Previous studies addressing this issue by measuring blood
acceleration have been inconclusive.42 Recent
modeling studies of the arterial system have, however,
shown that including an inertance term more completely characterizes
arterial function than models without
inertance.43 The present study is the first
to show an independent effect of an inertial term on LV remodeling. Our
data confirm that carotid artery diameter is also an independent albeit
fairly weak determinant of LVM,12 although why
this is so is not clear. It is possible that even the smaller mass of
blood in these very proximal vessels imposes sufficient load to induce
LVM-although further study is needed to verify this.
Carotid wall thickness, elastic modulus, and PWV (which is
directly related to wall stiffness) all affect the regression models
differently, confirming that they do not all convey the same
information about arterial wall properties. The two most
direct measures of stiffness, E and PWV, are not important predictors,
except in the absence of the BP terms. This indicates that wall
stiffness effects can essentially all be accounted for by BP. This may
reflect the direct relation between aortic pressure and wall
stiffness.2 33 Carotid wall thickness is an
independent but relatively minor predictor of LVM. Even though a
thicker wall is likely to be stiffer, the fact that E is not a
predictor suggests that the wall thickness effect is not attributable
to stiffness. Rather, a thicker carotid wall may be a manifestation of
a more generalized alteration in the arterial system that
loads the left ventricle. Regardless, the combined effects of aortic
and carotid artery diameter and wall thickness may explain, in part,
the higher fraction of LV variability accounted for by our study than a
comparable recent one.2 The lack of effect of
urine sodium suggests that the effects of salt intake arise through SBP
or other parameters.
The integrated indices of arterial function (TSVR, AGI, AC,
and Ea) were not independent predictors of LVM in the presence of SBP
and ESSV. TSVR, despite comprising >90% of the impedance, was not a
predictor of LVM in either the univariate or
multivariate analyses, thereby indicating that
it has little influence on cardiac remodeling. This is concordant with
studies showing resistance is not a good index of
hemodynamic loading.44 AGI is a
useful, noninvasive index of the timing and amplitude of
peripheral reflections and hence external load on the
heart.38 Our results show, however, that it
provides no additional predictive power over that of blood pressure for
LVM.
AC enters into the models differently than either PWV or E, which
indicates that wall stiffness and compliance do not have a direct
relation. This is not surprising because AC embodies
arterial size as well as stiffness. Because AC and Ea were
predictors only if ESSV was eliminated, this suggests that they exert
their influence on LVM by contractility rather than
vascular loading. For AC this may be due to the dominant effect of SV.
The relation between Ea and contractility is not
surprising because Ea has been shown to be closely related to LV
systolic function with aging, coronary artery disease,
and heart failure.36 45 Both AC and Ea, however,
entered the multiple regression models with signs we considered
inappropriate, which suggests that they were collinear with other
parameters in the model. In fact, the coefficient for Ea in
either the univariate or the multiple regression models had
a negative sign. Why this is so is unclear but may be due to the effect
of Ea being dominated by the SV term in the denominator. The fact that
it enters into the regression models differently than the wall
stiffness indexes, and only if SV is omitted, further confirms that it
is a different manifestation of arterial loading than wall
stiffness.
The importance of SV (or LVEDV, with which SV is highly correlated) and
ESSV in determining LVM confirms previous
findings.1 2 Their dominance in the presence of
the many other potential predictors, especially the
arterial load parameters, is impressive. The
relation between LV cavity size and LVM is understandable and expected.
Myocyte stretch is a powerful stimulus leading to a variety of
responses including changes in ionic homeostasis, activation of
intracellular messenger systems, increases in various proteins and
genes, and elaboration of growth factors.15 All
of these, acting singly or in combination, can lead to cardiac
hypertrophy. The strong, independent role of reduced
contractility-even in our population with no clinical
evidence of heart failure-suggests that although cell stretch may be a
unifying explanation for the effects of increased LV size and decreased
contractility, they each elicit different stimuli at
the cellular/molecular level.
In summary, this comprehensive study in a large, racially
homogeneous population found that the most important
independent determinants of LVM were LV size, SBP, LV
contractility, aortic size, and BMI. Independent
predictors of lesser importance were carotid artery diameter and wall
thickness, BSA, age, and heart rate. Many of the popular derived
vascular load parameters were examined, and none of them
added additional predictive power over and above the more direct
parameters listed above. The 9 independent factors
identified together thus constitute the most potent description of the
impact of mechanical loading factors, probably signaling in concert
with molecular growth factors, as determinants of LVM.
Received October 21, 1997;
revision received March 18, 1998;
accepted March 26, 1998.
2.
Devereux RB, Roman MJ, de Simone G, O'Grady MJ,
Paranicas M, Yeh JL, Fabsitz RR, Howard BV. Relations of left
ventricular mass to demographic and
hemodynamic variables in American Indians: the
Strong Heart Study. Circulation. 1997;96:1416–1423.
3.
Kelly R, Hayward C, Ganis J, Daley J, Avolio A,
O'Rourke M. Noninvasive registration of the arterial
pressure pulse waveform using high-fidelity applanation tonometry.
J Vasc Med Biol. 1989;1:142–149.
4.
Sharir T, Marmor A, Ting CT, Chen JW, Liu CP, Chang
MS, Yin FCP, Kass DA. Validation of a method for noninvasive
measurement of central arterial pressure.
Hypertension. 1993;21:74–82.
5.
Salonen R, Haapanen A, Salonen JT. Measurement of
intima-media thickness of common carotid arteries with high-resolution
B-mode ultrasonography: inter- and intra-observer variability.
Ultrasound Med Biol. 1991;17:225–230.[Medline]
[Order article via Infotrieve]
6.
Eliakim M, Sapoznikov D, Weinman J. Pulse wave
velocity in healthy subjects and in patients with various disease
states. Am Heart J. 1971;82:448–457.[Medline]
[Order article via Infotrieve]
7.
Gardin JM, Siscovick D, Anton-Culver H, Lynch JC,
Smith VE, Klopfenstein HS, Bommer WJ, Fried L, O'Leary D, Manolio TA.
Sex, age, and disease affect echocardiographic left
ventricular mass and systolic function in the
free-living elderly: the cardiovascular health study.
Circulation. 1995;91:1739–1748.
8.
Kawasaki T, Sasayama S, Yagi SI, Asakawa T, Hirai T.
Non-invasive assessment of the age related changes in stiffness of
major branches of the human arteries. Cardiovasc Res. 1987;21:678–687.[Medline]
[Order article via Infotrieve]
9.
Saba PS, Roman MJ, Ganau A, Pini R, Jones EC,
Pickering TG, Devereux RB. Relationship of effective
arterial elastance to demographic and arterial
characteristics in normotensive and hypertensive adults. J
Hypertens. 1995;13:971–977.[Medline]
[Order article via Infotrieve]
10.
Dart A, Silagy C, Dewar E, Jennings G, McNeil J. Aortic
distensibility and left ventricular structure and function
in isolated systolic hypertension. Eur Heart J. 1993;14:1465–1470.
11.
Bouthier JD, De Luca N, Safar ME, Simon AC. Cardiac
hypertrophy and arterial distensibility in
essential hypertension. Am Heart J. 1985;109:1345–1352.[Medline]
[Order article via Infotrieve]
12.
Roman MJ, Saba PS, Pini R, Spitzer M, Pickering TG,
Rosen S, Alderman MH, Devereux RB. Parallel cardiac and vascular
adaptation in hypertension. Circulation. 1992;86:1909–1918.
13.
Roman MJ, Pickering TG, Pini R, Schwartz JE, Devereux
RB. Prevalence and determinants of cardiac and vascular
hypertrophy in hypertension. Hypertension. 1995;26:369–373.
14.
Messerli FH, Sundgaard-Riise K, Ventura HO, Dunn FG,
Oigman W, Frohlich FD. Clinical and hemodynamic
determinants of left ventricular dimensions. Arch
Intern Med. 1984;144:477–481.
15.
Lakatta EG. Cardiovascular regulatory
mechanisms in advanced age. Physiol Rev. 1993;73:413–467.
16.
Shub C, Klein AL, Zachariah PK, Bailey KR, Tajik
AJ. Determination of left ventricular mass by
echocardiography in a normal population: effect of
age and sex in addition to body size. Mayo Clin Proc. 1994;69:205–211.[Medline]
[Order article via Infotrieve]
17.
Goble MM, Mosteller M, Moskowitz WB, Schieken RM. Sex
differences in the determinants of left ventricular mass in
childhood: the Medical College of Virginia Twin Study.
Circulation. 1992;85:1661–1665.
18.
de Simone G, Devereux RB, Roman MJ, Alderman MH, Laragh
JH. Relation of obesity and gender to left ventricular
hypertrophy in normotensive and hypertensive adults.
Hypertension. 1994;23:600–606.
19.
Marcus R, Krause L, Weder AB, Dominguez-Mejia A, Schork
NJ, Julius S. Sex-specific determinants of increased left
ventricular mass in the Tecumseh blood pressure study.
Circulation. 1994;90:928–936.
20.
Gardin JM, Wagenknecht LE, Anton-Culver H, Flack J,
Gidding S, Kurosaki T, Wong ND, Manolio TA. Relationship of
cardiovascular risk factors to
echocardiographic left ventricular mass in
healthy young black and white adult men and women.
Circulation. 1995;92:380–387.
21.
Savage DD, Levy D, Dannenberg AL, Garrison RJ, Castelli
WP. Association of echocardiographic left
ventricular mass with body size, blood pressure, and
physical activity (The Framingham study). Am J Cardiol. 1990;65:371–376.[Medline]
[Order article via Infotrieve]
22.
Hammond IW, Devereux RB, Alderman MH, Laragh JH.
Relation of blood pressure and body build to left
ventricular mass in normotensive and hypertensive employed
adults. J Am Coll Cardiol. 1988;12:996–1004.[Abstract]
23.
Daniels SR, Kimball TR, Morrison JA, Khoury P, Witt S,
Meyer RA. Effect of lead body mass, fat mass, blood pressure, and
sexual maturation on left ventricular mass in children and
adolescents: statistical, biological, and clinical significance.
Circulation. 1995;92:3249–3254.
24.
Lauer MS, Anderson KM, Levy D. Separate and joint
influences of obesity and mild hypertension on left
ventricular mass and geometry: the Framingham heart study.
J Am Coll Cardiol. 1992;19:130–134.[Abstract]
25.
Du Cailar G, Ribstein J, Daures JP, Mimran A. Sodium
and left ventricular mass in untreated hypertensive and
normotensive subjects. Am J Physiol. 1992;263:H177–H181.
26.
Kupari M, Koskinen P, Virolainen J. Correlates of left
ventricular mass in a population sample aged 36 to 37
years: focus on lifestyle and salt intake. Circulation. 1994;89:1041–1050.
27.
Iso H, Kiyama M, Doi M, Nakanishi N, Kitamura A, Naito
Y, Sato S, Iida M, Konishi M, Shimamoto T, Komachi Y. Left
ventricular mass and subsequent blood pressure changes
among middle-aged men in rural and urban Japanese populations.
Circulation. 1994;89:1717–1724.
28.
Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations
regarding quantitation in M-mode echocardiography:
results of a survey of echocardiographic measurements.
Circulation. 1978;58:1072–1083.
29.
Devereux RB, Lutas EM, Casale PN, Kligfield P,
Eisenberg RR, Hammond IW, Miller DH, Reis G, Alderman MH, Laragh JH.
Standardization of M-mode echocardiographic left
ventricular anatomic measurements. J Am Coll
Cardiol. 1984;4:1222–1230.[Abstract]
30.
Schiller NB, Shah PM, Crawford M, DeMaria A,
Devereux RB, Feigenbaum H, Gutgesell H, Reichek N, Sahn D, Schnittger
I, Silverman NH, Tajik AJ. Recommendations for quantitation of the left
ventricle by two-dimensional echocardiography.
J Am Soc Echocardiogr. 1989;2:358–367.[Medline]
[Order article via Infotrieve]
31.
Teichholz LE, Kreulen T, Herman MV, Gorlin R.
Problems in echocardiographic volume determinations:
echocardiographic-angiographic correlations in the
presence or absence of asynergy. Am J Cardiol. 1976;37:7–11.[Medline]
[Order article via Infotrieve]
32.
Avolio AP, Deng FQ, Li WQ, Luo YF, Huang ZD, Xing LF,
O'Rourke MF. Effects of aging on arterial distensibility
in populations with high and low prevalence of hypertension: comparison
between urban and rural communities in China. Circulation. 1985;71:202–210.
33.
Cox RH. Pressure dependence of the mechanical
properties of arteries in vivo. J Physiol. 1975;229:1371–1375.
34.
Randall OS, Westerhof N, Van Den Bos GC, Alexander B.
Reliability of stroke volume to pulse pressure ratio for estimating and
detecting changes in arterial compliance. J
Hypertens. 1986;4(suppl 5):S293–S296.
35.
Ferguson JJ, Julius S, Randall OS. Stroke volume: pulse
pressure relationships in borderline hypertension: a possible indicator
of decreased arterial compliance. J
Hypertens. 1984;2(suppl 3):397–399.
36.
Sunagawa K, Maughan WL, Burkhoff D, Sagawa K. Left
ventricular interaction with arterial load
studied in isolated canine ventricle. Am J Physiol. 1983;245:H773–H780.
37.
Kelly RP, Ting C, Yang TM, Liu CP, Maughan WL, Chang
MS, Kass DA. Effective arterial elastance as index of
arterial vascular load in humans. Circulation. 1992;86:513–521.
38.
Kelly R, Hayward C, Avolio A, O'Rourke M.
Noninvasive determination of age-related changes in the human
arterial pulse. Circulation. 1989;80:1652–1659.
39.
Chen CH, Tsai ST, Chuang JH, Chang MS, Wang SP, Chou P.
Population-based study of insulin, C-peptide, and blood pressure in
Chinese with normal glucose tolerance. Am J Cardiol. 1995;76:585–588.[Medline]
[Order article via Infotrieve]
40.
Dahan M, Siohan P, Viron B, Michel C, Paillole C,
Gourgon R, Mignon F. Relationship between left ventricular
hypertrophy, myocardial contractility, and
load conditions in hemodialysis patients: an
echocardiographic study. Am J Kidney
Dis. 1997;30:780–785.[Medline]
[Order article via Infotrieve]
41.
Gerstenblith G, Fredericksen J, Yin FCP, Fortuin NJ,
Lakatta EG, Weisfeldt ML. Echocardiographic assessment
of a normal adult aging population. Circulation. 1977;56:273–278.
42.
Noble MIM, Trenchard D, Guz A. Left
ventricular ejection in conscious dogs, I: measurement and
significance of the maximum acceleration of blood from the left
ventricle. Circ Res. 1966;19:139–147.
43.
Toy SM, Melbin J, Noordergraaf A. Reduced models of
arterial systems. IEEE Trans Biomed Eng. 1985;32:174–176.[Medline]
[Order article via Infotrieve]
44.
Devereux RB. Toward a more complete understanding of
left ventricular afterload. J Am Coll
Cardiol. 1991;17:122–124.[Medline]
[Order article via Infotrieve]
45.
Kass DA, Grayson R, Marino P. Pressure-volume
analysis as a method for quantifying simultaneous
drug (amrinone) effects on arterial load and contractile
state. J Am Coll Cardiol. 1990;16:726–732.[Abstract]
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Which Arterial and Cardiac Parameters Best Predict Left Ventricular Mass?
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—Many
cardiovascular and noncardiovascular
parameters are thought to be determinants of left
ventricular mass (LVM). Complicated interactions
necessitate the simultaneous measurement and consideration
of each to determine their individual and collective impact on LVM. We
undertook such a comprehensive study.
Key Words: hypertrophy • cardiac load • arterial system • vascular load • blood pressure
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
Left
ventricular mass is determined by integrated signaling of
multiple stimuli, among which are the mechanical loads imposed on the
heart. These loads are composed of both cardiac and
arterial components. The independent roles of cardiac
internal workload and contractility have been well
documented.1 2 Because the heart and
arterial system are physically coupled, some of the
workload of the heart is also determined by the arterial
system. Elevated BP, being the easiest index of arterial
load to measure, is a well-recognized cause of increased LVM. BP is
determined by cardiac factors such as heart rate,
contractility, and stroke volume on the one hand and
arterial size, wall properties, peripheral wave
reflections, and peripheral resistance on the other.
Technological advances allowing noninvasive measurement of several
arterial factors3 4 5 6 7 8 9 have spawned
many studies associating LVM with one or more
factors.2 10 11 12 13 14 Because of complicated
interrelations among the determinants of BP, however, without a
comprehensive study, it is difficult to discern the independent
predictors of LVM.
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Study Population
The study population consisted of 1315 subjects (698 men and 617
women) in Taiwan nearly equally distributed in the 3rd to 7th age
decades and older. None of the subjects were receiving antihypertensive
therapy, none had a history of angina pectoris, peripheral
vascular disease, or diabetes, and none had significant heart disease
disclosed by echocardiography. Each participant
underwent a 2-hour cardiovascular study including a
complete medical history, anthropometric measurements, physical
examination, echocardiography, carotid tonometry,
and Doppler flow examination.
LV dimensions were obtained in all subjects by the same
experienced sonographer using a Hewlett-Packard Sonos 500 U
(Hewlett-Packard) with a 2.5 MHz transducer in accordance with the
published recommendations.28 LV wall thickness
was the average of interventricular septal and posterior
wall thickness. LVM was calculated from the 2-D guided M-mode
echocardiogram29 as well as with a 2D
echo-derived truncated ellipsoidal formula.30 LV
end-systolic and end-diastolic volumes were
calculated with both the Teichholz method31 and a
simple spherical assumption, as well as from 2D echocardiograms. LV
internal load was the product of diastolic BP and
internal LV surface area.1 An index of LV
contractility, the ESSV, was calculated
as1
Sitting brachial artery SBP and DBP were measured with
conventional sphygmomanometry. Aortic root diameter was determined by
M-mode echocardiography. We also measured on-line
from frozen, digitized images (with a 7-MHz vascular probe incorporated
in the echocardiographic unit) the right common carotid
artery systolic and diastolic diameters (Ds and Dd)
and the intimal-medial thickness of the posterior
wall.5 
P)/([b2-a2]D),33
where a and b are the internal and external radii, and P and D
are the pressure and diameter changes over the cardiac cycle. Carotid
artery distensibility (Disten) and relative distensibility (RDist) were
also calculated as D/P and D/P/D, respectively.
Additionally, AC, (the ratio of SV to PP34,35),
arterial elastance (Ea, ratio of an index of ESP to
SV)36 were measured with ESP defined as
(2SBP+DBP)/3.37 AGI was calculated from the right
common carotid arterial pressure wave contour obtained
noninvasively with a tonometer.3 38
Peripheral vascular resistance was calculated as the ratio
of MBP to cardiac output.
In addition to age and sex, the effects of most other reported
modulators of LVM were accounted for. Lifestyle (urban or rural
domicile) and physical activity level (high for farmers or laborers and
low for retirees or white collar workers), were treated as categorical
variables. BMI and BSA were calculated from measurements of weight
and height (BMI
[kg/m2]=weight[kg]/height2[m];
BSA
[m2]=0.0001x71.84xweight0.425[kg]xheight0.725[cm]),
and WHR was measured.22
Results are expressed as mean±SD. Carotid E was logarithmically
transformed because of a skewed distribution. Unpaired Student
t tests were used to compare parameter means
between men and women. To delineate the independent determinants of
LVM, multiple regression models were created. Univariate
regression was first performed between LVM and each categorical and
continuous parameter. Only those variables having a
significant correlation with LVM were included in the subsequent
multiple regression models. We entered pairwise (omitting linear
combinations) SBP and DBP, SBP and PP, and MBP and PP and found little
difference among the pairs so we used SBP and PP. In the final reduced
model, variables with nonsignificant regression coefficients
(P>0.1) in the initial models were first removed. Then
variables that exhibited a high degree of collinearity were added
or deleted iteratively. This was to achieve a model with
consistency, conciseness, the highest adjusted total
variance and the least multicollinearity (assessed with the variance
inflation factor). Thus models in which the sign of a variable was
opposite that in the univariate analysis were
rejected. In some cases we could retain only one of several closely
related variables. For example, Disten and RDist could not both be
retained because a model with both rendered one of them insignificant.
Likewise, including LVEDV rendered SV insignificant. Whenever possible,
multicollinearity effects were further minimized by using variables
not derived from another variable in the model. For example, in
calculating ESSV, the LV volume used was that obtained by 2D
echocardiographic measurements rather than either of
the volumes derived from the M-mode parameters.
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
The mean values of most parameters differed by sex
(Table 1
). Table 2 summarizes the univariate
results. Every measured variable, with the exception of activity
level, hematocrit, and TSVR, was significantly correlated with LVM,
with the 4 strongest positive correlations being with LVEDV, LV load,
SV, and SW and the strongest negative correlation with ESSV. Table 3 lists correlation coefficients among
the anthropometric, cardiac, and vascular parameters.
View this table:
[in a new window]
Table 1. Characteristics of the Study Population
View this table:
[in a new window]
Table 2. Results of Univariate Analyses
for Predictors of LVM
View this table:
[in a new window]
Table 3. Correlation Coefficients Between Body Size Indexes
and Cardiovascular Variables depicts four different main
effects multiple regression models. LV load could not be retained
because it was colinear with another variable. The first column
lists the results of the optimal model with 9 main independent
predictors of LVM (adjusted r2=0.740).
Although many variables differed by sex, sex was itself not an
independent predictor of LVM. Likewise, fasting serum C-peptide, urine
sodium content, physical activity, and lifestyle were not independent
predictors. Except for ESSV, all other variables had positive
partial regression coefficients. The relative importance of each of the
independent determinants of LVM is shown in Figure 1. By far, the most important predictor
of LVM is SV. Together with SBP, aortic diameter, ESSV, and BMI, these
5 variables accounted for 98% of the model variance. The only
interaction that added significantly to the overall variance was
SVxESSV. Adding this term, however, only increased the adjusted
variance by 3% (from 0.740 to 0.764). Figure 2 is a 3-dimensional plot illustrating
the interacting effects of these two determinants of LVM.
View this table:
[in a new window]
Table 4. Four Different Main Effects of Multiple Regression
Models of LVM With Either SV or LVEDV as an Index of LV Size
View larger version (16K):
[in a new window]
Figure 1. Relative importance of each independent
variable in the final main effects multiple regression model is
shown. Model-adjusted r2 (=0.740)
is the proportion of variance of LVM that can be explained by the full
model. Partial r2 of each individual
variable was calculated from the stepwise multiple regression
approach. Ratio of partial r2 to model
r2 is the proportion of the model variance
that can be explained by the independent variable. Aorta indicates
aortic diameter.
View larger version (80K):
[in a new window]
Figure 2. Three-dimensional graph of the interaction effect
between SV and ESSV on LVM. Negative effect of ESSV on LVM is more
pronounced at smaller values of SV. Similarly, the positive effect of
SV on LVM is more pronounced at smaller values of ESSV. 
, some of the other
parameters became significant predictors of LVM only if SV,
ESSV, or SBP were deleted. For example, when SV was omitted, LVEDV
become a predictor, but with lower overall model variance. Omitting
ESSV did not allow PP, AC, or Ea to become significant predictors when
SV was in the model. AGI, PWV, or E became significant predictors only
when SBP was deleted. However, with these three parameters,
not only was the adjusted model r2
decreased to 0.658 and 0.661 for the SV and LVEDV models, respectively,
but PWV, AGI, and E together accounted for <4% of the total model
variance.
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
Because the M-mode formula for LVM includes
parameters, for example, LV diameters, used to calculate
other parameters, there may be some concern about
independence. To verify the reliability of the predictors of LVM, we
repeated the analyses by using LVM determined by a 2D
echocardiographic truncated ellipsoid
formula30 that does not contain M-mode–derived
terms in 1238 of the 1315 subjects who had adequate data for
analysis. We found a lower overall model variance (0.59 versus
0.74) but with the same top 5 predictors of LVM accounting for 98% of
the total variance (in order of partial
r2/model r2):
SV (63%), BMI (13%), ESSV (9%), aortic diameter (7%), and SBP
(6%). This result, combined with the LV wall thickness results
presented above, provides confidence that our results
present a realistic picture of the determinants of LVM in this
population. 
0.5% of the model
variance. Apparently, most of the age-related effects are accounted for
by the other parameters in the model. Despite the
significant sex differences in many parameters, sex itself
was not an independent predictor of LVM, which suggests that
essentially all of the previously reported sex influences on
LVM2 can be accounted for by body size, weight,
and so on.
Selected Abbreviations and Acronyms
AC
=
aortic compliance
AGI
=
carotid augmentation index
BMI
=
body mass index
BP
=
blood pressure
BSA
=
body surface area
DBP
=
diastolic BP
E
=
carotid elastic modulus
Ea
=
arterial elastance
ESP
=
end-systolic pressure
ESSV
=
end-systolic meridional stress to volume ratio
LV
=
left ventricular
LVEDV
=
LV end-diastolic volume
LVM
=
LV mass
MBP
=
mean blood pressure
PP
=
pulse pressure
PWV
=
pulse wave velocity
SBP
=
systolic BP
SV
=
stroke volume
SW
=
stroke work
TSVR
=
total systemic vascular resistance
WHR
=
waist to hip ratio
2-D
=
two-dimensional
Acknowledgments
This work was supported in part by Research and Development
Contract NO1-AG-1-2118 from the National Institute on Aging to Dr Yin.
Dr Chen is a research fellow from the Division of
Cardiology, Department of Medicine, Veterans General
Hospital–Taipei and National Yang-Ming University, ROC. The authors
wish to thank all the medical staff in the Pu-Li, Kuo-Hsin, and
Kin-Chen health stations for their support of manpower and space for
this study.
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
1.
Ganau A, Devereux RB, Pickering TG, Roman MJ,
Schnall PL, Santucci S, Spitzer MC, Laragh JH. Relation of left
ventricular hemodynamic load and
contractile performance to left ventricular mass in
hypertension. Circulation. 1990;81:25–36.
This article has been cited by other articles:
 |
|
 |
|
  S. M. Farasat, C. H. Morrell, A. Scuteri, C.-T. Ting, F. C.P. Yin, H. A. Spurgeon, C.-H. Chen, E. G. Lakatta, and S. S. Najjar Pulse Pressure Is Inversely Related to Aortic Root Diameter Implications for the Pathogenesis of Systolic Hypertension Hypertension, February 1, 2008; 51(2): 196 - 202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.D. Struthers and J. Davies Should we add screening for and treating left ventricular hypertrophy to the management of all patients needing secondary prevention of cardiovascular disease? QJM, June 1, 2003; 96(6): 449 - 452. [Full Text] [PDF]
|
|
|
|
 |
|
 E. G. Lakatta and D. Levy Arterial and Cardiac Aging: Major Shareholders in Cardiovascular Disease Enterprises: Part II: The Aging Heart in Health: Links to Heart Disease Circulation, January 21, 2003; 107(2): 346 - 354. [Full Text] [PDF]
|
|
|
|
|
|
A. Scuteri, C.-H. Chen, F. C.P. Yin, T. Chih-Tai, H. A. Spurgeon, and E. G. Lakatta Functional Correlates of Central Arterial Geometric Phenotypes Hypertension, December 1, 2001; 38(6): 1471 - 1475. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Palmieri, J. N. Bella, D. K. Arnett, M. J. Roman, A. Oberman, D. W. Kitzman, P. N. Hopkins, M. Paranicas, D. C. Rao, and R. B. Devereux Aortic Root Dilatation at Sinuses of Valsalva and Aortic Regurgitation in Hypertensive and Normotensive Subjects : The Hypertension Genetic Epidemiology Network Study Hypertension, May 1, 2001; 37(5): 1229 - 1235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Q. Galvan, F. Galetta, A. Natali, E. Muscelli, A. M. Sironi, G. Cini, S. Camastra, and E. Ferrannini Insulin Resistance and Hyperinsulinemia : No Independent Relation to Left Ventricular Mass in Humans Circulation, October 31, 2000; 102(18): 2233 - 2238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Roman, A. Ganau, P. S. Saba, R. Pini, T. G. Pickering, and R. B. Devereux Impact of Arterial Stiffening on Left Ventricular Structure Hypertension, October 1, 2000; 36(4): 489 - 494. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Hayward, R. P. Kelly, and P. Collins The roles of gender, the menopause and hormone replacement on cardiovascular function Cardiovasc Res, April 1, 2000; 46(1): 28 - 49. [Full Text] [PDF]
|
|
|
|
|
|
V. Palmieri, G. de Simone, M. J. Roman, J. E. Schwartz, T. G. Pickering, and R. B. Devereux Ambulatory Blood Pressure and M;etabolic Abnormalities in Hypertensive Subjects With Inappropriately High Left Ventricular Mass Hypertension, November 1, 1999; 34(5): 1032 - 1040. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||