(Circulation. 1995;91:734-740.)
© 1995 American Heart Association, Inc.
Articles |
From the Framingham Heart Study, Framingham, Mass (R.S.V., M.G.L., D.L.); the Divisions of Cardiology and Clinical Epidemiology, Beth Israel Hospital, Boston, Mass (D.L.); Boston (Mass) University School of Medicine (R.S.V., M.G.L., D.L.); and the National Heart, Lung, and Blood Institute, Bethesda, Md (D.L.).
| Abstract |
|---|
|
|
|---|
Methods and Results The study sample consisted of 1849 men and 2152 women in the Framingham Heart Study and Framingham Offspring Study who were free of clinically apparent cardiac disease when echocardiography was performed. Aortic root measurements were made by M-mode echocardiography using a leading-edge-to-leading-edge technique. The relations of age, height, weight, body surface area, and blood pressure variables (contemporary and those obtained 8 years before) to aortic root dimension were examined by use of sex-specific correlations and linear regression analyses. Age, height, weight, and sex emerged as the principal determinants of aortic root dimensions in adults (cumulative R2=.2085 in men and .2327 in women). The additional effect of contemporary or previous blood pressure measures was small and revealed direct associations of aortic root dimension with mean arterial and diastolic blood pressures and inverse associations with pulse and systolic blood pressures. Previous blood pressure measurements did not contribute significantly to prediction of aortic root size once contemporary blood pressure variables entered the models. Results of regression analyses using a sex-pooled data set showed that on average, the aortic root measurement in women was 2.4 mm smaller than that of men of comparable age, height, and weight. Logistic regression was used to assess the likelihood of aortic root enlargement according to blood pressure levels. After adjustment for age, height, and weight, the odds ratio of aortic dilation for a 1-SD increment in systolic pressure was 0.70 (95% CI, 0.52 to 0.95) in men and 0.79 (95% CI, 0.60 to 1.04) in women; the odds ratio for a 1-SD increment in diastolic pressure was 1.22 (95% CI, 0.91 to 1.63) in men and 1.33 (95% CI, 1.01 to 1.73) in women.
Conclusions Age, height, weight, and sex emerged as the principal determinants of aortic root dimensions. The additional influences of blood pressure measurements were small; direct associations of aortic root dimensions with mean arterial and diastolic blood pressures and inverse associations with pulse and systolic blood pressures were observed. Additional prospective studies are needed to confirm these observations and to assess the impact of aortic root dimensions on the incidence of hypertension.
Key Words: aorta echocardiography blood pressure
| Introduction |
|---|
|
|
|---|
The biological variables recognized to influence aortic root size include age, sex, indexes of body size, systolic and diastolic blood pressures, and stroke volume.8 9 10 Studies that evaluated the determinants of aortic root size, however, have not yielded uniform results. These previous studies were not population-based, and the numbers of subjects were small. No prior study included a sufficient number of subjects beyond the seventh decade of life, at which time age-related changes in the structural and functional properties of the thoracic aorta become more pronounced.11
In the present study, we examined the relations of age, height, weight, body surface area (BSA), sex, contemporary and long-term blood pressures (systolic, diastolic, mean), and pulse pressure to echocardiographically determined aortic root size in subjects from the Framingham Heart Study.
| Methods |
|---|
|
|
|---|
90 years (n=12); (5) morbid obesity defined as
body
mass index (BMI) >35 (n=119); (6) missing or unobtained blood
pressure
information (n=49); and (7) unavailable or technically inadequate
echocardiograms at the index examination (n=154).
Methods of Measurement and Definitions
Blood pressure data
were made up of two measurements of systolic
and diastolic pressures determined by the first and fifth Korotkoff
sounds, respectively, in the left arm of the seated subject obtained by
a physician using a mercury column sphygmomanometer. The following
definitions of blood pressure variables were used: (1) index systolic
blood pressure was the mean of the two systolic pressure readings
obtained at the index examination when the echocardiogram was obtained;
(2) index diastolic blood pressure was the mean of two diastolic
pressure readings obtained at the index examination; (3) index pulse
pressure was the difference between the index systolic and diastolic
blood pressures; (4) index mean arterial pressure was the sum of the
index diastolic blood pressure plus one third of the index pulse
pressure; and (5) previous blood pressures were the corresponding blood
pressure variables obtained at a clinic examination 8 years before the
index examination.
Body height and weight measurements obtained at the index examination were used to calculate BMI (weight in kilograms divided by the square of height in meters). BSA was calculated with the Du Bois14 formula.
All subjects underwent M-mode echocardiograms. More than 90% of the M-mode studies were obtained with two-dimensional (2D) echocardiographic guidance.15 All studies were performed with a 2.25 MHz, 1.25-cm diameter, unfocused transducer (K.B. Aerotech) and a Hoffrel 201 ultrasound receiver (Hoffrel Instruments) interfaced with a Honeywell 1856 strip chart recorder. Aortic root size was measured from the M-mode tracings in accordance with the American Society of Echocardiography (ASE) guidelines using a leading edge to leading edge measurement of the maximal distance between the anterior aortic root wall and the posterior aortic root wall at end diastole.16
Statistical Methods
All analyses were sex-specific. The
relations of age, height,
weight, BSA, stroke volume, and blood pressure (predictor variables) to
aortic root dimension (dependent variable) were initially examined
through calculation of simple and partial correlation
coefficients.17 Multiple linear regression analyses were
used to evaluate the relative influences of these predictor variables
on aortic root dimensions.17 Logistic
regression18 was used to assess the likelihood of aortic
root enlargement (defined as a value exceeding the 95th percentile
value) according to blood pressure variables, with adjustment for age
and anthropometric measures. A two-sided significance level of .05 was
used for each statistical test. All statistical analyses were performed
with the STATISTICAL ANALYSIS SYSTEM (SAS)19
on a SUN sparc 2 workstation.
| Results |
|---|
|
|
|---|
|
Simple and Partial Correlation Analyses
Table
2
gives the results of sex-specific
correlations. These analyses revealed highly significant but modest
correlations of each of the predictor variables with aortic root size,
except for height in women. After adjustment for age, all
anthropometric and blood pressure variables (other than systolic blood
pressure) remained significantly related to aortic root size.
|
Stepwise Linear Regression
Effects of Anthropometric
Measures on Aortic Root Dimensions
Linear regression models were
initiated with the inclusion of age
and anthropometric predictor variables. Models with (1) age, (2)
height, and (3) weight, BMI, or BSA were studied. When squares of these
variables were added to the models, only height squared in men was
statistically significant, but its contribution to the model was
negligible (the increment in R2 was .0017,
P=.046). To create a simple, parsimonious model, squared
terms of anthropometric measures were avoided. Table 3
lists the regression coefficients and the standard errors of the basic
regression model that included age, height, and weight.
|
Effects of Blood Pressure Variables
The blood
pressure variables were then added to the basic model in
a forward stepwise fashion. Both systolic and diastolic blood pressures
were entered into the model, and both were highly significant
(P<.0001). Addition of previous blood pressure terms
(variables from 8 years before) indicated that previous diastolic blood
pressures in men and women and previous systolic blood pressures in men
yielded statistically significant coefficients, but increments in
R2 were minimal. Ageblood pressure
interactions were examined; an age-related interaction term with
systolic blood pressure was significant but small in men (increment in
R2=.0036, P=.003) but not in women
(P>.40). There was no significant
agediastolic blood pressure interaction. Because in
samples this large, statistical significance can be achieved with very
small effects, predictor variables relating to previous blood pressure
and interaction terms were eliminated to avoid overfitting.
Additional
regression models were considered with age, height,
weight, and (1) systolic blood pressure, (2) diastolic blood pressure,
(3) systolic and diastolic blood pressures, (4) mean arterial
pressure, and (5) pulse pressure. Table 4
lists the regression coefficients, R2,
and incremental R2 values for these models.
Although blood pressure variables did contribute to the prediction of
aortic root size, the changes in R2 from the
basic model (age, height, weight) were small. For instance, the
inclusion of both systolic and diastolic blood pressures increased the
R2 of the model by only .0115 in men and .0042
in women. The coefficients for diastolic blood pressure were slightly
higher than those for systolic blood pressure and were similar in
magnitude in men and women. When considered together, the mathematical
signs of regression coefficients for systolic and diastolic blood
pressures were opposite, suggesting a "pulse pressure effect."
Collinearity is an unlikely explanation for this observation because
the correlation between systolic and diastolic blood pressures was only
0.49.
|
Figs 1
and 2
show observed data and
fitted regression lines for the equation
|
|
![]() |
based
on models 4 and 5 from Table 4
. Fig 1
indicates
that men and
women in the fourth quartile of mean arterial pressure have larger
predicted aortic root measurements than those in the first quartile.
Fig 2
shows that men and women in the first quartile of pulse
pressure
have larger aortic root measurements than subjects of the same sex in
the fourth quartile.
Sex-Related Differences
Although the primary analyses were
sex-specific, sex-related
interactions were explored in a sex-pooled data set. Results of these
analyses indicated that men and women have different aortic root
dimensions (holding all other variables equal and constant). On
average, the aortic root measurement in women was 2.4 mm smaller than
that of men of comparable age, height, and weight.
Aortic Root Dilatation
Multiple logistic regression was used
to determine the
contributions of blood pressure variables to the prevalence of aortic
root enlargement. Aortic root dilatation was defined as a value
exceeding the 95th percentile of the raw values of aortic root
dimensions. Logistic models were fitted that incorporated the
covariates of age, height, and weight. Blood pressure variables then
were entered in the following fashion: (1) contemporary systolic and
diastolic blood pressures, (2) contemporary mean arterial and pulse
pressures, (3) previous systolic and diastolic blood pressures, and (4)
contemporary and previous systolic and diastolic blood pressures. Table
5
gives the results. In general, contemporary and
previous systolic and diastolic blood pressures were significant
predictors of an enlarged aortic root in both sexes. After contemporary
blood pressure measurements were accounted for, previous measurements
of blood pressure did not contribute significantly. The regression
coefficients for systolic and diastolic blood pressures had opposite
mathematical signs: systolic blood pressure was inversely and diastolic
blood pressure was directly related to the probability of having an
enlarged aortic root measurement, suggesting a pulse pressure effect.
In both sexes, every 1-SD increment in pulse pressure was associated
with 25% lower odds of aortic root dilatation. When entered together
into the model, mean arterial pressure was directly related and pulse
pressure inversely related to the probability of having an enlarged
aortic root.
|
| Discussion |
|---|
|
|
|---|
Effect of Anthropometric Variables
The relations of
anthropometric variables to aortic dimensions
have been recognized for a long
time20 21 22 and were
confirmed in more recent echocardiographic studies of the aortic
root.9 10 23 24 It has been
suggested that height is the
most important determinant of aortic root size compared with other
measurements such as BSA or weight.10 24 In the
present study, height and weight (and their derivatives, BSA and
BMI) predicted aortic root dimensions. Results of linear regression
analyses revealed that a 10-cm increment in height was associated with
0.24- (men) and 0.38-mm (women) increments in aortic root size.
Corresponding increments in aortic root dimensions for a 10-kg
increment in weight were 0.87 and 0.68 mm. The correlation of aortic
root dimension with anthropometric measures yielded smaller correlation
coefficients in our study compared with those reported in the
literature.9 10 24 Failure to account for
sex differences
in the previous studies may be an important reason for this difference
because sex-pooled analyses can inflate correlations if the
distributions of the variables differ in men and women.25
Pooling of children and adults in one study10 is another
reason for the inflation of previously reported correlation
coefficients because much of the variation of aortic root size among
children is related to body growth.
Effect of Age
Age-related dilation of the aortic root has
been consistently
reported in autopsy
series26 27 28 29 30 31
and in clinical studies
where chest roentgenograms, computed tomography, and magnetic resonance
imaging22 32 33 34 were used.
In contrast, clinical
echocardiographic studies have yielded disparate results. An
age-related increase was evident in some
studies9 23 35
but not in others.10 36 Some of these studies
included a
predominantly pediatric population10 ; none included a
sufficient number of subjects beyond the seventh decade, when the
age-related structural and functional changes in the aorta become more
apparent. In the present study, aortic root dimensions increased
gradually and smoothly with age in subjects of both sexes. Results of
the linear regression models (Table 3
) revealed 0.8 (men) and
0.9 mm
(women) increases in aortic root dimension for each advancing decade of
age, after adjustment for height and weight. Such an age-related change
in vessel dimensions is seen in other vascular
territories37 38 39 40 and is
related to thinning and fracturing
of the elastic
laminae,11 41 42 43 44
presumably related to the
effects of cyclic stress.45 46 47 It has
been postulated that
because aortic distensibility declines with age, the aorta diameter
increases, so that the volume buffering capacity is
constant.39 48
Effect of Sex
While it is generally accepted that aortic root
dimensions are
smaller in women than in men,8 9 the basis for this
difference is not clear. Indexing for BSA has resulted in various
results; similar,9 larger,8 and
smaller29 values of indexed aortic root dimensions have
been reported in men compared with women. In the present study,
indexing the aortic root dimension produced variable results, depending
on the male/female ratio of the divisor selected for indexation.
Regression models that used a sex-pooled database indicated that women
on average have a 2.4 mm smaller aortic root dimension than men, even
after adjustment for anthropometric measures and age. Such sex-related
differences have also been observed in other vascular
territories.32 39 40
Effect of Blood Pressure Variables
Investigations of the
effects of systolic and diastolic pressures
on aortic root dimensions have produced inconsistent results. Direct
correlations with systolic24 and diastolic49
pressures and no significant correlations with either systolic or
diastolic pressure have been reported.9 Studies that
compared aortic root dimensions in hypertensive subjects with
normotensive age-matched control subjects have not found any consistent
difference.50 51 52
In the present
study, direct relations of aortic root size to
diastolic and mean arterial pressures were observed. The magnitude of
effect was very small after adjustment for the influence of age and
anthropometric measures. Interestingly, inverse relations of aortic
root diameter to systolic and pulse pressures were observed in the
present study. Results of linear regression models (Table 4
)
revealed that a 10 mm Hg increment in pulse pressure was associated
with a 0.25 (men) and 0.09 mm (women) decrease in aortic root size
after adjustment for age and anthropometric measures. Although the
direction of these pressure relations may appear counterintuitive, such
an inverse relation between pulse pressure and the arterial lumen is
predicted by the equation for estimation of elastic modulus of the
aorta:
![]() |
where
R is the radius of the aorta,
P is the pulse pressure,
and
R is the change in radius of the artery.53 A
growing body of evidence suggests that in contrast to mean arterial
pressure, pulse pressure may have an important effect on alteration of
vascular structure in hypertensive
subjects.54 55 56 Similar
inverse relations of pulse pressure and arterial diameter have been
observed in other vascular territories57 and may be
related to pulse pressuremediated intimomedial
hypertrophy58 59 or to reflex smooth muscle
activation in
the vessel wall, causing a decrease in vessel
diameter.60 61 In the present study, aortic
intimomedial thickness was not measured; we cannot exclude medial
hypertrophy as an explanation for the inverse relations of pulse
pressure and arterial diameter that we observed.
Although the findings of a direct relation of aortic root dimension to mean arterial pressure and an inverse relation to pulse pressure may seem paradoxical, they actually are consonant with normal physiology. This is because the hemodynamic determinants of mean arterial pressure and pulse pressure are different62 and these blood pressure variables have been shown to have differential effects on other physiological variables such as echocardiographic left ventricular mass.55 Further prospective studies are needed to confirm these observations.
Limitations
It has been stated that 2D echocardiography
offers advantages over
M-mode studies for the measurement of aortic root
dimensions.9 In the present study, M-mode measurements
of the aortic root were guided by 2D scanning, which increases the
quantitative reliability of the M-mode technique. However, only
standard left parasternal views were used to measure the aortic root
diameter. Ideally, high left parasternal, right parasternal, and other
modified views should be combined with the standard views for obtaining
the maximal aortic root diameter. Because these views are not routinely
performed in our laboratory, aortic root diameter may have been
underestimated in some patients. Interobserver and intraobserver
variabilities in the measurement of aortic root diameter were not
obtained in the present study; however, prior studies have
demonstrated the excellent reproducibility of M-mode echocardiographic
aortic root measurements.63
Because the ASE recommendations for measurement of aortic root diameter include the anterior wall of the aorta in the measure, the present study overestimates the true aortic root lumen. We have also assumed that blood pressure variables measured at the brachial artery are not substantially different from those at the aortic root. Whereas such an assumption is valid for mean arterial pressure (which remains the same throughout the vascular tree), pulse pressure is amplified at the level of the brachial artery because of earlier wave reflection,64 principally in younger subjects. We do not view this as a serious limitation of our study because other studies have demonstrated that aortic pulse pressure can be approximated to brachial pulse pressure in studies of central aortic distensibility.65
Conclusions
Aortic root dimension is highly dependent on age,
sex, and body
size measurements. The effect of blood pressure on aortic root
dimensions over and above these variables is small, with direct
associations with diastolic and mean blood pressures and inverse
associations with systolic and pulse pressures.
| Acknowledgments |
|---|
| Footnotes |
|---|
Received May 31, 1994; accepted September 23, 1994.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. M. Dart, B. A. Kingwell, C. D. Gatzka, K. Willson, Y.-L. Liang, K. L. Berry, L. M.H. Wing, C. M. Reid, P. Ryan, L. J. Beilin, et al. Smaller Aortic Dimensions Do Not Fully Account for the Greater Pulse Pressure in Elderly Female Hypertensives Hypertension, April 1, 2008; 51(4): 1129 - 1134. [Abstract] [Full Text] [PDF] |
||||
![]() |
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] |
||||
![]() |
R. S. Vasan Pathogenesis of Elevated Peripheral Pulse Pressure: Some Reflections and Thinking Forward Hypertension, January 1, 2008; 51(1): 33 - 36. [Full Text] [PDF] |
||||
![]() |
G. F. Mitchell, P. R. Conlin, M. E. Dunlap, Y. Lacourciere, J. M. O. Arnold, R. I. Ogilvie, J. Neutel, J. L. Izzo Jr, and M. A. Pfeffer Aortic Diameter, Wall Stiffness, and Wave Reflection in Systolic Hypertension Hypertension, January 1, 2008; 51(1): 105 - 111. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Segers, E. R. Rietzschel, M. L. De Buyzere, S. J. Vermeersch, D. De Bacquer, L. M. Van Bortel, G. De Backer, T. C. Gillebert, P. R. Verdonck, and on behalf of the Asklepios investigators Noninvasive (Input) Impedance, Pulse Wave Velocity, and Wave Reflection in Healthy Middle-Aged Men and Women Hypertension, June 1, 2007; 49(6): 1248 - 1255. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. F. Mitchell, C.-Y. Guo, E. J. Benjamin, M. G. Larson, M. J. Keyes, J. A. Vita, R. S. Vasan, and D. Levy Cross-Sectional Correlates of Increased Aortic Stiffness in the Community: The Framingham Heart Study Circulation, May 22, 2007; 115(20): 2628 - 2636. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. G. Svensson, E. H. Blackstone, J. Feng, D. de Oliveira, A. M. Gillinov, M. Thamilarasan, R. A. Grimm, B. Griffin, D. Hammer, T. Williams, et al. Are Marfan Syndrome and Marfanoid Patients Distinguishable on Long-Term Follow-Up? Ann. Thorac. Surg., March 1, 2007; 83(3): 1067 - 1074. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. C.Y. Tang, M. A. Coady, C. Lovoulos, A. Dardik, M. Aslan, J. A. Elefteriades, and G. Tellides Hyperplastic Cellular Remodeling of the Media in Ascending Thoracic Aortic Aneurysms Circulation, August 23, 2005; 112(8): 1098 - 1105. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Bekeredjian and P. A. Grayburn Valvular Heart Disease: Aortic Regurgitation Circulation, July 5, 2005; 112(1): 125 - 134. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. F. O'Rourke and W. W. Nichols Aortic Diameter, Aortic Stiffness, and Wave Reflection Increase With Age and Isolated Systolic Hypertension Hypertension, April 1, 2005; 45(4): 652 - 658. [Full Text] [PDF] |
||||
![]() |
C. J. Beller, M. R. Labrosse, M. J. Thubrikar, G. Szabo, F. Robicsek, and S. Hagl Increased aortic wall stress in aortic insufficiency: clinical data and computer model Eur. J. Cardiothorac. Surg., February 1, 2005; 27(2): 270 - 275. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. Pico, J. Labreuche, A. Cohen, P.-J. Touboul, P. Amarenco, and for the GENIC investigators Intracranial arterial dolichoectasia is associated with enlarged descending thoracic aorta Neurology, December 14, 2004; 63(11): 2016 - 2021. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. J. Nollen, M. Groenink, J. G.P. Tijssen, E. E. van der Wall, and B. J.M. Mulder Aortic stiffness and diameter predict progressive aortic dilatation in patients with Marfan syndrome Eur. Heart J., July 1, 2004; 25(13): 1146 - 1152. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. F. O'Rourke, M. E. Safar, W. W. Nichols, G. F. Mitchell, J. L. Izzo Jr, Y. Lacourciere, J.-P. Ouellet, J. Neutel, L. J. Kerwin, A. J. Block, et al. Proximal Aortic Diameter and Aortic Pressure-Flow Relationship in Systolic Hypertension * Response Circulation, May 25, 2004; 109(20): e227 - e228. [Full Text] [PDF] |
||||
![]() |
G. F. Mitchell, Y. Lacourciere, J.-P. Ouellet, J. L. Izzo Jr, J. Neutel, L. J. Kerwin, A. J. Block, and M. A. Pfeffer Determinants of Elevated Pulse Pressure in Middle-Aged and Older Subjects With Uncomplicated Systolic Hypertension: The Role of Proximal Aortic Diameter and the Aortic Pressure-Flow Relationship Circulation, September 30, 2003; 108(13): 1592 - 1598. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Agmon, B. K. Khandheria, I. Meissner, G. L. Schwartz, J. D. Sicks, A. J. Fought, W. M. O'Fallon, D. O. Wiebers, and A. J. Tajik Is aortic dilatation an atherosclerosis-related process?: Clinical, laboratory, and transesophageal echocardiographiccorrelates of thoracic aortic dimensions in the populationwith implications for thoracic aortic aneurysm formation J. Am. Coll. Cardiol., September 17, 2003; 42(6): 1076 - 1083. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. G. Svensson, K.-H. Kim, B. W. Lytle, and D. M. Cosgrove Relationship of aortic cross-sectional area to height ratio and the risk of aortic dissection in patients with bicuspid aortic valves J. Thorac. Cardiovasc. Surg., September 1, 2003; 126(3): 892 - 893. [Full Text] [PDF] |
||||
![]() |
V. T. Nkomo, M. Enriquez-Sarano, N. M. Ammash, L. J. Melton III, K. R. Bailey, V. Desjardins, R. A. Horn, and A. J. Tajik Bicuspid Aortic Valve Associated With Aortic Dilatation: A Community-Based Study Arterioscler. Thromb. Vasc. Biol., February 1, 2003; 23(2): 351 - 356. [Abstract] [Full Text] [PDF] |
||||