Author + information
- Received October 14, 2002
- Revision received April 30, 2003
- Accepted May 9, 2003
- Published online September 17, 2003.
- Yoram Agmon, MD*,
- Bijoy K Khandheria, MD*,* (, )
- Irene Meissner, MD†,
- Gary L Schwartz, MD‡,
- JoRean D Sicks, MS§,
- Angela J Fought, BS§,
- W.Michael O'Fallon, PhD§,
- David O Wiebers, MD† and
- A.Jamil Tajik, MD*
- ↵*Reprint requests and correspondence:
Dr. Bijoy K. Khandheria, Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA.
Objectives The study determined, in a population-based setting, whether dilatation of the thoracic aorta is an atherosclerosis-related process.
Background The role of atherosclerosis in thoracic aortic dilatation and aneurysm formation is poorly defined.
Methods The dimensions of the thoracic aorta were measured with transesophageal echocardiography in 373 subjects participating in a population-based study (median age 66 years; 52% men). The associations between clinical and laboratory atherosclerosis risk factors, aortic atherosclerotic plaques, and aortic dimensions were examined.
Results Age, male gender, and body surface area (BSA) jointly accounted for 41%, 31%, 38%, and 47% of the variability in diameters of the sinuses of Valsalva, ascending aorta, aortic arch, and descending aorta, respectively. Adjusting for age, gender, and BSA: 1) smoking was associated with a greater aortic arch diameter, and diastolic blood pressure and diabetes were each associated with a greater descending aorta diameter (p < 0.05); 2) atherosclerotic plaques in the descending aorta were associated with a greater descending aorta diameter (0.18 ± 0.08-mm increase in diameter per 1-mm increase in plaque thickness; p = 0.02); and 3) minor negative associations were noted between atherosclerotic plaques and risk factors for atherosclerosis and the dimensions of the proximal thoracic aorta. Notably, atherosclerosis risk factors and plaque variables each accounted for <2% of the variability in aortic dimensions, adjusting for age, gender, and BSA.
Conclusions Age, gender, and BSA are major determinants of thoracic aortic dimensions. Atherosclerosis risk factors and aortic atherosclerotic plaques are weakly associated with distal aortic dilatation, suggesting that atherosclerosis plays a minor role in aortic dilatation in the population.
Atherosclerosis typically causes occlusive arterial disease, but it may also result in arterial dilatation and aneurysm formation through destruction of normal arterial wall architecture (1). Although thoracic aortic aneurysms are commonly referred to as “atherosclerotic” (2), the role of atherosclerosis-related processes in thoracic aortic dilatation has not been determined.
The Stroke Prevention: Assessment of Risk in a Community (SPARC) study is an ongoing, National Institutes of Health–sponsored, population-based study designed to evaluate the prevalence of risk factors for stroke in the general population (3,4). Study participants—a sample of the Olmsted County (Minnesota) adult population—were examined with multiple modalities, including transesophageal echocardiography (TEE), thereby allowing a comprehensive evaluation of the size and morphology of the thoracic aorta. To test the hypothesis that atherosclerosis plays a role in aortic dilatation, we examined whether atherosclerosis risk factors and atherosclerotic vascular disease are associated with thoracic aortic dimensions in a population-based setting.
The study design and initial results of the first phase of SPARC have been published in detail (3,4). The initial study cohort consisted of 581 subjects, an age- and gender-stratified random sample of the Olmsted County (Minnesota) population 45 years or older. At a median interval of 4.6 years (range 3.8 to 5.4) after the initial evaluation, eligible participants were enrolled in the second phase of SPARC. Of 504 subjects eligible for the second phase (excluding 3 subjects lost to follow-up, 54 deceased, and 20 with severe medical disabilities precluding participation), 392 (78%) agreed to participate. In 388 subjects, TEE was repeated successfully (5), 373 of whom were included in the current analysis. Twelve subjects were excluded from the analysis because of aortic valve disease (bicuspid aortic valve [n = 3], aortic valve stenosis [transaortic flow velocities >2.5 m/s; n = 5], more than moderate aortic valve regurgitation [n = 1], and aortic valve prostheses [n = 3]), and three were excluded because measurements of aortic dimensions were unavailable. The clinical characteristics of the study population are presented in Table 1. This study was approved by the Mayo Clinic Institutional Review Board. Written, informed consent was obtained from all participants.
Clinical and laboratory data
Clinical data were collected by interviews and abstracting of medical records (Mayo Clinic and Olmsted Medical Center, the two primary health care providers in Olmsted County). During an office interview related to SPARC, two blood pressure (BP) measurements were taken with the patient sitting, 5 to 10 min apart, and averaged. Treatment of hypertension or hyperlipidemia was defined as the use of antihypertensive or lipid-lowering medications, respectively.
Fasting blood samples were collected (on the day TEE was performed for 97% of participants and within 20 days of TEE for all participants) and analyzed with standard, commercially available assays for plasma lipids, homocysteine, fibrinogen, and high-sensitivity C-reactive protein (high-sensitivity latex-enhanced immunoturbidimetric assay; Kamiya Biomedical Co., Seattle, Washington).
Multiplane TEE was performed, and the ascending aorta, aortic arch, and descending thoracic aorta were visualized in multiple short- and long-axis views using high-frequency (7-MHz) ultrasonographic imaging. The diameter of the aortic lumen was measured at four predefined locations (Fig. 1): sinuses of Valsalva, mid-ascending aorta, mid-aortic arch, and mid-descending thoracic aorta. All dimensions were measured at end diastole in long-axis views (sinuses of Valsalva and ascending aorta) or short-axis views (aortic arch and descending aorta). Proximal aortic dimensions (sinuses of Valsalva and ascending aorta) were measured in 373 subjects, and distal aortic dimensions were measured in the latter ∼80% of study participants (measurements of the aortic arch and descending aorta were available in 283 and 300 subjects, respectively).
The presence of aortic plaques in each aortic segment was documented. Among subjects with plaques, maximal plaque thickness and the presence of mobile debris were determined. For measurements of the aortic lumen diameter, atherosclerotic plaques (when present) were included within the vessel lumen, as demonstrated in Figure 2. The echocardiographic findings in the study population are presented in Table 1.
The associations between demographic, clinical, laboratory, and echocardiographic variables (independent variables) and aortic dimensions (dependent variables) were examined by linear regression. When necessary, variables with a skewed distribution were transformed logarithmically, and the results presented per twofold increase in these variables.
Four parallel series of analyses were performed, corresponding to the four sites of aortic diameter measurements. First, linear and quadratic functions of age and the interactions between age and gender were assessed to determine the best-fitting models (using the appearance of residuals, p values, and R2values to determine the best fit). Subsequently, the associations between body size variables (height, weight, body surface area [BSA], and body mass index [BMI]) and aortic dimensions were examined in a series of age- and gender-adjusted multivariate stepwise analyses (separate analysis for each aortic site). The body size variable that entered into the stepwise models most frequently was selected as the adjusting variable for all subsequent analyses at all aortic sites.
The associations between each atherosclerosis risk factor and aortic dimensions were examined separately, adjusting for age, gender, and body size. Subsequently, multivariate risk factor models were developed by stepwise analyses, allowing all atherosclerosis risk factors to compete for entry into the models, adjusting for age, gender, and body size. The associations between aortic atherosclerotic plaques, clinical cardiovascular disease, and aortic dimensions were examined, adjusting for age, gender, and body size and also adjusting for the respective multivariate risk factor models. The associations between inflammatory variables and aortic dimensions were examined, adjusting for age, gender, body size, and smoking status (current smoking and past smoking vs. never smoking).
The parameter estimates (±SE) of the linear regression analyses are presented, indicating the change in aortic diameter (in millimeters) per unit change in each continuous variable, or in the presence (vs. absence) of each categorical variable. The R2values, representing the proportion of variability in aortic dimensions that is explained by the respective models, were calculated.
Age, gender, body size, and aortic dimensions
The aortic dimensions at the four sites, stratified by age and gender, are presented in Table 2and Figure 3. On univariate analysis, age and male gender were associated with a greater diameter of the thoracic aorta (Table 3). The contribution of age to the aortic diameter was progressively greater in more distal aortic segments (partial R2= 0.009, 0.09, 0.12, and 0.21 for sinuses of Valsalva, ascending aorta, aortic arch, and descending aorta, respectively).
Height, weight, BSA, and BMI were associated with greater aortic dimensions (Table 3). By stepwise analyses, BSA was the body size variable that entered into the age- and gender-adjusted models most frequently; thus, it was chosen as the adjusting body size variable for all subsequent analyses. Notably, the differences in R2values between age- and gender-adjusted models that included all four body size variables and the respective models with BSA as the single adjusting body size variable were 0.02 for the sinuses of Valsalva and 0.01 for all other aortic sites, indicating that the incremental information added by the additional body size variables, beyond that explained by BSA, is minor. Overall, age, male gender, and BSA jointly accounted for 41%, 31%, 38%, and 47% of the variability in dimensions of the sinuses of Valsalva, ascending aorta, aortic arch, and descending aorta, respectively.
Atherosclerosis risk factors and aortic dimensions
The relationships between each atherosclerosis risk factor and aortic dimensions are presented in Table 4. Adjusting for age, gender, and BSA: 1) BP variables were negatively associated with the diameter of the proximal aortic segments (higher systolic BP, diastolic BP, and pulse pressure were associated with a smaller diameter of the sinuses of Valsalva, and higher pulse pressure was associated with a smaller diameter of the ascending aorta) and positively associated with the diameter of the descending aorta (higher diastolic BP and antihypertensive drug treatment, a surrogate of hypertension severity, were associated with a larger diameter of the descending aorta); 2) an atherogenic lipid profile was negatively associated with the diameter of the ascending aorta (higher high-density lipoprotein [HDL] cholesterol and apolipoprotein A-I levels were associated with a larger diameter, and higher triglycerides and apolipoprotein B-100 levels were associated with a smaller diameter of the ascending aorta); and 3) smoking and diabetes mellitus were associated with a larger diameter of the aortic arch and descending aorta, respectively. However, these associations were of minor significance, and each accounted for <2% of the variability in aortic dimensions, adjusting for age, gender, and BSA.
By multivariate analysis, systolic BP (negative association), HDL cholesterol, smoking, and diastolic BP (positive associations for the latter three variables) were independently associated with the diameter of the sinuses of Valsalva, ascending aorta, aortic arch, and descending aorta, respectively, adjusting for age, gender, and BSA (Table 4). Overall, these multivariate models accounted for 44%, 32%, 39%, and 48% of the variability in diameters of the sinuses of Valsalva, ascending aorta, aortic arch, and descending aorta, respectively, an increment of ≤3% compared with the variability explained by age, gender, and BSA alone.
Atherosclerosis, inflammation, and aortic dimensions
The relationships between aortic atherosclerotic plaques and aortic dimensions are presented in Table 5. Plaques were negatively associated with the aortic diameter at the sinuses of Valsalva and positively with the aortic diameter at the descending aorta. However, these associations were of minor significance, and each accounted for <2% of the variability in aortic dimensions, adjusting for age, gender, and BSA. The associations of aortic plaques with the diameter of the descending aorta remained significant after adjusting for the respective multivariate risk factor model (p < 0.05).
Clinical coronary artery disease (previous myocardial infarction, angina pectoris, percutaneous coronary angioplasty, coronary artery bypass graft surgery, and the combined variable of any of these clinical manifestations) was not associated with aortic dimensions, adjusting for age, gender, and BSA, except for a borderline association of angina pectoris with the diameter of the descending aorta (p = 0.07). Cerebrovascular disease (previous ischemic stroke, transient ischemic attack, carotid endarterectomy, and the combined variable of any of these clinical manifestations) was not associated with aortic dimensions, adjusting for age, gender, and BSA, except for an association of the combined variable of cerebrovascular disease with the aortic arch diameter (1.01 ± 0.46-mm increase in subjects with cerebrovascular disease; p = 0.03).
Fibrinogen and high-sensitivity C-reactive protein levels were not associated with aortic dimensions at any aortic sites, adjusting for age, gender, BSA, and smoking (p > 0.05).
Our population-based TEE study establishes reference values for thoracic aortic dimensions at multiple sites in a large sample of the general population. Age, gender, and body size account for a large proportion of the variability in aortic dimensions (31% to 47% of the variability in our cohort, a proportion greater than that previously reported in a population-based setting) (6). When these variables are considered, atherosclerosis risk factors and aortic atherosclerotic plaques are associated with larger dimensions of the distal thoracic aorta. However, these factors account for only a small proportion of the variability in aortic dimensions, suggesting that atherosclerosis plays a minor role overall in aortic dilatation. The lack of association between plasma levels of systemic inflammatory markers (which are associated with the presence and severity of aortic atherosclerotic plaques) (7)and aortic dimensions further supports these observations.
The following unique aspects of our study should be acknowledged. The SPARC study is the only large-scale, population-based TEE study performed to date. Study participants, a random sample of the Olmsted County population, are representative of a wide age range of the nonreferred adult population. Previous transthoracic echocardiographic studies have measured the dimensions of the proximal ascending aorta (6,8,9), the only aortic segment consistently visualized with this method. With TEE, we were able to measure the dimensions of the thoracic aorta at multiple sites and to examine the differential effects of various factors on the proximal and distal thoracic aorta. Moreover, TEE enabled us to assess both aortic dimensions and aortic morphology and to examine the associations between the presence and severity of aortic atherosclerotic plaques (4,7)and thoracic aortic dimensions—associations that have not been examined previously.
The relationship between age and thoracic aortic dimensions has been documented repeatedly (6,8,9). However, aging accounts only partially for the variability in aortic dimensions in the population (6,8), and as our findings suggest, age may have a differential effect on different aortic segments (a stronger relation to dimensions of distal thoracic aortic segments). Aortic size is greater in men, a difference partially related to larger body size in men (8). Various measures of body size are associated with aortic dimensions, and the strength of these associations varies at the different aortic segments (6,8,9). Nevertheless, our data suggest that the correction of aortic dimensions for BSA is a valid, simplified approach for uniformly correcting aortic dimensions for body size at all aortic sites (8).
The role of high BP as a major risk factor for thoracic aortic dilatation is unclear. Blood pressure showed no independent relation to two-dimensional measurements of the proximal aorta in normal subjects (8), but it was associated with minimal ascending aortic dilatation in hypertensive subjects (9), a subtle finding of questionable clinical importance. In the Framingham Heart Study, BP variables were inconsistently related to aortic root dimension (diastolic BP and calculated mean BP were positively related, and systolic BP and pulse pressure were negatively related), but overall these relationships were of minimal significance (6). We observed a weak positive association between BP variables and the diameter of the descending thoracic aorta, but, unexpectedly, BP was negatively associated with the diameter of the proximal aortic segments. As observed for BP, other atherosclerosis risk factors and atherosclerotic plaques showed weak positive and weak negative relationships, respectively, with the dimensions of the distal and proximal aortic segments. Together, these findings suggest that atherosclerosis-related processes have a minor role in distal thoracic aortic dilatation and may have an inverse minor relationship to the dimensions of the proximal thoracic aorta.
Atherosclerosis plays a major role in the pathogenesis of abdominal aortic aneurysms, as demonstrated by the association of atherosclerosis risk factors (10–12), atherosclerotic vascular disease (10–13), and systemic inflammatory markers (14)with abdominal aortic dilatation and aneurysm formation. However, the pathogenesis of thoracic aortic aneurysms is evidently more complex and multifactorial. Genetic abnormalities of the aortic wall structure, as in Marfan syndrome (15), are a major cause of aneurysms of the aortic root, but atherosclerosis has been hypothesized to have a role in the pathogenesis of distal thoracic aortic aneurysms (2). The frequent occurrence of both descending thoracic and abdominal aortic aneurysms (thoracoabdominal aneurysms) supports this hypothesis, but our results suggest that atherosclerosis-related mechanisms have a relatively limited role in thoracic aortic dilatation, as also suggested by a recent autopsy study (16). Atherosclerotic plaques in the thoracic aorta, including complex aortic plaques, are detected frequently in the general population (4,7), whereas thoracic aortic aneurysms are uncommon, a discrepancy further suggesting that additional mechanisms not directly related to atherosclerotic plaques are likely to have a role in the pathogenesis of thoracic aortic aneurysms. Although the explanation for the apparently counterintuitive negative association of BP (also reported by other investigators) (6,17), additional atherosclerosis risk factors, and atherosclerotic plaques with proximal aortic dimensions remains to be determined, it is plausible that atherosclerosis-related stiffening of the proximal aorta may prevent aortic dilatation due to other causes (e.g., genetic abnormalities of the aortic wall). However, in acknowledgment of the possible confounding effect of antihypertensive therapy on the relationship between BP measurements and aortic dimensions and the low TEE-observed frequency of plaques in the ascending aorta (as well as the relatively low accuracy of TEE for detecting plaques in the ascending aorta), these negative associations should be interpreted with caution.
Pathologically, aneurysmal aortic disease is primarily a disease of the media, whereas occlusive atherosclerotic vascular disease is predominantly an intimal process (18). A widely accepted hypothesis is that atherosclerosis-related inflammatory infiltrates destroy the media through proteolytic digestion of the extracellular matrix by various matrix metalloproteinases (1,18,19). However, the factors affecting the transition from intimal plaque accumulation to medial destruction (18–20), as well as the differential effect of these processes on various arterial segments (16), have not been determined.
Our study establishes the reference values and determinants of thoracic aortic dimensions in a population-based setting. Atherosclerosis risk factors and aortic atherosclerotic plaques are weakly associated with distal aortic dilatation, suggesting that atherosclerosis may have a role, albeit a relatively minor one, in aortic dilatation. Thus, additional nonatherosclerotic mechanisms of aortic dilatation and thoracic aortic aneurysm formation, including the genetic aspects of aortic remodeling (21), need to be determined.
☆ Dr. Agmon is now at the Rambam Medical Center, Haifa, Israel.
This study was supported in part by research grant NS-06663 from the National Institute of Neurological Disorders and Stroke.
- body mass index
- blood pressure
- body surface area
- high-density lipoprotein
- Stroke Prevention: Assessment of Risk in a Community study
- transesophageal echocardiography
- Received October 14, 2002.
- Revision received April 30, 2003.
- Accepted May 9, 2003.
- American College of Cardiology Foundation
- Meissner I.,
- Whisnant J.P.,
- Khandheria B.K.,
- et al.
- Agmon Y.,
- Khandheria B.K.,
- Meissner I.,
- et al.
- Agmon Y.,
- Khandheria B.K.,
- Meissner I.,
- et al.
- Vasan R.S.,
- Larson M.G.,
- Levy D.
- Agmon Y.,
- Khandheria B.K.,
- Meissner I.,
- et al.
- Alcorn H.G.,
- Wolfson S.K. Jr.,
- Sutton-Tyrrell K.,
- et al.
- Lee A.J.,
- Fowkes F.G.,
- Carson M.N.,
- et al.
- Rohde L.E.,
- Arroyo L.H.,
- Rifai N.,
- et al.
- Erbel R.,
- Alfonso F.,
- Boileau C.,
- et al.
- Jondeau G.,
- Boutouyrie P.,
- Lacolley P.,
- et al.
- Michel J.B.
- Davies M.J.
- Silence J.,
- Lupu F.,
- Collen D.,
- et al.
- Bella J.N.,
- MacCluer J.W.,
- Roman M.J.,
- et al.