Independent Association of High Blood Pressure and Aortic Atherosclerosis
A Population-Based Study
Background—Atherosclerosis of the thoracic aorta is associated with stroke. The association between hypertension, a major risk factor for stroke, and aortic atherosclerosis has not been determined in the general population.
Methods and Results—Transesophageal echocardiography was performed in 581 subjects, a random sample of the Olmsted County (Minnesota) population aged ≥45 years participating in the Stroke Prevention: Assessment of Risk in a Community (SPARC) study. Blood pressure was assessed by multiple office measurements and 24-hour ambulatory blood pressure monitoring. The association between blood pressure variables and aortic atherosclerosis was evaluated by multiple logistic regression, adjusting for other associated variables. Among subjects with atherosclerosis, blood pressure variables associated with complex aortic atherosclerosis (protruding plaques ≥4 mm thick, mobile debris, or ulceration) were determined. Age and smoking history were independently associated with aortic atherosclerosis of any degree (P≤0.001) and with complex atherosclerosis (P=0.002), whereas sex, diabetes mellitus, and body mass index were not. Multiple systolic and pulse pressure variables (office and ambulatory), but none of the diastolic blood pressure variables, were associated with atherosclerosis and complex atherosclerosis, adjusting for age and smoking. Among subjects with atherosclerosis, the odds of complex atherosclerosis increased as ambulatory out-of-bed systolic blood pressure increased (odds ratio 1.43 per 10 mm Hg increase, 95% CI 1.10 to 1.87) and with hypertension treatment, adjusting for age and smoking history.
Conclusions—High blood pressure is independently associated with aortic atherosclerosis. Among subjects with atherosclerosis, high blood pressure is associated with complex atherosclerosis.
Systemic hypertension is a major risk factor for cardiovascular morbidity in the general population.1 Evidence supporting the role of hypertension as an atherosclerosis risk factor is based on epidemiological data demonstrating an association between high blood pressure and cardiovascular events. Recent studies have also demonstrated an association between hypertension and ultrasonographically defined carotid atherosclerosis in representative samples of the population.2 To date, the association between high blood pressure and aortic atherosclerosis has not been assessed in the general population because of the semi-invasive nature of transesophageal echocardiography (TEE), the procedure of choice for high-resolution real-time imaging of the thoracic aorta.3
Stroke Prevention: Assessment of Risk in a Community (SPARC) is an ongoing population-based study designed to determine the prevalence of potential risk factors for stroke in the general population.4 5 Study participants, a random sample of the Olmsted County (Minnesota) population, were evaluated by TEE, thus defining the prevalence of aortic atherosclerosis in the general population. Blood pressure was assessed comprehensively in the SPARC population by multiple home and office measurements and by 24-hour ambulatory blood pressure monitoring (ABPM). The current analysis was performed to evaluate the association between various components of systemic blood pressure and both the presence and severity of atherosclerosis of the thoracic aorta in the general population.
The design and initial results of the SPARC study were recently published.4 5 The SPARC sampling process was designed to randomly select 580 residents of Olmsted County, Minnesota, stratified by sex and 5 age subgroups (45 to 54, 55 to 64, 65 to 74, 75 to 84, and ≥85 years). Of 1475 residents initially selected, 230 were ineligible because of predefined exclusion criteria (terminal illness, dementia, significant disability, or esophageal disease precluding TEE), and 607 refused to participate in the study. Of the 638 who participated in a home interview, 50 refused to participate further in the study. Thus, the final SPARC study sample consisted of 588 subjects (47% of eligible). TEE was performed successfully in 581 subjects (the present study population). The medical records of a random sample of 20% of eligible subjects who refused to participate were reviewed. The comorbidities of this group were not significantly different from those of the study group,5 confirming that study participants were a representative sample of the population. The present study was approved by the Institutional Review Board. Written informed consent was obtained from all participants.
Cardiovascular risk factors were assessed by home interviews and abstracting of medical records at the Mayo Clinic and Olmsted County Medical Center, the 2 primary healthcare providers in Olmsted County. Information on serum lipid levels (obtained within 1 year of TEE) was abstracted from medical records. Hypertension treatment (defined as use of antihypertensive drugs) and smoking status were self-reported during the interview.
Blood Pressure Measurements
Multiple blood pressure measurements were obtained during a home interview and 2 office appointments related to SPARC (all referred to as “office” blood pressure measurements).5 Blood pressure was measured with subjects in the sitting position; a mercury column sphygmomanometer with an appropriate-sized cuff was used. Korotkoff phases 1 and 5 established the levels of systolic and diastolic pressures, respectively. Two systolic and 2 diastolic readings from each of the 3 visits (total of 6 systolic and diastolic readings) were averaged.
Ambulatory blood pressure readings were obtained every 10 minutes during a 24-hour ABPM period with the use of commercially available instruments. Only technically satisfactory ABPM recordings obtained within 30 days of the TEE study were used for the present analysis. Blood pressure readings were averaged for the entire monitoring period (≈24 hours) and separately for daytime (out-of-bed) and nighttime (in-bed) measurements.
For office blood pressure measurements, hypertension was defined as mean systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or antihypertensive drug therapy. For ABPM, out-of-bed hypertension was defined as mean out-of-bed systolic blood pressure ≥135 mm Hg or out-of-bed diastolic blood pressure ≥85 mm Hg; in-bed hypertension was defined as mean in-bed systolic blood pressure ≥120 mm Hg or in-bed diastolic blood pressure ≥75 mm Hg.6 Pulse pressure, the difference between systolic and diastolic blood pressures, was calculated for office, 24-hour, out-of-bed, and in-bed blood pressure measurements.
TEE was performed according to standard practice guidelines with the use of commercially available ultrasonographic instruments. The ascending aorta, aortic arch, and descending thoracic aorta were imaged in short- and long-axis views. Atherosclerosis was defined as irregular intimal thickening with increased echogenicity. Atherosclerosis was defined as complex in the presence of protruding atheroma ≥4 mm thick,4 7 8 mobile atherosclerotic debris,9 or plaque ulceration, and it was defined as simple in the absence of complex morphological features (Figure 1⇓). For the present analysis, aortic atherosclerosis was defined as the presence of atherosclerosis (of any degree) in any (at least 1) segment of the thoracic aorta.8 Complex aortic atherosclerosis was defined as the presence of complex atherosclerosis in any aortic segment.8
Two separate series of multivariate analyses were performed. The first series of analyses compared subjects with atherosclerosis of any degree with subjects without atherosclerosis. The second series of analyses were performed among subjects with atherosclerosis, comparing subjects with complex atherosclerosis with those with simple atherosclerosis. Continuous variables were compared by unpaired Student t test (normal data) or Wilcoxon rank sum test (nonnormal data). Categorical data were compared by χ2 or Fisher exact test.
Logistic regression was used to assess the impact of age, sex, comorbid conditions, various blood pressure measurements, and hypertension treatment on the odds of aortic atherosclerosis (proportion with atherosclerosis of any degree/proportion without atherosclerosis). Similarly, the odds of complex atherosclerosis among subjects with atherosclerosis (proportion with complex atherosclerosis/proportion with simple atherosclerosis) were assessed. Initially, to assess the impact of various blood pressure measurements and hypertension definitions, each blood pressure variable was examined separately, adjusting for age and other significant comorbid conditions. Subsequently, the final age-adjusted models were identified by stepwise logistic regression, allowing all office blood pressure variables to compete for entry into the models. A value of P=0.05 was required to enter and to leave the stepwise models. All 2-way interactions between resulting variables were analyzed.
Initially, the total study population was examined only with the use of data available for the entire study population (including office blood pressure measurements). Subsequently, the subgroup for which ABPM data were available was examined. The impact of missing ABPM data on the odds of atherosclerosis was assessed by creating an indicator of subjects missing these data. Adjusting for age, the association between this variable and aortic atherosclerosis was examined by use of logistic regression. Likewise, the interactions between this indicator and the independent variables in the final model were examined.
The age and sex distribution of the SPARC population and the percentage of simple and complex aortic atherosclerosis in the various age and sex strata are presented in Figures 2⇓ and 3⇓, respectively. The distribution of aortic atherosclerosis and complex aortic atherosclerosis within the segments of the thoracic aorta is presented in Figure 4⇓. Aortic atherosclerosis of any degree was identified in 298 study participants (51.3%). Complex atherosclerosis was present in 44 subjects (14.8% of subjects with aortic atherosclerosis, 7.6% of the total study population).
Descriptive statistics of variables potentially associated with aortic atherosclerosis and complex atherosclerosis are presented in Table 1⇓. The estimated odds ratios of atherosclerosis and complex atherosclerosis, associated with selected variables, are presented in Table 2⇓. Age was significantly associated with increased risk of atherosclerosis and complex atherosclerosis. In subjects aged <75 years, the odds of atherosclerosis increased nearly 4-fold for every 10-year increase in age; beyond age 75, there was no significantly increased risk. Among subjects with atherosclerosis, the odds of complex atherosclerosis increased >2-fold for every 10-year increase in age.
Total Study Population (581 Subjects)
Predictors of Aortic Atherosclerosis
Pulse pressure, systolic blood pressure, and hypertension treatment were significantly associated with atherosclerosis, adjusting for age and smoking history, the 2 non–blood pressure variables associated with atherosclerosis (all P≤0.05). Stepwise logistic regression identified age, smoking history, and office pulse pressure as independently associated with aortic atherosclerosis (Table 3⇓). The odds of atherosclerosis increased by 23% per 10 mm Hg increase in pulse pressure.
Predictors of Complex Aortic Atherosclerosis
Among those with aortic atherosclerosis, the 44 subjects with complex atherosclerosis were compared with the 254 subjects with simple atherosclerosis. Pulse pressure, systolic blood pressure, and hypertension treatment were significantly associated with complex aortic atherosclerosis, adjusting for age and smoking history (all P≤0.01). Stepwise logistic regression identified age, smoking history, office pulse pressure, and hypertension treatment as independently associated with complex aortic atherosclerosis (Table 4⇓). Among those with atherosclerosis, the odds of complex atherosclerosis increased by 31% per 10 mm Hg increase in pulse pressure.
Subgroup With ABPM Data (473 Subjects)
Predictors of Aortic Atherosclerosis
Technically adequate ABPM recordings performed within 30 days of TEE were available for 473 study participants. After adjustment for age, an indicator of missing ABPM data was not significantly associated with atherosclerosis. Furthermore, this indicator did not interact with the other variables associated with atherosclerosis, suggesting that for these variables the odds of aortic atherosclerosis were the same in this subgroup as in the total study population.
Age and smoking were significantly associated with aortic atherosclerosis. The odds ratios of atherosclerosis for the various blood pressure variables, adjusted for age and smoking history (the 2 non–blood pressure variables associated with atherosclerosis), are presented in Table 5⇓. Twenty-four-hour systolic blood pressure, out-of-bed systolic blood pressure, in-bed hypertension, and hypertension treatment were associated with atherosclerosis, adjusting for age and smoking history. None of the diastolic blood pressure variables or the standard definition of hypertension (based on office blood pressure measurements) was significantly associated with atherosclerosis. Stepwise logistic regression identified age, smoking history, in-bed hypertension, and hypertension treatment as independently associated with aortic atherosclerosis (Table 6⇓). A significant interaction was observed between in-bed hypertension and hypertension treatment, the odds of atherosclerosis depending jointly on in-bed hypertension and hypertension treatment. Only subjects with both in-bed hypertension and hypertension treatment had increased odds of atherosclerosis: the odds were 3.5 times higher for these subjects than for those with neither of these conditions. The interaction between in-bed hypertension and hypertension treatment is probably related, in part, to the higher in-bed blood pressure (systolic blood pressure and pulse pressures) in treated in-bed hypertensives than in nontreated in-bed hypertensives (data not shown). When categorical variables were not included in the stepwise model (Table 6⇓), out-of-bed systolic blood pressure emerged as the blood pressure variable independently associated with aortic atherosclerosis. The odds of atherosclerosis increased by 18% per 10 mm Hg increase in out-of-bed systolic blood pressure.
Predictors of Complex Aortic Atherosclerosis
ABPM data were available for 230 of the 298 subjects with aortic atherosclerosis. An indicator of missing ABPM data was not associated with complex atherosclerosis and did not interact with the other variables associated with complex atherosclerosis.
Age and smoking were significantly associated with complex aortic atherosclerosis. The odds ratios of complex atherosclerosis for the various blood pressure variables, adjusting for age and smoking history, are presented in Table 5⇑. Multiple systolic blood pressure and pulse pressure variables (office, 24-hour, and out-of-bed ambulatory blood pressure measurements) and hypertension treatment were associated with complex atherosclerosis, adjusting for age and smoking history. None of the diastolic blood pressure variables or the standard definition of hypertension (based on office blood pressure measurements) was significantly associated with complex atherosclerosis. Stepwise logistic regression identified age, smoking history, out-of-bed systolic blood pressure, and hypertension treatment as independently associated with complex aortic atherosclerosis (Table 7⇓). Among those with atherosclerosis, the odds of complex atherosclerosis increased by 43% per 10 mm Hg increase in out-of-bed systolic blood pressure. The odds of complex atherosclerosis were >2 times greater for those with hypertension treatment than for those without hypertension treatment.
The SPARC study determines the association between high blood pressure and both the presence and severity of aortic atherosclerosis in the general population. Our detailed analysis of multiple blood pressure variables allows us to conclude the following: (1) Elevated systolic blood pressure and pulse pressure (which is quantitatively determined primarily by systolic blood pressure) are associated with aortic atherosclerosis, whereas diastolic blood pressure is not. (2) The association between high blood pressure and aortic atherosclerosis is apparent by both office (nonambulatory) and ambulatory blood pressure measurements. (3) High blood pressure is associated with both the presence of aortic atherosclerosis and its severity (ie, the presence of complex morphological features). (4) Although hypertension treatment is a general marker of hypertension severity, the association between hypertension treatment and atherosclerosis was independent of blood pressure level in some of our models, suggesting that blood pressure measurements alone may underestimate the true association between blood pressure and aortic atherosclerosis in treated hypertensive subjects.
An association between hypertension and aortic atherosclerosis has been demonstrated by autopsy studies.10 11 Previous echocardiographic studies addressing the association between hypertension and aortic atherosclerosis have yielded nonuniform results, probably because of patient selection, nonstandardized blood pressure measurements, and variable definitions of hypertension. An association between hypertension and aortic atherosclerosis and between pulse pressure and aortic atherosclerosis has been observed in selected patients with valvular disease12 and with atrial fibrillation,8 respectively. However, this association was not consistently observed in other studies.13 14 15 In contrast to all previously published TEE studies, the SPARC study enabled us to uniquely assess both aortic anatomy and blood pressure in a large nonreferred population, representative of the general population in a well-defined geographic area. Moreover, blood pressure was assessed in detail by multiple standardized nonambulatory and ambulatory measurements, thus enabling us to determine the association between multiple blood pressure variables and aortic atherosclerosis.
Sex was not associated with aortic atherosclerosis in the present study, probably because of age distribution of the SPARC population (mainly postmenopausal women). Diabetes mellitus was not independently associated with aortic atherosclerosis, probably because of the strong association of diabetes with age. We did not observe an association between serum lipids levels and aortic atherosclerosis. However, serum lipid levels were available in only a subgroup of our study population (Table 1⇑); therefore, we could not definitely exclude an association between hyperlipidemia and aortic atherosclerosis. An indicator of subjects missing lipid measurements was not associated with aortic atherosclerosis and did not interact with the variables associated with atherosclerosis in the final models. In addition, there were no significant differences in blood pressure measurements between subjects with or without serum lipid measurements (data not shown). These findings suggest that the associations observed between blood pressure and aortic atherosclerosis were not affected by the availability of serum lipid measurements. Similar results were obtained in an analysis of a larger subgroup of subjects in whom more remote (within 5 years of TEE) serum lipid measurements were available (data not shown), further strengthening the validity of our results in a smaller subgroup (Tables 1⇑ and 2⇑). Although an association between hyperlipidemia and aortic atherosclerosis has been previously described,16 this association was not demonstrated in a recently published large echocardiographic study.8
The association between high blood pressure and aortic atherosclerosis does not imply a cause-and-effect relationship. Hypertension may predispose to aortic atherosclerosis, as hypothesized for other atherosclerotic manifestations.17 Alternatively, aortic atherosclerosis may be associated with increased stiffness (reduced compliance) of the proximal aortic segments (the segments examined by TEE), resulting in secondary elevation of systemic blood pressure.13 18 The association between systolic blood pressure and pulse pressure (but not diastolic pressure) and aortic atherosclerosis supports the later hypothesis.
The association between high blood pressure and aortic atherosclerosis and the association between aortic atherosclerosis and stroke3 offer an additional potential mechanism of stroke in association with hypertension. The currently observed associations between systolic blood pressure, pulse pressure, and aortic atherosclerosis support the previously described associations between systolic hypertension and stroke19 and between pulse pressure and cardiovascular events.20 The findings of the present study should form the basis for future studies designed to evaluate the role of hypertension treatment in prevention of aortic atherosclerosis (as currently evaluated for carotid atherosclerosis21 ) and its potential embolic complications.
This study was supported in part by research grants NS06663 and AR30582 from the US Public Health Service.
Presented in part at the 72nd Scientific Sessions of the American Heart Association, Atlanta, Ga, November 7–10, 1999, and published in abstract form (Circulation 1999;100[suppl I]:I-231).
- Received April 18, 2000.
- Revision received June 5, 2000.
- Accepted June 8, 2000.
- Copyright © 2000 by American Heart Association
Ebrahim S, Papacosta O, Whincup P, et al. Carotid plaque, intima media thickness, cardiovascular risk factors, and prevalent cardiovascular disease in men and women. Stroke. 1999;30:841–850.
Meissner I, Whisnant JP, Sheps SG, et al. Detection and control of high blood pressure in the community: do we need a wake-up call? Hypertension. 1999;34:466–471.
Blackshear JL, Pearce LA, Hart RG, et al, for the Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography. Aortic plaque in atrial fibrillation: prevalence, predictors, and thromboembolic implications. Stroke. 1999;30:834–840.
Davila-Roman VG, Barzilai B, Wareing TH, et al. Atherosclerosis of the ascending aorta: prevalence and role as an independent predictor of cerebrovascular events in cardiac patients. Stroke. 1994;25:2010–2016.
Alexander RW. Hypertension and the pathogenesis of atherosclerosis. Hypertension. 1995;25:155–161.
Franklin SS, Khan SA, Wong ND, et al. Is pulse pressure useful in predicting risk for coronary heart disease?: the Framingham Heart Study. Circulation.. 1999;100:354–360.
Zanchetti A. Intermediate endpoints for atherosclerosis in hypertension. Blood Press Suppl. 1997;6:97–102.