- Email: [email protected]
Prevalence and risk factors for abdominal aortic aneurysms in older adults with and without isolated systolic hypertension
Prevalence and Risk Factors for Abdominal Aortic Aneurysms in Older Adults With and Without Isolated Systolic Hypertension Barbara L. Naydeck, MPH, Kim Sutton-Tyrrell, DrPH, Kathleen D. Schiller, RVT, Anne B. Newman, MD, MPH, and Lewis H. Kuller, MD, DrPH
RDMS, RTR,
An association between abdominal aortic aneurysm (AAA) and atherosclerotic disease has been recognized and may be due to shared risk factors. A consistent relation between blood pressure and AAA has not been found. AAA was compared between those with and without isolated systolic hypertension (ISH) and prevalence of aortic atherosclerosis was evaluated. Abdominal aortic ultrasound was performed in 266 people, 143 with ISH and 123 age-similar controls. AAA was defined as an infrarenal aortic diameter of >3.0 cm or an infrarenal-to-suprarenal diameter ratio of >1.2. The average age of participants was 73 years. Overall prevalence of AAA was 9.4%, 11.9% in those with ISH and 6.5% among normotensives (p 5 0.134). Multivariate analysis revealed male gender (p <0.001), higher lowdensity lipoprotein (p <0.001), higher pulse pressure (p 5 0.032), and current smoking (p 5 0.012) to be
independent predictors of AAA. When evaluating aortic atherosclerosis, those with AAA had significantly larger diameters of the iliac arteries along with greater intimamedia thickness of the iliac arteries. Those with and without aneurysms had a similar prevalence of plaque (89% to 96%), but measured plaques tended to be larger among those with than without AAA (p <0.001). Progression of AAA after 1 year was observed in 8 participants, with a mean diameter change of 3.42 mm. AAA was found to be independently associated with pulse pressure but not with systolic blood pressure. Patients with AAA also had greater wall thickness and greater diameter of the iliac arteries, which are probably associated with the underlying disease process. Q1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;83:759 –764)
eports of increasing prevalence of abdominal aortic aneurysm (AAA) and the severe conseR quences of aneurysm rupture suggest that population
Systolic Hypertension in the Elderly Program (SHEP) and 155 age-similar normotensive controls in those with and without AAA.
1
screening be implemented.2 To develop cost-effective screening mechanisms, the characteristics of those with AAA must be described. The abdominal aorta is easily accessible to noninvasive study, and both large and small AAAs can easily be detected using ultrasound techniques.3,4 Patients with aneurysms and atherosclerotic disease share similar risk factors including male gender,2 white race,5 older age,6 and a lipid profile that includes lower high-density lipoprotein and higher triglyceride and low-density lipoprotein concentrations,7 but a consistent relation between blood pressure and prevalence of AAA has not been found.8,9 To investigate the association between isolated systolic hypertension (ISH), other traditional cardiovascular risk factors and AAA, abdominal aortic ultrasound was performed in 267 older adults, 112 from the Pittsburgh center of the From the Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania. This study was supported by contract 5R01HL50439-04 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and was performed under the tenure of an Established Investigatorship from the American Heart Association. Manuscript received February 13, 1998; revised manuscript received and accepted October 6, 1998. Address for reprints: Barbara L. Naydeck, MPH, Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, 130 DeSoto Street, Pittsburgh, Pennsylvania 15261. E mail: [email protected]. ©1999 by Excerpta Medica, Inc. All rights reserved.
METHODS
Study population: As part of an ancillary study to the SHEP, ultrasound measures of subclinical atherosclerosis were obtained in 187 SHEP participants from the Pittsburgh center and 187 age-similar normotensive control subjects. SHEP was a multicenter randomized clinical trial designed to test the efficacy of treating ISH in adults aged $60 years. Participants were required to have a systolic blood pressure $160 mm Hg and a diastolic blood pressure ,90 mm Hg, and were randomly assigned to blood pressure treatment or placebo. Major exclusions from SHEP included recent myocardial infarction, stroke with residual paresis, uncontrolled congestive heart failure, peripheral arterial disease with evidence of tissue injury or loss, transient ischemic attacks with associated carotid bruits, and contraindication to study medications. Complete screening techniques and exclusion criteria have been reported previously.10 Data were collected on a total of 187 SHEP participants from January 1989 to November 1990. The SHEP trial found a 36% reduction in stroke rates and a significant reduction in coronary events for those assigned to blood pressure treatment.11 An additional 187 normotensive participants were recruited using the same screening mechanism from February 1989 to November 1991. All SHEP exclusion criteria were applied to these control 0002-9149/99/$–see front matter PII S0002-9149(98)00985-0
759
participants except that systolic blood pressure was required to be ,160 mm Hg and diastolic blood pressure ,90 mm Hg.12 All participants signed an informed consent, which was approved by the Biomedical Institutional Review Board of the University of Pittsburgh. Analysis group: Of the 374 subjects originally enrolled in the ancillary study of subclinical atherosclerosis, abdominal aortic ultrasound data were available for 267 participants. Missing data for 107 are due to 38 deaths, 25 refusals, and 44 cases in which technical problems prevented data collection (technical problems included poor imaging due to excessive intestinal gas or inability of the participant to travel to the clinic or lay on the examination table for the time required). One person had an aneurysm only in the iliac segment and was excluded from all analyses, which resulted in analysis of 266 participants (111 from SHEP and 155 from control subjects). Of the 155 initially normotensive controls, 32 (20.6%) had initiated blood pressure medication for ISH at the time of the abdominal scan. In the analyses presented here, normotensives who crossed over to blood pressure medication were reclassified as hypertensive. SHEP participants remained in the hypertensive category regardless of whether they were treated. Of the 111 SHEP participants, 64 (58%) were assigned to active blood pressure medication and 47 (42%) were assigned to placebo. After the end of SHEP, approximately 67% of participants were consistently on antihypertensive therapy. Aortic ultrasound evaluation: As part of the second ultrasound evaluation of these patients, an abdominal aortic ultrasound was performed. Examinations were performed 2 to 4 years after the end of the SHEP trial. Testing was done at the University of Pittsburgh, Department of Epidemiology, Ultrasound Research Laboratory (Pittsburgh, Pennsylvania), using a Toshiba SSA 270A duplex scanner with a 3.75-MHz convex probe was used. The protocol was based on that used in the Cardiovascular Health Study8 and was extended to include iliac artery imaging, measurement of plaque and wall thickness, measurement of residual lumen when thrombus was present and length of aneurysm. Gray-scale B-mode images of the abdominal aorta were obtained in both transverse and longitudinal projections. Aortic diameter was measured above the renal arteries and at the widest location below the renal arteries. Specifically, the suprarenal measure was taken 1 cm distal to the origin of the superior mesenteric artery, just above the level of the left renal artery. Either the superior mesenteric artery or the renal artery had to be visualized to accept the image for measurement. The infrarenal measure was taken at the site of the maximum aortic artery diameter (lumen plus wall) below the renal arteries. An AAA was defined as an infrarenal aortic diameter of $30 mm or a ratio of infrarenal to suprarenal diameter of $1.2, or as confirmed present by the physician. Aortic wall thickness was measured as the distance between the lumen-intima interface and the adventitiaperiadventitia interface using a transverse view. This 760 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
was done for the suprarenal aorta, the abdominal aorta, and the right and left iliac arteries. For each location, interfaces were identified across a 4-mm segment. These lines were then connected, generating a wall thickness measure for each pixel of the image. These measures were then averaged. Focal plaque was assessed in both suprarenal and infrarenal segments of the abdominal aorta and defined as an area protruding into the vessel lumen with at least 50% greater thickness relative to adjacent areas. The thickness of the largest plaque was measured. A plaque was considered to be calicified if shadowing due to mineralization originating from the plaque was present. Other measures of atherosclerosis: Intima-media thickness of the carotid arteries was measured across 10-mm segments of the near and far walls of the common carotid artery and far walls of the bulb and internal carotid artery on both sides. All measures were averaged. The presence and degree of focal plaque was evaluated along with the presence or absence of an internal carotid stenosis (by Doppler).13 Degree of focal plaque was summarized as the plaque index that was based on the number and size of plaques in the bulb and in the common and internal carotid arteries. Internal carotid stenosis was defined as an internal carotid to common carotid artery peak blood flow velocity ratio of $1.4, corresponding to a luminal diameter reduction of approximately 50%.13 The ratio of internal carotid to common carotid artery blood flow velocities is a measure of stenosis that controls for intersubject variation.14 The presence of lower extremity arterial disease was identified by the ankle-to-arm systolic blood pressure ratio, commonly called the ankle-arm index.15 An ankle-arm index of #0.90 was considered indicative of lower extremity artery disease. Other clinical data: Data on cigarette smoking and alcohol use were obtained by self-report. Alcohol use was reported by frequency and amount and was summarized as the average number of alcoholic drinks consumed per week. A fasting blood sample was drawn for the evaluation of lipids and glucose. Blood pressure was analyzed from an average of 4 readings taken during 2 separate clinic visits. Blood pressures were measured by trained observers using a randomzero manometer according to the SHEP protocol.10 Pulse pressure was calculated for each participant and is defined as systolic minus diastolic blood pressure. Prevalent cardiovascular disease was defined as history of angina, myocardial infarction, congestive heart failure, stroke, coronary bypass or vascular surgery, or the presence of low ankle-arm index or internal carotid artery stenosis. Follow-up aortic scans of aneurysms: Follow-up aortic ultrasound scans were performed 1 year after the baseline scan on 23 of the 25 participants classified as having AAA at baseline. Two participants died in the year after the baseline scan (death certificates listed chronic obstructive pulmonary disease and myocardial infarction for 1 participant and complications from lung cancer for the other participant). Progression or regression was defined as a change of $2.0 mm in the MARCH 1, 1999
TABLE I Characteristics for 25 Participants With Abdominal Aortic Aneurysm
Age (yr) & Sex
Baseline SBP/ DBP
Hypertensive Status at Time of AAA Scan
Total Cholesterol (mmol/L)
HDL (mmol/L)
62 M 71 M 76 F 77 F 71 M 73 F 77 M 82 F 75 M 70 F 73 M 68 F 74 M 75 M 69 M 75 M 78 M 76 M 76 M 67 M 82 M 67 M 78 F 78 M 86 M
124/69 133/72 162/77 137/69 162/80 130/75 173/74 173/76 129/59 165/77 130/68 170/84 157/67 117/59 141/82 165/76 135/62 157/80 163/74 165/71 129/77 143/72 174/88 195/76 182/61
Normotensive Normotensive Treated Normotensive Treated Treated Treated Treated Normotensive Treated Normotensive Treated Treated Treated Treated Treated Normotensive Normotensive Treated Treated Treated Normotensive Treated Treated Treated
5.72 6.70 6.57 8.41 7.06 7.58 5.25 6.96 4.97 6.26 6.49 7.55 4.45 6.59 6.75 5.46 7.03 4.40 4.65 6.41 4.91 5.43 9.05 4.42 4.81
0.82 1.20 1.63 1.19 1.63 1.06 1.26 1.60 0.94 1.39 1.51 1.39 1.40 0.83 1.10 1.54 1.36 0.85 0.75 1.36 1.18 1.48 1.51 4.57 1.13
Body Mass Index
Widest Infrarenal Diameter (mm)
Ratio of Infrarenal-toSuprarenal Diameter
21.5 23.7 23.3 20.0 32.1 24.4 23.9 24.8 25.8 24.9 26.2 26.5 30.9 27.7 24.8 26.7 25.4 27.2 29.8 23.1 24.5 32.3 23.8 22.7 28.4
20.8 22.2 23.9 24.6 25.0 25.1 25.6 25.8 26.4 26.5 27.4 27.4 27.7 29.0 30.6 31.9 32.3 34.6 34.7 34.9 36.9 37.6 46.7 49.1 50.8
1.3 1.2 1.2 1.3 1.4 1.3 1.2 1.4 1.3 1.2 1.3 1.5 1.4 1.2 1.5 1.3 1.5 1.3 1.5 1.3 1.5 1.5 2.1 2.1 2.0
Body Mass Index 5 weight (in kilograms) divided by height (in square meters); DBP 5 diastolic blood pressure; HDL 5 high-density lipoprotein; SBP 5 systolic blood pressure.
abdominal aortic lumen diameter at the widest point (which allowed for measurement error based on reproducibility studies). Statistical methods: The Pearson chi-square statistic was used to measure associations of AAA with demographic variables, smoking and alcohol use, prevalent clinical and subclinical cardiovascular disease, and laboratory values. Continuous variables were analyzed using the Student t test and were then categorized using clinically relevant categories where they have been established or data-based tertiles or quartiles otherwise. Age- and gender-adjusted p values are reported in addition to unadjusted values. Using only variables significantly associated with AAA after age and gender adjustments, logistic regression models were constructed to evaluate independent associations with AAA, and statistics for goodness-of-fit were used to assure the integrity of logistic models. Similar analyses steps were performed to detect statistically significant differences in baseline characteristics between those with and without progression.
RESULTS The analysis population ranged in age at the time of the scan from 60 to 88 years (mean 73). Sixty-four percent were women and all but 1 were white. The prevalence of AAA was 9.4% (25 of 266). All aneurysms had an infrarenal-to-suprarenal ratio of $1.2, but only 11 were also $30 mm (Table I). The infrarenal-to-suprarenal ratio ranged from 1.2 to 2.1 (mean 1.4). The aortic diameter at the widest point for those
with AAA ranged from 20.8 to 50.8 mm (mean 31.1). Nine aneurysms (36%) were found in segment 3 only, 8 (32%) were found in segment 4 only, and 8 (32%) involved both segments 3 and 4. Aneurysm length ranged from 22.7 to 71.7 mm (mean of 38.3). Thrombus was present in 6 aneurysms. AAA was found in 8 normotensive (6.5%) and 17 (11.9%) hypertensive participants (p 5 0.134). Those with hypertension were significantly older than normotensives (p ,0.001) and were more likely to have prevalent cardiovascular disease (p 5 0.001), a worse lipid profile, and higher mean weight (p 5 0.018) than normotensives. Table II compares those with and without AAA. Those with AAA were slightly older than those without AAA (p 5 0.220). AAA was found in 5.2% of those aged 60 to 69 years, in 12.8% of those aged 70 to 79, and in 8.8% of those aged $80 (p 5 0.128). The lower prevalence in the oldest age group may be due to selective survival. Aneurysms were more prevalent in men than in women overall (p 5 0.001) and in each age category. However, statistical significance was not reached in those aged .80 years (1 of 11 men [9.1%] vs 1 of 25 women [4.0%], p 5 0.539). AAA was more prevalent among current smokers than former or never smokers, and a nonsignificant “U-shaped” relation was seen between aneurysm prevalence and level of alcohol use. Diabetics formed only 5% of the study population, and all aneurysms were found among nondiabetics. Systolic blood presMISCELLANEOUS/ABDOMINAL AORTIC ANEURYSM
761
prevalent cardiovascular disease, or a low ankle blood pressure. However, Age-Gender this reached statistical significance Characteristic No AAA AAA Adjusted only for the mean ankle-arm index (n 5 266) (n 5 241) (n 5 25) % p Value p Value (age-gender adjusted p 5 0.030). Baseline characteristics Evidence of atherosclerosis in the Age at scan (mean) 72.8 74.5 0.220 0.077 aorta itself is also presented in Table Gender II. At the site of the aneurysm, the Women 163 7 4.1 abdominal aortic wall was signifiMen 78 18 18.8 0.001 0.0003 Smoking history cantly thicker among those with Never 137 9 6.2 AAA (p 5 0.0001) because this meaFormer 87 11 11.2 surement included thrombus, if Current 17 5 22.7 0.034 0.061 present. However, wall thickness Alcohol use was also increased in the nonaneuNever 145 16 9.9 Occasional/moderate 69 5 6.8 rysmal sections. After multivariate Frequent 27 4 12.9 0.574 0.321 adjustments (age, gender, low-denPrevalent CVD sity lipoprotein, pulse pressure, and Not present 149 14 8.6 smoking), the diameter of the iliac Present 92 11 10.7 0.569 0.826 Baseline clinical measures (mean) arteries was significantly larger in Systolic BP (mm Hg) 143.9 152.2 0.098 0.146 those with than without AAA. Diastolic BP (mm Hg) 73.4 72.7 0.743 0.731 Plaque was present in similar proporPulse pressure (mm Hg) 70.6 79.4 0.033 0.047 tions in both those with and without Cholesterol (mmol/L) 5.68 6.15 0.034 0.0001 aneurysms: 96% vs 89%, respecHDL (mmol/L) 1.39 1.27 0.113 0.488 LDL (mmol/L) 3.52 4.24 0.0008 0.0001 tively. The largest focal plaque averApolipoprotein(b) (mmol/L) 0.978 1.14 0.005 0.0001 aged 8.3 mm among those with AAA Ankle-arm index (mean) 0.91 0.74 0.143 0.030 and 3.9 mm among those without Internal carotid artery measures n n % AAA (p 5 0.0001, age/gender adInternal carotid stenosis Not present 214 19 8.2 justed); however, the largest focal Present 27 6 18.2 0.065 0.168 plaque could have been at the aneuPlaque index rysm site. Calcification was present Low (,3) 118 7 5.6 in 88% of those with and in 77% of Medium (3–5) 78 9 10.3 those without aneurysm (p 5 0.063, High 45 9 16.7 0.062 0.255 Aortic measures (mean) multivariate adjustment). Maximum far wall thickness (mm) 2.40 3.92 ,0.0001 0.0001 Multivariate analysis revealed Suprarenal far wall thickness (mm) 2.43 2.60 0.407 0.420 male gender (odds ratio [OR] 23.0, Right iliac far wall thickness (mm) 2.09 2.58 0.011 0.027 95% confidence interval [CI] 6.3 to Left iliac far wall thickness (mm) 1.88 2.56 0.0001 0.006 Percent focal plaque 89.3 96.2 0.557 0.277 84.4, p ,0.0001), higher low-density Percent calcified plaque 76.8 88.0 0.428 0.111 lipoprotein (OR 2.49/0.65 mmol/L, 95% CI 1.7 to 3.7, p ,0.0001), higher BP 5 blood pressure; CVD 5 cardiovascular disease; HDL 5 high-density lipoprotein; LDL 5 lowdensity lipoprotein. pulse pressure (OR 1.4/10 mm Hg, 95% CI 1.0 to 1.8, p 5 0.032), and current smoking (OR 6.4, 95% CI 1.5 sure was higher and diastolic pressure lower in those to 27.1, p ,0.012) to be independently associated with with AAA, but these differences were not statistically AAA. A 10-year increment in age was nonsignifidifferent. However, pulse pressure remained associ- cantly associated with an OR of 1.8. Variables not ated after age and sex adjustment (p 5 0.047). Those entering the final model were use of blood pressure with AAA had higher mean cholesterol levels than medication, height, weight, apolipoprotein(b), cholesthose without AAA, but high-density lipoprotein was terol, ankle-arm index, presence of internal carotid not significantly lower in those with AAA. After age stenosis, carotid plaque index, presence of cardiovasand sex adjustment, total cholesterol, low-density li- cular disease, and use of alcohol. Separate risk-factor analyses were performed using poprotein, and apolipoprotein(b) were significantly associated with AAA. Those with AAA were taller and as the definition of aneurysm either a transverse dihad larger waist-to-hip ratios, but these differences ameter $30 mm or an infrasuprarenal ratio of $1.5, were explained by age and gender. Blood glucose, with similar results as mentioned above. Baseline and 1-year follow-up scans for the 23 hematocrit, hemoglobin, and homocysteine were not participants with AAA were compared, and the mean found to be significantly associated with AAA. Risk factor analysis also included the assessment overall change in diameter was 10.91 mm (SD 2.39). of atherosclerosis in other vascular beds. Disease in Progression (a $2-mm increase in diameter) occurred the lower extremities and carotid arteries was evalu- in 8 cases (35%) and, for these, the mean diameter ated as well as the presence or absence of prevalent change was 13.42 mm (range 2.22 to 7.01, SD 1.63). cardiovascular disease (Table II). AAA was higher Three participants had regression, with a mean diamamong those with carotid stenosis, carotid plaque, or eter change of 22.82 mm (range 23.96 to 22.06, SD TABLE II Relation Between AAA and Selected Characteristics
762 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
MARCH 1, 1999
1.00). Far wall thickness measures were not significantly different in those with and without progression (3.88 vs 3.85 mm). Of those with progression, 4 participants had ISH and a mean diameter change of 4.32 (SD 1.98), and 4 participants were normotensive with a mean change in diameter of 2.53 mm (SD 0.38) (p 5 0.167). The maximum far wall thickness measures for these 8 participants by hypertensive status were not significantly different at baseline or at follow-up. Four women (4 of 7, 57%) and 4 men (4 of 16, 25%) had progression (p 5 0.136). Although not statistically significant, those with progression had higher total cholesterol (6.89 vs 5.96 mmol/L, p 5 0.136) and higher low-density lipoprotein (4.9 vs 4.11 mmol/L, p 5 0.11 than those without progression. Because of the small numbers, multivariate analysis was not possible.
DISCUSSION The prevalence of AAA varies widely due to differing definitions and dissimilar study populations.16 Our definition used a combination of absolute diameter and a ratio to identify early aneurysms. This study reports a prevalence of 9.4%, which is similar to the 9.5% reported in the Cardiovascular Health Study.8 According to the definition of aortic diameter $30 mm, the prevalence of AAA is 5.6% and is slightly lower than that reported by Krohn et al (5.8% with small AAA in men aged 60)17 and Morris et al (6.3% with aortic diameter $25 mm in men aged 50 to 64 years),18 and slightly higher than that reported by the Cardiovascular Health Study (4.6%)18 and the Veterans Administration study (4.7%).9 Hypertension has been reported to be both independently associated with AAA9 and not associated with AAA.19 Although our results did not show an association between AAA and ISH, we did see an independent and significant association with pulse pressure. The underlying cause of ISH is a stiffening of the central arteries, and increasing pulse pressure is a measure of ISH. A number of studies have implicated degradation of elastin and increasing aortic wall stiffness as the initiating event in AAA formation.20 It is likely that a degradation of elastin in the aorta and other central vessels is related to the cause of both AAA and ISH. Our study found that those with AAA had increased wall thickness in the iliac arteries. This may reflect a biologic mechanism in response to increased pulse pressure. Etiologic factors for AAA include genetic and biochemical processes as well as atherosclerotic factors such as gender and smoking. Populations with peripheral or cardiovascular disease have been found to have a higher prevalence of AAA.21 This study did not find consistent significant associations between AAA and evidence of peripheral atherosclerosis, although a trend toward higher AAA rates among those with peripheral disease was seen. It is possible that the lack of an association here is due to the exclusion of those with advanced atherosclerotic disease from the SHEP trial or due to higher mortality rates among those with both AAA and peripheral atherosclerosis.
Although many studies support the hypothesis that AAA is a disease process separate from atherosclerosis,22 aneurysms and obstructive disease frequently coexist. Shared risk factors may be an important link. If the underlying pathophysiology of aneurysm development is a degeneration of the media, then cardiovascular risk factors that contribute to this process may add to an underlying predisposition to AAA. Whereas a causal link between atherosclerosis and AAA cannot be assumed, the co-presence of these disease processes may allow for the identification of a high-risk population for screening purposes. Other studies have found a high prevalence of AAA in those referred for peripheral vascular disease assessment.21,23 It has been suggested that if the patient is already receiving ultrasound assessment for peripheral or carotid artery disease, the cost of adding an abdominal aortic scan is modest.23 The progression found in this study (3.25 mm mean diameter) compares with a median expansion rate (mm/ year) of 3.1 mm reported by Bengtsson et al24 and 2.6 mm/year reported by Stonebridge et al.25 Grimshaw et al26 suggests a suitable rescan interval for aortas with diameters ,40 mm to be 2 years. Our study confirms fairly slow growth rates of these small aneurysms. It is clear that if mortality rates associated with AAA are to be reduced, aneurysms must be identified early so that medical or surgical intervention can be initiated before rupture.27 Early detection is particularly valuable with the advent of new techniques for minimally invasive surgery.28 Screening for AAA in an asymptomatic population is controversial,29 but there is some indication that a single ultrasonographic screen for AAA in men aged 60 to 80 is cost-effective when compared with screening by palpation.30 Another recent study suggests that screening in men be performed as early as age 50 to allow for earlier intervention.18 Frame et al30 also mention that the cost-effectiveness ratio may be improved by targeting those with coronary or peripheral vascular disease. Targeting screening to those with cardiovascular disease is supported by the results of other studies.17 Wolf et al23 described costs involved when AAA screening is performed on those already receiving vascular laboratory assessment of peripheral vascular disease. Our study did not show the use of peripheral atherosclerosis in a screening algorithm to be justified, but does suggest that the screening strategy could target men aged .60 years and those with an atherosclerotic risk profile including higher low-density lipoprotein and elevated pulse pressure. In summary, pulse pressure, male sex, smoking, and elevated low-density lipoprotein appear to be independently associated with AAA. Those with AAA also had greater wall thickness and greater diameter of the iliac arteries, which are probably associated with the underlying disease process. Acknowledgment: We thank the participants, physicians, ultrasound technicians, and data management personnel who made this study possible, especially Fern Schwartz, BS, for her assistance in data manageMISCELLANEOUS/ABDOMINAL AORTIC ANEURYSM
763
ment. This study was performed during the tenure of an Established Investigatorship from the American Heart Association. 1. Stevens K. The incidence of abdominal aortic aneurysms [Review]. Br J Clin
Pract 1993;47:208 –210. 2. Van der Vliet JA, Boll APM. Abdominal aortic aneurysm [seminar]. Lancet
1997;349:863– 866. 3. Leopold G, Goldberger L, Bernstein E. Ultrasonic detection and evaluation of abdominal aortic aneurysms. Surgery 1972;72:939 –945. 4. Lederle F, Walker J, Reinke D. Selective screening for abdominal aortic aneurysms with physical examination and ultrasound. Arch Intern Med 1988; 148:1753–1756. 5. Johnson G, Avery A, McDougal E, Burnham S, Keagy B. Aneurysms of the abdominal aorta. Arch Surg 1985;120:1138 –1140. 6. Collin J. The epidemiology of abdominal aortic aneurysm. Br J Hosp Med 1988;40:65– 68. 7. Norrgard O, Angquist K, Johnson O. Familial aortic aneurysms: serum concentrations of triglyceride, cholesterol, HDL-cholesterol and (VSLD1LDL)cholesterol. Br J Surg 1985;72:113–116. 8. Alcorn HG, Wolfson SK, Sutton-Tyrrell K, Kuller LH, O’Leary DH. Risk factors for abdominal aortic aneurysms in older adults enrolled in the Cardiovascular Health Study. Circulation 1995;92:8:I-152. 9. Lederle FA, Johnson GR, Wilson SE, Chute EP, Littooy FN, Bandyk D, Krupski WC, Barone GW, Acher CW, Ballard DJ, for the Aneurysm Detection and Management (ADAM) Veterans Affairs Cooperative Study Group. Prevalence and associations of abdominal aortic aneurysm detected through screening. Ann Intern Med 1997;126:441– 449. 10. Petrovitch H, Byington R, Bailey G, Borhani P, Carmody S, Goodwin L, Harrington H, Johnson HA, Johnson P, Jones M, Levin J, Sugars C, Probstfield JL. Systolic Hypertension in the Elderly Program (SHEP): baseline characteristics of the randomized sample, II: screening and recruitment. Hypertension 1991;17(suppl II):II-16 –II-23. 11. SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA 1991;265: 3255–3264. 12. Sutton-Tyrrell K, Alcorn HG, Herzog H, Kelsey SF, Kuller LH. Morbidity, mortality and antihypertensive treatment effects by extent of atherosclerosis in older adults with isolated systolic hypertension. Stroke 1995;26:1319 –1324. 13. Sutton KC, Wolfson SK Jr, Kuller LH. Carotid and lower extremity disease in elderly adults with isolated systolic hypertension. Stroke 1987;18:817– 822.
764 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
14. Sutton-Tyrrell K, Wolfson SK Jr, Thompson T, Kelsey SF. Measurement variability in duplex scan assessment of carotid atherosclerosis. Stroke 1992;23: 215–220. 15. Newman AB, Sutton-Tyrrell K, Rutan GH, Locher J, Kuller LH. Lower extremity arterial disease in elderly subjects with systolic hypertension. J Clin Epidemiol 1991;44:15–20. 16. Moher D, Cole CW, Hill GB. Epidemiology of abdominal aortic aneurysm: the effect of differing definitions. Eur J Vasc Surg 1992;6:647– 650. 17. Krohn CD, Kullmann G, Kvernebo K, Rosen L, Kroese A. Ultrasonographic screening for abdominal aortic aneurysm. Eur J Surg 1992;158:527–530. 18. Morris GE, Hubbard CS, Quick CR. An abdominal aortic aneurysm screening program for all males over the age of 50 years. Eur J Vasc Surg 1994;8:156 –160. 19. Lindholm L, Ejlertsson G, Forsberg L, Norgren L. Low prevalence of abdominal aortic aneurysm in hypertensive patients. Acta Med Scan 1985;218: 305–310. 20. MacSweeney ST, Young G, Greenhalgh RM, Powell JT. Mechanical properties of the aneurysmal aorta. Br J Surg 1992;79:1281–1284. 21. Karanjia PN, Madden KP, Lobner S. Coexistence of abdominal aortic aneurysm in patients with carotid stenosis. Stroke 1994;25:627– 630. 22. Martin P. On abdominal aortic aneurysms. J Cardiovasc Surg 1978;19:597– 598. 23. Wolf YG, Otis SM, Schwend RB, Bernstein EF. Screening for abdominal aortic aneurysms during lower extremity arterial evaluation in the vascular laboratory. J Vasc Surg 1995;22:417– 421. 24. Bengtsson H, Bergqvist D, Ekberg O, Ranstam J. Expansion pattern and risk of rupture of abdominal aortic aneurysms that were not operated on. Eur J Surg 1993;159:461– 467. 25. Stonebridge PA, Draper T, Kelman J, Howlett J, Allan PL, Prescott R, Ruckley CV. Growth rate of infrarenal aortic aneurysms. Eur J Vasc Endovasc Surg 1996;11:70 –73. 26. Grimshaw GM, Thompson JM, Hamer JD. A statistical analysis of the growth of small abdominal aortic aneurysms. Eur J Vasc Surg 1994;8:741–746. 27. Scott RA, Wilson NM, Ashton HA, Kay DN. Influence of screening on the incidence of ruptured abdominal aortic aneurysm: 5-year results of a randomized controlled study. Br J Surg 1995;82:1066 –1070. 28. Harris PL. Endovascular grafting for abdominal aortic aneurysms. Ann R Coll Surg Engl 1996;78(suppl 1):23–24. 29. Cheatle TR. The case against a national screening programme for aortic aneurysms [comment]. Ann R Coll Surg Engl 1997;79:310 [Source: Ann R Coll Surg Engl 1997;79:90 –95]. 30. Frame PS, Gryback DG, Patterson C. Screening for abdominal aortic aneurysm in men ages 60 to 80 years. Ann Intern Med 1993;119:411– 416.
MARCH 1, 1999
RDMS, RTR,
An association between abdominal aortic aneurysm (AAA) and atherosclerotic disease has been recognized and may be due to shared risk factors. A consistent relation between blood pressure and AAA has not been found. AAA was compared between those with and without isolated systolic hypertension (ISH) and prevalence of aortic atherosclerosis was evaluated. Abdominal aortic ultrasound was performed in 266 people, 143 with ISH and 123 age-similar controls. AAA was defined as an infrarenal aortic diameter of >3.0 cm or an infrarenal-to-suprarenal diameter ratio of >1.2. The average age of participants was 73 years. Overall prevalence of AAA was 9.4%, 11.9% in those with ISH and 6.5% among normotensives (p 5 0.134). Multivariate analysis revealed male gender (p <0.001), higher lowdensity lipoprotein (p <0.001), higher pulse pressure (p 5 0.032), and current smoking (p 5 0.012) to be
independent predictors of AAA. When evaluating aortic atherosclerosis, those with AAA had significantly larger diameters of the iliac arteries along with greater intimamedia thickness of the iliac arteries. Those with and without aneurysms had a similar prevalence of plaque (89% to 96%), but measured plaques tended to be larger among those with than without AAA (p <0.001). Progression of AAA after 1 year was observed in 8 participants, with a mean diameter change of 3.42 mm. AAA was found to be independently associated with pulse pressure but not with systolic blood pressure. Patients with AAA also had greater wall thickness and greater diameter of the iliac arteries, which are probably associated with the underlying disease process. Q1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;83:759 –764)
eports of increasing prevalence of abdominal aortic aneurysm (AAA) and the severe conseR quences of aneurysm rupture suggest that population
Systolic Hypertension in the Elderly Program (SHEP) and 155 age-similar normotensive controls in those with and without AAA.
1
screening be implemented.2 To develop cost-effective screening mechanisms, the characteristics of those with AAA must be described. The abdominal aorta is easily accessible to noninvasive study, and both large and small AAAs can easily be detected using ultrasound techniques.3,4 Patients with aneurysms and atherosclerotic disease share similar risk factors including male gender,2 white race,5 older age,6 and a lipid profile that includes lower high-density lipoprotein and higher triglyceride and low-density lipoprotein concentrations,7 but a consistent relation between blood pressure and prevalence of AAA has not been found.8,9 To investigate the association between isolated systolic hypertension (ISH), other traditional cardiovascular risk factors and AAA, abdominal aortic ultrasound was performed in 267 older adults, 112 from the Pittsburgh center of the From the Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania. This study was supported by contract 5R01HL50439-04 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and was performed under the tenure of an Established Investigatorship from the American Heart Association. Manuscript received February 13, 1998; revised manuscript received and accepted October 6, 1998. Address for reprints: Barbara L. Naydeck, MPH, Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, 130 DeSoto Street, Pittsburgh, Pennsylvania 15261. E mail: [email protected]. ©1999 by Excerpta Medica, Inc. All rights reserved.
METHODS
Study population: As part of an ancillary study to the SHEP, ultrasound measures of subclinical atherosclerosis were obtained in 187 SHEP participants from the Pittsburgh center and 187 age-similar normotensive control subjects. SHEP was a multicenter randomized clinical trial designed to test the efficacy of treating ISH in adults aged $60 years. Participants were required to have a systolic blood pressure $160 mm Hg and a diastolic blood pressure ,90 mm Hg, and were randomly assigned to blood pressure treatment or placebo. Major exclusions from SHEP included recent myocardial infarction, stroke with residual paresis, uncontrolled congestive heart failure, peripheral arterial disease with evidence of tissue injury or loss, transient ischemic attacks with associated carotid bruits, and contraindication to study medications. Complete screening techniques and exclusion criteria have been reported previously.10 Data were collected on a total of 187 SHEP participants from January 1989 to November 1990. The SHEP trial found a 36% reduction in stroke rates and a significant reduction in coronary events for those assigned to blood pressure treatment.11 An additional 187 normotensive participants were recruited using the same screening mechanism from February 1989 to November 1991. All SHEP exclusion criteria were applied to these control 0002-9149/99/$–see front matter PII S0002-9149(98)00985-0
759
participants except that systolic blood pressure was required to be ,160 mm Hg and diastolic blood pressure ,90 mm Hg.12 All participants signed an informed consent, which was approved by the Biomedical Institutional Review Board of the University of Pittsburgh. Analysis group: Of the 374 subjects originally enrolled in the ancillary study of subclinical atherosclerosis, abdominal aortic ultrasound data were available for 267 participants. Missing data for 107 are due to 38 deaths, 25 refusals, and 44 cases in which technical problems prevented data collection (technical problems included poor imaging due to excessive intestinal gas or inability of the participant to travel to the clinic or lay on the examination table for the time required). One person had an aneurysm only in the iliac segment and was excluded from all analyses, which resulted in analysis of 266 participants (111 from SHEP and 155 from control subjects). Of the 155 initially normotensive controls, 32 (20.6%) had initiated blood pressure medication for ISH at the time of the abdominal scan. In the analyses presented here, normotensives who crossed over to blood pressure medication were reclassified as hypertensive. SHEP participants remained in the hypertensive category regardless of whether they were treated. Of the 111 SHEP participants, 64 (58%) were assigned to active blood pressure medication and 47 (42%) were assigned to placebo. After the end of SHEP, approximately 67% of participants were consistently on antihypertensive therapy. Aortic ultrasound evaluation: As part of the second ultrasound evaluation of these patients, an abdominal aortic ultrasound was performed. Examinations were performed 2 to 4 years after the end of the SHEP trial. Testing was done at the University of Pittsburgh, Department of Epidemiology, Ultrasound Research Laboratory (Pittsburgh, Pennsylvania), using a Toshiba SSA 270A duplex scanner with a 3.75-MHz convex probe was used. The protocol was based on that used in the Cardiovascular Health Study8 and was extended to include iliac artery imaging, measurement of plaque and wall thickness, measurement of residual lumen when thrombus was present and length of aneurysm. Gray-scale B-mode images of the abdominal aorta were obtained in both transverse and longitudinal projections. Aortic diameter was measured above the renal arteries and at the widest location below the renal arteries. Specifically, the suprarenal measure was taken 1 cm distal to the origin of the superior mesenteric artery, just above the level of the left renal artery. Either the superior mesenteric artery or the renal artery had to be visualized to accept the image for measurement. The infrarenal measure was taken at the site of the maximum aortic artery diameter (lumen plus wall) below the renal arteries. An AAA was defined as an infrarenal aortic diameter of $30 mm or a ratio of infrarenal to suprarenal diameter of $1.2, or as confirmed present by the physician. Aortic wall thickness was measured as the distance between the lumen-intima interface and the adventitiaperiadventitia interface using a transverse view. This 760 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
was done for the suprarenal aorta, the abdominal aorta, and the right and left iliac arteries. For each location, interfaces were identified across a 4-mm segment. These lines were then connected, generating a wall thickness measure for each pixel of the image. These measures were then averaged. Focal plaque was assessed in both suprarenal and infrarenal segments of the abdominal aorta and defined as an area protruding into the vessel lumen with at least 50% greater thickness relative to adjacent areas. The thickness of the largest plaque was measured. A plaque was considered to be calicified if shadowing due to mineralization originating from the plaque was present. Other measures of atherosclerosis: Intima-media thickness of the carotid arteries was measured across 10-mm segments of the near and far walls of the common carotid artery and far walls of the bulb and internal carotid artery on both sides. All measures were averaged. The presence and degree of focal plaque was evaluated along with the presence or absence of an internal carotid stenosis (by Doppler).13 Degree of focal plaque was summarized as the plaque index that was based on the number and size of plaques in the bulb and in the common and internal carotid arteries. Internal carotid stenosis was defined as an internal carotid to common carotid artery peak blood flow velocity ratio of $1.4, corresponding to a luminal diameter reduction of approximately 50%.13 The ratio of internal carotid to common carotid artery blood flow velocities is a measure of stenosis that controls for intersubject variation.14 The presence of lower extremity arterial disease was identified by the ankle-to-arm systolic blood pressure ratio, commonly called the ankle-arm index.15 An ankle-arm index of #0.90 was considered indicative of lower extremity artery disease. Other clinical data: Data on cigarette smoking and alcohol use were obtained by self-report. Alcohol use was reported by frequency and amount and was summarized as the average number of alcoholic drinks consumed per week. A fasting blood sample was drawn for the evaluation of lipids and glucose. Blood pressure was analyzed from an average of 4 readings taken during 2 separate clinic visits. Blood pressures were measured by trained observers using a randomzero manometer according to the SHEP protocol.10 Pulse pressure was calculated for each participant and is defined as systolic minus diastolic blood pressure. Prevalent cardiovascular disease was defined as history of angina, myocardial infarction, congestive heart failure, stroke, coronary bypass or vascular surgery, or the presence of low ankle-arm index or internal carotid artery stenosis. Follow-up aortic scans of aneurysms: Follow-up aortic ultrasound scans were performed 1 year after the baseline scan on 23 of the 25 participants classified as having AAA at baseline. Two participants died in the year after the baseline scan (death certificates listed chronic obstructive pulmonary disease and myocardial infarction for 1 participant and complications from lung cancer for the other participant). Progression or regression was defined as a change of $2.0 mm in the MARCH 1, 1999
TABLE I Characteristics for 25 Participants With Abdominal Aortic Aneurysm
Age (yr) & Sex
Baseline SBP/ DBP
Hypertensive Status at Time of AAA Scan
Total Cholesterol (mmol/L)
HDL (mmol/L)
62 M 71 M 76 F 77 F 71 M 73 F 77 M 82 F 75 M 70 F 73 M 68 F 74 M 75 M 69 M 75 M 78 M 76 M 76 M 67 M 82 M 67 M 78 F 78 M 86 M
124/69 133/72 162/77 137/69 162/80 130/75 173/74 173/76 129/59 165/77 130/68 170/84 157/67 117/59 141/82 165/76 135/62 157/80 163/74 165/71 129/77 143/72 174/88 195/76 182/61
Normotensive Normotensive Treated Normotensive Treated Treated Treated Treated Normotensive Treated Normotensive Treated Treated Treated Treated Treated Normotensive Normotensive Treated Treated Treated Normotensive Treated Treated Treated
5.72 6.70 6.57 8.41 7.06 7.58 5.25 6.96 4.97 6.26 6.49 7.55 4.45 6.59 6.75 5.46 7.03 4.40 4.65 6.41 4.91 5.43 9.05 4.42 4.81
0.82 1.20 1.63 1.19 1.63 1.06 1.26 1.60 0.94 1.39 1.51 1.39 1.40 0.83 1.10 1.54 1.36 0.85 0.75 1.36 1.18 1.48 1.51 4.57 1.13
Body Mass Index
Widest Infrarenal Diameter (mm)
Ratio of Infrarenal-toSuprarenal Diameter
21.5 23.7 23.3 20.0 32.1 24.4 23.9 24.8 25.8 24.9 26.2 26.5 30.9 27.7 24.8 26.7 25.4 27.2 29.8 23.1 24.5 32.3 23.8 22.7 28.4
20.8 22.2 23.9 24.6 25.0 25.1 25.6 25.8 26.4 26.5 27.4 27.4 27.7 29.0 30.6 31.9 32.3 34.6 34.7 34.9 36.9 37.6 46.7 49.1 50.8
1.3 1.2 1.2 1.3 1.4 1.3 1.2 1.4 1.3 1.2 1.3 1.5 1.4 1.2 1.5 1.3 1.5 1.3 1.5 1.3 1.5 1.5 2.1 2.1 2.0
Body Mass Index 5 weight (in kilograms) divided by height (in square meters); DBP 5 diastolic blood pressure; HDL 5 high-density lipoprotein; SBP 5 systolic blood pressure.
abdominal aortic lumen diameter at the widest point (which allowed for measurement error based on reproducibility studies). Statistical methods: The Pearson chi-square statistic was used to measure associations of AAA with demographic variables, smoking and alcohol use, prevalent clinical and subclinical cardiovascular disease, and laboratory values. Continuous variables were analyzed using the Student t test and were then categorized using clinically relevant categories where they have been established or data-based tertiles or quartiles otherwise. Age- and gender-adjusted p values are reported in addition to unadjusted values. Using only variables significantly associated with AAA after age and gender adjustments, logistic regression models were constructed to evaluate independent associations with AAA, and statistics for goodness-of-fit were used to assure the integrity of logistic models. Similar analyses steps were performed to detect statistically significant differences in baseline characteristics between those with and without progression.
RESULTS The analysis population ranged in age at the time of the scan from 60 to 88 years (mean 73). Sixty-four percent were women and all but 1 were white. The prevalence of AAA was 9.4% (25 of 266). All aneurysms had an infrarenal-to-suprarenal ratio of $1.2, but only 11 were also $30 mm (Table I). The infrarenal-to-suprarenal ratio ranged from 1.2 to 2.1 (mean 1.4). The aortic diameter at the widest point for those
with AAA ranged from 20.8 to 50.8 mm (mean 31.1). Nine aneurysms (36%) were found in segment 3 only, 8 (32%) were found in segment 4 only, and 8 (32%) involved both segments 3 and 4. Aneurysm length ranged from 22.7 to 71.7 mm (mean of 38.3). Thrombus was present in 6 aneurysms. AAA was found in 8 normotensive (6.5%) and 17 (11.9%) hypertensive participants (p 5 0.134). Those with hypertension were significantly older than normotensives (p ,0.001) and were more likely to have prevalent cardiovascular disease (p 5 0.001), a worse lipid profile, and higher mean weight (p 5 0.018) than normotensives. Table II compares those with and without AAA. Those with AAA were slightly older than those without AAA (p 5 0.220). AAA was found in 5.2% of those aged 60 to 69 years, in 12.8% of those aged 70 to 79, and in 8.8% of those aged $80 (p 5 0.128). The lower prevalence in the oldest age group may be due to selective survival. Aneurysms were more prevalent in men than in women overall (p 5 0.001) and in each age category. However, statistical significance was not reached in those aged .80 years (1 of 11 men [9.1%] vs 1 of 25 women [4.0%], p 5 0.539). AAA was more prevalent among current smokers than former or never smokers, and a nonsignificant “U-shaped” relation was seen between aneurysm prevalence and level of alcohol use. Diabetics formed only 5% of the study population, and all aneurysms were found among nondiabetics. Systolic blood presMISCELLANEOUS/ABDOMINAL AORTIC ANEURYSM
761
prevalent cardiovascular disease, or a low ankle blood pressure. However, Age-Gender this reached statistical significance Characteristic No AAA AAA Adjusted only for the mean ankle-arm index (n 5 266) (n 5 241) (n 5 25) % p Value p Value (age-gender adjusted p 5 0.030). Baseline characteristics Evidence of atherosclerosis in the Age at scan (mean) 72.8 74.5 0.220 0.077 aorta itself is also presented in Table Gender II. At the site of the aneurysm, the Women 163 7 4.1 abdominal aortic wall was signifiMen 78 18 18.8 0.001 0.0003 Smoking history cantly thicker among those with Never 137 9 6.2 AAA (p 5 0.0001) because this meaFormer 87 11 11.2 surement included thrombus, if Current 17 5 22.7 0.034 0.061 present. However, wall thickness Alcohol use was also increased in the nonaneuNever 145 16 9.9 Occasional/moderate 69 5 6.8 rysmal sections. After multivariate Frequent 27 4 12.9 0.574 0.321 adjustments (age, gender, low-denPrevalent CVD sity lipoprotein, pulse pressure, and Not present 149 14 8.6 smoking), the diameter of the iliac Present 92 11 10.7 0.569 0.826 Baseline clinical measures (mean) arteries was significantly larger in Systolic BP (mm Hg) 143.9 152.2 0.098 0.146 those with than without AAA. Diastolic BP (mm Hg) 73.4 72.7 0.743 0.731 Plaque was present in similar proporPulse pressure (mm Hg) 70.6 79.4 0.033 0.047 tions in both those with and without Cholesterol (mmol/L) 5.68 6.15 0.034 0.0001 aneurysms: 96% vs 89%, respecHDL (mmol/L) 1.39 1.27 0.113 0.488 LDL (mmol/L) 3.52 4.24 0.0008 0.0001 tively. The largest focal plaque averApolipoprotein(b) (mmol/L) 0.978 1.14 0.005 0.0001 aged 8.3 mm among those with AAA Ankle-arm index (mean) 0.91 0.74 0.143 0.030 and 3.9 mm among those without Internal carotid artery measures n n % AAA (p 5 0.0001, age/gender adInternal carotid stenosis Not present 214 19 8.2 justed); however, the largest focal Present 27 6 18.2 0.065 0.168 plaque could have been at the aneuPlaque index rysm site. Calcification was present Low (,3) 118 7 5.6 in 88% of those with and in 77% of Medium (3–5) 78 9 10.3 those without aneurysm (p 5 0.063, High 45 9 16.7 0.062 0.255 Aortic measures (mean) multivariate adjustment). Maximum far wall thickness (mm) 2.40 3.92 ,0.0001 0.0001 Multivariate analysis revealed Suprarenal far wall thickness (mm) 2.43 2.60 0.407 0.420 male gender (odds ratio [OR] 23.0, Right iliac far wall thickness (mm) 2.09 2.58 0.011 0.027 95% confidence interval [CI] 6.3 to Left iliac far wall thickness (mm) 1.88 2.56 0.0001 0.006 Percent focal plaque 89.3 96.2 0.557 0.277 84.4, p ,0.0001), higher low-density Percent calcified plaque 76.8 88.0 0.428 0.111 lipoprotein (OR 2.49/0.65 mmol/L, 95% CI 1.7 to 3.7, p ,0.0001), higher BP 5 blood pressure; CVD 5 cardiovascular disease; HDL 5 high-density lipoprotein; LDL 5 lowdensity lipoprotein. pulse pressure (OR 1.4/10 mm Hg, 95% CI 1.0 to 1.8, p 5 0.032), and current smoking (OR 6.4, 95% CI 1.5 sure was higher and diastolic pressure lower in those to 27.1, p ,0.012) to be independently associated with with AAA, but these differences were not statistically AAA. A 10-year increment in age was nonsignifidifferent. However, pulse pressure remained associ- cantly associated with an OR of 1.8. Variables not ated after age and sex adjustment (p 5 0.047). Those entering the final model were use of blood pressure with AAA had higher mean cholesterol levels than medication, height, weight, apolipoprotein(b), cholesthose without AAA, but high-density lipoprotein was terol, ankle-arm index, presence of internal carotid not significantly lower in those with AAA. After age stenosis, carotid plaque index, presence of cardiovasand sex adjustment, total cholesterol, low-density li- cular disease, and use of alcohol. Separate risk-factor analyses were performed using poprotein, and apolipoprotein(b) were significantly associated with AAA. Those with AAA were taller and as the definition of aneurysm either a transverse dihad larger waist-to-hip ratios, but these differences ameter $30 mm or an infrasuprarenal ratio of $1.5, were explained by age and gender. Blood glucose, with similar results as mentioned above. Baseline and 1-year follow-up scans for the 23 hematocrit, hemoglobin, and homocysteine were not participants with AAA were compared, and the mean found to be significantly associated with AAA. Risk factor analysis also included the assessment overall change in diameter was 10.91 mm (SD 2.39). of atherosclerosis in other vascular beds. Disease in Progression (a $2-mm increase in diameter) occurred the lower extremities and carotid arteries was evalu- in 8 cases (35%) and, for these, the mean diameter ated as well as the presence or absence of prevalent change was 13.42 mm (range 2.22 to 7.01, SD 1.63). cardiovascular disease (Table II). AAA was higher Three participants had regression, with a mean diamamong those with carotid stenosis, carotid plaque, or eter change of 22.82 mm (range 23.96 to 22.06, SD TABLE II Relation Between AAA and Selected Characteristics
762 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
MARCH 1, 1999
1.00). Far wall thickness measures were not significantly different in those with and without progression (3.88 vs 3.85 mm). Of those with progression, 4 participants had ISH and a mean diameter change of 4.32 (SD 1.98), and 4 participants were normotensive with a mean change in diameter of 2.53 mm (SD 0.38) (p 5 0.167). The maximum far wall thickness measures for these 8 participants by hypertensive status were not significantly different at baseline or at follow-up. Four women (4 of 7, 57%) and 4 men (4 of 16, 25%) had progression (p 5 0.136). Although not statistically significant, those with progression had higher total cholesterol (6.89 vs 5.96 mmol/L, p 5 0.136) and higher low-density lipoprotein (4.9 vs 4.11 mmol/L, p 5 0.11 than those without progression. Because of the small numbers, multivariate analysis was not possible.
DISCUSSION The prevalence of AAA varies widely due to differing definitions and dissimilar study populations.16 Our definition used a combination of absolute diameter and a ratio to identify early aneurysms. This study reports a prevalence of 9.4%, which is similar to the 9.5% reported in the Cardiovascular Health Study.8 According to the definition of aortic diameter $30 mm, the prevalence of AAA is 5.6% and is slightly lower than that reported by Krohn et al (5.8% with small AAA in men aged 60)17 and Morris et al (6.3% with aortic diameter $25 mm in men aged 50 to 64 years),18 and slightly higher than that reported by the Cardiovascular Health Study (4.6%)18 and the Veterans Administration study (4.7%).9 Hypertension has been reported to be both independently associated with AAA9 and not associated with AAA.19 Although our results did not show an association between AAA and ISH, we did see an independent and significant association with pulse pressure. The underlying cause of ISH is a stiffening of the central arteries, and increasing pulse pressure is a measure of ISH. A number of studies have implicated degradation of elastin and increasing aortic wall stiffness as the initiating event in AAA formation.20 It is likely that a degradation of elastin in the aorta and other central vessels is related to the cause of both AAA and ISH. Our study found that those with AAA had increased wall thickness in the iliac arteries. This may reflect a biologic mechanism in response to increased pulse pressure. Etiologic factors for AAA include genetic and biochemical processes as well as atherosclerotic factors such as gender and smoking. Populations with peripheral or cardiovascular disease have been found to have a higher prevalence of AAA.21 This study did not find consistent significant associations between AAA and evidence of peripheral atherosclerosis, although a trend toward higher AAA rates among those with peripheral disease was seen. It is possible that the lack of an association here is due to the exclusion of those with advanced atherosclerotic disease from the SHEP trial or due to higher mortality rates among those with both AAA and peripheral atherosclerosis.
Although many studies support the hypothesis that AAA is a disease process separate from atherosclerosis,22 aneurysms and obstructive disease frequently coexist. Shared risk factors may be an important link. If the underlying pathophysiology of aneurysm development is a degeneration of the media, then cardiovascular risk factors that contribute to this process may add to an underlying predisposition to AAA. Whereas a causal link between atherosclerosis and AAA cannot be assumed, the co-presence of these disease processes may allow for the identification of a high-risk population for screening purposes. Other studies have found a high prevalence of AAA in those referred for peripheral vascular disease assessment.21,23 It has been suggested that if the patient is already receiving ultrasound assessment for peripheral or carotid artery disease, the cost of adding an abdominal aortic scan is modest.23 The progression found in this study (3.25 mm mean diameter) compares with a median expansion rate (mm/ year) of 3.1 mm reported by Bengtsson et al24 and 2.6 mm/year reported by Stonebridge et al.25 Grimshaw et al26 suggests a suitable rescan interval for aortas with diameters ,40 mm to be 2 years. Our study confirms fairly slow growth rates of these small aneurysms. It is clear that if mortality rates associated with AAA are to be reduced, aneurysms must be identified early so that medical or surgical intervention can be initiated before rupture.27 Early detection is particularly valuable with the advent of new techniques for minimally invasive surgery.28 Screening for AAA in an asymptomatic population is controversial,29 but there is some indication that a single ultrasonographic screen for AAA in men aged 60 to 80 is cost-effective when compared with screening by palpation.30 Another recent study suggests that screening in men be performed as early as age 50 to allow for earlier intervention.18 Frame et al30 also mention that the cost-effectiveness ratio may be improved by targeting those with coronary or peripheral vascular disease. Targeting screening to those with cardiovascular disease is supported by the results of other studies.17 Wolf et al23 described costs involved when AAA screening is performed on those already receiving vascular laboratory assessment of peripheral vascular disease. Our study did not show the use of peripheral atherosclerosis in a screening algorithm to be justified, but does suggest that the screening strategy could target men aged .60 years and those with an atherosclerotic risk profile including higher low-density lipoprotein and elevated pulse pressure. In summary, pulse pressure, male sex, smoking, and elevated low-density lipoprotein appear to be independently associated with AAA. Those with AAA also had greater wall thickness and greater diameter of the iliac arteries, which are probably associated with the underlying disease process. Acknowledgment: We thank the participants, physicians, ultrasound technicians, and data management personnel who made this study possible, especially Fern Schwartz, BS, for her assistance in data manageMISCELLANEOUS/ABDOMINAL AORTIC ANEURYSM
763
ment. This study was performed during the tenure of an Established Investigatorship from the American Heart Association. 1. Stevens K. The incidence of abdominal aortic aneurysms [Review]. Br J Clin
Pract 1993;47:208 –210. 2. Van der Vliet JA, Boll APM. Abdominal aortic aneurysm [seminar]. Lancet
1997;349:863– 866. 3. Leopold G, Goldberger L, Bernstein E. Ultrasonic detection and evaluation of abdominal aortic aneurysms. Surgery 1972;72:939 –945. 4. Lederle F, Walker J, Reinke D. Selective screening for abdominal aortic aneurysms with physical examination and ultrasound. Arch Intern Med 1988; 148:1753–1756. 5. Johnson G, Avery A, McDougal E, Burnham S, Keagy B. Aneurysms of the abdominal aorta. Arch Surg 1985;120:1138 –1140. 6. Collin J. The epidemiology of abdominal aortic aneurysm. Br J Hosp Med 1988;40:65– 68. 7. Norrgard O, Angquist K, Johnson O. Familial aortic aneurysms: serum concentrations of triglyceride, cholesterol, HDL-cholesterol and (VSLD1LDL)cholesterol. Br J Surg 1985;72:113–116. 8. Alcorn HG, Wolfson SK, Sutton-Tyrrell K, Kuller LH, O’Leary DH. Risk factors for abdominal aortic aneurysms in older adults enrolled in the Cardiovascular Health Study. Circulation 1995;92:8:I-152. 9. Lederle FA, Johnson GR, Wilson SE, Chute EP, Littooy FN, Bandyk D, Krupski WC, Barone GW, Acher CW, Ballard DJ, for the Aneurysm Detection and Management (ADAM) Veterans Affairs Cooperative Study Group. Prevalence and associations of abdominal aortic aneurysm detected through screening. Ann Intern Med 1997;126:441– 449. 10. Petrovitch H, Byington R, Bailey G, Borhani P, Carmody S, Goodwin L, Harrington H, Johnson HA, Johnson P, Jones M, Levin J, Sugars C, Probstfield JL. Systolic Hypertension in the Elderly Program (SHEP): baseline characteristics of the randomized sample, II: screening and recruitment. Hypertension 1991;17(suppl II):II-16 –II-23. 11. SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA 1991;265: 3255–3264. 12. Sutton-Tyrrell K, Alcorn HG, Herzog H, Kelsey SF, Kuller LH. Morbidity, mortality and antihypertensive treatment effects by extent of atherosclerosis in older adults with isolated systolic hypertension. Stroke 1995;26:1319 –1324. 13. Sutton KC, Wolfson SK Jr, Kuller LH. Carotid and lower extremity disease in elderly adults with isolated systolic hypertension. Stroke 1987;18:817– 822.
764 THE AMERICAN JOURNAL OF CARDIOLOGYT
VOL. 83
14. Sutton-Tyrrell K, Wolfson SK Jr, Thompson T, Kelsey SF. Measurement variability in duplex scan assessment of carotid atherosclerosis. Stroke 1992;23: 215–220. 15. Newman AB, Sutton-Tyrrell K, Rutan GH, Locher J, Kuller LH. Lower extremity arterial disease in elderly subjects with systolic hypertension. J Clin Epidemiol 1991;44:15–20. 16. Moher D, Cole CW, Hill GB. Epidemiology of abdominal aortic aneurysm: the effect of differing definitions. Eur J Vasc Surg 1992;6:647– 650. 17. Krohn CD, Kullmann G, Kvernebo K, Rosen L, Kroese A. Ultrasonographic screening for abdominal aortic aneurysm. Eur J Surg 1992;158:527–530. 18. Morris GE, Hubbard CS, Quick CR. An abdominal aortic aneurysm screening program for all males over the age of 50 years. Eur J Vasc Surg 1994;8:156 –160. 19. Lindholm L, Ejlertsson G, Forsberg L, Norgren L. Low prevalence of abdominal aortic aneurysm in hypertensive patients. Acta Med Scan 1985;218: 305–310. 20. MacSweeney ST, Young G, Greenhalgh RM, Powell JT. Mechanical properties of the aneurysmal aorta. Br J Surg 1992;79:1281–1284. 21. Karanjia PN, Madden KP, Lobner S. Coexistence of abdominal aortic aneurysm in patients with carotid stenosis. Stroke 1994;25:627– 630. 22. Martin P. On abdominal aortic aneurysms. J Cardiovasc Surg 1978;19:597– 598. 23. Wolf YG, Otis SM, Schwend RB, Bernstein EF. Screening for abdominal aortic aneurysms during lower extremity arterial evaluation in the vascular laboratory. J Vasc Surg 1995;22:417– 421. 24. Bengtsson H, Bergqvist D, Ekberg O, Ranstam J. Expansion pattern and risk of rupture of abdominal aortic aneurysms that were not operated on. Eur J Surg 1993;159:461– 467. 25. Stonebridge PA, Draper T, Kelman J, Howlett J, Allan PL, Prescott R, Ruckley CV. Growth rate of infrarenal aortic aneurysms. Eur J Vasc Endovasc Surg 1996;11:70 –73. 26. Grimshaw GM, Thompson JM, Hamer JD. A statistical analysis of the growth of small abdominal aortic aneurysms. Eur J Vasc Surg 1994;8:741–746. 27. Scott RA, Wilson NM, Ashton HA, Kay DN. Influence of screening on the incidence of ruptured abdominal aortic aneurysm: 5-year results of a randomized controlled study. Br J Surg 1995;82:1066 –1070. 28. Harris PL. Endovascular grafting for abdominal aortic aneurysms. Ann R Coll Surg Engl 1996;78(suppl 1):23–24. 29. Cheatle TR. The case against a national screening programme for aortic aneurysms [comment]. Ann R Coll Surg Engl 1997;79:310 [Source: Ann R Coll Surg Engl 1997;79:90 –95]. 30. Frame PS, Gryback DG, Patterson C. Screening for abdominal aortic aneurysm in men ages 60 to 80 years. Ann Intern Med 1993;119:411– 416.
MARCH 1, 1999