Abdominal aortic aneurysm in women: Prevalence, risk factors, and implications for screening

Methods 

All patients were screened through the Stroke and Aneurysm Vascular Evaluation (SAVE) Program sponsored and funded by Medtronic Inc (Minneapolis, Minn) and conducted by Life Line Screening (Cleveland, Ohio). The objective of the program was to increase awareness of vascular disease and to identify patients with AAA, carotid artery stenosis, or peripheral vascular disease. Individuals eligible for free screening included men >60 years and women >65 years of age with at least one risk factor associated with AAA (hypertension, history of tobacco use, known cardiovascular disease) or men or women >50 years of age with a history of AAA in a first-degree relative. Individuals with a known diagnosis of AAA were excluded from this analysis. Patients were recruited through advertisements that emanated from more than 100 individual hospitals or clinics, which served as screening centers and were equally distributed throughout the Northeastern, Southern, Midwestern, and Western United States. All patients were screened between May 2004 and December 2006. Screening events were coordinated by Medtronic and arranged with the hospitals and clinics by local vascular surgeons. Sponsoring hospitals agreed to provide floor space and local advertising in the form of flyers and poster advertisements, as well as spreading awareness among referring physicians. Prospective patients contacted Life Line Screening to determine eligibility, schedule a screening appointment, and receive prescreening preparation instructions. All equipment and technologists for the screening exams were provided by Life Line Screening. Translation services were provided locally by the screening centers when necessary.

On the day of screening, all patients completed a brief questionnaire, which provided demographic information and potential risk factors for aneurysmal disease (see Appendix, online only). Smoking history was defined as greater than or equal to 100 cigarettes in a lifetime. Patients were considered to have hypertension, hypercholesterolemia, or diabetes if they had been given these diagnoses by a physician or were treated for these conditions. Known cardiovascular disease was defined as having had a prior myocardial infarction, a history of percutaneous or surgical coronary revascularization, or other unspecified cardiac surgery. A family history of AAA was defined as a first-degree relative who carried a diagnosis of AAA.

Duplex ultrasonography of the abdominal aorta after a 6-hour fasting period was performed with a 3.5-MHz real-time ultrasound scanner. B-mode grey-scale images of the aorta were obtained in longitudinal and transverse orientations with the patient in the supine position. Measurements were obtained at the area of largest diameter in the supra-celiac, supra-renal, and infra-renal aorta, as well as the distal common iliac arteries. An infra-renal aortic diameter of 3.0 cm or larger was defined as an AAA for both men and women, and aneurysms were further segregated into those measuring 3 to 5 cm or >5 cm. All abnormal studies were confirmed by a second technologist, and all studies were reviewed by a physician. A results package was sent to each subject’s home address and to the office of the primary care physician identified by the patient when clinically relevant findings were discovered.

All data analysis and interpretation was conducted independently by our institution. For comparisons of demographic data and prevalence of AAA between subjects, continuous variables were expressed as mean ± standard deviation and compared with the Fisher exact t-test. Comparisons between dichotomous and nominal data were performed with χ2 analysis. Univariate analysis was performed to identify risk factors for AAA in women and odds ratios were developed by χ2 analysis. A multivariable logistic regression model with odds ratio and 95% confidence intervals (CI) was developed based on the results of the univariate analysis, with inclusion variables having a value of P ≤ .25.

In order to generate a valid statistical logistic regression model with a sufficient ratio of events to variables analyzed, variables including history of myocardial infarction, coronary revascularization (percutaneous procedure or bypass operation), and unspecified cardiac surgery were defined and analyzed as the composite variable termed “heart disease” for multivariable analysis. Incomplete responses on subject questionnaires were excluded from subgroup analysis, resulting in exclusion of 7.9% of female patients for failure to report family history and between 4.1% and 4.8% of female patients for all other variables examined on univariate analysis. Of the 10,012 total female patients, 831 (8.3%) were excluded from multivariable analysis because of incomplete reporting.

Results 

A total of 17,540 subjects were screened for AAA with duplex ultrasound and completed the prescreening questionnaire. This group included 10,012 women and 7528 men. Demographic information and comorbidities for both men and women are listed in Table I. Mean age was similar for men and women at 70.0 and 69.6 years, respectively. Men were significantly more likely to have a history of tobacco use (>100 cigarettes per lifetime) and to carry a diagnosis of heart disease or diabetes. Women were more likely to have hypertension or a family history of AAA.

Table I. Demographics and Comorbidities of Population Screened in the SAVE Program

		Women	Men	All	P-value
	Total	10,012	7,528	17,540
	Age (years)	69.6	70.0	69.7	0.75
	Race				<0.0001
	Caucasion	83.9%	88.9%		—
	African American	5.4%	3.1%		—
	Hispanic	2.1%	1.8%		—
	Asian	1.1%	1.4%		—
	Native American	3.8%	2.5%		—
	Other	3.7%	2.3%		—
	Tobacco use	47.4%	66.1%	55.4%	<0.0001
	Hypertension	63.7%	61.0%	62.6%	0.0003
	Hypercholesterolemia	63.5%	62.0%	62.9%	0.04
	Diabetes	13.9%	17.3%	15.3%	<0.0001
	Family history of AAA	10.7%	6.8%	9.0%	<0.0001
	Heart disease	12.0%	27.7%	18.7%	<0.0001
	Myocardial infarction	6.1%	15.2%		<0.0001
	Coronary revascularization	8.5%	24.1%		<0.0001
	Other cardiac surgery	5.5%	15.3%		<0.0001

As shown in Table II, a total of 74 aneurysms were detected in women while 291 were detected in men, for an overall prevalence of a previously unrecognized aneurysm of 0.7% and 3.9%, respectively. Included in these were four aneurysms larger than 5 cm in women and 31 in men. Table III displays the prevalence of aneurysms in women of different age cohorts, ethnic groups, and with various comorbidities. The prevalence of aneurysms increased with age, with rates as high as 1.4% in women over the age of 75 years and 2.7% in the oldest age cohort of women >85 years.

Table II. Prevalence of AAA (3-5 cm and >5 cm) in Men and Women Screened for Abdomninal Aortic Aneurysm

		All (n = 17,540)	Women (n = 10,012)	Men (n = 7,528)	P-value
	Any aneurysm	365(2.1)	74(0.7)	291(3.9)	<0.0001
	3-5 cm	330(1.9)	70(0.7)	260(3.5)	<0.0001
	>5 cm	35(0.2)	4(0.04)	31(0.4)	<0.0001

Table III. Prevalence of AAA (3-5 cm and >5 cm) in Women Stratified by Age and Comorbidities

		N	Any AAA	>5 cm
	Age
	<55 years (n = 499)	499	3(0.6)	0(0.0)
	55-64 years (n = 1856)	1856	2(0.1)	0(0.0)
	65-74 years (n = 4840)	4840	27(0.6)	0(0.0)
	75-85 years (n = 2598)	2598	36(1.4)	2(0.1)
	>85 years (n = 219)	219	6(2.7)	2(0.9)
	Race
	Caucasian	8404	52(0.6)	1(0.0)
	African American	540	6(1.1)	2(0.4)
	Hispanic	209	0(0.0)	0(0.0)
	Asian	102	1(1.0)	0(0.0)
	Native American	382	5(1.3)	0(0.0)
	Other	375	10(2.7)	1(0.2)
	Tobacco use	4485	47(1.0)	1(0.02)
	Hypertension	6098	45(0.7)	2(0.03)
	Hypercholesterolemia	6047	42(0.7)	2(0.03)
	Diabetes	1322	10(0.8)	0(0.0)
	Family history of AAA	924	9(1.0)	0(0.0)
	Heart disease	1201	24(2.0)	1(0.1)
	Myocardial infarction	587	15(2.6)	1(0.2)
	Coronary revascularization	812	18(2.2)	1(0.1)
	Other cardiac surgery	523	12(2.3)	1(0.2)

Univariate analysis in women revealed that age, heart disease, and history of tobacco use were significant risks factors for aneurysmal disease (Table IV). Women who were >65 years of age had a greater than fourfold risk of developing an AAA compared with women younger than 65 (odds ratio [OR] 4.57; 95% confidence interval [CI] of 1.98 to 10.54, P < .0001). Similarly, heart disease and a history of tobacco use were both associated with a greater than threefold increased risk of developing an AAA in women (OR 3.57; 95% CI 2.19 to 5.84, P < 0.0001 and OR 3.29; 95% CI 1.86 to 5.80, P < .0001, respectively). Race, family history, and other comorbidities, including hypertension and diabetes, did not appear to be associated with AAA in women.

Table IV. Univariate Analysis of Risk Factors Associated with AAA >3.0 cm in Women

		OR	95% CI	P-value
	Age
	>65 years	4.57	[1.98,10.54]	<0.0001
	>75 years	3.39	[2.15,5.37]	<0.0001
	>85 years	4.03	[1.72,9.38]	<0.00005
	Native American Race (vs Caucasion)	2.13	[0.85,5.36]	0.10
	Tobacco use	3.29	[1.86,5.80]	<0.0001
	Hypertension	1.35	[0.79,2.31]	0.27
	Hypercholesterolemia	1.04	[0.63,1.74]	0.85
	Diabetes	1.20	[0.60,2.36]	0.61
	Family history of AAA	1.50	[0.74,3.07]	0.25
	Heart disease (composite variable)	3.57	[2.19,5.84]	<0.0001
	Myocardial infarction	4.90	[2.73,8.80]	<0.0001
	Coronary revascularization	4.29	[2.26,8.12]	<0.0001
	Cardiac operation (not specified)	4.39	[2.53,7.63]	<0.0001

Each risk factor identified on univariate analysis remained significant on multivariable analysis (P < .0001, Table V). For example, increasing age was associated with AAA on both univariate and multivariable analysis; each 1-year increase in age increased the likelihood of AAA in women by 10% (OR 1.10; 95% CI 1.06 to 1.14, P < .0001). Multivariable analysis revealed that heart disease as a composite was a strong independent predictor of AAA in women (OR = 3.62, 95% CI 2.08 to 6.29, P <.0001) when adjusted for age, smoking history, family history, or race. History of tobacco use was found to be the strongest independent predictor of AAA in women on multivariable analysis as female smokers had a fourfold increased risk of developing an AAA relative to nonsmokers (OR = 4.02, 95% CI 2.17 to 7.44, P < .0001). The prevalence of AAA in women with a history of tobacco was 1.05% compared with 0.32% in nonsmokers. A trend toward an association between aneurysms and a family history of AAA in women was noted, but this did not reach statistical significance on multivariable analysis (P = .09).

Table V. Multivariable Analysis of Risk Factors Associated with AAA >3.0 cm in Women

		OR	95% CI	P-value
	Age (1 year incremental odds ratio)	1.10	[1.06,1.14]	<0.0001
	Native American Race (vs Caucasion)	1.41	[0.43,4.63]	0.57
	Tobacco use	4.02	[2.17,7.44]	<0.0001
	Family history of AAA	1.95	[0.90,4.22]	0.09
	Heart disease	3.62	[2.08,6.29]	<0.0001

Although overall prevalence of AAA in women was only 0.7%, women with multiple risk factors for aneurysmal disease had much higher rates (Table VI). For example, women over the age of 65 years with a history of tobacco use had a prevalence rate (1.7%) twice that seen in the entire female subject pool. When the additional risk factor of heart disease was added, a prevalence rate of 3.4% was observed. The highest rate of 6.8% observed was in women with three risk factors and a family history of AAA.

Table VI. Prevalence of AAA in Patients with Multiple Risk Factors

	Age >65 years	Tobacco use	Heart disease	Family history	N	Any AAA
	+	+	−	−	3,180	45(1.4)
	+	−	+	−	1,005	23(2.3)
	+	+	+	−	464	16(3.4)
	+	+	+	+	31	2(6.4)

Discussion 

Centers for Disease Control and Prevention (CDC) statistics report that 0.47% of the 1,246,324 deaths in females in 2003 were attributable to aortic aneurysms and dissections, compared with 0.74% of deaths in males.1 Prevalence rates for AAA in women have generally been found to be between 0.35% and 2.2% in the published literature.8, 13, 16, 29, 30 Overall prevalence for AAA in women in this study was 0.7% for women and 3.9% for men, rates that correlate closely with those of other studies. In the published literature, this disease is significantly more common in males, with up to four times higher prevalence rates in men across all age groups, and up to a 10 times higher prevalence for aneurysms greater than 5 cm in diameter.

Risk factors associated with AAA in women in this study included increasing age, a history of tobacco use, and cardiovascular disease. These risk factors have previously been shown to be associated with AAA in men, and are similar to those cited for women in other studies.8, 9, 16 An association between AAA in women and hypertension or hypercholesterolemia has been reported previously but was not observed in this analysis.8, 19, 29 Although some have suggested a negative correlation between diabetes and aneurysmal disease, we found no correlation between AAA and this comorbidity.9, 16 A history of tobacco use was found to be the strongest independent predictor of aneurysmal disease in women and should be considered an important avoidable risk factor. A family history of AAA has been shown to be a risk factor in previous studies; however, this association did not reach statistical significance in this analysis. A significant association between Native American race and aneurysmal disease in women was demonstrated in an earlier subset of these data.27 While a similar trend was detected in this full data set on univariate analysis (P =.10), this association was not present on multivariable analysis.

This study represents the largest cohort of women screened for abdominal aortic aneurysms to date and offers increased insight into the prevalence rates and risk factors in this previously under-represented gender group. The next largest studies regarding screening for AAA in women include those by Lederle et al (2001) involving 3450 women, Scott et al (2002), which screened 3052 women, and Pleumeekers et al (1995) involving 3143 women.13, 16, 29

Lederle et al reported findings in 3450 women between the ages of 50 and 79 years who were screened in the Veterans Affairs Aneurysm Detection And Management (ADAM) study.16 Overall AAA prevalence rates in that study were 1.0% for women and 4.3% for men, which compares similarly with our rates of 0.7% and 4.3%, respectively. Like our results, the ADAM study confirmed that subgroups of women exist that have a higher prevalence of AAA, such as women with increased age, cerebrovascular disease, and a history of smoking.

Screening for AAA with abdominal duplex has been shown to be cost-effective and has led to reduction in aneurysm-related death in men.10, 11, 14, 37, 38 However, the role of screening for AAA in women is less clear due to the lack of substantive data about prevalence and the impact of risk factors on prevalence. Although associated risk factors were not highlighted by Scott et al, their manuscript represents an important randomized trial assessing the impact of screening for AAA in unselected populations. Prevalence rates in women were higher in that study than we observed (1.3% vs 0.7%), but not outside the range cited by other authors.8, 16, 29, 30 This trial failed to demonstrate a difference in rupture rates between the screened and control populations at 5- and 10-year follow-up intervals and concluded that screening for women was neither clinically indicated nor economically viable.17 However, this relatively large study screened an unselected population of women without consideration of risk factors for aneurysmal disease, and over 50% of those identified did not undergo aneurysm repair.

For screening to be cost-effective and to result in reduced mortality for women as has been demonstrated for men, it will likely require limiting screening to appropriate high-risk populations and directing it toward those individuals willing and fit for aneurysm repair. Studies such as this one have now more clearly outlined the subgroups of women at risk for aneurysmal disease and defined the prevalence rates in these various subsets. For example, despite the overall prevalence of AAA of 0.7% in our study, women with multiple risk factors were found to have a prevalence of 6.4%. When these rates are considered in conjunction with studies revealing an increased risk of death in women with ruptured AAA, increased rate of aneurysm enlargement in women, and an increased risk of rupture of smaller diameter aneurysms compared with men, screening for aneurysms in women appears quite justifiable.23, 24, 31, 32, 33, 34, 36 In fact, a recent publication from Sweden using a Markov decision-analysis model suggested that even with a prevalence as low as 1.1%, the increased risk of rupture and its associated complications makes screening for AAA in women cost-effective.35 Conversely, women without any of these risk factors clearly have very low rates of aneurysm formation and do not warrant screening. Interestingly, only 4.6% of women in the present study had coexistence of three risk factors, and screening programs directed at such specific subsets of the female population could keep the overall costs associated with screening to a minimum. Obviously, further investigations including randomized trials of screening programs aimed at higher-risk subgroups are needed to demonstrate the cost-effectiveness and clinical utility of AAA screening in women.

The data presented in this manuscript has certain limitations. Selection bias may affect the prevalence discovered in the study population, and this could explain the relatively high percentage of patients with a family history of aneurysmal disease. On the contrary, however, healthy individuals may be more likely to seek and complete a screening program compared with sicker individuals, thus, leading to a lower prevalence rate than may be found in a randomly chosen sample of the population. Second, this population does not represent an unselected group but instead included only individuals who carry a diagnosis of at least one risk factor previously associated with AAA (ie, hypertension, history of tobacco use, etc) and may therefore be more likely to harbor an aneurysm. Third, this study relied upon self-report questionnaires for the characterization of risk factors, and bias can be introduced by the manner in which patients report their comorbid illnesses. Finally, clinical outcomes for these patients were not available for analysis, and this prevents us from drawing conclusions regarding the clinical utility or cost-effectiveness of this particular screening program.

In conclusion, despite the low prevalence of abdominal aortic aneurysms in women overall, this study has identified specific risk factors that increase the likelihood of aneurysmal disease in subgroups of women. These include increasing age (>65 years), tobacco use, and cardiovascular disease. Women with these risk factors are more likely to benefit from programs designed to reduce aneurysm-related mortality and should be considered for duplex ultrasound AAA screening.

Author contributions