A comparison of women and men with small abdominal aortic aneurysms

Abstract

Background

Sex-related differences between men and women with small abdominal aortic aneurysms (AAAs) have implications for patient care and research. Women are less likely to be affected by AAA than men in the United States. This study explored sex-related differences in patient characteristics, aneurysm growth and clinical outcomes in the Non-invasive Treatment of Abdominal Aortic Aneurysm Clinical Trial (N-TA³CT).

Method

: We conducted a retrospective cohort study performing a secondary analysis of data from N-TA³CT (NCT01756833), a placebo-controlled, RCT assessing doxycycline treatment of AAA. We included 166 N-TA³CT participants with baseline AAA maximum transverse diameter ≤ 4.5 cm and at least one follow-up CT scan. We compared patient characteristics, aneurysm growth rates, adverse events, and clinical outcomes between men and women.

Results

Among 166 participants, 35 (21.1%) were women and 131 (78.9%) were men. Women were more likely than men to have a family history of AAA (31.4%% versus 16.0%, P = 0.05). Women were less likely to take statins (68.6% versus 84.7%, P = 0.05) or any antiplatelet agent (48.6% versus 77.9%, P = 0.001). Women had significantly faster annual AAA growth rates than men (0.21 versus 0.16 cm/year, P < 0.001).

Conclusions

Within the size range of 3.5–4.5 cm, AAAs grow faster in women than in men. Women with AAA report a family history of AAA more frequently than men do.

This study explored sex-related differences in patient characteristics, aneurysm growth and clinical outcomes in the Non-invasive Treatment of Abdominal Aortic Aneurysm Clinical Trial (N-TA3CT). We included 166 N-TA3CT participants with baseline AAA maximum transverse diameter < 4.5 cm and at least one follow-up CT scan. We found that women had significantly faster annual AAA growth rates than men (0.21 versus 0.16 cm/year, P < 0.001).

Introduction

Abdominal aortic aneurysms (AAA) are present in 0.5–1.3% of women and 4.0–7.6% of men 65 years of age and older^1–4. Although the prevalence of AAA is higher in men, women who develop AAAs have been reported to have faster growth and increased rupture rates at smaller diameters^5,6. The proportion of AAA that meet guideline recommendations for surgical intervention that are untreated is greater in women than men; and women’s risk of AAA rupture within 5 years has been reported to be higher than men’s⁷.

Our understanding of AAA disease in women is poorly informed, as exemplified by the uncertainty of the diameter threshold for elective repair of asymptomatic patients⁸. This is partly related to the small number of women enrolled in AAA clinical trials^9–11. The importance of expanding this understanding is underscored by the significantly higher mortality of women versus men with AAA¹².

The Non-invasive Treatment of Abdominal Aortic Aneurysm Clinical Trial (N-TA³CT, NCT01756833) was a placebo-controlled RCT assessing doxycycline to reduce AAA growth in patients with small infrarenal AAA, that is 3.5–5.0 cm (men) and 3.5–4.5 cm (women). The present analysis, using data from N-TA³CT, compares sex-related differences in patient characteristics, aneurysm growth assessed using serial CT scans, clinical outcomes, and adverse events¹³.

Methods

Study design

We conducted a retrospective cohort study using data from N-TA³CT. The clinical trial design and methods have been previously reported^9,13. The protocol was approved by the institutional review boards at all clinical sites, core laboratories, and coordinating centres. All randomized participants provided written informed consent. The trial reported that among patients with small infrarenal AAA, doxycycline compared with placebo did not significantly reduce maximum transverse diameter (MTD) growth over the course of 2 years⁹.

Study population

For the current study we restricted analyses to men and women with baseline MTD ≤4.5 cm who had at least two CT scans to calculate growth rate. The MTD restriction was made because the N-TA³CT eligibility criteria were set to accommodate standard of care surgical interventions at 5.0 cm MTD for women. We restricted our study population to those with MTD ≤4.5 cm at baseline so that men and women would be comparable. As doxycycline did not affect AAA growth⁹, the two treatment arms were pooled for analyses. Patients with fusiform and saccular appearance of aneurysms were included.

Growth rate assessment

Abdominal and pelvic CT scans were obtained using a standardized image acquisition (section thickness less than or equal to 2.5 mm, reconstruction overlap minimum 50%, narrow field-of-view centred on aorta) protocol at baseline and approximately every 6 months after randomization^13,14. CT scans were analysed in a central Imaging Core Laboratory with validated measurement techniques by a single reader with an intraclass correlation coefficient exceeding 0.99^9,13. The single reader underwent quarterly proficiency testing compared to three other readers in the corelab to ensure there was no systemic drift over time, and measurements were performed without comparison to prior values. Infrarenal AAA MTD was measured from outer wall to outer wall with a clinically validated technique using electronic callipers in an image perpendicular to the centreline using the double-oblique technique¹⁵. Individual growth rates (cm/year) were assessed with ordinary least-squares linear regression across all available CT measurements (that is up to 6). Linear growth was defined by an R² value of at least 0.90 for the regression model.

Clinical outcomes and adverse events

Clinical outcomes in the current analysis were defined as surgical repair of AAA or death within the two-year N-TA³CT trial period. Adverse events were defined as frequent joint pain, headache, gastric or intestinal upset, bleeding or bruising, spells of dizziness, visual disturbance, moderate to severe sunburn, skin rash, tooth discoloration, or persistent fever at any point during the two-year N-TA³CT trial. We also assessed discontinuation of treatment (placebo or doxycycline).

Statistical analyses

Continuous variables are reported as mean(s.d.) or median (i.q.r.). Categorical data are reported as count (%). Two-sample Student’s t-tests (two-tailed) were used to compare continuous variables and Fisher’s exact tests were used to compare categorical data.

Comparison of growth rates between men and women was performed using linear mixed models. Baseline variables associated with sex at P ≤ 0.1 were added to the regression model. The final comparison in both adjusted and unadjusted models was based on the interaction between sex and time for MTD with both main effects (sex and time) in the model. We plotted growth rate as a function of baseline MTD by sex and tested the significance of the sex term by regressing growth rate on baseline MTD and sex in a linear model. Additionally, we examined a three-way interaction (sex, time, and treatment assignment) for MTD. Interaction was examined between sex and treatment assignment for all adverse events using logistic regression. Statistical analysis was performed using Stata MP 17.0 (StataCorp LLC). P < 0.05 was taken to be statistically significant.

Results

Of the 261 people randomized in the N-TA³CT trial, 95 were excluded (93 with MTD > 4.5 cm and 2 with <2 CT scans), leaving 166 (131 men and 35 women) who met study criteria and who were included in the present analysis. Baseline characteristics are summarized in Table 1. Mean age was 70.9(s.d. 7.4) years for men and 70.4(s.d. 7.5) for women (P = 0.73). Mean baseline MTD did not differ significantly between men (4.1 cm(s.d. 0.3)) and women (4.0 cm(s.d. 0.3)) (P = 0.17). Women were more likely than men to report a family history of AAA (31.4%% versus 16.0%, P = 0.05). Women were less likely to take statins (68.6% versus 84.7%, P = 0.05) or any antiplatelet (48.6% versus 77.9%, P = 0.001) compared to men. Women were less likely to report a history of diabetes (17.1% versus 23.7%), but this difference was small and not statistically significant (P = 0.50).

Table 1.

Baseline characteristics of participants in the non-invasive treatment of abdominal aortic aneurysm (AAA) clinical trial with maximum transverse diameter (MTD) between 3.5 and 4.5 cm at baseline and more than one CT scan, n = 166

	Men (n = 131)	Women (n = 35)	P*
Age, years	70.9 (7.4)	70.4 (7)	0.73
AAA MTD	4.1 (0.3)	4.0 (<1)	0.17
Height, cm	176.5 (7.8)	163.9 (8)	<0.001
BMI	29.5 (5.5)	28.0 (5)	0.10
Plasma biomarkers
MMP-9	36.0 (27.1, 54.2)	44.8 (31, 95)	0.12
CRP	2.2 (1.2, 4.5)	3 (1, 6)	0.31
Treatment assignment			0.52
Doxycycline	69 (52.7)	18 (51)
Placebo	62 (47.3)	17 (49)
Tobacco use			0.40
History of smoking	79 (60.3)	17 (49)
Current smoker	45 (34.4)	15 (43)
Never smoker	7 (5.3)	3 (9)
Co-morbidities
Atrial fibrillation	17 (13.0)	4 (11)	0.99
Cancer	38 (29.0)	11 (31)	0.84
COPD†	27 (20.6)	9 (26)	0.50
Coronary artery disease	55 (42.0)	10 (29)	0.18
Diabetes	31 (23.7)	6 (17)	0.50
Family history of AAA	21 (16.0)	11 (31)	0.05
Hypercholesterolaemia	104 (79.4)	23 (66)	0.12
Heart failure	9 (6.9)	2 (6)	0.99
Kidney disease	23 (17.6)	4 (11)	0.45
Myocardial infarction	24 (18.3)	3 (9)	0.20
Stroke	13 (9.9)	3 (9)	0.99
Medications
Antihypertensives
Beta blocker	70 (53.4)	18 (51)	0.85
Diuretics	42 (32.1)	12 (34)	0.84
ACE‡ inhibitor	47 (35.9)	8 (23)	0.16
Calcium channel blocker	31 (23.7)	7 (20)	0.82
Angiotensin receptor blocker	21 (16.0)	8 (23)	0.33
Statin	111 (84.7)	24 (69)	0.05
Any antiplatelet	102 (77.9)	17 (49)	0.001
Daily aspirin	94 (71.8)	17 (49)	0.02

Data are presented as n (%), mean ± standard deviation, or median (i.q.r.). *P from Fisher’s exact test, Student’s t-test, or Wilcoxon rank sum. †Chronic obstructive pulmonary disease. ‡Angiotensin-converting enzyme.

Aneurysm growth

Over an average of 3.1 study visits with 92% of men and 83% of women having ≥4 visits, mean growth rate was significantly faster in women (0.21 cm/year; 95% c.i. 0.19, 0.23) than in men (0.16 cm/year; 95% c.i. 0.15, 0.17), P < 0.001. This difference in AAA growth rate (0.05 cm/year; 95% c.i. 0.03, 0.07) remained unchanged after adjustment for history of hyperlipidaemia, height, family history of AAA, statins, and any antiplatelet therapy. When we forced smoking, diabetes, and treatment assignment into our regression model, the estimates for the difference, statistical significance and 95% confidence interval remained unchanged. Figure 1 displays growth rate as a function of baseline AAA MTD in men and women. On average, women had faster growth at every MTD value relative to men (P = 0.02). The three-way interaction between sex, treatment assignment, and time was statistically significant (P = 0.003). Women on doxycycline grew faster (0.25 cm/year; 95% c.i. 0.22, 0.28) than women on placebo (0.18 cm/year; 95% c.i. 0.15, 0.21). There was no difference in growth by treatment assignment in men.

Fig. 1.

Baseline maximum transverse diameter (MTD) and growth rate in men and women in the non-invasive treatment of abdominal aortic aneurysm clinical trial with MTD between 3.5 and 4.5 cm and more than one CT scan, n = 166

A linear growth pattern was observed in 83% of women and 70.2% of men (P = 0.20).

Clinical outcomes and adverse events

The frequency of aneurysm repair among women during the 2-year follow-up was not significantly different from that among men (3% versus 0.8%, P = 0.38) (Table 2). Women were significantly more likely than men to experience frequent headaches (38% versus 15.3%, P = 0.007), dizziness (41% versus 22.9%, P = 0.05), and gastrointestinal upset (65% versus 41.2%, P = 0.02). Interaction between sex and study treatment was not significant for any of the individual adverse effects. Based on capsule counts and prescriptions, women were more likely than men to permanently discontinue doxycycline treatment (61% versus 18.8%, P < 0.001). There was no significant difference in mortality rate (3% versus 1.5%, P = 0.51).

Table 2.

Adverse events reported by participants in the non-invasive treatment of abdominal aortic aneurysm clinical trial with maximum transverse diameter between 3.5 and 4.5 cm at baseline and more than one CT scan, n = 166

	Men, n = 131	Women, n = 35	P*
Adverse events
Frequent joint pain	83 (63.4)	18 (53)	0.32
Frequent headache	20 (15.3)	13 (38)	0.007
Frequent gastric or intestinal upset	54 (41.2)	22 (65)	0.02
Frequent bleeding or bruising	57 (43.1)	9 (27)	0.08
Frequent spells of dizziness	30 (22.9)	14 (41)	0.05
Visual disturbance	33 (25.2)	10 (29)	0.66
Moderate to severe sunburn	26 (19.9)	6 (18)	0.99
Skin rash	38 (29.0)	7 (21)	0.39
Tooth discoloration	13 (9.9)	4 (12)	0.75
Fever	14 (10.7)	3 (9)	0.99
Permanent discontinuation of treatment†
Placebo (n = 79 (men = 62, women = 17))	7 (11.3)	4 (24)	0.20
Doxycycline (n = 87 (men = 69, women = 18))	13 (18.8)	11 (61)	<0.001
Aneurysm repair	1 (0.8)	1 (3)	0.38
Mortality	2 (1.5)	1 (3)	0.51

Data are presented as n (%). *P from Fisher’s exact test. †Denominator.

Discussion

Small AAA grew on average 0.05 cm/year faster in women than men. The strong evidence for this finding stands out in literature that has been less definitive because of the precision of CT scan measurements in the N-TA³CT core laboratory and the sensitivity of the generalized linear model used in analysis. This difference in growth rate would have clinical consequences over the course of 5–10 years, placing women over an interventional threshold at 5.0 cm more quickly than men (at 5.5 cm)^16,17.

Although our follow-up is limited, these data suggest that distinct follow-up intervals for women may be appropriate and potentially beneficial. As the evidence for elective repair threshold is based on clinical trials enrolling primarily men, the advocacy of a lower threshold in women is less informed. In the absence of direct evidence in women, thresholds have been based on ratio hypotheses. A threshold of 5.2 cm has been suggested for elective repair among women, based on targeting a similar ratio of elective/emergent repair rates as observed in men¹⁸. Absent any ruptures observed in this study population, current repair guidelines seem reasonable.

The present findings are consistent with those from previous studies that reported faster AAA growth rates in women compared to men^19,20. A recent larger study of women with AAA, but without men for concurrent comparison, reported that mean annual growth rate increased with MTD size and AAA pathology²¹. Average annual growth rates reported for women with AAA diameter between 3 and 4.9 cm were consistent with those observed in this study²¹. N-TA³CT did not include women with AAA MDT >4.5 cm. However, based on an average of more than three CT scans per participant, a linear model fit the data well for over 80% of women and 70% of men.

N-TA³CT findings are consistent with those in RESCAN. The meta-analysis performed by the RESCAN investigators reported faster growth in women with small AAA, but the difference was not statistically significant²². Their standard error of the estimate of the difference in growth was larger than any other standard error they observed. There was a large variability across RESCAN studies (I² = 78%); these studies collected data between 1976 and 2008 using primarily ultrasound technology. The upper 95% confidence interval of the RESCAN estimate for the difference in AAA growth between men and women across studies was 0.04 cm/year, not far from our estimate of 0.05 cm/year considering the differences in study performance²².

With greater precision of AAA measurement in N-TA³CT due to its strict CT assessment protocol across sites, we were able to detect a significant effect of sex on growth. A difference of 0.05 cm/year is less than the usual CT scan resolution from one year to the next in an individual, but it has an impact over years of follow-up. By measuring MTD perpendicular to the aneurysms’ walls, the present study did not encounter extremely large or small outlier values that were found in ultrasound-based studies such as in Mofidi et al.¹⁹. These orthogonal MTD measurements are more reproducible and accurate than axial CT or ultrasound images used in the previous studies^14,23. Additionally, the present analysis was restricted to one congruent range of baseline AAA diameters, avoiding the loss of precision entailed in including men at MTD that excluded women from enrolment.

Most small AAAs in both men and women grew linearly. Although women were more likely to have a linear growth pattern (83% of women and 70.2% of men), this difference was not statistically significant. These results add to the findings of Olson et al.¹⁴ showing that the majority (70%) of AAA undergo linear and predictable growth within a 2-year interval.

Women were more likely than men to report a family history of AAA as has been reported previously²⁴. Akai et al. found that family history of aortic aneurysm was an independent risk factor for faster-growing small AAAs in Japan²⁵. The present results are interesting. Future studies could integrate biomarker or genomic data to explore potential mechanisms.

Both statins and antiplatelet agents were self-reported to be used more often in men than in women. There is evidence suggesting that both statins and antiplatelet agents are associated with slower AAA growth^26,27. Our data show that the undertreatment of women with cardiovascular disease is a serious and ongoing problem for the medical community to address²⁸. However, in the present study the differences between men and women in reported use of statins or antiplatelet agents did not influence the adjusted comparison.

There were observed differences in reported adverse events such as dizziness, headaches, and gastrointestinal effects between men and women that were not significantly associated with doxycycline. However, women were much more likely than men to permanently stop treatment while on doxycycline. Women appear to tolerate full doses of doxycycline less well than men. The high proportion of women who discontinued doxycycline creates uncertainty as to whether the observation of greater growth rate among women assigned to doxycycline is an unusual chance observation or a true biological effect. Although there is little clinical implication for our patients as the N-TA³CT results do not support the use of chronic doxycycline regimens, this finding may have importance for other patients. Additional studies on the natural history of AAA growth in women are needed to confirm our estimated growth rate in the placebo group.

These data addressing characteristics and growth of AAA in women represent one of the larger published cohorts of women with AAA and the largest we know of with a prespecified CT scan protocol assessed in a core laboratory. However, this study has limitations worth considering. First, the number of women in the study is much smaller than the number of men, limiting efficiency. Given that our findings were statistically significant, questions about power for a test of hypothesis are superseded by a 95% confidence interval that was 0.03–0.07 cm/year for the difference in growth rate. As N-TA³CT did not include individuals without small AAAs, we cannot discern whether significant differences were a result of sex or a sex–AAA interaction. Follow-up was limited to 2 years. This period is too short to come to conclusions about rupture or mortality rate; longer-term follow-up would be necessary to determine if the growth findings persist over time.

On average, AAA in women grew 0.05 cm/year faster than in men. Future studies could determine optimal observation intervals according to sex and MTD and inform insight into mechanisms of the difference in growth between men and women with biomarker and genomic studies.

Funding

This work was supported by grants R34AG028684, R01AG037120, and P30AG0288747 from the National Institute on Aging to the University of Maryland, Baltimore, the National Institute of Diabetes and Digestive and Kidney Diseases (T35DK095737) and the Office of Women’s Health of the US National Institutes of Health.

Author contributions

Drs Kazmers, Shames, Tsai, Matsumura, and Baxter contributed to study design, data collection, results interpretation and drafted and approved the final manuscript. Drs Curci, Blackwelder and Terrin contributed to study design, statistical analysis, results interpretation and drafted and approved the final manuscript. Drs Ngai, Olson and Albrecht contributed to statistical analysis, results interpretation and drafted and approved the final manuscript. All authors agree to be accountable for all factual aspects of the work and for ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Disclosure

None.

Data availability

Investigators interested in analysing the N-TA³CT data are invited to reach out to Dr Michael Terrin at mterrin@som.umaryland.edu.