Summary

Introduction

Treatment guidelines recommend that an abdominal aortic aneurysm in men should be considered for elective repair once the aneurysm has reached a diameter of 5·5 cm, based on randomised trials that compared early elective repair with surveillance; these trials included only a small proportion of women.¹ The decision to offer repair is made when the probable risk of rupture exceeds the risk of repair. Since the rate of rupture of small abdominal aortic aneurysm is four times higher in women than men at the same diameter,² women are often considered for repair at diameters smaller than 5·5 cm. Several studies^{3, 4} have suggested that the prognosis of individuals with abdominal aortic aneurysm is worse for women than men, with the operative mortality following repair of intact aneurysm being higher in women than men.

To determine whether women, compared with men, are being offered adequate management of their intact aneurysms once they have reached the aneurysm diameter threshold for intervention, three pieces of crucial information are required: the proportion of each sex morphologically suitable for endovascular aneurysm repair (EVAR; often preferred to open repair because it is minimally invasive and has a lower operative mortality), the proportion of each sex not offered repair (for which the suitability for EVAR might be influential), and the 30-day mortality after elective aneurysm repair, either EVAR or open repair.

Randomised trials have reported that 30-day operative mortality for elective repair of abdominal aortic aneurysm is about three times higher after open repair than after EVAR,⁵ but again, women were poorly represented in these trials, perhaps partly because the early generation endografts were not suitable for most women.^{6, 7} However, data from the large Medicare database in the USA suggest that this relative difference in 30-day mortality between EVAR and open repair is observed for both men and women, even though the absolute mortality was higher in women than men.⁸ An earlier systematic review,⁹ including studies published by July, 2009, showed that for elective open repairs the operative mortality (30 day or in-hospital) was 7·6% for women versus 5·1% for men. For EVAR, the mortality for women was 2·9% versus 1·5% for men, but included only 1014 women undergoing EVAR.

The aim of this study was to quantify how the prognosis of women with abdominal aortic aneurysm compares with the prognosis in men. Therefore, we have systematically searched for contemporary evidence that describes the outcome of women with an intact abdominal aortic aneurysm of the size for which repair is considered in three sequential stages of the care pathway.

Methods

Results

After our search and evaluation (appendix), five papers based on five studies^{13, 14, 15, 16, 17} were eligible for inclusion in the meta-analysis regarding suitability for endovascular repair. Another study¹⁸ also included suitability for endovascular sealing technology, but used a selected population. The characteristics of the included studies are summarised in the appendix. One of these studies included more than 100 women, most studies were small, and the quality of these studies was assessed as fair. The criteria of morphological suitability for EVAR were different in each study; three studies^{13, 15, 16} included all the aneurysms diagnosed (including those not offered intervention), one study¹⁷ did not specify which patients were being considered for EVAR, and one study¹⁴ only considered patients who had undergone elective repair. The largest study¹⁵ has published two further updates but neither provided sufficient information to merit inclusion in the review.^{19, 20} Threshold abdominal aortic aneurysm diameter for inclusion ranged from 4 cm to 5 cm. In total, information was available for 1507 men and 400 women in whom the suitability for EVAR ranged from 46% to 64% for men and from 25% to 47% for women. The overall pooled estimate of suitability for EVAR was 34% in women compared with 54% in men; both overall estimates had significant heterogeneity (figure 1A). For the proportion of women compared with men who were suitable for EVAR, the OR was 0·44 (95% CI 0·32–0·62), with moderate heterogeneity (figure 1B).

Open in a separate window

Figure 1

Forest plots of the proportion of aneurysm patients morphologically suitable for EVAR

(A) Women and men separately. (B) Women versus men. EVAR=endovascular repair.

Some studies considered relaxation of the morphological criteria for EVAR, which increased the proportion of women suitable for EVAR. For instance, in the largest study including 251 women,¹⁵ relaxation of the neck length to longer than 7·5 mm (from >15 mm) increased the suitability for EVAR from 63 (25%) women to 113 (45%) women. In one study,¹³ relaxation of the minimum iliac diameter from 7·5 mm or 8 mm to 6 mm increased the proportion of 41 women eligible for EVAR from 11 (27%) to 16 (39%). The type of endograft considered also affects the proportion of women suitable for EVAR. For example, in one excluded study¹⁸ of 78 women that considered both conventional endografts and endovascular sealing, 32 (41%) women were suitable for the Gore Excluder graft (Gore Medical, Flagstaff, AZ, USA), but 61 (78%) would have been eligible for endovascular-sealing technology.¹⁸

The largest study¹⁵ investigated how suitability for EVAR declined with increasing aneurysm diameter, with almost no women being considered suitable for EVAR if their abdominal aortic aneurysm diameter cutoff exceeded 6·5 cm compared with 30% of men being suitable. The other four studies^{13, 14, 16, 17} provided few demographic or clinical details, so it was not possible to investigate how the suitability for EVAR in women might depend on age or other characteristics.

Our search and evaluation of non-intervention rates (appendix) yielded four publications, all from the UK, including two papers^{21, 22} and two abstracts^{23, 24} (both with additional information provided by the authors). All four studies were retrospective, with a total of 1365 men and 247 women; the studies were assessed as of fair quality (appendix). The overall results showed greater heterogeneity for men than for women and suggested that a third (34%) of potentially eligible women were either not offered or refused abdominal aortic aneurysm repair. The proportion was about double the non-intervention rate in men (19%; OR 2·27, 95% CI 1·21–4·23; figure 2). The difference in non-intervention rates between men and women was highest for the earliest study²² at a specialist tertiary referral centre.

Open in a separate window

Figure 2

Forest plots of non-intervention rates in cohorts of patients assessed for intact abdominal aortic aneurysm repair

(A) Women and men separately. (B) Women versus men.

After our search and evaluation for studies on elective operative mortality (appendix), eight papers based on eight studies met the inclusion criteria.^{25, 26, 27, 28, 29, 30, 31, 32} For one study³¹ reporting from 1992 to 2012, the 30-day operative mortality data for the late era (2003–12) were obtained from the corresponding author. Similarly, the data for 30-day mortality for the time period 2006–10 in the Swedish study²⁸ were supplied by the corresponding author. All studies included consecutive patients undergoing EVAR or open repair, or both, for infrarenal abdominal aortic aneurysm between January, 2000, and March, 2015. One further study⁵ was identified and we excluded the few patients with repairs before 2000. Therefore, nine studies were included in the meta-analysis. All nine studies provided data for intact infrarenal aneurysms only: two^{5, 31} studies excluded symptomatic abdominal aortic aneurysm. The English study³² was population-based, but for EVAR only. We excluded a similar study,³³ including open repair but covering an earlier time period, because much of the 30-day mortality was not aneurysm-related.

Full details of the included studies are summarised in the appendix. Three population-based studies^{28, 30, 32} were done in Sweden (765 women), the USA (5421 women), and the UK (2304 women). Other, mostly smaller, studies,^{25, 26, 27, 29, 31} based on either single centre or voluntary registries, were all from the USA and included 2438 women. Individual patient data meta-analysis⁵ of four prospective randomised controlled trials contributed data for 148 women. These data gave a total of 11 076 women operated on in nine separate studies, with an overall 30-day mortality of 2·31% (95% CI 1·99–2·68) after EVAR with no heterogeneity (figure 3A) and 5·37% (95% CI 4·18–6·88) after open repair with slight heterogeneity (figure 4A). These data contrast with the more heterogeneous results for a much larger cohort of 52 018 men operated on in these same studies with an overall 30-day mortality of 1·37% (95% CI 1·12–1·68) after EVAR and 2·82% (95% CI 1·88–4·22) after open repair (Figure 3, Figure 4). The Medicare study³⁰ provided the most data for the results of open repair. When this study was removed from the meta-analysis, the change in 30-day mortality was little: 4·72% for women (95% CI 3·83–5·82) and 2·50% for men (1·76–3·55). Similarly for EVAR, after exclusion of the two largest studies,^{30, 32} the change in 30-day mortality was little: 2·55% for women (95% CI 1·83–3·55) and 1·24% for men (0·81–1·89).

Open in a separate window

Figure 3

Forest plots of 30-day mortality after EVAR for intact abdominal aortic aneurysm

(A) Women and men separately. (B) Women versus men. EVAR=endovascular repair.

Open in a separate window

Figure 4

Forest plots of 30-day mortality after open repair for intact abdominal aortic aneurysm

(A) Women and men separately. (B) Women versus men.

In the nine included studies, 30-day mortality was significantly higher in women than in men for both types of repair: the overall OR for EVAR was 1·67 (95% CI 1·38–2·04) and for open repair 1·76 (1·35–2·30; Figure 3, Figure 4). Confounding factors such as age, abdominal aortic aneurysm diameter, number of symptomatic abdominal aortic aneurysms included, and comorbidities were reported inconsistently, but the adjusted OR continued to show a higher mortality in women of almost double (table); only three studies reported separately for EVAR and open repair.

Discussion

Prognosis of abdominal aortic aneurysm is often regarded as worse in women than men,^{3, 4} with adverse factors likely to include the older age of women, and higher morbidity and mortality after elective aneurysm repair. The data from our systematic review do not suggest that the proportion of women morphologically suitable for EVAR has improved since earlier this century.^{6, 7} However, the operative mortality in both men and women has fallen substantially since a 2010 systematic review,⁹ particularly for open repair; although 30-day operative mortality remains almost twice as high in women as in men. Some of the reduction in operative mortality in women might have come at the expense of the high proportion of women (a third) who are assigned to a non-intervention policy; although, the literature on non-intervention policy remains sparse.

Surgical registries and national databases do not record or report the numbers of patients with abdominal aortic aneurysm who either are morphologically suitable for EVAR or denied elective repair. The only non-intervention data identified came from four single centres in the UK, with decisions about repair being made at a multidisciplinary team meeting: presumably the women not offered repair had extensive comorbidities and had a high risk of early postoperative death. Only one of these series²¹ has provided detailed follow-up data for those initially assigned to a non-intervention policy. The study showed that after 3 years, only about a third of these patients remained alive and that 37% had died of rupture. Similarly, in the other full paper,²² only 35% of non-intervention patients were alive at 2 years.

Since EVAR is a less invasive procedure than open repair and is associated with a third to half of the 30-day mortality, the rate of morphological eligibility for EVAR is likely to effect the non-intervention rates for elective abdominal aortic aneurysm surgery. In 2001, using early-generation endografts, one study⁷ reported that only 39% of women versus 60% of men were morphologically eligible for EVAR. Our systematic review shows no improvement, with an overall estimate of 34% of women being morphologically suitable for EVAR, based on the manufacturers' instructions for use. The principal morphological criteria rendering women unsuitable for EVAR include both short and angulated proximal aneurysm necks and the dimensions of the access iliac arteries. Many endovascular specialists are willing to relax the strict requirements of the instructions for use,³⁴ and adjunct techniques do exist, such as iliac conduits, to use EVAR in narrow or heavily calcified iliac arteries. Nowadays, lower profile devices suitable for smaller access vessels and devices that can adapt to shorter and more angulated necks are available. Potentially, this availability should allow more women to be treated by EVAR. However, summary information from the Characterization of Human Aortic Anatomy Project²⁰ for 2012–13 lower profile devices (including Aorfix [Lombard Medical, Didcot, UK], Endurant II [Medtronic, Minneapolis, MN, USA], and Ovation [Endologix, Irvine, CA, USA]) suggests that only 40% of women remain suitable for EVAR based on the instructions for use. Other information from this project indicates that in women older than 80 years, EVAR eligibility is particularly low (29%).¹⁹ No endografts have been designed specifically for women, but preliminary data suggest that the endovascular-sealing technology is the approach most likely to increase the proportion of women suitable for endovascular repair.¹⁸

A comprehensive systematic review and meta-analysis⁹ showing the sex-specific differences in mortality after either EVAR or open repair of abdominal aortic aneurysm was published in 2010. This review included publications from 1995 to July, 2009, for operative mortality (either 30-day or in-hospital mortality; in-hospital mortality is usually lower than 30-day mortality). Operative mortality was higher in women than in men (overall OR for EVAR 2·51, 95% CI 1·72–3·69 vs OR for open repair 1·50, 1·33–1·69). The results were dominated by a 20-year review (1980–2000) from the USA with 81 384 women³⁵ and contained few data for patients who underwent EVAR. By contrast, we focused on 30-day mortality only in studies published since Jan 1, 2009, which included more data for EVAR. Again the mortality for both EVAR and open repair was almost twice as high in women as in men, with mortality for EVAR being much lower than for open repair. Overall 30-day mortality for EVAR was 2·3% and for open repair was 5·4% for women versus 1·4% for EVAR and 2·8% for open repair in men. Although the overall mortality has decreased since the 2010 systematic review, the OR for women versus men has changed little. The mortality following open repair in women appears to be unacceptably high.

Since our analysis of 30-day mortality, we have become aware of two sources of recent data, covering the period 2011–14. A retrospective cohort study³⁶ of 6661 patients from the American College of Surgeons National Surgical Quality Improvement Program has shown that, for either EVAR or open repair of intact abdominal aortic aneurysm, 30-day mortality in women is double that of men and women have a 30% increase in complication rate, even after adjustment for age and aneurysm diameter: mortality in women was 2·9% for EVAR and 8·2% for open repair versus 1·1% for EVAR and 4·0% for open repair in men. Analysis of the similar size English administrative dataset (Hospital Episode Statistics) showed that 30-day mortality for elective admissions was 2·3% for EVAR and 7·0% for open repair in women versus 1·4% for EVAR and 4·1% for open repair in men (Michaels J A, University of Sheffield, personal communication). The apparent recent increase in mortality for open repairs might be explained by the decreasing proportion of operations undertaken as open repairs, which diminishes expertise at a time when the remaining open repairs are likely to be the most technically challenging cases.

These studies have several limitations. First and foremost, factors that might confound the comparison of men and women, such as age, aneurysm diameter, and comorbidities, were inconsistently reported. Operative mortality is known to increase with age, and women appear to develop clinically relevant aneurysms at an older age than men.^{37, 38} Since prevalence increases rapidly with age, especially in women,³⁹ this factor is an important consideration. Nevertheless, when adjusted data were available, the much higher 30-day mortality in women remained. Second, there was a varying amount of heterogeneity in the meta-analyses, but the results for all outcomes consistently showed sex-specific differences. Third, it can take several years for publications to appear in print, which might cover several generations of endografts and experience for EVAR, so obtaining contemporary data is challenging. Fourth, the patients selected for EVAR and open repair in the various studies are likely to be different, including differences in aortic morphology, which might affect mortality from both EVAR and open repair. Particularly when used outside the instructions for use, the complication, re-intervention, and mortality for EVAR could all increase.^{34, 40} Finally, a relative paucity of data is available about the number of women versus men either eligible for EVAR or not considered for any intervention, which underscores the importance of reporting primary data by sex.⁴¹

Moreover, our review has several strengths. All the studies included in this series had to include both men and women and present the data for each separately, so that men and women from the same population and the same time period are compared in each study. Furthermore, we presented results as OR and compared women with men in each study (and overall), since these OR should be less susceptible to heterogeneity than the results for men and women separately.

In summary, the results of these reviews indicate that the prognosis of women with abdominal aortic aneurysm is worse than that of men; although, we have not considered the management of cardiovascular risk (often worse in women),^{42, 43} complications after repair, and longer term quality or length of life. Women have smaller aortas than men,⁴⁴ and perhaps if a smaller threshold for both diagnosis and intervention were introduced, compared with those recommended for men, women might have a better chance of being offered and surviving intervention at a younger age. Given the relatively high mortality following open abdominal aortic aneurysm repair in women, it would seem to be crucial either to design an endograft for women or to accept that most women need to be referred to centres that either offer endovascular sealing technology or are specialists in open repair. Such measures might see an improved prognosis for women.

Acknowledgments

SWAN (Screening Women for Aortic aNeurysm) collaborators

Contributors

Declaration of interests

Contributor Information

SWAN collaborators^†:

References