Principal findings
This study was undertaken to investigate the outcome of women undergoing surgical intervention for severe aortic regurgitation, looking specifically at concomitant aortic pathology. The reason to focus on the aorta was the provocative data from Klodas and colleagues, 3 who found women had a much worse survival after operations for isolated aortic regurgitation than did men or age-matched controls. The data pointed to aortic pathology as a possible explanation for reduced survival. An aortic aneurysm was present in 51% of the women compared with 25% of the men, yet a similar percentage of men and women underwent aortic root replacement. Late fatal aortic rupture or dissection occurred in 17% of the women versus 2% of the men. 3
In our series most of the valves were morphologically normal, with aortic enlargement the likely cause of the aortic regurgitation. No valve-sparing operations were performed. This reflects the calendar years of the study and perhaps reflects that the surgical focus was primarily on aortic regurgitation and not aortic enlargement.
Aneurysmal changes of the ascending aorta were present in our study in 58% of the patients as defined by the surgeon, yet 38% of the patients not having ascending replacements had aortas that measured 4 cm or greater. This group was at particularly high risk, with 32% experiencing an aortic event, 71% of these being in the ascending aorta. The risk was most strongly associated with size indexed to the patient’s size and not absolute aortic size.
From our data, 2.4 cm/m2 or greater than 4.0 cm appears to be a reasonable threshold to consider aortic replacement in this group of patients. It is important to remember that the risk of aortic events is also related to age and is increased at similar aortic sizes in older patients. The risks of replacement will need to be balanced against the risks of future events.
General discussion
Why is there such heterogeneity in the threshold for aortic replacement in this series? One reason may have been concern about the increased risks of adding an aortic replacement to the valve procedure. More likely is the general lack of consensus in the surgical literature of the appropriate size for aortic replacement in aneurysmal disease. For example, the Ad Hoc Committee of Reporting Standards, Society for Vascular Surgery, has recommended defining the presence of an aneurysm when it is 50% larger than the normal size. 13 The normal size of the ascending aorta in this report was based on a 1949 radiographic study of aortic dimension. More recent literature has recommended replacement when the ascending aorta is larger than 5.5 to 6.0 cm. 6,14 Depending on the criteria used, less than 25% to over 75% of the patients in this study would have had aortic replacement. In these publications there was no recommendation to adjust dimension criteria to the patient’s size, age, sex, or aortic pathology during valve or coronary procedures.
In a similar series from Japan of 86 patients with severe aortic regurgitation Natsuaki and colleagues 15 found a 59% 10-year survival and 75% freedom from aortic events in patients with ascending aortas of greater than 4.0 cm at the time of valve replacement. Aortoplasty appeared to be protective. In this study the average BSA was 1.6 m2. In the group with the aortic dimension/BSA ratio of less than 2.0 cm/m2, there was one late death and no complications compared with the group with a ratio of greater than 2.9 in which there were 10 aortic complications and 9 late deaths. Thus, the ratio for aortic replacement of 2.4 cm/m2 as determined in our article coincides with the 2.5 cm/m2 (4.0 cm) as recommended in Natsuaki’s article.
Is the risk of rupture found in women only a size issue, or is there a sex-based predilection to aortic enlargement, especially with aging? A recent study highlights sex differences in thoracic aortic aneurysm occurrence and rupture risk. A population-based cohort study from Olmsted County, Minnesota, found that the incidence of thoracic aortic aneurysm was the same for men and women, but women had similar sized aneurysms (4.9 cm on average) much later in life: 63 years for men versus 76 years for women. Of the ruptures that occurred, 79% were in women, with a 5-year rupture risk of 33% versus 9% for men. 16 The reason for this increased rupture risk associated with older age in women is unknown. One possible explanation is that the standard size criteria for intervening may not be neutral to sex, and if indexed to BSA, women’s aortic sizes would reach the threshold for intervention at a smaller absolute size. We now know from a recently published study that the aortic enlargement associated with bicuspid valve disease is related to severe degenerative changes in the aortic wall and not the valve disease. 17 Baron and Galea 18 have found that the media/ intima ratio of collagen and elastin content of the carotid artery decreases in postmenopausal women, suggesting that wall structure and strength may deteriorate with age. How these findings, aging, and the hormonal state of postmenopausal women affect aortic wall structure and composition requires further study.
Limitations
The study is retrospective in nature, and the size criteria for defining aneurysm disease were not uniform. This fact, however, allowed us to investigate the outcome of patients who did and did not have aortic replacement yet had similar sized ascending aortas. Determination of the exact cause of death was problematic with the limited number of autopsies. Detailed interviews with patients’ families and physicians strongly support the designation of aortic rupture as the cause of death but cannot prove it. Similarly, the site of rupture is inferred from the clinical data and is labeled as a likely site of rupture in the article to highlight this limitation. Without a direct comparison with a similar group of men, the effect of sex on outcome remains unclear.
Clinical inferences
The data suggest that our current size criteria for aortic replacement are too conservative in women. As aortic operation has become increasingly refined and safe, liberalizing the size criteria for replacement is appropriate and potentially life-saving. Cohn has recommended that during aortic valve replacement for aortic stenosis, moderately enlarged (4.5 cm) aortas should be replaced. 19 Because dimension is the most common criteria used, we advocate normalizing to BSA and using 2.4 to 2.5 cm/m2 as the threshold to consider intervention when operating on aortic valve pathology. Recognizing the abnormal aortic wall pathology in patients with a bicuspid aortic valve supports an aggressive approach in these patients. Age appears to be an important risk factor for future events. Whether sex itself is a risk and whether intervention is required in an asymptomatic woman with a normally functioning aortic valve but an aorta around 2.4 cm/m2 requires further study. Finally, postoperative radiographic surveillance of the remaining aorta is imperative as ascending aortic replacement reduces but does not eliminate the risk of future aortic events.
Appendix: Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
Appendix: Discussion
References
Robert J. Rizzo (Boston, Mass). Dr McDonald and her coauthors direct our attention to the problem of late aortic rupture after aortic valve replacement for aortic regurgitation in women. Dr McDonald, you report a 25% incidence of aortic rupture at 10 years in this group. You also correlate ascending aortic diameter to the risk of aortic rupture but not specifically to the risk of ascending aortic rupture, which would be more useful. I have tried to extract this ascending aortic rupture event rate from the data in the article.
In patients who had an ascending aortic diameter of more than 5.0 cm and did not have aortic replacement, 3 (60%) of 5 patients had ascending aortic rupture. If the ascending aortic diameter was between 4.0 and 5.0 cm and not replaced, 2 (14%) of 14 had ascending aortic rupture. If the aortic diameter was less than 4.0 cm and not replaced, none of the patients experienced a rupture. Interestingly, if the aortic diameter was greater than 4.0 cm and was replaced, 2 (3%) of 58 patients still had ascending aortic rupture. Clearly, your data show that the risk of ascending aortic rupture after aortic valve replacement rises with increasing ascending aortic diameter and is decreased by replacement of the dilated aorta but not eliminated.
I agree with your conclusions that the threshold for aortic replacement should be lower for patients expected to have a normally smaller aorta, which in most patients are the small, young, or female patients. The threshold for ascending aortic replacement should also be lower for patients in whom the chest is opened for other cardiac procedures in which the dilated aorta may not handle manipulation and suturing very well, leading to trouble. Thus, it is convenient for both the patient and the surgeon to replace the dilated aorta to prevent this later trouble. I also agree strongly that any patient with a history of a dilated aorta, whether repaired or not, should have careful follow-up to watch for progression of aortic disease and allow elective repair, and this follow-up should include both the thoracic and abdominal aorta.
My questions are as follows. First, how many of the patients in this study had the diagnosis of aortic rupture confirmed by autopsy versus by clinical or radiographic signs versus by merely having sudden death? Two, why was there rupture in the ascending aorta despite previous ascending aortic replacement, and how can this be prevented? Three, is the diameter of the ascending aorta that you currently use as your threshold for ascending aortic replacement in women after aortic valve replacement, which you have described at the end here, different for patients with aortic valve stenosis versus regurgitation, and is it different for men?
View this table:
[in this window]
[in a new window]
Appendix Table I. Characteristics of women with aortic regurgitation
View this table:
[in this window]
[in a new window]
Appendix Table II. Diameter of aorta at various levels from preoperative echocardiograms
View this table:
[in this window]
[in a new window]
Appendix Table III. Variables examined in multivariable analyses
Dr McDonald. I thank Dr Rizzo for reviewing this article. To answer the first question, we did an extensive review of each death. Although we had only one autopsy confirmation, we investigated every death certificate and were confident that we had sorted out people with sudden cardiac deaths, of whom there were 8, versus those dying suddenly from aortic rupture.
In answer to your second question, some of the ascending aortic ruptures were late technical complications. For example, one patient died of a ruptured pseudoaneurysm. We found that women undergoing isolated valve replacement experienced early deaths, being primarily from ascending aortic rupture, whereas those undergoing an aortic operation experienced later deaths, mainly in the descending aorta. We currently do not have enough information on the preoperative aortic pathology in the cause of aortic rupture to identify a mechanism for these findings. However, we believe that the ascending aorta should be considered for replacement at the time of the operation, and the descending thoracic aorta needs to be monitored because of incomplete protection against future events.
View this table:
[in this window]
[in a new window]
Appendix Table IV. Univariable comparisons of survival (actuarial method)
The final question is about indications for aortic replacement in women undergoing a median sternotomy for either aortic stenosis or coronary grafting. In brief, we do not know the answer to this, and we do not know whether this is a phenomenon unique to the female sex or caused by the aortic regurgitation. My recommendation would be that surgeons consider replacing the aorta when it is approaching 2.4 cm/m2, regardless of the indication for an operation. Further studies are needed to support this recommendation.
severity impact on the aneurysm. Any ideas anyone? Thanks!!!!
