- Email: [email protected]
Who Should We Operate On and How Do We Decide: Predicting Rupture and Survival in Patients with Aortic Aneurysm
Who Should We Operate On and How Do We Decide: Predicting Rupture and Survival in Patients with Aortic Aneurysm Mark Fillinger, MD The decision to operate on a patient with an aortic aneurysm is based on the risk of aneurysm rupture versus the risk of aneurysm repair, within the context of the patient’s overall life expectancy. Risk of rupture is still primarily based on the maximum aneurysm diameter, with some allowances made for factors that modify rupture risk, such as gender and current smoking. Newer methods for determining rupture risk, such as aneurysm-wall stress analysis, appear promising, but are not yet broadly available. Until then, diameterbased prediction rules for rupture risk will “fail” 10% to 25% of patients with both small and large abdominal aortic aneurysms. With regard to predicting operative mortality and life expectancy after open or endovascular aneurysm repair, multiple risk-stratification algorithms have been created. The best of these algorithms are accurate in 75% to 80% of patients, meaning that they fail in 20% to 25% of cases. Prediction algorithms provide significant guidance, but cannot take the place of an experienced clinician at this point. Somehow, experienced surgeons are able to sift through a massive amount of information and properly select patients who are appropriate for surgery, with quite reasonable perioperative and long-term mortality rates. Semin Vasc Surg 20:121-127 © 2007 Elsevier Inc. All rights reserved.
D
URING TRAINING, A surgical resident or vascular fellow naturally wants to focus on the technical aspects of a procedure. This “truism” certainly applies to aortic aneurysm repairs, perhaps more than any other procedure in vascular surgery. As any experienced vascular surgeon will attest, however, it is much harder to learn when to operate than it is to learn how to operate. Experienced surgeons understand that outcomes are closely linked to patient selection, in many cases learning this the “hard way” early in their careers. But why is patient selection so difficult, especially with regard to aneurysm repair? In a patient with an aortic aneurysm, the decision to operate is based on three primary factors: risk of aneurysm rupture, risk of aneurysm repair, and patient’s life expectancy. There is a fourth factor, the patient’s personal preference, but this can be heavily influenced by the information they assimilate with regard to the other three primary factors. The patient’s personal preference is also the easiest to deter-
Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, NH. Address reprint requests to Mark F. Fillinger, Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03750. E-mail: [email protected]
0895-7967/07/$-see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1053/j.semvascsurg.2007.04.001
mine: the patient will give that to the surgeon once they have discussed the three “primary” factors. Deciding who we should operate on should be simple then, shouldn’t it? Can’t we simply calculate the aneurysm rupture risk, the procedure risk and the patient’s life expectancy based on the literature and actuarial tables? As Shakespeare wrote, “Ay, there’s the rub . . .,” for understanding any of these seemingly simple factors can be quite difficult in any given patient. Decades of work have provided precious little in our understanding of these three issues, much less the interaction of the three factors in a given patient. Of course, this problem will not be solved within this article, but at least the issues can be framed in a potentially helpful manner.
Rupture Risk For the past five decades, the primary determinant of rupture risk has been maximum aneurysm diameter, based on the work of Szilagyi et al in the 1960s.1 These authors compared the outcomes of small-diameter and large-diameter aneurysms, and found that patients with larger aneurysms (⬎6 cm) were much more likely to die of a ruptured aneurysm. In that era, the diameter had to be determined by physical exam 121
122 and abdominal x-ray, which are both now known to overestimate the actual diameter, such that the actual threshold for aneurysm size was approximately 5 cm. The 5-cm threshold was adopted when ultrasonography became widely available, and has been used ever since, with minor modification throughout the decades based on clinical evidence that variables such as hypertension, pulmonary disease, gender, and smoking modified aneurysm rupture risk somewhat.2-7 Only recently has there been even a minor change in the diameter threshold for proposing surgery, because of randomized clinical trials, such as the United Kingdom Small Aneurysm Trial (UK SAT) and the Aneurysm Detection and Management study, suggesting the threshold for the typical patient should be 5.5 cm.8,9 Even with these results, some question whether women, who had a three- to fourfold risk of rupture under observation in UK SAT,2 should be repaired at 5 cm. Data from our own institution including 122 ruptured abdominal aortic aneurysms (AAA) indicate that the average size of ruptured aneurysms in women is 5 mm smaller than in men,5 lending further evidence that the size threshold should be slightly lower in women. Given the relative stability of the diameter threshold for considering surgery, one might reasonably assume that maximum aortic aneurysm diameter is an extremely reliable determinant for patient selection. Unfortunately, this is not the case. Although maximum aneurysm diameter works relatively well in general (as evidenced by summary data involving large numbers of patients), all clinicians know that this “metric” often fails in individual patients. Some aneurysms rupture at an unusually small size.4,10,11 In some series, 10% to 24% of the ruptured aneurysms were ⱕ5 cm in maximum diameter.2,11 Ultimately, large studies were conducted to determine whether it is safe to observe AAAs with a maximum diameter ⬍5.5 cm.8,9 In an effort to prevent rupture, these studies used frequent observation including ultrasonography or computed tomography (CT) every 6 months, with surgical intervention for symptoms, expansion, or growth to 5.5 cm. This resulted in a surgical intervention rate of approximately 60% in the “observation” group within 4 years, and ⬎75% of the observation patients had surgery within 6 years. Even with a high rate of intervention in a patient population willing to have frequent and reliable surveillance, the rupture rate can still be up to 2% per year, depending on gender and aneurysm size.2,9,12 Moreover, while diameter-threshold issues receive widespread study in patients with smaller aneurysms, there has not been as much study of larger aneurysms. In older and sicker patients, however, decision-making can be difficult, and repair might not be recommended until the aneurysm reaches a larger size. In patients with aneurysms ⬎5.5 cm, however, ⬎50% will rupture when surgery is deferred because of high operative risk.13 In this high-risk group, the median time to rupture was only 19 months for patients with 5.5- to 5.9-cm AAAs, and only 9 months for patients with AAAs ⬎7 cm.13 Even in patients under close surveillance, the aneurysm rupture rate is substantial when ⬎5.5 cm, and rises exponentially with size.14,15 While this might lead one to consider repair for nearly all patients with large aneurysms,
M. Fillinger studies such as EVAR 2 make it clear that there is no simple strategy.15 Lastly, even the method of determining maximum aneurysm diameter has methodological problems. Most aneurysms have an elliptical cross-section on axial CT imaging, and large studies have suggested that rupture risk is more closely associated with the minor axis.5,16 This makes sense with regard to three-dimensional reconstructions of CT imaging data, which demonstrates that the elliptical cross-section in most aneurysms is due to tortuosity rather than a truly asymmetrical shape.5 For those aneurysms with a truly asymmetrical shape, the larger “diameter” in the elliptical crosssection may be a more appropriate estimate of rupture risk.5 Despite these issues, however, many studies fail to precisely report how maximum aneurysm diameter was determined, and many centers still use the “major axis” diameter to estimate rupture risk for all aortic aneurysms. As an example of how influential such an issue can be, the UK SAT used the maximal antero-posterior (AP) diameter of the aneurysm by ultrasound, which will overestimate the “true” size of an aneurysm that bows anteriorly (which is not uncommon). Thus, even though maximum aneurysm diameter remains the gold standard for estimating rupture risk, it is far from ideal. We have demonstrated that finite element analysis of AAA wall stress using three-dimensional CT reconstructions is better than diameter for differentiating AAAs near the time of rupture,17 and that wall stress is superior to AAA diameter for predicting rupture risk in patients under observation.18 In 2002, in the first report using finite element analysis on a large cohort of AAA patients with and without rupture, our group reported that finite element analysis of AAA wall stress using three-dimensional CT reconstructions is better than diameter for differentiating AAAs near the time of rupture (Fig 1). Moreover, we found that calculated indices previously suggested to better predict rupture risk (eg, ratio of maximum AAA diameter to normal infrarenal diameter) were not helpful.17 Lack of utility for calculated indices was later confirmed in much larger cohort of 259 patients, 122 of them with documented rupture.5 In 2003, our group reported on the first large series of patients with AAAs under observation, looking at the natural history of aneurysms that had elective repair deferred or were felt to be at low risk of rupture. In that study, we found that wall stress is superior to AAA diameter for predicting rupture risk in patients under observation. It was clear that aneurysm wall stress is elevated well in advance of the time of rupture, allowing sufficient time to repair the aneurysm prior to a catastrophic outcome in almost all patients.18 This study demonstrated for the first time that wall stress can be a useful clinical tool with the potential to replace the method we have used for the past 4 decades. At this point, the strength of the data and the size of the patient cohort already rival or exceed that of the data initially used to determine the clinical use of aneurysm diameter to estimate rupture risk in the 1960s. Despite this, wall stress remains to be validated in a large multicenter cohort using a standardized, broadly applicable technique (currently one such study is underway).19
Predicting rupture and survival in aortic aneurysm
123
Figure 1 Abdominal aortic aneurysm wall stress analysis is performed on a three-dimensional reconstruction from computed tomography or magnetic resonance data and displayed graphically, with the highest stresses shown in red, the lowest in blue, and the remainder on a color gradient. The peak stress is determined at systole, and is displayed in a numerical readout.
Until studies are completed, the gold standard will continue to be based on 5.5-cm diameter, modifying the recommendation for repair based on patient factors, such as gender and current smoking, the procedure risk and the
life expectancy.2,3,5,7,20 One way of incorporating these factors in a standard reference fashion is demonstrated in Table 1, illustrating a table from a current multicenter study.19
M. Fillinger
124 Table 1 Known Risk Factors for Aneurysm Rupture Apart from Diameter and Wall Stress, Demonstrating How They Can Be Displayed for an Actual Patient Risk Factor Rate of growth Gender Family history Pulmonary disease/ COPD Hypertension history Steroids Higher blood pressure Today’s blood pressure Smoking
Your Information mm/year Female None On meds
Risk Relative to Average NA Higher than average Lower than average Higher than average
Yes Higher than average Not taking Average steroids 180/72 mm Hg Higher than average 168/70 mm Hg Higher than average Current (within Higher than average year)
Abbreviations: COPD, chronic obstructive pulmonary disease; NA, not applicable. In this case, the modifiable risk factors are displayed along with other factors. Strategy modified from Brewster DC, Cronenwett JL, Hallett JW, Jr, Johnston KW, Krupski WC, Matsumura JS. Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 37:1106-1117, 2003.
Operative Risk Operative mortality is the most widely reported of all aneurysmrelated variables. A literature review by Blankensteijn et al21 found that population-based studies reported mortality rates as high as 8% (prospective), and as a whole are significantly higher than single-center reports averaging 3.8%. A review by Hallin et al22 found a weighted operative mortality for elective open AAA repair of 5%, which is consistent with the UK SAT of 5.6%,23 US hospital discharge data (5.6% for a review of 360,000 repairs),24 and the Canadian Aneurysm Study (4.8%).25 The large volume of data regarding open AAA repair provides perhaps the best opportunity for creating a prediction rule regarding operative mortality. A meta-analysis by Steyerberg et al26 found seven independent risk factors for operative mortality (in order: creatinine ⬎1.8 mg/dL, congestive heart failure (CHF), electrocardiogram (ECG) ischemia, pulmonary dysfunction, older age, and female gender). They created a prediction rule, taking into account a surgeon’s specific operative mortality. Unfortunately, this model has not been validated in other trials, including the UK SAT, where the Steyerberg prediction rule did not work well. This may have been due, in part, to incomplete data on CHF, but other studies have also found slightly different variable sets as independent risk factors. The Glasgow Aneurysm Score is a more recent prediction algorithm for operative mortality that also shows promise, but also has an accuracy rate in the 75% range, which means it will fail a substantial number of patients.27-29 Multiple scoring systems have been developed,
but none are better than the 75% to 80% range.27 Consistent themes in the studies by Steyerberg, Hallin et al, the Canadian Aneurysm Study, L’Italien, and UK SAT suggest that cardiac disease (CHF, ECG ischemia), pulmonary disease, and renal insufficiency are all strong predictors of operative mortality, with lesser effects of age and gender if adjusting for other comorbid factors. Coronary bypass within 5 years and negative cardiac stress testing appears to have a protective effect.30 Factors such as suprarenal versus infrarenal aortic cross-clamp for open repair are also important, but will not be discussed here due to the focus on infrarenal AAAs. With such large studies, why is it still so difficult to create prediction rule for operative mortality that can be independently validated? One reason may be institutional variation, as demonstrated in the Dartmouth Atlas of Vascular Healthcare in 1996.24 This analysis demonstrated that surgeon volume is important, with operative mortality rates of 4% when a surgeon performed ⬎10 AAA repairs/year, as compared to a mortality rate of 8% for surgeons performing ⬍4 AAAs per year. Surgical specialty was also found to have an impact, with the lowest mortality rates achieved by vascular surgeons (4.4%) compared to cardiac surgeons (5.4%) and general surgeons (7.3%). Dimick et al31 found that hospital volume also played a role, with 30-day mortality rates of 3% in hospitals performing ⬎35 AAA repairs/year and 5.5% in hospitals performing ⬍35 AAA repairs/year. As there is an association between surgeon volume and hospital volume and specialty, Dimick et al31 used a multivariate analysis to account for this, and found that surgeon volume, hospital volume and surgical specialty were independently associated with operative mortality rates in elective open AAA repair. Much of the impact of the aforementioned studies has been overshadowed by the impact of endovascular abdominal aortic aneurysm repair (EVAR). Some of the most heated debate of the past decade has been whether endovascular AAA repair truly has a lower operative mortality rate than open repair. Initially there were suggestions to this effect, but nonrandomized clinical trials did not always show much difference in operative mortality. As one might expect, the older sicker patients in the nonrandomized trials tended to be in the endovascular groups. Initially, decision analysis models were brought to bear using existing data, including a study by Finlayson et al32 in 1998 attempting to quantitatively compare open repair and EVAR in terms of the aneurysm size that would benefit from repair. This study calculated the size threshold where repair would be appropriate based on Quality Adjusted Life Years, and found that endovascular repair alters size threshold very little for young, healthy patients, but has a significant impact on the appropriate size threshold in older sicker patients.32 Slowly but surely, the controversy regarding these two options appears to be resolving. As progressively larger studies were reported, the mortality rate for EVAR was consistently lower than for open AAA repair, including large statewide and nationwide studies. New York State data reported by Anderson et al33 indicated in-hospital mortality in 2001 was 3.55% for open repair and 1.14% for EVAR (P ⫽ .0018), and in 2002 these rates were
Predicting rupture and survival in aortic aneurysm 4.21% versus 0.8% (P ⬍ .0001), respectively, despite a higher frequency of comorbidities in the EVAR patients. Lee et al34 compared open repair and EVAR in a United States National Inpatient Sample from 2001 for over 7,000 patients. They found that, despite a significantly greater proportion of octogenarians in the EVAR group, EVAR demonstrated benefits in a number of variables, including morbidity (18% v 29%; P ⫽ .0001), mortality (1.3% v 3.8%; P ⫽ .0001), median length of stay (2 v 7 days; P ⫽ .0001), and the rate of discharge to an institutional facility versus home (6% v 14%; P ⫽ .0001). Multivariate analysis indicated only open AAA repair and age older than 80 years were strong independent predictors (P ⫽ .0001 for all) for mortality or discharge to an institutional facility.34 More recently, Dillavou et al35 found similar results from a 5% US national Medicare sample over 3 years, including a significantly lower mortality rate for EVAR (odds ratio for 30-day mortality ⫽ 0.34; 95% confidence interval, .22-.50, P ⬍ .001). Despite these significant differences for population-based data in the “real world,” critics suggested the results were not indicative of the true differences in open repair and EVAR due to nonrandom samples and selection bias. Ultimately, the question was at least partially answered by randomized clinical trials, primarily EVAR 1 and the Dutch Randomized Endovascular Aneurysm Management.36,37 Both of these studies compared open repair and EVAR in a randomized, prospective fashion. The Dutch Randomized Endovascular Aneurysm Management and EVAR 1 trials randomized patients with ⬎5.5-cm AAAs appropriate for open repair between endovascular repair (EVAR) and open repair. Both trials showed lower 30-day mortality for EVAR, but similar all-cause mortality at later time points. The larger of these two studies was EVAR 1, which involved 1,082 patients. By perprotocol analysis in EVAR 1, 30-day mortality for EVAR was 1.6% (8/512) versus 4.6% (23/496) for open repair (0.33, 95% confidence interval: 0.15-0.74, P ⫽ .007).36 There was a higher rate of secondary intervention in the EVAR group, but a persistent reduction in aneurysm-related death at 4 years. As already mentioned, all-cause mortality was similar in the two groups by 2 years in both studies,36,38 which raises the issue of life expectancy in AAA patients with a large number of comorbidities.
Life Expectancy Of the three factors determining who should have an aneurysm repair, life expectancy is perhaps the most difficult to have a “feel” for, but nonetheless critical to determining whether the patient will benefit from repair. One of the difficulties of determining life expectancy is that it is not a simple linear function. A typical 60-year-old surviving AAA repair has a 13-year life expectancy, but a 70-year-old surviving a AAA repair has a 10-year life expectancy, and an 80-yearold has a 6-year life expectancy.39 These values are not as good as the age-matched general population because of the associated comorbidities typical of aneurysm patients. Therein lies the second major problem with determining life expectancy—aneurysm patients have a relatively high inci-
125 dence of coronary artery disease, pulmonary disease (chronic obstructive pulmonary disease), hypertension, renal insufficiency, hyperlipidemia, cerebrovascular disease, and cancer. One extensive review of 32 articles by Norman et al40 found that mean 5-year survival after AAA repair was 70%, compared to 80% for the general population. The key, of course, is how to incorporate the patient’s comorbidities into some reasonable predictor of life expectancy. Age, gender, cardiac disease, chronic obstructive pulmonary disease, renal insufficiency, and smoking have all been associated with predicting late mortality after aneurysm repair.8,9,41,42 Putting this data into a quantitative, unbiased, prediction rule that results in a straightforward recommendation about longevity and benefit from repair is far from simple, however. One such attempt is the Glasgow Aneurysm Score, which has been shown to detect significant differences in short-term and long-term mortality after both open and endovascular aneurysm repair.27-29 Unfortunately, even though the utility of this scale can be replicated, the best accuracy of the Glasgow Aneurysm Score is on the order of 75% to 80%, and the sensitivity for some key outcomes is as low as 55%.27-29 Thus, the best quantitative measures available still fall far short of what is needed. As with essentially all of the “prediction rules” for aneurysm patients, the Glasgow Aneurysm Score works fairly well for large populations, but falls short when trying to make a decision for an individual. The difficulty in making decisions for individual patients with significant comorbidities is well demonstrated by the EVAR 2 study.15 In EVAR 2, patients with AAAs ⬎5.5 cm were randomized to EVAR or no intervention. The patients were all considered “unfit for open repair,” with substantial comorbidities accounting for an all-cause mortality rate at 4 years of 64%. A uniform strategy for all patients in this category could only be deemed a failure, as neither group performed according to expectations: nearly half of all deaths in the group assigned to EVAR were from rupture prior to elective repair. Operative mortality was 9% for the EVAR-assigned group, which one could ascribe to the “unfit” status of the patients, but 47 patients in the “no-intervention group” crossed over to repair with only a 2% operative mortality. Thus, it appears that the focus needs to be on determining which of these patients will rupture early, and how to get patients into a condition that will allow repair with a low operative mortality rate. We also need a better way to identify the patients with “one foot in the grave and another on a banana peel,” as one author put it, meaning those that cannot be rehabilitated enough to withstand surgery and need to be left alone. This is not simply due to operative mortality rates, but also recovery times. The US National Inpatient Sample and the Medicare data all suggest that 6% to 9% of EVAR patients will not be discharged to home, and a striking 14% to 25% of open repair patients will not be discharged to home.34,35 No patient wants to spend 6 months of a 2-year life expectancy in a rehabilitation facility, or be put into a nursing home for the remainder of his/her life. More work needs to focus on where the patients are living and how they are doing 1 year after surgery rather than simply whether they are alive or not.
M. Fillinger
126
Putting the Factors Together: Simple, Right? So how do we “operationalize” the concept that intervention for an AAA is appropriate when the cumulative risk of rupture exceeds the risk of repair, within the context of patient life expectancy? Of course, this is often not a simple matter, but it is not always difficult, either. For a young, healthy patient with a large aneurysm, the recommendation for intervention is a relatively easy decision. In healthy patients with aneurysms ⬍5.5 cm in diameter, and in high surgical risk patients with large aneurysms, however, the decision is not as simple. Recently, large clinical trials have demonstrated relative safety for observation of AAAs with a maximum diameter ⬍5.5 cm.8,9,12,43 In an effort to prevent rupture, however, these studies required frequent observation including ultrasound or CT scan every 6 months, with surgical intervention for symptoms, rapid expansion, or growth to 5.5 cm. This resulted in a surgical intervention rate of ⬎60% in the “observation” group within several years for both of the major trials. Even with a high rate of intervention in a patient population willing to have frequent and reliable surveillance, the rupture rate may still be ⬎2% per year in some patient populations.2,12 Although observation is appealing in older, high-risk patients, ⬎50% of patients with aneurysms ⬎5.5 cm will rupture when surgery is deferred due to high operative risk,13 many within the first year of observation.13,14 These issues illustrate the importance of the ability to predict AAA rupture risk. It appears that finite element analysis of maximum aneurysm wall stress will be a better predictor than diameter, but this has yet to be applied to large patient cohorts in a multicenter trial using a standardized broadly-applicable technique (currently one such study is underway). Until we have something better, the gold standard will continue to be 5.5-cm diameter, modifying the recommendation for repair based on factors such as gender, aneurysm anatomy, and current smoking, the procedure risk and the patient’s life expectancy.2,3,5,7 Despite all of this, current data suggest that aneurysm diameter as a criteria will fail 10% to 25% of patients.2,4,11,13-15,17,18 Once we have determined that a patient should be considered for repair, how does procedure risk influence the decision? Studies such as EVAR 1 suggest that in the majority of patients, operative mortality significantly favors EVAR. Despite the issues of durability, the majority of secondary procedures after EVAR are endovascular. A decision-analysis study by Schermerhorn et al44 used EUROSTAR data to calculate Quality Adjusted Life Expectancy found that the relative benefits of EVAR versus open repair were most dependent on operative mortality rate. Other factors were also potentially important, but the difference between open repair and EVAR was small across the plausible range of most of these variables. This suggests that the key factor is the operative mortality rate, which is statistically in favor of EVAR based on randomized clinical trials. Thus, with modern-generation endografts, the best short-term and intermediateterm outcomes will be generated by EVAR in most patients. By the same token, however, mathematical models also demonstrate that EVAR should not substantially alter the diame-
ter threshold in the majority of patients.32 We also do not yet know what the long-term results will be, and look forward to the ongoing randomized trials. Given the mortality rates in these elderly patients, however, endograft problems at late follow-up will need to be significant in order to offset the early mortality advantage of EVAR. Whether we determine the patient is best served by open or endovascular aneurysm repair, a number of algorithms exist to estimate perioperative and long-term survival, but none of them are more accurate than 75% to 80% of patients.23,26-29 This once again brings us back to the role of life expectancy. EVAR 2 demonstrates that we cannot assume benefit just because we have a relatively noninvasive procedure and a patient with a large aneurysm. EVAR 2 also suggests, however, that if one focuses efforts on improving a patient’s overall health, their risk/benefit for an operative procedure can change. Other studies evaluating “high-risk” patients demonstrate that we cannot simply ignore patients in this category, as many will have quite acceptable operative mortality rates.44-46 The sometimes conflicting message in all of these studies demonstrates that the definition of “high risk” is rather nebulous at present, and that no study has come up with a clear algorithm that adequately stratifies life expectancy or operative risk in an individual patient. Moreover, we need to focus on quality of life, discharge home versus an institutional facility, and other factors that are critical to these elderly patients. Clearly we need an accurate, quantitative, and unbiased method of incorporating information on rupture risk, operative risk, and life expectancy into a single metric regarding the risk-to-benefit ratio for observation for repair. This issue has received a great deal more attention of late, and we will likely have substantially better tools within the next decade. Nothing is likely to completely substitute for experience, however. Somehow, experienced surgeons are able to sift through a massive amount of information and properly select patients who are appropriate for surgery, with quite reasonable perioperative and long-term mortality rates.45-47 Interestingly, they seem to do this better than mathematical prediction rules or trials enforcing a single treatment strategy on a seemingly homogeneous cohort of patients. Until better tools come along, we all need to pay close attention to those with the most experience making the crucial patient selection decision, learn how they estimate risk based on history, physical exam, “the eyeball test” and other data, and internalize our own results in the context of the literature.
References 1. Szilagyi DE, Smith RF, DeRusso FJ, Elliott JP, Sherrin FW: Contribution of abdominal aortic aneurysmectomy to prolongation of life. Ann Surg 164:678-699, 1966 2. Brown LC, Powell JT: Risk factors for aneurysm rupture in patients kept under ultrasound surveillance. UK Small Aneurysm Trial Participants. Ann Surg 230:289-296; discussion 296-287, 1999 3. Cronenwett JL: Variables that affect the expansion rate and rupture of abdominal aortic aneurysms. Ann N Y Acad Sci 800:56-67, 1996 4. Cronenwett JL, Murphy TF, Zelenock GB, et al: Actuarial analysis of variables associated with rupture of small abdominal aortic aneurysms. Surgery 98:472-483, 1985 5. Fillinger MF, Racusin J, Baker RK, et al: Anatomic characteristics of
Predicting rupture and survival in aortic aneurysm
6. 7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24. 25.
26.
ruptured abdominal aortic aneurysm on conventional CT scans: implications for rupture risk. J Vasc Surg 39:1243-1252, 2004 Strachan DP: Predictors of death from aortic aneurysm among middleaged men: the Whitehall study. Br J Surg 78:401-404, 1991 Smoking, lung function and the prognosis of abdominal aortic aneurysm. The UK Small Aneurysm Trial Participants. Eur J Vasc Endovasc Surg 19:636-642, 2000 Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. The UK Small Aneurysm Trial Participants. Lancet 352(9141): 1649-1655, 1998 Lederle FA, Wilson SE, Johnson GR, et al: Immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 346:1437-1444, 2002 Darling RC, Messina CR, Brewster DC, Ottinger LW: Autopsy study of unoperated abdominal aortic aneurysms. The case for early resection. Circulation 56:II161-III164, 1977 Nicholls SC, Gardner JB, Meissner MH, Johansen HK: Rupture in small abdominal aortic aneurysms. J Vasc Surg 28:884-888, 1998 Brown PM, Pattenden R, Vernooy C, Zelt DT, Gutelius JR: Selective management of abdominal aortic aneurysms in a prospective measurement program. J Vasc Surg 23:213-220; discussion 221-212, 1996 Conway KP, Byrne J, Townsend M, Lane IF: Prognosis of patients turned down for conventional abdominal aortic aneurysm repair in the endovascular and sonographic era: Szilagyi revisited? J Vasc Surg 33: 752-757, 2001 Lederle FA, Johnson GR, Wilson SE, et al: Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA 287:2968-2972, 2002 Endovascular aneurysm repair and outcome in patients unfit for open repair of abdominal aortic aneurysm (EVAR trial 2): Randomised controlled trial. Lancet 365(9478):2187-2192, 2005 Ouriel K, Green RM, Donayre C, Shortell CK, Elliott J, DeWeese JA: An evaluation of new methods of expressing aortic aneurysm size: relationship to rupture. J Vasc Surg 15:12-18; discussion 19-20, 1992 Fillinger MF, Raghavan ML, Marra SP, Cronenwett JL, Kennedy FE: In vivo analysis of mechanical wall stress and abdominal aortic aneurysm rupture risk. J Vasc Surg 36:589-597, 2002 Fillinger MF, Marra SP, Raghavan ML, Kennedy FE: Prediction of rupture risk in abdominal aortic aneurysm during observation: wall stress versus diameter. J Vasc Surg 37:724-732, 2003 Fillinger M: Aortic aneurysm wall stress analysis: a description and invitation for participation in a multicenter evaluation. Endovascular Today 22:44-48, 2003. Brewster DC, Cronenwett JL, Hallett JW Jr, Johnston KW, Krupski WC, Matsumura JS: Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 37:1106-1117, 2003 Blankensteijn JD, Lindenburg FP, Van der Graaf Y, Eikelboom BC: Influence of study design on reported mortality and morbidity rates after abdominal aortic aneurysm repair. Br J Surg 85:1624-1630, 1998 Hallin A, Bergqvist D, Holmberg L: Literature review of surgical management of abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 22:197-204, 2001 Brady AR, Fowkes FG, Greenhalgh RM, Powell JT, Ruckley CV, Thompson SG: Risk factors for postoperative death following elective surgical repair of abdominal aortic aneurysm: results from the UK Small Aneurysm Trial. On behalf of the UK Small Aneurysm Trial participants. Br J Surg 87:742-749, 2000 Cronenwett JL, Birkmeyer JD: The Dartmouth Atlas of Vascular Healthcare. Chicago, AHA Press, 2000 Johnston KW: Multicenter prospective study of nonruptured abdominal aortic aneurysm. Part II. Variables predicting morbidity and mortality. J Vasc Surg 9:437-447, 1989 Steyerberg EW, Kievit J, de Mol Van Otterloo JC, van Bockel JH, Eijkemans MJ, Habbema JD. Perioperative mortality of elective abdominal aortic aneurysm surgery. A clinical prediction rule based on literature and individual patient data. Arch Intern Med 155:1998-2004, 1995
127 27. Nesi F, Leo E, Biancari F, et al: Preoperative risk stratification in patients undergoing elective infrarenal aortic aneurysm surgery: evaluation of five risk scoring methods. Eur J Vasc Endovasc Surg 28:52-58, 2004 28. Biancari F, Leo E, Ylonen K, Vaarala MH, Rainio P, Juvonen T: Value of the Glasgow Aneurysm Score in predicting the immediate and longterm outcome after elective open repair of infrarenal abdominal aortic aneurysm. Br J Surg 90:838-844, 2003 29. Biancari F, Hobo R, Juvonen T: Glasgow Aneurysm Score predicts survival after endovascular stenting of abdominal aortic aneurysm in patients from the EUROSTAR registry. Br J Surg 93:191-194, 2006 30. L’Italien GJ, Paul SD, Hendel RC, et al: Development and validation of a Bayesian model for perioperative cardiac risk assessment in a cohort of 1,081 vascular surgical candidates. J Am Coll Cardiol 27:779-786, 1996 31. Dimick JB, Cowan JA Jr, Stanley JC, Henke PK, Pronovost PJ, Upchurch GR Jr: Surgeon specialty and provider volumes are related to outcome of intact abdominal aortic aneurysm repair in the United States. J Vasc Surg 38:739-744, 2003 32. Finlayson SR, Birkmeyer JD, Fillinger MF, Cronenwett JL: Should endovascular surgery lower the threshold for repair of abdominal aortic aneurysms? J Vasc Surg 29:973-985, 1999 33. Anderson PL, Arons RR, Moskowitz AJ, et al: A statewide experience with endovascular abdominal aortic aneurysm repair: rapid diffusion with excellent early results. J Vasc Surg 39:10-19, 2004 34. Lee WA, Carter JW, Upchurch G, Seeger JM, Huber TS: Perioperative outcomes after open and endovascular repair of intact abdominal aortic aneurysms in the United States during 2001. J Vasc Surg 39:491-496, 2004 35. Dillavou ED, Muluk SC, Makaroun MS: Improving aneurysm-related outcomes: nationwide benefits of endovascular repair. J Vasc Surg 43: 446-451; discussion 451-442, 2006 36. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 365(9478):2179-2186, 2005 37. Prinssen M, Verhoeven EL, Buth J, et al: A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 351:1607-1618, 2004 38. Blankensteijn JD, de Jong SE, Prinssen M, et al: Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N Engl J Med 352:2398-2405, 2005 39. Schermerhorn ML, Cronenwett JL: Abdominal aortic and iliac aneurysms, in Rutherford RB, (ed): Vascular Surgery. Philadelphia, Elsevier Saunders, 2005, pp 1408-1452 40. Norman PE, Semmens JB, Lawrence-Brown MM: Long-term relative survival following surgery for abdominal aortic aneurysm: a review. Cardiovasc Surg 9:219-224, 2001 41. Newman AB, Arnold AM, Burke GL, O’Leary DH, Manolio TA: Cardiovascular disease and mortality in older adults with small abdominal aortic aneurysms detected by ultrasonography: the cardiovascular health study. Ann Intern Med 134:182-190, 2001 42. Singh K, Bonaa KH, Jacobsen BK, Bjork L, Solberg S: Prevalence of and risk factors for abdominal aortic aneurysms in a population-based study: the Tromso Study. Am J Epidemiol 154:236-244, 2001 43. Long-term outcomes of immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 346:1445-1452, 2002 44. Schermerhom ML, Finlayson SR, Fillinger MF, Buth J, van Marrewijk C, Cronenwett JL: Life expectancy after endovascular versus open abdominal aortic aneurysm repair: results of a decision analysis model on the basis of data from EUROSTAR. J Vasc Surg 36(6):1112-1120, 2002 45. Bush RL, Johnson ML, Collins TC, et al: Open versus endovascular abdominal aortic aneurysm repair in VA hospitals. J Am Coll Surg 202:577-587, 2006 46. Sicard GA, Zwolak RM, Sidawy AN, White RA, Siami FS: Endovascular abdominal aortic aneurysm repair: long-term outcome measures in patients at high-risk for open surgery. J Vasc Surg 44:229-236, 2006 47. Zarins CK, Crabtree T, Bloch DA, Arko FR, Ouriel K, White RA: Endovascular aneurysm repair at 5 years: does aneurysm diameter predict outcome? J Vasc Surg 44:920-929; discussion 929-931, 2006
D
URING TRAINING, A surgical resident or vascular fellow naturally wants to focus on the technical aspects of a procedure. This “truism” certainly applies to aortic aneurysm repairs, perhaps more than any other procedure in vascular surgery. As any experienced vascular surgeon will attest, however, it is much harder to learn when to operate than it is to learn how to operate. Experienced surgeons understand that outcomes are closely linked to patient selection, in many cases learning this the “hard way” early in their careers. But why is patient selection so difficult, especially with regard to aneurysm repair? In a patient with an aortic aneurysm, the decision to operate is based on three primary factors: risk of aneurysm rupture, risk of aneurysm repair, and patient’s life expectancy. There is a fourth factor, the patient’s personal preference, but this can be heavily influenced by the information they assimilate with regard to the other three primary factors. The patient’s personal preference is also the easiest to deter-
Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, NH. Address reprint requests to Mark F. Fillinger, Section of Vascular Surgery, Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03750. E-mail: [email protected]
0895-7967/07/$-see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1053/j.semvascsurg.2007.04.001
mine: the patient will give that to the surgeon once they have discussed the three “primary” factors. Deciding who we should operate on should be simple then, shouldn’t it? Can’t we simply calculate the aneurysm rupture risk, the procedure risk and the patient’s life expectancy based on the literature and actuarial tables? As Shakespeare wrote, “Ay, there’s the rub . . .,” for understanding any of these seemingly simple factors can be quite difficult in any given patient. Decades of work have provided precious little in our understanding of these three issues, much less the interaction of the three factors in a given patient. Of course, this problem will not be solved within this article, but at least the issues can be framed in a potentially helpful manner.
Rupture Risk For the past five decades, the primary determinant of rupture risk has been maximum aneurysm diameter, based on the work of Szilagyi et al in the 1960s.1 These authors compared the outcomes of small-diameter and large-diameter aneurysms, and found that patients with larger aneurysms (⬎6 cm) were much more likely to die of a ruptured aneurysm. In that era, the diameter had to be determined by physical exam 121
122 and abdominal x-ray, which are both now known to overestimate the actual diameter, such that the actual threshold for aneurysm size was approximately 5 cm. The 5-cm threshold was adopted when ultrasonography became widely available, and has been used ever since, with minor modification throughout the decades based on clinical evidence that variables such as hypertension, pulmonary disease, gender, and smoking modified aneurysm rupture risk somewhat.2-7 Only recently has there been even a minor change in the diameter threshold for proposing surgery, because of randomized clinical trials, such as the United Kingdom Small Aneurysm Trial (UK SAT) and the Aneurysm Detection and Management study, suggesting the threshold for the typical patient should be 5.5 cm.8,9 Even with these results, some question whether women, who had a three- to fourfold risk of rupture under observation in UK SAT,2 should be repaired at 5 cm. Data from our own institution including 122 ruptured abdominal aortic aneurysms (AAA) indicate that the average size of ruptured aneurysms in women is 5 mm smaller than in men,5 lending further evidence that the size threshold should be slightly lower in women. Given the relative stability of the diameter threshold for considering surgery, one might reasonably assume that maximum aortic aneurysm diameter is an extremely reliable determinant for patient selection. Unfortunately, this is not the case. Although maximum aneurysm diameter works relatively well in general (as evidenced by summary data involving large numbers of patients), all clinicians know that this “metric” often fails in individual patients. Some aneurysms rupture at an unusually small size.4,10,11 In some series, 10% to 24% of the ruptured aneurysms were ⱕ5 cm in maximum diameter.2,11 Ultimately, large studies were conducted to determine whether it is safe to observe AAAs with a maximum diameter ⬍5.5 cm.8,9 In an effort to prevent rupture, these studies used frequent observation including ultrasonography or computed tomography (CT) every 6 months, with surgical intervention for symptoms, expansion, or growth to 5.5 cm. This resulted in a surgical intervention rate of approximately 60% in the “observation” group within 4 years, and ⬎75% of the observation patients had surgery within 6 years. Even with a high rate of intervention in a patient population willing to have frequent and reliable surveillance, the rupture rate can still be up to 2% per year, depending on gender and aneurysm size.2,9,12 Moreover, while diameter-threshold issues receive widespread study in patients with smaller aneurysms, there has not been as much study of larger aneurysms. In older and sicker patients, however, decision-making can be difficult, and repair might not be recommended until the aneurysm reaches a larger size. In patients with aneurysms ⬎5.5 cm, however, ⬎50% will rupture when surgery is deferred because of high operative risk.13 In this high-risk group, the median time to rupture was only 19 months for patients with 5.5- to 5.9-cm AAAs, and only 9 months for patients with AAAs ⬎7 cm.13 Even in patients under close surveillance, the aneurysm rupture rate is substantial when ⬎5.5 cm, and rises exponentially with size.14,15 While this might lead one to consider repair for nearly all patients with large aneurysms,
M. Fillinger studies such as EVAR 2 make it clear that there is no simple strategy.15 Lastly, even the method of determining maximum aneurysm diameter has methodological problems. Most aneurysms have an elliptical cross-section on axial CT imaging, and large studies have suggested that rupture risk is more closely associated with the minor axis.5,16 This makes sense with regard to three-dimensional reconstructions of CT imaging data, which demonstrates that the elliptical cross-section in most aneurysms is due to tortuosity rather than a truly asymmetrical shape.5 For those aneurysms with a truly asymmetrical shape, the larger “diameter” in the elliptical crosssection may be a more appropriate estimate of rupture risk.5 Despite these issues, however, many studies fail to precisely report how maximum aneurysm diameter was determined, and many centers still use the “major axis” diameter to estimate rupture risk for all aortic aneurysms. As an example of how influential such an issue can be, the UK SAT used the maximal antero-posterior (AP) diameter of the aneurysm by ultrasound, which will overestimate the “true” size of an aneurysm that bows anteriorly (which is not uncommon). Thus, even though maximum aneurysm diameter remains the gold standard for estimating rupture risk, it is far from ideal. We have demonstrated that finite element analysis of AAA wall stress using three-dimensional CT reconstructions is better than diameter for differentiating AAAs near the time of rupture,17 and that wall stress is superior to AAA diameter for predicting rupture risk in patients under observation.18 In 2002, in the first report using finite element analysis on a large cohort of AAA patients with and without rupture, our group reported that finite element analysis of AAA wall stress using three-dimensional CT reconstructions is better than diameter for differentiating AAAs near the time of rupture (Fig 1). Moreover, we found that calculated indices previously suggested to better predict rupture risk (eg, ratio of maximum AAA diameter to normal infrarenal diameter) were not helpful.17 Lack of utility for calculated indices was later confirmed in much larger cohort of 259 patients, 122 of them with documented rupture.5 In 2003, our group reported on the first large series of patients with AAAs under observation, looking at the natural history of aneurysms that had elective repair deferred or were felt to be at low risk of rupture. In that study, we found that wall stress is superior to AAA diameter for predicting rupture risk in patients under observation. It was clear that aneurysm wall stress is elevated well in advance of the time of rupture, allowing sufficient time to repair the aneurysm prior to a catastrophic outcome in almost all patients.18 This study demonstrated for the first time that wall stress can be a useful clinical tool with the potential to replace the method we have used for the past 4 decades. At this point, the strength of the data and the size of the patient cohort already rival or exceed that of the data initially used to determine the clinical use of aneurysm diameter to estimate rupture risk in the 1960s. Despite this, wall stress remains to be validated in a large multicenter cohort using a standardized, broadly applicable technique (currently one such study is underway).19
Predicting rupture and survival in aortic aneurysm
123
Figure 1 Abdominal aortic aneurysm wall stress analysis is performed on a three-dimensional reconstruction from computed tomography or magnetic resonance data and displayed graphically, with the highest stresses shown in red, the lowest in blue, and the remainder on a color gradient. The peak stress is determined at systole, and is displayed in a numerical readout.
Until studies are completed, the gold standard will continue to be based on 5.5-cm diameter, modifying the recommendation for repair based on patient factors, such as gender and current smoking, the procedure risk and the
life expectancy.2,3,5,7,20 One way of incorporating these factors in a standard reference fashion is demonstrated in Table 1, illustrating a table from a current multicenter study.19
M. Fillinger
124 Table 1 Known Risk Factors for Aneurysm Rupture Apart from Diameter and Wall Stress, Demonstrating How They Can Be Displayed for an Actual Patient Risk Factor Rate of growth Gender Family history Pulmonary disease/ COPD Hypertension history Steroids Higher blood pressure Today’s blood pressure Smoking
Your Information mm/year Female None On meds
Risk Relative to Average NA Higher than average Lower than average Higher than average
Yes Higher than average Not taking Average steroids 180/72 mm Hg Higher than average 168/70 mm Hg Higher than average Current (within Higher than average year)
Abbreviations: COPD, chronic obstructive pulmonary disease; NA, not applicable. In this case, the modifiable risk factors are displayed along with other factors. Strategy modified from Brewster DC, Cronenwett JL, Hallett JW, Jr, Johnston KW, Krupski WC, Matsumura JS. Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 37:1106-1117, 2003.
Operative Risk Operative mortality is the most widely reported of all aneurysmrelated variables. A literature review by Blankensteijn et al21 found that population-based studies reported mortality rates as high as 8% (prospective), and as a whole are significantly higher than single-center reports averaging 3.8%. A review by Hallin et al22 found a weighted operative mortality for elective open AAA repair of 5%, which is consistent with the UK SAT of 5.6%,23 US hospital discharge data (5.6% for a review of 360,000 repairs),24 and the Canadian Aneurysm Study (4.8%).25 The large volume of data regarding open AAA repair provides perhaps the best opportunity for creating a prediction rule regarding operative mortality. A meta-analysis by Steyerberg et al26 found seven independent risk factors for operative mortality (in order: creatinine ⬎1.8 mg/dL, congestive heart failure (CHF), electrocardiogram (ECG) ischemia, pulmonary dysfunction, older age, and female gender). They created a prediction rule, taking into account a surgeon’s specific operative mortality. Unfortunately, this model has not been validated in other trials, including the UK SAT, where the Steyerberg prediction rule did not work well. This may have been due, in part, to incomplete data on CHF, but other studies have also found slightly different variable sets as independent risk factors. The Glasgow Aneurysm Score is a more recent prediction algorithm for operative mortality that also shows promise, but also has an accuracy rate in the 75% range, which means it will fail a substantial number of patients.27-29 Multiple scoring systems have been developed,
but none are better than the 75% to 80% range.27 Consistent themes in the studies by Steyerberg, Hallin et al, the Canadian Aneurysm Study, L’Italien, and UK SAT suggest that cardiac disease (CHF, ECG ischemia), pulmonary disease, and renal insufficiency are all strong predictors of operative mortality, with lesser effects of age and gender if adjusting for other comorbid factors. Coronary bypass within 5 years and negative cardiac stress testing appears to have a protective effect.30 Factors such as suprarenal versus infrarenal aortic cross-clamp for open repair are also important, but will not be discussed here due to the focus on infrarenal AAAs. With such large studies, why is it still so difficult to create prediction rule for operative mortality that can be independently validated? One reason may be institutional variation, as demonstrated in the Dartmouth Atlas of Vascular Healthcare in 1996.24 This analysis demonstrated that surgeon volume is important, with operative mortality rates of 4% when a surgeon performed ⬎10 AAA repairs/year, as compared to a mortality rate of 8% for surgeons performing ⬍4 AAAs per year. Surgical specialty was also found to have an impact, with the lowest mortality rates achieved by vascular surgeons (4.4%) compared to cardiac surgeons (5.4%) and general surgeons (7.3%). Dimick et al31 found that hospital volume also played a role, with 30-day mortality rates of 3% in hospitals performing ⬎35 AAA repairs/year and 5.5% in hospitals performing ⬍35 AAA repairs/year. As there is an association between surgeon volume and hospital volume and specialty, Dimick et al31 used a multivariate analysis to account for this, and found that surgeon volume, hospital volume and surgical specialty were independently associated with operative mortality rates in elective open AAA repair. Much of the impact of the aforementioned studies has been overshadowed by the impact of endovascular abdominal aortic aneurysm repair (EVAR). Some of the most heated debate of the past decade has been whether endovascular AAA repair truly has a lower operative mortality rate than open repair. Initially there were suggestions to this effect, but nonrandomized clinical trials did not always show much difference in operative mortality. As one might expect, the older sicker patients in the nonrandomized trials tended to be in the endovascular groups. Initially, decision analysis models were brought to bear using existing data, including a study by Finlayson et al32 in 1998 attempting to quantitatively compare open repair and EVAR in terms of the aneurysm size that would benefit from repair. This study calculated the size threshold where repair would be appropriate based on Quality Adjusted Life Years, and found that endovascular repair alters size threshold very little for young, healthy patients, but has a significant impact on the appropriate size threshold in older sicker patients.32 Slowly but surely, the controversy regarding these two options appears to be resolving. As progressively larger studies were reported, the mortality rate for EVAR was consistently lower than for open AAA repair, including large statewide and nationwide studies. New York State data reported by Anderson et al33 indicated in-hospital mortality in 2001 was 3.55% for open repair and 1.14% for EVAR (P ⫽ .0018), and in 2002 these rates were
Predicting rupture and survival in aortic aneurysm 4.21% versus 0.8% (P ⬍ .0001), respectively, despite a higher frequency of comorbidities in the EVAR patients. Lee et al34 compared open repair and EVAR in a United States National Inpatient Sample from 2001 for over 7,000 patients. They found that, despite a significantly greater proportion of octogenarians in the EVAR group, EVAR demonstrated benefits in a number of variables, including morbidity (18% v 29%; P ⫽ .0001), mortality (1.3% v 3.8%; P ⫽ .0001), median length of stay (2 v 7 days; P ⫽ .0001), and the rate of discharge to an institutional facility versus home (6% v 14%; P ⫽ .0001). Multivariate analysis indicated only open AAA repair and age older than 80 years were strong independent predictors (P ⫽ .0001 for all) for mortality or discharge to an institutional facility.34 More recently, Dillavou et al35 found similar results from a 5% US national Medicare sample over 3 years, including a significantly lower mortality rate for EVAR (odds ratio for 30-day mortality ⫽ 0.34; 95% confidence interval, .22-.50, P ⬍ .001). Despite these significant differences for population-based data in the “real world,” critics suggested the results were not indicative of the true differences in open repair and EVAR due to nonrandom samples and selection bias. Ultimately, the question was at least partially answered by randomized clinical trials, primarily EVAR 1 and the Dutch Randomized Endovascular Aneurysm Management.36,37 Both of these studies compared open repair and EVAR in a randomized, prospective fashion. The Dutch Randomized Endovascular Aneurysm Management and EVAR 1 trials randomized patients with ⬎5.5-cm AAAs appropriate for open repair between endovascular repair (EVAR) and open repair. Both trials showed lower 30-day mortality for EVAR, but similar all-cause mortality at later time points. The larger of these two studies was EVAR 1, which involved 1,082 patients. By perprotocol analysis in EVAR 1, 30-day mortality for EVAR was 1.6% (8/512) versus 4.6% (23/496) for open repair (0.33, 95% confidence interval: 0.15-0.74, P ⫽ .007).36 There was a higher rate of secondary intervention in the EVAR group, but a persistent reduction in aneurysm-related death at 4 years. As already mentioned, all-cause mortality was similar in the two groups by 2 years in both studies,36,38 which raises the issue of life expectancy in AAA patients with a large number of comorbidities.
Life Expectancy Of the three factors determining who should have an aneurysm repair, life expectancy is perhaps the most difficult to have a “feel” for, but nonetheless critical to determining whether the patient will benefit from repair. One of the difficulties of determining life expectancy is that it is not a simple linear function. A typical 60-year-old surviving AAA repair has a 13-year life expectancy, but a 70-year-old surviving a AAA repair has a 10-year life expectancy, and an 80-yearold has a 6-year life expectancy.39 These values are not as good as the age-matched general population because of the associated comorbidities typical of aneurysm patients. Therein lies the second major problem with determining life expectancy—aneurysm patients have a relatively high inci-
125 dence of coronary artery disease, pulmonary disease (chronic obstructive pulmonary disease), hypertension, renal insufficiency, hyperlipidemia, cerebrovascular disease, and cancer. One extensive review of 32 articles by Norman et al40 found that mean 5-year survival after AAA repair was 70%, compared to 80% for the general population. The key, of course, is how to incorporate the patient’s comorbidities into some reasonable predictor of life expectancy. Age, gender, cardiac disease, chronic obstructive pulmonary disease, renal insufficiency, and smoking have all been associated with predicting late mortality after aneurysm repair.8,9,41,42 Putting this data into a quantitative, unbiased, prediction rule that results in a straightforward recommendation about longevity and benefit from repair is far from simple, however. One such attempt is the Glasgow Aneurysm Score, which has been shown to detect significant differences in short-term and long-term mortality after both open and endovascular aneurysm repair.27-29 Unfortunately, even though the utility of this scale can be replicated, the best accuracy of the Glasgow Aneurysm Score is on the order of 75% to 80%, and the sensitivity for some key outcomes is as low as 55%.27-29 Thus, the best quantitative measures available still fall far short of what is needed. As with essentially all of the “prediction rules” for aneurysm patients, the Glasgow Aneurysm Score works fairly well for large populations, but falls short when trying to make a decision for an individual. The difficulty in making decisions for individual patients with significant comorbidities is well demonstrated by the EVAR 2 study.15 In EVAR 2, patients with AAAs ⬎5.5 cm were randomized to EVAR or no intervention. The patients were all considered “unfit for open repair,” with substantial comorbidities accounting for an all-cause mortality rate at 4 years of 64%. A uniform strategy for all patients in this category could only be deemed a failure, as neither group performed according to expectations: nearly half of all deaths in the group assigned to EVAR were from rupture prior to elective repair. Operative mortality was 9% for the EVAR-assigned group, which one could ascribe to the “unfit” status of the patients, but 47 patients in the “no-intervention group” crossed over to repair with only a 2% operative mortality. Thus, it appears that the focus needs to be on determining which of these patients will rupture early, and how to get patients into a condition that will allow repair with a low operative mortality rate. We also need a better way to identify the patients with “one foot in the grave and another on a banana peel,” as one author put it, meaning those that cannot be rehabilitated enough to withstand surgery and need to be left alone. This is not simply due to operative mortality rates, but also recovery times. The US National Inpatient Sample and the Medicare data all suggest that 6% to 9% of EVAR patients will not be discharged to home, and a striking 14% to 25% of open repair patients will not be discharged to home.34,35 No patient wants to spend 6 months of a 2-year life expectancy in a rehabilitation facility, or be put into a nursing home for the remainder of his/her life. More work needs to focus on where the patients are living and how they are doing 1 year after surgery rather than simply whether they are alive or not.
M. Fillinger
126
Putting the Factors Together: Simple, Right? So how do we “operationalize” the concept that intervention for an AAA is appropriate when the cumulative risk of rupture exceeds the risk of repair, within the context of patient life expectancy? Of course, this is often not a simple matter, but it is not always difficult, either. For a young, healthy patient with a large aneurysm, the recommendation for intervention is a relatively easy decision. In healthy patients with aneurysms ⬍5.5 cm in diameter, and in high surgical risk patients with large aneurysms, however, the decision is not as simple. Recently, large clinical trials have demonstrated relative safety for observation of AAAs with a maximum diameter ⬍5.5 cm.8,9,12,43 In an effort to prevent rupture, however, these studies required frequent observation including ultrasound or CT scan every 6 months, with surgical intervention for symptoms, rapid expansion, or growth to 5.5 cm. This resulted in a surgical intervention rate of ⬎60% in the “observation” group within several years for both of the major trials. Even with a high rate of intervention in a patient population willing to have frequent and reliable surveillance, the rupture rate may still be ⬎2% per year in some patient populations.2,12 Although observation is appealing in older, high-risk patients, ⬎50% of patients with aneurysms ⬎5.5 cm will rupture when surgery is deferred due to high operative risk,13 many within the first year of observation.13,14 These issues illustrate the importance of the ability to predict AAA rupture risk. It appears that finite element analysis of maximum aneurysm wall stress will be a better predictor than diameter, but this has yet to be applied to large patient cohorts in a multicenter trial using a standardized broadly-applicable technique (currently one such study is underway). Until we have something better, the gold standard will continue to be 5.5-cm diameter, modifying the recommendation for repair based on factors such as gender, aneurysm anatomy, and current smoking, the procedure risk and the patient’s life expectancy.2,3,5,7 Despite all of this, current data suggest that aneurysm diameter as a criteria will fail 10% to 25% of patients.2,4,11,13-15,17,18 Once we have determined that a patient should be considered for repair, how does procedure risk influence the decision? Studies such as EVAR 1 suggest that in the majority of patients, operative mortality significantly favors EVAR. Despite the issues of durability, the majority of secondary procedures after EVAR are endovascular. A decision-analysis study by Schermerhorn et al44 used EUROSTAR data to calculate Quality Adjusted Life Expectancy found that the relative benefits of EVAR versus open repair were most dependent on operative mortality rate. Other factors were also potentially important, but the difference between open repair and EVAR was small across the plausible range of most of these variables. This suggests that the key factor is the operative mortality rate, which is statistically in favor of EVAR based on randomized clinical trials. Thus, with modern-generation endografts, the best short-term and intermediateterm outcomes will be generated by EVAR in most patients. By the same token, however, mathematical models also demonstrate that EVAR should not substantially alter the diame-
ter threshold in the majority of patients.32 We also do not yet know what the long-term results will be, and look forward to the ongoing randomized trials. Given the mortality rates in these elderly patients, however, endograft problems at late follow-up will need to be significant in order to offset the early mortality advantage of EVAR. Whether we determine the patient is best served by open or endovascular aneurysm repair, a number of algorithms exist to estimate perioperative and long-term survival, but none of them are more accurate than 75% to 80% of patients.23,26-29 This once again brings us back to the role of life expectancy. EVAR 2 demonstrates that we cannot assume benefit just because we have a relatively noninvasive procedure and a patient with a large aneurysm. EVAR 2 also suggests, however, that if one focuses efforts on improving a patient’s overall health, their risk/benefit for an operative procedure can change. Other studies evaluating “high-risk” patients demonstrate that we cannot simply ignore patients in this category, as many will have quite acceptable operative mortality rates.44-46 The sometimes conflicting message in all of these studies demonstrates that the definition of “high risk” is rather nebulous at present, and that no study has come up with a clear algorithm that adequately stratifies life expectancy or operative risk in an individual patient. Moreover, we need to focus on quality of life, discharge home versus an institutional facility, and other factors that are critical to these elderly patients. Clearly we need an accurate, quantitative, and unbiased method of incorporating information on rupture risk, operative risk, and life expectancy into a single metric regarding the risk-to-benefit ratio for observation for repair. This issue has received a great deal more attention of late, and we will likely have substantially better tools within the next decade. Nothing is likely to completely substitute for experience, however. Somehow, experienced surgeons are able to sift through a massive amount of information and properly select patients who are appropriate for surgery, with quite reasonable perioperative and long-term mortality rates.45-47 Interestingly, they seem to do this better than mathematical prediction rules or trials enforcing a single treatment strategy on a seemingly homogeneous cohort of patients. Until better tools come along, we all need to pay close attention to those with the most experience making the crucial patient selection decision, learn how they estimate risk based on history, physical exam, “the eyeball test” and other data, and internalize our own results in the context of the literature.
References 1. Szilagyi DE, Smith RF, DeRusso FJ, Elliott JP, Sherrin FW: Contribution of abdominal aortic aneurysmectomy to prolongation of life. Ann Surg 164:678-699, 1966 2. Brown LC, Powell JT: Risk factors for aneurysm rupture in patients kept under ultrasound surveillance. UK Small Aneurysm Trial Participants. Ann Surg 230:289-296; discussion 296-287, 1999 3. Cronenwett JL: Variables that affect the expansion rate and rupture of abdominal aortic aneurysms. Ann N Y Acad Sci 800:56-67, 1996 4. Cronenwett JL, Murphy TF, Zelenock GB, et al: Actuarial analysis of variables associated with rupture of small abdominal aortic aneurysms. Surgery 98:472-483, 1985 5. Fillinger MF, Racusin J, Baker RK, et al: Anatomic characteristics of
Predicting rupture and survival in aortic aneurysm
6. 7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24. 25.
26.
ruptured abdominal aortic aneurysm on conventional CT scans: implications for rupture risk. J Vasc Surg 39:1243-1252, 2004 Strachan DP: Predictors of death from aortic aneurysm among middleaged men: the Whitehall study. Br J Surg 78:401-404, 1991 Smoking, lung function and the prognosis of abdominal aortic aneurysm. The UK Small Aneurysm Trial Participants. Eur J Vasc Endovasc Surg 19:636-642, 2000 Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. The UK Small Aneurysm Trial Participants. Lancet 352(9141): 1649-1655, 1998 Lederle FA, Wilson SE, Johnson GR, et al: Immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 346:1437-1444, 2002 Darling RC, Messina CR, Brewster DC, Ottinger LW: Autopsy study of unoperated abdominal aortic aneurysms. The case for early resection. Circulation 56:II161-III164, 1977 Nicholls SC, Gardner JB, Meissner MH, Johansen HK: Rupture in small abdominal aortic aneurysms. J Vasc Surg 28:884-888, 1998 Brown PM, Pattenden R, Vernooy C, Zelt DT, Gutelius JR: Selective management of abdominal aortic aneurysms in a prospective measurement program. J Vasc Surg 23:213-220; discussion 221-212, 1996 Conway KP, Byrne J, Townsend M, Lane IF: Prognosis of patients turned down for conventional abdominal aortic aneurysm repair in the endovascular and sonographic era: Szilagyi revisited? J Vasc Surg 33: 752-757, 2001 Lederle FA, Johnson GR, Wilson SE, et al: Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA 287:2968-2972, 2002 Endovascular aneurysm repair and outcome in patients unfit for open repair of abdominal aortic aneurysm (EVAR trial 2): Randomised controlled trial. Lancet 365(9478):2187-2192, 2005 Ouriel K, Green RM, Donayre C, Shortell CK, Elliott J, DeWeese JA: An evaluation of new methods of expressing aortic aneurysm size: relationship to rupture. J Vasc Surg 15:12-18; discussion 19-20, 1992 Fillinger MF, Raghavan ML, Marra SP, Cronenwett JL, Kennedy FE: In vivo analysis of mechanical wall stress and abdominal aortic aneurysm rupture risk. J Vasc Surg 36:589-597, 2002 Fillinger MF, Marra SP, Raghavan ML, Kennedy FE: Prediction of rupture risk in abdominal aortic aneurysm during observation: wall stress versus diameter. J Vasc Surg 37:724-732, 2003 Fillinger M: Aortic aneurysm wall stress analysis: a description and invitation for participation in a multicenter evaluation. Endovascular Today 22:44-48, 2003. Brewster DC, Cronenwett JL, Hallett JW Jr, Johnston KW, Krupski WC, Matsumura JS: Guidelines for the treatment of abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 37:1106-1117, 2003 Blankensteijn JD, Lindenburg FP, Van der Graaf Y, Eikelboom BC: Influence of study design on reported mortality and morbidity rates after abdominal aortic aneurysm repair. Br J Surg 85:1624-1630, 1998 Hallin A, Bergqvist D, Holmberg L: Literature review of surgical management of abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 22:197-204, 2001 Brady AR, Fowkes FG, Greenhalgh RM, Powell JT, Ruckley CV, Thompson SG: Risk factors for postoperative death following elective surgical repair of abdominal aortic aneurysm: results from the UK Small Aneurysm Trial. On behalf of the UK Small Aneurysm Trial participants. Br J Surg 87:742-749, 2000 Cronenwett JL, Birkmeyer JD: The Dartmouth Atlas of Vascular Healthcare. Chicago, AHA Press, 2000 Johnston KW: Multicenter prospective study of nonruptured abdominal aortic aneurysm. Part II. Variables predicting morbidity and mortality. J Vasc Surg 9:437-447, 1989 Steyerberg EW, Kievit J, de Mol Van Otterloo JC, van Bockel JH, Eijkemans MJ, Habbema JD. Perioperative mortality of elective abdominal aortic aneurysm surgery. A clinical prediction rule based on literature and individual patient data. Arch Intern Med 155:1998-2004, 1995
127 27. Nesi F, Leo E, Biancari F, et al: Preoperative risk stratification in patients undergoing elective infrarenal aortic aneurysm surgery: evaluation of five risk scoring methods. Eur J Vasc Endovasc Surg 28:52-58, 2004 28. Biancari F, Leo E, Ylonen K, Vaarala MH, Rainio P, Juvonen T: Value of the Glasgow Aneurysm Score in predicting the immediate and longterm outcome after elective open repair of infrarenal abdominal aortic aneurysm. Br J Surg 90:838-844, 2003 29. Biancari F, Hobo R, Juvonen T: Glasgow Aneurysm Score predicts survival after endovascular stenting of abdominal aortic aneurysm in patients from the EUROSTAR registry. Br J Surg 93:191-194, 2006 30. L’Italien GJ, Paul SD, Hendel RC, et al: Development and validation of a Bayesian model for perioperative cardiac risk assessment in a cohort of 1,081 vascular surgical candidates. J Am Coll Cardiol 27:779-786, 1996 31. Dimick JB, Cowan JA Jr, Stanley JC, Henke PK, Pronovost PJ, Upchurch GR Jr: Surgeon specialty and provider volumes are related to outcome of intact abdominal aortic aneurysm repair in the United States. J Vasc Surg 38:739-744, 2003 32. Finlayson SR, Birkmeyer JD, Fillinger MF, Cronenwett JL: Should endovascular surgery lower the threshold for repair of abdominal aortic aneurysms? J Vasc Surg 29:973-985, 1999 33. Anderson PL, Arons RR, Moskowitz AJ, et al: A statewide experience with endovascular abdominal aortic aneurysm repair: rapid diffusion with excellent early results. J Vasc Surg 39:10-19, 2004 34. Lee WA, Carter JW, Upchurch G, Seeger JM, Huber TS: Perioperative outcomes after open and endovascular repair of intact abdominal aortic aneurysms in the United States during 2001. J Vasc Surg 39:491-496, 2004 35. Dillavou ED, Muluk SC, Makaroun MS: Improving aneurysm-related outcomes: nationwide benefits of endovascular repair. J Vasc Surg 43: 446-451; discussion 451-442, 2006 36. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 365(9478):2179-2186, 2005 37. Prinssen M, Verhoeven EL, Buth J, et al: A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 351:1607-1618, 2004 38. Blankensteijn JD, de Jong SE, Prinssen M, et al: Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N Engl J Med 352:2398-2405, 2005 39. Schermerhorn ML, Cronenwett JL: Abdominal aortic and iliac aneurysms, in Rutherford RB, (ed): Vascular Surgery. Philadelphia, Elsevier Saunders, 2005, pp 1408-1452 40. Norman PE, Semmens JB, Lawrence-Brown MM: Long-term relative survival following surgery for abdominal aortic aneurysm: a review. Cardiovasc Surg 9:219-224, 2001 41. Newman AB, Arnold AM, Burke GL, O’Leary DH, Manolio TA: Cardiovascular disease and mortality in older adults with small abdominal aortic aneurysms detected by ultrasonography: the cardiovascular health study. Ann Intern Med 134:182-190, 2001 42. Singh K, Bonaa KH, Jacobsen BK, Bjork L, Solberg S: Prevalence of and risk factors for abdominal aortic aneurysms in a population-based study: the Tromso Study. Am J Epidemiol 154:236-244, 2001 43. Long-term outcomes of immediate repair compared with surveillance of small abdominal aortic aneurysms. N Engl J Med 346:1445-1452, 2002 44. Schermerhom ML, Finlayson SR, Fillinger MF, Buth J, van Marrewijk C, Cronenwett JL: Life expectancy after endovascular versus open abdominal aortic aneurysm repair: results of a decision analysis model on the basis of data from EUROSTAR. J Vasc Surg 36(6):1112-1120, 2002 45. Bush RL, Johnson ML, Collins TC, et al: Open versus endovascular abdominal aortic aneurysm repair in VA hospitals. J Am Coll Surg 202:577-587, 2006 46. Sicard GA, Zwolak RM, Sidawy AN, White RA, Siami FS: Endovascular abdominal aortic aneurysm repair: long-term outcome measures in patients at high-risk for open surgery. J Vasc Surg 44:229-236, 2006 47. Zarins CK, Crabtree T, Bloch DA, Arko FR, Ouriel K, White RA: Endovascular aneurysm repair at 5 years: does aneurysm diameter predict outcome? J Vasc Surg 44:920-929; discussion 929-931, 2006