- Email: [email protected]
Influencing factors of sac shrinkage after endovascular aneurysm repair
Influencing factors of sac shrinkage after endovascular aneurysm repair Florent Lalys, PhD,a Anne Daoudal, MD,b,c,d Juliette Gindre, PhD,b,c,e Cemil Göksu, PhD,a Antoine Lucas, MD,b,c,d and Adrien Kaladji, MD, PhD,b,c,d Rennes and Villeurbanne, France
ABSTRACT Objective: Sac shrinkage is considered a reliable surrogate marker of success after endovascular aneurysm repair (EVAR). Whereas sac shrinkage is the best expected outcome, predictive factors of sac shrinkage remain unclear. The aim of this study was to identify the role of preoperative and postoperative influencing factors of sac reduction after EVAR. Methods: Online searches across MEDLINE, Embase, and Cochrane Library medical databases were simultaneously performed. Study effects were pooled using a random-effects model, and forest plots were generated for every potential influencing factor. Results: A total of 24 studies with 14,754 patients were included (mean age, 73.4 years; 76% male). At a mean follow-up of 24 months, the pooled shrinkage proportion was 47%. Random-effects meta-analysis revealed that renal impairment (odds ratio [OR], 0.74; 95% confidence interval [CI], 0.57-0.96), type I endoleaks (OR, 0.17; 95% CI, 0.08-0.39), type II endoleaks (OR, 0.21; 95% CI, 0.14-0.33), and combined type I and type II endoleaks (OR, 0.32; 95% CI, 0.22-0.47) were found to prevent sac shrinkage, whereas hypercholesterolemia (OR, 1.24; 95% CI, 1.02-1.51) and smoking (OR, 1.32; 95% CI, 1.17-1.49) have a significant positive impact on sac shrinkage. In addition, there was a trend toward the association between shrinkage and statin therapy (OR, 4.07; 95% CI, 1.02-16.32) and nearly significant negative impacts of coronary artery disease (OR, 0.84; 95% CI, 0.70-1.01), diabetes (OR, 0.79; 95% CI, 0.60-1.04), and sac thrombus (OR, 0.88; 95% CI, 0.77-1.01) on sac shrinkage. Conclusions: In this large meta-analysis of patients undergoing EVAR, we found that several comorbidity and postoperative factors were associated with postoperative sac shrinkage. These findings may contribute to a better understanding of the shrinkage process of patients undergoing EVAR. (J Vasc Surg 2017;-:1-9.)
The best expected outcome after endovascular aneurysm repair (EVAR) is a significant shrinkage of the aneurysmal sac and freedom from graft-related complications. Sac shrinkage is currently the best indicator of aneurysm exclusion and durable treatment success.1,2 Reporting standards in vascular surgery have been developed in an attempt to unify methods of conveying size changes after EVAR.3 Diameter change of 5 mm between two consecutive computed tomography (CT) scans is considered significant for sac shrinkage or expansion. Several studies4,5 have shown that patients who have significant sac shrinkage in the early postoperative phase will experience fewer complications and consequently require less intensive imaging surveillance. To this end, detecting patients having a high probability of sac shrinkage is becoming almost as important as detecting patients at risk for development of complications.
From the Therenva,a INSERM U1099,b Signal and Image Processing Laboratory (LTSI), University Rennes 1,c and Department of Cardiothoracic and Vascular Surgery, CHU Rennes,d Rennes; and the ANSYS, Villeurbanne.e Author conflict of interest: none. Correspondence: Florent Lalys, PhD, Therenva, 4 rue Jean Jaures, Rennes 35000, France (e-mail: fl[email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2017 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.12.131
Even if sac shrinkage is now considered a reliable surrogate marker of durable successful EVAR, there is a lack of consensus on the mechanisms by which sac changes occur and on real influencing factors. It is likely that demographic, anatomic, or physiologic variables may influence the behavior of the aneurysm sac after EVAR, but previous retrospective studies have shown different and contradictory results of determining influencing factors. A more detailed understanding of the sac shrinkage process is essential to enable clinicians to understand how aneurysms are expected to behave after EVAR. The aim of this meta-analysis was therefore to identify and to evaluate the role of influencing factors of sac shrinkage after EVAR.
METHODS Search strategy. Searches were simultaneously conducted across MEDLINE, Embase, and Cochrane Library medical databases. The search keywords “endovascular,” “EVAR,” “sac shrinkage,” “sac regression,” “aneurysm shrinkage,” and “aneurysm regression” were used in combination with the Boolean operators AND and OR. References of the full-text articles were also screened for other records of interest. Three authors independently performed the search and assessed the full text of all identified records for eligibility. The last search was performed on February 1, 2016. Studies published before 2000 were not considered, nor were studies in languages other than English or studies reporting fewer than 40 patients to reach correct statistical power. Publications 1
2
Journal of Vascular Surgery
Lalys et al
---
were also excluded if they focused on subgroups of patients with specific treatment, if median follow-up was <6 months, or if studies focused on subgroups of patients with endoleak or other adverse events. In case of patients’ database or registry duplicates, the more detailed publication was preferred. When articles reported only P value, results were kept for the discussion. When data on both 5-mm and 10-mm sac regression were reported, results of 5-mm regression were preferred according to the reporting standard.3 Similarly, when data at multiple time points were available, longer follow-up was preferred. Data extraction and quality assessment. Methodologic, population characteristic, risk factor, and outcome data were extracted and tabulated into a computerized spreadsheet. When available, the following were rigorously extracted: year of publication, range years of inclusion, study design, sample size, patient follow-up (median, by default mean), mean age, mean preoperative abdominal aortic aneurysm (AAA) diameter, sac shrinkage definition in terms of size and follow-up, sac shrinkage incidence, and information on stent graft models. The quality of the observational studies was assessed using the Newcastle-Ottawa Scale. Based on a 9-point system, patient selection, comparability of the study groups, and ascertainment of study outcome are assessed. A score of 9 stars indicates a study with a low risk of bias; a score of 7 or 8 indicates medium bias risk, whereas a score of 6 indicates a high chance of bias. For each influencing factor, the effect estimate was extracted, regardless of the reported strength of association. The effect estimate of interest for each predictor was the odds ratio (OR), which is the preferred measure of association between a potential predictor and an outcome. An OR >1 represents a positive effect of the predictor on the outcome, whereas an OR <1 represents a negative effect of the predictor on the outcome. Reported adjusted ORs from multivariate analyses were preferred. For univariate analyses, ORs were either directly extracted from text or tables or computed with reported percentages or numbers. We pooled results only on factors reported in at least three studies. For better readability, predictors have been grouped into demographic-, treatment-, comorbidity-, morphologic-, and postoperative-related factors. To minimize risk factor measurement bias, only studies with objective categorization and clear reporting were included in the analysis of individual risk factors. Statistical analysis. Study effects were pooled using a random-effects model, and forest plots were generated for every such factor to have a better understanding of the influence of the included studies. Cohen k statistics was used to evaluate the inter-rater reliability for study eligibility, and intraclass correlations were employed to evaluate inter-reviewer reliability for data collection. Because of the high variability of included
2017
studies in terms of outcome definition, sample size, or duration of follow-up, a deep sensitivity analysis of methodologic factors was performed to study all potential biases. In particular, publication bias was visually assessed by a funnel plot and statistically assessed by Egger test6; attrition bias was investigated by analyzing duration of follow-up; performance bias was evaluated by analyzing the median year of study recruitment; and selection biases were assessed by analyzing study design, sample size, mean age, and mean preoperative AAA diameter.
RESULTS Study selection and characteristics. The online database search identified 1145 records, and 38 additional records were identified from review of the included studies following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology (Fig 1). We excluded 1041 studies after reviewing abstracts and 42 after reviewing full text; 24 studies finally met our full inclusion criteria.1,2,4,7-27 Population characteristics are shown in Fig 2. No prospective study was identified, whereas seven were multicenter studies. A total of 14,754 patients (range, 49-8638) with an average followup of 24 months (range, 6-54.8 months) were included in the analysis, with a mean age of 73.4 years (range, 69.2-79.9 years), a majority of men (76%), and a mean preoperative AAA diameter of 56.4 cm (range, 47.7-62 cm). The overall pooled sac shrinkage proportion was 47% (95% confidence interval [CI], 38%-56%). Based on available data of stent graft models, 32.8% of patients were implanted with a Zenith (Cook Medical, Bloomington, Ind), 14.7% with a Talent (Medtronic, Santa Rosa, Calif), 17.8% with an AneuRx (Medtronic), 13% with a Vanguard (Boston Scientific, Marlborough, Mass), 11.9% with a Gore Excluder (W. L. Gore & Associates, Flagstaff, Ariz), and 9.8% with other types of stent grafts. Data extraction and quality assessment. There was a correct strength of agreement for the study eligibility (k ¼ 0.76; 95% CI, 0.68-0.99). Divergence in study eligibility was settled by consensus between reviewers. Overall, 94% of reported ORs were computed from univariate analysis and only 6% from multivariate analysis (linear or logistic multiple regression models). The mean Newcastle-Ottawa Scale score was 7.3, with a minimum of 6 and a maximum of 9. The sensitivity analysis did not identify any methodologic factors that influence sac shrinkage incidence, being selection, attrition, or performance bias (Fig 3). There was also no evidence of publication bias based on the Egger test (P ¼ .10; Fig 4). Influencing factors. A total of 17 factors related to demography, treatment, comorbidity, morphology, and postoperative events were statistically pooled. Fig 5 summarizes results for all factors. Based on pooled ORs, renal impairment (OR, 0.74; 95% CI, 0.57-0.96), type I
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
-
Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of the selection process. Excluded studies can have multiple exclusion criteria. EVAR, Endovascular aneurysm repair; T2Es, type II endoleaks; TEVAR, thoracic endovascular aortic repair.
Fig 2. Clinical and methodologic characteristics of included studies. The forest plot of shrinkage proportion is shown on the right. Proportions are shown with 95% confidence interval (CI). AAA, Abdominal aortic aneurysm; NOS, Newcastle-Ottawa Scale.
3
4
Journal of Vascular Surgery
Lalys et al
---
2017
Fig 3. Sensitivity analysis of methodologic factors. Pooled proportions of sac shrinkage are shown in percentage. AAA, Abdominal aortic aneurysm.
Fig 4. Funnel plot of effect size against sample size for all included studies. Three colors are used to differentiate sac shrinkage definitions. CI, Confidence interval; SD, standard deviation.
endoleaks (OR, 0.17; 95% CI, 0.08-0.39; I2, 0.90), type II endoleaks (OR, 0.21; 95% CI, 0.14-0.33), and combined type I and type II endoleaks (OR, 0.32; 95% CI, 0.22-0.47) were found to significantly prevent sac shrinkage, whereas hypercholesterolemia (OR, 1.24; 95% CI, 1.02-1.51) and smoking (OR, 1.32; 95% CI, 1.17-1.49) were found to have a positive impact on sac shrinkage. From these factors, only pooled ORs from type II endoleaks showed high heterogeneity (I2, 0.03). There was also a trend toward the association between shrinkage and statin therapy (OR, 4.07; 95% CI, 1.02-16.32) despite the high statistical heterogeneity (I2, 0.011). Coronary artery disease (OR, 0.84; 95% CI, 0.70-1.01), diabetes (OR, 0.79; 95% CI, 0.60-1.04), and sac thrombus (OR, 0.88; 95% CI,
0.77-1.01) were found to have a nearly significant negative impact on sac shrinkage. All others factors were found not to influence sac shrinkage.
DISCUSSION For patients with AAA who have suitable anatomy, EVAR is more and more considered the “gold standard” rather than open repair, and regression of the aneurysm sac has been used as a surrogate marker for technical success. However, persistent sac shrinkage is preferred because of the dynamic behavior of the aneurysmal shrinkage process. Specifically, Lee et al28 demonstrated that a reliable predictor of success is a volume reduction of 10% or more at 6 months and continuing regression
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
5
-
Fig 5. Pooled odds ratios (ORs) of sac shrinkage for identified demographic, treatment, comorbidity, morphologic, and postoperative factors. Number of included studies, dedicated number of patients, associated P value, and heterogeneity test (I2) are also shown. Pooled ORs are reported with 95% confidence interval (CI). AAA, Abdominal aortic aneurysm.
over time. Bastos Gonçalves et al4 later found similar results with a larger contemporary series. In that study, patients were divided in three groups: minor, major sac regression, and no sac regression. This last group included only 2.3% of patients with sac expansion (>5 mm) and consisted therefore of the majority of patients with “sac stabilization.” Results of the multivariate analysis showed that risk of late complications increased by 3.1 times for patients without sac regression (ie, almost only patients with sac stabilization) compared with major shrinkage (>10 mm), suggesting that sac regression (especially >10 mm 1 year after device implantation) is a strong predictor of long-lasting success after EVAR, probably rather than sac stabilization. Houbballah et al1 and Bisdas et al5 showed that AAA shrinkage >5 mm after EVAR was also associated with statistically significantly longer survival. Based on the present random-effects meta-analysis, several independent factors have been clearly identified. Endoleaks are well-recognized factors of sac enlargement29; it was therefore consistent to find that the absence of type I endoleaks, type II endoleaks, or combined type I and type II endoleaks can facilitate sac shrinkage. Paradoxically, two cardiovascular risks factors (hypercholesterolemia and smoking) showed a positive effect on shrinkage. According to the European Collaborators on Stent/graft Techniques for aortic Aneurysm Repair (EUROSTAR) registry,30 smokers seem to be less exposed to late endograft-related complications, especially to endoleaks. This result was confirmed in the latest study12 and could explain why smokers are more likely to present with sac regression. Other associations found in the present meta-analysis should be interpreted with
caution because the majority of the included studies are retrospective. Actually, recurrent factors are identified in the included studies, and we can only make a hypothesis to explain such results. Even if aneurysmal calcifications are rarely reported, it is acknowledged that patients with renal impairment are more likely to present with arterial calcification, including the aneurysmal sac. Therefore, arterial compliance is decreasing,31 and it would be reasonable to expect that the regression capacity of an aneurysm is partly dependent on these factors in the absence of any endoleaks. Although it is not systematically associated, patients suffering from hypercholesterolemia are often treated with statin therapy. As found in this meta-analysis, statin therapy is also associated with aneurysmal sac regression. Raux et al13 tried to explain how statin therapy may have a role in sac regression through the reduction of the inflammatory matrix metalloproteinases in the aortic wall particularly. It was reported in several publications12,30 that tobacco had paradoxically a protective effect on the occurrence of type II endoleaks, probably through the prothrombogenic effect of tobacco, which can explain why aneurysmal regression is more likely to occur in smokers. Whereas the natural history of AAA sac evolution after EVAR in the presence of a type II endoleak remains to be defined, this meta-analysis clearly identified the absence of endoleaks of type I, type II, or combination of both as a strong predictor of sac shrinkage and by definition of technical success. However, because sac shrinkage developed in some patients over time in spite of the presence of type II endoleaks, not all type II endoleaks have an impact on sac change. In the metaregression by Karthikesalingam et al,32 there was no
6
Lalys et al
Journal of Vascular Surgery ---
2017
Fig 6. Number of studies quantifying influence of continuous factors on sac shrinkage. AAA, Abdominal aortic aneurysm.
evidence that any strategy for isolated type II endoleak intervention improved sac regression compared with using a conservative approach. It is in contradiction to the results of this meta-analysis. Other factors have not been included in the metaanalysis. Influence of continuous factors on sac shrinkage is, for instance, extremely difficult to review. Because of the high heterogeneity in the definition of those factors across studies, we did not succeed in finding enough studies with available ORs and similar definition. A count of studies reporting P values of different continuous factors is proposed, however, in Fig 6. For age, proximal neck length, and proximal neck diameter, no consensus was found. For the preoperative maximum AAA diameter, contradictory results have been published (nine studies describing nonassociation vs eight studies showing significant association), and further studies will be needed on a large cohort to determine its real impact. We also did not include device-specific clinical outcomes, despite differences found in incidence of device migration, frequency of endoleaks, or risk of limb thrombosis. In particular, early results have shown that sac size change after EVAR is device specific.33,34 Different trials have also indicated that rates of sac expansion or regression were correlated to graft models.2,26 However, for the majority of them, these studies compared models of the oldest generation of stent grafts and the next one. These differences between first and second generation seem to be less important with the latest generation. Two recent studies have compared the last two generations of stent grafts,35,36 and it seems that there is no major difference between them in AAA-related outcomes (sac behavior, endoleaks). Both studies have shown that the most recent generation allowed more challenging anatomic cases to be treated. Since these studies have been reported, new endografts are available, and our potential conclusion about the difference between endografts would have not been relevant because the models of endografts in the meta-analysis are no longer used. A large number of factors have not been included in this meta-analysis because of lack of comparison data.
Within clinical or biologic factors, preoperative C-reactive protein level and atherosclerosis appear to be associated with an increased likelihood of sac regression,37 as are calcium channel blockers,15 angiotensin-converting enzyme inhibitors,15 and tranexamic acid.38 Chikazawa et al39 found that an additional thoracic aortic aneurysm had a negative impact on sac shrinkage. In morphology-related factors, the absences of calcification and thrombus at the level of the iliac arteries and aortic neck have been identified as significant factors affecting aneurysm shrinkage.14 Neck thrombus or plaque was found to be significant,40 despite being nonsignificant in another study.7 Overall increased proportion of thrombus on preoperative CT also resulted in greater likelihood of sac shrinkage.41 Whereas patencyrelated factors have been mostly highlighted in predictive models of endoleaks, some studies made the association with sac shrinkage. Sheehan et al42 and Kray et al43 found a significant association with patent inferior mesenteric artery and coiling of the lumbar or inferior mesenteric artery, whereas Back et al27 identified the overall patency of aortic side branches as a significant predictor, contrary to Torsello et al.22 Bastos Gonçalves et al44 significantly stratified their cohort into a low-risk group (with proximal and distal seal length of 10 mm and no endoleak) and a high-risk group (with insufficient seal, presence of endoleak, or both). The mean diameter of the infrarenal aorta,1 endograft oversizing,45 patients without postoperative migration,46 mean pressure index,47 and effective seal length and surface area of circumferential apposition between the aorta and the graft material48 are other factors of interest that will need further consideration. Finally, shrinkage rates seem to be different among hospitals, as shown using the same endograft and protocol-defined patient selection criteria.49 Identification of influencing factors is also a first step toward predictive models. Preoperative prediction of complications or success is critical to help physicians in decision-making, to inform patients, and to adjust follow-up surveillance. As we develop a better understanding of expected aneurysmal sac evolution, similar
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
7
-
predictive models could help preoperatively stratify patients and tailor postoperative surveillance to the expected aneurysmal sac evolution. On the one hand, patients having a low likelihood of shrinking of their sac might be at high risk of complications and thus require intensive and long-term surveillance. It may be necessary to monitor those patients carefully through continued imaging, whereas on the contrary for patients with a high likelihood of sac regression, imaging follow-up intensiveness might be reduced. According to Soler et al,7 late failures might still occur in patients with significant sac shrinkage, so even for those patients, lifelong follow-up should be provided. To improve surveillance, another potential surveillance technique is the serial assessment of intrasac pressure. It has been shown to be feasible50 and strongly correlated with type II endoleaks,47 and it could help the decision-making process related to reintervention after EVAR. However, further studies are still needed to validate its utility and reproducibility. Whereas sac shrinkage is desired and remains a key factor, another important consideration is the morphologic change in length and angulation introduced by the shrinkage. Morphologic changes are regularly evaluated by using serial CT angiography to identify complications, and information about the sac diameter or volume change can be retrieved. In particular, studies have shown that volume measurements should be performed in addition to diameter measurements on CT angiography51 for better robustness. Those changes may have an impact on attachment of the stent graft components and overall graft shape, which can be the cause of other complications. This study has some limitations. Even if the sensitivity analysis did not find any evidence of bias, and even if the study quality assessment showed that all included papers were methodologically robust, slight clinical heterogeneity, arising from differences in study populations and consequential distribution risk factors, may be present. A large range of EVAR experience, hospital routines, and complication definitions may also hamper the generalizability of the results. Obviously, we do not recommend to our patients to smoke and stop therapy against hypercholesterolemia. For example, diabetes is a well-recognized protective factor against aneurysm rupture (not a statistical error), and nobody says that diabetes should not be controlled for in patients presenting with AAAs. Currently, data are not strong enough to determine if these factors are a statistical finding or have a real impact on aneurysmal sac behavior after EVAR.
CONCLUSIONS As our current understanding of influencing factors of sac shrinkage after EVAR is weak, this meta-analysis aimed to determine significant factors to identify in
which patients sac shrinkage is more likely to be high. Whereas diabetes, coronary artery disease, renal impairment, sac thrombus, and type I and type II endoleaks seem to prevent sac shrinkage, patients receiving statin therapy, patients with hypercholesterolemia, and active smokers may have a better chance to observe sac regression and therefore to have a lower risk of reintervention.
AUTHOR CONTRIBUTIONS Conception and design: FL, CG, AL, AK Analysis and interpretation: FL, AK Data collection: FL, AD, JG Writing the article: FL, AD, JG, CG, AL, AK Critical revision of the article: FL, AD, JG, CG, AL, AK Final approval of the article: FL, AD, JG, CG, AL, AK Statistical analysis: FL Obtained funding: Not applicable Overall responsibility: FL
REFERENCES 1. Houbballah R, Majewski M, Becquemin JP. Significant sac retraction after endovascular aneurysm repair is a robust indicator of durable treatment success. J Vasc Surg 2010;52: 878-83. 2. Rhee RY, Eskandari MK, Zajko AB, Makaroun MS. Long-term fate of the aneurysmal sac after endoluminal exclusion of abdominal aortic aneurysms. J Vasc Surg 2000;32:689-96. 3. Chaikof EL, Blankensteijn JD, Harris PL, White GH, Zarins CK, Bernhard VM, et al. Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2002;35:1048-60. 4. Bastos Gonçalves F, Baderkhan H, Verhagen HJ, Wanhainen A, Björck M, Stolker RJ, et al. Early sac shrinkage predicts a low risk of late complications after endovascular aortic aneurysm repair. Br J Surg 2014;101:802-10. 5. Bisdas T, Weiss K, Eisenack M, Austermann M, Torsello G, Donas KP. Durability of the Endurant stent graft in patients undergoing endovascular abdominal aortic aneurysm repair. J Vasc Surg 2014;60:1125-31. 6. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34. 7. Soler RJ, Bartoli MA, Mancini J, Lerussi G, Thevenin B, SarlonBartoli G, et al. Aneurysm sac shrinkage after endovascular repair: predictive factors and long-term follow-up. Ann Vasc Surg 2015;29:770-9. 8. Pini R, Faggioli G, Mascoli C, Gallitto E, Freyrie A, Gargiulo M, et al. Influence of statin therapy on type 2 endoleak evolution. Ann Vasc Surg 2015;29:1167-73. 9. Nakai M, Ikoma A, Sato H, Sato M, Nishimura Y, Okamura Y. Risk factors associated with late aneurysmal sac expansion after endovascular abdominal aortic aneurysm repair. Diagn Interv Radiol 2015;21:195-201. 10. Tsuyuki Y, Matsushita S, Dohi S, Yamamoto T, Tambara K, Inaba H, et al. Factors for sac size change of abdominal aortic aneurysm after endovascular repair. Ann Thorac Cardiovasc Surg 2014;20:1016-20. 11. Cieri E, De Rango P, Isernia G, Simonte G, Verzini F, Parlani G, et al. Effect of stentgraft model on aneurysm shrinkage in 1, 450 endovascular aortic repairs. Eur J Vasc Endovasc Surg 2013;46:192-200. 12. Koole D, Moll FL, Buth J, Hobo R, Zandvoort H, Pasterkamp G, et al. The influence of smoking on endovascular abdominal aortic aneurysm repair. J Vasc Surg 2012;55:1581-6.
8
Lalys et al
Journal of Vascular Surgery ---
13. Raux M, Cochennec F, Becquemin JP. Statin therapy is associated with aneurysm sac regression after endovascular aortic repair. J Vasc Surg 2012;55:1587-92. 14. Kaladji A, Cardon A, Abouliatim I, Campillo-Gimenez B, Heautot JF, Verhoye JP. Preoperative predictive factors of aneurysmal regression using the reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2012;55: 1287-95. 15. Bailey MA, Sohrabi S, Flood K, Griffin KJ, Rashid ST, Johnson AB, et al. Calcium channel blockers enhance sac shrinkage after endovascular aneurysm repair. J Vasc Surg 2012;55:1593-9. 16. Hogg ME, Morasch MD, Park T, Flannery WD, Makaroun MS, Cho JS. Long-term sac behavior after endovascular abdominal aortic aneurysm repair with the Excluder lowpermeability endoprosthesis. J Vasc Surg 2011;53:1178-83. 17. El Sayed HF, Meier GH, Mendoza B, Sprouse LR, Parent FN, Panneton JM. Aneurysm regression after endovascular aneurysm repair: what should we expect? Vasc Endovascular Surg 2009;42:545-50. 18. Becquemin JP, Aksoy M, Marzelle J, Roudot-Thoraval F, Desgranges P, Allaire E, et al. Abdominal aortic aneurysm sac behavior following Cook Zenith graft implantation: a fiveyear follow-up assessment of 212 cases. J Cardiovasc Surg (Torino) 2008;49:199-206. 19. Broker HS, Foteh KI, Murphy EH, Davis CM, Clagett GP, Modrall JG, et al. Device-specific aneurysm sac morphology after endovascular aneurysm repair: evaluation of contemporary graft materials. J Vasc Surg 2008;47:702-6; discussion: 707. 20. Higashiura W, Greenberg RK, Katz E, Geiger L, Bathurst S. Predictive factors, morphologic effects, and proposed treatment paradigm for type II endoleaks after repair of infrarenal abdominal aortic aneurysms. J Vasc Interv Radiol 2007;18: 975-81. 21. Golledge J, Parr A, Boult M, Maddern G, Fitridge R. The outcome of endovascular repair of small abdominal aortic aneurysms. Ann Surg 2007;245:326-33. 22. Torsello G, Osada N, Florek HJ, Horsch S, Kortmann H, Luska G, et al. Long-term outcome after Talent endograft implantation for aneurysms of the abdominal aorta: a multicenter retrospective study. J Vasc Surg 2006;43:277-84; discussion: 284. 23. Biebl M, Hakaim AG, Oldenburg WA, Klocker J, Lau LL, Neuhauser B, et al. Does chronic oral anticoagulation with warfarin affect durability of endovascular aortic aneurysm exclusion in a midterm follow-up? J Endovasc Ther 2005;12: 58-65. 24. Zarins CK, Bloch DA, Crabtree T, Matsumoto AH, White RA, Fogarty TJ. Aneurysm enlargement following endovascular aneurysm repair: AneuRx clinical trial. J Vasc Surg 2004;39: 109-17. 25. Sampaio SM, Panneton JM, Mozes GI, Andrews JC, Bower TC, Karla M, et al. Proximal type I endoleak after endovascular abdominal aortic aneurysm repair: predictive factors. Ann Vasc Surg 2004;18:621-8. 26. Sternbergh WC, Conners MS, Tonnessen BH, Carter G, Money SR. Aortic aneurysm sac shrinkage after endovascular repair is deviceedependent: a comparison of Zenith and AneuRx endografts. Ann Vasc Surg 2003;17: 49-53. 27. Back MR, Bowser AN, Johnson BL, Schmacht D, Zwiebel B, Bandyk DF. Patency of infrarenal aortic side branches determines early aneurysm sac behavior after endovascular repair. Ann Vasc Surg 2003;17:27-34. 28. Lee JT, Aziz IN, Lee JT, Haukoos JS, Donayre CE, Walot I, et al. Volume regression of abdominal aortic aneurysms and its
2017
relation to successful endoluminal exclusion. J Vasc Surg 2003;38:1254-63. 29. Schanzer A, Greenberg RK, Hevelone N, Robinson WP, Eslami MH, Goldberg RJ, et al. Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair. Circulation 2011;123:2848-55. 30. Lottman PE, Van Marrewijk CJ, Fransen GA, Laheij RJ, Buth J. Impact of smoking on endovascular abdominal aortic aneurysm surgery outcome. Eur J Vasc Endovasc Surg 2004;27:512-8. 31. Speelman L, Bohra A, Bosboom EM, Schurink GW, van de Vosse FN, Makaroun MS, et al. Effects of wall calcifications in patient-specific wall stress analyses of abdominal aortic aneurysms. J Biomech Eng 2007;129:105-9. 32. Karthikesalingam A, Thrumurthy SG, Jackson D, Choke E, Sayers RD, Loftus IM, et al. Current evidence is insufficient to define an optimal threshold for intervention in isolated type II endoleak after endovascular aneurysm repair. J Endovasc Ther 2012;19:200-8. 33. Ouriel K, Clair DG, Greenberg RK, Lyden SP, O’Hara PJ, Sarac TP, et al. Endovascular repair of abdominal aortic aneurysms: device-specific outcome. J Vasc Surg 2003;37:991-8. 34. Bertges DJ, Chow K, Wyers MC, Landsittel D, Frydrych AV, Stavropoulos W, et al. Abdominal aortic aneurysm size regression after endovascular repair is endograft dependent. J Vasc Surg 2003;37:716-23. 35. Sobocinski J, Briffa F, Holt PJ, Martin Gonzalez T, Spear R, Azzaoui R, et al. Evaluation of the Zenith low-profile abdominal aortic aneurysm stent graft. J Vasc Surg 2015;62: 841-7. 36. Maudet A, Daoudal A, Cardon A, Clochard E, Lucas A, Verhoye JP, et al. Endovascular treatment of infrarenal aneurysms: comparison of the results of second- and thirdgeneration stent grafts. Ann Vasc Surg 2016;34:95-105. 37. Ishibashi H, Ishiguchi T, Ohta T, Sugimoto I, Iwata H, Yamada T, et al. Mid-term results of endovascular abdominal aortic aneurysm repair: is it possible to predict sac shrinkage? Surg Today 2011;41:1605-9. 38. Aoki A, Suezawa T, Yamamoto S, Sangawa K, Irie H, Mayazaki N, et al. Effect of antifibrinolytic therapy with tranexamic acid on abdominal aortic aneurysm shrinkage after endovascular repair. J Vasc Surg 2014;59:1203-8. 39. Chikazawa G, Hiraoka A, Totsugawa T, Tamura K, Ishida A, Sakaguchi T, et al. Influencing factors for abdominal aortic aneurysm sac shrinkage and enlargement after EVAR: clinical reviews before introduction of preoperative coil embolization. Ann Vasc Dis 2014;7:280-5. 40. Fairman RM, Nolte L, Snyder SA, Chuter TA, Greenberg RK; Zenith Investigators. Factors predictive of early or late aneurysm sac size change following endovascular repair. J Vasc Surg 2006;43:649-56. 41. Sadek M, Dexter DJ, Rockman CB, Hoang H, Mussa FF, Cayne NS, et al. Preoperative relative abdominal aortic aneurysm thrombus burden predicts endoleak and sac enlargement after endovascular aneurysm repair. Ann Vasc Surg 2013;27:1036-41. 42. Sheehan MK, Hagino RT, Canby E, Wholey MH, Postoak D, Suri R, et al. Type 2 endoleaks after abdominal aortic aneurysm stent grafting with systematic mesenteric and lumbar coil embolization. Ann Vasc Surg 2006;20:458-63. 43. Kray J, Kirk S, Franko J, Chew DK. Role of type II endoleak in sac regression after endovascular repair of infrarenal abdominal aortic aneurysms. J Vasc Surg 2015;61:869-74. 44. Bastos Gonçalves F, van de Luijtgaarden KM, Hoeks SE, Hendriks JM, ten Raa S, Rouwet EV, et al. Adequate seal and no endoleak on the first postoperative computed tomography angiography as criteria for no additional imaging up to
Journal of Vascular Surgery Volume
45.
46.
47.
48.
-,
Number
Lalys et al
9
-
5 years after endovascular aneurysm repair. J Vasc Surg 2013;57:1503-11. Sternbergh WC, Money SR, Greenberg RK, Chuter TA; Zenith Investigators. Influence of endograft oversizing on device migration, endoleak, aneurysm shrinkage, and aortic neck dilation: results from the Zenith Multicenter Trial. J Vasc Surg 2004;39:20-6. Lee JT, Lee J, Aziz I, Donayre CE, Walot I, Kopchok GE, et al. Stent-graft migration following endovascular repair of aneurysms with large proximal necks: anatomical risk factors and long-term sequelae. J Endovasc Ther 2002;9:652-64. Dias NV, Ivancev K, Resch TA, Malina M, Sonesson B. Endoleaks after endovascular aneurysm repair lead to nonuniform intra-aneurysm sac pressure. J Vasc Surg 2007;46:197-203. Arthurs ZM, Ouriel K, Welborn MB III, McDaniel HB. Extended proximal seal zone during EVAR predicts aneurysm sac regression. J Vasc Surg 2014;60:546.
49. Kritpracha B, Beebe HG, Criado FJ, Comerota AJ. Postendograft abdominal aortic aneurysm shrinkage varies among hospitals: observations from multicenter trials. J Endovasc Ther 2004;11:454-9. 50. Ellozy SH, Carroccio A, Lookstein RA, Jacobs TS, Addis MD, Teodorescu VJ, et al. Abdominal aortic aneurysm sac shrinkage after endovascular aneurysm repair: correlation with chronic sac pressure measurement. J Vasc Surg 2006;43:2-7. 51. van Keulen JW, van Prehn J, Prokop M, Moll FL, van Herwaarden JA. Potential value of aneurysm sac volume measurements in addition to diameter measurements after endovascular aneurysm repair. J Endovasc Ther 2009;16: 506-13.
Submitted Aug 2, 2016; accepted Dec 16, 2016.
ABSTRACT Objective: Sac shrinkage is considered a reliable surrogate marker of success after endovascular aneurysm repair (EVAR). Whereas sac shrinkage is the best expected outcome, predictive factors of sac shrinkage remain unclear. The aim of this study was to identify the role of preoperative and postoperative influencing factors of sac reduction after EVAR. Methods: Online searches across MEDLINE, Embase, and Cochrane Library medical databases were simultaneously performed. Study effects were pooled using a random-effects model, and forest plots were generated for every potential influencing factor. Results: A total of 24 studies with 14,754 patients were included (mean age, 73.4 years; 76% male). At a mean follow-up of 24 months, the pooled shrinkage proportion was 47%. Random-effects meta-analysis revealed that renal impairment (odds ratio [OR], 0.74; 95% confidence interval [CI], 0.57-0.96), type I endoleaks (OR, 0.17; 95% CI, 0.08-0.39), type II endoleaks (OR, 0.21; 95% CI, 0.14-0.33), and combined type I and type II endoleaks (OR, 0.32; 95% CI, 0.22-0.47) were found to prevent sac shrinkage, whereas hypercholesterolemia (OR, 1.24; 95% CI, 1.02-1.51) and smoking (OR, 1.32; 95% CI, 1.17-1.49) have a significant positive impact on sac shrinkage. In addition, there was a trend toward the association between shrinkage and statin therapy (OR, 4.07; 95% CI, 1.02-16.32) and nearly significant negative impacts of coronary artery disease (OR, 0.84; 95% CI, 0.70-1.01), diabetes (OR, 0.79; 95% CI, 0.60-1.04), and sac thrombus (OR, 0.88; 95% CI, 0.77-1.01) on sac shrinkage. Conclusions: In this large meta-analysis of patients undergoing EVAR, we found that several comorbidity and postoperative factors were associated with postoperative sac shrinkage. These findings may contribute to a better understanding of the shrinkage process of patients undergoing EVAR. (J Vasc Surg 2017;-:1-9.)
The best expected outcome after endovascular aneurysm repair (EVAR) is a significant shrinkage of the aneurysmal sac and freedom from graft-related complications. Sac shrinkage is currently the best indicator of aneurysm exclusion and durable treatment success.1,2 Reporting standards in vascular surgery have been developed in an attempt to unify methods of conveying size changes after EVAR.3 Diameter change of 5 mm between two consecutive computed tomography (CT) scans is considered significant for sac shrinkage or expansion. Several studies4,5 have shown that patients who have significant sac shrinkage in the early postoperative phase will experience fewer complications and consequently require less intensive imaging surveillance. To this end, detecting patients having a high probability of sac shrinkage is becoming almost as important as detecting patients at risk for development of complications.
From the Therenva,a INSERM U1099,b Signal and Image Processing Laboratory (LTSI), University Rennes 1,c and Department of Cardiothoracic and Vascular Surgery, CHU Rennes,d Rennes; and the ANSYS, Villeurbanne.e Author conflict of interest: none. Correspondence: Florent Lalys, PhD, Therenva, 4 rue Jean Jaures, Rennes 35000, France (e-mail: fl[email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2017 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.12.131
Even if sac shrinkage is now considered a reliable surrogate marker of durable successful EVAR, there is a lack of consensus on the mechanisms by which sac changes occur and on real influencing factors. It is likely that demographic, anatomic, or physiologic variables may influence the behavior of the aneurysm sac after EVAR, but previous retrospective studies have shown different and contradictory results of determining influencing factors. A more detailed understanding of the sac shrinkage process is essential to enable clinicians to understand how aneurysms are expected to behave after EVAR. The aim of this meta-analysis was therefore to identify and to evaluate the role of influencing factors of sac shrinkage after EVAR.
METHODS Search strategy. Searches were simultaneously conducted across MEDLINE, Embase, and Cochrane Library medical databases. The search keywords “endovascular,” “EVAR,” “sac shrinkage,” “sac regression,” “aneurysm shrinkage,” and “aneurysm regression” were used in combination with the Boolean operators AND and OR. References of the full-text articles were also screened for other records of interest. Three authors independently performed the search and assessed the full text of all identified records for eligibility. The last search was performed on February 1, 2016. Studies published before 2000 were not considered, nor were studies in languages other than English or studies reporting fewer than 40 patients to reach correct statistical power. Publications 1
2
Journal of Vascular Surgery
Lalys et al
---
were also excluded if they focused on subgroups of patients with specific treatment, if median follow-up was <6 months, or if studies focused on subgroups of patients with endoleak or other adverse events. In case of patients’ database or registry duplicates, the more detailed publication was preferred. When articles reported only P value, results were kept for the discussion. When data on both 5-mm and 10-mm sac regression were reported, results of 5-mm regression were preferred according to the reporting standard.3 Similarly, when data at multiple time points were available, longer follow-up was preferred. Data extraction and quality assessment. Methodologic, population characteristic, risk factor, and outcome data were extracted and tabulated into a computerized spreadsheet. When available, the following were rigorously extracted: year of publication, range years of inclusion, study design, sample size, patient follow-up (median, by default mean), mean age, mean preoperative abdominal aortic aneurysm (AAA) diameter, sac shrinkage definition in terms of size and follow-up, sac shrinkage incidence, and information on stent graft models. The quality of the observational studies was assessed using the Newcastle-Ottawa Scale. Based on a 9-point system, patient selection, comparability of the study groups, and ascertainment of study outcome are assessed. A score of 9 stars indicates a study with a low risk of bias; a score of 7 or 8 indicates medium bias risk, whereas a score of 6 indicates a high chance of bias. For each influencing factor, the effect estimate was extracted, regardless of the reported strength of association. The effect estimate of interest for each predictor was the odds ratio (OR), which is the preferred measure of association between a potential predictor and an outcome. An OR >1 represents a positive effect of the predictor on the outcome, whereas an OR <1 represents a negative effect of the predictor on the outcome. Reported adjusted ORs from multivariate analyses were preferred. For univariate analyses, ORs were either directly extracted from text or tables or computed with reported percentages or numbers. We pooled results only on factors reported in at least three studies. For better readability, predictors have been grouped into demographic-, treatment-, comorbidity-, morphologic-, and postoperative-related factors. To minimize risk factor measurement bias, only studies with objective categorization and clear reporting were included in the analysis of individual risk factors. Statistical analysis. Study effects were pooled using a random-effects model, and forest plots were generated for every such factor to have a better understanding of the influence of the included studies. Cohen k statistics was used to evaluate the inter-rater reliability for study eligibility, and intraclass correlations were employed to evaluate inter-reviewer reliability for data collection. Because of the high variability of included
2017
studies in terms of outcome definition, sample size, or duration of follow-up, a deep sensitivity analysis of methodologic factors was performed to study all potential biases. In particular, publication bias was visually assessed by a funnel plot and statistically assessed by Egger test6; attrition bias was investigated by analyzing duration of follow-up; performance bias was evaluated by analyzing the median year of study recruitment; and selection biases were assessed by analyzing study design, sample size, mean age, and mean preoperative AAA diameter.
RESULTS Study selection and characteristics. The online database search identified 1145 records, and 38 additional records were identified from review of the included studies following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology (Fig 1). We excluded 1041 studies after reviewing abstracts and 42 after reviewing full text; 24 studies finally met our full inclusion criteria.1,2,4,7-27 Population characteristics are shown in Fig 2. No prospective study was identified, whereas seven were multicenter studies. A total of 14,754 patients (range, 49-8638) with an average followup of 24 months (range, 6-54.8 months) were included in the analysis, with a mean age of 73.4 years (range, 69.2-79.9 years), a majority of men (76%), and a mean preoperative AAA diameter of 56.4 cm (range, 47.7-62 cm). The overall pooled sac shrinkage proportion was 47% (95% confidence interval [CI], 38%-56%). Based on available data of stent graft models, 32.8% of patients were implanted with a Zenith (Cook Medical, Bloomington, Ind), 14.7% with a Talent (Medtronic, Santa Rosa, Calif), 17.8% with an AneuRx (Medtronic), 13% with a Vanguard (Boston Scientific, Marlborough, Mass), 11.9% with a Gore Excluder (W. L. Gore & Associates, Flagstaff, Ariz), and 9.8% with other types of stent grafts. Data extraction and quality assessment. There was a correct strength of agreement for the study eligibility (k ¼ 0.76; 95% CI, 0.68-0.99). Divergence in study eligibility was settled by consensus between reviewers. Overall, 94% of reported ORs were computed from univariate analysis and only 6% from multivariate analysis (linear or logistic multiple regression models). The mean Newcastle-Ottawa Scale score was 7.3, with a minimum of 6 and a maximum of 9. The sensitivity analysis did not identify any methodologic factors that influence sac shrinkage incidence, being selection, attrition, or performance bias (Fig 3). There was also no evidence of publication bias based on the Egger test (P ¼ .10; Fig 4). Influencing factors. A total of 17 factors related to demography, treatment, comorbidity, morphology, and postoperative events were statistically pooled. Fig 5 summarizes results for all factors. Based on pooled ORs, renal impairment (OR, 0.74; 95% CI, 0.57-0.96), type I
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
-
Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of the selection process. Excluded studies can have multiple exclusion criteria. EVAR, Endovascular aneurysm repair; T2Es, type II endoleaks; TEVAR, thoracic endovascular aortic repair.
Fig 2. Clinical and methodologic characteristics of included studies. The forest plot of shrinkage proportion is shown on the right. Proportions are shown with 95% confidence interval (CI). AAA, Abdominal aortic aneurysm; NOS, Newcastle-Ottawa Scale.
3
4
Journal of Vascular Surgery
Lalys et al
---
2017
Fig 3. Sensitivity analysis of methodologic factors. Pooled proportions of sac shrinkage are shown in percentage. AAA, Abdominal aortic aneurysm.
Fig 4. Funnel plot of effect size against sample size for all included studies. Three colors are used to differentiate sac shrinkage definitions. CI, Confidence interval; SD, standard deviation.
endoleaks (OR, 0.17; 95% CI, 0.08-0.39; I2, 0.90), type II endoleaks (OR, 0.21; 95% CI, 0.14-0.33), and combined type I and type II endoleaks (OR, 0.32; 95% CI, 0.22-0.47) were found to significantly prevent sac shrinkage, whereas hypercholesterolemia (OR, 1.24; 95% CI, 1.02-1.51) and smoking (OR, 1.32; 95% CI, 1.17-1.49) were found to have a positive impact on sac shrinkage. From these factors, only pooled ORs from type II endoleaks showed high heterogeneity (I2, 0.03). There was also a trend toward the association between shrinkage and statin therapy (OR, 4.07; 95% CI, 1.02-16.32) despite the high statistical heterogeneity (I2, 0.011). Coronary artery disease (OR, 0.84; 95% CI, 0.70-1.01), diabetes (OR, 0.79; 95% CI, 0.60-1.04), and sac thrombus (OR, 0.88; 95% CI,
0.77-1.01) were found to have a nearly significant negative impact on sac shrinkage. All others factors were found not to influence sac shrinkage.
DISCUSSION For patients with AAA who have suitable anatomy, EVAR is more and more considered the “gold standard” rather than open repair, and regression of the aneurysm sac has been used as a surrogate marker for technical success. However, persistent sac shrinkage is preferred because of the dynamic behavior of the aneurysmal shrinkage process. Specifically, Lee et al28 demonstrated that a reliable predictor of success is a volume reduction of 10% or more at 6 months and continuing regression
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
5
-
Fig 5. Pooled odds ratios (ORs) of sac shrinkage for identified demographic, treatment, comorbidity, morphologic, and postoperative factors. Number of included studies, dedicated number of patients, associated P value, and heterogeneity test (I2) are also shown. Pooled ORs are reported with 95% confidence interval (CI). AAA, Abdominal aortic aneurysm.
over time. Bastos Gonçalves et al4 later found similar results with a larger contemporary series. In that study, patients were divided in three groups: minor, major sac regression, and no sac regression. This last group included only 2.3% of patients with sac expansion (>5 mm) and consisted therefore of the majority of patients with “sac stabilization.” Results of the multivariate analysis showed that risk of late complications increased by 3.1 times for patients without sac regression (ie, almost only patients with sac stabilization) compared with major shrinkage (>10 mm), suggesting that sac regression (especially >10 mm 1 year after device implantation) is a strong predictor of long-lasting success after EVAR, probably rather than sac stabilization. Houbballah et al1 and Bisdas et al5 showed that AAA shrinkage >5 mm after EVAR was also associated with statistically significantly longer survival. Based on the present random-effects meta-analysis, several independent factors have been clearly identified. Endoleaks are well-recognized factors of sac enlargement29; it was therefore consistent to find that the absence of type I endoleaks, type II endoleaks, or combined type I and type II endoleaks can facilitate sac shrinkage. Paradoxically, two cardiovascular risks factors (hypercholesterolemia and smoking) showed a positive effect on shrinkage. According to the European Collaborators on Stent/graft Techniques for aortic Aneurysm Repair (EUROSTAR) registry,30 smokers seem to be less exposed to late endograft-related complications, especially to endoleaks. This result was confirmed in the latest study12 and could explain why smokers are more likely to present with sac regression. Other associations found in the present meta-analysis should be interpreted with
caution because the majority of the included studies are retrospective. Actually, recurrent factors are identified in the included studies, and we can only make a hypothesis to explain such results. Even if aneurysmal calcifications are rarely reported, it is acknowledged that patients with renal impairment are more likely to present with arterial calcification, including the aneurysmal sac. Therefore, arterial compliance is decreasing,31 and it would be reasonable to expect that the regression capacity of an aneurysm is partly dependent on these factors in the absence of any endoleaks. Although it is not systematically associated, patients suffering from hypercholesterolemia are often treated with statin therapy. As found in this meta-analysis, statin therapy is also associated with aneurysmal sac regression. Raux et al13 tried to explain how statin therapy may have a role in sac regression through the reduction of the inflammatory matrix metalloproteinases in the aortic wall particularly. It was reported in several publications12,30 that tobacco had paradoxically a protective effect on the occurrence of type II endoleaks, probably through the prothrombogenic effect of tobacco, which can explain why aneurysmal regression is more likely to occur in smokers. Whereas the natural history of AAA sac evolution after EVAR in the presence of a type II endoleak remains to be defined, this meta-analysis clearly identified the absence of endoleaks of type I, type II, or combination of both as a strong predictor of sac shrinkage and by definition of technical success. However, because sac shrinkage developed in some patients over time in spite of the presence of type II endoleaks, not all type II endoleaks have an impact on sac change. In the metaregression by Karthikesalingam et al,32 there was no
6
Lalys et al
Journal of Vascular Surgery ---
2017
Fig 6. Number of studies quantifying influence of continuous factors on sac shrinkage. AAA, Abdominal aortic aneurysm.
evidence that any strategy for isolated type II endoleak intervention improved sac regression compared with using a conservative approach. It is in contradiction to the results of this meta-analysis. Other factors have not been included in the metaanalysis. Influence of continuous factors on sac shrinkage is, for instance, extremely difficult to review. Because of the high heterogeneity in the definition of those factors across studies, we did not succeed in finding enough studies with available ORs and similar definition. A count of studies reporting P values of different continuous factors is proposed, however, in Fig 6. For age, proximal neck length, and proximal neck diameter, no consensus was found. For the preoperative maximum AAA diameter, contradictory results have been published (nine studies describing nonassociation vs eight studies showing significant association), and further studies will be needed on a large cohort to determine its real impact. We also did not include device-specific clinical outcomes, despite differences found in incidence of device migration, frequency of endoleaks, or risk of limb thrombosis. In particular, early results have shown that sac size change after EVAR is device specific.33,34 Different trials have also indicated that rates of sac expansion or regression were correlated to graft models.2,26 However, for the majority of them, these studies compared models of the oldest generation of stent grafts and the next one. These differences between first and second generation seem to be less important with the latest generation. Two recent studies have compared the last two generations of stent grafts,35,36 and it seems that there is no major difference between them in AAA-related outcomes (sac behavior, endoleaks). Both studies have shown that the most recent generation allowed more challenging anatomic cases to be treated. Since these studies have been reported, new endografts are available, and our potential conclusion about the difference between endografts would have not been relevant because the models of endografts in the meta-analysis are no longer used. A large number of factors have not been included in this meta-analysis because of lack of comparison data.
Within clinical or biologic factors, preoperative C-reactive protein level and atherosclerosis appear to be associated with an increased likelihood of sac regression,37 as are calcium channel blockers,15 angiotensin-converting enzyme inhibitors,15 and tranexamic acid.38 Chikazawa et al39 found that an additional thoracic aortic aneurysm had a negative impact on sac shrinkage. In morphology-related factors, the absences of calcification and thrombus at the level of the iliac arteries and aortic neck have been identified as significant factors affecting aneurysm shrinkage.14 Neck thrombus or plaque was found to be significant,40 despite being nonsignificant in another study.7 Overall increased proportion of thrombus on preoperative CT also resulted in greater likelihood of sac shrinkage.41 Whereas patencyrelated factors have been mostly highlighted in predictive models of endoleaks, some studies made the association with sac shrinkage. Sheehan et al42 and Kray et al43 found a significant association with patent inferior mesenteric artery and coiling of the lumbar or inferior mesenteric artery, whereas Back et al27 identified the overall patency of aortic side branches as a significant predictor, contrary to Torsello et al.22 Bastos Gonçalves et al44 significantly stratified their cohort into a low-risk group (with proximal and distal seal length of 10 mm and no endoleak) and a high-risk group (with insufficient seal, presence of endoleak, or both). The mean diameter of the infrarenal aorta,1 endograft oversizing,45 patients without postoperative migration,46 mean pressure index,47 and effective seal length and surface area of circumferential apposition between the aorta and the graft material48 are other factors of interest that will need further consideration. Finally, shrinkage rates seem to be different among hospitals, as shown using the same endograft and protocol-defined patient selection criteria.49 Identification of influencing factors is also a first step toward predictive models. Preoperative prediction of complications or success is critical to help physicians in decision-making, to inform patients, and to adjust follow-up surveillance. As we develop a better understanding of expected aneurysmal sac evolution, similar
Journal of Vascular Surgery Volume
-,
Number
Lalys et al
7
-
predictive models could help preoperatively stratify patients and tailor postoperative surveillance to the expected aneurysmal sac evolution. On the one hand, patients having a low likelihood of shrinking of their sac might be at high risk of complications and thus require intensive and long-term surveillance. It may be necessary to monitor those patients carefully through continued imaging, whereas on the contrary for patients with a high likelihood of sac regression, imaging follow-up intensiveness might be reduced. According to Soler et al,7 late failures might still occur in patients with significant sac shrinkage, so even for those patients, lifelong follow-up should be provided. To improve surveillance, another potential surveillance technique is the serial assessment of intrasac pressure. It has been shown to be feasible50 and strongly correlated with type II endoleaks,47 and it could help the decision-making process related to reintervention after EVAR. However, further studies are still needed to validate its utility and reproducibility. Whereas sac shrinkage is desired and remains a key factor, another important consideration is the morphologic change in length and angulation introduced by the shrinkage. Morphologic changes are regularly evaluated by using serial CT angiography to identify complications, and information about the sac diameter or volume change can be retrieved. In particular, studies have shown that volume measurements should be performed in addition to diameter measurements on CT angiography51 for better robustness. Those changes may have an impact on attachment of the stent graft components and overall graft shape, which can be the cause of other complications. This study has some limitations. Even if the sensitivity analysis did not find any evidence of bias, and even if the study quality assessment showed that all included papers were methodologically robust, slight clinical heterogeneity, arising from differences in study populations and consequential distribution risk factors, may be present. A large range of EVAR experience, hospital routines, and complication definitions may also hamper the generalizability of the results. Obviously, we do not recommend to our patients to smoke and stop therapy against hypercholesterolemia. For example, diabetes is a well-recognized protective factor against aneurysm rupture (not a statistical error), and nobody says that diabetes should not be controlled for in patients presenting with AAAs. Currently, data are not strong enough to determine if these factors are a statistical finding or have a real impact on aneurysmal sac behavior after EVAR.
CONCLUSIONS As our current understanding of influencing factors of sac shrinkage after EVAR is weak, this meta-analysis aimed to determine significant factors to identify in
which patients sac shrinkage is more likely to be high. Whereas diabetes, coronary artery disease, renal impairment, sac thrombus, and type I and type II endoleaks seem to prevent sac shrinkage, patients receiving statin therapy, patients with hypercholesterolemia, and active smokers may have a better chance to observe sac regression and therefore to have a lower risk of reintervention.
AUTHOR CONTRIBUTIONS Conception and design: FL, CG, AL, AK Analysis and interpretation: FL, AK Data collection: FL, AD, JG Writing the article: FL, AD, JG, CG, AL, AK Critical revision of the article: FL, AD, JG, CG, AL, AK Final approval of the article: FL, AD, JG, CG, AL, AK Statistical analysis: FL Obtained funding: Not applicable Overall responsibility: FL
REFERENCES 1. Houbballah R, Majewski M, Becquemin JP. Significant sac retraction after endovascular aneurysm repair is a robust indicator of durable treatment success. J Vasc Surg 2010;52: 878-83. 2. Rhee RY, Eskandari MK, Zajko AB, Makaroun MS. Long-term fate of the aneurysmal sac after endoluminal exclusion of abdominal aortic aneurysms. J Vasc Surg 2000;32:689-96. 3. Chaikof EL, Blankensteijn JD, Harris PL, White GH, Zarins CK, Bernhard VM, et al. Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2002;35:1048-60. 4. Bastos Gonçalves F, Baderkhan H, Verhagen HJ, Wanhainen A, Björck M, Stolker RJ, et al. Early sac shrinkage predicts a low risk of late complications after endovascular aortic aneurysm repair. Br J Surg 2014;101:802-10. 5. Bisdas T, Weiss K, Eisenack M, Austermann M, Torsello G, Donas KP. Durability of the Endurant stent graft in patients undergoing endovascular abdominal aortic aneurysm repair. J Vasc Surg 2014;60:1125-31. 6. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34. 7. Soler RJ, Bartoli MA, Mancini J, Lerussi G, Thevenin B, SarlonBartoli G, et al. Aneurysm sac shrinkage after endovascular repair: predictive factors and long-term follow-up. Ann Vasc Surg 2015;29:770-9. 8. Pini R, Faggioli G, Mascoli C, Gallitto E, Freyrie A, Gargiulo M, et al. Influence of statin therapy on type 2 endoleak evolution. Ann Vasc Surg 2015;29:1167-73. 9. Nakai M, Ikoma A, Sato H, Sato M, Nishimura Y, Okamura Y. Risk factors associated with late aneurysmal sac expansion after endovascular abdominal aortic aneurysm repair. Diagn Interv Radiol 2015;21:195-201. 10. Tsuyuki Y, Matsushita S, Dohi S, Yamamoto T, Tambara K, Inaba H, et al. Factors for sac size change of abdominal aortic aneurysm after endovascular repair. Ann Thorac Cardiovasc Surg 2014;20:1016-20. 11. Cieri E, De Rango P, Isernia G, Simonte G, Verzini F, Parlani G, et al. Effect of stentgraft model on aneurysm shrinkage in 1, 450 endovascular aortic repairs. Eur J Vasc Endovasc Surg 2013;46:192-200. 12. Koole D, Moll FL, Buth J, Hobo R, Zandvoort H, Pasterkamp G, et al. The influence of smoking on endovascular abdominal aortic aneurysm repair. J Vasc Surg 2012;55:1581-6.
8
Lalys et al
Journal of Vascular Surgery ---
13. Raux M, Cochennec F, Becquemin JP. Statin therapy is associated with aneurysm sac regression after endovascular aortic repair. J Vasc Surg 2012;55:1587-92. 14. Kaladji A, Cardon A, Abouliatim I, Campillo-Gimenez B, Heautot JF, Verhoye JP. Preoperative predictive factors of aneurysmal regression using the reporting standards for endovascular aortic aneurysm repair. J Vasc Surg 2012;55: 1287-95. 15. Bailey MA, Sohrabi S, Flood K, Griffin KJ, Rashid ST, Johnson AB, et al. Calcium channel blockers enhance sac shrinkage after endovascular aneurysm repair. J Vasc Surg 2012;55:1593-9. 16. Hogg ME, Morasch MD, Park T, Flannery WD, Makaroun MS, Cho JS. Long-term sac behavior after endovascular abdominal aortic aneurysm repair with the Excluder lowpermeability endoprosthesis. J Vasc Surg 2011;53:1178-83. 17. El Sayed HF, Meier GH, Mendoza B, Sprouse LR, Parent FN, Panneton JM. Aneurysm regression after endovascular aneurysm repair: what should we expect? Vasc Endovascular Surg 2009;42:545-50. 18. Becquemin JP, Aksoy M, Marzelle J, Roudot-Thoraval F, Desgranges P, Allaire E, et al. Abdominal aortic aneurysm sac behavior following Cook Zenith graft implantation: a fiveyear follow-up assessment of 212 cases. J Cardiovasc Surg (Torino) 2008;49:199-206. 19. Broker HS, Foteh KI, Murphy EH, Davis CM, Clagett GP, Modrall JG, et al. Device-specific aneurysm sac morphology after endovascular aneurysm repair: evaluation of contemporary graft materials. J Vasc Surg 2008;47:702-6; discussion: 707. 20. Higashiura W, Greenberg RK, Katz E, Geiger L, Bathurst S. Predictive factors, morphologic effects, and proposed treatment paradigm for type II endoleaks after repair of infrarenal abdominal aortic aneurysms. J Vasc Interv Radiol 2007;18: 975-81. 21. Golledge J, Parr A, Boult M, Maddern G, Fitridge R. The outcome of endovascular repair of small abdominal aortic aneurysms. Ann Surg 2007;245:326-33. 22. Torsello G, Osada N, Florek HJ, Horsch S, Kortmann H, Luska G, et al. Long-term outcome after Talent endograft implantation for aneurysms of the abdominal aorta: a multicenter retrospective study. J Vasc Surg 2006;43:277-84; discussion: 284. 23. Biebl M, Hakaim AG, Oldenburg WA, Klocker J, Lau LL, Neuhauser B, et al. Does chronic oral anticoagulation with warfarin affect durability of endovascular aortic aneurysm exclusion in a midterm follow-up? J Endovasc Ther 2005;12: 58-65. 24. Zarins CK, Bloch DA, Crabtree T, Matsumoto AH, White RA, Fogarty TJ. Aneurysm enlargement following endovascular aneurysm repair: AneuRx clinical trial. J Vasc Surg 2004;39: 109-17. 25. Sampaio SM, Panneton JM, Mozes GI, Andrews JC, Bower TC, Karla M, et al. Proximal type I endoleak after endovascular abdominal aortic aneurysm repair: predictive factors. Ann Vasc Surg 2004;18:621-8. 26. Sternbergh WC, Conners MS, Tonnessen BH, Carter G, Money SR. Aortic aneurysm sac shrinkage after endovascular repair is deviceedependent: a comparison of Zenith and AneuRx endografts. Ann Vasc Surg 2003;17: 49-53. 27. Back MR, Bowser AN, Johnson BL, Schmacht D, Zwiebel B, Bandyk DF. Patency of infrarenal aortic side branches determines early aneurysm sac behavior after endovascular repair. Ann Vasc Surg 2003;17:27-34. 28. Lee JT, Aziz IN, Lee JT, Haukoos JS, Donayre CE, Walot I, et al. Volume regression of abdominal aortic aneurysms and its
2017
relation to successful endoluminal exclusion. J Vasc Surg 2003;38:1254-63. 29. Schanzer A, Greenberg RK, Hevelone N, Robinson WP, Eslami MH, Goldberg RJ, et al. Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair. Circulation 2011;123:2848-55. 30. Lottman PE, Van Marrewijk CJ, Fransen GA, Laheij RJ, Buth J. Impact of smoking on endovascular abdominal aortic aneurysm surgery outcome. Eur J Vasc Endovasc Surg 2004;27:512-8. 31. Speelman L, Bohra A, Bosboom EM, Schurink GW, van de Vosse FN, Makaroun MS, et al. Effects of wall calcifications in patient-specific wall stress analyses of abdominal aortic aneurysms. J Biomech Eng 2007;129:105-9. 32. Karthikesalingam A, Thrumurthy SG, Jackson D, Choke E, Sayers RD, Loftus IM, et al. Current evidence is insufficient to define an optimal threshold for intervention in isolated type II endoleak after endovascular aneurysm repair. J Endovasc Ther 2012;19:200-8. 33. Ouriel K, Clair DG, Greenberg RK, Lyden SP, O’Hara PJ, Sarac TP, et al. Endovascular repair of abdominal aortic aneurysms: device-specific outcome. J Vasc Surg 2003;37:991-8. 34. Bertges DJ, Chow K, Wyers MC, Landsittel D, Frydrych AV, Stavropoulos W, et al. Abdominal aortic aneurysm size regression after endovascular repair is endograft dependent. J Vasc Surg 2003;37:716-23. 35. Sobocinski J, Briffa F, Holt PJ, Martin Gonzalez T, Spear R, Azzaoui R, et al. Evaluation of the Zenith low-profile abdominal aortic aneurysm stent graft. J Vasc Surg 2015;62: 841-7. 36. Maudet A, Daoudal A, Cardon A, Clochard E, Lucas A, Verhoye JP, et al. Endovascular treatment of infrarenal aneurysms: comparison of the results of second- and thirdgeneration stent grafts. Ann Vasc Surg 2016;34:95-105. 37. Ishibashi H, Ishiguchi T, Ohta T, Sugimoto I, Iwata H, Yamada T, et al. Mid-term results of endovascular abdominal aortic aneurysm repair: is it possible to predict sac shrinkage? Surg Today 2011;41:1605-9. 38. Aoki A, Suezawa T, Yamamoto S, Sangawa K, Irie H, Mayazaki N, et al. Effect of antifibrinolytic therapy with tranexamic acid on abdominal aortic aneurysm shrinkage after endovascular repair. J Vasc Surg 2014;59:1203-8. 39. Chikazawa G, Hiraoka A, Totsugawa T, Tamura K, Ishida A, Sakaguchi T, et al. Influencing factors for abdominal aortic aneurysm sac shrinkage and enlargement after EVAR: clinical reviews before introduction of preoperative coil embolization. Ann Vasc Dis 2014;7:280-5. 40. Fairman RM, Nolte L, Snyder SA, Chuter TA, Greenberg RK; Zenith Investigators. Factors predictive of early or late aneurysm sac size change following endovascular repair. J Vasc Surg 2006;43:649-56. 41. Sadek M, Dexter DJ, Rockman CB, Hoang H, Mussa FF, Cayne NS, et al. Preoperative relative abdominal aortic aneurysm thrombus burden predicts endoleak and sac enlargement after endovascular aneurysm repair. Ann Vasc Surg 2013;27:1036-41. 42. Sheehan MK, Hagino RT, Canby E, Wholey MH, Postoak D, Suri R, et al. Type 2 endoleaks after abdominal aortic aneurysm stent grafting with systematic mesenteric and lumbar coil embolization. Ann Vasc Surg 2006;20:458-63. 43. Kray J, Kirk S, Franko J, Chew DK. Role of type II endoleak in sac regression after endovascular repair of infrarenal abdominal aortic aneurysms. J Vasc Surg 2015;61:869-74. 44. Bastos Gonçalves F, van de Luijtgaarden KM, Hoeks SE, Hendriks JM, ten Raa S, Rouwet EV, et al. Adequate seal and no endoleak on the first postoperative computed tomography angiography as criteria for no additional imaging up to
Journal of Vascular Surgery Volume
45.
46.
47.
48.
-,
Number
Lalys et al
9
-
5 years after endovascular aneurysm repair. J Vasc Surg 2013;57:1503-11. Sternbergh WC, Money SR, Greenberg RK, Chuter TA; Zenith Investigators. Influence of endograft oversizing on device migration, endoleak, aneurysm shrinkage, and aortic neck dilation: results from the Zenith Multicenter Trial. J Vasc Surg 2004;39:20-6. Lee JT, Lee J, Aziz I, Donayre CE, Walot I, Kopchok GE, et al. Stent-graft migration following endovascular repair of aneurysms with large proximal necks: anatomical risk factors and long-term sequelae. J Endovasc Ther 2002;9:652-64. Dias NV, Ivancev K, Resch TA, Malina M, Sonesson B. Endoleaks after endovascular aneurysm repair lead to nonuniform intra-aneurysm sac pressure. J Vasc Surg 2007;46:197-203. Arthurs ZM, Ouriel K, Welborn MB III, McDaniel HB. Extended proximal seal zone during EVAR predicts aneurysm sac regression. J Vasc Surg 2014;60:546.
49. Kritpracha B, Beebe HG, Criado FJ, Comerota AJ. Postendograft abdominal aortic aneurysm shrinkage varies among hospitals: observations from multicenter trials. J Endovasc Ther 2004;11:454-9. 50. Ellozy SH, Carroccio A, Lookstein RA, Jacobs TS, Addis MD, Teodorescu VJ, et al. Abdominal aortic aneurysm sac shrinkage after endovascular aneurysm repair: correlation with chronic sac pressure measurement. J Vasc Surg 2006;43:2-7. 51. van Keulen JW, van Prehn J, Prokop M, Moll FL, van Herwaarden JA. Potential value of aneurysm sac volume measurements in addition to diameter measurements after endovascular aneurysm repair. J Endovasc Ther 2009;16: 506-13.
Submitted Aug 2, 2016; accepted Dec 16, 2016.