Aortoiliac Elongation after Endovascular Aortic Aneurysm Repair Venita Chandra,1 Martin Rouer,1,1 Trit Garg,1 Dominik Fleischmann,2 and Matthew Mell,1 Stanford, California
Background: Aortoiliac elongation after endovascular aortic aneurysm repair (EVAR) is not well studied. We sought to assess the long-term morphologic changes after EVAR and identify potentially modifiable factors associated with such a change. Methods: An institutional review boardeapproved retrospective review was conducted for 88 consecutive patients who underwent EVAR at a single academic center from 2003 to 2007 and who also had at least 2 follow-up computed tomography angiograms (CTAs) available for review up to 5 years after surgery. Standardized centerline aortic lengths and diameters were obtained on Aquarius iNtuition 3D workstation (TeraRecon Inc., San Mateo, CA) on postoperative and all-available follow-up CTAs. Relationships to aortic elongation were determined using Wilcoxon rank-sum test or linear regression (Stata version 12.1, College Station, TX). Changes in length over time were determined by mixed-effects analysis (SAS version 9.3, Cary, NC). Results: The study cohort was composed of mostly men (88%), with a mean age of (76 ± 8) and a mean follow-up of 3.2 years (range, 0.4e7.5 years). Fifty-seven percent of patients (n ¼ 50) had devices with suprarenal fixation and 43% (n ¼ 38) had no suprarenal fixation. Significant lengthening was observed over the study period in the aortoiliac segments, but not in the iliofemoral segments. Aortoiliac elongation over time was not associated with sex (P ¼ 0.3), hypertension (P ¼ 0.7), coronary artery disease (P ¼ 0.3), diabetes (P ¼ 0.3), or tobacco use (P ¼ 0.4), but was associated with the use of statins (P ¼ 0.03) and the presence of chronic obstructive pulmonary disease (P ¼ 0.02). Significant aortic lengthening was associated with increased type I endoleaks (P ¼ 0.03) and reinterventions (P ¼ 0.03). Over the study period, 4 different devices were used; Zenith (Cook Medical Inc., Bloomington, IN), Talent (Medtronic, Minneapolis, MN), Aneuryx (Medtronic), and Excluder (W. L. Gore and Associates Inc., Flagstaff, AZ). After adjusting for differences in proximal landing zone, significant differences in aortic lengthening over time were observed by device type (P ¼ 0.02). Conclusions: Significant aortoiliac elongation was observed after EVAR. Such morphologic changes may impact long-term durability of EVAR, warranting further investigation into factors associated with these morphologic changes.
1
Division of Vascular Surgery, Stanford University Medical Center, Stanford, CA. 2
Division of Radiology, Stanford University Medical Center, Stanford, CA. 1
Present Address: Inserm UMRS 698-Paris VII University, H^opital Xavier Bichat, Paris, France. Correspondence to: Venita Chandra, MD, Division of Vascular Surgery, Stanford University Medical Center, 300 Pasteur Drive, Suite H3600, Stanford, CA 94305, USA; E-mail: [email protected] Ann Vasc Surg 2015; 29: 891–897 http://dx.doi.org/10.1016/j.avsg.2014.12.041 Ó 2015 Elsevier Inc. All rights reserved. Manuscript received: August 3, 2014; manuscript accepted: December 17, 2014; published online: March 7, 2015.
INTRODUCTION Endovascular abdominal aortic aneurysm repair (EVAR) has become the preferred technique for repair of abdominal aortic aneurysm (AAA) given the shorter hospital stays and decreased perioperative morbidity and mortality compared with open surgical repair.1e5 Durability of these increasingly complex endovascular aortic interventions can be impacted by dimensional or morphologic changes in the aorta. Accordingly, patients who undergo EVAR require regular postoperative imaging and follow-up to monitor for signs of stent-graft distortion 891
892 Chandra et al.
or disruption (including fraction or occlusion), device migration, endoleak development, aneurysm growth, and potentially, even rupture.6e10 Given these risks, it is not surprising that there has been substantial focus on the morphologic changes that occur in the aorta and within an aneurysm after EVAR. Aortic diameter has been studied extensively, and aortic neck dilation has been documented to occur frequently and to be associated with increased stent-graft migration and endoleak formation.11e15 Although many studies focused on stent-graft migration in relation to aortic diameter changes, far fewer have looked at aortic length changes, although one could argue that length may be an equally important morphologic variable.7,11 Initially the thought was that there might be decreases in aortic length after EVAR given the decreasing aneurysmal diameter often seen after EVAR. This was not, however, found to be the case.7,12,13 In contrary to this original belief, there is some biological and natural history evidence that the aorta likely lengthens over time.14e17 However, there have been no comprehensive studies assessing the overall abdominal aortic length over time after EVAR. The purpose of our study was to take advantage of the postoperative monitoring in EVAR patients to better determine the morphologic changes that occur in the aorta, primarily focusing on changes in aortic length over time. In addition, we sought to determine the impact of such aortic morphologic changes on clinically significant parameters such as the formation of endoleaks and the need for reintervention.
METHODS We conducted an institutional review boarde approved retrospective review of patients who underwent EVAR at a single academic institution between 2004 and 2007. We chose this time frame to maximize the number of postoperative images available for review and to measure changes over longitudinal follow-up. Patients who had a postoperative computed tomography angiogram (CTA) (with at least 1-mm thick cuts) and at least 2 follow-up CTAs available for review were included in the study. Patient characteristics including age, sex, comorbidities, and medications along with perioperative details and outcomes including devices used, endoleak rates, and reinterventions were recorded and analyzed for all patients. Reference CTAs were those obtained within 1 month after EVAR placement. Serial follow-up CTAs were then reviewed. Each CTA was analyzed
Annals of Vascular Surgery
by 2 independent reviewers. Standardized centerline aortic lengths and diameters were obtained on Aquarius iNtuition 3D workstations (Terarecon Inc., San Mateo, CA). Both reviewers were vascular surgeons with extensive experience with Aquarius iNtuition and centerline measurements. Measurements included maximum aortic diameter and the following lengths: superior mesenteric artery (SMA) to top of stent, SMA to flow divider, SMA to right and left iliac bifurcations, top of stent to right and left iliac bifurcations, and right and left iliac bifurcations to right and left femoral bifurcations. Statistical Analysis Interobserver concordance was assessed with correlation coefficients, coefficients of concordance, and a BlandeAltman plot. Pearson correlation coefficients were performed to analyze the relationship between aortic diameter changes and total aortoiliac elongation. For most of the statistical analysis, the primary length measurement used was the distance from the SMA to the iliac bifurcations, as this value was not affected by device placement. The relationship of variables analyzed to differences in aortic length was determined using Wilcoxon rank-sum test or linear regression. Statistical analyses were performed using Stata (version 12.1, College Station, TX).
RESULTS During the study period, 341 consecutive patients underwent EVAR. Among them, 88 met the selection criteria and made up the study cohort. Demographics and operative and device details of the study cohort are shown in Table I. Fifty-seven percent of the patients (n ¼ 50) had devices placed with suprarenal fixation and 43% (n ¼ 38) had no suprarenal fixation. Two patients had aortouniiliac devices placed, and 12 patients (12.6%) underwent embolization of one or both internal iliac arteries during the index procedure with extension of the device limb into the external iliac artery. Mean follow-up was 3.2 years (range, 0.4e 7.5 years). Of the 88 patients, 5 (5.7%) had open surgical conversion (1 at 1 year, 2 at 2 years, and 2 at 4 years), and 10 patients (11.4%) had proximal cuff deployment. Ten patients (11.4%) developed type I endoleaks during follow-up, 4 of which already had the type I endoleak at the time of the reference CTA. Thirty-seven (42%) patients developed type II endoleaks, the vast majority of which (34 of 37) were present on reference CTA. No patients were diagnosed with type III endoleaks during
Vol. 29, No. 5, July 2015
Aortoiliac elongation after EVAR 893
Table I. Patient demographics and operative characteristics Variable
N ¼ 88, n (%)
Age, years Male HTN DM HLD CAD TOB COPD Suprarenal fixation Medtronic Talent Cook Zenith No suprarenal fixation Medtronic Aneuryx Gore Excluder Mean follow-up, years
76 77 62 12 53 29 75 26 50 37 13 33 29 9 3.2
(±8) (88) (70) (14) (60) (33) (85) (30) (57) (42) (15) (43) (33) (10) (0.4e7.5)
HTN, hypertension; DM, diabetes mellitus; CAD, coronary artery disease; HLD, hyperlipidemia; TOB, tobacco use.
the study period. Twenty-nine patients (33%) underwent some form of reintervention at our institution during the follow-up period, 17 (58.6%) within the first 3 years after EVAR. Interobserver concordance between reviewers of the centerline-based measurements is depicted in Figure 1; 91.8% of the measurements fell within the 95% limits of the agreement. The concordance correlation coefficient was 0.980, and the coefficient of variation was 1.5%. Table II lists aortic measurements at the reference CTA in comparison with the last available CTA. Figure 2 depicts these findings graphically. When looking at the mean differences in aortic lengths at the various measurement points over time, increases in length are seen in the aortoiliac segments but not in the iliofemoral regions. These changes in length were statistically significant beginning at 3 years. Aortic diameters did not demonstrate significant change between baseline measurements and that of the last follow-up image, having decreased and then increased over the follow-up period. Conversely, aortoiliac lengths gradually increased over time (Fig. 3). When comparing the changes in aortic lengths with changes in aortic diameter, we found a modest correlation with a Pearson correlation coefficient of 0.34. Univariate analysis was performed to identify patient and operative variables associated with aortic elongation. Table III lists patient demographics and their association with aortic length. There was a trend to increased lengthening in men compared with women, although this was not statistically
Fig. 1. BlandeAltman interobserver agreement plot. Concordance correlation coefficient, 0.980 and coefficient of variation, 1.5%; 91.8% within the 95% limits of agreement.
significant. The presence of chronic obstructive pulmonary disease (COPD) and the use of statins appeared to have a protective effect on aortic lengthening. No other patient demographics were associated with aortic lengthening. When looking at operative variables, we found that aortic elongation over time was inversely related to proximal landing zone length (P ¼ 0.007). We did not find a significant difference in aortic lengthening when comparing devices with suprarenal fixation to no suprarenal fixation (P ¼ 0.7). Over the study period, 4 different devices were used (Table IV); Zenith (Cook Medical Inc., Bloomington, IN), Talent (Medtronic, Minneapolis, MN), Aneuryx (Medtronic), and Excluder (W. L. Gore and Associates Inc., Flagstaff, AZ). After adjusting for differences in proximal landing zone, significant differences in aortic lengthening over time were, however, observed by device type (P ¼ 0.02). This difference remained significant (P ¼ 0.04) after adjusting for reinterventions. Table V demonstrates the relationship between aortic lengthening and endoleaks or reinterventions. We found that aortic lengthening was significantly increased in patients who developed type I endoleaks compared with those who did not. In addition, those who underwent reinterventions during follow-up had more elongation that those who did not. There was no significant increase in aortic lengths in patients who developed type II endoleaks. Aortic diameter increase and aortic lengthening were both independently and positively associated with both increase rates of type I endoleaks and reinterventions by multivariable logistic regression analysis (Table VI). Every 5-mm increase in length was
894 Chandra et al.
Annals of Vascular Surgery
Table II. Aortic measurements
Aortic maximum diameter SMAetop of stent graft SMAeflow divider SMAeiliac bifurcation Top of stent grafteiliac bifurcation Iliac bifurcationefemoral bifurcation
Reference CTA, mean ± SD, mm
Last follow-up CTA, mean ± SD, mm
P value
57.4 11.0 67.9 212.5 201.5 164.2
56.0 17.5 76.4 221.4 205.4 163.4
0.25 0.001
Background: Aortoiliac elongation after endovascular aortic aneurysm repair (EVAR) is not well studied. We sought to assess the long-term morphologic changes after EVAR and identify potentially modifiable factors associated with such a change. Methods: An institutional review boardeapproved retrospective review was conducted for 88 consecutive patients who underwent EVAR at a single academic center from 2003 to 2007 and who also had at least 2 follow-up computed tomography angiograms (CTAs) available for review up to 5 years after surgery. Standardized centerline aortic lengths and diameters were obtained on Aquarius iNtuition 3D workstation (TeraRecon Inc., San Mateo, CA) on postoperative and all-available follow-up CTAs. Relationships to aortic elongation were determined using Wilcoxon rank-sum test or linear regression (Stata version 12.1, College Station, TX). Changes in length over time were determined by mixed-effects analysis (SAS version 9.3, Cary, NC). Results: The study cohort was composed of mostly men (88%), with a mean age of (76 ± 8) and a mean follow-up of 3.2 years (range, 0.4e7.5 years). Fifty-seven percent of patients (n ¼ 50) had devices with suprarenal fixation and 43% (n ¼ 38) had no suprarenal fixation. Significant lengthening was observed over the study period in the aortoiliac segments, but not in the iliofemoral segments. Aortoiliac elongation over time was not associated with sex (P ¼ 0.3), hypertension (P ¼ 0.7), coronary artery disease (P ¼ 0.3), diabetes (P ¼ 0.3), or tobacco use (P ¼ 0.4), but was associated with the use of statins (P ¼ 0.03) and the presence of chronic obstructive pulmonary disease (P ¼ 0.02). Significant aortic lengthening was associated with increased type I endoleaks (P ¼ 0.03) and reinterventions (P ¼ 0.03). Over the study period, 4 different devices were used; Zenith (Cook Medical Inc., Bloomington, IN), Talent (Medtronic, Minneapolis, MN), Aneuryx (Medtronic), and Excluder (W. L. Gore and Associates Inc., Flagstaff, AZ). After adjusting for differences in proximal landing zone, significant differences in aortic lengthening over time were observed by device type (P ¼ 0.02). Conclusions: Significant aortoiliac elongation was observed after EVAR. Such morphologic changes may impact long-term durability of EVAR, warranting further investigation into factors associated with these morphologic changes.
1
Division of Vascular Surgery, Stanford University Medical Center, Stanford, CA. 2
Division of Radiology, Stanford University Medical Center, Stanford, CA. 1
Present Address: Inserm UMRS 698-Paris VII University, H^opital Xavier Bichat, Paris, France. Correspondence to: Venita Chandra, MD, Division of Vascular Surgery, Stanford University Medical Center, 300 Pasteur Drive, Suite H3600, Stanford, CA 94305, USA; E-mail: [email protected] Ann Vasc Surg 2015; 29: 891–897 http://dx.doi.org/10.1016/j.avsg.2014.12.041 Ó 2015 Elsevier Inc. All rights reserved. Manuscript received: August 3, 2014; manuscript accepted: December 17, 2014; published online: March 7, 2015.
INTRODUCTION Endovascular abdominal aortic aneurysm repair (EVAR) has become the preferred technique for repair of abdominal aortic aneurysm (AAA) given the shorter hospital stays and decreased perioperative morbidity and mortality compared with open surgical repair.1e5 Durability of these increasingly complex endovascular aortic interventions can be impacted by dimensional or morphologic changes in the aorta. Accordingly, patients who undergo EVAR require regular postoperative imaging and follow-up to monitor for signs of stent-graft distortion 891
892 Chandra et al.
or disruption (including fraction or occlusion), device migration, endoleak development, aneurysm growth, and potentially, even rupture.6e10 Given these risks, it is not surprising that there has been substantial focus on the morphologic changes that occur in the aorta and within an aneurysm after EVAR. Aortic diameter has been studied extensively, and aortic neck dilation has been documented to occur frequently and to be associated with increased stent-graft migration and endoleak formation.11e15 Although many studies focused on stent-graft migration in relation to aortic diameter changes, far fewer have looked at aortic length changes, although one could argue that length may be an equally important morphologic variable.7,11 Initially the thought was that there might be decreases in aortic length after EVAR given the decreasing aneurysmal diameter often seen after EVAR. This was not, however, found to be the case.7,12,13 In contrary to this original belief, there is some biological and natural history evidence that the aorta likely lengthens over time.14e17 However, there have been no comprehensive studies assessing the overall abdominal aortic length over time after EVAR. The purpose of our study was to take advantage of the postoperative monitoring in EVAR patients to better determine the morphologic changes that occur in the aorta, primarily focusing on changes in aortic length over time. In addition, we sought to determine the impact of such aortic morphologic changes on clinically significant parameters such as the formation of endoleaks and the need for reintervention.
METHODS We conducted an institutional review boarde approved retrospective review of patients who underwent EVAR at a single academic institution between 2004 and 2007. We chose this time frame to maximize the number of postoperative images available for review and to measure changes over longitudinal follow-up. Patients who had a postoperative computed tomography angiogram (CTA) (with at least 1-mm thick cuts) and at least 2 follow-up CTAs available for review were included in the study. Patient characteristics including age, sex, comorbidities, and medications along with perioperative details and outcomes including devices used, endoleak rates, and reinterventions were recorded and analyzed for all patients. Reference CTAs were those obtained within 1 month after EVAR placement. Serial follow-up CTAs were then reviewed. Each CTA was analyzed
Annals of Vascular Surgery
by 2 independent reviewers. Standardized centerline aortic lengths and diameters were obtained on Aquarius iNtuition 3D workstations (Terarecon Inc., San Mateo, CA). Both reviewers were vascular surgeons with extensive experience with Aquarius iNtuition and centerline measurements. Measurements included maximum aortic diameter and the following lengths: superior mesenteric artery (SMA) to top of stent, SMA to flow divider, SMA to right and left iliac bifurcations, top of stent to right and left iliac bifurcations, and right and left iliac bifurcations to right and left femoral bifurcations. Statistical Analysis Interobserver concordance was assessed with correlation coefficients, coefficients of concordance, and a BlandeAltman plot. Pearson correlation coefficients were performed to analyze the relationship between aortic diameter changes and total aortoiliac elongation. For most of the statistical analysis, the primary length measurement used was the distance from the SMA to the iliac bifurcations, as this value was not affected by device placement. The relationship of variables analyzed to differences in aortic length was determined using Wilcoxon rank-sum test or linear regression. Statistical analyses were performed using Stata (version 12.1, College Station, TX).
RESULTS During the study period, 341 consecutive patients underwent EVAR. Among them, 88 met the selection criteria and made up the study cohort. Demographics and operative and device details of the study cohort are shown in Table I. Fifty-seven percent of the patients (n ¼ 50) had devices placed with suprarenal fixation and 43% (n ¼ 38) had no suprarenal fixation. Two patients had aortouniiliac devices placed, and 12 patients (12.6%) underwent embolization of one or both internal iliac arteries during the index procedure with extension of the device limb into the external iliac artery. Mean follow-up was 3.2 years (range, 0.4e 7.5 years). Of the 88 patients, 5 (5.7%) had open surgical conversion (1 at 1 year, 2 at 2 years, and 2 at 4 years), and 10 patients (11.4%) had proximal cuff deployment. Ten patients (11.4%) developed type I endoleaks during follow-up, 4 of which already had the type I endoleak at the time of the reference CTA. Thirty-seven (42%) patients developed type II endoleaks, the vast majority of which (34 of 37) were present on reference CTA. No patients were diagnosed with type III endoleaks during
Vol. 29, No. 5, July 2015
Aortoiliac elongation after EVAR 893
Table I. Patient demographics and operative characteristics Variable
N ¼ 88, n (%)
Age, years Male HTN DM HLD CAD TOB COPD Suprarenal fixation Medtronic Talent Cook Zenith No suprarenal fixation Medtronic Aneuryx Gore Excluder Mean follow-up, years
76 77 62 12 53 29 75 26 50 37 13 33 29 9 3.2
(±8) (88) (70) (14) (60) (33) (85) (30) (57) (42) (15) (43) (33) (10) (0.4e7.5)
HTN, hypertension; DM, diabetes mellitus; CAD, coronary artery disease; HLD, hyperlipidemia; TOB, tobacco use.
the study period. Twenty-nine patients (33%) underwent some form of reintervention at our institution during the follow-up period, 17 (58.6%) within the first 3 years after EVAR. Interobserver concordance between reviewers of the centerline-based measurements is depicted in Figure 1; 91.8% of the measurements fell within the 95% limits of the agreement. The concordance correlation coefficient was 0.980, and the coefficient of variation was 1.5%. Table II lists aortic measurements at the reference CTA in comparison with the last available CTA. Figure 2 depicts these findings graphically. When looking at the mean differences in aortic lengths at the various measurement points over time, increases in length are seen in the aortoiliac segments but not in the iliofemoral regions. These changes in length were statistically significant beginning at 3 years. Aortic diameters did not demonstrate significant change between baseline measurements and that of the last follow-up image, having decreased and then increased over the follow-up period. Conversely, aortoiliac lengths gradually increased over time (Fig. 3). When comparing the changes in aortic lengths with changes in aortic diameter, we found a modest correlation with a Pearson correlation coefficient of 0.34. Univariate analysis was performed to identify patient and operative variables associated with aortic elongation. Table III lists patient demographics and their association with aortic length. There was a trend to increased lengthening in men compared with women, although this was not statistically
Fig. 1. BlandeAltman interobserver agreement plot. Concordance correlation coefficient, 0.980 and coefficient of variation, 1.5%; 91.8% within the 95% limits of agreement.
significant. The presence of chronic obstructive pulmonary disease (COPD) and the use of statins appeared to have a protective effect on aortic lengthening. No other patient demographics were associated with aortic lengthening. When looking at operative variables, we found that aortic elongation over time was inversely related to proximal landing zone length (P ¼ 0.007). We did not find a significant difference in aortic lengthening when comparing devices with suprarenal fixation to no suprarenal fixation (P ¼ 0.7). Over the study period, 4 different devices were used (Table IV); Zenith (Cook Medical Inc., Bloomington, IN), Talent (Medtronic, Minneapolis, MN), Aneuryx (Medtronic), and Excluder (W. L. Gore and Associates Inc., Flagstaff, AZ). After adjusting for differences in proximal landing zone, significant differences in aortic lengthening over time were, however, observed by device type (P ¼ 0.02). This difference remained significant (P ¼ 0.04) after adjusting for reinterventions. Table V demonstrates the relationship between aortic lengthening and endoleaks or reinterventions. We found that aortic lengthening was significantly increased in patients who developed type I endoleaks compared with those who did not. In addition, those who underwent reinterventions during follow-up had more elongation that those who did not. There was no significant increase in aortic lengths in patients who developed type II endoleaks. Aortic diameter increase and aortic lengthening were both independently and positively associated with both increase rates of type I endoleaks and reinterventions by multivariable logistic regression analysis (Table VI). Every 5-mm increase in length was
894 Chandra et al.
Annals of Vascular Surgery
Table II. Aortic measurements
Aortic maximum diameter SMAetop of stent graft SMAeflow divider SMAeiliac bifurcation Top of stent grafteiliac bifurcation Iliac bifurcationefemoral bifurcation
Reference CTA, mean ± SD, mm
Last follow-up CTA, mean ± SD, mm
P value
57.4 11.0 67.9 212.5 201.5 164.2
56.0 17.5 76.4 221.4 205.4 163.4
0.25 0.001
