Effect of challenging neck anatomy on mid-term migration rates in AneuRx endografts

Methods 

Prospectively maintained databases at the University of North Carolina (UNC) and Washington University (WU) were used to identify 595 patients (UNC, n = 230; WU, n = 365) who underwent EVAR with the AneuRx bifurcated stent graft. Because mid- and long-term migration rates were of primary interest, patients with at least 30 months of follow-up (range, 30-55 months) were identified and underwent further assessment of migration (UNC, n = 25; WU, n = 59). The period of initial EVAR extended from October 1999 to March 2001, occurring in the first 18 months after Food and Drug Administration approval, and included only patients treated after commercial availability of the AneuRx device. No patient was treated as part of a clinical trial. Institutional review board approval was obtained at each institution.

Patients were evaluated with preoperative contrast-enhanced computed tomography (CT) imaging with 3-mm reconstructions. Preoperative imaging was not available for review for 4 patients (4.8%). In these cases, the preimplantation intraoperative angiogram was used to estimate neck length angle and diameter, and the 1-month postoperative CT scan was used to determine all other measurements. Patients were followed up after surgery with scheduled physical examinations, multiple-view abdominal radiographs, and CT scans at 1, 6, and 12 months and yearly thereafter. Additional imaging was obtained as clinically indicated.

Multiplanar reconstructed axial image sets of the preoperative and postoperative CT scans were examined by a single observer. Orthogonal reconstructions that were perpendicular to the flow axis, not to the patient, were performed on a workstation. The level of the inferior aspect of the most caudal patent renal artery was marked. Next, the most cephalad level at which a complete ring of the AneuRx device could be visualized was marked. The extent of caudal device migration was based on these measurements. Using similar multiplanar reconstructions, additional measurements were obtained, including neck length, neck diameter, maximal aneurysm diameter, terminal aorta diameter, neck-body angle, the extent of neck thrombus and calcification, and the presence of endoleak.

All diameter measurements were taken from adventitia to adventitia. Neck length was measured as the distance from the most caudal renal artery to the point of initial aneurysmal dilation. Neck angle was defined by the angle formed between the flow axes of the neck and body of the aneurysm.

A minimum of three CT scans were analyzed for each patient, including the preoperative, 1-month postoperative, and most recent scan. If a patient required a secondary intervention, the CT scan before the intervention was also examined. In patients exhibiting device migration, the preceding CT scans were examined to determine the earliest evidence of migration. Medical records were reviewed to identify the presence or absence of potential risk factors, including age, sex, diabetes mellitus, hypertension, coronary artery disease, smoking history, chronic obstructive pulmonary disease, chronic renal insufficiency, and use of β-blockers.

Freedom from migration over time was expressed as Kaplan-Meier life-table analysis by using the log-rank test to determine significance. Bivariate analysis was completed to study the relationship between migration and the following variables: age, sex, diabetes mellitus, hypertension, coronary artery disease, smoking, chronic obstructive pulmonary disease, renal insufficiency (RI), preoperative use of β-blockers, aneurysm size, percentage oversizing, neck diameter, neck length, neck angle, and graft deployment distance from the most caudal renal artery. The χ2 test was used for categorical variables, and the t test was used for continuous variables. Multivariate analysis, controlling for each of the possible confounding variables, was also conducted by using both the Cox regression model and the logistic regression model. A P value <.05 was considered significant. Data are presented as mean ± SE. Statistical analysis was conducted with SAS software (SAS Institute Inc, Cary, NC).

Results 

Eighty-four patients with follow-up of more than 30 months were studied, consisting of 71 men (84%) and 13 women (15%), with a mean age of 73.1 ± 1.1 years and a mean average aneurysm size of 5.7 ± 1.2 cm (range, 4.4-10.2 cm). The patient risk factors are summarized in Table I. Aneurysm characteristics are displayed in Table II. The overall average neck length was 25.6 ± 1.1 mm; nonmigrators had an average neck length of 26.7 ± 1.2 mm, whereas the average neck length in migrators was 20.5 ± 3.0 mm. The mean neck angle was 26.5° ± 2.6° (nonmigrators, 24.1° ± 2.8°; migrators, 39.5° ± 6.9°). Devices were inserted at a mean distance of 3.4 ± 0.5 mm below the renal arteries (nonmigrators, 2.8 ± 0.5 mm; migrators, 6.5 ± 1.9 mm).

Table I. Patient demographics

	Variable	Migrators (n=14)	Nonmigrators (n=70)	P value
	Mean age (y)	72.4±2.1	73.3±0.8	NS
	Male	11(79%)	59(84%)	NS
	Female	3(21%)	11(16%)	NS
	DM	2(14%)	10(14%)	NS
	HTN	9(64%)	38(54%)	NS
	CAD⁎	7(50%)	41(59%)	NS
	COPD	3(21%)	11(16%)	NS
	Renal insufficiency†	0	5(7%)	NS
	Beta blockade	1(7%)	7(10%)	NS
	Smoker	8(57%)	49(70%)	NS
	UFNA	10(71%)	17(24%)	.0003

NS, Not significant; UFNA, unfavorable neck anatomy; DM, diabetes mellitus; HTN, hypertension; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease.

⁎ Previous myocardial infarction, angina, or electrocardiogram evidence of ischemia. † Preoperative serum creatinine >1.5 mg/Dl or on dialysis.

Table II. Patient characteristics (migrators vs nonmigrators) by multivariate analysis

	Variable	All patients (n=84)	Migrators (n=14)	Nonmigrators (n=70)	P value
	Aneurysm diameter (mm)	56.6±1.2	55.9±5.3	56.8±1.3	NS
	Male:female	6.6:1	3.7:1	7.8:1	NS
	Neck length (mm)	25.6±1.1	20.5±3.0	26.7±1.2	.053
	Neck diameter (mm)	23.4±2.4	23.2±6.7	23.5±2.6	NS
	Neck angle (°)	26.5±2.6	39.5±6.9	24.1±2.8	.025
	% Oversizing	11.6±1.2	10.3±2.4	11.9±1.3	NS
	Initial distance to renal artery (mm)	3.4±0.5	6.5±1.9	2.8±0.5	.011
	Type Ia endoleak	7(8%)	6(43%)	1(1.4%)	<.0001
	Any endoleak	24(29%)	6(43%)	18(26%)	NS

NS, Not significant. Data are mean ± SE unless otherwise noted.

Upon retrospective review, 70% of the patients (n = 59) met all inclusion criteria for neck anatomy (length, angle, diameter, and quality) as defined by the revised instructions for use (IFU) guidelines described previously and are referred to as those with favorable neck anatomy (FNA). The remaining 30% (n = 25) had neck characteristics that fell outside of the revised IFU guidelines and are referred to as those with unfavorable neck anatomy (UFNA). Thirteen of these patients had a neck-body angle between 45° and 60°, whereas 6 patients had a neck-body angle greater than 60°. Four patients had a neck length between 10 and 15 mm, whereas two patients’ necks were less than 10 mm. Two patients had endograft oversizing less than 10%. Distal attachment zone characteristics were not analyzed. Overall, late graft-related complications occurred in 38% of patients (FNA, 27%; UFNA, 64%; P = .003; relative risk, 1.7). Complications included endoleak of any type (n = 24; 29%), migration (n = 14; 17%), and component separation (n = 3; 4%). There was no incidence of rupture or aneurysm-related death.

The incidence of type Ia endoleak was 16% (4/25) in UFNA patients and 5% (3/54) in FNA patients. All were successfully repaired with placement of a proximal cuff, and all but 1 patient (6/7) exhibited concomitant migration.

Migration 

The mean follow-up time was 40.3 months (range, 30-55 months). Life-table analysis for FNA patients revealed a migration rate of 0% and 6.1% at 2 and 4 years, respectively (Fig 1). For UFNA patients, the migration rate was 24.0% and 42.1% at 2 and 4 years, respectively. With use of the log-rank test for comparison, this reached statistical significance (P < .0001). The overall (FNA and UFNA) migration rate was 7.1%, 17.1%, and 17.1% at 2, 3, and 4 years, respectively (Fig 2). Migration in the FNA subset occurred in 4 (6.8%) of 59 patients. Ten (40%) of the 25 patients with challenging neck anatomy (UFNA) experienced a migration. Among the 14 migrators (FNA, n = 4; UFNA, n = 10), 1 patient had movement less than 10 mm with a secondary intervention. The relative risk of migration for UFNA patients was 2.5 compared with FNA patients (P = .0003).

View Large Image Download to PowerPoint

Fig 1. Freedom from migration by Kaplan-Meier analysis for those with favorable neck anatomy (FNA) and unfavorable neck anatomy (UFNA).

View Large Image Download to PowerPoint

Fig 2. Freedom from migration by Kaplan-Meier analysis for all patients.

Bivariate analysis of possible risk factors associated with migration was performed by using the χ2 and t-test methods. Neck-body angle (P = .04), distance of initial placement from the most caudal renal artery (P = .01), and overall compliance with revised IFU recommendations (P = .0003) were found to influence migration. Because of the possible confounding nature of several risk factors, multivariate analysis was implemented by using both Cox regression and logistic regression models. The results were similar to those with bivariate analysis. Shorter neck length approached but did not reach statistical significance for predicting migration with multivariate analysis (P = .053). Migration did not seem to be influenced by initial aneurysmal diameter, percentage oversizing, or neck diameter. No other patient demographic variable, such as age, β-blockade, medical history, or social history, influenced the risk of migration.

We then studied the effect of neck-body angle on migration. Those with a neck-body angle greater than 45° had a 4 times greater risk of migration than those with an angle less than 45°. It is interesting to note that when 60° was used, the odds ratio was only 3.75.

Discussion 

Reports of the migration of aortic endografts have raised concern over their long-term durability. Relatively high rates of migration for the AneuRx device have been described previously. Cao et al6 reported 113 patients with a minimum of 24 months of follow-up after EVAR with the AneuRx device. Of these 113 patients, 17 had movement of greater than 10 mm. The probability of migration by life-table analysis was 20% at 24 months and 27% at 36 months, which is noticeably higher than our results. Of the patients who had migration, 47% underwent secondary procedures, and two patients had open conversion. Through surveillance and early intervention, rupture and abdominal aortic aneurysm (AAA)-related deaths were avoided.

A higher migration rate was reported by Conners et al,7 who analyzed 91 patients after AneuRx implantation and used movement of 5 mm or more to define device migration. Their migration rates were 20.4%, 42.1%, and 66.7% at 2, 3, and 4 years after implantation, respectively. However, few patients were available for 3- and 4-year follow-up (19 and 3, respectively). Additionally, use of a 5-mm threshold for device migration may increase the number of false-positive results, taking into consideration the margin of error when examining 2.5- to 3-mm-slice axial CT scans.

Most recently, Zarins et al13 reviewed 1119 patients who were enrolled in the multicenter AneuRx clinical trial. Stent-graft migration was reported in 94 patients. Kaplan-Meier estimates of migration rates were 7% at 2 years and 19% at 3 years. Only 8% of patients were at risk at the end of the 3-year interval. Twenty-three (24%) of 94 patients with migration underwent secondary intervention. One patient experienced aneurysm rupture associated with graft migration (successful open repair), and there were no aneurysm-related deaths.

In this study, all 84 patients had follow-up of at least 30 months, with a mean follow-up time of 40.3 months. Life-table analysis reveals relatively low migration rates of 0%, 6.1%, and 6.1% at 2, 3, and 4 years, respectively, when the revised IFU guidelines are followed. These results, although they represent a subset of patients, are lower than those previously published. Not surprisingly, patients with neck anatomy that did not comply with current IFU guidelines had much more frequent migration rates of 24%, 42%, and 42% at 2, 3, and 4 years, respectively. Overall migration rates were 7.1%, 17.1%, and 17.1%, at 2, 3, and 4 years, with 18 (21%) patients remaining at risk after 48 months. These overall results are similar to those reported by Zarins et al.13 These data reinforce the importance of patient selection in patients undergoing EVAR with the AneuRx stent graft.

Of additional note, these reported migration rates are all accompanied by a notable low incidence of rupture and AAA-related deaths. Even when examining implantation of devices into aneurysm necks that did not meet recommended criteria, the results indicate safety, because these patients have avoided aneurysm rupture by undergoing secondary procedures and placement of aortic cuffs and transrenal components to rectify their potential problems. In our series of patients with challenging neck anatomy, no patient experienced rupture, AAA-related death, or open conversion. All necessary interventions were performed via an endovascular approach. EVAR should therefore not be discounted in patients with comorbidities that prohibit open operation even when their neck anatomy does not fall within strict IFU guidelines. At the time of their treatment, next-generation devices were not yet available, and many of these patients could now be treated with alternative devices.

Previous articles implicate various factors as influencing device migration.8, 11, 15, 16, 17 Cao et al6 found an initial aortic neck diameter of 25 mm or more, a preoperative AAA size of greater than 55 mm, and aortic neck enlargement of greater than 10% to be risk factors for migration. Conners et al7 demonstrated a trend toward greater endograft oversizing in migrators and also found that aortic neck dilatation was a predictor of migration; however, neck angle, size of the AAA, and initial neck diameter were not. Finally, Zarins et al13 reported that the proximal fixation length and the renal artery to stent graft distance were significant in device migration. Our data support the importance of neck-body angulation and the initial distance from the device to the renal arteries. In our study, neck length approached, but did not achieve, statistical significance. We did not identify initial neck or aneurysmal diameter or percentage oversizing as influencing migration. Of note, the degree of oversizing was not excessive (11.6% ± 10.6%).

In this study, several migrators (29%) did not undergo secondary interventions because of physician discretion or patient preference. They were closely followed up clinically and radiographically because of shrinking aneurysm size and/or a long aneurysm neck providing for a maintained zone of proximal fixation and sealing. As shown in Table II, all other patients (71%) successfully received a proximal aortic cuff, with one patient requiring an embolization of a type II endoleak. Cases with concomitant type I endoleak (n = 6) were repaired. No patient required open conversion. Our secondary intervention rate is greater than that reported in other studies. In the AneuRx clinical trial, only 32% of patients with migration underwent treatment. Cao et al6 repaired 8 (47%) of 17 migrators: 6 patients were treated with placement of a proximal cuff, and 2 patients underwent late conversion to open repair. Similarly, Conners et al7 treated 5 (33%) of 15 patients with device migration, all with placement of proximal aortic cuffs. This variation may be explained in part by the diversity in defining migration or the distribution of adverse neck characteristics (Table III).

Table III. Patients with migration: timing of migration is from time of implantation

	Patient No.	Timing of migration (mo)	Migration (mm)	Neck anatomy	Clinical sequelae	Intervention
	1	1	10	UFNA	N	Proximal cuff⁎
	2	12	12	UFNA	I,II,AE	Embolization, proximal cuff
	3	12	10	UFNA	N	Proximal cuff
	4	14	15	UFNA	I	Proximal cuff
	5	22	14	UFNA	N	Observation
	6	23	10	UFNA	I,AE	Proximal cuff
	7	30	22	UFNA	I,AE	Proximal cuff
	8	30	6	FNA	I	Proximal cuff
	9	33	15	FNA	I	Proximal cuff
	10	35	18	UFNA	N	Proximal cuff
	11	36	10	UFNA	N	Observation
	12	36	11	FNA	N	Observation
	13	36	15	UFNA	N	Observation
	14	49	18	FNA	I,AE	Proximal cuff

I, Type I endoleak; II, type II endoleak; AE, aneurysm expansion; N, none; UFNA, unfavorable neck anatomy; FNA, favorable neck anatomy.

⁎ Migration discovered at the 1-month follow-up computed tomographic scan.

Although demonstrating encouraging success rates, the durability of these secondary procedures is not well documented. Component separation or continued migration could still potentially occur. Therefore, patients experiencing device migration, including those repaired with secondary procedures, should undergo enhanced surveillance.

Conclusion 

In conclusion, the AneuRx bifurcated device under the revised instructions for use exhibits a low mid-term migration rate. Patients who have unfavorable aneurysm neck anatomy experience significantly higher migration, device-related complication, and secondary intervention rates. However, there was no incidence of open conversion, rupture, or AAA-related death, thereby supporting the use of the AneuRx device as a feasible alternative to open repair even in those patients with challenging neck characteristics. Enhanced surveillance should be used in these high-risk patients.

Author Contributions