The correlation of aortic neck length to early and late outcomes in endovascular aneurysm repair patients
Article Outline
- Abstract
- Methods
- Results
- Neck length and fate of early endoleak and type of early intervention
- Neck length and late endoleak
- Neck length and late reintervention
- Neck length and fate of late endoleak and type of late intervention
- Survival rates by neck length
- EVAR patency by neck length
- Multivariate analysis of neck length and adverse clinical outcomes
- Discussion
- Conclusions
- Author contributions
- Acknowledgment
- References
- Copyright
Background
Initially, patients with a short angulated aortic neck were considered unfit for endovascular aneurysm repair (EVAR). Recently, however, more liberal use of EVAR has been advocated. This study analyzes the correlation of aortic neck length to early and late outcomes.
Methods
We analyzed 238 patients who underwent EVAR during a recent 7-year period. All patients were followed up clinically and underwent postoperative duplex ultrasound imaging or computed tomography angiography, which were repeated every 6 months. Aortic neck length was classified into ≥15 mm (L1, n = 195), 10 to <15 mm (L2, n = 24), and <10 mm (L3, n = 17). Kaplan-Meier methods were used to estimate freedom from late endoleak, early and late reintervention, and survival.
Results
Analyzed were 49 Ancure, 47 AneuRx, 104 Excluder, and 38 Zenith grafts. The mean follow-up was 24.7 months (range, 1-87 months). The initial technical success was 99%. The perioperative complication rates for groups L1, L2, and L3 were 13%, 21%, and 24%, respectively (P = .289). Proximal type I early endoleaks occurred in 12%, 42%, and 53% in groups L1, L2, and L3, respectively (P < .001). Intraoperative proximal aortic cuffs were needed to seal proximal type I endoleaks in 10%, 38%, and 47% in L1, L2, and L3 groups, respectively (P < .0001). However, the rate of late reintervention was comparable in all groups. Postoperatively, the size of the abdominal aortic aneurysm decreased or remained unchanged in 95%, 94%, and 88% in L1, L2, and L3, respectively (P = .660). Rates of freedom from late type I endoleak at 1, 2, and 3 years were 84%, 82%, and 80% for L1; 68%, 54%, and 54% for L2; and 71%, 71%, and 53% for L3 (P = .0263). Rates of freedom from late intervention at 1, 2, and 3 years were 96%, 94%, and 92% for L1; and 94%, 83%, and 83% for L2; and 93%, 93%, and 93% for L3 (P = .5334).
Conclusions
EVAR can be used for patients with a short aortic neck; however, it was associated with a significantly higher rate of early and late type I endoleaks, resulting in an increased use of proximal aortic cuffs for sealing the endoleaks.
The standard of care for the treatment of abdominal aortic aneurysms (AAA) has been open surgical repair since it was first described in 1951.1 The treatment of AAA was revolutionized in the last decade with the advent of endovascular aneurysm repair (EVAR) with stents in 1991.2
The suitability for EVAR, based on the manufacturers instructions for use (IFU), requires a specific length of infrarenal aorta free of aneurysm and severe infrarenal aortic angulation. Initially, the proximal neck criteria for EVAR included a neck length of ≥15 mm, a neck diameter between 17 and 25 mm, an aortic neck angulation of <45°, and 10% to 20% oversizing. However, with the liberal use of endovascular stent grafting over the last few years, and the addition of larger-diameter devices, a neck length of <15 mm, an aortic neck diameter of >25 mm, and an aortic neck angulation of ≥45° have been used. The recommendations are a minimum of 10 mm of infrarenal aorta free of aneurysm for the Talent stent graft3 (Medtronic, Inc, Santa Rosa, Calif) and 15 mm for all other stent grafts.
Strict adherence to the IFU specific for each stent graft carries a low mortality and morbidity. Conversely, deviation from the specific IFU may lead to negative effects on late results and outcomes.4 The envelope to treat AAA is pushed beyond the recommendations and IFU; hence, a significant number of patients with AAAs in the United States are presently being treated by EVAR. The proportion of AAA patients found suitable for EVAR has increased significantly, from a reported 20% in the early stage to 45% to 80% more recently.5, 6, 7, 8
This present study analyzes the correlation of aortic neck length to early and late clinical outcomes.
Methods
The study protocol was approved by the Institutional Review Board of West Virginia University/Charleston Area Medical Center.
Patients
Between January 2000 and July 30, 2007, 523 patients underwent EVAR at our institution using a variety of devices approved by the United States Food and Drug Administration. The grafts were Ancure (Guidant Corp, Indianapolis, Ind), AneuRx (Medtronic Corp, Santa Rosa, Calif), Excluder (W. L. Gore & Assoc, Flagstaff, Ariz), and Zenith (Cook Corp, Indianapolis, Ind). This study analyzed only 238 patients whose EVAR procedure was done by two of our full-time academic faculty (A. F. A. and P. A. S.). Patients treated by nonacademic physicians were excluded because we had no control over their follow-up. Also excluded were patients who lacked preoperative infrarenal aortic neck measurements (primarily due to lack of good quality computed tomography [CT] scanning), patients undergoing EVAR for ruptured AAA, and patients who lacked at least 1 year of follow-up.
Prospectively collected data were supplemented with a retrospective review of medical records (Vascular Center of Excellence charts) and radiologic images obtained before the operation. The preoperative workup of these patients included CT angiography (CTA), color duplex ultrasound (DUS) imaging, and arteriography to select patients for EVAR. Preoperative arteriograms were obtained in all patients using a marked pigtail catheter. The demographic and clinical characteristic profiles, aortic neck anatomic characteristics, operative details, and intraoperative and 30-day perioperative complications or events were recorded.
Every effort was made to follow the selection criteria recommended by the manufacturers of these devices. Device selection was based on physician preference. All procedures were performed in an independent Circulatory Dynamic Laboratory with the patient under epidural or general anesthesia, based on the physician's choice, using conventional fluoroscopy (General Electric Medical, Milwaukee, Wisc). All devices were inserted, and patients were followed up, according to the manufacturers' recommendations. Every effort was made to deploy the endovascular device flush with the level of the lowest renal artery.
All patients were encouraged to participate in our postoperative surveillance protocol, which included CTA or color DUS imaging, or both, and plain abdominal radiography at 1, 3, 6, and 12 months, then every 12 months thereafter. In our early experience, we were strict about combining CTA and color DUS imaging, which were repeated at these time intervals. During the last few years, however, our protocol was modified so that all patients were followed up using color DUS imaging every 6 months, and CTA was only done if the DUS image showed a significant change in the size of the residual AAA sac (≥0.5 cm).
CT scanning protocol
The standard CT follow-up protocol, with and without intravenous contrast material, required a CT section thickness of 3 mm. The proximal aortic neck diameter was recorded in the minor access from adventitia to adventitia, just below the lowest renal artery. Another measurement was made at 15 mm below the lowest renal artery or at the distal end of the aortic neck in patients with a short neck (<15 mm). The infrarenal aortic neck was measured on CTA using multiplanar reconstruction as the distance between the lowest renal artery and the point of the initial aneurysm dilatation, or where the infrarenal aortic diameter increased to >3 mm of the proximal neck diameter. The aneurysmal sac size was defined as the maximum transverse diameter and was also measured from adventitia to adventitia (the outer diameter). Details of these measurements, specifically the neck anatomy, were reviewed separately without the knowledge of the early and late clinical outcome of these patients.
AAA largest minor access diameter at 12 months and at the last follow-up
Pretreatment images were used to measure AAA stability or shrinkage, as recommended by the Ad Hoc Committee for the Standardized Reporting Practice in Vascular Surgery.9 Significant AAA diameter change was defined as ≥5 mm, and shrinkage was defined as ≤5 mm. Endograft migration was determined by measuring the distance from the lowest renal artery and the most cephalad portion of the stent graft, as seen on CTA images. Significant migration was defined as displacement of ≥10 mm from the predischarge study or any displacement requiring secondary intervention.
For clinical applications and statistical analysis, the aortic neck lengths in our study were classified into ≥15 mm (L1), 10 to <15 mm (L2), and <10 mm (L3).
Endoleak was determined using CT, based on extravasation of contrast between the prosthesis and the aneurysm wall or by color DUS imaging, or both, where the flow and spectral signals were outside the prosthesis. If the CT and DUS results differed, contrast arteriography was done to confirm the endoleak. In general, CT was the gold standard. Primary or early endoleak was defined as a leak detected ≤30 days of the procedure, and a secondary or late endoleak was defined as a leak observed >30 days.
The primary end points for analysis for this study included early outcome (30 days postoperatively); specifically, the incidence of early proximal aortic endoleak (type I), use of proximal aortic neck extension or cuff, other secondary interventions, technical success, and surgical conversion. A proximal aortic cuff was used to provide greater radial force and, if necessary in some patients, to extend the graft closer to the renal arteries. Secondary early outcome included operative blood loss, transfusion requirement, volume of contrast used during implantation, stent graft patency, other endoleaks (distal type I, type II, III, and IV), and other perioperative morbidity or mortality.
Technical success was defined from the periprocedural period through the first 24 hours postoperatively, which requires successful introduction and deployment of the stent graft in the absence of surgical conversion, death, or type I endoleak persisting beyond the initial surgery. Late clinical outcomes included type I endoleak, other types of endoleak, stent graft patency, aortic sac expansion, conversion to open repair, stent graft migration, aneurysm rupture, secondary interventions, and aneurysm-related death.
Statistical methods
The data analysis was performed using SAS 9.1 software (SAS Institute, Cary, NC). Basic descriptive statistics, such as means and standard deviations for continuous variables and proportions and frequencies for categoric variables, were used to analyze the data. Comparisons between the groups were performed using contingency table analysis with a χ2 or Fisher exact test for categoric variables and t tests for continuous variables to determine statistically significant differences. Logistic regression was used to predict which risk factors were associated with early/late intervention and early/late endoleak. Each factor (early endoleak and early intervention) was looked at separately. The model included age, gender, race, coronary artery disease, chronic renal disease, type of graft, neck length, neck angle, and at least one complication. The Kaplan-Meier method was used to estimate survival distributions (survival, freedom from late endoleak, freedom from late intervention, and EVAR patency) for the groups. The test statistic for comparison between these two survival distributions was based on the log-rank test. An α ≤ 0.05 was used to determine statistical significance.
Results
EVAR procedures in 238 patients were analyzed, including 49 Ancure, 47 AneuRx, 104 Excluder, and 38 Zenith devices. The mean follow-up was 24.7 months (range, 1-87 months). The overall initial technical success rate was 99%. All AneuRx, Excluder, and Zenith devices were successfully deployed. Failures were recorded with two Ancure devices: One was secondary to failure of introduction for the original model of the Ancure device, which was converted to AneuRx in the same procedure, and the other failed due to iliac artery obstruction/rupture and the patient underwent an open repair.
The demographic and clinical characteristics were comparable for all devices except for the Ancure device, where there were a statistically significantly higher number of men, patients with chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, and chronic kidney disease. This was probably a reflection of our early selection process, wherein we selected patients for the Ancure device who were too high-risk for an open procedure.
Table I summarizes the device type and AAA characteristics. Patients who underwent a repair with a Zenith device had a significantly larger AAA size than the other devices (6.5 cm, P = .0223); however, the mean difference between perioperative and postoperative AAA sizes was not statistically significant. The size of the AAA sac decreased or stayed the same in most of these patients. A higher number of patients with the Excluder devices also had a shorter neck length (<10 mm or 10 to <15 mm, P = .0705).
Table I. Abdominal aortic aneurysm characteristics by type of device
| Characteristica | Excluder | Zenith | AneuRx | Ancure | P |
|---|---|---|---|---|---|
| (n = 104) | (n = 38) | (n = 47) | (n = 49) | ||
| AAA size, cm | |||||
 Pre-op | 6.03(4.2-9.1) | 6.5(4.6-8.5) | 5.74(4.5-7.1) | 5.76(4.2-8.5) | .0223 |
| Post-op | 5.44(3-9.1) | 6.01(2.8-10.6) | 5.22(3.2-7.5) | 4.98(3.2-8.5) | .0524 |
| Mean difference | –0.57(–3.7to1.9) | –0.5(–1.9to2.3) | –0.52(–2.3to2.3) | –0.8(–3.1to0.8) | .375 |
| Neck size, mm | 23.8(16-28) | 25.7(16-32) | 24.1(17-26) | 22.9(18-26) | .1316 |
| AAA size | |||||
| Increased | 6(6) | 3(9) | 2(5) | 1(3) | .8627 |
| No change | 47(48) | 14(41) | 21(57) | 17(50) | |
| Decreased | 45(46) | 17(50) | 14(38) | 16(47) | |
| Neck length | |||||
| <10 mm | 12(12) | 4(11) | 1(2) | 0 | .0705 |
| 10 to <15 mm | 13(13) | 3(8) | 4(9) | 4(9) | |
| ≥15 mm | 77(75) | 31(82) | 42(89) | 43(91) | |
| Neck angle | |||||
| <45° | 57(56) | 22(58) | 32(68) | 37(80) | .0482 |
| 45 to <60° | 23(23) | 5(13) | 9(19) | 6(13) | |
| >60° | 21(21) | 11(29) | 6(13) | 3(7) |
aContinuous data are presented as mean (range); categoric data as number (%). |
Ancillary procedures were done more often in patients with Ancure: 38% vs 9% for AneuRx, 10% for Excluder, and 20% for Zenith (P < .001). Most of these procedures were percutaneous transluminal angioplasty (PTA) and stenting or common femoral artery endarterectomy with patch angioplasty, or both.
The demographic and clinical characteristics according to neck length are summarized in Table II. These characteristics were somewhat similar, except for gender and the incidence of coronary artery disease.
Table II. Demographics and clinical characteristics by neck length
| Variablea | L1 (≥15 mm) | L2 (10 to <15 mm) | L3 (<10 mm) | P |
|---|---|---|---|---|
| (n = 195) | (n = 24) | (n = 17) | ||
| Male | 163(84) | 14(58) | 12(71) | .0064 |
| Age, y | 74.16(53-88) | 74.79(48-91) | 71.71(48-91) | .4264 |
| Hypertension | 158(81) | 23(96) | 15(88) | .1849 |
| COPD | 83(43) | 7(29) | 6(35) | .4046 |
| Current tobacco use | 50(26) | 4(17) | 2(12) | .3012 |
| Previous tobacco use | 98(50) | 13(54) | 11(65) | .5034 |
| Congestive heart failure | 35(18) | 3(13) | 0 | .1416 |
| Coronary artery disease | 131(67) | 11(46) | 7(41) | .0185 |
| Home oxygen use | 17(9) | 1(4) | 0 | .4853 |
| Diabetes mellitus | 48(25) | 6(25) | 2(12) | .4842 |
| Chronic renal disease | 57(29) | 5(21) | 3(18) | .4374 |
| Hyperlipidemia | 112(57) | 14(58) | 7(41) | .4225 |
| Follow-up, mon | 26.31(0.1-60.8) | 17.82(0.1-61.4) | 18.14(0-86.9) | .0826 |
aContinuous data are presented as mean (range); categoric data as number (%). |
Table III summarizes aortic neck length and aortic characteristics. There were no statistically significant differences in these parameters according to neck length. Postoperatively, the size of the AAA decreased or remained unchanged in 95%, 94%, and 88% in L1, L2, and L3 patients, respectively (P = .660). Similarly, the difference in aortic neck angle in the three groups was not significant.
Table III. Abdominal aortic aneurysm characteristics by neck length
| Variablea | L1 (≥15 mm) | L2 (10 to <15 mm) | L3 (<10 mm) | P |
|---|---|---|---|---|
| (n = 195) | (n = 24) | (n = 17) | ||
| AAA size, cm | ||||
| Pre-op | 5.95(4.2-8.5) | 6.23(4.5-9.1) | 6.22(4.3-8.1) | .2136 |
| Post-opb | 5.36(2.8-10.6) | 5.74(4.2-9.1) | 5.64(3.3-9.8) | .3539 |
| Mean difference | –0.60(–3.7to2.3) | –0.56(–1.9to1.9) | –0.53(–3.1to9.8) | .952 |
| Neck size, mm | 23.89(16-32) | 23.65(18-30) | 25.38(20-32) | .0989 |
| Post-op AAA size | ||||
| Increasedb | 9(5) | 1(6) | 2(13) | .6606 |
| No change | 83(49) | 9(50) | 6(38) | |
| Decreased | 77(46) | 8(44) | 8(50) | |
| Neck angle | ||||
| <45°c | 128(66) | 12(52) | 9(53) | .3603 |
| ≥45 to <60° | 35(18) | 4(17) | 4(24) | |
| >60° | 31(16) | 7(30) | 4(24) |
aContinuous data are presented as mean (range); categoric data as number (%). |
bAt end of follow-up. |
c21 patients (52.5%) with <45° angle had <15 mm neck length (L2 and L3) vs 19 patients (47.5%) with ≥45° angle had L2 and L3 neck length (P = .1065). |
Table IV summarizes the correlation of aortic neck length and intraoperative/hospital variables. The mean fluoroscopy time, mean amount of contrast, and mean transfusion were not statistically significant.
Table IV. Intraoperative and hospital variables by neck length
| Variables, mean (range) | L1 (≥15 mm) | L2 (10-<15 mm) | L3 (<10 mm) | P |
|---|---|---|---|---|
| (n = 195) | (n = 24) | (n = 17) | ||
| Fluoroscopic time, min | 25.5(8-81) | 27.4(13-61) | 29.5(16-50) | .4667 |
| Estimated blood loss, mL | 263.3(0-2100) | 254(40-1000) | 224(100-400) | .7844 |
| Transfusion amount, U | 0.72(0-11) | 1.25(0-6) | 0.71(0-6) | .3641 |
| Contrast amount, mL | 135(30-317) | 117(45-184) | 145(54-272) | .3274 |
| Length of stay, d | 5.1(1-43) | 4.6(1-16) | 4.4(2-10) | .7609 |
The overall perioperative complications were somewhat similar in regards to neck length; however, the perioperative mortality rate was statistically higher in patients with a neck length of <10 mm and 10 to <15 mm (6% and 8%, P = .0173, Table V). This increase in death may be secondary to the complexity of the procedure. There were four perioperative deaths, three secondary to myocardial infarction, and one secondary to multisystem failure. There was no renal artery loss; however, a small percentage of patients sustained acute renal failure in the various groups, which was not statistically significant.
Table V. Perioperative complications by neck lengtha
| Complication, No. (%) | L1 (≥15 mm) | L2 (10 to <15 mm) | L3 (<10 mm) | P |
|---|---|---|---|---|
| (n = 195) | (n = 24) | (n = 17) | ||
| Myocardial infarction | 3(2) | 2(8) | 0 | .1238 |
| Iliac rupture | 1(1) | 0 | 0 | 1 |
| Graft limb thrombosis/acute limb ischemia | 9(5) | 1(4) | 1(6) | .8337 |
| Deep vein thrombosis | 2(1) | 0 | 0 | 1 |
| Systemic embolization | 0 | 1(4) | 0 | .1737 |
| Hematoma/bleeding | 4(2) | 0 | 0 | 1 |
| Wound infection | 4(2) | 0 | 1(6) | .3785 |
| Sepsis | 1(1) | 0 | 0 | 1 |
| Colon ischemia | 5(3) | 1(4) | 1(6) | .3503 |
| Acute renal failure | 3(2) | 1(4) | 1(6) | .2087 |
| Paralysis | 0 | 1(4) | 1(6) | .0296 |
| Perioperative death | 1(1) | 2(8) | 1(6) | .0173 |
| All complications | 26(13) | 5(21) | 4(24) | .2893 |
aAt least one complication. |
Table VI summarizes the correlation of neck length and the incidence of endoleak and intervention. The incidence of early type I endoleak and all early endoleaks was 53% in patients with a short neck (<10 mm), which was statistically significantly higher (P ≤ .0001). Similarly, the incidence of late type I endoleak was higher in patients with a short neck (12%, P = .05). The differences in the outcomes between the length of the neck and the graft type could not be determined because only 17 patients had a short neck (<10 mm), and the Excluder was used for 12 of these (ie, the Excluder device was used for most short necks). Seven of the 12 with the Excluder device had type I early endoleaks.
Table VI. Endoleak and intervention by neck length
| Endoleak/intervention, No. (%) | L1 (≥15 mm) | L2 (10 to <15 mm) | L3 (<10 mm) | P |
|---|---|---|---|---|
| (n = 195) | (n = 24) | (n = 17) | ||
| Early endoleaks | ||||
| Type I | 24(12) | 10(42) | 9(53) | <.0001 |
| Type II | 19(10) | 1(4) | 0 | |
| Type IV | 5(3) | 0 | 0 | |
| All early endoleaks | 48(25) | 11(46) | 9(53) | .0071 |
| Type I Rx w/aortic cuff | 20(10) | 9(38) | 8(47) | <.0001 |
| Late endoleaks | ||||
| Type Ia | 11(6) | 3(15) | 2(12) | .044 |
| Type II | 18(11) | 4(20) | 2(12) | |
| Type III | 0 | 0 | 1(6) | |
| All late endoleaks | 29(17) | 7(35) | 5(31) | .086 |
| Early interventionb | 30(15) | 9(38) | 8(47) | <.0001 |
| Late intervention | 10(5) | 2(8) | 2(12) | .321 |
aTwo L1 patients with late type I endoleak were secondary to migration. |
bMost of these were done for type I proximal endoleak noted during the primary surgery, which were corrected by proximal aortic cuffs. |
Proximal aortic cuff extensions were used to seal early type I endoleaks in 47% of patients with a short neck vs 10% in patients with a ≥15-mm neck (P < .0001). Overall, there were more early interventions in patients with a short neck (P < .0001). Proximal cuff extensions were needed in 27% of Excluder grafts and 13% of the AneuRx grafts, in contrast to 0% for the Zenith graft. However, the Excluder group had more patients with shorter (P = .07) and angulated (P = .048) necks.
No correlation was found between the mean neck size and the incidence of type I endoleak. The mean neck size for patients with no early endoleak was 23.96 mm compared with 23.88 mm for patients with type I early endoleaks and 24.2 mm for type II early endoleaks.
Neck length and fate of early endoleak and type of early intervention
As noted in Table VI, a proximal aortic cuff extension was inserted in eight of nine patients with L3 neck anatomy to seal their early type I endoleak, except three who needed Palmaz stents in addition to aortic cuff extensions. These early proximal type I endoleaks were noted during the primary surgery. In the remaining patient, the type I endoleak (detected postoperatively) decreased, with no significant change in the aortic sac size at 12 months. The completion angiogram also documented 10 patients with a L2 neck who had type I early endoleak. Eight of these patients were treated with proximal aortic cuff extension and sealed immediately, and one was treated with aortic cuff extensions that decreased the endoleak, and it sealed 6 months later. Early type I endoleaks occurred in 24 patients with L1 neck anatomy; of these, 20 were treated with proximal aortic cuff extensions only, and all sealed immediately. One patient each of the remaining four patients sealed at 3 months, 6 months, 12 months, and 24 months postoperatively, with no significant change in the AAA sac size. An early type II endoleak in one patient with L2 neck anatomy sealed 6 months later. There were 19 patients with L1 neck anatomy who had type II endoleak, 13 of which sealed spontaneously ≤30 days and 1 patient each sealed at 3 months, 16 months, and 46 months. The three remaining patients had persistent type II endoleaks, two at 12 months and one at 36 months, with no significant change in the AAA sac size.
Overall, eight L3 patients and nine L2 patients required early 30-day perioperative intervention with proximal aortic extensions to seal type I proximal endoleaks. Thirty patients with L1 neck anatomy had early intervention. These included 20 proximal aortic cuff extensions to seal proximal type I early endoleaks, 2 groin explorations for bleeding, 2 lysis/PTA, 4 thrombectomies, and 2 patients with Ancure devices required stent grafts for treatment of iliac rupture. Overall, proximal aortic cuff extensions were used seal to 37 of 43 (86%) of the proximal type I early endoleaks detected during the primary surgery, with no significant renal artery loss or noticeable complications. However, three of these proximal aortic cuffs required an extra Palmaz stent to aid in sealing proximal endoleaks in patients with L3 neck anatomy.
Neck length and late endoleak
Rates of freedom from late endoleak at 1, 2, and 3 years were, respectively, 84%, 82%, and 80% for L1 patients; 68%, 54%, and 54% for L2 patients; and 71%, 71%, and 53% for L3 patients (P = .0263, Fig 1).
Fig 1.
Freedom from late endoleaks according to neck length at 1, 2, and 3 years.
Neck length and late reintervention
Rates of freedom from late intervention at 1, 2, and 3 years were, respectively, 96%, 94%, and 92% for L1 patients; 94%, 83%, and 83% for L2 patients; and 93%, 93%, and 93% for L3 patients (P = .533, Fig 2).
Fig 2.
Freedom from late intervention according to neck length at 1, 2, and 3 years.
Neck length and fate of late endoleak and type of late intervention
Five patients with L3 anatomy had late endoleaks, two of which were type I endoleaks. One was an early type I endoleak that was present at 12 months, and one was noted at 6 months; neither was associated with significant AAA sac size changes. Two patients had late type II endoleaks, of which one sealed at 6 months postoperatively and one sealed at 12 months. The fifth patient had a type III endoleak, which was noted at 40 months and treated with an aortic cuff extension.
Seven patients with L2 neck anatomy had late endoleak; three were late type I endoleaks, and four were type II endoleaks. The three late type I endoleaks were early endoleaks that were treated with proximal aortic cuffs, reappeared as late endoleaks, but sealed between 3 and 6 months later. In the four patients with late type II endoleaks, one sealed 3 months later and three were present at 9, 18, and 48 months. All were associated with a stable AAA sac size.
Late endoleaks occurred in 29 patients with L1 neck anatomy; 12 were type I and 17 were type II. Four of the late type I endoleaks were originally early endoleaks and sealed as described earlier in 3 to 24 months. Eight others were new late type I endoleaks. Two were distal iliac endoleaks at 12 and 18 months and were treated with a distal iliac extender. One was noted at 12 months later (graft migration) and was associated with an increasing AAA sac size, but the patient refused further treatment. The AAA eventually ruptured, and the patient died. One endoleak was noted at 6 months, but the patient died of gastrointestinal hemorrhage. Three endoleaks were noted at 6 and 12 months and were treated with proximal aortic cuff extenders. The remaining endoleak had graft migration at 30 months and was treated with a proximal aortic cuff.
Late type II endoleaks occurred in 17 patients with L1 neck anatomy. Six were present as early endoleaks and three sealed (from 3 to 46 months) with no intervention; the other three persisted at 36 months with no significant change in AAA sac size. The remaining 11 patients had late type II endoleaks. Six sealed between 9 and 27 months with no intervention, and the endoleaks in the remaining five patients persisted for 6 to 48 months with no significant AAA sac size.
Two patients with L3 neck anatomy had late intervention. One patient had a type III endoleak at 36 months, which was treated successfully with two cuffs, and the other had an infected endovascular graft that was removed. Two other patients had L2 neck anatomy. One had late limb occlusion at 15 months and was treated with a femorofemoral bypass graft. A left iliac limb distal type I endoleak developed in the other and was treated successfully with an iliac extender.
There were also 10 late interventions in patients with L1 neck anatomy, including 1 patient with PTA/stent with lysis for limb ischemia at 3 months, 1 aortic cuff extension for type I endoleak for graft migration at 30 months, 1 iliac extension for distal type I endoleak, 1 conversion of patient with a unibody Cook device with femorofemoral bypass graft, 1 coil embolization for type II endoleak of the hypogastric artery at 18 months, 1 distal limb extender for distal type I endoleak at 12 months, 2 patients with proximal aortic cuff for proximal type I endoleak at 12 months, 1 aortic cuff extension for type I proximal endoleak at 6 months, and 1 graft limb thrombectomy with femorofemoral bypass graft.
Survival rates by neck length
The survival rates at 1, 2, and 3 years were 90%, 82%, and 77% for L1 patients; 76%, 76%, and 76% for L2 patients; and 81%, 72%, and 72% for L3 patients, respectively (P = .3595, Fig 3).
Fig 3.
Survival rates according to neck length at 1, 2, and 3 years.
EVAR patency by neck length
The EVAR patency rates at 1, 2 and 3 years were, respectively, 95%, 94%, and 93% for L1 patients; 100%, 91%, and 91% for L2 patients; and and 100%, 89%, and 89% for L3 patients (P = .9534, Fig 4).
Fig 4.
Patency of endovascular aneurysm repair (EVAR) according to neck length at 1, 2, and 3 years.
Overall, 13 EVAR procedures failed, eight perioperatively and five at a later date. Five were Zenith devices, 4 were Ancure, 3 were Excluder, and 1 was AneuRx. Patients with perioperative graft thrombosis were successfully treated with thrombectomy or lysis, or both, and PTA/stenting, except for three patients who needed femorofemoral bypass grafting. In the remaining five grafts, one Ancure graft failed at 1 month and was successfully treated with PTA/stenting and lysis. Graft infection at 12 months caused one Zenith graft to fail and it was removed, but the patient died 1 month later from multiple organ system failure. Another Zenith graft failed 7 months later, and the patient was treated with a femorofemoral bypass graft. Two other Zenith grafts failed at 8 and 12 months later, with no further treatment (patients were asymptomatic).
Multivariate analysis of neck length and adverse clinical outcomes
Logistic regression analysis showed that early endoleak was highly associated with a shorter neck or angulated neck (>60°), or both. Early intervention was also significantly associated with neck length or angulation.
Discussion
The manufacturers of EVAR devices have generally recommended that an infrarenal neck length of ≥15 mm is needed to secure proximal graft fixation. However, a threshold of 10 mm has been proposed recently as sufficient to produce adequate sealing.10, 11 Overall, hostile neck anatomy is the primary reason that patients with infrarenal aortic aneurysm must undergo open repair.12 Because of patients' preference and referral patterns, a significant portion (53%) of this series of patients treated with EVAR at our institution had hostile aneurysm neck anatomy. The endograft is usually held in its proximal position by friction dependent on the radial force of the graft against the aneurysmal wall and the contact surface between the graft and the aortic wall. The length of the proximal attachment in patients with short necks will decrease, leading to a smaller contact surface and lower friction forces. Therefore, stent graft migration may occur if displacement forces exceed the friction forces.13
Device technology, such as suprarenal fixation and larger diameter grafts, continues to evolve to allow additional patients to be treated with EVAR who would not have previously been candidates. However, it is important to assess outcomes in patients with hostile aneurysm neck anatomy to weigh the risks and benefits of EVAR and open repair, especially considering the possibility of the need for increased late intervention for endoleaks, as well as increased complexity during the procedure that could result in increased perioperative complications.
Our results showed no significant increase in fluoroscopic time, mean amount of contrast used during the procedure, and mean amount of transfusion required in patients with a short neck. Fairman et al14 also evaluated if complicated neck anatomy affected blood loss, transfusion requirements, and volume of contrast required for EVAR. There was no significant difference in the procedural variables of transfusion required or the volume of contrast used; however, an important difference between the two studies is that they combined six distinct neck features—short (<15 mm), very short (<10 mm), dilated (>28 mm), angulated (≥45°), calcified, and thrombus-lined—to describe a complicated neck.
Another important outcome to consider is whether perioperative complications increase with a short neck anatomy. Our study did not demonstrate a significant increase in the incidence of graft limb thrombosis or acute limb ischemia and hematoma or bleeding complications with a short neck. However, perioperative death was significantly increased with short necks in our study.
One landmark study analyzed the outcomes of EVAR with neck length from the European Collaborators on Stent-Graft Techniques for Aortic Aneurysm Repair (EuroSTAR) registry.8 Leurs et al 8 reported that the 30-day mortality rate was higher in patients with a neck of <15 mm compared with patients with an aortic neck of ≥15 mm. This finding was also supported by an earlier study by Hovsepian et al15 that concluded that a shorter proximal neck was a significant risk factor for increasing intraoperative and postoperative complications and deaths.15 Other studies have not shown an increased risk of perioperative mortality with hostile neck anatomy.12, 14 The difference in outcomes could possibly be secondary to the fact that these studies included other anatomic criteria for the classification of a hostile neck, whereas our study only focused on neck length.
The clearest difference with decreasing neck length, when compared with favorable neck anatomy in our series, was the incidence of type I early endoleaks. More than one-half of patients with neck lengths <10 mm had type I early endoleaks. Most of these required intraoperative placement of aortic cuffs; however, three patients required proximal aortic Palmaz stents to seal proximal type I early endoleaks.
Others have used adjunctive balloon-expandable stents to treat patients with a hostile neck anatomy.16 Cox et al16 treated 19 patients with proximal balloon-expandable stents (17 were Cordis/Palmaz stents) for hostile neck anatomy, classified as a neck length of <15 mm, neck diameter of ≥26 mm, circumferential thrombus at the proximal neck, angulated neck ≥60°, and a neck bulge or reverse tapered neck. From this experience, they concluded that EVAR may be offered to an expanded patient population with hostile neck anatomy with the use of prophylactic balloon-expandable stents.
Decreased neck length was also associated with significantly increased rates of late type I endoleaks, although this did not correlate with a significant increase in late reintervention. Similar results were noted by Leurs et al,8 who found that short neck lengths predicted early as well as late type I endoleaks. During a long-term follow-up of 4 years, the hazard of proximal endoleak was 2 and 2.3 times higher in patients with a neck of 11 to 15 mm and patients with a neck of ≤10 mm, respectively, than patients with a neck of >15 mm. Although proximal endoleak was strongly associated with rupture of AAA, the incidence of rupture was too low to reveal any significant association with short neck or severe aortic neck angulation in the entire study group.17
Fulton et al,18 also using the AneuRx endograft, concluded that patients with unfavorable neck anatomy had significantly higher migration, device-related complications, and secondary intervention rates. However, there was no incidence of open conversion, rupture, or AAA-related deaths. Their conclusions supported the use of AneuRx devices as a feasible alternative to open repair, even in patients with a challenging neck anatomy. Stanley et al,10 using the Zenith device, also found increasing rates of proximal endoleak.
Other studies that have used multiple anatomic factors to classify hostile neck anatomy have not found a significant difference in patients treated with EVAR with hostile neck anatomy vs “good” neck anatomy in regards to early and late type I endoleaks.12, 19 Again, it is likely that the differences in these study results, when compared with ours, are that multiple anatomic factors were included to define a hostile neck.
Our series found no correlation between EVAR patency and neck length. Similar results were also noted by Fairman et al,14 who concluded that the primary graft limb patency was 100%, regardless of the neck anatomy.
This study has demonstrated that patients with short neck anatomy may be treated with EVAR. However, short necks were clearly associated with an increasing incidence of type I early endoleak that necessitated early intervention during the initial procedure to achieve proximal aortic neck seals. Similar conclusions were also noted by Choke et al.19
Our study was limited by the heterogeneity of devices used for EVAR. The Ancure device is no longer available on the market and was not a modular device. We also included devices with and without suprarenal fixation, which may affect the outcome with hostile neck anatomy.20 Another limitation was that device selection for each case was not randomized, and that there was likely a preference for certain devices for patients with hostile neck anatomy and preference for other devices in those with good neck anatomy.
Another limitation of our study was that the follow up for the L2 and L3 groups was much shorter than the L1 group, because in our earlier experience, we were more strict in the criteria of neck selection, as indicated by the IFU. During the last few years, however, we became more liberal in using these devices, specifically for shorter necks, as seen in the results.
Conclusions
Although early and late type I endoleaks increased in patients with short necks, there was not an increased need for late intervention, only an increased need for early placement of aortic cuffs or early intervention. We did, however, document an increased rate of perioperative complications and death in patients with short neck anatomy. This has to be weighed when making the decision to offer open repair vs EVAR in these patients.
Author contributions
We acknowledge Mary Emmett, PhD, Director, Center for Health Services & Outcomes Research, Charleston Area Medical Center, for her assistance on this manuscript.
References
- . Resection of an aneurysm of the abdominal aorta: re-establishment of the continuity by a preserved human arterial graft, with result after five months. AMA Arch Surg. 1952;64:405–408
- . Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg. 1991;5:491–499
- . Important safety information. http://www.medtronic.com/your-health/abdominal-aortic-aneurysm/important-safety-information/index.htmAccessed April 30, 2009
- Outcomes following endovascular abdominal aortic aneurysm repair (EVAR): an anatomic and device-specific analysis. J Vasc Surg. 2008;48:19–28
- The anatomical suitability of abdominal aortic aneurysms for endovascular repair. Br J Surg. 1997;84:178–180
- Will endovascular repair replace open surgery for abdominal aortic aneurysm repair?. Ann Surg. 2000;232:501–507
- Impact of exclusion criteria on patient selection for endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2001;34:1050–1054
- . Influence of infrarenal neck length on outcome of endovascular abdominal aortic aneurysm repair. J Endovasc Ther. 2006;13:640–648
- Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg. 2002;35:1048–1060
- Evaluation of patient selection guidelines for endoluminal AAA repair with the Zenith stent-graft: the Australasian experience. J Endovasc Ther. 2001;8:457–464
- . Endovascular grafting in patients with short proximal necks: an analysis of short-term results. Cardiovasc Surg. 2000;8:350–354
- Does hostile neck anatomy preclude successful endovascular aortic aneurysm repair?. J Vasc Surg. 2003;38:657–663
- . Influence of severe infrarenal aortic neck angulation on complications at the proximal neck following endovascular AAA repair: an EUROSTAR study. J Endovasc Ther. 2007;14:1–11
- Midterm pivotal trial results of the Talent Low Profile System for repair of abdominal aortic aneurysm: Analysis of complicated versus uncomplicated aortic necks. J Vasc Surg. 2004;40:1074–1082
- Endovascular abdominal aortic aneurysms repair in 144 patients: correlation of aneurysm size, proximal aortic neck length, and procedure-related complications. J Vasc Interv Radiol. 2001;12:1373–1382
- . Outcomes of endovascular AAA repair in patients with hostile neck anatomy using adjunctive balloon-expandable stents. Vasc Endovascular Surg. 2006;40:35–40
- Rupture of infrarenal aortic aneurysm after endovascular repair: a series from EUROSTAR registry. Eur J Vasc Endovasc Surg. 2003;26:487–493
- Effect of challenging neck anatomy on mid-term migration rates in AneuRx endografts. J Vasc Surg. 2006;44:932–937
- Outcomes of endovascular abdominal aortic aneurysm repair in patients with hostile neck anatomy. Cardiovasc Intervent Radiol. 2006;29:975–980
- . Suprarenal endograft fixation avoids adverse outcomes associated with aortic neck angulation. Ann Vasc Surg. 2005;19:172–177
Competition of interest: none.
PII: S0741-5214(09)00998-7
doi:10.1016/j.jvs.2009.04.061
© 2009 Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
