Spinal cord ischemia after TEVAR in patients with abdominal aortic aneurysms
Article Outline
Objective
To examine the incidence of and the anatomic factors that may contribute to spinal cord ischemia (SCI) in patients with a history of abdominal aortic aneurysms (AAA) after thoracic endovascular aortic repair (TEVAR).
Methods
The medical records, computed tomography (CT) angiograms, and a prospectively maintained clinical database of all TEVAR patients at a single institution between 2000 and 2007 were reviewed. Select preoperative demographics, thoracoabdominal aortoiliac anatomy, intraoperative procedural variables, and postoperative outcomes were examined. Univariate and multivariate analyses were performed and odds ratio estimates were reported with 95% confidence intervals.
Results
Of the 261 patients who underwent TEVAR, 27 developed SCI (10%). Thirteen (48%) of these 27 patients were completely reversed with spinal drainage, and 14 (52%) were permanent. Patients with SCI tended to be older (P = .006), male (P = .049), and required more emergent procedures (P = .051) performed under general anesthesia (P = .004). Interestingly, while prior AAA repair (50/261, 19%) alone was not associated with SCI (P = .44), a history of either repaired or unrepaired AAA (101/261, 39%) was a predictor of SCI on multivariate analysis (odds ratio [OR] = 4.35 [1.43, 14.3], P = .10), independent of thoracic aortic coverage (P = .001) and lumbar artery patency (P = .008), both of which were also associated with SCI.
Conclusion
Although the causes of SCI after TEVAR are multifactorial, abdominal aortic anatomy appears to be associated with development of this complication. Patients with either prior AAA repair or those with unrepaired AAA appear to be at increased risk for SCI.
Spinal cord ischemia (SCI) is a devastating complication in patients undergoing surgical or endovascular repair of the thoracic aorta. The incidence of SCI has been reported to be between 0-14% for patients undergoing thoracic endovascular aortic repair (TEVAR).1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 Perioperative risk factors contributing to this complication have been reported to include length of aortic coverage,8, 10, 12 prior abdominal aortic aneurysm (AAA) repair,1, 5, 7 hypotension,3, 13, 14 and left subclavian artery coverage.15, 16, 17, 18 Although the putative mechanism of loss of lumbar collateral perfusion in those who had prior aortic repairs appears reasonable, occurrence of SCI in this subset of patients has not been consistent. In this study, we examined the incidence of and the anatomic factors that may contribute to SCI in patients with a history of AAAs after TEVAR.
Methods
Between September 2000 and September 2007, 261 TEVARs were performed at a single tertiary-care university medical center. We retrospectively reviewed a prospectively maintained thoracic endovascular database, as well as the medical records and imaging studies of all patients who underwent stent graft repair of their thoracic aorta. Preoperative demographics and thoracoabdominal aortoiliac anatomy, intraoperative procedure-related measures, device usage, and postoperative outcomes were assessed. This study had been approved by the Institutional Review Board.
All patients underwent preoperative and postoperative computed tomographic angiography (CTA) (Toshiba Aquilion, Tustin, Calif, 16-64 multislice detectors depending on the year of the study) using a timed-bolus, intravenous contrast technique acquired at 2-3 mm collimations. The dataset was then reconstructed and post-processed using the Aquarius Workstation (TeraRecon, San Mateo, Calif). This was used to assess preoperative aortoiliac morphology, dimensions, branch vessel patency, and postoperative aortic coverage.
An AAA was defined using a maximum transverse aortic diameter of ≥30 mm. Although this definition may seem arbitrary, the 30-mm threshold has numerous precedents in the published literature and has been more commonly used than the 1.5X diameter of adjacent “normal” artery criterion.19, 20, 21, 22, 23 Furthermore, all of the aneurysms showed focal dilation of the aorta and did not simply represent ectasia or arteriomegaly.
SCI was defined as any new lower extremity motor and/or sensory deficit. It was considered “transient” (vs “permanent”) when a clear deficit was documented and then completely reversed to the patient's baseline functional status. Even when a patient's symptoms had improved but his or her functional status was not restored to preoperative levels, the complication was considered “permanent”.
Prophylactic spinal drainage was primarily based on operator preference and no clear protocol was followed during the study period. The only exceptions to this practice included patients who had focal pathologies, such as penetrating ulcers or traumatic transections, and required ≤10 cm of thoracic coverage or had coagulopathy (international normalized ratio, [INR] >1.3) that could not be corrected in a timely manner precluding safe placement of a spinal catheter.
All patients were admitted postoperatively to the Cardiac Intensive Care Unit, where neurologic exams were performed hourly. Upon detection of symptoms, the blood pressure was elevated to a systolic pressure of >160 mm Hg (mean arterial pressure of >100 mm Hg) with vasopressors, and a spinal catheter was promptly (<2 hours of symptoms) placed by a qualified cardiac anesthesiologist (in those patients who were not drained preoperatively). The spinal drainage catheter was placed at 10 cm above the level of the heart and adjusted higher or lower depending on the amount of spinal fluid drainage and therapeutic effect. The catheter was left in place for 72 hours. After reversal of SCI, the catheter was clamped for another 24 hours in case the symptoms returned, and then removed.
Continuous data were analyzed using a two-tailed t test and categorical variables using Fisher's exact test, with P < .05 considered statistically significant. Univariate and multivariate logistic regression models were constructed using a set of 20 potential predictors: age, gender, BMI (body mass index), American Society of Anesthesiologists (ASA) classification, type, maximum diameter, and length of thoracic aortic pathology being treated, endograft type, history of AAA repair, lumbar and hypogastric artery patency, left subclavian artery (LSA) coverage, urgency of the procedure, preoperative spinal drainage, anesthetic technique, iliac conduit, thoracic aortic length (from left common carotid artery to celiac artery) and endograft coverage, duration of procedure, and blood loss. Odds ratio estimates were reported with 95% confidence intervals.
Results
Of the 261 patients who underwent TEVAR, 67% (176/261) were men and the mean age was 66.2 ± 15.2 years with a BMI of 26.8 ± 5.5 kg/m2. The majority of the patients were designated as ASA IV (160/261, 61%) and had either degenerative aneurysms (52%, 136/261) or dissections (21%, 55/261). The TAG endograft (W. L. Gore, Flagstaff, Ariz) was the only commercially-available thoracic aortic device during the study period and, therefore, 78% (204/261) patients were repaired using this system, with the balance comprised of investigational devices. Eighty-eight (34%) of the 261 procedures were performed emergently for acute or symptomatic conditions, such as ruptures or complicated dissections, and general anesthesia was used in 69% (179/261) cases. Iliac conduits were required in 19% (50/261) of cases and only 12% (31/261) had prophylactic spinal drainage. The mean duration of the procedures was 113 ± 67 minutes and blood loss 308 ± 365 mL. Overall, there were 33 intraoperative complications in 27 patients for an event rate of 10.3% and 1.2 events/person (Table I). Postoperatively, there were 177 30-day or in-hospital complications in 92 patients for an overall rate of 35% and 1.9 events/patient. Early mortality was 6.5% (17/261) (Table II).
Table I. Intraoperative complications
| Complication | N (%) |
|---|---|
| Vascular access-related | 18 (6.9) |
| Endograft-related | 2(0.8) |
| Stroke | 6(2.3) |
| Cardiac arrest | 2(0.8) |
| Aortic dissection | 1(0.4) |
| Other1 | 4(1.5) |
1-unplanned LSA occlusion (1), bowel injury during retroperitoneal exposure (1), thromboembolism (1), mesenteric ischemia (1). |
Table II. Postoperative (in-hospital or 30-day) complications
| Complication | N (%) |
|---|---|
| Wound1 | 7(1.5) |
| Bleeding2 | 11(4.2) |
| Cardiac3 | 9(3.4) |
| Pulmonary4 | 25(9.6) |
| Neurologic5 | 43(16.5) |
| Ischemic6 | 18(6.9) |
| Gastrointestinal7 | 9(3.4) |
| Renal8 | 14(5.4) |
| Death | 17(6.5) |
| Other9 | 24(9.2) |
1-access site infections and incisional dehiscence; |
2-pseudoaneurysms and hematomas; |
3-myocardial infarction and arrhythmia; |
4-pneumonia and prolonged respiratory insufficiency; |
5-SCI, strokes, or peripheral neuropathy; |
6-peripheral ischemia from any cause; |
7-ileus, bleeding, or colitis; |
8-decreased estimated glomerular filtration rate and need for dialysis; |
9-infectious, hematologic, vascular, or device-related events. |
SCI occurred in 10.3% (27/261) of cases. Thirteen (48.1%) were completely reversed with spinal drainage, and 14 (51.9%) were permanent. Patients who developed SCI were older (74 ± 9 vs 65 ± 16 years, P = .006), had a higher proportion of men (85% vs 65%, P = .049), lower BMI (23.6 vs 27.0 kg/m2, P = .003), more emergent procedures (52% vs 32%, P = .051) performed under general anesthesia (93% vs 66%, P = .004), and longer operating times (143 ± 79 vs 110 ± 65 minutes, P = .014) than those who did not have SCI. However, there were no significant differences in the ASA class, endograft type, intraoperative blood loss, incidence of LSA coverage, use of iliac conduits, or prophylactic spinal drainage between the two cohorts (Table III).
Table III. Univariate analysis
| Variable | No SCI (n = 234) | SCI (n = 27) | P |
|---|---|---|---|
| Age (years) | 65.2 ± 15.6 | 73.8 ± 8.8 | .006 |
| Males | 153(65%) | 23(85%) | .049 |
| BMI (kg/m2) | 27.0 ± 5.1 | 23.6 ± 8.6 | .003 |
| ASA IV | 140(60%) | 20(74%) | .21 |
| Aortic pathology | .24 | ||
| TAA | 122(52%) | 14(52%) | |
| Dissection | 46(20%) | 9(33%) | |
| Penetrating ulcer | 37(16%) | 4(15%) | |
| Traumatic transection | 15(6%) | 0(0%) | |
| Other | 14(6%) | 0(0%) | |
| Endograft(TAG) | 181(77%) | 23(85%) | .46 |
| Patent LSA | 143(61%) | 18(67%) | .68 |
| Emergent procedure | 74(32%) | 14(52%) | .051 |
| Prophyl spinal drain | 29(13%) | 2(7%) | .75 |
| General anesthesia | 154(66%) | 25(93%) | .004 |
| Iliac conduit | 44(19%) | 6(22%) | .61 |
| Proc duration (minutes) | 110 ± 65 | 143 ± 79 | .014 |
| Blood loss (mL) | 304 ± 380 | 339 ± 196 | .64 |
The mean length of the thoracic aorta for the entire group was 289 ± 39 mm. Patients with SCI had a larger maximum thoracic aortic diameter (P = .014), longer thoracic aortic length (P = .036), and higher fractional endograft coverage of their thoracic aorta (P = .0001) than those without SCI. At the time of their index endovascular thoracic procedures, 39% (101/261) of the patients had either a repaired (51/101) or unrepaired (50/100) AAA. The mean size of the unrepaired AAA in the SCI and no-SCI groups were similar (38.9 ± 9.4 vs 41.2 ± 10.4 mm, P = .27). On univariate analysis, while a prior AAA repair alone was not associated with SCI (odds ratio [OR] = 1.56 [0.63, 3.85], P = .34), a history of either repaired or unrepaired AAA was strongly predictive of SCI (OR = 3.57 [1.56, 8.33], P = .003) (Table IV).
Table IV. Aortoiliac anatomy and SCI
| Variable | No SCI (n = 234) | SCI (n = 27) | Odds Ratio (95% CI) | P |
|---|---|---|---|---|
| AAA (un/repaired) | 83(35%) | 18(67%) | 3.6(1.6,8.3) | .003 |
| History of repaired AAA | 43(18%) | 7(26%) | 1.6(0.6,3.8) | .34 |
| No. patent lumbar arteries | 6.57±2.58 | 5.08±2.62 | 0.83(0.7,0.9)1 | .008 |
| No. patent hypogastric arteries | 1.92±0.32 | 1.85±0.37 | 0.58(0.2,1.6) | .29 |
| Diameter of thoracic pathology(mm) | 60.0±15.3 | 67.7±13.9 | 1.4(1.0,1.9)2 | .045 |
| Length of thoracic pathology (mm) | 123.5±89.4 | 122.2±79.9 | 0.99(0.8,1.3) | .95 |
| Length of thoracic aorta | 282.7±38.8 | 299.2±37.0 | 1.7(0.9,3.2) | .09 |
| Covered thoracic aorta (% total) | 66.1±23.8 | 88.1±12.0 | 1.9(1.3,2.8)3 | .001 |
1-for each additional patent lumbar artery; |
2-for every 10 mm increase in diameter; |
3-for each incremental 10% (∼2.9-cm) thoracic aortic coverage. |
Axial images from 254 preoperative CTA were serially reviewed from the celiac artery to the aortic bifurcation and the number of patent lumbar and hypogastric arteries counted. Seven scans were unavailable for analysis. The mean number of patent lumbar arteries was significantly lower in the SCI group compared to the no-SCI group (P = .008), while hypogastric artery patency was similar (Fig). Univariate analysis showed that each additional patent lumbar artery conferred approximately a 17% reduction in the risk of SCI.
Fig.
Histogram of patent lumbar arteries in the SCI and no SCI groups.
Multivariate logistic regression analysis was performed to eliminate potential confounders. Again, a history of either repaired or unrepaired AAA was highly independently predictive of SCI (OR = 4.35 [1.43, 14.3], P = .10). Other variables which remained independently associated with SCI included extent of thoracic aortic coverage (OR = 2.0 [1.3, 3.1], P = .002, for every 10% or ∼2.9 cm incremental coverage) and general anesthesia (OR = 7.7 [1.6, 37], P = .012).
Discussion
Endovascular repair has become a minimally-invasive alternative to the treatment of a variety of thoracic aortic pathologies with decreased perioperative mortality, respiratory failure, renal insufficiency, and hospital stay compared to open repair.24, 25, 26 Spinal cord ischemia after endovascular repair is a devastating complication whose occurrence has been difficult to predict preoperatively. One of the risk factors that has been previously reported to be associated with SCI has been prior AAA repair.1, 5, 7
In this study, we showed that not prior AAA repair, per se, but rather a history of either repaired or unrepaired AAA was associated with SCI. Our finding is consistent with the recent report by Baril et al1 who also noted an increased incidence of SCI in patients with concomitant AAA or previously repaired AAA (14.3%) vs patients with no AAA (1.0%). In our study, the incidence of SCI in patients with either repaired or unrepaired AAA was 18% (18/101) vs 6% (9/160) in patients without any history of AAA (P = .003). Unlike in aortoiliac occlusive disease where many of the lumbar arteries are frequently patent and enlarged serving as important collateral supply to the pelvis, in patients with AAA one or more lumbar arteries are occluded. Indeed, the mean number of lumbar arteries in those with a history of AAA (repaired or unrepaired) was significantly less than in those without an AAA (4.5 ± 3.0 vs 7.6 ± 1.3, P < .0001), and 78% of the former cohort had at least two occluded lumbar arteries. On the other hand, as shown by the multivariate analysis, decreased lumbar artery perfusion alone does not account for the pathophysiology of SCI in patients with unrepaired AAA.
Above and beyond its association with AAA, patients with SCI had a decreased number of patent lumbar arteries compared to those who did not develop SCI. The incidence of SCI was 22% (13/59) if there were less than five patent lumbar arteries compared to 6.7% (13/195) in those who had five or more (P = .002). A limitation of our CTA-based analysis is that identification of patent arteries is obviously dependent on the technique and quality of the image acquisition. The possibility of inadequate contrast delivery to a lumbar artery resulting in the appearance of an occluded vessel cannot be ruled out, as well as the failure to detect smaller lumbar vessels (<1 mm) and differentiation between direct vs collateral perfusion.
Finally, despite numerous prior publications suggesting an association between LSA coverage and SCI,15, 16, 17, 18, 27 neither our study nor the one by Amabile et al28 found such a correlation. In the current series, SCI occurred in 11% (18/161) of those with and in 9% (9/100) of those without a patent LSA (P = .68). Despite its apparent anatomic basis, the actual contribution to overall spinal cord perfusion by one or both vertebral arteries via the anterior spinal artery is difficult to determine, and we do not believe justifies prophylactic revascularization in every case of LSA coverage. Currently, the main indications in our practice for preoperative subclavian artery bypass include a dominant left vertebral artery with a diminutive right vertebral artery and/or a patent left internal mammary artery (LIMA) graft to the left anterior descending (LAD) coronary artery. To this end, CTA imaging of the intracranial circulation is routinely performed as part of our preoperative imaging protocol. The incidence of symptomatic left arm claudication is extremely small. Of the 115 of 261 (44%) cases that had LSA coverage, 7 had prophylactic preoperative revascularizations and only 2 (1.7%) required LSA bypass for upper extremity ischemia.
Conclusion
Symptoms of SCI can occur in up to 10% of those undergoing TEVAR, only about half of which are permanent. Patients with a history of either repaired or unrepaired AAA face over four times increased risk of developing SCI as compared to those without an AAA and is independent of the length of thoracic aorta covered. Prophylactic measures for spinal cord protection should be considered in the subset of individuals with this preoperative risk factor.
Author contributions
References
- Endovascular thoracic aortic repair and previous or concomitant abdominal aortic repair: is the increased risk of spinal cord ischemia real?. Ann Vasc Surg. 2006;20:188–194
- . Mid-term results for second-generation thoracic stent grafts. Br J Surg. 2003;90:811–817
- . Spinal cord ischemia after elective stent-graft repair of the thoracic aorta. J Vasc Surg. 2005;42:11–17
- Endovascular repair of the thoracic aorta: lessons learned. Ann Thorac Surg. 2005;80:857–863
- . The “first generation” of endovascular stent-grafts for patients with aneurysms of the descending thoracic aorta. J Thorac Cardiovasc Surg. 1998;116:689–703
- Challenges of endovascular tube graft repair of thoracic aortic aneurysm: midterm follow-up and lessons learned. J Vasc Surg. 2003;38:676–683
- Risk of spinal cord ischemia after endograft repair of thoracic aortic aneurysms. J Vasc Surg. 2001;34:997–1003
- Endovascular repair of descending thoracic aortic aneurysms: an early experience with intermediate-term follow-up. J Vasc Surg. 2000;31:147–156
- Endovascular repair of thoracic aortic lesions with the Zenith TX1 and TX2 thoracic grafts: intermediate-term results. J Vasc Surg. 2005;41:589–596
- Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis. J Vasc Surg. 2005;41:1–9
- Mid-term results after endovascular repair of the atherosclerotic descending thoracic aortic aneurysm. Eur J Vasc Endovasc Surg. 2004;28:146–153
- Looking for the artery of Adamkiewicz: a quest to minimize paraplegia after operations for aneurysms of the descending thoracic and thoracoabdominal aorta. J Thorac Cardiovasc Surg. 1996;112:1202–1213
- Postoperative risk factors for delayed neurologic deficit after thoracic and thoracoabdominal aortic aneurysm repair: a case-control study. J Vasc Surg. 2003;37:750–754
- Delayed paraplegia after thoracic and thoracoabdominal aneurysm repair: a continuing risk. Ann Thorac Surg. 2003;75:113–119
- Results of endovascular repair of the thoracic aorta with the Talent Thoracic stent graft: the Talent Thoracic Retrospective Registry. J Thorac Cardiovasc Surg. 2006;132:332–339
- . Management of the left subclavian artery during endovascular repair of the thoracic aorta. J Endovasc Ther. 2008;15:168–176
- . Utility of left subclavian artery revascularization in association with endoluminal repair of acute and chronic thoracic aortic pathology. J Vasc Surg. 2006;43:433–439
- Early results of endovascular treatment of the thoracic aorta using the Valiant endograft. Cardiovasc Intervent Radiol. 2007;30:1130–1138
- . Abdominal aortic aneurysm as an incidental finding in abdominal ultrasonography. Br J Surg. 1991;78:1261–1263
- . Opportunistic screening for abdominal aortic aneurysm. J Med Screen. 1994;1:220–222
- Yield of repeated screening for abdominal aortic aneurysm after a 4-year interval. Arch Intern Med. 2000;160:1117–1121
- . The long-term benefits of a single scan for abdominal aortic aneurysm (AAA) at age 65. Eur J Vasc Endovasc Surg. 2001;21:535–540
- Expansion rates and outcomes for the 3.0-cm to the 3.9-cm infrarenal abdominal aortic aneurysm. J Vasc Surg. 2002;35:666–671
- . Endovascular stent grafting versus open surgical repair of descending thoracic aortic aneurysms in low-risk patients: a multicenter comparative trial. J Thorac Cardiovasc Surg. 2007;133:369–377
- Endoluminal versus open treatment of descending thoracic aortic aneurysms. J Vasc Surg. 2002;36:732–737
- Stent-graft versus open-surgical repair of the thoracic aorta: mid-term results. J Vasc Surg. 2006;44:1188–1197
- Neurologic complications associated with endovascular repair of thoracic aortic pathology: incidence and risk factors (A study from the European Collaborators on Stent/Graft Techniques for Aortic Aneurysm Repair (EUROSTAR) registry). J Vasc Surg. 2007;46:1103–1110
- . Incidence and determinants of spinal cord ischaemia in stent-graft repair of the thoracic aorta. Eur J Vasc Endovasc Surg. 2008;35:455–461
Competition of interest: none.
CME article
PII: S0741-5214(08)01633-9
doi:10.1016/j.jvs.2008.08.119
© 2009 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.