Delayed permanent paraplegia after endovascularrepair of abdominal aortic aneurysmLee J. Goldstein, MD,a Combiz Rezayat, MD,b Gautam V. Shrikhande, MD,b

and Harry L. Bush Jr, MD,b New York, NY; and Miami, Fla

Spinal cord ischemia is a rare complication after abdominal aortic surgery and has been attributed to surgicaldevascularization of the spinal cord, atheroembolization of the cord circulation, or hypoperfusion of cord structuressecondary to hypotension or cord edema. We present a diabetic, hypertensive 75-year-old male with endstage renal diseasewho presented with a 5.5 cm asymptomatic infrarenal abdominal aortic aneurysm, and concomitant 3.5 cm right commoniliac artery aneurysm. After undergoing successful endovascular repair with an aorto-uni-iliac device, unilateral hypogas-tric artery embolization, and femoral-femoral bypass, he was discharged to a rehabilitation facility neurologically intactwith a stage 2 decubitus ulcer. He returned on postoperative day 21 with a large stage 4 septic decubitus ulcer, fever,leukocytosis, hypotension, and paraplegia. We hypothesize that the compromised blood flow from the initial reconstruc-tion, combined with the delayed hypotension imposed by sepsis, resulted in spinal cord infarction. He was eventuallydischarged to a nursing facility with no improvement in his neurologic status. We report the first case of significantly

delayed permanent paraplegia after endovascular abdominal aortic aneurysmorrhaphy. (J Vasc Surg 2010;51:725-8.)

Endovascular aneurysm repair (EVAR) has become thefirst-line therapy for elective repair of abdominal aorticaneurysms (AAA) due to the significant decrease in periop-erative morbidity of the procedure compared to the opensurgical technique. Spinal cord ischemia after traditionalopen abdominal aortic surgery is a rare and devastatingcomplication occurring in approximately 0.25% of cases.1

The incidence of neurologic complications after EVAR issimilarly low, with the Eurostar Database finding an inci-dence of 0.21% and other smaller EVAR series listing noneurologic complications.2-4

Several mechanisms have been proposed for acute para-plegia after abdominal aortic surgery, including surgicalinterruption of the vascular supply to the spinal cord,atheroembolization of debris into the pelvic circulationsupplying the cord, and cord hypoperfusion secondaryto intraoperative and postoperative hypotension or cordedema.1,5-7 Each of these mechanisms contributes to avarying degree in each case, with the particular cause ineach case depending heavily on whether open or endovas-cular techniques were employed. We present the only re-port of significantly delayed, permanent paraplegia afterEVAR in a patient presenting 3 weeks postoperatively withsepsis and hypotension from a sacral decubitus wound.

From the Division of Vascular Surgery, The DeWitt Daughtry FamilyDepartment of Surgery, Leonard M. Miller School of Medicine, Univer-sity of Miami Jackson Memorial Medical Center,a and the Division ofVascular Surgery, New York Presbyterian Hospital, Weill Cornell MedicalCollege.b

Competition of interest: none.Reprint requests: Harry L. Bush Jr, MD, New York Presbyterian Hospital,

Weill Cornell Medical College, Division of Vascular Surgery, 525 E 68thStreet, P708, New York, NY 10065 (e-mail: hlb2001@med.cornell.edu).

The editors and reviewers of this article have no relevant financial relationshipsto disclose per the JVS policy that requires reviewers to decline review of anymanuscript for which they may have a competition of interest.

0741-5214/$36.00Copyright © 2010 by the Society for Vascular Surgery.

doi:10.1016/j.jvs.2009.09.023

CASE REPORT

A 75-year-old man with a medical history significant fordiabetes mellitus, coronary artery disease, hypertension, peripheralvascular disease, and endstage renal disease (secondary to polycys-tic kidney disease) recently started on hemodialysis, presented witha 5.5 cm asymptomatic infrarenal AAA, and concomitant 3.5 cmright common iliac artery aneurysm. He had previously undergonecoronary artery bypass grafting and mechanical aortic valve replace-ment. A preoperative computed tomography (CT) scan demon-strated a fusiform aneurysm starting approximately 3 cm below thelowest (right) renal artery. The infrarenal aortic neck was ectaticmeasuring 32-34 mm throughout its length, precluding placementof standard endovascular grafts (Fig 1). The region of the aortafrom the top of the left renal artery to the bottom of the superiormesenteric artery (SMA) orifice measured 11 mm in length, and30-32 mm in diameter, which was considered an acceptable prox-imal fixation site, with planned coverage of both renal arteries. Asthe distal aorta was too severely narrowed and calcified to accom-modate a bifurcated graft, the operative plan included the use of anaorto-uni-iliac (AUI) graft, with femoral-femoral bypass, and con-tralateral common iliac artery occlusion. Open surgical repair ofthe aneurysm was not attempted due to the patient’s multipleco-morbid conditions, and the patient’s desire to begin homeperitoneal dialysis.

Under general anesthesia, bilateral transverse groin cutdownswere used to expose the common femoral arteries. The proximalright hypogastric artery was occluded using embolization coils(Cook Medical, Bloomington, Ind). A 36 mm AUI endograft(Zenith Renu, Cook Medical) was deployed starting proximally atthe orifice of the SMA (with the “suprarenal” bare metal stentsextending above the SMA) and landing distally in the right com-mon iliac artery. There was no patent inferior mesenteric arteryseen before graft deployment. The endograft was extended using12 mm limb extensions to land in the right external iliac artery,thus treating the right common iliac artery aneurysm. A 20 mmZenith iliac occluder plug was then deployed to occlude the left

common iliac artery. Finally, a femoral-femoral bypass was per-

725

JOURNAL OF VASCULAR SURGERYMarch 2010726 Goldstein et al

formed to restore circulation to the left lower extremity and pelvis.At the completion of the procedure no endoleak was observed,pedal pulses were present in both extremities, and retrograde fillingof the left hypogastric artery was demonstrated angiographically(Fig 2, A, B). The entire procedure took approximately 6 hours tocomplete.

On the morning of the first postoperative day, the patient wasnoted to have a 10 cm stage 1 sacral decubitus ulcer that had notbeen present preoperatively. The wound was treated conservativelywith wound care and weight shifting, and the patient was dis-charged to a subacute rehabilitation facility on postoperative day10 after being anticoagulated for his mechanical aortic valve. Ondischarge, the decubitus ulcer was stage 2, the patient was neuro-logically intact, and ambulating with assistance (using a walker dueto mild deconditioning and discomfort from his sacral decubituswound).

The patient presented in septic shock on postoperative day 21to an outside facility with fever, rigors, hypotension (70/50 mmHg), leukocytosis (23,300/uL) and bilateral lower extremityparesis. After he was stabilized hemodynamically and grossly de-brided, he was transferred to our institution for further manage-ment approximately 2.5 days later. On arrival, he had a foul-smelling, grossly infected, 17 � 10 cm stage 4 sacral decubitusulcer extending onto the right buttock. His neurologic examconsisted of paraplegia, areflexia, and loss of sensation at the T8level. Spinal cord drainage was not attempted due to the period oftime that the patient had been paraplegic at presentation to ourinstitution. Both lower extremities remained well perfused, and theleft hypogastric artery remained patent by CT angiogram (Fig 3). Amagnetic resonance imaging (MRI) scan confirmed diffuse edemaof the distal spinal cord consistent with ischemia. The patient

Fig 1. Preoperative non-contrast coronal abdominal computedtomography (CT) scan demonstrating the ectatic infrarenal aorticneck above the abdominal aortic aneurysm.

endured a prolonged recovery from his sepsis including multiple

operative sacral debridements, a diverting colostomy, a tracheos-tomy, and a percutaneous endoscopic gastrostomy tube. He wasdischarged to a skilled nursing facility on hospital day 124 withoutimprovement in his neurologic exam. On discharge, the sacralwound bed was lined with healthy granulation tissue and main-tained with a vacuum dressing.

DISCUSSION

Spinal cord ischemia (SCI) after abdominal aortic sur-gery is an infrequent catastrophe and is associated with ahigh mortality. Potential factors contributing to the devel-opment of acute spinal cord ischemia include prolongedaortic clamping (and concurrent cord ischemia), intraoper-ative hypotension, and coverage or interruption of theartery of Adamkiewicz (greater radicular artery).8 Delayedspinal cord ischemia has been postulated to be due toreperfusion, cord edema, and subsequent hypoperfusion ofthe spinal cord.9 Systemic hypotension resulting in cordhypoperfusion has also been implicated in the pathogenesisof SCI. Cerebrospinal fluid drainage and raising spinal cordperfusion pressure have successfully treated these mecha-nisms of ischemia.10

EVAR involves routine coverage of the inferior mesen-teric artery (IMA) and lumbar arteries, and instrumentationof the aneurysmal vasculature with stiff wires and devices,yet SCI after infrarenal EVAR remains exceptionally un-common. This is presumably due to the placement ofendografts at a safe distance from the artery of Adamk-iewicz (originating between T8 and T12 in over 80% ofpatients) and the preservation of the hypogastric circulationin most patients.6,11 Additional factors implicated in SCIafter EVAR include intentional or inadvertent coverage ofone or both hypogastric arteries, and atheroembolizationdue to wire manipulation.5,12,13 This patient had two sig-nificant anatomic compromises to the spinal cord circula-tion: the endograft was situated more proximally than iscustomary (originating at the level of the SMA potentiallycovering more lumbar arteries), and one hypogastric arterywas electively embolized at its origin. There were no prom-inent lumbar arteries or a patent IMA noted on initialangiography before endograft deployment (Fig 4).

Spinal cord ischemia has been reported after EVAR bythe Eurostar collaborators in nearly the same frequency asfor open repair (0.21% for EVAR vs 0.26% for open sur-gery).3 A higher incidence of SCI has been observed inruptured AAA (rEVAR) using an AUI device (as used inour patient). Spinal cord ischemia occurred in 11.5% in aseries of 35 rEVAR patients, with a significantly higher riskincurred by coverage of a hypogastric artery with an AUIdevice.14 Peppelenbosch et al14 hypothesized that the useof an AUI device prolongs the functional aortic occlusiontime during completion of the femoral-femoral bypass andretrograde reperfusion of the contralateral hypogastric ar-tery. The time required for this portion of the procedureprolongs the potential ischemic insult to a compromisedspinal cord. This EVAR technique should be avoided whenpossible, especially in the presence of hemorrhagic shock or

contralateral pelvic embolization. The embolization of the

JOURNAL OF VASCULAR SURGERYVolume 51, Number 3 Goldstein et al 727

right hypogastric artery and dependence on retrogradeperfusion of the left hypogastric artery compromised theperfusion of the pelvis and spinal cord in our patient,perhaps contributing to the development of the early sacraldecubitus ulcer.

SCI after elective EVAR has been reported as an imme-diate finding upon discontinuation of general or regionalanesthesia.5,6 Several reports also exist of patients emergingfrom elective surgery intact, and then acutely developingvarying degrees of lower extremity paralysis on the opera-tive day.8,10,11 Bajwa et al12 successfully reversed the symp-

Fig 2. A, Completion digital subtraction angiogram deaneurysm (AAA) and flow into the external iliac artery aftsuperior mesenteric artery (SMA). B, Digital subtractiofemoral arterial sheath demonstrating effective occlusionpatency of the left hypogastric artery.

Fig 3. Postoperative axial contrast-enhanced computed tomogra-phy (CT) scan demonstrating continued patency of the left hypogas-tric artery (arrow) after the patient presented with paraplegia.

toms by implementing cerebrospinal fluid drainage to raise

the perfusion pressure of the spinal cord. In all cases, thepatients eventually ambulated successfully. Bhama et al12

and Kwok et al13 reported delayed SCI occurring approxi-mately 2 days after the elective procedure. Kwok et al13

reported bilateral lower extremity weakness with bowel andbladder incontinence in a patient with bilateral hypogastricembolizations. These symptoms resolved by 1 month.Bhama et al12 reported bilateral flaccid paralysis treatedwith cerebrospinal fluid drainage, intravenous steroid ther-apy, and pharmacologic augmentation of the blood pres-sure. That patient demonstrated mild improvement in neu-rologic function after 8 months.

Reid et al15 reported the only case of significantlydelayed (greater than 2 weeks) temporary paraplegia afterEVAR in a 74-year-old woman with preoperative imagingsuggestive of fibromuscular dysplasia of the renal arteries.She underwent EVAR with a residual type I endoleak at thedistal right limb seen on CT scan on postoperative day 4.On postoperative day 19, she returned with a 2-day historyof nausea, vomiting, renal failure, dehydration, and bilat-eral lower limb flaccidity. After 3.5 years, the endoleakhad spontaneously resolved, the patient remained hemo-dialysis-dependent, and ambulated with a cane.

This case describes the only known report of signifi-cantly delayed permanent paraplegia after EVAR. The timecourses exhibited by our patient and that reported by Reidet al15 reiterate the precarious blood supply to the spinalcord after elective coverage of the IMA and lumbar arteriesafter EVAR. Both patients were suffering from renal failure

strating endovascular exclusion of the abdominal aorticcement of aorto-uni-iliac (AUI) device. Arrow indicatesiogram performed via retrograde injection from a left

he left common iliac artery via the occluder device, and

moner plan angof t

and exhibited acute hypoperfusion secondary to dehydra-

JOURNAL OF VASCULAR SURGERYMarch 2010728 Goldstein et al

tion or sepsis. As evidenced by Peppelenbosch et al14 afterrEVAR, and by Kwok et al,13 the hypogastric collateralcirculation is critical to the perfusion of the spinal cord afterEVAR. The combination of a high proximal graft place-ment, and unilateral hypogastric artery embolization left atenuous collateral supply to this patient’s spinal cord. Thehypotension secondary to the sacral decubitus ulcer sepsisreduced the perfusion to an already compromised circula-tion resulting in spinal cord infarction.

This report highlights the importance of the fragilecirculation to the spinal cord that results after EVAR.Specifically in cases after hypogastric embolization, highplacement of grafts, or ruptures with significant hemor-rhagic shock, the spinal cord circulation is even furthercompromised. Similar reports have documented how main-tenance of spinal cord perfusion pressure with appropriatesystolic blood pressure control is essential, and that cere-

Fig 4. Lateral aortogram before deployment of the aorto-uni-iliac endograft. There were no prominent lumbar arteries noted inthe suprarenal segment of aorta to be covered by the endograft.

brospinal fluid drainage or intravenous steroid therapy may

be helpful adjuncts in the acute setting to reverse symp-toms.

REFERENCES

1. Szilagyi DE, Hageman JH, Smith RF, Elliott JP. Spinal cord damage insurgery of the abdominal aorta. Surgery 1978;83:38-56.

2. Lee CW, Kaufman JA, Fan CM, Geller SC, Brewster DC, Cambria RP,et al. Clinical outcome of internal iliac artery occlusions during endo-vascular treatment of aortoiliac aneurysmal diseases. J Vasc InterventRadiol 2000;11:567-71.

3. Berg P, Kaufmann D, van Marrewijk CJ, Buth J. Spinal cord ischaemiaafter stent-graft treatment for infra-renal abdominal aortic aneurysms.Analysis of the Eurostar database. Eur J Vasc Endovasc Surg 2001;22:342-7.

4. Zarins CK, White RA, Schwarten D, Kinney E, Diethrich EB, HodgsonKJ, Fogarty TJ. AneuRx stent graft versus open surgical repair ofabdominal aortic aneurysms: multicenter prospective clinical trial. JVasc Surg 1999;29:292-305; discussion 306-8.

5. Rockman CB, Riles TS, Landis R. Lower extremity paraparesis orparaplegia subsequent to endovascular management of abdominal aor-tic aneurysms. J Vasc Surg 2001;33:178-80.

6. Garcia ND, Tehrani H, Morasch M, Pearce W, Matsumura J. Spinalcord ischemia following endovascular repair of an infrarenal aorticaneurysm. Ann Vasc Surg 2002;16:509-12.

7. Rosenthal D. Spinal cord ischemia after abdominal aortic operation: is itpreventable? J Vasc Surg 1999;30:391-7.

8. Riess KP, Gundersen SB 3rd, Ziegelbein KJ. Delayed neurologic deficitafter infrarenal endovascular aortic aneurysm repair. Am Surg 2007;73:385-7.

9. Naslund TC, Hollier LH, Money SR, Facundus EC, Skenderis BS 2nd.Protecting the ischemic spinal cord during aortic clamping. The influ-ence of anesthetics and hypothermia. Ann Surg 1992;215:409-15;discussion 415-6.

10. Bajwa A, Davis M, Moawad M, Taylor PR. Paraplegia following electiveendovascular repair of abdominal aortic aneurysm: reversal with cere-brospinal fluid drainage. Eur J Vasc Endovasc Surge 2008;35:46-8.

11. Fortes DL, Atkins BZ, Chiou AC. Delayed paraplegia following infra-renal abdominal aortic endograft placement: case report and literaturereview. Vascular 2004;12:130-5.

12. Bhama JK, Lin PH, Voloyiannis T, Bush RL, Lumsden AB. Delayedneurologic deficit after endovascular abdominal aortic aneurysm repair.J Vasc Surg 2003;37:690-2.

13. Kwok PC, Chung TK, Chong LC, Chan SC, Wong WK, Chan MK,Chu WS. Neurologic injury after endovascular stent-graft and bilateralinternal iliac artery embolization for infrarenal abdominal aortic aneu-rysm. J Vasc Interv Radiol 2001;12:761-3.

14. Peppelenbosch N, Cuypers PW, Vahl AC, Vermassen F, Buth J. Emer-gency endovascular treatment for ruptured abdominal aortic aneurysmand the risk of spinal cord ischemia. J Vasc Surg 2005;42:608-14.

15. Reid JA, Mole DJ, Johnston LC, Lee B. Delayed paraplegia afterendovascular repair of abdominal aortic aneurysm. J Vasc Surg 2003;37:1322-3.

Submitted Jul 27, 2009; accepted Sep 12, 2009.