Annals of Vascular Surgery
Volume 24, Issue 4 , Pages 550.e5-550.e9, May 2010

Paraparesis after Thoracic Stent-Graft Relining for an Unrecognized Type III Endoleak

  • David Volders

      Affiliations

    • Department of Vascular Surgery, University Hospital Leuven, Leuven, Belgium
    • ,
  • Inge Fourneau

      Affiliations

    • Department of Vascular Surgery, University Hospital Leuven, Leuven, Belgium
    • Corresponding Author InformationCorrespondence to: Inge Fourneau, MD, PhD, Department of Vascular Surgery, University Hospital Leuven, Herestraat 49, B-3000 Leuven, Belgium
    • ,
  • Kim Daenens

      Affiliations

    • Department of Vascular Surgery, University Hospital Leuven, Leuven, Belgium
    • ,
  • Sabrina Houthoofd

      Affiliations

    • Department of Vascular Surgery, University Hospital Leuven, Leuven, Belgium
    • ,
  • Geert Maleux

      Affiliations

    • Department of Interventional Radiology, University Hospital Leuven, Leuven, Belgium
    • ,
  • André Nevelsteen

      Affiliations

    • Department of Vascular Surgery, University Hospital Leuven, Leuven, Belgium

published online 02 February 2010.

Article Outline

Background

We examined the reasons for missing a type III endoleak on conventional imaging and the pathophysiology of paraparesis after relining this stent graft.

Methods and Results

A 46-year-old man was treated with a thoracic stent graft for thoracic rupture of a chronic type B thoracoabdominal dissection with aneurysm formation. In a second intervention, retrograde revascularization of the visceral and renal arteries was performed in combination with insertion of an abdominal stent graft. After initial shrinkage of the aneurysmal sac, the thoracic aortic diameter started increasing again. Consecutive three-phase helical computed tomographic scans did not reveal any endoleak. Because of unbearable back pain, an open surgical exploration was performed. This showed a type III endoleak. Relining of the thoracic stent graft was performed, but paraparesis developed.

Conclusion

In patients with unexplained increase of the aneurysmal sac contrast-enhanced magnetic resonance imaging could help to illuminate the underlying endoleak. The collateral network concept can explain spinal cord injury by even minor hemodynamic changes.

 

Back to Article Outline

Case Report 

A 46-year-old man was transferred to our university hospital because of exacerbating acute thoracic back pain. At arrival the patient was in a stable hemodynamic condition. Contrast-enhanced computed tomographic (CT) scanning showed a thoracoabdominal aneurysm type II, based on a chronic type B thoracoabdominal dissection. The maximum diameter in the descending aorta was 80 mm. The maximum diameter in the supraceliac thoracic segment was 40 mm and that in the infrarenal abdominal segment was 86 mm. There was a contained rupture at the level of the sixth thoracic vertebra. There were no signs of malperfusion of the viscera. He had a history of morbid obesity in spite of a gastric bypass 2 years before and bad tolerance of effort.

To avoid extracorporeal circulation and deep hypothermia in this unfit patient, a hybrid therapy was planned in two stages, giving priority to the rupture. In a first stage an 8 mm silver-coated Dacron carotid–carotid–subclavian bypass (InterGard Silver®; Intervascular, La Ciotat, France) was performed as an emergency procedure, immediately followed by deployment of three thoracic stent grafts (Valiant® straight graft; Medtronic, Minneapolis, MN) from just distal to the brachiocephalic trunk to the level of the twelfth thoracic vertebra, proximal to the celiac trunk.

Two weeks later, a control CT scan showed a further increase of the maximal thoracic diameter to 100 mm as a result of the persistent distal dissection and a stable maximum abdominal diameter. The residual abdominal aneurysm was treated as an elective procedure. First, through a median laparotomy, an iliorenal–hepatic bypass was performed to the right using an 8 mm Dacron graft (Gelsoft®; Vascutek, Inchinnin, UK). Second, an iliorenal–superior mesenteric bypass was performed to the left using an 8 mm Dacron graft (Gelsoft). The procedure was finished with extension of the previously deployed thoracic stent graft to the common iliac arteries by an abdominal bifurcation stent graft (Valiant straight and Talent® bifurcated graft, Medtronic) (Fig. 1). The patient was discharged 1 month after this second intervention in good condition. Recovery was uneventful.

Follow-up was planned with three-phase helical CT scan and plain X-ray.

Three months after the first procedure, a control CT scan showed good function of the stent grafts with complete thrombosis of the false lumen, a slight reduction of the maximal diameter of both the descending and abdominal aorta (89 and 81 mm, respectively), no signs of endoleak, and good opacification of the visceral bypasses. A control CT scan at 7 months confirmed further shrinkage of the false lumen in both the thoracic and the abdominal aorta (86 and 68 mm, respectively).

Fourteen months after the first procedure, a new control CT scan showed an increased maximal diameter of the descending aorta to 93 mm with appearance of mural thrombus in the true lumen on the right (Fig. 2). No endoleaks were observed. The maximal diameter of the abdominal aorta was decreased to 60 mm. At this point no intervention was planned.

A new control CT scan 4 months later showed a further increase of the maximal diameter of the thoracic aorta to 98 mm without endoleaks and further enlargement of the mural thrombus. Another control CT scan 1 month later showed no further increase of the maximal diameter but enlargement of the mural thrombus. Since the patient was asymptomatic and there were no signs of endoleak, this was considered as endotension and we decided to follow the patient carefully with 3-monthly CT scans.

Another control CT scan 21 months after the first intervention showed a further increase of the maximal diameter of the descending aorta to 110 mm, still without detectable endoleak. Erosive damage of the anterior side of the ninth and tenth thoracic vertebrae was detected (Fig. 3). At none of these time points did plain X-ray show any sign of stent fracture or migration.

  • View full-size image.
  • Fig. 3 

    Transverse CT section through the ninth thoracic vertebra (venous phase) at 7 months (A) and 21 months (B) showing increase of the diameter of the descending aorta and development of erosion of the vertebra.

Twenty-two months after the first admission, our patient presented with unbearable paravertebral pain. A new three-phase helical CT showed enlargement of the known mural thrombus on the right of the true lumen of the descending aorta as the only change compared to the same investigation 1 month earlier. The tentative diagnosis was endotension with erosion of the vertebra.

At this time we decided to intervene and to drain the aneurysm in order to relieve the pressure on the vertebral column. We had no evidence at all for a structural graft defect. A left thoracotomy was performed. Due to extensive stickiness of the lung, a resection of the lower left lobe was needed. After incision of the thoracic aneurysm, the thrombotic content was removed and major arterial bleeding occurred, which could be stopped only by clamping the endoprosthesis. Inspection showed a major (>2 mm) fabric disruption in the graft material (endoleak type III). The hole was located in the middle portion of the second thoracic stent graft at the transition zone with the proximal bare stent of the third stent graft. This endoleak was sutured with polypropylene 3/0 reinforced with Teflon patches and spouted with a fibrin glue (Tissucol®; Baxter, Vienna, Austria). It was felt safer to try to reline the original stent graft with a new stent graft. However, it was impossible to introduce a new stent graft, either by a left or a right femoral approach, due to severe tortuosity and loss of flexibility at the transition zone with the abdominal stent graft.

Two days later, the patient developed acute thoracic hemorrhage with hemodynamic shock. The patient was intubated, resuscitated, and immediately brought to the operating theater, where we made another attempt to reline the thoracic stent graft. Through a sternotomy, a 10 mm Dacron graft (Gelsoft) was sutured on the ascending aorta as a conduit for two Talent stent grafts that relined the previously inserted thoracic stent graft precisely.

After extubation on the sixth postoperative day, paraparesis with sensibility disturbances extending to the back and the abdomen was obvious. Magnetic resonance imaging (MRI) of the spine at 1 month showed no defects in the spinal cord. MRI of the skull did not reveal an explanation for the neurological defects.

Over the next months, the patient improved through multidisciplinary revalidation. The sensibility and strength in both legs improved. At the moment the patient is able to walk with walkers. The paravertebral pain disappeared. Every 6 months a follow-up three-phase helical CT scan is performed. The last CT showed good position of the stent grafts without endoleaks and with good vascularization of the viscera. The aneurysmal dilatation of the descending aorta remained stable at a maximum diameter of 98 mm.

Back to Article Outline

Discussion 

This case drew our attention for three reasons. First, endoleak type III is a rare type of endoleak, especially if caused by a fabric tear. Second, how could that the diagnosis of a type III endoleak be missed in spite of close follow-up with conventional imaging? Third, why did paraparesis develop after relining of the thoracic stent graft, although no additional side branches were sacrificed?

Type III endoleaks occur when there is a structural failure of the stent graft. This includes stent-graft fractures, fabric tears, and junctional separations seen with modular devices. However, due to continuous improvement of the stent grafts, type III endoleaks are currently quite unusual. Type III endoleaks are responsible for 20% of all endoleaks.1 We hypothesize that the fabric tear was caused by repetitive friction between the fabric of the second stent graft and the bare stent of the third stent graft. To our knowledge this is the first case of a type III endoleak due to a fabric tear in a Valiant stent graft.

For our patient, follow-up was organized with three-phase helical CT scan and plain X-ray. Helical CT angiography is a standard method for the postprocedural assessment of aortic stent grafts.2 However, there is evidence that MRI is superior to CT angiography at detecting endoleaks, at least for nitinol stent grafts. The Valiant stent graft we used consists of a nitinol wire scaffolding attached to a low-profile polyester monofilament weave and is suitable for MRI-based surveillance.3 According to Pitton et al.,4 the sensitivity of endoleak detection in nitinol stent grafts is 92.9%, 44.0%, 34.8%, and 38.3% for contrast-enhanced MRI, biphasic CT, uniphasic arterial CT, and uniphasic late CT, respectively. The corresponding negative predictive values are 91.7%, 58.4%, 54.7%, and 56.1%. A review by Stavropoulos and Charagundla in 20075 on imaging techniques for the detection and management of endoleaks after endovascular aortic aneurysm repair concluded that although CT angiography is currently the standard imaging modality for endoleak detection, contrast-enhanced MRI and ultrasound will continue to have an expanding role in the future. As in our patient CT angiography showed an unexplained increase of the aneurysm diameter, contrast-enhanced MRI could have been a good supplement to the diagnostic work-up and could have helped to illuminate the phenomenon of “endotension” in our patient.

Endotension is defined as a persistent or recurrent pressurization of the aneurysm sac following endovascular repair. Gilling-Smith et al.6 recognized three types of endotension: grade I (high flow), grade II (low flow), and grade III (no flow). They hypothesized that one of the possible explanations for grade III endotension was transmission of pressure through the thrombus that seals an endoleak. During surgery in our patient we found thrombus sealing the type III endoleak as an explanation for why the endoleak was not visualized on consecutive CT scans. However, as pressure was transmitted through this thrombus, the aneurysm diameter started increasing after an initial tendency to shrink.

Spinal cord injury is the most dreaded complication of repair of descending thoracic and thoracoabdominal aneurysms and is often explained by sacrifice of critical intercostal and lumbar arteries. The incidence of paraplegia/paraparesis after thoracic stent-graft procedures as reported in the literature varies between 1.5% and 8.7%, the most important risk factor being the length of the covered aorta.7, 8, 9 However, our patient did not develop paraplegia after the first intervention, although all intercostal and lumbar arteries had been covered at that moment, but 2 years later during relining of an existing stent graft without sacrifice of additional arteries. According to the literature, delayed-onset paraplegia hours or even weeks postoperatively accounts for up to one-third of cases of postoperative permanent spinal cord injury.10

To study potential explanations for this phenomenon, Etz et al.11 retrospectively looked at 10 cases of paraplegia that developed within 48 hr after surgical intervention despite intact somatosensory evoked potentials throughout the operation and compared them with 10 matched control patients who recovered without spinal cord injury. They hypothesized that differences in postoperative management distinguished those patients who subsequently had paraplegia from those who recovered seemingly normal function and that paraplegia can be caused by minor differences in postoperative hemodynamic and fluid management. Notice that the indication for semiurgent relining of the thoracic stent graft was hemodynamic and respiratory instability.

That even quite subtle changes in blood pressure can affect the development of spinal cord injury fits in the collateral circulation concept as synthesized by Griepp and Griepp12 and based on many laboratory studies as well as clinical experience.9, 13, 14 After surgical sacrifice of segmental arteries, perfusion of the spinal cord depends on the stabilization of a collateral network fed from below by the hypogastric arteries and from above by the internal thoracic artery and other branches from the subclavian arteries. This explains why routine sacrifice of segmental aortic branches can be justified in surgical and endovascular therapy, without risking postoperative neurological damage. On the other hand, the inflow of such a collateral network is very precarious and depends principally on arterial pressure, which is largely determined by cardiac output, blood volume, and the competing demands of viscera and muscle tissue connected to the same collateral network. This concept is also supported by the findings of Buth et al.9 that perioperative paraplegia or paraparesis is significantly associated with blockage of the left subclavian artery without revascularization.

In our patient the first stent graft started distal of the truncus brachiocephalicus and ended on the level of the capping of thoracic vertebra 12. The second one relined precisely the first one without sacrificing extra vessels. We hypothesize that with the placement of the first stent graft, the sacrifice of segmental arteries was compensated for by the collateral network and no neurological disturbances occurred. However, hemodynamic instability preceding the relining procedure disturbed this precarious balance in blood supply to the spinal cord. The absence of neurological disturbances after a thoracic stent-graft procedure therefore does not absolutely exclude the possibility of spinal cord ischemia during relining of the same stent graft.

Perfusion-weighted MRI has been used increasingly in the investigation of acute cerebral infarcts.15 Unfortunately, there are few reports of the spinal cord. However, Yanaka et al.16 demonstrated vasogenic edema of the spinal cord using perfusion-weighted MRI. We suggest that transient hypoperfusion of the spinal cord can create vasogenic edema, which can resolve once the circulation has restored itself. This would explain why the paraparesis was transient.

There are several studies suggesting techniques to reduce the chances of spinal cord ischemia. Jacobs and Mess17 described in 2003 that the combination of monitoring motor-evoked potentials, cerebrospinal fluid drainage, distal aortic perfusion, and epidural cooling could prevent neurological deficit in 98% of patients with open thoracoabdominal aortic aneurysm repair. They confirmed this in an extensive study in 2006.18 Rigorous perioperative hemodynamic and fluid management could help to overcome delayed-onset paraplegia.11 Aware of these findings, we always use prophylactic cerebrospinal fluid drainage when covering the distal third of the descending aorta. However, in this patient we did not as we wrongly felt that a simple relining of the graft would have no consequences on the medullar circulation.

In conclusion, a fabric tear resulting in a type III endoleak can happen. In patients with unexplained increase of the aneurysmal sac, contrast-enhanced MRI could help to illuminate the underlying endoleak. The collateral network concept can explain spinal cord injury by even minor hemodynamic changes.

Back to Article Outline

References 

  1. Parmer S, Carpenter J, Stavropoulos W, et al. Endoleaks after endovascular repair of thoracic aortic aneurysms. J Vasc Surg. 2006;44:447–452
  2. Thurnher S, Cejna M. Imaging of aortic stent-grafts and endoleaks. Radiol Clin North Am. 2002;40:799–833
  3. van der Laan M, Bartels L, Bakker C, et al. Suitability of 7 aortic stent-graft models for MRI-based surveillance. J Endovasc Ther. 2004;11:366–371
  4. Pitton MB, Schweitzer H, Herber S, et al. MRI versus helical CT for endoleak detection after endovascular aneurysm repair. AJR Am J Roentgenol. 2005;185:1275–1281
  5. Stavropoulos SW, Charagundla SR. Imaging techniques for detection and management of endoleaks after endovascular aortic aneurysm repair. Radiology. 2007;243:641–655
  6. Gilling-Smith G, Brennan J, Harris P, et al. Endotension after endovascular aneurysm repair: definition, classification, and strategies for surveillance and intervention. J Endovasc Surg. 1999;6:305–307
  7. Khoynezhad A, Donayre C, Bui H, et al. Risk factors of neurologic deficit after thoracic aortic endografting. Ann Thorac Surg. 2007;83:S882–S889
  8. Fairman R, Criado F, Farber M, et al. Pivotal results of the Medtronic vascular Talent thoracic stent graft system: the VALOR trial. J Vasc Surg. 2008;48:546–554
  9. Buth J, Harris PL, Hobo R, et al. 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
  10. Wong DR, Coselli JS, Amerman K, et al. Delayed spinal cord deficits after thoracoabdominal aortic aneurysm repair. Ann Thorac Surg. 2007;83:1345–1355
  11. Etz CD, Luehr M, Kari FA, et al. Paraplegia after extensive thoracic and thoracoabdominal aortic aneurysm repair: does critical spinal cord ischemia occur postoperatively?. J Thorac Cardiovasc Surg. 2008;135:324–330
  12. Griepp RB, Griepp EB. Spinal cord perfusion and protection during descending thoracic and thoracoabdominal aortic surgery: the collateral network concept. Ann Thorac Surg. 2007;83:865–869
  13. Jacobs MJ, de Mol BA, Ellenbaas T, et al. Spinal cord blood supply in patients with thoracoabdominal aortic aneurysms. J Vasc Surg. 2002;35:30–37
  14. Schurink GW, Nijenhuis RJ, Backes WH, et al. Assessment of spinal cord circulation and function in endovascular treatment of thoracic aortic aneurysms. Ann Thorac Surg. 2007;83:S877–S881
  15. Neumann-Haefelin T, Wittsack HJ, Fink GR, et al. Diffusion- and perfusion-weighted MRI: influence of severe carotid artery stenosis on the DWI/PWI mismatch in acute stroke. Stroke. 2000;31:1311–1317
  16. Yanaka K, Matsumaru Y, Uemura K, et al. Perfusion-weighted MRI of spinal dural arteriovenous fistula. Neuroradiology. 2003;45:744–747
  17. Jacobs MJ, Mess WH. The role of evoked potential monitoring in operative management of type I and type II thoracoabdominal aortic aneurysms. Semin Thorac Cardiovasc Surg. 2003;15:353–364
  18. Jacobs MJ, Mess W, Mochtar B, et al. The value of motor evoked potentials in reducing paraplegia during thoracoabdominal aneurysm repair. J Vasc Surg. 2006;43:239–246

PII: S0890-5096(09)00325-2

doi:10.1016/j.avsg.2009.08.016

Annals of Vascular Surgery
Volume 24, Issue 4 , Pages 550.e5-550.e9, May 2010

Article Tools

* Image Usage & Resolution