CASE REPORTS

Salvage options after stent collapse in thethoracic aortaCaroline D. Rodd, MB Bch, FRCS, Sharmini Desigan, MB BCh, BAO (NUI), LRCP, SI, FRCS, FRCR,Mohamad S. Hamady, MD, FRCR, Richard G. Gibbs, MD, FRCS, and Michael P. Jenkins, BSc MS,FRCS, London, United Kingdom

The endograft was originally developed to repair aneurismal disease of the infra-renal aorta and has since realised manyother applications, including the treatment of arterial trauma. Traumatic transection of the thoracic aorta is a conditionassociated with a high mortality and affected patients often have multiple injuries. Endovascular repair of thoracictransection is an attractive option in those patients for whom open surgical repair would be highly dangerous and othergroups have reported early technical success. However, we report 3 cases of young patients with traumatic thoracic aortictransection, initially treated successfully by endoluminal stenting, who developed the complication of stent collapse. We

discuss here the options available to treat the complication. (J Vasc Surg 2007;46:780-5.)

The use of the endovascular technique for the treat-ment of arterial trauma was initially reported by Parodi in1999.1 The endovascular technique had originally beendeveloped for the high risk patient requiring an infra-renalaortic aneurysm repair. The endoluminal option has dra-matically altered the approach to aneurysm repair through-out the vascular tree, from the aorta to the external iliacvessels. As experience developed, so the applications of thetechnique have broadened to include traumatic pathologyin addition to the aneurysmal pathology.2-4

We report 3 cases of young patients with traumaticthoracic aortic transection, initially treated successfully byendoluminal stenting, who developed the complication ofstent collapse. We discuss here the options available to treatthe complication.

CASE REPORT

Case 1. A 20-year-old man who had been involved in a roadtraffic accident presented with multiple injuries, including multiplerib fractures, bilateral lung contusions, a left hemothorax, fracturesof the right iliac blade and acetabulum, and a thoracic aortictransection. The native thoracic aorta was 22.7 mm in diameter. (Inthis case and the following cases, measurements of the aorta wereobtained from multidetector computed tomography angiography[MDCTA]. The measurements were taken from one inner wall tothe other inner wall of the aorta, in the axial plane.) The patientproceeded to endoluminal stent grafting through a right femoral

From the Regional Vascular Unit, St. Mary’s NHS Trust, London.Competition of interest: none.Reprint requests: Caroline D. Rodd, MB Bch, FRCS, Regional Vascular

Unit, St Mary’s NHS Trust, Praed Street, London W2 1NY, UK (e-mail:Caroline.rodd@glos.nhs.uk).

0741-5214/$32.00Copyright © 2007 by The Society for Vascular Surgery.

doi:10.1016/j.jvs.2007.03.059

780

artery cutdown. A Gore TAG (W.L. Gore & Associates, Inc,Flagstaff, Ariz) graft (28 � 15 mm) was placed that covered the leftsubclavian artery origin. A good technical result was achieved andcan be seen in the postoperative MDCTA images in Fig 1. The28-mm graft was used because it was the smallest available in thedepartment at the time, and the patient was too unstable fortreatment to be delayed. Adjunctive ballooning was not performedin this case or any of the following cases.

After a period of ventilation in the intensive care unit, thepatient made an uncomplicated recovery. He was reviewed in theoutpatient department 5 months after discharge and was found tobe clinically well; results of examination of his vascular system werenormal. In our institution, all patients undergoing endovascularaneurysm repair undergo routine computed tomographic surveil-lance with scans before discharge, at 6 months and 1 year afterstenting, and annually thereafter if there are no complications. The6-month surveillance MDCTA scan of this patient revealed partialcollapse of the thoracic stent graft (Fig 2).

The patient was immediately readmitted, and a decision wasmade to explant the stent graft. An endovascular solution for thisproblem was initially considered but then dismissed for two rea-sons. First, the stent appeared collapsed during study of the com-puted tomographic images. Second, the patient was now medicallyfit, and we considered that he would be able to tolerate a thora-cotomy. He underwent a clamshell thoracotomy on bypass withopen removal of the stent graft (Fig 3) and aortic tube graft repairof the aortic arch and descending thoracic aorta (Fig 4). Theprocedure was performed with the patient on full cardiac bypass; asplit circuit was used with cannulas in the femoral artery and aorta.At operation, the stent was found to be partially incorporated intothe aortic wall, and extensive dissection was required to enableextraction of the endograft. The proximal aortic cross-clamp hadto be placed proximal to the left common carotid artery to facilitate

stent explantation, after which the clamp was moved to a position

JOURNAL OF VASCULAR SURGERYVolume 46, Number 4 Rodd et al 781

distal to the left common carotid artery but proximal to the leftsubclavian artery origin. The patient made a good recovery and wasdischarged on day 10.

Case 2. A 17-year-old woman had been a front-seat passen-ger in a road traffic accident and sustained multiple injuries,including bilateral lung contusions, a right-sided pneumothoraxnecessitating a chest drain, a splenic laceration, a small liver lacer-ation, a pelvic fracture with hematoma, and a closed head injury. Inaddition, she had a transection of the thoracic aorta. She wasreferred as an emergency transfer from another region. However,just before transfer, she developed circulatory collapse and re-quired on-site laparotomy and splenectomy. She was transferredonce stable. The native thoracic aorta was measured at 21.5 mm indiameter from the MDCTA. We placed two overlapping GoreTAG stents (26 � 10 mm and 26 � 15 mm) into the aortic arch

Fig 1. Multidetector computed tomography angiography imagesof a stent deployed in the thoracic aorta.

(covering the left subclavian artery origin) via a left transfemoral

approach (Fig 5). During deployment of the first endograft, thestent jumped back as a result of a “wind sock effect,” thus resultingin incomplete coverage of the target area. Hence, a second over-lapping stent was deployed.

Her associated injuries necessitated a prolonged admission tothe intensive care unit. However, there were no stent-relatedcomplications, and she was subsequently discharged and followedup in her local hospital. Eleven months later, she presented to herlocal medical team with vomiting and abdominal pain. Investiga-tions revealed that she was profoundly acidotic with a base deficit of�15 and was also 16 weeks pregnant. A transthoracic echocardio-gram suggested that there was a 100 mm Hg pressure gradient acrossher thoracic stent, and she was transferred back to our care forinvestigation. She denied any vascular symptoms, and on clinicalexamination of the vascular system, all upper limb pulses werepresent, but the femoral pulses were reduced. A chest radiographshowed that the proximal end of the stent was partially collapsedwith central crimping, thus producing a double-barreled appear-ance of the stent (Fig 6). A transthoracic echocardiogram con-

Fig 2. Multidetector computed tomography angiography imageof a collapsed stent deployed in the thoracic aorta.

Fig 3. Collapsed stent in the thoracic aorta.

firmed these findings and the pressure gradient of 100 mm Hg

JOURNAL OF VASCULAR SURGERYOctober 2007782 Rodd et al

across the stent. She was reviewed and fully investigated by boththe obstetric and medical teams. A history of possible anorexia andvomiting was elicited. No other cause for the acidosis was identi-fied, and the acidosis was attributed to hyperemesis related toeither pregnancy or anorexia nervosa. The acidosis resolved afterintravenous rehydration and an improved dietary intake.

In view of her pregnancy, open thoracotomy was not consid-ered to be a safe treatment option. However, we were concernedthat the combination of the hyperdynamic state of pregnancy andan impaired outflow tract might compromise the pregnancy.

She therefore proceeded to endoluminal ballooning of theproximal end of the collapsed stent with deployment of aPalmaz stent (P40414; Cordis Europa N.V., Roden, TheNetherlands). The Palmaz stent was mounted over a MAXI LDballoon (Cordis Europa N.V.), and the balloon was inflated to 2atm with an inflation device. Controlled hypotension (mean pres-sure, 90-100 mm Hg) was induced during the balloon inflationand stent deployment in the arch, as is our routine practice. Thecollapsed stent was successfully opened, and intraoperative esoph-ageal echocardiogram demonstrated that the pressure gradientacross the stent had disappeared.

We had obtained advice from the radiation protection adviserfor our institution before the operation regarding the fetus. As aresult, a lead apron (0.25 mm Pb equivalent) was applied to herabdomen and pelvis during the procedure. Her pregnancy anddelivery were otherwise uneventful. We remain uncertain as towhether the placement of the Palmaz stent will be a long-termsolution for this patient or whether the original endograft has anincreased chance of collapsing again. This patient may ultimatelyrequire open surgery and stent removal. We continue to follow her upwith regular transthoracic echocardiograms, and the most recent scanshows no evidence of a pressure gradient across the stent.

Case 3. A 32-year-old man was admitted with multiple inju-ries after a fall of 40 feet from a balcony. His injuries includedmultiple bilateral rib fractures with a flail segment, a left hemotho-rax and right pneumothorax, bilateral lung contusions, traumaticright-sided lung cysts, a fractured wrist, and a wedge fracture of the12th thoracic vertebral body associated with a fracture of itstransverse process. He also had a traumatic transection of his

Fig 4. Completed open repair of the thoracic aorta.

thoracic aorta just distal to the left subclavian artery origin (Fig 7).

The native thoracic aorta was measured at 23.2 mm fromMDCTA. He underwent a technically successful endoluminalstent insertion (Gore TAG; 28 � 15 mm) to the aortic arch(covering the left subclavian artery origin) and descending aorta viaa right transfemoral approach. As in case 1, the 28-mm-diametergraft was used because it was the smallest available in our depart-ment at that time, and he was too unstable for treatment to bedelayed.

An MDCTA scan was performed on day 4, and this confirmedthe endograft to be in a good position. Despite initial clinicalimprovement, the patient developed progressive resistant hyper-

Fig 5. Multidetector computed tomography angiography imagesof a stent deployed in the thoracic aorta.

tension in the intensive care unit 7 days later. An MDCTA revealed

JOURNAL OF VASCULAR SURGERYVolume 46, Number 4 Rodd et al 783

that the endograft had collapsed (Fig 8). Again, an endovascularsolution for this problem was considered and reluctantly dismissedon this occasion because the proximal end of the stent graftappeared very distorted and fully collapsed when studied on thecomputed tomographic images. The patient immediately pro-ceeded to a clamshell thoracotomy on bypass, explantation of thestent graft (Fig 9), and an interposition aortic tube graft repair ofthe aortic arch and descending thoracic aorta. His postoperativecourse was difficult because of his associated injuries; however, hehas now been discharged home and is making good progress.

DISCUSSION

Traumatic transection of the thoracic aorta is a bluntdeceleration injury, usually occurring as a result of a motorvehicle accident.5,6 Injuries of the thoracic aorta result fromthe effect of shear forces at points of anatomic fixity—eg,the left subclavian artery (65%), arch (10%), descendingaorta (12%), or multiple sites (13%).7 The injury has anassociated mortality of more than 80%,8 with an overallmortality of thoracoabdominal aortic traumatic injury of94% to 98% in the first 2 hours.6,9 Those patients reachingthe hospital fall into two groups: those who will remainstable and can be treated conservatively10 and those whorequire surgery.11-14 Of those treated conservatively, 42%may develop evidence of aneurysmal disease within 5 yearsof injury, and 85% may do so within 20 years.15

The reported mortality of open repair varies from 16%to 31%,5,14,16-18 and associated morbidity such as paraple-gia varies from 8.7% to 19%.5,16,17,19 As shown by ourcases, there are often significant associated injuries, such asliver laceration (61%), head injuries (42%), multiple ribfractures (78%), heart laceration (34%), splenic laceration(36%), and thoracic spine fractures (20%).6 Therefore, anendovascular repair of thoracic transection is an attractiveoption in patients who have other life-threatening comor-bidities and for whom open surgical repair would be highlydangerous, particularly given the necessity of anticoagula-

Fig 6. Chest x-ray image of double-barreled appearance of acollapsed stent in the thoracic aorta.

tion during the procedure.

Many other groups have had early technical successwith this technique, with few complications and no mortal-ity using Talent (Medtronic, Minneapolis, Minn),3,20

Gore,20 Min Tec (Boston Scientific, Natick Mass),3,20 andGianturco Z (Cook Inc, Bloomington, Ind)21 stents.However, recently there have been reports of late stentcollapse.22 Thoracic transection can be treated with en-doluminal stent placement with good technical success butseems to have a risk of subsequent early stent collapse. Oncethis has occurred, the stent either must be removed viaopen thoracotomy or can be reopened as a temporizingmeasure with an endoluminal approach. Even in youngpatients, an open thoracotomy can be associated with mor-tality and morbidity. If the collapse occurs in the early

Fig 7. Multidetector computed tomography angiography imagesof thoracic aortic transection.

postdeployment period, these risks will be higher, because the

JOURNAL OF VASCULAR SURGERYOctober 2007784 Rodd et al

patient may have significant ongoing polytrauma comorbid-ity. Although two of our cases have undergone open surgicalremoval of the stent, we were forced to consider anotheroption for the pregnant woman, which to date has beensuccessful and more durable than we anticipated.

The group of patients with thoracic transections is notthe same as those who undergo thoracic stenting for aneu-rysms or dissections of the thoracic aorta. The patients areoften younger, and their aortas are normal, with none ofthe associated calcification, stiffness, and atherosclerosisassociated with intrinsic aortic disease. The aortic diameteris smaller, and it has been proposed that the collapse issecondary to oversizing. The oversizing in these cases was23.6%, 21.4%, and 20.95% for cases 1, 2, and 3, respec-tively; these results are similar to those in other reports.22

Although it is conceivable that oversizing should beavoided in these cases, it is not feasible for most depart-ments to stock a range of stents for aortas of approximately20 mm in diameter. The other possible explanation forstent collapse may be that the hemodynamics in this group,

Fig 8. Multidetector computed tomography angiography imageof a collapsed stent in the thoracic aorta.

Fig 9. Explanted collapsed stent.

with smaller and more compliant vessels after trauma, are

very different from those in the patients with older, larger,calcified aneurysmal vessels. In addition, the aortic arch ofyoung patients has a more acute angle, and it is possible thatthe endograft is unable to conform tightly to the innercurvature of the arch. If the proximal stent and aortic wallhave poor apposition, a type Ia endoleak will result, and thismay cause the collapse of the endograft. The collapse can beeither an early or late complication.

Until now, we have treated thoracic transectionswith stents that were designed for a different pathology.We require a new range of grafts that are smaller andmore compliant and have lower profiles to meet theneeds of the younger patient with a smaller, more peaked,and more compliant aortic arch and smaller access vessels.

We also need to consider the use of adjunctive technol-ogy to assist with fixation of the proximal endograft in thesepatients. We have already used a Palmaz stent as a secondaryprocedure, but we should be prepared to use this techniqueas an adjunct in the primary procedure to avoid rather thantreat the complication. Alternative solutions to proximalgraft fixation problems may include modifications to theproximal end of the stent and perhaps the use of endo-staples. However, the aorta in these patients is very thinwalled, and such fixation might result in a risk of perfora-tion. The deployment of multiple, shorter grafts in theaorta may offer an increase in resistance and reduce the riskof stent collapse. However, currently available stents aretoo large, and such a practice will increase the problemswith oversizing. Developments in technology are necessaryto refine the highly desirable endovascular treatment op-tion for patients with multiple injuries and aortic transec-tion.

REFERENCES

1. Parodi JC, Schonholz C, Ferreira LM, Bergan J. Endovascular stent-graft treatment of traumatic arterial lesions. Ann Vasc Surg 1999;13:121-9.

2. Bell RE, Taylor PR, Aukett M, Sabharwal T, Reidy JF. Results of urgentand emergency thoracic procedures treated by endoluminal repair. EurJ Vasc Endovasc Surg 2003;25:527-31.

3. Rousseau H, Soula P, Perreault P, Bui B, Janne DB, Massabuau P, et al.Delayed treatment of traumatic rupture of the thoracic aorta withendoluminal covered stent. Circulation 1999;99:498-504.

4. Taylor PR, Gaines PA, McGuinness CL, Cleveland TJ, Beard JD,Cooper G, et al. Thoracic aortic stent grafts—early experience from twocentres using commercially available devices. Eur J Vasc Endovasc Surg2001;22:70-6.

5. Fabian TC, Richardson JD, Croce MA, Smith JS Jr, Rodman G Jr,Kearney PA, et al. Prospective study of blunt aortic injury: MulticenterTrial of the American Association for the Surgery of Trauma. J Trauma1997;42:374-80.

6. Williams JS, Graff JA, Uku JM, Steinig JP. Aortic injury in vehiculartrauma. Ann Thorac Surg 1994;57:726-30.

7. Wall MJ Jr, Hirshberg A, LeMaire SA, Holcomb J, Mattox K. Thoracicaortic and thoracic vascular injuries. Surg Clin North Am 2001;81:1375-93.

8. Smith RS, Chang FC. Traumatic rupture of the aorta: still a lethal injury.Am J Surg 1986;152:660-3.

9. Kalmar P, Puschel K, Stubbe HM, Gultekin E. Delayed surgical therapyof acute aortic rupture [in German]. Zentralbl Chir 1996;121:750-5.

10. Fisher RG, Oria RA, Mattox KL, Whigham CJ, Pickard LR. Conserva-tive management of aortic lacerations due to blunt trauma. J Trauma

1990;30:1562-6.

JOURNAL OF VASCULAR SURGERYVolume 46, Number 4 Rodd et al 785

11. Kipfer B, Leupi F, Schuepbach P, Friedli D, Althaus U. Acute traumaticrupture of the thoracic aorta: immediate or delayed surgical repair? EurJ Cardiothorac Surg 1994;8:30-3.

12. Stulz P, Reymond MA, Bertschmann W, Graedel E. Decision-makingaspects in the timing of surgical intervention in aortic rupture. EurJ Cardiothorac Surg 1991;5:623-7.

13. Pate JW, Fabian TC, Walker W. Traumatic rupture of the aorticisthmus: an emergency? World J Surg 1995;19:119-25.

14. Akins CW, Buckley MJ, Daggett W, McIlduff JB, Austen WG. Acutetraumatic disruption of the thoracic aorta: a ten-year experience. AnnThorac Surg 1981;31:305-9.

15. Finkelmeier BA, Mentzer RM Jr, Kaiser DL, Tegtmeyer CJ, Nolan SP.Chronic traumatic thoracic aneurysm. Influence of operative treatmenton natural history: an analysis of reported cases, 1950-1980. J ThoracCardiovasc Surg 1982;84:257-66.

16. Cowley RA, Turney SZ, Hankins JR, Rodriguez A, Attar S, Shankar BS.Rupture of thoracic aorta caused by blunt trauma. A fifteen-year expe-rience. J Thorac Cardiovasc Surg 1990;100:652-60.

17. von Oppell UO, Dunne TT, De Groot MK, Zilla P. Traumatic aorticrupture: twenty-year metaanalysis of mortality and risk of paraplegia.

Ann Thorac Surg 1994;58:585-93.

18. Schmidt CA, Wood MN, Razzouk AJ, Killeen JD, Gan KA. Primaryrepair of traumatic aortic rupture: a preferred approach. J Trauma1992;32:588-92.

19. Nocolosi AC, Almassi GH, Bousamra M, Haasler GB, Olinger GN.Mortality and neurologic morbidity after repair of traumatic aorticdisruption. Ann Thorac Surg 1996;61:875-8.

20. Scheinert D, Krankenberg H, Schmidt A, Gummert JF, Nitzsche S,Scheinert S, et al. Endoluminal stent-graft placement for acute ruptureof the descending thoracic aorta. Eur Heart J 2004;25:694-700.

21. Lobato AC, Quick RC, Phillips B, Vranic M, Rodriguez-Lopez J,Douglas M, et al. Immediate endovascular repair for descending tho-racic aortic transection secondary to blunt trauma. J Endovasc Ther2000;7:16-20.

22. Steinbauer MG, Stehr A, Pfister K, Herold T, Zorger N, Topel I, et al.Endovascular repair of proximal endograft collapse after treatment forthoracic aortic disease. J Vasc Surg 2006;43:609-12.

Submitted Dec 8, 2006; accepted Mar 1, 2007.