Spinal cord ischemia may be reduced via a novel technique of intercostal artery revascularization during open thoracoabdominal aneurysm repair

Methods 

We retrospectively reviewed our records for patients having received this procedure. Overall, eight patients were repaired under this manner in the course of one year (2005-2006). Patient ages ranged from 46 to 70 with a mean of 56 years of age. All patients had a previous aortic dissection with aneurysmal degeneration from the left subclavian artery to the visceral segment (type I TAAA). Indications for repair included symptomatic pain, rupture, rapid enlargement, or size greater than 6.5 cm.

All operations were approached from a left thoracoabdominal incision. Cerebrospinal fluid drainage was utilized in all cases. With the exception of one case, all were performed under deep hypothermic circulatory arrest (DHCA) due to lack of a proximal clamp site and a difficult proximal reconstruction. A standard Dacron graft was used to replace the aorta as an end-end anastomosis proximally and a beveled anastomosis distally. The aorta was replaced from the distal arch to the visceral vessels. An 8 mm limb was sewn end-side to the proximal graft prior to opening the aorta. Under circulatory arrest or partial bypass the aorta was opened and the proximal and distal anastomoses were performed. While rewarming the patient, the 8 mm limb was opened longitudinally and then sutured to multiple intercostal artery pairs in a side-side manner. The distal end of the 8 mm limb was then sutured to the distal graft in an end-side manner with a side-biting clamp on the aortic graft to allow continuous aortic perfusion (Fig 1, Fig 2). Intraoperative electroencephalograms (EEG) and upper and lower somatosensory evoked potentials (SSEP) were monitored throughout the procedure. In cases employing deep hypothermic circulatory arrest, patients were cooled on cardiopulmonary bypass and 10-channel EEG was used to confirm electrocerebral silence prior to the initiation of circulatory arrest at an average nasopharyngeal temperature of 14 degrees Celsius.

View Large Image Download to PowerPoint

Fig 1. A graphic representation of the intercostal patch graft.

View Large Image Download to PowerPoint

Fig 2. An intraoperative photograph of the intercostal patch graft.

Patients were followed by history and physical exam at 1 month postprocedure. They were subsequently followed at 6-month intervals with a history, physical exam, and a CT scan. Follow-up ranged from 6 to 14 months.

Results 

All patients were ambulatory at the time of admission although one patient reported mild unilateral leg weakness from a previous spinal cord infarction from a type A dissection. The average aneurysm size was 7.5 cm. Four patients were symptomatic upon presentation.

Seven operations were performed with deep hypothermic circulatory arrest. One case was performed under left atrial-left femoral artery bypass. Upon opening the aorta, there was an average of seven pairs of patent intercostal arteries. We reimplanted an average of five pairs of intercostal arteries extending to the T10-T11 level. The remainder of the patent intercostal arteries was over sewn.

In all seven cases involving DHCA, the typical pattern of burst-suppression of the EEG was followed by electrocerebral silence with cooling. There was also the typical pattern of loss of the cortical evoked potentials from both the arms and legs. In four cases recovery of continuous, symmetrical EEG and cortical SSEP was seen following rewarming after circulatory arrest. In three cases of DHCA spinal SSEP did not recover with rewarming and in the one LA-FA case, SSEP were lost with aortic clamping These changes in the spinal SSEP were consistent with spinal ischemia. Prompt reimplantion of intercostal arteries with the described technique was undertaken. Following reimplantation, there was a return of lower extremity spinal SSEP to baseline. These patients had a normal postoperative exam.

There were no perioperative complications. No patients suffered any significant hemodynamic changes postoperatively. No patients were taken back to the operating room for bleeding. No patients suffered transient or permanent paraplegia during the postoperative or follow-up periods. All patients were discharged ambulating. The one patient who had unilateral leg weakness reported subjective improvement in strength of the affected leg.

Follow-up ranged from 6 months to 14 months. All patients remained ambulatory throughout the follow-up period. There was no evidence of pseudoaneurysm formation or patch aneurysm formation. All intercostal reconstructions remained patent through follow-up as demonstrated by CT (Fig 3, Fig 4, Table). Furthermore, follow-up CTA demonstrated persistent flow into intercostal arteries from the graft (Fig 5, Table).

View Large Image Download to PowerPoint

Fig 3. CT angiogram demonstrating an aortic graft with a patent intercostal patch graft.

View Large Image Download to PowerPoint

Fig 4. Three-dimensional reconstruction demonstrating patency in the follow-up period.

Table. Latest follow-up imaging study after the procedure

	Patient	Follow-up study	Graft	Patent intercostals
	1	12moCTA	patent	5
	2	3moCT	patent	N/A
	3	14moCTA	patent	4
	4	6moCT	patent	N/A
	5	3moCT	patent	N/A
	6	6moCTA	patent	7
	7	12moCTA	patent	5
	8	3moCTA	patent	5

Either a traditional CT scan or CT angiogram was performed. Intercostal patency was only able to be reliably assessed on the CT angiograms.

View Large Image Download to PowerPoint

Fig 5. CT angiogram demonstrating patent intercostal arteries off the patch graft.

Discussion 

Repair of thoracoabdominal aortic aneurysms has been associated with substantial morbidity. Due to the anatomic location and extent of disease, paraplegia is a considerable risk. Despite maneuvers to reduce this risk, such as distal aortic perfusion, cerebrospinal fluid drainage, elevation of mean arterial pressure, and intraoperative neuromonitoring, the risk of paraplegia remains significant.6, 9, 10, 11, 12, 13, 14, 18 As a result, we have devised a novel technique to improve spinal perfusion during TAAA repair.

We applied this technique to patients who had previous aortic dissections and subsequent aneurysmal degeneration. Traditionally, patients with previous abdominal aortic repairs and patients requiring full aortic replacement (type II TAAA) have been at highest risk for paraplegia.5, 19 Although, some studies have not demonstrated chronic dissection as a risk factor for paraplegia, it is our experience, and that of others, that these patients with previous aortic dissections seem to be at increased risk.20, 21 This may be due to decreased thrombus formation within the aneurysm sac. This is likely due to a combination of flow dynamics within the true and false lumen as well as rapid aneurysmal enlargement which decreases the amount of time for static flow and thrombus formation. As a result, there are often more patent intercostal arteries.15 This may result in a less mature or well-formed collateral circulation pathway to the spinal cord and increased dependence upon direct perfusion of the intercostal arteries.

In order to maximize adequate spinal cord perfusion during and after repair, attempts have been made to identify the artery of Adamkiewicz or the great radicular artery via multiple imaging techniques.22, 23 Nevertheless, reimplantation of these vessels has not yielded markedly reduced rates of paraplegia.15, 16, 17 In addition, exclusion of the Adamkiewicz artery does not necessarily correlate with spinal cord dysfunction.24 Rather, reimplantation of as many intercostal arteries as possible, whether or not they feed the Adamkiewicz or great radicular artery, may be critical. Furthermore, in cases of aortic dissection and subsequent aneurysmal degeneration, there can often be many patent intercostal arteries. Creating separate bypass grafts to each vessel or vessel pair would be cumbersome and not feasible. As a result, reimplanting multiple branches as one patch would seem most beneficial in terms of reperfusing the spinal cord and minimizing ischemic times. In general, we sought to reimplant as many intercostals as possible at the T6-T11 level. This also served to potentially revascularize the Artery of Adamkiewicz, which is most often found in this location.25 Because of the potential increased outflow resistance due to the small caliber of these vessels and subsequent potential for thrombosis, we reconnected the distal aspect of the side-limb to the main tube graft. As a result, the outflow for the side-limb becomes that of the tube graft (the entire lower body). Based upon the conservation of the principles of inertia and energy, antegrade flow is maintained within the side-limb allowing for a continuous flow loop and continuous flow to the intercostal arteries. Also, by avoiding a direct anastomosis of the main tube graft to the intercostals, aortic ischemia is minimized. Aortic perfusion is restored immediately after the distal aortic anastomosis is completed. While reimplanting the intercostal arteries, there is continuous aortic perfusion through the main tube graft. In addition, the distal end-side anastomosis of the side-limb is performed with a side-biting clamp which does not interrupt aortic flow. As a result, continued perfusion through spinal collaterals (internal iliacs, viscerals, vertebrals) is maintained.

We found that this technique was able to be performed with technical success. The reconstructions did not lengthen the case as they were performed during patient rewarming. Postoperatively, no patients required replacement of spinal drains after their removal. Furthermore, no patients required blood pressure augmentation due to lower extremity weakness. All reconstructions remained patent throughout the follow-up period. All patients also had full preservation of lower extremity strength and sensation.

The importance of reimplanting multiple intercostal vessels was recognized in this study. Persistent loss of spinal SSEP from the legs has been shown to be predictive of lower extremity neurologic dysfunction.12 In our cohort, four patients had evidence of a persistent intraoperative injury on lower extremity SSEP during the aortic repair. This change is the typical finding of ischemia and dysfunction of the spinal sensory pathways from the legs. Shortly following reimplantation of intercostal arteries there was return of the lower extremity SSEP to baseline. It is our experience and others that patients developing persistent or transient spinal SSEP changes are typically at high risk for developing postoperative spinal ischemia.12 These patients had no evidence of spinal cord ischemia upon arousal from anesthesia nor did the patients develop an episode of postoperative spinal cord ischemia. It is possible that reimplantation not only reversed these patients’ intraoperative ischemia but also added an additional level of postoperative protection. It is certainly possible and likely that these patients would have had significant neurologic dysfunction without reimplantation.

Conclusion 

Paraplegia is a devastating complication of TAAA repair. Moreover, certain types of TAAA are at increased risk for paraplegia when repaired. Many techniques have been devised for eliminating spinal cord injury. While these techniques have decreased the incidence, spinal cord ischemia still occurs at an unacceptably high rate. We report a technique which improves spinal cord perfusion and may help reduce this incidence.

Author contributions