Magnetic resonance angiography of collateral blood supply to spinal cord in thoracic and thoracoabdominal aortic aneurysm patients

Objective

Preservation of spinal cord blood supply during descending thoracic (TAA) and thoracoabdominal aortic aneurysm (TAAA) surgery is mandatory to prevent neurologic complications. Although collateral arteries have been identified occasionally and are considered crucial for maintaining spinal cord function in the individual patient, their critical functionality is poorly understood and very little experience exists with visualization. This study investigated whether the preoperative and postoperative presence or absence of collateral arteries detected by magnetic resonance angiography (MRA) is related to spinal cord function during the intraoperative exclusion of the segmental supply to the Adamkiewicz artery.

Methods

Spinal cord MRA was used to localize the Adamkiewicz artery and its segmental supplier in 85 patients scheduled for open elective surgery for TAA or TAAA. The segmental artery to the Adamkiewicz artery was inside the cross-clamped aortic area in 55 patients, and spinal cord supply was consequently dependent on collateral supply. In these 55 patients the presence of collaterals originating from arteries outside the cross-clamped aortic segment was related to changes in the intraoperative motor-evoked potentials (MEPs) that occurred before corrective measures. Twenty-one patients returned for postoperative MRA.

Results

A highly significant (P < .0015) relation was found between the presence of collaterals and intraoperative spinal cord function. In 30 of 31 patients (97%) in whom collaterals were identified, MEPs remained stable. The collaterals in most patients originated caudally to the distal clamp (eg, from the pelvic arteries), which were perfused by means of extracorporeal circulation during cross-clamping. The MEPs declined in 9 of 24 patients (38%) in whom no collaterals were preoperatively visualized. Postoperatively, the 21 patients who had MRA, including 10 in whom preoperatively no collaterals were found, displayed a well-developed collateral network.

Conclusion

Collateral arteries supplying the spinal cord can be systematically visualized using MRA. Spinal cord blood supply during open aortic surgery may crucially depend on collateral arteries. Preoperatively identified collateral supply was 97% predictive for stable intraoperative spinal cord function. Patients in whom no collaterals can be depicted preoperatively are at increased risk for spinal cord dysfunction.