The risk of ischemic spinal cord injury in patients undergoing graft replacement for thoracoabdominal aortic...

We developed a monitoring system to detect spinal cord ischemia during aortic cross-clamping (AXC). This system was used to prospectively determine in which patients ischemia occurs, in which patients reimplantation of intercostal arteries is unnecessary or mandatory, and when reperfusion of intercostal arteries (ICAs) is urgent.

Two hundred sixty patients underwent thoracoabdominal aortic aneurysm (TAA) repair with simple AXC. In 167 patients, two electrocatheters were placed before the onset of anaesthesia at level L1/L2 (stimulation) and level T5/T6 (recording) within the epidural space. During surgery, spinal cord function was monitored by recording spinal somatosensory evoked potentials (sSSEP). According to the extent of aortic replacement, most patients were expected to have a high risk of paraplegia.

In group A (59 patients), sSSEP remained normal throughout surgery, and in 54 of these patients ICAs were not reattached outside the proximal aortic anastomosis. In the other five patients ICAs were reimplanted separately because of possible anatomic relation to spinal cord blood supply. No patient in group A had postoperative neurologic deficit. In group B (54 patients) sSSEP remained normal until 15 minutes after AXC but were impaired thereafter. Nineteen patients had early reimplantation of ICAs. Of the 19, three had paraparesis and two had paraplegia. Neurologic deficit developed in the patients without early reimplantation of ICAs. In four patients separate reimplantation of ICAs was performed late in the procedure because of incomplete sSSEP recovery. Subsequently, the sSSEP returned to normal and only one of the four patients had mild paraparesis. The total rate of neurologic deficits in this group was 13% (paraplegia, 3.5%; paraparesis, 9.5%). All 54 patients in group C showed rapid loss of sSSEP within 15 minutes of AXC. In 28 patients ICAs were reimplanted only within the proximal anastomosis. Twenty-one of these patients showed prompt signal recovery after blood-flow release into the reimplanted ICAs, and none had neurologic deficit. Seven patients had no or very late and incomplete sSSEP recovery. Of the seven, three had paraplegia and four had paraparesis. In 26 patients ICAs were reimplanted separately to the proximal anastomosis. This was done early during the procedure in 17 patients, of whom 13 had full recovery of sSSEP and normal neurologic status. Four patients had incomplete or no recurrence of sSSEP, followed by paraplegia in one and paraparesis in three. In nine patients ICAs were reimplanted after the aortic replacement had been completed because of sSSEP recovery was not satisfactory. In all patients in this subgroup sSSEP returned to normal. Six patients had a normal neurologic status and three had mild paraparesis. The total neurologic complication rate in group C was 26% (paraplegia, 7.5%; paraparesis, 18.5%).

The risk of ischemic spinal cord injury during replacement for TAA can be assessed continuously by monitoring the sSSEP directly from the spinal cord. Patients without sSSEP changes during aortic reconstruction do not require ICA reattachment and will not have neurologic deficit. Patients who lose sSSEP after AXC are at risk for paraplegia. Patients with impairment or loss of sSSEP >15 minutes after AXC have some collateral vessels, and must have ICAs reimplanted only if sSSEP do not return within normal recovery time after blood-flow release into the proximal anastomosis. Loss of sSSEP within 15 minutes of AXC shows poor collateralization and mandates early restoration of spinal cord blood supply. If the surgeon can achieve the return of sSSEP to normal by subsequent separate reimplantation of ICAS, paraplegia will not occur and paraparesis will be rare and mild. Spinal cord monitoring is a valuable guide to detect whether the spinal cord is at risk and to take measures against par