Effects of propofol, propofol-nitrous oxide and midazolam on cortical somatosensory evoked potentials during sufentanil anaesthesia for major spinal surgery.

Recording of cortical somatosensory evoked potentials (CSEP) enables monitoring of spinal cord function. We studied the effects of propofol, propofol-nitrous oxide or midazolam during sufentanil anaesthesia on CSEP monitoring during major spinal surgery. Thirty patients with normal preoperative CSEP were allocated randomly to one of the following anaesthesia regimens: propofol (2.5 mg kg-1 followed by 10-6 mg kg-1 h-1) with or without nitrous oxide, or midazolam (0.3 mg kg-1 followed by 0.15 mg kg-1 h-1) combined with sufentanil 0.5 microgram kg-1 h-1 in the propofol and midazolam groups, or 0.25 microgram kg-1 h-1 in the propofol-nitrous oxide group. CSEP were elicited by alternate right and left tibial posterior nerve stimulation and recorded before and after induction (15 min, 1, 2 and 3 h), and during skin closure. CSEP latencies were not significantly modified in the three groups. CSEP amplitude decreased significantly in the propofol-nitrous oxide group (from mean 2.0 (SEM 0.3) to 0.6 (0.1) microV; P < 0.05) but not in the propofol (from 1.8 (0.6) to 2.2 (0.3) microV) or midazolam (1.7 (0.5) to 1.6 (0.5) microV) groups. The time to the first postoperative voluntary motor response (recovery) delay was significantly greater in the midazolam group (115 (19) min) compared with the propofol and propofol-nitrous oxide groups (43 (8) and 41 (3) min, respectively). Consequently, the use of propofol without nitrous oxide can be recommended during spinal surgery when CSEP monitoring is required.