• Toshinori Horiuchi, Masahiko Kawaguchi, Satoki Inoue, Hironobu Hayashi, Ryuichi Abe, Nobuoki Tabayashi, Shigeki Taniguchi, and Hitoshi Furuya.
• Department of Anesthesiology, Nara Medical University, 840 Shijo-Cho, Kashihara, Nara 634-8522, Japan. thoriuch@naramed-u.ac.jp
• J Anesth. 2011 Feb 1;25(1):18-28.

PurposeMonitoring motor evoked potentials (MEPs) has been recognized as a highly reliable method to detect intraoperative spinal cord ischemia (SCI) in aortic repair. However, the data regarding the sensitivity and specificity of MEPs for predicting postoperative paraplegia are limited. We retrospectively assessed the value of intraoperative MEP amplitudes for predicting postoperative paraplegia.MethodsThe medical records of 44 patients were reviewed. A train-of-five stimulation was delivered to C3-C4, and MEPs were recorded from the abductor pollicis brevis and the tibialis anterior muscles. The cutoff point for detecting SCI was set at 75% decrease of the baseline MEP. Receiver operating characteristic curves were applied at various cutoff points.ResultsThree patients (6.8%) had postoperative paraplegia. The minimum MEP during surgery had 100% sensitivity and 64.9% specificity in predicting paraplegia, and the MEP at the end of surgery had 66.7% sensitivity and 78.0% specificity in predicting paraplegia: only 1 patient, who had borderline paraplegia (right monoparesis), showed a false-negative result. Receiver operating characteristic curves indicated that adequate cutoff points for the minimum MEP during surgery and for the MEP amplitude at the end of surgery were a 75-90% decrease and a 64-75% decrease of the baseline MEP, respectively.ConclusionMonitoring MEPs had relatively high sensitivity and acceptable specificity, with the cutoff point set at 75% decrease of the baseline MEP, for predicting paraplegia and paraparesis. Because of the small sample in our study, further investigations would be necessary to investigate an adequate cutoff point that could predict postoperative paraplegia.

32 keywords...

hide…