Intraoperative cerebral perfusion monitoring by evoked potential and transcranial doppler ultrasonography in carotid endarterectomy: a comparative study

Abstract

Objective To investigate the effectiveness of evoked potential in monitoring cerebral perfusion during carotid endarterectomy by a retrospective cohort study. Methods A retrospective cohort survey of clinical data of 59 patients, who underwent carotid endarterectomy in our hospital from September 2013 to December 2016, was performed. All patients were monitored by transcranial doppler ultrasonography (TCD) and somatosensory evoked potential (SEP), motor evoked potential (MEP) for changes of cerebral blood flow; the monitoring of TCD was defaulted as gold standard, recording monitoring results at different time points. By drawing the four table, the sensitivity, specificity, positive predictive value, negative predictive value of SEP and MEP during cerebral hypoperfusion were calculated, and correlation analyses between SEP and TCD, and between MEP and TCD were performed. Through receiver operating characteristic (ROC) curve, SEP monitoring threshold of cerebral hypoperfusion was predicted. Results The sensitivity, specificity, positive predictive value, and negative predictive value of SEP in predicting intra-operative cerebral hypoperfusion were 60.00%, 91.84%, 60.00%, and 91.84% when SEP amplitude declined beyond 50%; the sensitivity, specificity, positive predictive value and negative predictive value of SEP in predicting intra-operative cerebral hypoperfusion were 70.00%, 93.88%, 70.00% and 93.88% when SEP latency extended beyond 10%. The sensitivity, specificity, positive predictive value, and negative predictive value of MEP in predicting intra-operative cerebral hypoperfusion were 30.00%, 97.96%, 75.00%, and 87.27% when MEP amplitude declined beyond 50%. Decline of SEP and MEP amplitudes and extension of SEP latency have positive correlations with TCD. ROC indicated that the amplitude of somatosensory evoked potentials could predict intraoperative hypoperfusion when it decreased by 51.0% than the baseline monitoring value after carotid artery occlusion, with sensitivity of 100.0%, specificity of 89.8% and area under the curve of 0.918 (95%CI: 0.846~0.990, P=0.024). ROC also showed that the latent phase of somatosensory evoked potentials could predict intraoperative hypoperfusion when it was prolonged >9.8% compared to the baseline monitoring value after carotid artery occlusion, with sensitivity of 100.0%, specificity of 92.9% and area under the curve of 0.941 (95%CI 0.878~0.995, P=0.014). Conclusion Evoked potentials can real-time monitor intra-operative cerebral hypoperfusion continuously in carotid endarterectomy, enjoying high sensitivity and specificity; in addition, decline of SEP amplitude and extension of SEP latency have higher sensitivity than MEP. Key words: Carotid endarterectomy; Intraoperative monitoring; Evoked potential; Transcranial doppler ultrasonography