Timing of evoked potentials forecasting the prognosis of severe stroke patients

Shu-ying XIAO, Ying-ying SU, Yan ZHANG, Hong ZHAO, Ying PANG, Yi-fei LIU


Objective To investigate the best assessment time of short-latency somatosensory-evoked potential (SLSEP) and brain stem auditory-evoked potential (BAEP) in predicting the prognosis of patients with acute severe stroke. Methods Fifty-two patients who were diagnosed as supratentorial massive cerebral infarction or large-volume cerebral hemorrhage by brain CT and/or MRI examination with Glasgow Coma Scale (GCS) ≤ 12 were selected as observation subjects. GCS, SLSEP and BAEP were recorded at 1-3 and 4-7 d after onset. Outcomes were examined 6 months later using the modified Rankin Scale (mRS). A mRS score of 0-4 was considered as favorable outcome while a score of 5-6 was considered as unfavorable. The correlation between different predictive indexes (GCS, SLSEP and BAEP) and outcome (mRS) was analyzed. The predictive accuracy was also analyzed. Results At 1-3 d after onset, there was no correlation between all the predictors and outcome (P > 0.05, for all). At 4-7 d after onset, SLSEP and BAEP were significantly correlated with mRS (P < 0.01, for all; C > 0.400). At 4-7 d after onset, the prognostic sensitivity of SLSEP and BAEP Ⅴ wave was 85.71%-97.62% ; prognostic specificity of BAEP was 80.00%-90.00%; positive predictive value of all predictors was 89.13%-96.88%; negative predictive value of SLSEP was 83.33%-85.71% ; total predictive accuracy of SLSEP was 88.46%-90.38%. The predictive accuracy of both SLSEP and BAEP achieved the clinical expectation, and the former is better than the latter. Conclusions SLSEP and BAEP have a high accuracy rate in predicting the unfavorable prognosis of patients with acute severe stroke 4-7 d after onset.


DOI: 10.3969/j.issn.1672-6731.2015.12.004


Stroke; Evoked potentials, somatosensory; Evoked potentials, auditory, brain stem; Prognosis; Sensitivity and specificity

Full Text: PDF

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.