PURPOSE: To evaluate the prognostic value of multimodal evoked potentials in predicting both awakening and failure to awake from coma due to cardiac arrest. METHODS: Multimodal evoked potentials composed of somatosensory evoked potentials (SEPs), visual evoked potentials (VEPs) and brainstem auditory evoked potentials (BAEPs) were recorded for 46 patients with anoxicischemic encephalopathy who had coma duration>24 h. Patients with trauma, stroke, malignancy, or age<14 y were excluded. Outcomes were categorized as awakening (GOS 3-5) vs. vegetative state or death (GOS 1-2). RESULTS: Forty-one percent of patients regained consciousness, and 59% died or remained vegetative. Bilaterally absent SEP N20 peaks predicted non-awakening with a sensitivity of 52% and a specificity of 100%. Unilaterally or bilaterally absent BAEP III-V predicted non-awakening with a sensitivity of 12% and a specificity of 100%. Using bilaterally absent SEP N20 peaks, unilaterally or bilaterally absent BAEP III-V, or both of the above predicted non-awakening with a sensitivity of 56% and a specificity of 100%. Bilaterally present SEP N20 peaks predicted awakening with a sensitivity of 100% and a specificity of 63%. However, the combination of bilateral presence of SEP N20 and VEP P100 in predicting awakening increased the specificity to 74%, reaching a sensitivity of 89%, a PPV of 71% and a NPV of 91%. CONCLUSION: Non-awakening in postanoxic coma can be reliably predicted with SEPs and BAEPs. Bilaterally preserved SEPs and VEPs predicted awakening with an accuracy of 80% at 1~4 days after cardiopulmonary resuscitation.
Coma ; Consciousness ; Evoked Potentials ; Evoked Potentials, Auditory, Brain Stem ; Evoked Potentials, Somatosensory ; Evoked Potentials, Visual ; Heart Arrest ; Humans ; Persistent Vegetative State ; Prognosis ; Sensitivity and Specificity ; Stroke ; Survivors