Is brain computed tomography combined with somatosensory evoked potentials useful in the prediction of brain death after cardiac arrest?

BACKGROUND: Brain death (BD) in coma after cardiac arrest (CA) is difficult to predict. Basal ganglia gray matter/white matter (GM/WM) ratio density and somatosensory evoked potentials (SEPs) may differentiate patients evolving toward BD.

METHODS: We used SEPs and brain computed tomography (CT) after coma onset, within the first 24hours.

RESULTS: Of the 160 patients included in the study, 22 (14%) evolved toward BD. SEP patterns predicted BD (ROC area=0.82, P<0.0001). The combination of SEP patterns, bilaterally absent (AA) and absent on one hemisphere and pathological on the other (AP), predicted BD with a sensitivity of 100% and a specificity of 62.3%, with a positive likelihood ratio of 2.65. The GM/WM ratio predicted BD (ROC area=0.68, P=0.01). A GM/WM ratio <1.07 had a sensitivity of 30.4%, a specificity of 94.9%, and a positive likelihood ratio of 6.27. The combination of SEP and CT findings did not increase the prediction of BD.

CONCLUSION: SEPs and brain CT within 24hours predicted BD after CA. Severe SEP findings (SEP patterns: AA, AP) identified a subset of patients in whom BD could occur. Brain CT (GM/WM ratio in basal ganglia) predicted an early evolution toward BD with high specificity but lower sensitivity.