The association between admission coma score and eventual outcome was assessed using a coma scale developed for children with a variety of central nervous system injuries. As opposed to the Glasgow coma scale, this scale does not demand assessment of verbalization, and thus can be applied to the preverbal or previously intubated child. Cortical function is graded from 6 (purposeful, spontaneous movements) to 0 (flaccid), and brainstem function is graded from 3 (intact) to 0 (absent and apneic). Maximum total score is 9. In 91 children treated for intracranial hypertension, the association was moderately good. The scale was better in predicting the outcome of patients with hypoxic encephalopathy and head trauma than that of patients with Reye's syndrome, meningitis, or encephalitis. No child with a score of less than 3 survived in spite of intensive therapy. Most of these children were flaccid with depressed or absent brainstem reflexes. No child with flaccidity on admission survived.