Medical and cognitive outcome in children with traumatic brain injury.

BACKGROUND: Head injury is an important cause of morbidity and mortality in pediatrics. Comprehensive studies on outcome are scarce despite significant clinical concern that multiple areas of functioning may be impaired following moderate to severe head injury. The literature suggests that sequelae include not only medical problems but also impairments in cognitive functioning.
METHODS: A retrospective medical and psychology chart review of patients, age 1-18 years, admitted to the Children's Hospital of Eastern Ontario with moderate (Glasgow Coma Scale [GCS] 9-12) or severe head injury (GCS < or = 8) from November 1, 1993 until December 31, 1998 was conducted. Correlations were performed between medical variables (i.e., GCS, Pediatric Risk of Mortality [PRISM] III score, duration of ICU and hospital stay) and measures of intelligence and memory functioning.
RESULTS: Eighty-three children age 1 to 18 were included. Seventy percent of the children were classified as having a severe head injury. There was a mortality rate of thirteen percent. Younger age at injury, lower GCS, and higher PRISM III scores predicted higher mortality. Medical complications were documented systematically. Forty-four patients underwent at least one cognitive assessment and 17 of these children had intelligence testing at three points in time: baseline (< four months), early recovery (five to 15 months) and follow-up (16 to 38 months). The mean intelligence and memory scores fell within the average range at the latest point in follow-up. For those children who underwent three serial assessments, the mean verbal and performance IQ fell within the low average range at baseline improving significantly to fall within the average range by early recovery. Continued improvements were apparent in verbal memory beyond early recovery, with the mean obtained at follow-up falling within 1 SD of the normative mean. Despite the return to normal ranges for the group means the proportion of scores falling below 1.5 standard deviations from the mean was greater than population norms for verbal IQ, performance IQ and verbal memory. Lower GCS scores and longer duration of stay in ICU or hospital were predictive of lower nonverbal intelligence. Lower GCS was also predictive of lower visual memory scores.
CONCLUSIONS: This study describes a population of Canadian children who suffered moderate or severe traumatic brain injury. Initial GCS was the best predictor of mortality and cognitive outcome. These children demonstrated a temporal improvement in intelligence and memory functioning, with their mean performance on these cognitive measures falling within the average range at 16 to 38 months postinjury, although there was considerable variability in the outcomes between individuals.