Early hyperthermia after traumatic brain injury in children: Risk factors, influence of length of stay, and effect on short-term neurologic status. Natale JE, Joseph JG, Helfaer MA, Shaffner DH. Crit Care Med 2000; 28:2608–2615. Hyperthermia worsens outcome in virtually all animal models of acute brain injury. This is true whether hyperthermia occurs during or after the insult. In adult humans, the natural history of acute brain injury (for example, cardiac arrest, subarachnoid hemorrhage, and traumatic brain injury) includes a high incidence of spontaneous hyperthermia (Albrecht et al. Mayo Clin Proc 1998; 73:629—35). It has been argued that this hyperthermia should be treated to reduce severity of secondary insults after hospitalization. There is no data on the frequency of hyperthermia in children suffering from acute brain injury. Accordingly, the authors of this report retrospectively examined the medical records of 117 children hospitalized for traumatic brain injury (TBI). The cases occurred between 1995 and 1997. The median age of the children was 5.4 years. Patients with TBI were hospitalized in the pediatric intensive care unit (PICU) if their Glasgow Coma Scale (GCS) score was <15 or if there were other co-existing medical indications. Induction of mild hypothermia was not part of the treatment regimen. In addition to examining for evidence of fever, the investigators noted the following data: demographic characteristics, preinjury condition (especially recent preinjury antibiotic use), extracranial injuries, injury severity, initial computerized tomography (CT) scan, early physiologic response (such as glycemic and hemodynamic status), and medical care (for example, requirements for surgery). When possible, established scoring systems were used for each attribute. Outcome was measured by the PICU discharge GCS score and by the length of PICU stay. The data were subjected to bivariate analysis and logistic regression procedures to define risk factors for hyperthermia and to segregate the outcome predictive value of hyperthermia from known determinants other than length of PICU stay. All patients were neurologically intact before injury. Admission median GCS score was 12. Early hyperthermia (defined as a recorded value >38.5oC) was observed in 29.9% of patients. Early risk factors for hyperthermia were limited to injury severity, initial head CT, and early physiologic response to injury (primarily arterial hypotension). Factors associated with a low GCS score at PICU discharge were injury severity, early physiologic response (including hyperglycemia, leukocytosis, hypoxia, hypotension, and hyperthermia), and surgery during PICU stay. Factors associated with prolonged PICU stay were extracranial injuries, injury severity, early physiologic response including hyperthermia, and surgery during PICU stay. Logistic regression analysis showed that early hyperthermia predicted length of PICU stay independent from injury severity, hypotension, leukocytosis, or hyperglycemia. The results of this retrospective analysis in pediatric patients are in agreement with those found in adults. A hyperthermic response to TBI is common. However, the incidence of hyperthermia observed in this series was about one-half that observed in adults. This might be explained by the fact that this data was collected after it had become recognized that hyperthermia might be detrimental and therefore was more aggressively treated. Regardless, the incidence in children remained relatively high. There are no prospective controlled trials that have examined the effects of prophylactic avoidance of hyperthermia on outcome in patients with TBI and therefore any recommendations remain speculative. However there is no evidence for a beneficial effect of hyperthermia on the injured brain. In contrast, there is consistent evidence from laboratory models of acute brain injury that demonstrate adverse outcome effects from either early or delayed hyperthermia. This allows a reasonable conclusion to be made that hyperthermia, now known to frequently accompany TBI, should be aggressively treated. (DSW)