Kuppermann N, Holmes JF, Dayan PS, Hoyle JD Jr, et al; Pediatric Emergency Care Applied Research Network (PECARN). Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009 Oct 3;374(9696):1160-70.
What we already know about the topic
According to the CDC, there is an estimated 600,000 pediatric emergency department visits a year are for head trauma. This rate has been increasing over the last 2 decades with most age groups seeing an increase of 150 to 200%. Suffice it to say, it is common in the daily practice of emergency department to evaluate a head injury in a pediatric patient.
Why this study is important
Prior to this study, some CDC data suggested that 50% of children with head injury seen in the emergency department were receiving a CT head; this rate seemed to increase during the 1990s to early 2000s. Cumulative ionizing radiation is known to be a risk factor for the development of malignancy. This study sought to create and validate a decision rule to identify patients at low risk for traumatic brain lesions that require acute treatment and thus decrease the number of unnecessary CT scans.
Brief overview of the study
This was a prospective cohort study that enrolled patients <18 years of age with head trauma at one of 25 emergency departments within the first 24 hours of injury. Patients were excluded if trauma was penetrating, they had a known brain tumor, prior neuro disorder complicating assessment, bleeding disorder, vp shunts, or a GCS of <14.
Clinically important TBI was defined as: death, neurosurgical intervention, intubation for >24 hours (<24 hours were excluded), hospitalization for 2 nights. Patients that were discharged were followed up by telephone survey at 7 days and 90 days to evaluate for missed ciTBI.
In the derivation study, a total of 33785 patients were enrolled; 8502 age <2 years of age and 25283 > 2 years of age. Total number of ciTBI by age: <2 years- 73 (0.86%) and >2 years 215 (0.85%).
For children <2 years of age, the authors looked at 6 factors that were predictive of ciTBI in these patients: (1)altered mental status, (2)non-frontal scalp hematoma, (3)loss of consciousness, (4)severe injury mechanism (MVC with passenger ejection or death or rollover, MVC vs. pedestrian/bike without helmet, fall >3 feet, (5)palpable skull fracture, or (6)not acting normal per parent. The derivation study enrolled 8502 patients, of whom 4527 had none of these 6 factors. Only 1 patient (0.02%) with no risk factors had ciTBI giving this rule a sensitivity of 98.6% (CI 92.6-99.97). The validation study enrolled 2191 children, 1175 of whom had none of these predictors, and none of them went on to have a ciTBI for a sensitivity of 100%(CI 86-100).
For children >2 years of age, the authors looked at 6 factors that were predictive of ciTBI in these patients: (1)altered mental status, (2)loss of consciousness, (3)history of vomiting, (4)severe injury mechanism (MVC with passenger ejection or death or rollover, MVC vs. pedestrian/bike without helmet, fall >5 feet, (5)clinical signs of basilar skull fracture, or (6)severe headache. Of the 25283 patients, 14440 had none of these predictors. Of these patients, only 7 patients (0.05%) had a ciTBI giving this rule a sensitivity of 96.7% (CI 93.4-98.7). In the validation study, 6411 patients were enrolled, 3698 of whom had none of these predictors and only 2 of these patients (0.05%) had a ciTBI for a sensitivity of 96.8% (CI 89-99.6) with a negative predictive value of 99% (CI 89-99.6).
Overall this was a high quality multicenter study with a very large number of enrolled pediatric patients. A limitation of this study was that 21% of the patients that were discharged were lost to follow up. In a study that already had a low incidence in the number of ciTBIs, especially in those with no risk factors, this could have impacted the sensitivity of the decision rules. As the authors note, the intention of this study was to identify TBI that required acute treatment and therefore it is unclear how this decision rule may impact long term neurologic outcomes in these patients. This rule should also be used cautiously in patients 3 months old or younger as (1) they have thinner skulls and thus their brains are more susceptible to damage and (2) it may be difficult to assess a 3 mo for AMS or changes in mental status after an observation period. Thus consider a CT head in these patients early in their course.
Take home points
PECARN, when used in the appropriate patient population, is a powerful tool to quickly risk stratify pediatric patients into low risk categories that have a high NPV for clinically important traumatic brain injuries. If a child is 3 months or less, you should have a low threshold to order a CT.