Repeat CT selectively in children with traumatic brain injuries

View on the News

Small but enticing study

Intuitively, this makes sense, and it’s something we want to believe.

No patient in the selective CT group developed progression of neurologic symptoms or required delayed intervention. In contrast, 7% of children undergoing scheduled or routine CT demonstrated progression, and a majority had epidural hematoma. Two required operative intervention, both as a result of expanding epidural hematoma. From this, the authors conclude that routine or scheduled CT is not indicated in absence of progression or epidural hematoma.

As I understand it, the purpose for repeat CT is to detect progression and the need for intervention. A consistent observation in a review of the literature demonstrates a wide discrepancy in the reported rates of CT progression with or without neurologic progression. Can we safely observe all lesions other than subdural hematoma and obtain CT only if neurologic symptoms progress? If so, for how long do we observe these patients?

The study has a limited number of patients. An appropriately powered noninferiority study to show a new treatment, selective CT, is not worse than existing CT would require nearly 6,000 patients in each arm.

The study provides a description of the lesions noted on CT in the two groups by type, yet it is difficult to determine if the two groups were truly comparable, since there’s a great deal of variability between lesions. Further complicating the analysis is not only the type of lesion but the location of the lesion. Outcomes between supratentorial and posterior fossa lesions can be significantly different regardless of the volume of blood. The findings also could have been influenced by the volume of blood and the time from initial injury to evaluation and imaging.

Dr. Denis D. Bensard is chief of pediatric surgery and trauma at Denver Health Medical Center. These are excerpts of his remarks as discussant of the study at the meeting. He reported having no financial disclosures.



SAN FRANCISCO – Repeat CT scans in children with mild traumatic brain injury were not justified when done routinely but may be warranted for epidural hematomas, according to a retrospective study of 120 patients.

A comparison of the 106 patients who underwent repeat CT scans and the 14 who did not found that the two groups did not differ significantly based on their Injury Severity Score (ISS), the mechanism of injury, or the type of brain injury. Neurologic symptoms did not worsen in patients who did not get a repeat CT scan, none of whom needed surgery, Dr. Jarett K. Howe and his associates reported.

Repeat CT scans showed injury progression in 7 patients (7%), including 2 with subarachnoid hemorrhage and 5 with epidural hematoma –a third of the 15 epidural hematomas in the study. Two patients with epidural hematoma required craniotomy, one of whom showed worsening of clinical symptoms. All 106 patients who had repeat CT scans were discharged without sequelae, said Dr. Howe of Cardinal Glennon Children’s Medical Center, St. Louis University.

The investigators analyzed records for 435 children admitted with traumatic brain injury (TBI) between July 2004 and July 2012 who had CT evidence of an intracranial hemorrhage and a Glasgow Coma Scale score of 14-15, 120 of whom had complete data and met no exclusion criteria.

Children who got a repeat CT scan were significantly older (8 years of age) than those who didn’t (3 years), Dr. Howe reported at the annual meeting of the American Association for the Surgery of Trauma. "This is particularly interesting since CT scan usage is often justified by the difficulty in examining a young child," he said.

The mean Injury Severity Score was similar between groups: 15 in those scanned and 13 in those not scanned.

Although the type of injury did not influence the likelihood of repeat CT, all 15 patients with epidural hematoma got a repeat CT scan (100%), he noted. Repeat CT scans also were ordered in 94% of patients with subdural hematoma, 71% with contusion, 88% with intraparenchymal hemorrhage, 77% with subarachnoid hemorrhage, and 67% with intraventricular hemorrhage.

The TBIs were caused primarily by falls or motor vehicle accidents but also by assaults, sports, or being hit by a car as a pedestrian.

Every year in the United States approximately 642,000 children are evaluated in emergency departments for TBI, 65,000 are admitted, and 7,400 die of TBI. As many as 70% of these children get imaged, with head CT, the diagnostic modality of choice, Dr. Howe said. A total of 4%-8% of children undergoing a CT will have a skull or intracranial abnormality, and less than 1% will have an injury requiring neurosurgical intervention.

Radiation from the CT scan, however, may cause a fatal cancer later on in 1 of every 1,200 children scanned, an extrapolation of historical data from Hiroshima, Japan suggests, he said.

"A majority of our children with minor physiologic insults can be managed without repeat imaging," he said. The investigators now are organizing a multi-institutional study to validate the findings.

A previous prospective study that evaluated 42,412 children with head injuries found that 90% had isolated head trauma and 97% had a Glasgow Coma Scale score of 15 upon arrival in emergency rooms. Still, 35% underwent head CT scans, which identified TBI in 5%, and 0.4% underwent surgery. No patients died (Lancet 2009;374:1160-70).

Previous studies in adults also suggest that repeat CT is not necessary for mild TBI because patients requiring intervention will show deterioration on physical exam, Dr. Howe added. Repeat CT showed injury progression in 21% of 179 adults in one study, only 7 of whom needed surgical intervention (J. Trauma 2006;60:494-9).

Dr. Howe reported having no financial disclosures.

[email protected]

On Twitter @sherryboschert

Next Article: