It is not clear whether clinical symptoms of head injury in infants are reliable and whether radiographic imaging is a necessary part of the diagnostic evaluation. Greenes and Schutzman prospectively studied the historical features and physical findings of head-injured infants to determine how well clinical features predict intracranial injury. The primary goals were to determine the role of radiographic imaging in head-injured infants. The study premises were the following: (1) whether clinical signs are reliable, (2) whether radiographic imaging in an asymptomatic infant with a hematoma is useful and (3) whether asymptomatic infants can be safely managed without radiographic imaging.
Over a one-year period, all children less than two years of age who were seen in an emergency department for head trauma were included in this study. Information was initially collected about the mechanism of injury, symptoms reported by parents and signs of head injury or neurologic abnormalities as noted on physical examination. Patients were classified as “symptomatic” if they had a history of loss of consciousness, lethargy, irritability or seizures; if they had vomited at least twice; if neurologic abnormalities were present on examination; or if they had a bulging fontanel, abnormal vital signs or focal neurologic findings. Mental status was rated on a scale of 1 (normal mental status, “alert and interactive”) to 4 (severely depressed mental status, “arousable to painful stimuli” or “unarousable”).
If present, scalp hematomas were rated as small, moderate or large. Infants less than one year of age were classified as having a significant hematoma if any hematoma was present; moderate or large hematomas were considered significant in children between one and two years of age. Low-risk was defined as an asymptomatic infant without a significant scalp hematoma. A skull radiograph or computed tomographic (CT) scan was then obtained at the discretion of the physician following previously developed guidelines. These guidelines included obtaining a CT for symptomatic infants, a skull radiograph for an asymptomatic infant presenting with a significant scalp hematoma and a CT for any infant whose skull radiograph revealed a skull fracture. The clinical course of all patients was followed, and a follow-up telephone call was made to all parents two weeks following the emergency department visit.
Of the 608 patients enrolled, 177 (29 percent) were symptomatic and 431 (71 percent) were asymptomatic. Of the 431 asymptomatic patients, 166 (39 percent) had significant scalp hematomas and 265 (61 percent) had no significant scalp hematoma. These 265 patients (44 percent of the entire study) were considered low-risk. Thirty (5 percent) of the 608 patients were diagnosed with an intracranial injury, and 63 (10 percent) were diagnosed with an isolated skull fracture with no associated intracranial injury. Age was a significant factor in relation to intracranial injury. Of 92 infants zero to two months of age, 12 (13 percent) had intracranial injury compared with 13 of 224 infants (6 percent) three to 11 months of age and five of 292 infants (2 percent) 12 months of age or older. Four of 17 infants who had no reported history of head trauma were found to have an intracranial injury, and each was thought to be a victim of child abuse. Children who fell from heights of three feet or more or those who fell down stairs were also more likely to have intracranial injury.
Intracranial injury was more likely in children who had the following symptoms or signs (odds ratio [OR] among all 608 study participants): a history of lethargy (OR: 9.19), irritability (OR: 2.41), depressed mental status (OR: 4.90), bulging fontanel (OR: 22.1) and vital signs consistent with increased intracranial pressure (OR: 20.57). No significant association between loss of consciousness, seizures or vomiting and a finding of intracranial injury was evident.
Overall, a significant scalp hematoma was a more sensitive indicator of intracranial injury than signs and symptoms of head or neurologic injury. Almost one half of infants with an intracranial injury were found to be asymptomatic. Almost all infants (96 percent) who had significant scalp hematomas and an intracranial injury also had a skull fracture. Only one infant who was classified as low-risk was found to have an intracranial injury; no treatment was required for this small epidural hematoma. At two-week follow-up, none of the 420 infants who had not undergone head CT was found to have any complication. None of the low-risk patients deteriorated clinically or required any treatment for the injury.
The authors conclude that head-injured infants with symptoms or asymptomatic infants with significant scalp hematomas should undergo radiographic imaging. Also, the index of suspicion should be lower for younger infants, especially those younger than three months of age. Radiographic imaging in an asymptomatic infant older than three months of age without a significant scalp hematoma is generally unnecessary.