Background: Emergency department (ED) management of skull fractures in children remains controversial. Because infants incurring head trauma have a high incidence of skull fracture, we chose to describe fractures in this subset of patients and to determine if there are clinical predictors of associated intracranial injury (ICI) that may have utility in developing more efficient management schemes in these patients.
Methods: A retrospective medical record review was conducted on all awake patients < 13 months of age with an acute skull fracture from non-birth trauma, presenting to the ED of a university-affiliated children's hospital during a three-year period. Clinical and radiographic data extracted were used to describe skull fractures in these patients. The ability of various characteristics to determine the presence of ICI was assessed by calculating sensitivity, specificity, positive predictive value, and negative predictive value for each.
Results: The predominant mechanism of injury for the 102 infants was falls (91%). Suspicion of abuse was found in only one case. The parietal bone was fractured in 87 infants, and 34 had nonparietal fractures. The most prevalent fracture type was linear (92 infants), and 31 had > 1 cranial bone fractured. CT scans obtained on 32 infants (CT group) revealed 21 ICIs in 15 patients. Two with temporoparietal fractures required emergent evacuation of epidural blood. In the CT group, seven of the 15 (47%) with ICI (ICI group) were lethargic compared to two of the 17 (12%) without ICI (No ICI group) (P = 0.035). Five (33%) in the ICI group had temporal bone fractures compared to 0 in the No ICI group (P = 0.015). The presence of any sign or symptom had a sensitivity and negative predictive value of 100%, but only a specificity of 35%. The presence of lethargy had a positive predictive value of 78%. The presence of temporal and frontal bone fractures had positive predictive values of 100 and 75%, respectively.
Conclusion: This study reports a high prevalence of fracture characteristics often associated with inflicted injury in other studies when virtually all injuries in our sample were accidental. Several clinical characteristics were demonstrated to be potentially useful in predicting ICI associated with skull fracture; however, prospective study is recommended to validate these findings prior to clinical application.