Purpose: This study examines the epidemiologic data of pediatric craniofacial (CF) fractures secondary to violence, comparing these data to CF fractures sustained from all other causes. There is a paucity of data on pediatric CF fractures due to violence, although it is not an uncommon cause in children. Data regarding the demographics, mechanism, and fracture pattern of these patients may aid in the characterization and prevention of injury to those children most at risk.
Methods: An IRB-approved retrospective review was completed for all patients diagnosed with an ICD-9 code indicative of CF fracture after presenting to the Emergency Department of a major urban children’s hospital from 2000 to 2005. Socioeconomic analysis was performed using Geographic Information System mapping and 2000 US Census data by postal code.
Results: During the study period, 1528 patients (0.5% of the 317,453 ED visits) were diagnosed with cranial and/or facial fractures. Isolated cranial fractures were excluded, leaving 793 patients. There were 98 cases of a facial fracture due to a violent cause (V) (97 peer assaults and 1 case of child abuse) and 695 from a non-violent cause (NV). The most common NV causes were fall and motor vehicle crash. V patients were older (16y vs. 10y; p<.001), more likely to be male (84.7% vs. 66.5%; p<.001), more likely to be Non-Caucasian (51% vs. 14.4%; p<.001), and more likely to be living below the poverty level (14.5% vs. 11%; p<.001).  V patients exhibited a greater number of nasal bone (40.8% vs. 27.3%; p=.006) and mandible angle fractures (14.3% vs. 3.5%; p<.001) whereas NV patients had more calvarial (24.5% vs. 5.1%; p<.001) and orbital fractures (45.6% vs. 34.7%; p=.03).  NV patients were significantly more likely to posses concomitant neurological (50.1% vs. 26.5%; p<.001), respiratory (3.7% vs. 0%; p=.008), musculoskeletal (13.2% vs. 2%; p<.001), and abdominal injuries (5.3% vs. 0%; p=.002) compared to V patients. NV patients had a significantly greater hospital admission rate (70% vs. 46%; p<.001), ICU admission rate (21.2% vs. 3.1%; p<.001), and longer length of stay (3.3d vs. 1d; p<.001). There was no difference between V and NV patients regarding treatment (rate of observation vs. open reduction and internal fixation).
Conclusions: V patients are a distinct group exhibiting many significant demographic and socioeconomic differences compared to NV patients. V patients experienced less morbidity as evidenced by the fewer concomitant organ system injuries, the decreased hospital admission rate, ICU rate, and shorter length of stay. This suggests that V mechanisms yield lower force and produce a less diffuse injury than NV mechanisms. V patients had a different fracture pattern, likely also related to the differing mechanism of injury; there was no difference in treatment between V and NV patients.