Session II - Pediatrics/Spine


Friday, October 22, 1999 Session II, Paper #16, 11:02 am

The Epidemiology of Accidental and Non-accidental Femur Fractures in Children

Paul Tornetta, III, MD; Susan Scherl, MD; Lisa Miller, MD; Christopher Sullivan, MD, Boston Medical Center, Boston, MA

Purpose: The epidemiology of diaphyseal femur fractures at two Level 1 pediatric trauma centers, with particular emphasis on the characteristics of accidental versus nonaccidental injury, was studied.

Methods: A retrospective review of 438 charts of patients age 12 or under who sustained nonpathologic diaphyseal femur fractures between the years 1986 and 1996 was made. The morphology of the fracture, mechanism of injury, information regarding whether the case was referred for investigation of possible child abuse and the results of such investigation were recorded.

Results: There were 456 fractures in 438 patients (283 male, 151 female). The average age of the patients was 5.7 years; 287 fractures were middiaphyseal, 60 were distal and 103 were proximal. Mechanisms were pedestrian struck (253), falls (120), MVA (23). There was no history of trauma in 21 cases. Eighty cases (18%) were investigated for possible child abuse. Fifteen investigations were positive (3.4% of the total cases and 18.3% of those investigated). Falls accounted for 26% of the injuries overall, but represented 54% of those investigated. The most common morphologies of the fractures cases that were investigated, and positive investigations are listed in table #1.

Table #1

Fracture Morphology (3 most common types)

Fracture Pattern 

 All Cases

 Investigated Cases

 (+) Investigations

 Transverse

 46%

 29%

 44%

 Spiral

 37%

 37%

 31%

 Oblique

 17%

 16%

 6%

Discussion: Many practitioners believe that spiral fractures are "pathogno-monic" of child abuse. Our data supports the premise that transverse fractures are the most common pattern seen in abuse cases. In fact, transverse fractures were more common in the series overall, in falls (the most commonly investigated mechanism), and in the investigation positive cohort. Most importantly, however, they were underrepresented in the investigated cohort, comprising 46% of the overall fractures but only 29% of those investigated. This suggests that practitioners are more likely to view transverse fractures as accidental and have a lower index of suspicion for abuse in these cases. Such bias may lead to missed cases of abuse in children with transverse fractures.

Conclusion: Transverse, not spiral or oblique fracture is the most common pattern seen in cases of child abuse but is underrepresented in those referred for investigation.