Background: In the published reports of the developed society, subdural hematoma and/or retinal hemorrhages, in the absence of documented history of major trauma, should be considered diagnostic of child abuse. Many people used the above criteria for diagnosis, but subsequently found that retinal hemorrhages were more common in non-accidental injuries (NAI). To what extent is the proposed pathognomonic association between unexplained subdural hematoma/retinal hemorrhages and child abuse a self-fulfilling prophecy?
Methods: Clinical details of nine children under 2 years with unexplained subdural hematoma admitted to Prince of Wales Hospital between 1995 and 1998 were reviewed.
Results: Four had no other physical signs of injury, five had retinal hemorrhages and one had multiple bruises over the body. Following multidisciplinary case conferences for seven children, a diagnosis of NAI was concluded in four cases, but in no case could the abuser be definitely identified. Clinical outcome was poor with seven children showing either profound disability (n = 5) or evidence of developmental delay (n = 2).
Conclusion: In this series, NAI were not established in three of the seven cases. Did we underdiagnose child abuse in these cases? Despite a magnitude of opinion to the contrary, the issue of whether "trivial" head injury can cause subdural hemorrhages and/or retinal hemorrhages is yet unresolved. Clearly much more information on this very sensitive and serious issue is required and these data should be collected with an open mind.