Article Text


Retinal and subdural haemorrhages: Aoki revisited
  1. H B Gardner
  1. Manitou Springs, CO, USA; horacebgardner{at}

    Statistics from

    In the United States and much of the English speaking world, the combination of retinal and subdural haemorrhages in a very young child is highly suggestive of non-accidental injury and felt to be incompatible with a history of a minor fall. Thus, when Aoki et al1 published 26 cases of infants, age 3–14 months, with this combination following minimal falls, even though they stated in their article “in this series, none of the patients had a significant medical history or were subjects of child abuse before the trauma,” there was immediate concern2 that these may have been undiagnosed cases of shaken baby syndrome (SBS). Aoki responded to this by saying “we do not believe the injuries were caused by battering or shaking” although noting their “series might include some injuries caused by abuse.” In 1987, Ikeda et al3 added cases to the Japanese experience noting these were “usually caused by a simple domestic head injury” and indicating they were aware of the battered child syndrome.

    A comparison of the age distributions of the Aoki and Ikeda series shows them to be very similar (a simple t test not being able to rule they are different with even a 20% probability). However, if either Japanese series or a combination of the two are compared to Kivlin’s4 SBS series in the United States, it is apparent these are not the same (the probability by a simple t test shows there to be between less than one chance in 400 and a chance in a million that these age distributions are the same) (Tables 1 and 2).

    Recent articles5,6 have again brought into question the assumption that a history of a minimal fall is incompatible with the finding of subdural haematoma and retinal haemorrhage in an infant. However, even if some of the Japanese cases are from abuse, some of the American cases may be from minimal falls. The United States assumption would require all the Japanese cases to be misdiagnoses and all 26–40 independent stories in Japan (which are similar to many of the United States stories) to be similar prevarications. Instead of discarding the Japanese experience, an inquiry as to why the age distributions are so different seems appropriate.

    Table 1

    Age distributions

    Table 2



    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.