Shaken babies

Since the "shaken baby syndrome" was first described by Caffey in 1972, as a combination of subdural and subarachnoid haemorrhage with traction metaphyseal fractures and retinal haemorrhages, there has been disagreement about which features comprise the syndrome and how shaking might cause them. Duhaime and colleagues, writing in the New England Journal of Medicine recently went as far as to suggest that because the majority of infants who are injured or killed in this way also have clinical, radiological, or necropsy evidence of blunt impact to the head, the syndrome should be called the "shaken-impact syndrome". The situation is further confused by intracerebral damage being caused by sudden angular deceleration rather than specific contact forces. If the contact surface is soft, there may be no external signs of injury to the head.

With such an uncertain definition, it is not surprising that the syndrome has been misdiagnosed and that some children with subdural haematomas due to other causes have been suspected of having "shaken-baby syndrome" with parents or carers accused of assault. The consequences of such an error are appalling. One UK paediatrician, Dr David Southall, told a television programme, Newsnight, in January this year, "It's quite clear there is a consensus on this . . . with regard to acute subdural haemorrhage in babies, if there are retinal haemorrhages and other injuries, then this is child abuse. This accounts for 95% of cases". On the same programme, Dr Bernard Knight, a senior forensic pathologist denied there was any such consensus and went on to say that "there are medically witnessed cases, not many of them, of kids rolling off a settee or a chair and getting a subdural . . . And saying 95% are due to NAI [non-accidental injury] still leaves you with the 5% that are not. How do we know in any particular case whether it's the 95% or the 5%?" The Newsnight programme was concerned with an infant whose subdural haematoma was mistakenly attributed to non-accidental injury. The mother was astounded: "And then I said, 'Well, you know, I'm going home' and at that point the social worker pointed out to me that if I removed (my child) from the ward, the police would be contacted."

Violent assaults on children were brought to medical attention in the early 1960s. Injuries to the head pose special difficulties, especially in the absence of any external signs of violence. No imaging or clinical investigations can differentiate with certainty between accidental and inflicted injury. It is thus all the more important that the "shaken-baby syndrome" be accorded closer attention and an acceptable definition decided. It might be useful, for example, to exclude from the diagnosis any evidence of blunt impact, and to decide how diagnostic retinal haemorrhages are of being shaken. If, 26 years after Caffey's description, doctors are still undecided about the "shaken-baby syndrome", the difficulties faced by experts in presenting medical evidence in court, and by the judge and jury in making sense of it, are readily imaginable.

Recent successful murder prosecutions of women said to have caused an infant's death by shaking have increased public and professional awareness of the "shaken-baby syndrome". Let us hope that such increased awareness is tempered with caution against over-diagnosis. The lessons of Cleveland, UK, in the late 1980s, must stay well-learned. In that episode, two paediatricians claimed to have uncovered an epidemic of child sexual abuse. Relying on physical signs which proved unreliable, doctors sequestered child after child from their parents. The full enormity of these misjudgements took months to emerge. The deaths of Matthew Eappen, Joseph Makin, and Caroline Jongen should not lead to the premature diagnosis of "shaken-baby syndrome".