Editor—The “serious data gaps, flaws of logic, and inconsistency of case definition” shown up by the evidence based case report of the shaken baby syndrome (p 754) and highlighted in the accompanying editorials (pp 719 and 720) will be of interest to the many parents who over the past 10 years have maintained that they have been wrongly accused and convicted of causing their children’s injuries.1-3
Furthermore, the recent evidence emphasized by Geddes and Plunkett that trivial falls and other minor injuries can give rise to the allegedly characteristic signs of subdural and retinal hemorrhages is consistent with a triad of possible alternative explanations for shaken baby syndrome. This triad has emerged from an analysis of 98 parental accounts reported to the support group the Five Percenters, each of the three being compatible with a distinct type of neuropathology.
The first is minor trauma (37% of cases). This group gives a history of minor trauma (such as a fall from a bed or sofa) with either immediate loss of consciousness or delayed presentation of an acute subdural bleed and retinal hemorrhages. This is in line with the recently reported series from the United States of independently witnessed minor falls resulting in an acute intracranial bleed, the retinal hemorrhages being caused by a sudden rise in retinal venous pressure as in Terson’s syndrome.4
The second is birth injury (29% of cases). The clinical presentation in the second group is quite different. There is a general period of variable length of non-specific symptoms such as vomiting and lethargy warranting repeated medical consultations until computed tomography shows the presence of a chronic subdural haemorrhage. The most likely aetiology is a subdural bleed at birth, which, though usually associated with prematurity or a difficult labour, can follow a normal delivery.5
The third is respiratory arrest (22% of cases). In this group the precipitating event is suggestive of respiratory arrest—often followed by attempts at resuscitation—that could result in the subdural and retinal haemorrhages characteristic of hypoxic encephalopathy. The findings that severe traumatic brain damage is not, as previously thought, present in these cases contradicts the assumption that such injuries could only have been induced by violent shaking.6
A fourth type of presentation, epileptiform seizures (12%) is presumably secondary to underlying intracranial disease—and is thus uninformative about possible aetiology.
These three patterns of clinical events—in the absence of other circumstantial evidence for non-accidental injury—offer a more credible explanation than shaken baby syndrome for the presence of subdural and retinal hemorrhages. It should be noted that shaking has never been directly observed or proved to cause such injuries but is rather an inference based on (contested) theories of biomechanics.7
By contrast, consistent parental testimony tallies with descriptions from independent witnesses. Furthermore, each pattern of clinical events is consistent with a distinctive type of neuropathology of acute subdural, chronic subdural, or the thin subdurals of hypoxic encephalopathy.
While we recognise the limitations of the volunteered parental testimony on which this analysis is based, the same triad of presentations—designated as acute encephalopathic, idiopathic subdural, and hyperacute presentation—has also been independently identified from an extended database of cases of suspected non-accidental injury (see previous letter).8
These findings necessarily raise disturbing questions about the validity of the opinions expressed by medical experts in the courts. They warrant further, urgent, and appropriate scientific investigation.