The Swedish shaken baby syndrome report and review: what does the latest knowledge tell us?

The topic has never ceased generating heated controversy and acrimonious debates. At the core of the dispute is not whether shaking can cause devastating lethal injury, as it can certainly do 2, 3. The main controversy resides in the value of the ‘triad’ as categoric evidence of abusive head trauma (AHT) by shaking when no other credible explanations are found. The mechanism of SBS/AHT as initially proposed involves rotational and acceleration/deceleration movement of the brain within the cranium, producing shearing of the nerve fibres of the brain and tearing of the cortical bridging veins into the potential subdural ‘space’ 1. Geddes et al. 4 showed that the brain pathology in infants thought to have been shaken was due to hypoxia and ischaemia rather than traumatic tearing of axons and that damage to the craniocervical junction could account for apnoea, leading to hypoxia and brain swelling. The debate further increased when Geddes et al. 5 suggested that the cause of the subdural bleeding in some cases of infant head injury is not the traumatic rupture of bridging veins, but a phenomenon of immaturity involving intradural bleeding. What became known as the ‘Geddes hypothesis’ is that thin film SDH and RH could be caused by severe hypoxia in association with brain swelling and raised central venous pressure. Significantly, in cases with no evidence of trauma, the possibility of a natural disease process precipitating the same sequence of events was implied. The Geddes ‘hypothesis’ 5 was heavily criticised by traditionalists who argued that traumatic disruption of bridging cerebral veins and vitreoretinal shearing injury explain RH and SDH. To add to the debate, in 2005, the Court of Appeal found the Geddes Hypothesis ‘not credible’, which greatly influenced the medicolegal community, even though the Courts do not have medical expertise and the courtroom is not a suitable forum in which to debate contentious medical hypotheses. Interestingly, recent studies have shown that (i) there is a rich intradural vascular plexus, (ii) SDH may result from tears in the thin-walled segments of the large bridging veins as they pass through the dural border layer as well as from rupture of the narrow vessels of the dural venous plexuses, leading to thin film bleeding in the subdural ‘compartment’ and that (iii) thin film SDH can also be seen in young infants with hypoxic–ischaemic encephalopathy not related to trauma 6. Therefore, the thin film of SDH in the triad appears to be unreliable as evidence of AHT/SBS. It is fascinating that despite a shift in the paradigm, the debate concerning the power of the ‘triad’ as evidence of AHT remains unchanged. The Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) conducted a systematic literature review aiming to determine the diagnostic accuracy of the triad in detecting whether an infant had been shaken 7. Lynøe et al. 7 identified 3773 abstracts, 1064 were considered relevant and read as full texts, and 30 studies were ultimately included. Of these, 28 were assessed as having a high risk of bias related to methodological shortcomings and circular reasoning when classifying shaken baby cases and controls. Only two studies had a moderate risk of bias; both used confessions and convictions when classifying shaken baby cases, but their different designs made a meta-analysis impossible. None of the studies had a low risk of bias, and circular reasoning was a common feature. The conclusions of the review study 7 were that there was insufficient scientific evidence to diagnose isolated shaking using the triad as the diagnostic criterion. In a more recent publication, Lynøe et al. 8 reiterated that the two main conclusions of their previous study 7 were as follows: (i) the insufficient scientific evidence on which to assess the diagnostic accuracy of the triad in identifying traumatic shaking and (ii) the limited scientific evidence that the triad and can be associated with traumatic shaking. Lynøe's et al. 7 results fuelled the discussion and widened the differences, involving influential medical societies and leading to the publication of numerous responses both in support and in disagreement 9, 10. Notably, one of the studies that followed 10 reported a low incidence of AHT in Sweden. In a national study covering a 20-year period, the authors were unable to identify any infant with SDH without perinatal risk factors, significant disease or findings consistent with nonimpact head trauma. Moreover, the authors of this study 10 raised the question that previously reported higher incidence of fatal AHT in other Western countries has been exaggerated. Before the Swedish Report was published 7, SBU received pressure from relevant organisations, including the American Academy of Pediatrics, the European Society of Radiology, the Society for Pediatric Radiology, the Royal College of Paediatrics and Child Health, the American Society of Pediatric Neurosurgeons, the Norwegian Pediatric Society and the Swedish Pediatric Society, all requesting to review the draft report and raising ethical considerations. SBU did not surrender to the pressure, publishing the report in 2016 in Swedish and in 2017 in English 7. Not surprisingly, most critiques of Lynøe et al. 9 study were authored by clinicians and radiologists involved in child protection teams. One of these studies 9 stated that ‘radiologists do not conclude abuse merely because of the presence of the triad. Other conditions are always excluded either before or following referral to the child protection team’. This statement merits some consideration: (i) the radiologists report on shadows, not reaching the gold standards of autopsy combined with microscopic examination; (ii) radiologists are happy to diagnose abuse when other conditions are excluded, but do not take into consideration the unknowns: sudden infant death is the leading cause of death in the developed world, affecting infants of the same age and with the same gender ratio (M:F 2:1) as SBS/AHT and whose cause remains unknown and (iii) the presence of fractures (i.e. ‘other conditions’) is irrelevant, as the SBU review focussed specifically on the validity of the ‘triad’ as evidence of SBS/AHT. The SBU report is a further addition to previous large-scale judicial reviews undertaken in the UK (Goldsmith review, 2006) and in Canada (Goudge Inquiry, 2008), which concluded that the presence of the triad alone should not automatically lead to a diagnosis of SBS 3. Since the 1971 proposal of shaking as a mechanism leading to the ‘triad’ by Norman Guthkelch 1, some elements have subtly disappeared from the statements of those who propose that the triad is evidence of AHT: (i) that the SDH is caused by the rupture of bridging veins (the existence of an intradural venous plexus and the presence of haemorrhage in the dura are currently accepted) and that (ii) the SDH develops in the potential subdural space (it is accepted that this is not a potential space, but a compartment within the dura). More importantly, despite the Court of Appeal ruling on the Geddes hypothesis in 2005, the current view is that damage of the cervical spine leading to anoxic encephalopathy appears to play a role in AHT 2, 3. Most of the criticisms of the SBU review focus on the methodology rather than challenging the conclusion with solid scientific evidence capable of demonstrating that the presence of the ‘triad’ is indeed evidence of SBS/AHT. This seems a shortcoming, and a reflection of circular reasoning and groupthink. Medicine is not an exact science, and today's truth might not be so tomorrow. The SBU report, alongside the conclusions from previous judicial reviews (Goldsmith Review in the UK -2006- and Goudge Inquiry in Canada -2008-), reaffirms the need for future research to be undertaken in this field. The scientific community needs to understand the mechanism of injury during AHT and to identify the physiopathology of anoxic encephalopathy and SDH in the absence of trauma. No conflict of interest declared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.