SHAKEN BABY SYNDROME
Summary
Medical and legal controversy has surrounded the diagnosis and, indeed, existence of shaken baby syndrome. Does shaken baby syndrome exist and how certain can we be in assessing whether a head injury in an infant is accidental or non-accidental? Dr Robert Minns reviews the evidence base underpinning this controversial and emotive diagnosis.
Key Points
- The first reported association of multiple long bone fractures in infants with chronic subdural haematomas without a history of injury was made in 1946. Following animal studies that demonstrated that repeated acceleration/deceleration without direct impact could cause haemorrhage, the hypothesis was made in 1971 that these injuries might be due to shaking.
- The association of subdural haematomas, retinal haemorrhages and small avulsion fractures of metaphysial bones and their attributions to shaking in 1972 led to the definition of shaken baby syndrome (SBS) in its current form.
- Clinical observations and testimony that non-accidental head injuries may be due to shaking underpins the literature on SBS but further evidence comes from animal studies, computer and biomechanical modelling and ultrasound, CT and MRI neuroimaging.
- These are almost always unwitnessed injuries which may be incurred by a variety of mechanisms, singly or in combination, and the clinician may diagnose a generic ‘non-accidental head injury’ with various degrees of certainty, but may not attribute a specific mechanism of injuring such as ‘shaking’ with certainty, as there is no single pathognomonic sign. The diagnosis of non-accidental head injury is made on a combination of the history of the circumstances of the injury, the pattern of the injuries, and the social context in which they occur.
- Non-accidental head injury (including shaking injury) accounts for most of the handicaps caused by child abuse. A half suffer long-term neurological injury, a fifth die of their injuries, and only 10% have no long-term sequela.
Declaration of interests: No conflict of interests declared
Recent medical and legal controversy over the Shaken Baby Syndrome (SBS) has led to questions regarding its diagnosis and even its existence. Decisions in this area have serious consequences for parents or carers and children, and circumstances in which SBS is suspected are associated with strong emotions. Recent decisions in UK courts have questioned the ability of ‘experts’ to diagnose SBS, and there have been reports that paediatricians in the UK now feel intimidated in respect of diagnosing child-abuse. This controversy has important implications for parent/carer protection and for child protection, and accordingly JRCPE commissioned Dr Robert Minns, an expert in the field, to review SBS for Behind the Medical Headlines. Dr Minns has produced an important and extensive review which we believe will be invaluable to paediatricians. However, because it is important that all physicians, not to say the public, have an awareness of SBS, we have provided this summary, approved by Dr Minns. The full article complete with references will be published in the Journal of the Royal College of Physicians in February 2005.
HISTORICAL BACKGROUND
The story begins when Caffey (1946) reported multiple long bone fractures associated with chronic subdural haematomas in six infants in the absence of any history of injury. He recommended that unexplained long bone fractures warranted investigation for subdural haematoma. In the late 1960s, Ommaya and colleagues reported experimental work in animals showing that high-speed rotation of the head on the neck could produce concussion and bleeding over the brain, and they also described subdural haematomas in two children caused by whiplash injury alone. It was thought that injuries in these situations were caused by an easily deformable brain moving inside a rigid skull resulting in tearing at the attachments, especially of the veins, between the two. Guthkelch (1971) then reported two infants with subdural haematoma without external signs of any injury and, based on the work of Ommaya and colleagues, suggested that these injuries were due to acceleration/deceleration during shaking. Thus, it was suggested that subdural haematomas could occur due to shaking with no fractures of limbs or spine.
Caffey (1972) then published a seminal article reporting 27 infants in whom shaking was considered to have produced subdural haematomas, bilateral retinal haemorrhage, and metaphyseal avulsions in which small fragments of bone had been torn off where the periosteum covering the bone and the cortical bone itself are most tightly bound together. These injuries were all attributed to acceleration-deceleration traction or whiplash stresses. Thus, the basic definition of SBS (infantile subdural and subarachnoid haemorrhage, traction metaphyseal fractures, and retinal haemorrhages) was established, and possibly long-term consequences such as cerebral palsy, developmental delay and epilepsy were considered. Duhaime and her colleagues (1987) later reported on 48 children with SBS, and they carried out doll experiments suggesting that the production of SBS required impact as well as acceleration-deceleration injury (i.e. Shaken-Impact Baby Syndrome).
Numerous reports of SBS are now available and confusion and controversy has arisen. The central controversies are whether shaking can injure the infant brain, what degree of certainty can be put on a diagnosis of SBS, and whether accidental injury or non-traumatic conditions can simulate SBS.
CAN SHAKING INJURE THE INFANT BRAIN?
The literature on SBS consists largely of clinical observations linked to testimony that the injuries could only have been related to shaking. Overall, the literature points to shaking alone being the cause of the injuries in about one-third of cases. A Scottish database (2004) of 124 cases is consistent with these findings, and single carefully documented cases show that shaking alone can lead to head injury. It may be noted in passing that Shaken Adult Syndromes have occasionally been reported in relation to prisoner abuse and to domestic violence. Shaken baby syndrome, however, is much more common than its adult equivalent because the head in infants is much bigger in relation to the body and the neck muscles in infants are too weak to prevent whiplash during shaking. There are, however, several other lines of evidence related to the matter of shaking and brain injury.
ANIMAL MODELS
Studies of whiplash injury in primates have shown that rotational forces can produce cerebral concussion and gross haemorrhage of the brain, and that haematomas were usually related to ruptured veins. The acceleration magnitudes and frequencies in these experiments were probably greater than could be produced by shaking an infant manually, but there is evidence that the brain in infants may be more susceptible to injury than in adults.
COMPUTER MODELLING
The development of computer models of the human head, including the skull, the brain and its connective tissue coverings and anchor points, the venous structures, and the cerebrospinal fluid (CSF) have allowed three-dimensional studies on the mechanisms of injury. These studies have shown that linear trauma, especially to the front of the skull, can cause severe fractures without significant brain damage but that the human brain is very susceptible to severe injury from rotational acceleration forces. Furthermore, the CSF, which protects against linear trauma, facilitates rotational injury by lubricating brain rotation. Finally, computer modelling shows that the brain is very sensitive to the frequency of shaking. A small increase in shaking frequency can cause a disproportionate increase in brain injury.
BIOMECHANICAL MODELLING
The construction of models (dolls) to study SBS has proved difficult as relevant information has primarily been derived from experiments on adult animals. Furthermore, doll models have varied in their construction making comparison of results difficult. However, it has been shown that linear forces produce focal brain injury while rotational forces produce diffuse injury. In fact, shaking a two-month-old infant with force sufficient to produce injury requires considerable exertion. Adults could only shake such a doll for an average of 20 seconds, and children 3–10 years of age could only do so for 20 seconds and at lower force levels. Causing SBS is an exhausting business.
NEURO-IMAGING, ULTRASOUND, CT and MRI IMAGING IN NON-ACCIDENTAL INJURY
These studies have shown that young children without evidence of impact injury all show subdural haemorrhage, and in addition tearing of veins on the surface of the brain and injury to the brain itself. These injuries are those to be expected from experimental studies.
Overall, case series including rare cases in adults, animal experiments, computer and bio-mechanical model experiments, and the results of neuroimaging in non-accidental injury all support the view that adults do shake infants and produce extensive brain injury.
CAN SBS BE DIAGNOSED?
Shaking can cause a variety of brain injuries but because these injuries can be produced by other mechanisms, the injuries cannot be attributed to shaking unless the perpetrator admits that shaking occurred. In reality, most of these injuries occur unwitnessed. Accordingly, injuries should be diagnosed as non-accidental head injury (NAHI). This diagnosis requires an appropriate history, a clinical-radiological-ophthalmological syndrome, and a supportive social background.
HISTORY
The first important step is a full history of trauma. Vague and inconsistent explanations of the circumstances of injury, a history inadequate to explain the severity of the injuries, or no history at all are important. Delay in seeking medical help is also suspicious.
SYNDROME
The rapidity of development and severity of the syndrome varies from acute to chronic, but the common features include encephalopathy, epileptic fits, subdural and retinal haemorrhage, bruising and rib and metaphyseal fractures. Malicious injuries such as cuts, scalds, and cigarette burns, and injuries of different ages are also important. It is important that there is probably no single feature that is diagnostic of NAHI, so even a very complete syndrome should still be considered as suspected NAHI.
SOCIAL CONTEXT
Non-accidental head injury occurs in all walks of life and knows no ethnic, social, or occupational boundaries. However, common social risk factors include a young or unmarried parent, cohabitation, previous child or domestic abuse, alcohol or drug abuse, mental ill-health, a premature infant, failure to thrive, multiple previous medical consultations or hospital admissions.
Overall, NAHI is based on multiple findings and sometimes has to remain just a suspicion as when children present with subdural or retinal haemorrhages only and no or a minor history of trauma. Final diagnosis in this difficult area should be in the hands of doctors with extensive experience of NAHI.
CAN OTHER CONDITIONS SIMULATE NAHI?
Many conditions can predispose to individual components of NAHI and need to be considered.
ISOLATED SUBDURAL HAEMATOMA
Infections, inborn/metabolic disorders, thrombophilia, and low intracranial pressure syndromes can all be causes and should be considered but are usually readily identified. More difficult is whether a ‘short fall’ (chair, bed, baby bouncer, adults’ arms), particularly onto a hard floor, can produce this injury. In general, ‘short falls’ can cause serious local bruising, skull fracture and brain damage but not concussion or encephalopathy. Subdural haematoma requires rotational injury and this needs to be explained by the circumstances of an accidental injury.
RETINAL HAEMORRHAGE
These may occur after normal birth (but resolve within three to four weeks), in meningitis, raised intracranial pressure, coagulation disorders, blood pressure disorders or drowning but haemorrhages in these disorders are few and restricted. Accidental head injury, except for high velocity side-impact road traffic accidents, rarely cause retinal haemorrhages while in NAHI the haemorrhages tend to be severe and extensive due to the sheering forces in the form of injury.
SKELETAL LESIONS
Many conditions affecting the skeleton can simulate the associated injuries of NAHI. Most of these conditions would not easily be confused with NAHI but suggestive radiological appearances need careful consideration.
COMBINATIONS
Subdural haematomas, retinal haemorrhage, and fractures together point reliably to non-accidental injury. Only birth and accidental injury need to be considered otherwise.
OUTCOME
Non-accidental head injury represents less than 0.5% of all child abuse but accounts for most of the child handicaps covered by child abuse. Overall, one-fifth die of their injuries and a half are left with neurological injury while less than 10% escape unscathed. Apart from the pain and suffering, these children have learning difficulties, motor disability, blindness, epilepsy and behaviour problems to say nothing of long-term emotional damage. Further research is needed, but regrettably we know that adults do seriously injure infants’ brains. To be fair to the child as well as to the carers, suspected NAHI should be referred quickly to those with the relevant knowledge and experience.