1.3.2.3.1

Risk factors related to the child

The gender ratio: there is male predominance, with a boy/girl ratio of between 1.3 and 2.6.

Prematurity: there is a higher proportion of premature infants among SBS victims (11 to 21%, with 11% in the series reported by Mireau ) than among the general population (7 to 8%).

Multiple pregnancies are more frequent in SBS (in 5% of cases reported by Mireau ) than in the general population (1.5% of all births).

Crying cannot be considered as a risk factor per se but may trigger abuse of the infant , given that parents’ tolerance of crying in a child is very variable. A consultation for crying in a young baby should not only search for the cause but also evaluate the parents’ feelings and reactions.

1.3.2.3.2

Risk factors related to the shaker

In cases where the perpetrator has been identified (regardless of whether he/she has admitted the act), the latter is usually (70%) male (more often the child’s father than the mother’s partner).

Unrelated adults also constitute a significant category of potential shakers: in the series of 151 AHI cases examined by Starling’s group , the mother’s partner was implicated in 20.5% of cases and a female carer/babysitter was implicated in 17.3% of cases.

1.3.2.3.3

Risk factors related to the parents

In terms of the socio-economic context, the results are very contradictory. All backgrounds are concerned by SBS but the suggested vulnerability factors remain to be documented (a first child, a new pregnancy, a return to work, poor knowledge of a child’s needs or normal behaviour, social and family isolation, a history of domestic violence, past or current psychiatric disorders, drug or alcohol abuse, etc.). In 2005, Mireau noted that the parents had very poor knowledge of a child’s needs or normal behaviour. Low parental age is frequently emphasized by researchers and has been reported as a potential risk factor .

A prospective study examined the socio-economic context of 25 cases of non-accidental HI observed over almost 10 years (January 1998 to September 2006) in a region of Scotland: 76% of the cases came from the most deprived areas in terms of education, learning and social capacities, with 72% in areas with the highest crime rates, 68% in areas with poor healthcare facilities, 60% in low-income areas, 52% in areas with poor housing and 48% in areas with high unemployment.

These results are in sharp contrast to those of earlier studies: of the two studies published in 2000 on the same cohort of children, one noted that most parents had a stable job (81% of the mothers were in work) and the other found that the majority of the parents had been in secondary or higher education.

Ruling out a diagnosis of AHI because of an apparently favourable sociofamilial context may lead to so-called “missed” diagnoses .

However, in general, a healthcare professional should pay attention to a difficult sociofamilial context for the child’s carers.

The relationship between SBS and ethnic factors has been studied. Fortin reviewed the literature on this topic and concluded that ethnic origin is no more significantly associated with SBS than it is with accidental HI. In fact, the data suggested rather that the “ethnic origin” parameter constituted a risk factor for other social markers that increased in the risk of AHI in the child.

In summary , the currently available data (although fragmented and sometimes contradictory) suggest that children who are male, first born, aged under 6 months, born prematurely after a complicated or multiple pregnancy, living with parents with a history of psychoactive substance abuse (alcohol, drugs) or family violence and/or with poor knowledge of strategies for managing the relationship with their infant are at a greater risk of being SBS victims. Of course, this violent act can occur in the absence of any socio-economic and cultural risk factors. Even though it is right to consider that SBS victims are more likely to belong to one or more of the above-mentioned risk groups, it is wrong to believe that the majority of children presenting these characteristics are victims of this type of abuse.

1.3.3.1.1

The meninges (with subdural or subarachnoid haemorrhage), the brain, the eyes and the spinal cord are likely to be damaged in SBS

Other lesions may also be observed: fractures of the limbs, ribs or skull; bruising of the scalp, hematoma of the neck muscles and posterior spinal lesions.

Although the most detailed data have been provided by autopsy series, the latter obviously correspond to the most serious forms of SBS because they have led to the child’s death.

On the basis of 93 neuropathological examinations of SBS victims lacking visible cranial trauma, Billette described:

• •
  

  81 cases of subdural hematoma (SDH), 65 cases of subarachnoid haemorrhage, 14 cases of intraparenchymatous haemorrhage, 69 cases of brain oedema and 41 cases of cerebral herniation;

  
  
• •
  

  56 cases of intraocular haemorrhage;

  
  
• •
  

  21 spinal cord lesions.

The state of the spine was not mentioned. The lesions observed here were not specific for the mechanism of death.

Another study found cervical epidural haemorrhages and focal axonal lesions of the brain stem and the roots of the spinal nerves in 11 of 37 AHI cases and in none of 14 control cases (i.e. deaths from other causes).

Different types of damage to the brain parenchyma can be observed:

• •
  

  anoxic lesions of the cortex, grey nuclei and thalamus: these lesions translate into (rarely haemorrhagic) hypodensities associated with a loss of contrast between white matter and grey matter ;

  
  
• •
  

  brain oedema, translating into a decrease in the volume of liquid-filled spaces;

  
  
• •
  

  contusions, in particular in the frontal and temporal regions and at the white matter–grey matter junction.

Intracranial extra-axial, blood collections (SDH, potentially combined with subarachnoid haemorrhage) present some particular features in SBS ( Figs. 1–3 ). They are generally multifocal, bilateral and faint (with no mass effect). They cover the convexity of the brain hemispheres and accumulate in the longitudinal fissure in an inclined position and along the insertion of the tentorium cerebelli. The detection of a haemorrhage in the falx cerebri is very suggestive of SBS , as are subdural collections in the posterior fossa. It is nevertheless noteworthy that SDH is not observed in all shaken babies: it was only seen in 72–93% of the cases in the articles reviewed by Christophe .

Multifocal subdural haematoma of the tentorium cerebelli (indicated by an arrow) in a case of abusive head injury (AHI) caused by shaking.

Multifocal interhemisphere subdural haematoma (hyperdensity indicated by an arrowhead) and multifocal left pericerebral subdural haematoma (hyperdensity indicated by an arrow) in a case of abusive head injury (AHI) due to shaking.

Multifocal subdural haematoma of the vertex (indicated by arrows) in a case of abusive head injury (AHI) caused by shaking.

EDH results more from accidental HI than SBS, where it is extremely rare.

1.3.3.1.2

Eye lesions

RHs are NOT always present in SBS: according to Defoort-Dhellemmes , they are seen in about 80% of cases (ranging from 50 to 100%, depending on the series). They are described in terms of:

• •
  

  whether they are bilateral or not: most are bilateral, which contrasts with the generally unilateral RH that can be observed in accidental HI . Nevertheless, RH are unilateral in 10 to 17% of cases of SBS;

  
  
• •
  

  their appearance, size (small, large, more or less than two optic discs) and shape (flame-shaped, blots, punctiform, dome-shaped), which depend on their location in the eye;

  
  
• •
  

  their location at the posterior pole of the eye (peripapillar, macular, along the vascular arcades) or at the periphery of the fundus (near to the periphery or extending out across the very edge to the ora serrata);

  
  
• •
  

  their location relative to the retinal layers ( Table 2 ):

  
  
  
  
  • ∘
    

    dome-shaped, preretinal haemorrhages situated just under the internal limiting membrane, whether small (classic, pearl-shaped RH) or large (haemorrhagic retinoschisis),

    
    
  • ∘
    

    superficial haemorrhages (that disappear very rapidly, sometimes in less than 24 hours) or deep intraretinal haemorrhages,

    
    
  • ∘
    

    subretinal haemorrhages.

  
  

Defoort-Dhellemmes distinguishes between three types of RH, depending on their number and their extent ( Table 3 ):

• •
  

  type 1: intraretinal haemorrhages, flame-shaped, blots or punctiform, situated at the posterior pole of the eye;

  
  
• •
  

  type 2: preretinal dome-shaped haemorrhages that are small (no larger than the diameter of the optic disc) and pearl-shaped, situated at the posterior pole, around the optic disk and along the vascular arcades or mid-way out towards the periphery. These haemorrhages may occur alone or in combination with type 1 RH;

  
  
• •
  

  type 3: profuse, multiple haemorrhages of all types (intra-, pre- or subretinal), coating the whole retina or flecked out to its periphery, combined with unilateral or bilateral premacular haemorrhagic plaques (which are sometimes immediately suggestive of haemorrhagic retinoschisis). Type 3 haemorrhages are extremely suggestive of SBS. They can be considered as almost pathognomonic, especially when combined with SDH, massive brain oedema or bone lesions that are very suggestive of abuse. However, they can be observed extremely rarely in violent, accidental HI (road accidents).

Other lesions can be seen in the fundus: vitreous matter and choroidal haemorrhage and papillary oedema due to intracranial hypertension.

Intra-orbital haemorrhage (scleral haemorrhage and haemorrhage of the optic nerve sheath, muscles and orbital fat) can be detected on autopsy .

1.3.3.1.3

Lesions of the neck muscles, spine or spinal cord

Neck damage was noted in 4% of the children studied by King et al. . Billette reported that several observations of spinal cord SDH have been described in the literature. Christophe pointed out that violent shaking can provoke: (i) widespread axonal lesions near the brain stem and the upper spinal cord and (ii) epidural haematoma near the neck/head junction.

1.3.3.1.4

Skin lesions

In the absence of medical causes, bruising is very suggestive of abuse in an infant that cannot yet walk unaided . In the latter article, only 0.6% of the children under 6 months and 1.7% of the children under 9 months of age had one or several bruises. It is particularly important to look for bruising on the scalp: in a study on HI in infants, Greenes and Schutzman found that 93% of the infants presenting bruising of the scalp also had intracranial lesions.

1.3.3.1.5

Bone lesions

Bone lesions are particularly suggestive of abuse:

• •
  

  rib fractures (in the absence of prior, aggressive, respiratory physiotherapy) are posterior, at the costovertebral junction. There are generally multiple fractures on contiguous, symmetric ribs;

  
  
• •
  

  metaphyseal fractures;

  
  
• •
  

  periosteal spurs;

  
  
• •
  

  some skull fractures: multiple fractures and depressed occipital fractures.

1.3.3.2.1

Clinical assessments

Check and update the height, weight and head circumference curves.

Thorough clinical screening for trauma, which must be photographed if found.

A neurological examination is, of course, essential; one should note the head circumference (in comparison with earlier figures), the state of the fontanel, axial tone and possible motor impairments.

1.3.3.2.2

Additional assessments

When faced with neurological clinical signs or a combination of the signs described above, the following additional examinations are required:

• •
  

  a computed tomography (CT) brain scan is the first-line examination in an emergency . It is a sensitive method for detecting haemorrhagic lesions: SDH, subarachnoid haemorrhage and (more rarely) haemorrhages of the brain parenchyma. The brain scan can also define the extent of any oedema. If symptoms persist (and even when the first scan is normal) a second scan can be performed 12 to 24 hours later;

  
  
• •
  

  an ophthalmological examination: it must be performed after dilatation, by an experienced ophthalmologist, within 48 to 72 hours at the latest (due to the rapid resorption of some types of RH). Photos must be taken whenever possible;

  
  
• •
  

  the value of MRI:

  
  
  
  
  • ∘
    

    when performed in the acute phase as soon as permitted by the child’s state, MRI is of significant diagnostic value for revealing lesions that are not visible on CT (i.e. small SDH, oedema and hypoxic lesions). Performance of an MRI scan depends on the child’s clinical state (stability). This is the examination of choice for having a complete overview of axial and extra-axial lesions, whether haemorrhagic or not . MRI enables the brain stem, spinal cord and neck region to be assessed, in addition to the brain itself.

    
    

    Kemp et al. reviewed the literature on children with severe spinal cord damage (24 children described by 15 studies) and recommended that head and neck MRI should be performed in any infant in whom HI is suspected, especially if there is unexplained deformation of the spine, focal neurological signs or skeletal lesions.

    
    

    Useful conventional sequences include the T1- and T2-weighted spin-echo sequences and the T2* echo gradient sequence. All are sensitive to the paramagnetic effect of the haemoglobin degradation products and enable determination of the approximate age of intraparenchymatous haemorrhage.

    
    

    The fluid-attenuated inversion recovery (FLAIR) sequence is very good for detecting subarachnoid haemorrhage and small SDH .

    
    

    So-called “advanced” techniques (such as magnetic susceptibility imaging, spectroscopy and diffusion imaging) have further increased the sensitivity and the diagnostic and prognostic value of MRI:

    
    
    
    
    • –
      

      magnetic susceptibility imaging can visualize very small areas of bleeding, whether recent or old ,

      
      
    • –
      

      spectroscopy can provide information on anatomical and functional damage to neurons and axons ,

      
      
    • –
      

      diffusion imaging can estimate changes in the volume and configuration of extracellular spaces and/or intracellular viscosity ,

      
      
    • –
      

      studies on SBS have demonstrated that early anomalies in diffusion imaging are compatible with cytotoxic-type oedema – probably as part of an associated ischaemic, hypoxic encephalopathy ;

    
    
    
    
  • ∘
    

    it is less urgent to perform MRI as part of the lesion screening process but it must be performed before discharge from hospital. The brain stem, spine and spinal cord must be studied and not just the brain. Furthermore, the MRI can reveal hypoxic lesions and lesions of different ages;

  
  
  
  
• •
  

  other necessary examinations:

  
  
  
  
  • ∘
    

    a complete blood count + platelet count, PT, APTT, coagulation factors,

    
    
  • ∘
    

    X-rays of the whole skeleton, performed according to the American Academy of Pediatrics guidelines . A skeletal survey is mandatory in all cases of suspected physical abuse in children younger than 2 years (not a “whole body” X-ray): radiographies of the axial skeleton (anteroposterior and lateral views of the thorax, and possibly an oblique view to see the ribs, the upper dorsal and lumbar spine; anteroposterior and lateral views of the pelvis, in order to see the medial lumbar spine; a lateral view of the lumbosacral spine, anteroposterior and lateral views of the cervical spine, anteroposterior and lateral views of the skull [and other views, if necessary]) and all limb segments. One should thus screen for recent or older fractures – particularly in the locations suggestive of SBS cited earlier.

    
    

    There are then three possibilities: the bone lesions are (i) very suggestive of abuse or (ii) not typical or (iii) absent. In the last two cases and if abuse is strongly suspected, scintigraphic examination can reveal lesions that are not visible on X-rays (e.g. rib fractures that do not yet show callus formation, very small diaphyseal fractures and early periosteal thickening). Another option (if the child can be placed in a safe environment) is to repeat the skeletal X-rays 10 or 15 days later and see whether there are any changes .

  
  

If the child is dead on admission to the hospital, the family/carers must be questioned as to any abnormal signs in the hours or days before death. In February 2007, the French National Authority for Health (HAS) recommended performing the following examinations (in addition to an analysis of the circumstances in which the child was found and the medical history) in all cases of unexpected infant death: a thorough clinical examination, a fundoscopy and an X-ray examinations by a paediatric radiologist (X-rays of the skull, spine, pelvis, all four limbs and the thorax) and CT or MRI imaging of the brain and, if possible, the whole body. Autopsy is essential but requires the parents’ consent (unless otherwise decided by the district prosecutor for forensic and legal reasons) and so must always be suggested. It must include a fundoscopy and a neuropathological examination of the brain, eyes and spinal cord. Several authors have described autopsy techniques in the young child . Ehrlich et al. insist on the need to evidence rupture of the corticodural veins. Several techniques for this have been suggested .

Some time after shaking, certain neurological or neuropsychological symptoms and some neuroradiological images can suggest shaking a posteriori. It is then necessary to look for a drop in the head circumference curve (at this stage, microcephaly is observed) and perform MRI.

1.3.4.1.1

Disorders of haemostasis

Congenital coagulation disorders (factor V, X or XIII deficiencies and haemophilia A) can lead to intraparenchymatous or extra-axial haemorrhage .

Severe thrombopenia can lead to intracranial haemorrhage (mainly intraparenchymatous haemorrhage).

1.3.4.1.2

Arteriovenous malformations

These are extremely rare below the age of 1 year and trigger subarachnoid haemorrhage (often associated with an intracerebral or intraventricular haemorrhage) rather than subdural haemorrhage.

1.3.4.1.3

Metabolic diseases

Depending on the context, one must screen for:

• •
  

  type 1 glutaric aciduria: 1 in 30,000 births. This condition often manifests itself by acute neurological distress in the first months of life, with a pseudoencephalitic clinical picture in children with macrocephaly and pre-existing hypotonia. Imaging can reveal suggestive anomalies: a broad lateral sulcus and lesions in the central grey nuclei. During disease progression, SDH is frequent and RHs are reported in 20 to 30% of cases . Faced with this characteristic clinical and radiological picture, a diagnosis of this metabolic disease is confirmed by the chromatographic assay of urinary organic acids ;

  
  
• •
  

  Menkes disease: this affects boys only (as a recessive, X-linked disease) and is also rare (1 per 250,000–300,000 births). It is a metabolic disease of copper absorption and induces multiple bone lesions, SDH , hypotonia, early convulsions and severe mental retardation in children surviving beyond the neonatal period. The twisted appearance of the hair is suggestive. Serum copper and ceruloplasmin assays make a biochemical diagnosis very easy.

1.3.4.1.4

Osteogenesis imperfecta

Two articles have stated that SDH is possible in this syndrome but did not indicate its frequency and a causal relationship has not been established. The clinical picture is quite different; the constitutive bone fragility results in diaphyseal fractures and not metaphyseal damage.

1.3.4.3.1

Different clinical situations can suggest a diagnosis of SBS

When an infant is brought in dead and does not correspond to terminal progression of a known pathology , a diagnosis of AHI must always be considered as a potential cause of unexpected death. The interview with the family must be performed with respect but must also be detailed, in order to establish the circumstances surrounding the death.

Some signs must be given special attention:

• •
  

  implausible explanations;

  
  
• •
  

  statements that change over time (although the parents’ emotional state can be an explanation);

  
  
• •
  

  suspicion of previous abuse and other poorly explained deaths in the family.

For all cases of unexpected infant death, it is advisable to obtain the parents’ consent for an autopsy that will potentially enable collection of data on shaking (bearing in mind that any indication of SBS should be reported to the district prosecutor, who will be able to order an autopsy as part of a forensic procedure).

When faced with inaugural, acute, neurological distress , the clinician should perform a brain CT scan as soon as possible and, if intracranial haemorrhage is evidenced, look for the simultaneous presence of other signs of shaking:

• •
  

  RH;

  
  
• •
  

  signs of abuse (skin and/or bone lesions);

  
  
• •
  

  explanations given by the family that are implausible or change over time.

If all these signs are found together, the diagnosis of SBS is highly probable, or even certain.

When faced with signs that strongly suggest neurological damage , such as those described above (changes in behaviour, poor food intake, poor contact, less smiling, decrease in the child’s capabilities; change in tone [axial hypotonia] or certain non-specific signs (vomiting, respiratory disorders [pauses, apnoea], pallor or an infant who appears to be in pain), it is essential to consider shaking:

• •
  

  palpation of the fontanel, measurement of the head circumference and examination of the growth curve (acts that must always be performed when examining an infant) can evidence a bulging fontanel and enlargement of the skull with a sharp upwards shift on the head circumference curve;

  
  
• •
  

  look for:

  
  
  
  
  • ∘
    

    a combination of the above-mentioned signs ( Table 1 ), found in a significant percentage of cases of shaking,

    
    
  • ∘
    

    other signs of abuse via clinical examination and analysis of the medical history;

  
  
  
  
• •
  

  note during the interview:

  
  
  
  
  • ∘
    

    the existence of any delay in seeking medical care,

    
    
  • ∘
    

    the fact that the explanations given by the family/carers are implausible, barely plausible or change over time or from one person to another;

  
  
  
  
• •
  

  look for RH;

  
  
• •
  

  perform a brain CT scan immediately (MRI being performed later, when permitted by the child’s state).

Some of these elements may be absent and it is impossible (given our current state of knowledge) to assess the relative importance of each of those elements present. However, certain data have more specific value:

• •
  

  in clinical terms: in addition to the interview data, the existence of apnoea, bruising (on the scalp or elsewhere);

  
  
• •
  

  the CT results: subdural or subarachnoid haemorrhages are more frequent (refer to the description above) than intraparenchymatous bleeding and EDH is extremely rare in cases of SBS;

  
  
• •
  

  the fundoscopy: the existence of RH of all types (even on one side only; Table 3 ), especially when profuse or flecked across the retina to the outer periphery, combined with one or several large, dome-shaped or plaque-shaped haemorrhages (sometimes with the characteristic aspect of haemorrhagic retinoschisis) or a perimacular retinal fold (Defoort-Dhellemmes type 3, which is almost completely pathognomonic for SBS).

One of these signs alone will not enable the physician to affirm a diagnosis of SBS. The combination of two or more of these signs is strongly suggestive, as long as differential diagnoses have been ruled out. The description of an act of shaking (sometimes by the perpetrator but more often by an eye witness) is, of course, a key element.

In any case, any suspicion of shaking must prompt the physician to tell the parents about his/her concern and the absolute need to hospitalize the child.

1.3.4.3.2

At the end of the clinical and radiological assessments (whatever the initial symptoms), the probability of a diagnosis of SBS will vary according to the lesions observed

In a child under 1 year of age and after having ruled out differential diagnoses:

• •
  

  a diagnosis of shaking is highly probable (or even certain) in cases with:

  
  
  
  
  • ∘
    

    multifocal extra-axial haemorrhages (SDH, subarachnoid haemorrhage) ( Figs. 1–3 ),

    
    
  • ∘
    

    AND RH that is profuse or flecked across the retina out to its periphery (Defoort-Dhellemmes type 3) ( Table 3 ),

    
    
  • ∘
    

    AND a clinical history that is absent, incoherent, changes over time or is incompatible with the observed lesions or the child’s age.

    
    

    The coexistence of these three diagnostic elements (as described here) prompts a diagnosis of AHI (probably by shaking).

    
    

    Other elements can be present and reinforce the diagnosis of shaking:

    
    
    
    
    • –
      

      hypoxic brain lesions,

      
      
    • –
      

      cervical lesions (spinal canal haematoma, spinal cord lesions and lesions of the occipitovertebral or cervicodorsal junctions),

      
      
    • –
      

      description of violent shaking by an eye witness;

    
    

  
  
  
  
• •
  

  a diagnosis of shaking is probable in cases with:

  
  
  
  
  • ∘
    

    EITHER multifocal extra-axial haemorrhages ( Figs. 1–3 ), in the presence or absence of RH of any type ( Table 3 ),

    
    
  • ∘
    

    OR localized extra-axial haemorrhage with type 2 or 3 RH,

    
    
  • ∘
    

    AND a clinical history that is absent, incoherent, changes over time or is incompatible with the observed lesions or the child’s age;

  
  
  
  
• •
  

  in cases of localized SDH and RH limited to the posterior pole (type 1 RH), with a clinical history that is absent, incoherent, changes over time or is incompatible with the lesions observed or the child’s age, there was no consensus within the Hearing Commission as to whether diagnosis of shaking must be considered as probable or possible;

  
  
• •
  

  a diagnosis of shaking is possible in cases with:

  
  
  
  
  • ∘
    

    localized SDH,

    
    
  • ∘
    

    AND a clinical history that is absent, incoherent, changes over time or is incompatible with the observed lesions or the child’s age;

  
  
  
  
• •
  

  a diagnosis of shaking can be ruled out in cases with:

  
  
  
  
  • ∘
    

    localized SDH ( Fig. 4 ), with a linear fracture and adjacent bruising in some cases,

    
    

    
    

    
    

    

    
    

    
    

    Localized right frontoparietal SDH and adjacent bruising of the scalp (indicated by an arrow) in a case of accidental head injury (HI).

    
    
    
    
  • ∘
    

    AND an unchanging clinical history compatible with the observed lesions and the child’s age and which features a description of a violent, accidental HI.

  
  

In all cases, the observation of any of the following is strongly suggestive of a diagnosis of abuse and should prompt the implementation of appropriate measures:

• •
  

  abuse-specific bone lesions and bruising (particularly on the scalp) in a child too young to walk unaided;

  
  
• •
  

  a history of abuse, or of unexpected poorly explained sibling death(s);

  
  
• •
  

  a delay in seeking medical assistance.

1.4.1.1.1

Subdural haematoma

Biomechanical evidence.– In their first study in 1987, Duhaime et al. estimated that shaking in the absence of an impact could not cause brain lesions or SDH. However, the reference values used in the study had been obtained in primates and the transposition to infants had not been validated.

Some more recent biomechanical studies have indicated that shaking in the absence of an impact is sufficient to rupture the corticodural veins. Roth et al. , by using a finite element model of the infant head to show that even though the pressure and shearing values are significantly higher in cases of impact, the relative extension of the corticodural veins is similar in the presence and absence of impact (180%) and exceeds the degree of stretching need to rupture the said veins in children under 3 months of age (150%) .

Clinical evidence.– Several clinical studies have confirmed that SDH may be caused by shaking in the absence of an impact . Moreover, SDH are more frequent in confirmed cases of AHI by shaking than in accidental HI in children of the same age .

Autopsy data.– Autopsy data confirm also the existence of SDH in the absence of any signs of external impact .

In conclusion.– There are sufficiently clinical, radiological, autopsy and biomechanical arguments to affirm that SDH can occur with shaking in the absence of any impact.

1.4.1.1.2

Retinal haemorrhage

Shaking in the absence of an impact can cause RH. In fact, RHs seem more strongly linked to the mechanism of shaking than to the occurrence of an impact.

Biomechanical elements.– Modelling of the infant eye has shown that during shaking, significant stress is applied to the posterior pole as a result of the strong traction exerted by the optic nerve on the retina. This explains the predominance of lesions to the posterior part of the retina .

Clinical elements.– RHs are found in cases of SBS both with and without cranial impacts and may be even more frequent during shaking alone (in the absence of an evidenced impact) .

Further evidence in favour of a link between the shearing mechanism and RH is the fact that in cases of AHI (compared with accidental HI), RH are:

• •
  

  more frequent: 80%, on average (ranging from 50 to 100%, depending on the study) in cases of shaking vs. 8.9% on average for all the cases in the literature on road accidents collated by Kivlin et al. ;

  
  
• •
  

  bilateral in the great majority of cases (83 to 90%) ;

  
  
• •
  

  often more abundant ;

  
  
• •
  

  more abundant in cases of certain SBS than in cases of probable SBS ( P < 0.0001) .

Type 3 RH extending out to the periphery are extremely rare after accidental HI:

• •
  

  they are extremely rare in road accidents ;

  
  
• •
  

  they have been reported only five times in other circumstances that were not simple falls (a television set or a 63 kg person falling on the child’s head; a child falling off a platform in a play area ), some of which are controversial because there were no independent eye witnesses, no autopsy data and no fundoscopy data from an ophthalmologist .

According to Betz et al. , massive RH covering more than 20 to 30% of the whole retina’s surface area cannot be explained by a simple, accidental HI and a banal fall in particular. According to Defoort-Dhellemmes , type 3 RH are considered as almost pathognomonic for shaking .

In conclusion.– Type 3 RH (but also types 1 and 2 RH) can occur during shaking in the absence of an impact and seem even more related to the mechanism of shaking than to the existence of an impact.

Type 3 RH are extremely rare in other circumstances and are thus almost pathognomonic for shaking.

1.4.1.2.1

Falls are the main explanation given by adults

It is particularly important to study the mechanism of falls because the latter explanation is that most frequently alleged by adults to justify the lesions observed in an infant .

It is probably often accepted in error in the absence of other plausible explanations, as very strongly suggested by several studies:

• •
  

  in 1991, Chadwick et al. reviewed the medical records of 317 consecutive admissions to a children’s trauma centre in San Diego between 1984 and 1988, for which a history of a fall was reported by the parents as being the cause of HI. The researchers analysed the data on the history as reported and did not evaluate their probability or credibility with respect to the diagnosis and the outcome. For the 283 children for whom the height of the fall was known, they found 7% of deaths for falls reportedly from a height of less than 1 metre (7 out of 100), compared with a rate mortality of 0% for falls from a height of between 1 and 3 m (0 out of 65) and of 0.8% for falls from a height of between 3 and 12 m (1 out of 118). The age of the child was not always specified but of the seven deaths for alleged falls from a height of less than 1 metre, two children had fallen from their own height (and thus were under the age of four because a 4-year-old is 1 metre tall, on average), two had fallen from a bed or a table (which suggests a young age) and two children (aged 6 weeks and 13 months) had fallen from someone’s arms. Lastly, the remaining case (an 11-month-old child) had fallen down the stairs;

  
  
• •
  

  in a prospective study of 398 falls by children admitted to Oakland Children’s Hospital over a 2-year period, Williams compared two populations of under-3s for whom the alleged causal mechanism was either a fall reported by family members or carers (53 cases) or by a neutral eye witness or several eye witnesses (106 cases). The researchers observed a death rate of 3.8% and a serious HI rate of 34% for cases with related eye witnesses and no deaths (other than a child who fell 21 m) and no serious HIs for cases with neutral or several eye witnesses;

  
  
• •
  

  in an editorial , Chadwick insisted on the fact that reliably witnessed cases in mechanistic studies are those occurring in hospital or an authorized care setting under certain conditions. He considered that cases for which the eye witnesses are carers or other children should not be included in databases concerning alleged mechanisms.

1.4.1.2.2

Mortality due to falls from a low height

In the child under 5 years of age, the immediate or differed mortality rate after a fall from a low height (< 1.5 m) is very low:

• •
  

  on the basis of a review of three databases in the United States, five book chapters, the work of two learned societies, seven literature reviews and 177 articles published in peer-reviewed journals, Chadwick et al. estimated the mortality rate at less than 0.48 per million under-5s per year (some cases were included in this study despite doubtful circumstances of occurrence). The estimation of the incidence under the age of 1 year was not specified;

  
  
• •
  

  in five studies, none of the 708 children who fell in a hospital setting (at least 94 of whom were under the age of one) died .

Denton and Mileusnic published the case of a 9-month-old child (looked after by its grandmother) who died 72 hours after a backwards fall from a 76-cm-high bed where it was sitting. The child (which did not have any clinical symptoms until the evening before death) was found dead in the morning. The autopsy evidenced mild parenchymatous brain lesions and slight SDH above a linear parietal fracture, without disjunction of the edges in spite of massive brain oedema (1035 g for an expected weight of 750 g). There were no RHs. However, this case (not considered by Chadwick et al. as resulting from a fall from a low height) cannot be taken into account since (i) the eye witness was not neutral and (ii) the absence of disjunction of the sutures (despite a large oedema) argues in favour of a very recent HI – which casts doubt on the alleged mechanism of a fall from a low height.

The article by Plunkett on 75,000 playground accident reports over a period of 11.5 years showed that death is possible but extremely rare after a fall from a low height. Eighteen children (0.024%) aged from 12 months to 13 years had died “following a fall from playground equipment from a height of between 0.6 to 3 m” (the height was judged in terms of the part of the body nearest to the ground at the time of the fall, rather than the distance above ground of the child’s centre of gravity). However, the following points must be noted for the eight deaths of children aged three or under (cases 1 to 8):

• •
  

  none of the children was under 1 year of age and the four youngest cases (cases 1 to 4) were 12 to 20 months old;

  
  
• •
  

  the eye witnesses were either family members in five of the eight cases (cases 1 to 4 and 6). This was particularly the case for the four youngest children and one other child (case 8);

  
  
• •
  

  an autopsy was not performed in three cases (cases 1, 2 and 7). In a fourth case (case 4), the autopsy was “limited”;

  
  
• •
  

  only one of the four youngest children received a full autopsy (case 3);

  
  
• •
  

  being swung or rocked was associated with the fall in three of the eight cases (cases 1, 3 and 6), including two of the four youngest children;

  
  
• •
  

  the lucid interval between the fall and the first symptoms was always less than 15 minutes and was even zero in three cases (cases 1, 3 and 6);

  
  
• •
  

  brain oedema was indicated as the cause of death – even for cases lacking an autopsy.

In the eight youngest children in this series, only case 5 (a 23-month-old child having fallen over a barrier around a platform situated 0.7 m above the ground) appears to be an incontestable HI due to a fall from a low height (a filmed case, with an autopsy). There was a 10-minute lucid interval before coma. The lesions observed before death were:

• •
  

  bilateral RH evidenced 24 hours after admission but with no other details (the fundoscopy was not performed by an ophthalmologist) and not described later at autopsy either;

  
  
• •
  

  a large right SDH with disappearance of the lateral ventricle and mild subfalcine herniation on the initial CT scan. Soft tissues were not studied. The SDH had been evacuated.

At autopsy, the following were found:

• •
  

  a right frontal impact;

  
  
• •
  

  a small residual SDH;

  
  
• •
  

  a right parietal parenchymatous contusion;

  
  
• •
  

  brain oedema with cerebellar herniation.

The low risk of death by falling from a low height is emphasized by the fact that even falls from a great height have a very low mortality rate:

• •
  

  two studies found no deaths for falls from heights of less than three storeys:

  
  
  
  
  • ∘
    

    Smith et al. : a study on falls from three storeys or less by 70 children aged from 10 months to 15 years (50% of whom were under 3 years of age),

    
    
  • ∘
    

    Barlow et al. : 61 children under 16 years of age;

  
  
  
  
• •
  

  in the series by Chadwick et al. , only one of the 118 children admitted after a fall from more than 3 m died (it was 11 months old). None of the 65 children having fallen from 1 to 3 m died.

1.4.1.2.3

Which clinical signs result from falls from a low height?

A few studies provide reliable data falls from a low height (corresponding to a mild HI) that have occurred in well-established circumstances (as observed by objective or several eye witnesses):

• •
  

  five studies concerning falls of hospitalized children found the same results: no deaths and one case with vigilance disorders (a neonate having fallen 1 m off a delivery table) among 708 children in total, including 493 children under the age of seven. On the basis of the details given by the articles, one can establish that at least 94 of the children were under the age of one ;

  
  
• •
  

  a prospective study of 106 children under 3 years of age with neutral eye witnesses did not find any serious clinical manifestations (bruising, abrasions, cuts or neurological disorders) and found only three skull fractures without loss of consciousness (a fall against a hard edge) for falls from a height of less than 1.5 m;

  
  
• •
  

  other data came from questionnaires put to parents on possible previous falls and consequences. Kravitz et al. questioned the mothers of children under 1 year of age ( n = 536) and established that over half of these children had fallen from a low height at some moment in their life, with very few serious injuries and no fatalities . Fifteen children had been hospitalized. The reported symptoms were lethargy in 14 cases, a loss of consciousness in two cases, vomiting in eight cases and convulsions in two cases. The total number of consultations has not been indicated;

  
  
• •
  

  of 3357 falls (including 97% concerning the head) in 2554 children monitored longitudinally from birth to the age of 6 months, an injury was reported in 437 cases (as bruising in 244 cases and as a fracture or mild HI [commotion] in less than 1% of cases [21 times]) .

On the basis of these studies, it is clear that a fall from a height of 1.5 m hardly ever leads to death (fewer than 0.48 per million under-5s per year ) and rarely produces clinical signs.

1.4.1.2.4

Can subdural haematomas be observed after falls from a low height?

Available data.– A few articles are available:

• •
  

  on the basis of two studies of infant death [first study: 63 accidental deaths, including 10 by accidental HI and six by falls, excluding deaths due to drowning, fires, burns and road accidents, with 25 (39%) children under 1 year of age, 18 (28%) between 1 and 2 years of age and five (8%) between 2 and 3 years of age; second study: 21 accidental deaths, including two due to HI and with nine children (43%) under 1 year of age] performed in a city with two million inhabitants over a 24-year period (1975 to 1985 and 1986 to 1999) and other studies in the literature, Case concluded that in cases of falls from a height of less than 1.8 m (6 feet), fractures of the skull were observed in 1 to 3% of cases. The factures are generally linear and are not accompanied by intracranial haemorrhage or neurological impairments. Fewer than 1% of these fractures caused EDH (1–3 per 10,000) and even fewer caused a contact subdural haemorrhage. When these haemorrhages are sufficiently voluminous to create a mass effect, death can occur by intracranial hypertension. In all these cases, the EDH and SDH were focal and were situated at or adjacent to the fracture;

  
  
• •
  

  based on his experience and a literature review, Dias found that:

  
  
  
  
  • ∘
    

    falls from a height of less than 1.5 m (5 feet) can cause intracranial damage, albeit very rarely (usually skull fractures, EDH and subarachnoid haemorrhage but also localised SDH),

    
    
  • ∘
    

    diffuse SDH and brain hypodensities “are extraordinarily rare if they exist at all” (Dias did not find any case reports),

    
    
  • ∘
    

    the lesions are often silent or poorly symptomatic and do not leave sequelae,

    
    
  • ∘
    

    falls from a low height are even less likely to be lethal. Sixty-seven to 75% of the deaths were due to lesions with a mass effect (EDH, SDH);

  
  
  
  
• •
  

  Matschke et al. found just one case of SDH due to accidental HI (a road accident) in a study of 715 consecutive autopsies of children under 1 year of age [with deaths due to malformation (almost 300), perinatal complications (175 cases), infections (under 100 cases) and metabolic damage (30 cases)]. The SDH was localized.

Biomechanical aspects.– Bertocci et al. have studied linear accelerations of the head for falls from the height of a bed onto different types of surfaces. Even though further research is still necessary to determine the lesional limits of the child (since the values obtained [55 to 418 m/s ² ] are very much lower than the lesional threshold in the adult [900 m/s ² ]), the work suggests that (for the falls studied) serious brain lesions (such as acute SDH or intracerebral haemorrhage) cannot occur, since the linear acceleration is too low.

Conclusions concerning SDHs.– SDH after falls from a low height are extremely rare and are localised. They are mostly adjacent to the fracture line. Given the extreme rarity of this situation, the observation of SDH after an alleged fall from a low height should prompt the healthcare professional to first consider AHI as the cause.

1.4.1.2.5

Retinal haemorrhage after a fall from a low height

Regardless of the height of the fall, RHs are rarely described in falls with one or more neutral, reliable eye witnesses.

After a fall from a low height, RHs (when observed) are never widespread in terms of either surface area or depth, as described in cases of AHI. The RHs are then more likely to be associated with an EDH and are moderate: small intraretinal or preretinal haemorrhages, situated on the posterior pole of the eye . They are usually unilateral but can be bilateral and often asymmetric (in the series reported by Defoort-Dhellemmes ).

1.4.1.2.6

Subdural haematoma and retinal haemorrhages after a fall from a low height

Christian et al. have reported three cases of traumatic SDH occurring at home, including two cases with RHs that were limited to the posterior pole (with pre-, intra- or subretinal haemorrhage) in one eye only. However, none of these cases corresponded to a fall from a low height:

• •
  

  a fall down 13 steps by a 13-month-old child, with local signs of impact (skin contusions on the forehead, nose abrasion) and RH on the same side as the SDH;

  
  
• •
  

  a fall through banisters onto a concrete floor by a 7-month-old child, with a skull fracture and right-side brain parenchymatous contusion but no RH;

  
  
• •
  

  a fall suffered by a 9-month-old infant after being swung: it was stated that the father lost his grip while swinging the child in his arms. The rear of the child’s head had then hit the ground. This case cannot be considered as an example of a fall from a low height because:

  
  
  
  
  • ∘
    

    the eye witness (a family friend) was not neutral,

    
    
  • ∘
    

    the combination of swinging and a loss of grip during swinging cannot be likened to a simple fall from a low height,

    
    
  • ∘
    

    the height was very low (30 cm),

    
    
  • ∘
    

    there was no lucid interval,

    
    
  • ∘
    

    in contrast to the significant internal lesions (parieto-occipital SDH and RH), there were no external lesions, bruising or soft tissue damage observed in the other two cases to suggest an impact.

  
  

Vinchon et al. gathered 3 years of prospective data on 150 under-2s having suffered an HI. Seventy-three cases were considered to be accidental. There were 12 road accidents. The trauma occurred at home in 55 cases and was been reported as a fall in 53 of these: 13 falls from a seat, 10 down the stairs, nine from the arms of an adult, nine from a table, eight falls when standing or from a great height and two from a bed. The eye witness (when there was one) was a family member or friend in all cases. In some cases, the traumatic nature of the incident has been deduced from the observed lesions. RHs were observed in five of the 73 children but Vinchon et al. did not state which children were concerned, how old they were and what the associated intracranial lesions were. On the basis of this article, it is impossible to identify any case of a child with both SDH and RH after a fall from a low height (especially in the absence of any neutral eye witnesses).

For the four cases of children aged between 12 to 20 months of age who died after playground accidents described by Plunkett , only one (case 4) had undergone fundoscopy and presented SDH and bilateral RH (RH in several layers). The fundoscopy had not been performed by an ophthalmologist. The eye witness was not neutral.

In summary, there are no literature cases of children under the age of one presenting both SDH and RH after a fall from a low height.

1.4.1.4.1

Biomechanical aspects

Adults sometimes mention a mechanism that involves shaking the infant in “baby bouncer” seat by an older child.

Jones et al. studied a manikin corresponding to a 5-week-old child and submitted it to rocking in a “baby bouncer” seat. The acceleration values ranged from 6 to 16 G (compared with the 50 G peak acceleration considered to be lesional, 750 G resulting from an impact against a wall and 177 G corresponding to violent shaking). Moreover, when the duration of action is taken into account, the resulting index ranged from 2 to 52 (in comparison with a value of 1000 for a 50% chance of severe neurological damage in an adult). Even though if further studies are necessary to determine the lesional limits in the child, the values obtained (2 to 52) are well below the lesional threshold in the adult (1000) and suggest that it is very likely that serious brain lesions cannot occur through shaking in a “baby bouncer” seat.

1.4.1.4.2

Clinical aspects

There are no literature reports or expert feedback on any cases of HI with RH or SDH having occurred during play.

1.4.1.5.1

Subdural haematoma

Biomechanical elements.– The subdural lesions occurring during childbirth correspond to “static” compression phenomena, rather than impacts.

Clinical aspects.– The type of delivery influences the frequency of occurrence of intracranial haemorrhage, which occurs above all in cases of abnormal labour. Towner et al. studied 583,340 consecutive live, first births of children weighing 2500 to 4000 g (excluding multiple pregnancies) registered in a Californian database between 1992 and 1994, and examined the relationship between morbidity and the type of delivery. The rate of intracranial haemorrhages is higher after vacuum-assisted delivery, force forceps delivery and caesarean section performed after induction of labour than for non-instrumented delivery or caesarean section performed before induction of labour. This suggests that the risk is above all related to an abnormal course of labour.

We reviewed prospective studies on asymptomatic children . None of these children had undergone fundoscopy.

Asymptomatic SDH can be seen soon after childbirth (within 72 h), with a variable frequency of 9 to 46% depending on the investigational technique (ultrasound vs. MRI), the MRI field strength and the use of coronal slice to differentiate between supratentorial and subtentorial locations, the date of the examination and the type of delivery.

SDH is found more frequently when investigated soon after birth, with a more powerful MRI system, after vaginal delivery (14 to 33%), for vacuum-assisted delivery (40 to 77%) or forceps delivery (30 to 33%), for higher birth weights , lesions in the genital area , premature deliveries (24% in the study by Sezen ) and for first children (20% in the study by Sezen ).

These SDHs are situated on the subtentorial level, the supratentorial level or both. Whitby et al. and Looney et al. described a preponderance of subtentorial sites – in contrast to Rooks et al. .

Supratentorial SDH can be occipital but also parietal or temporal. In the series reported by Rooks et al. , all 46 cases of supratentorial SDH were situated in the posterior half of the skull: 30 out of 46 (65%) were interhemispheric and posterior, 29 out of 46 (63%) were occipital and 22 out of 46 (22%) were close to the tentorium cerebelli. Twenty children (43%) had also SDH of the posterior fossa. None of the cases of asymptomatic, supratentorial SDH in the series was frontal, as also reported by Whitby et al. . In the series reported by Rooks et al., 74% of the cases (34 out of 46) had SDH in two or three areas. The SDH was always homogeneous in all the MRI sequences . In the study by Looney et al. (using a 3 Tesla MRI machine), the 12 cases of multiple SDH were homogeneous and the lesions had the same age. The haemorrhage was often mild and appeared in the form of a thin film. None of the 46 cases described by Rooks et al. had an extradural subarachnoid or intraparenchymatous haemorrhage.

In summary, in the context of childbirth, SDH can be evidenced by brain imaging in asymptomatic neonates. The SDH is generally situated generally in a supratentorial site in the posterior half of the skull (not the anterior half) or in the posterior fossa – sites which are found in SBS. These SDHs are often multifocal. They are homogeneous and have the same age (when this information is given).

In the referenced studies, the follow-up data (when complete) showed that these asymptomatic SDHs did not progress to chronic SDH and resolved spontaneously within 1 month , except for one case in Rooks et al. which initially had (i) bilateral occipital SDH and (ii) SDH of the posterior fossa and then developed (at 26 days of age) a new left-side, frontal SDH. An examination at 5 months confirmed the disappearance of the SDH but enlargement of the pericerebral spaces.

1.4.1.5.2

Retinal haemorrhage

On the basis of several prospective studies, it seems that about one in three term neonates have RHs. They are found after all types of childbirth but occur more frequently after forceps or (even more so) vacuum-assisted deliveries. These haemorrhages are almost always superficial and deep intraretinal and uni- or bilateral (52%, i.e. 26 out of 50 of the cases reported by Emerson). They are often numerous (> 10) and extend out to the periphery in a third of cases. In 15 to 25% of cases, the RH has a white centre. They disappear very rapidly – often in less than 3 days . The RHs are very rarely still found at 1 month of age (two out of 202 cases, both of which involved vacuum-assisted delivery) and are never found after 2 months .

In summary, in a context of childbirth (particularly after forceps and vacuum-assisted deliveries) in asymptomatic neonates, RH of various types exist in a third of cases. These are mainly Defoort-Dhellemmes type 1 or 2 and disappear in less than a month (and mostly within a few days).

RH and even haemorrhages of the vitreous matter can also occur as a complication of prematurity-associated retinopathy in 1 to 2% of premature infants and can last for longer.

1.4.1.6.1

Subdural haematoma

Radiological aspects.– Three retrospective studies on the radiological signs of serious hypoxia have shown the absence of SDH in children (including some with prolonged cardiac arrest ).

Autopsy data.– Byard et al. did not evidence any macroscopic SDH in a series of 82 autopsies with histological proof of hypoxic-ischaemic encephalopathy in children under the age of three, other than in a context of SBS. They did not differentiate between foetuses and neonates.

Matschke et al. (who did not differentiate between foetuses and neonates either) found only 50 macroscopically observable cases of SDH in 715 consecutive autopsies performed in children under 1 year of age, most of whom had suffered from hypoxia before death [malformation (40.4%), the outcome of a perinatal complication (24.3%), infection (12.2%), metabolic disease (4.5%), sudden infant death (3.4%), inflicted trauma (2.4%) and other non-natural causes (2.1%)]. A cause was found for all the cases of SDH when a full dataset was available, i.e. in 48 cases out of 50: 14 cases of SBS, one case of accidental injury, 13 haemostasis disorders, 13 perinatal deaths, four deaths due to metabolic causes, three due to infectious disease and cases of two very localised SDH after sudden death. The most frequent cause was inflicted trauma (14 cases) and the latter mechanism was also mentioned for the two cases with incomplete data.

Thus, SDH in a hypoxic context is extremely rare. This is also what clinical studies show .

Upon histological examination, Geddes et al. found 36 intradural haemorrhages (IDH) in 50 foetuses and children but only one (unilateral) SDH was found macroscopically (in a child born after 25 weeks of pregnancy, who died of an infection after 1 week of life). In the 30 children born alive, only 13 had a histologically manifest IDH. Twelve of the 13 deaths occurred very early in perinatal period (during the first week for 11 of them) or the neonatal period (the first month of life for the 12th case) and only one of the children with clear IDH corresponded to the average age for SBS. Childbirth had been excluded as a possible cause of bleeding in 72% of the cases (36 out of 50: in 11 of the 17 in utero deaths but also in 13 of the 18 children who had lived more than 5 days [median: 23 days]), whereas haemorrhages due to childbirth can last up to 1 month. The difference in the frequency of occurrence of IDH (pooling “+” cases and “++” cases) according to whether there was hypoxia and anoxia or not was not significant ( P = 0.15) but it was stated that this lack of significance was due to the small number of cases. Despite this lack of significance, a “unified hypothesis” was presented. It supposed that subdural bleeding seen in some cases of HI in the child can result from passage of blood from the intracranial veins to the subdural space, due to a combination of severe hypoxia, brain oedema and an elevation in the central venous pressure. According to Geddes et al., subdural bleeding may thus not be due to rupture of the corticodural veins but may testify to immaturity – with no need for an impact or considerable force.

Cohen and Scheimberg evidenced clear IDH and signs of hypoxia of various degrees of severity in 25 foetuses and 30 children who died within the 19 first days of life, including 25 in the first week. SDH was found in two thirds of cases (16 foetuses and 20 neonates). The IDH was predominantly in the posterior part of the falx cerebri and the tentorium cerebelli.

In summary, hypoxia:

• •
  

  is likely, in a particular population constituting of foetuses and children who died in the first month of life (mostly in the first week), to cause, contribute to or be associated with histologically detectable IDH and, at most, very mild posterior supratentorial and subtentorial subdural effusion (a thin film). Intradural suffusions may be missed on imaging or even on autopsy when the sample is inadequate;

  
  
• •
  

  does not produce any macroscopic SDH in children over the age of 1 month.

1.4.1.6.2

Retinal haemorrhages

Acute hypoxia (such as that produced during suffocation, which frequently produces petechiae at the surface of the lungs, the heart or other internal organs) does not cause any RH.

1.4.1.7.1

Subdural haematoma

There are no known literature reports of an association between SDH and cardiorespiratory resuscitation.

The lesions observed following resuscitation are above all related to whatever circumstances had justified the use of resuscitation manoeuvres. The most frequently encountered lesions after resuscitation are anoxic-ischaemic lesions with brain oedema, diffuse brain ischemia that predominates in the less well vascularised regions, depending on the age of the child, and ischaemic lesions of the central grey nuclei. In the series, no SDH was reported .

In the retrospective study by Bush et al. of 211 children (mean age: 19 months) having required cardiorespiratory resuscitation (a study inclusion criterion) lasting 45 minutes on average (range: 3 to 180 minutes) for sudden infant death (56%), drowning (8%), congenital cardiopathy (7%) and pneumonia (4%), brain lesions were not reported.

The discovery of SDH after a cardiorespiratory resuscitation does not at all prove that resuscitation manoeuvres were responsible for this haematoma.

1.4.1.7.2

Retinal haemorrhage

There are few available studies concerning pre-hospital or hospital-based resuscitation by healthcare professionals. RH are extremely rare in young babies with neither HI nor haemorrhagic disease and who have received cardiorespiratory resuscitation (even when prolonged). In such cases, the RH are few in number, punctiform or flame-shaped, localized (sometimes with a white centre) and situated in the posterior pole in one or both eyes .

1.4.1.8.1

Retinal haemorrhage after convulsions

None of 32 children admitted after convulsions had any RH .

Only one child (aged 8 months) out of 153 (aged from 2 months to 2 years) examined after an epileptic attack had RH. The RH was unilateral and flame-like around the optic disc .

None of 32 children (mean age: 30 months) hospitalized after an epileptic attack had any RH, despite the need for resuscitation in a few cases .

1.4.1.8.2

Retinal haemorrhage after coughing

No RH was observed in a series of 100 consecutive children with a persistent cough .

1.4.1.8.3

Retinal haemorrhage after vomiting

No RH was observed in a series of 100 children with vomiting due to pyloric stenosis .

1.4.2.2.1

Retinal lesions

It is difficult to date RH .

RH located at the posterior pole disappears within a few days, whereas massive haemorrhages (of all types, i.e. intra-, pre- or subretinal) that are profuse, coating the whole retina or flecked out to the periphery and associated with a unilateral or bilateral premacular haemorrhagic plaque, sometimes immediately suggesting haemorrhagic retinoschisis) take several weeks or months to disappear.

RH with a white centre (usually considered to be old) can be seen on the first day post-trauma.

Since intraretinal RH disappears within a few days, only the association with scarring, the sequelae of preretinal or subretinal haemorrhages (circular white scar on the retinal folds, macular retraction syndrome, zones of pigmentation and retinal atrophy – particularly in macular areas or right out at the edge of the retina – and fibroglial scarring) can be considered as reflecting the coexistence of older and recent lesions.

1.4.2.2.2

Subdural haematoma and subarachnoid haemorrhage

Time windows for dating.– On imaging, the change over time of intraparenchymatous haemorrhages is known:

• •
  

  on a CT scan and for up to 2 days, the haemorrhage is hyperdense (compared with the adjacent parenchyma) and then (from 2 days to 1 month) becomes isodense and then (beyond 1 month) hypodense;

  
  
• •
  

  on MRI, the characteristics of the intraparenchymatous haemorrhage are determined by the paramagnetic effects of the haemoglobin degradation products, the strength of the magnetic field and the type of sequence. However, this change over time is not applicable to SDH and subarachnoid haemorrhages in the infant, where it depends on the haemoglobin degradation products and physiological process such as blood clot formation and retraction, the change over time in the red blood cell concentration, the suffusion of serous liquid through the damaged dura mater and the extravasation of CSF and/or blood from the subarachnoid spaces through a torn arachnoid membrane .

The time window for dating acute SDH is wide for both CT and MRI: for CT, the hyperdensity can persist up to 11 days after the trauma; using MRI, the T1 hyperintensity can extend beyond day 13 .

According to a CT and MRI study of HIs of known date in adults, the time window for hyperintensity on T1, FLAIR and even diffusion sequences was wide: the FLAIR hypersignal is visible for up to 3 weeks and the T1 hypersignal becomes visible after 24 hours and persists for up to 4 months.

In conclusion, it is difficult to achieve precise dating by imaging .

Furthermore, the images are not always easy to interpret:

• •
  

  in the acute stage of cases of compressive, diffuse brain oedema, SDH in a case of SBS can be absent or missed in the first imaging dataset . It may also not be detected if it is of low abundance and only located at the vertex;

  
  
• •
  

  hyperdensity of the longitudinal fissure can reflect congestion or acute, intradural bleeding but also subarachnoid haemorrhage, a normal falx or relative hypodensity of the oedematous brain parenchyma .

Lastly, it should be noted that:

• •
  

  heterogeneous SDH (suggesting repeated trauma and a diagnosis of SBS according to some authors) has also been reported in accidental HI ;

  
  
• •
  

  the heterogeneous character of SDH can reflect rebleeding from the capillaries in a neoformed membrane ;

  
  
• •
  

  some children can be subjected to several episodes of shaking, with the possible superposition of lesions of different ages (i.e. acute and older lesions). The presence of several SDHs with different density then becomes highly informative.

In summary, it is difficult to date lesions precisely with imaging.

It would be advisable to launch multicentre studies with contrasted-enhanced CT and MRI in all cases of accidental or non-accidental HI, in order to determine the potential for evidencing venous rupture and determining narrower time windows for dating.

1.4.2.3.1

Examination of SDH

Macroscopic examination.– Dating is based on the appearance of an adherence between the SDH and the dura mater. For the first 3 days, the blood is liquid. Adherence starts on the fourth day. A neomembrane forms around the bleeding and becomes visible at between 7 to 10 days and is well developed from week 2 to week 4. From week 6 onwards, the contents of the blood clot take on a classic “engine oil” aspect. Encapsulation is complete at around week 8.

Histology.– The dura mater and the blood clot must be studied as a whole. One must study the junction zone between the healthy dura mater and the zone of the dura mater to which the clot begins to adhere. The dating involves the appearance of the clot, the dura mater and the brain’s meningeal surface. Schematically, the change over time is as follows :

• •
  

  in the first 2 days, polynuclear neutrophils invade the clot and then progressively increase in number;

  
  
• •
  

  on days 2–3, macrophages arrive and the nuclei swell;

  
  
• •
  

  at around days 4–5, red blood cell lysis begins;

  
  
• •
  

  from day 2 onwards, the neomembrane starts to form via the attachment of a layer of fibroblasts to the dural surface. Roughly, one layer of fibroblasts forms each day. Hence, one can distinguish between the following times: from 2 to 5 days, with the presence of two to five layers of fibroblasts at the dural surface;

  
  
• •
  

  by week 1, the angiofibroblastic invasion of the clot begins; the dural membrane has 12 to 14 layers of fibroblasts at this point;

  
  
• •
  

  by week 2, giant neovessels appear. The neomembrane is then about a third or half as thick as the dura mater;

  
  
• •
  

  in parallel, Perls staining is always used for all samples, in order to identify iron in blood and siderotic macrophages which appear on days 2–3 in the subdural compartment following blood contact. This corresponds to metabolism of the heme group from haemoglobin after its release into the tissues by bleeding.

1.4.2.3.2

Examination of brain contusions

The presence of possible brain contusions and their change over time (albeit not very precise) can aid the dating process.

Histology.– In the child, three time phases can be distinguished:

• •
  

  in the acute phase (during which there is little bleeding), one observe pycnosis of the nuclei of astrocytes and nearby neurons and the appearance of a few polynuclear cells;

  
  
• •
  

  at 36 hours and up to week 3, the second phase starts with a cell reaction that intensifies progressively: it includes vacuolization of the capillaries and the oligodendrocytes, an inflammatory cell reaction, the activation and swelling of glial cells and activation of the microglia;

  
  
• •
  

  after week 3, the third phase (gliosis) occurs.

Immunohistochemistry.– Oehmichen et al. have described the immunohistochemical change over time in the macrophages appearing in contusion areas as a function of the time since the trauma. This may enable dating on the basis of this marker.

1.4.2.3.3

Examination of the thymus

Another essential element in early-stage dating is the presence or absence of an acute thymic involution. This thymic reaction (mediated by corticoids synthesized and released by the adrenals) is the body’s reaction to stress. Four stages of acute thymic involution have been described, depending on the histological appearance. These stages are significantly correlated with the time since the onset of stress :

• •
  

  in the first 12 hours, the appearance of the thymus does not change;

  
  
• •
  

  Grade I acute involution corresponds to stress lasting for 12 to 24 hours: the thymic cortex takes on a faint “starry sky” appearance, due to the arrival of macrophages that come to phagocytose apoptotic lymphocytes in the cortex;

  
  
• •
  

  Grade II corresponds to stress lasting for 24 to 48 hours and is characterized by a very marked “starry sky” aspect, associated with necrosis of Hassall’s corpuscles;

  
  
• •
  

  Grade III corresponds to stress lasting for 48 to 72 hours: the differentiation between the cortical and spinal cord regions is no longer visible due to lymphocyte depopulation (their re-release into the circulation and intracortical apoptosis);

  
  
• •
  

  Grade IV, for stress lasting for 3 days or more: there is complete lymphoid depopulation of the cortical zones; hence, the medulla appears to occupy the whole thymic surface and Hassall’s corpuscles are large and very close together.

This gradation of the thymus’s appearance is very reliable and very useful.

1.4.2.3.4

Analysis of bruising

This analysis can be performed in different ways : inspection with the naked eye or on photographs, histological analysis, biochemical assays (rarely used in routine practice) and objective evaluation of the colour by colorimetry and spectrometry.

Dating is random, due to high inter- and intra-individual variability. One can consider than a violet colour means that the bruising is less than 24 hours old and that the yellow colour does not appear before 18 hours at least .

By pooling and cross-checking all the clinical, imaging and pathology lab data, one can narrow the estimated window during which the trauma occurred to a day or even half a day.

1.4.3.1.1

First hypothesis: expansion of the subarachnoid spaces is a factor favouring the occurrence of SDH

The first hypothesis is that this widening is part of an expansion of the subarachnoid spaces (ESAS), a clinical entity which may predispose to the occurrence of SDH.

Characteristics of ESAS.– ESAS – a transient disorder of CSF resorption – has sometimes been called “benign external hydrocephalus of the infant”. ESAS is a clinical entity that is accepted by the paediatric and neurosurgical communities and is characterised by widening of the arachnoid spaces in an infant (in boys more than girls) in the absence of developmental abnormalities or clinical neurological signs.

The main clinical sign is macrocrania, with an increase in head circumference and a large, bulging fontanel. Some studies have reported a high frequency of children with ESAS but no macrocrania, corresponding almost completely to a physiological particularity . However, these studies were retrospective, the brain CT scan was performed in hospitalized children who were all symptomatic (retarded development, epilepsy or tone disorders ) and, lastly, clinical, radiological and ophthalmological signs of abuse were not always screened for.

In this entity, head circumference is often (but not always) high at birth and steadily increases to more than 2 standard deviations (SD) between 3 and 6 months. It generally stabilizes between the age of 1 and 2 years and then moves smoothly back to around +2 SD afterwards . The subarachnoid spaces normalize after 2 years .

There may be a family history of the condition.

The diagnosis of ESAS is based on the detection (in brain imaging) of liquid-filled expansion of the pericerebral spaces and the cerebral fissures (testifying to subarachnoid involvement) that predominates in frontal regions (mostly on both sides of the body) and the longitudinal fissure (more than 3 mm at the convexity, more than 4 mm in the corticocranial area near the foramina and more than 6 mm for the interhemisphere gap). The effusion is isodense with the CSF, the arachnoid cannot be distinguished and the parenchyma is normal. The expansion then decreases but moderate skull-brain detachment may persist .

Analysis of the literature concerning the hypothesis whereby ESAS favours the occurrence of SDH.– Clinical aspects:

• •
  

  this hypothesis is rarely and poorly documented in the literature ;

  
  
• •
  

  the main obstacles to progressing our knowledge of this precise aspect of the pathogenesis of SDH in the infant are as follows: sources of methodological bias, old studies that ignore the potential role of abuse and (above all) the difficulty in proving the prior existence of ESAS and the absence of trauma;

  
  
• •
  

  the case reports with coexisting SDH and ESAS have only very rarely been explored thoroughly in terms of a potential AHI: fundoscopy and radiological bone examinations were rarely performed, an investigation was not always performed and the head circumference curve had not always been plotted. Whether the eyewitness was a family member/carer or not was not always stated;

  
  
• •
  

  eight studies (either series or case reports) have reported follow-up data for children presenting ESAS , including just one prospective study . In all, 215 patients were followed up for an average of 40.3 months (cumulative follow-up: 524 patient years). Only two studies reported a total of four patients, including one in which the performance of fundoscopy and an investigation was not stated and three patients who definitely did not undergone fundoscopy or an investigation.

Biomechanical aspects:

• •
  

  a biomechanical model suggests that ripping of the corticodural veins is promoted by ESAS . In contrast, a recent finite element analysis of the head of a 6-month-old infant by Raul et al. showed that ESAS has a shock-absorbing effect by reducing the relative displacement of the brain relative to the skull .

In summary, there is nothing in the literature to support the hypothesis whereby ESAS favours SDH in the infant.

1.4.3.1.2

Second hypothesis: expansion of the pericerebral space is the consequence of a previously undiagnosed head injury

The second hypothesis is that the widening of the pericerebral space is, in fact, a consequence of a previously undiagnosed HI.

Ewing-Cobbs et al. prospectively compared the initial imaging results for two groups of HI victims aged between 1 month to 6 years (30 children with AHI and 29 with accidental HI). The two groups were similar in terms of data on the perinatal period (apart from the birth weight) and the neonatal period. In the 30 children with AHI, there were nine cases of brain atrophy (bilateral, cortical and supratentorial), four cases of subdural hygroma (unilateral and affecting the convexity and the longitudinal fissure) and seven cases of ventricular dilatation. None of these anomalies were found in the 29 children with AHI. The nine children with atrophy all had some other associated brain anomalies: four had ventricular dilatation, four others had ventricular dilatation and hygroma and the ninth had subdural hygroma but normal-sized ventricles. Several haemorrhages were present in 68% of the AHI group vs. 32% in the accidental HI group. The mean number of extra-axial haemorrhages was 2.1 in the AHI group and 1.3 in the accidental HI group.

In a prospective study of children under the age of three with SDH, Feldman et al. found chronic SDH or acute SDH on chronic mixed SDH in 17 (44%) of 29 children with AHI, in eight (67%) children with trauma of unknown origin and in none of the 15 (0%) children with an accidental HI.

These studies suggest that the widening of the pericerebral spaces observed during the initial phase corresponds to the sequelae of an undiagnosed HI, rather than ESAS.

Follow-up studies on a cohort of infants with ESAS and shaken infants (evaluating the change over time in the brain imaging data) would be very useful.

1.5.1.3.1

Civil proceedings

Following the report, the district prosecutor can deliver an ordonnance de placement provisoire (OPP, temporary care order) if he/she feels that it is necessary to protect the child without delay (even during a hospital stay, in view of the risk of early recovery of the child by its parents, while enabling continuation of medical care).

The report also enables the protection of other children (whether siblings or not) likely to be in danger by virtue of the risk of recurrence, whoever the perpetrator is.

A protective OPP does not influence any subsequent legal proceedings and is strongly recommended to give the child immediate protection. This decision has no right of appeal and is valid for 8 days. It may involve referring the case to a juvenile court judge.

The district prosecutor who pronounces an OPP has “ must petition the appropriate judge within one week, who will maintain, modify or revoke the order ” (article 375-5 of the French Civil Code). Hence, the juvenile court judge can decide to prolong or not protective measures for the child or to modify it if needed. He/she can pronounce a care order (which is likely to be appealed by the holders of parental responsibility) that can last several months and may be renewed. The judge can also implement educational assistance measures.

When the judge is petitioned by the district prosecutor having pronounced the emergency, temporary care order, “he/she summons the parties and pronounces within two weeks of being petitioned. Otherwise, the child is handed back on request to its father, mother or guardian or to person or service to which it had been confided” (Civil Proceedings Code, article 1184, paragraph 3).

“The decision on pursual or not of the case must be taken within six months of the decision to issue temporary measures. Otherwise, the child is handed back on request to its father, mother or guardian or to person or service to which it had been confided.

If the inquiry has not finished within the term set out in the previous paragraph, the judge may, after seeking advice from the district prosecutor, extend this term for a period not exceeding six months” (Civil Proceedings Code, article 1185).

1.5.1.3.2

Criminal proceedings

Referring the matter to the district prosecutor in parallel enables the evidence to be examined in terms of a criminal offence and triggers an immediate criminal inquiry, in order to identify and potentially prosecute the perpetrator(s), decide on whether or not the enquiry should be pursued to establish the facts or identify the perpetrator(s) and assess whether the explanations given by the family/carers are plausible or change over time. The criminal inquiry may also exonerate other people involved. During the criminal inquiry, one must not lose sight of the need to protect the child.

Submission of the case to a judge enables the child’s future rights to be preserved (notably its right to compensation for the injury, even if the latter was caused by one of its parents). It is essential to maintain the possibility of compensation that will help pay for future care (private-practice occupational therapy and psychomotor therapy not covered by the social security system, individual help at school, etc.), limit the sequelae and enable the provision of appropriate support during childhood and then during adulthood in the event of handicap (a third-party carer, in particular).

In the event of death and medical and forensic opposition to burial, the police office leading the criminal inquiry and/or the district prosecutor must be informed. The latter opens an inquiry or proceedings to establish the cause of death (article 74 of the Code of Criminal Procedure) and can order a forensic autopsy.

The case must be submitted to the district prosecutor as soon as possible, so that the inquiry can start rapidly.

1.5.1.9.1

Risks incurred by the physician

Article 44 of the French Code of Medical Ethics obliges the physician (except under particular circumstances that he/she shall consider in all conscience) to alert the judicial, medical or administrative authorities in the event of abuse or privation in a child aged 15 or under. For cases of severe abuse, a report to the State Prosecutor is recommended by the French Medical Council. If the physician does not alert any of these authorities, he/she may be sanctioned under article 223-6 of the Penal Code.

The fact that this obligation is accompanied by a conscience clause does not make it any less imperative. Hence, if the physician does not alert the authorities, he/she must explain the exceptional circumstances that prompted this decision. Consideration of these exceptional circumstances in all conscience does not exempt the physician from explaining the absence of a report but does protect him/her against being sanctioned. However, in the latter case, repeated refusal to report a suspected case (which would appear to be the result of a predetermined policy) would prevent the physician from invoking the conscience clause because the supposedly exceptional circumstances would then have become general.

If a physician does invoke the conscience clause, he/she must nevertheless ensure that effective protective measures are implemented in order to avoid recurrence of the shaking or other abuse.

If further abuse occurs, the physician may be accused of inaction, on the basis of article 223-6 of the Penal Code ⁸

8 “Anyone who, being able without risk to himself or to third parties to prevent by immediate action a felony or a misdemeanour against the bodily integrity of a person, wilfully abstains from doing so, is punished by five years’ imprisonment and a fine of € 75,000. The same penalties apply to anyone who wilfully fails to render to a person in danger any assistance which, without risk to himself or to third parties, he could render him either by his own action, or by initiating rescue operations.”

In the event of death following highly probable (or even certain) shaking, a physician who attests that the infant died of natural causes and that the death does not pose any legal or forensic problems commits the offence of falsification, which is punishable by 3 years imprisonment, a fine of €45,000 and being struck off (article 441-1 of the Penal Code).

The physician also exposes him/herself to disciplinary sanctions for violation of the following dispositions: article 3 of the French Code of Medical Ethics as codified by article R. 4127-3 of the Code of Public Health (“ The physician must, under all circumstances, comply with the principles of morality, probity and devotion that are essential for practicing medicine. ”) and article 28 of the French Code of Medical Ethics (codified by article R. 4127-28 of the Code of Public Health (“ The issuance of a biased report or a certificate out of kindness is prohibited. ”).

As soon as a diagnosis of shaking appears to be the cause or a possible cause of the infant’s death, the physician who delivered the death certificate must mention that there are forensic and legal problems.

1.5.1.9.2

Risks incurred by other professionals

Articles 434-1 and 434-3 of the Penal Code on the obligation to report crimes and abuse include an exception for the people subject to the professional duty of confidentiality and specify that the latter may not be convicted on this basis. Nevertheless, these people may choose not to report the abuse but always have a legal obligation to assist a person in danger (articles 223-5 to 223-7 of the Penal Code).

1.5.1.14.1

The criminal inquiry

The goal of the criminal inquiry is to determine the exact circumstances leading to injury of the child and the causes and motives for the shaking, to evaluate the risk of recurrence, to establish whether a crime has been committed ¹⁴

14 There are two kinds of penal charges: offences are assaults causing a temporary work incapacity for less than a week or more than a week or involuntary acts of violence causing a temporary work incapacity of more than three months or less than three months; crimes are murder with intent and manslaughter.

The possible perpetrators must be confronted with the contradiction between their version of the facts and the medical diagnosis. This confrontation is essential but complex. The date or probable date of the injury is an important piece of information for the criminal inquiry. It prompts the investigation of all persons likely to have been in the presence of the child during the suggested period (police custody, a hearing before the judge, seizure of documents, comments, etc.). The police custody is decided by police officer in the criminal investigation department.

The members of the hospital team and professionals in contact with the family may be interviewed – notably in the event of contradiction with the parents’ explanations or the absence of an explanation. They must state their knowledge of the situation and transmit any documents required for the purposes of the inquiry . Hospital staff and social workers cannot claim an overriding professional duty of confidentiality when information is requested by a crime squad officer duly mandated and supervised by a legal authority.

If the physician receives a summons as part of a preliminary inquiry, he/she may communicate the facts known to him/her but is not obliged to do so – he/she can plead a professional duty of confidentiality.

The criminal investigation department performs all the procedural acts, under the supervision of the district prosecutor’s office (to which regular reports will be made). In this respect, an autopsy or an expert medical opinion may be ordered and the members of the hospital service may be called for interview.

Importantly, experts who become involved after the start of the judicial inquiry should be competent in the field concerned: paediatrics, ophthalmology or even biomechanics. In the absence of appointed medical experts, information on the context, compatibility of accounts and the drafting of medical reports are necessary for the investigating officers.

During the inquiry or investigation, the medical records may be requisitioned. This requisition is the only way for criminal investigation department officers to obtain the medical information held by the physicians and healthcare establishment. This procedure must comply with procedures that maintain physician-patient confidentiality. The requisition must be performed in the presence of the establishment’s Director and the Head of Service concerned.

The medical records are generally handed over voluntarily by the hospital service to a criminal investigation department officer bearing letters rogatory (i.e. without the need for seizure) (appeal judgement 95-811.40 dated 20/09/1995 and Circular 97-13 dated 24/11/1997).

It is advisable to photocopy requested documents prior to their transmission, in order to (i) comply with the healthcare establishment’s obligation to store its medical data and (ii) counter the risk of not having the requisitioned items returned.

The presence of a Medical Council representative is recommended during the requisition of medical records (Circular 97-13 dated 24/11/1997 ).

An inventory of requisitioned items should always be made before the latter are placed under seal. However, if drawing up an inventory at the time is problematic, it is possible to have the items placed under seal temporarily (article 56 of the Code of Criminal Procedure).

It is advisable to request the return of requisitioned items. This request must be sent to the investigating judge if it is made during the legal proceedings (article 99, paragraph 1 of the Code of Criminal Procedure) or to the ruling jurisdiction if the case is closed (Article 373 and article 484, paragraph 2 of the Code of Criminal Procedure).

1.5.1.14.2

The outcomes of the criminal inquiry

The procedure may ¹⁵

15 Article 40-1 of the Code of Criminal Procedure: “Where he considers that facts brought to his attention in accordance with the provisions of article 40 constitute an offence committed by a person whose identity and domicile are known, and for which there is no legal provision blocking the implementation of a public prosecution, the district prosecutor with territorial jurisdiction decides if it is appropriate: (1) to initiate a prosecution; (2) implement alternative proceedings to a prosecution, in accordance with the provisions of articles 41-1 or 41-2; (3) or to close the case without taking any further action, where the particular circumstances linked to the commission of the offence justify this.”

• •
  

  be abandoned forthwith if the offense cannot be proven due to lack of evidence or lack of an identifiable perpetrator;

  
  
• •
  

  prompt the opening of a judicial investigation (obligatory for offences classified as crimes) so that an examining judge can pursue the inquiry. The later may request assistance from medical experts. At the end of the investigation, the case may be:

  
  
  
  
  • ∘
    

    dismissed if the offence is not sufficiently characterized,

    
    
  • ∘
    

    judged before a Magistrates Court (if the act is considered to be an offence with relatively minor consequences),

    
    
  • ∘
    

    judged before Court of Assizes (for offences with serious consequences);

  
  
  
  
• •
  

  lead to judgement of the person or persons accused:

  
  
  
  
  • ∘
    

    by a Magistrates Court, as a result of a direct summons, convocation by a criminal investigation department, through a statement or by an immediate appearance in court in the few cases where the facts are simple, can be dealt with by the magistrate and can be ascribed with certainty to the accused,

    
    
  • ∘
    

    by Court of Assizes if the act is considered to be a crime; the accused person may appear in court prior to arrest or after a period of provisional custody.

  
  

1.5.1.14.3

Initiation of a child protection procedure

Initiation of a child protection procedure is intended to provide the child with judicial protection. The district prosecutor can (regardless of whether an OPP is ordered or not) refer the case to a juvenile court judge. The latter generally orders investigation and a 6-month care review procedure ( investigation et orientation éducative ¹⁶

16 The social inquiry and the IOE are due to be renamed as a “mesure judiciare d’investigation educative” (MJIE).

17 Art. 375 of the Civil Code: “Where the health, security or morality of a not emancipated minor are imperilled, or where the conditions of his education are seriously endangered, measures of educational assistance may be judicially ordered.”

In theory, these measures are independent of the outcome of any legal proceedings seeking to establish whether the child’s family or carers are guilty or not. If the presumption of innocence leads to the dismissal of legal proceedings (in the event of doubt as to the intention to cause harm or the identity of the perpetrator of a violent act), this does not means that the child may no longer require protection.

1.5.1.14.4

What are the potential criminal charges and sentences for perpetrators of shaking?

There is no specific criminal charge of “shaking” per se but the act always constitutes a violation of the law. The usual criminal charges relate to involuntary or voluntary assault with aggravating circumstances. It is important to differentiate between the intent to shake and consideration of the consequences of this act.

The severity of sentences handed down by the courts illustrates very clearly the clear disapproval of all violent acts inflicted on a child under the age of 15 by an ascendant (the legal, natural, or adoptive parents or the grand-parents, if the child has been entrusted to them) or a person with responsibility for the child (a child-minder, the latters’ spouse or partner, a partner of one of the parents or any other person charged with looking after the child).

In fact, the sentences are even heavier for perpetrators who have a duty to look after the child’s health and safety. The maximum prison sentence for perpetrators of shaking as set out below, depending on whether the act is classified as an offense or a crime (fines are voluntarily not referred to):

• •
  

  for a crime to be tried before a Court of Assizes:

  
  
  
  
  • ∘
    

    manslaughter of a child aged 15 or under: a 20-year prison sentence, or 30 years if the perpetrator is an ascendant or a person with authority over the child,

    
    
  • ∘
    

    assault leading to mutilation or permanent infirmity in a child aged 15 or under: a 15-year prison sentence, or 20 years if the perpetrator is an ascendant or a person with authority over the child;

  
  
  
  
• •
  

  for an offence to be tried before a Magistrates Court:

  
  
  
  
  • ∘
    

    assault leading to ITT for more than 1 week in a child aged 15 or under: a 5-prison sentence, or 10 years if the perpetrator is an ascendant or a person with authority over the child,

    
    
  • ∘
    

    assault leading to a ITT for a week or less in a child aged 15 or under: a 3-year prison sentence, or 5 years if the perpetrator is an ascendant or a person with authority over the child.

  
  

There is also a charge relating to repeated violence to a child aged 15 or under; depending on the outcome, this charge may be qualified as a crime or an offence:

• •
  

  the act is qualified as a crime if the repeated violence led to death of the child (with a 30-year prison sentence) or mutilation or permanent infirmity (with a 20-year prison sentence);

  
  
• •
  

  the act is qualified as an offence if the repeated violence led to ITT of more than a week (with a 10-year prison sentence) or less than a week (with a 5-year prison sentence).

In addition to the main perpetrator, other people in the child’s environment may also be prosecuted for other charges, notably:

• •
  

  not having prevented a crime or offence against the child, failure to assist a person in danger: a 5-year prison sentence;

  
  
• •
  

  lack of care and/or endangerment of a child’s state of health by an ascendant or a person with authority over a child aged 15 or under (for example, this is the case for a passive mother who is aware of her child’s precarious status and fails to act, thus worsening the situation): a 7-year prison sentence.

It has been suggested that shaking an infant can be qualified as involuntary damage to a person’s well-being caused by clumsiness, imprudence, inattention, negligence or a failure to ensure safety or prudence imposed by the law (this type of qualification mainly applies to traffic offenses, labour law offences and injuries caused by animals). These charges are unlikely to be used in the context of SBS because the act itself is always qualified as voluntary violence (with intent; as stated by the Court of Cassation’s judgement dated January 7, 1998, which criticized the Court of Appeal for not having sufficiently established that the accused [an ascendant charged with manslaughter of a child aged 15 or under] was aware that his actions could damage the child’s physical integrity of its child but added the facts pertaining to the accusation were indeed considered as a crime under the law and that the proceedings were legitimate). There do not appear to be any cases of jurisprudence that qualify shaking an infant as an involuntary act.

If the act is qualified as an offense or crime by a judge, an ad hoc legal guardian ¹⁸

18 Article 706-50 of the Code of Criminal Procedure: “The district prosecutor or the investigating judge seized for intentional offences committed against a minor appoint an ad hoc administrator where the protection of the interests of the minor is not completely ensured by his legal representatives or by one of them. The ad hoc administrator ensures the protection of the interests of the minor and exercises if necessary in the latter’s name the rights open to a civil party. In the event of a civil party petition, the judge has an advocate appointed officially for the minor if none has been already chosen. The previous provisions are applicable before the trial court.”

Additional sentences may be handed down. In cases where shaking is admitted, the most adequate additional sentence is a temporary or permanent professional ban on the person charged with looking after the child (5 years at most, in this instance).

It is important to note that qualifying the act of shaking is of huge importance. The whole process of protection is enabled or stopped according to whether the act is qualified or not. The same is therefore true for the compensation.

1.5.1.14.5

Under which conditions can the victim receive compensation?

Compensation is possible once the act of shaking is established. It is therefore important that it be established, even if the perpetrator is never identified. Reporting the shaking to the legal authorities is essential for enabling the possible award of long-term compensation to the victim of shaking.

Since shaking may cause severe, invalidating injuries, the question of compensation for the victim arises. The compensation process is rendered more complex by two factors: the victim’s young age (which prevents the child from acting in a legal capacity) and the fact that the perpetrator of the injury is often the child’s legal representative, charged with safeguarding his/her interests. Furthermore, the perpetrator’s economic situation may not correspond at all to the amount of compensation awarded as damages to the victim. Lastly, the inquiry does not always enable identification of the perpetrator with certainty.

The law has solved the first problem by instituting the ad hoc legal guardian, enabling to resolve the contradiction generated by the fact that the parent is the legal representative of the abused child. Hence, it is important for the latter to be appointed.

The second series of difficulties (resulting from the perpetrator’s meagre financial resources or, more generally, when then perpetrator cannot be identified) has been taken into account by legislative introduction of the CIVI. As in the field of road accidents or terrorist attacks, the law replaces the traditional mechanism of individual responsibility (where the perpetrator of the injury is the only person liable to pay damages) by shared responsibility, where the state takes charge of compensating the victims of certain types of damage or injury (although the state can pursue the perpetrator, if ever the latter becomes known).

The powers and attributions of the CIVI are defined in articles 706-3 to 706-15 of the Code of Criminal Procedure. This commission exists within all county courts and is composed of two professional magistrates and one lay assessor. It can be petitioned by “ any person having suffered injury as a result of confirmed criminal acts, whether voluntary or not in nature ” and which have led to death, permanent incapacity or temporary incapacity of more than a month or which constitute a crime of rape, sexual assault, sexual injury without violence or slave trading. This definition covers all the charges likely to be brought in cases of SBS. The compensation system also concerns situations where the perpetrator has not been identified with certainty, although the shaking has been confirmed. The only case in which a child will not be compensated by the CIVI is when the inquiry has quashed the diagnosis of shaking and excluded any criminal behaviour leading to injury.

The matter can be referred to the CIVI immediately, without waiting for the outcome of the judicial proceedings. A request for compensation must be presented within 3 years of the date of the offence or in the year following the final verdict issued in a public or civil prosecution before the criminal courts, in compliance with article 706-5 of the Code of Criminal Procedure. The plaintiff may, however, be freed from this debarment if he/she was unable to exercise his/her rights within the prescribed time interval, if the injury worsens or “ for any other legitimate reason ”. The fact that the victim is a child suspends the time restriction; the victim can pursue the matter before the CIVI up until his/her 21st birthday.

Any compensation awarded by the CIVI is paid out of the national Victims of Terrorism and other Offenses Fund. Given that the cases for compensation are quite burdensome for the Fund, the CIVI tends to appoint a medical expert to re-evaluate the sequelae.

Proceedings before the CIVI do not prevent the victim from acting as a plaintiff in court against the perpetrator of the acts, although the commission is not obliged to wait for the end of legal proceedings before compensating the victim. The CIVI may be petitioned immediately, without waiting for the end of legal proceedings, by the ad hoc legal guardian or even by a parent. If the court awards more compensation that has been granted by the CIVI, the victim can petition the latter for additional compensation.

After the trial (before the Magistrates Court or the Court of Assizes), the matter can be referred to the CIVI with a view to obtaining full compensation on the basis of the expert medical opinion. Petitioning the CIVI before the trial is always possible but it is advisable to be in possession of the latest legal and medical evidence.

1.5.3.1.1

Team coordination

For a diagnosis of “probable” SBS, the medical team (at least two healthcare professionals) must meet and review the symptoms and the reported facts, pool and compare the gathered information, make a diagnosis, consider (as appropriate) any other associated mistreatment, clarify each person’s positions and discuss whether a report should be filed. This meeting must take place without delay must be summarized in writing.

In addition, a multidisciplinary medical, psychological and social evaluation of the situation is essential, in order to analyse the child’s situation, identify risk factors, consider the child’s outcome and consider child protection measures. Consulting with colleagues is never a waste of time. The healthcare professional should not confuse speed with haste.

1.5.3.1.2

Reporting

Each hospital service should define the disclosure procedures to be followed by its staff in cases of shaking, as follows:

• •
  

  establish an initial, emergency report that includes the essential information (see the textbox above: Section 1.5.1.6 ), and then add further information after a multidisciplinary evaluation;

  
  
• •
  

  the report should be signed by the attending hospital physician and social services staff. In some cases, it may be completed by the psychologist or psychiatrist who has interviewed the family. It is essential to provide investigators with information on the context and the latter’s compatibility with the medical facts;

  
  
• •
  

  the report should be sent by fax and by registered post with signed-for delivery, while keeping a copy for the hospital records;

  
  
• •
  

  if the report is sent directly to the district prosecutor’s office, a copy must be sent to the Chairperson of the County Council.

An ordonnance of placement provisoire (OPP) ordered à titre conservatoire by the district prosecutor will not influence subsequent legal proceedings and is recommended – if only to provide the child with immediate protection.

Disclosure should be made as quickly as possible, in order to provide the child with immediate protection and not jeopardize the inquiry: the sooner after the event the report is filed, the sooner an effective criminal inquiry can start.

In this respect, the County’s procedure (established in compliance with the 2007 Child Protection Act and involving the council itself, the district prosecutor’s office, hospitals and other parties likely to be aware of children in danger in general and shaking victims in particular) should contain advice that is clear and is applicable to the county as a whole.

As soon as a diagnosis of SBS is stated as the cause of (or one of the causes of) an infant’s death, the physician who delivers the death certificate must mention the fact that this death raises legal issues.

1.5.3.1.3

Dialogue with the parents and the family

In the diagnostic phase, dialogue with the parents is essential – notably in order to note their explanations concerning the child’s injuries. Issues such as forming an opinion on the circumstances of and justifications for shaking, identifying the perpetrator and deciding whether the latter should be prosecuted are not diagnostic objectives and are the sole responsibility of the district prosecutor.

The physician should inform the parents that in light of the severity of their child’s condition, the district prosecutor’s office will be contacted. He/she should also specify that the district prosecutor will decide on any further actions to be taken. This approach, which means taking time and making an effort to listen, is best led by a senior physician in the presence of another staff member. It is essential to try to make the parents understand that the goal is not simply to denounce them but to draw attention to a dangerous situation and protect their child.

The parents must be informed of the report (except when counter to the child’s interest), the possibility that an OPP may be issued by the district prosecutor and the possibility of filing a complaint against person or persons unknown if they deny the shaking and attribute it to someone else.

However, this necessary dialogue and the obligation to provide the parents with information must not hinder or jeopardize the criminal inquiry.

1.5.3.1.4

Exchanging information with the district prosecutor’s office

We recommend the use of a form for exchanging information between the district prosecutor’s office and the reporting person. To initiate this practice, the form should feature in the reporting procedure.

1.5.3.1.5

Requisition of medical records

It is advisable to photocopy requisitioned documents prior to their transmission, in order to (i) comply with the healthcare establishment’s obligation to store its medical data and (ii) counter the risk of not having the seized items returned.

It is advisable to request the return of requisitioned items. This request must be sent to the investigating judge if the latter is involved in the legal proceedings (article 99, paragraph 1 of the Code of Criminal Procedure) or to the ruling jurisdiction if the prosecution is over (Article 373 and article 484, paragraph 2 of the Code of Criminal Procedure).

1.5.3.1.6

Disclosure of concern

To disclose concern to the County Child Abuse Prevention Office (CRIP), the physician should follow the county’s procedure, in compliance with article L. 226-3 of the Family and Social Action Code ²³

23 “… procedures to this end must be established by the Chairperson of the County Council, the State’s representative in the county and the legal authorities with a view to centralizing disclosures of concern within a specialist unit for gathering, processing and evaluating this information within the County Child Abuse Prevention Office”.

1.5.3.4.1

For professionals dealing with infants

Initial training, continuous professional education and staff support must be implemented in order to maintain high levels of awareness.

In general, any professional likely to become aware of shaking in some way must receive at least basic training on this issue: social services staff, child care managers, magistrates, police officers, etc.

When justified by the context, it is advisable to identify all the resources that can potentially be leveraged in the family environment to help parents better fulfil their role. In general, personal problems experienced by people looking after children should act as a warning sign for social services staff and healthcare professionals.

1.5.3.4.2

For parents and families

It is advisable to organise frequent, appropriate awareness-raising and information campaigns (targeted at people in contact with infants, above all) on the dangers of shaking and the precautions that should be taken.

In maternity wards and in the days following discharge to home, parents should always be made aware of the danger of shaking.

“Parent and child drop-in points” at which the relationship between the parents and the child can be built or strengthened are recommended solutions.

It is important to offer appropriate support to parents and implement support measures when requested. Referral to child care professionals (notably with home visits) must be envisaged. The same is true for various forms of home help (provided by the social services).

1.5.3.4.3

In the child’s interest

We recommend that France borrows elements of Canadian law in which an ad hoc legal guardian can be appointed whenever necessary in the child’s interest and in defence of the child’s rights.

2.3.2.3.1

Facteurs de risque liés à l’enfant

Sex-ratio : prédominance des garçons (sex-ratio G/F entre 1,3 et 2,6).

Prématurité : taux de prématurité plus élevé que dans la population générale : 11 à 21 % (11 % dans la série de Mireau ) versus 7 à 8 % dans la population générale.

Grossesses multiples : elles seraient plus fréquentes (gémellité dans 5 % des cas de Mireau ) que dans la population générale (1,5 % de grossesses multiples).

Les pleurs ne peuvent être considérés comme un facteur de risque à proprement parler, mais ils peuvent être un élément déclenchant de comportement maltraitant envers les nourrissons , sachant que la tolérance des parents aux pleurs de l’enfant est très variable : une consultation pour pleurs d’un nourrisson ne devrait pas avoir pour seul but de trouver la cause des pleurs, mais également d’évaluer le ressenti et les réactions des parents à ces pleurs.

2.3.2.3.2

Facteurs de risque liés à l’auteur des faits

Dans les séries où l’auteur des faits a été identifié, qu’il y ait eu ou non aveu de sa part, cet auteur est dans la majorité des cas (70 %) un homme, le père plus souvent que le compagnon de la mère.

Les adultes non apparentés constituent également une catégorie significative d’auteurs potentiels de maltraitance : dans la série de 151 cas de TCI examinés par l’équipe de Starling , les boyfriends de la mère sont en cause dans 20,5 % des cas, les baby-sitters de sexe féminin dans 17,3 % des cas.

2.3.2.3.3

Facteurs de risque liés aux parents

En ce qui concerne le contexte socio-économique, les études sont très contradictoires, tous les milieux peuvent être concernés ; les facteurs de vulnérabilité évoqués restent à documenter (premier enfant, nouvelle grossesse, reprise du travail, méconnaissance des besoins et des comportements normaux de l’enfant, isolement social et familial, histoire de violences familiales, troubles psychiatriques passés ou présents, abus de drogue illicite et d’alcool, etc.). Mireau en 2005 note que les parents ont une importante méconnaissance des besoins et comportements normaux de l’enfant. Le jeune âge des parents est fréquemment souligné par les auteurs et signalé comme un facteur de risque potentiel .

Une étude prospective réalisée de janvier 1998 à septembre 2006 détaille le contexte socio-économique de 25 cas de TC non accidentels observés en dix ans dans une région écossaise : 76 % des cas provenaient des quartiers les plus défavorisés du point de vue de l’éducation, de l’instruction et des habiletés sociales, 72 % de ceux où la criminalité était la plus élevée, 68 % de ceux où la santé était la moins bonne, 60 % de ceux où le revenu était le plus bas, 52 % de ceux où la qualité des habitations était la plus mauvaise, 48 % de ceux où le chômage était le plus important.

Ces résultats sont battus en brèche par des études antérieures : deux études de 2000 portant sur la même cohorte d’enfants notent, l’une que la majorité des parents a une situation professionnelle stable (81 % des mères sont sur le marché du travail), l’autre qu’une grande partie des parents a un niveau d’études secondaire ou supérieur.

Écarter le diagnostic de TCI en raison d’un contexte sociofamilial apparemment favorable peut en tout cas être à l’origine de diagnostics « manqués » .

Cependant, d’une façon générale, les difficultés liées au contexte de vie des personnes s’occupant de l’enfant doivent attirer l’attention du professionnel qui en a connaissance.

La relation du SBS avec un facteur ethnique a été étudiée. Fortin fait la synthèse des articles et conclut que l’origine ethnique ne constitue pas un facteur significativement plus associé au SBS qu’au TC accidentel. Les données suggèrent plutôt que la variable « origine ethnique » constitue un facteur de risque lié à d’autres marqueurs sociaux qui augmentent chez l’enfant le risque de blessures.

En résumé , les données disponibles à ce jour, bien que parcellaires et parfois contradictoires, suggèrent que les enfants de sexe masculin, premiers-nés, âgés de moins de six mois, nés prématurément d’une grossesse compliquée ou multiple, vivant avec des parents ayant une histoire présente ou passée d’abus de substances psychoactives (alcool, drogues) ou de violence familiale et/ou ayant une méconnaissance des stratégies de gestion de la relation avec leur nourrisson, sont plus à risque d’être victimes du SBS. Cet acte de violence peut bien sûr survenir sans qu’il y ait de tels facteurs de risque socio-économiques et culturels. De plus, s’il peut être juste de penser que l’enfant victime de SBS appartient plus souvent au groupe à risque ci-dessus, il est faux de croire que la majorité des enfants présentant ces caractéristiques sont victimes de cette forme de maltraitance.

2.3.3.1.1

Sont susceptibles d’être lésés dans le SBS les méninges (hémorragies sous-durales ou sous-arachnoïdiennes), l’encéphale, l’œil et la moelle épinière

D’autres lésions peuvent être associées : fractures des membres, des côtes, du crâne, ecchymoses du scalp, hématomes des muscles du cou, lésions rachidiennes postérieures.

Les données les plus précises sont apportées par les séries neuro-anatomopathologiques, mais elles concernent évidemment les formes les plus graves puisque responsables du décès de l’enfant.

Billette , à partir de 93 examens neuropathologiques d’enfants victimes de secouement sans impact crânien visible, décrit :

• •
  

  81 hématomes sous-duraux (HSD), 65 hémorragies sous-arachnoïdiennes, 14 hémorragies intraparenchymateuses, 69 cas d’œdème cérébral, 41 engagements cérébraux ;

  
  
• •
  

  56 hémorragies intraoculaires ;

  
  
• •
  

  21 lésions de la moelle épinière ;

  
  
• •
  

  l’état du rachis n’est pas mentionné ;

  
  
• •
  

  les lésions observées ne sont pas spécifiques du mécanisme du décès.

Une autre étude trouve des hémorragies épidurales cervicales et des lésions axonales focales du tronc cérébral et des racines des nerfs spinaux dans 11 cas parmi 37 de TCI et aucun chez 14 cas-témoins, morts d’autres causes.

Différents types de lésions du parenchyme cérébral peuvent être vus :

• •
  

  lésions anoxiques du cortex, des noyaux gris, des thalamus : ces lésions se traduisent par des hypodensités, rarement hémorragiques, associées à une perte de contraste entre les substances blanche et grise ;

  
  
• •
  

  œdème cérébral se traduisant par une diminution des espaces liquidiens ;

  
  
• •
  

  contusions en particulier dans les régions frontales et temporales, et à la jonction substance blanche–substance grise.

Les collections hématiques extra-axiales (HSD, éventuellement associés à des hémorragies sous-arachnoïdiennes) présentent des caractères particuliers dans le SBS ( Fig. 1–3 ). Elles sont habituellement plurifocales, bilatérales et fines, sans effet de masse. Elles recouvrent la convexité des hémisphères cérébraux, s’accumulent dans la scissure interhémisphérique en position déclive et longent l’insertion de la tente du cervelet. La reconnaissance d’une hémorragie de la faux du cerveau est très évocatrice du diagnostic de SBS , de même que les collections sous-durales de la fosse postérieure (cf. figures ci-dessous). Il faut cependant remarquer que l’HSD n’est pas observé chez tous les bébés secoués : il n’existe que dans 72 à 93 % des cas dans les articles revus par Christophe .

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

A public hearing “Shaken baby syndrome: Guidelines on establishing a robust diagnosis and the procedures to be adopted by healthcare and social services staff”. Guidelines issued by the Hearing Commission A public hearing. “Shaken baby syndrome: Guidelines on establishing a robust diagnosis and the procedures to be adopted by healthcare and social services staff”. Scoping report Éditorial A public hearing. “Shaken baby syndrome: Guidelines on establishing a robust diagnosis and the procedures to be adopted by healthcare and social services staff”. Scoping report A public hearing. “Shaken baby syndrome: Guidelines on establishing a robust diagnosis and the procedures to be adopted by healthcare and social services staff”. Scoping report A public hearing “Shaken baby syndrome: Guidelines on establishing a robust diagnosis and the procedures to be adopted by healthcare and social services staff”. Guidelines issued by the Hearing Commission

Stay updated, free articles. Join our Telegram channel

Join