Objective To describe the range of intracranial injuries encountered in 0–2-year-olds in cases of accidental head injury where the mechanism of trauma was well characterised and to assess the clinical consequences.
Design A retrospective review of imaging and clinical data.
Setting Two tertiary paediatric referral centres.
Patients All children aged 0–2 undergoing cranial CT as indicated by National Institute for Health and Clinical Excellence guidance for head injury from 2006 to 2011. After exclusion criteria, 149 patients were included.
Main outcome measures Rates of skull fracture, intracranial haemorrhage, parenchymal injuries and ischaemic change per type of mechanism of injury. Rates of neurological sequelae on follow-up.
Results Skull fractures were demonstrated in 54 (36%) patients of whom 17 (11%) had thin underlying subdural haemorrhage (SDH). Extradural haemorrhage complicated one fracture and two cases of isolated subdural haematomas were seen without skull fracture. Radiologically evident brain parenchymal injuries were present in three patients, all with mechanisms of injury involving high levels of force; severe neurological sequelae were only seen in one patient, who had diffuse hypoxic–ischaemic damage at presentation and whose (accidental) mechanism of injury involved extensive acceleration/deceleration/translational forces.
Conclusions Skull fractures and focal SDH are relatively common following minor trauma in this age group but in the vast majority of cases there are no long-term neurological sequelae. Conversely, diffuse brain injury with severe subsequent neurological impairment was only seen in patients with a correspondingly severe mechanism of injury.
- Child Abuse
- Accident & Emergency
Statistics from Altmetric.com
What is already known on this topic
Head injuries are common in children and in the majority they do not cause radiologically evident craniocerebral trauma
While there are very rare reports of fatal injury from low-level falls, the vast majority are clinically benign
Previous case series have not concentrated on findings of AHI in infancy—the age range with the highest prevalence of non-accidental injury
What this study adds
The vast majority of cross-sectional imaging findings of domestic head injury in infancy consist of either normal scans or simple skull fracture/focal SDH
In the vast majority of cases in this age group there are no significant clinical/neurological sequelae
Brain parenchymal or hypoxic-ischaemic damage in accidental head trauma is only encountered with mechanisms of injury involving severe forces in this age group
Accidental head injury (AHI) in childhood is a frequent occurrence causing approximately 650,000 emergency department attendances and 50 000 hospital admissions per year in the USA.1 Low-level falls are common and largely benign, the majority of injuries not reaching medical attention.2 Previous publications addressing the risks of low-level falls in children estimated that the risk of fatality is less than one in a million up to the age of five.3 Other case series reporting fatalities after short-level falls contained large numbers of cases of inflicted injury4 ,5; a very heterogeneous population of fall heights, defining ‘low level’ up to 15 feet/4.5 m6 ,7; and paucity of witnessed accounts.6 ,8 When cases of proven inflicted injury are excluded, the outcome from radiologically evident head injury (skull fracture, intracranial haemorrhage) in short vertical falls appears benign.9
The literature base regarding the consequences of AHI in early childhood (0–2-year age group) is sparser, perhaps because it is relatively uncommon in this subgroup. This fact is important because the presentation of a child in this age group with head injury often prompts consideration of inflicted or non-AHI (NAHI).
Certain patterns of imaging findings are known to be highly associated with NAHI.3–,5 In such cases, a well-established pathway of further investigation, including ophthalmological assessment and radiographic skeletal survey, along with multi-agency family/history verification, is followed to establish further evidence of abuse.10 A question that frequently confronts radiologists and paediatricians is whether or not imaging findings of skull fracture and/or intracranial haemorrhage can be explained by the mechanism of injury described by the caregiver. In addition, particularly in the context of concerns regarding potentially inappropriate mechanism of injury and potential judicial proceedings, it remains uncertain whether causality can be determined on the basis of radiological imaging. The minimum level of force required to cause intracranial injury remains unknown.
A body of opinion frequently referred to by legal defence teams is that low-level (generally considered to be one metre or less) falls have a significant risk of fatality.1,4 However, this is not supported by experience in routine clinical practice.
Since the introduction of the National Institute for Health and Clinical Excellence (NICE) guidance for management of head injury in children in the UK in 2003(updated in 200711 ,12), the use of cranial CT is more common in the UK, especially among those who have sustained low-level falls.2 This shift in clinical practice provides the opportunity to address the knowledge gap regarding the patterns of cranial imaging abnormality in young children, and in particular, to study the patterns of intracranial injury in those who sustain low-level falls in which the injury mechanism(s) is known.
In this multicentre retrospective study, we systematically review the cranial imaging findings in children of 0–2 years presenting with AHI to two major (level 1 and level 2) paediatric accident and emergency departments in the UK, and relate the imaging findings to clinical outcome.
Materials and methods
A retrospective review was undertaken of cranial CT performed for trauma infants aged 0–2 years at two large tertiary referral teaching hospital trusts in the East Midlands from January 2006 to August 2011. Indications for cranial CT scanning in children under 2 years of age were as per the NICE guidelines.11 ,12 A change in the NICE criteria adopting a more permissive approach to imaging was implemented in mid-2007; all but six patients in the cohort were scanned under the updated guidance.
Cases were identified from the radiology information system by two of the researchers (AT, SH) by manual searching of the database according to inclusion and exclusion criteria listed in box 1. The emergency department clinical information system (EDIS) was interrogated for details of presenting history. Clinical follow-up information was obtained from electronic discharge summaries and/or any follow-up neuroimaging investigations. Demographic details were recorded, along with details of injury mechanism, whether or not independent witnesses corroborated the history and the details of any further imaging investigation(s), such as skeletal survey or MRI of the brain and spine. Indications for subsequent MRI in cases of accidental injury were ongoing focal neurological signs or further evaluation of suspected brain parenchymal injuries demonstrated on CT.
Study inclusion and exclusion criteria
Age 0–2 years at the time of initial cranial CT scan
Clear history of accidental head injury, or if initial history insufficient to explain injury but subsequent MRI and/or skeletal survey negative for evidence of non-accidental injury and no persisting clinical concerns regarding non-accidental injury
Insufficient clinical history to determine a clear mechanism of injury and where initial or subsequent investigation proved compatible with, or where there was ongoing clinical suspicion of, non-accidental injury
All imaging was reviewed by two experienced paediatric neuroradiologists (TJ, RD) who were blinded to the clinical status of the child and the reported mechanism of injury. Imaging findings of skull fracture (location/type), scalp injury or abnormality, intracranial haemorrhage (location/type) and brain parenchymal injury or ischaemic changes were recorded.
Mechanism of injury was obtained from the EDIS and subsequent clinical notes. The reported height of fall was recorded. The distribution of heights was compared between those with and without cranial or intracranial injury using the non-parametric Kruskal–Wallis test. Statistical analysis was performed using SPSS V.19 (Chicago, Illinois, USA).
The study was granted permission by the local research ethics committee chairman after the requirement for formal ethical review was waived.
One hundred and fifty-six children in the 0–2 age group were identified as having had a cranial CT scan for head injury that fulfilled the NICE criteria within the study period (median age 12 months, range 2 days to 23 months, M:F=91:65). Of these, seven children were excluded from further analysis based on the exclusion criteria given in box 1. The final study population consisted of 149 children (median age 11 months, range 2 days to 23 months, M:F=86:63).
All children had cranial CT on admission. Eighteen subsequently underwent skeletal survey, with no other fractures or injuries identified. Ten children underwent follow-up MRI of brain and spine.
The distribution of injuries as demonstrated on CT is shown in table 1.
Mechanism of injury
The distribution of mechanism of injury is detailed in table 2.
Of the 122 reported falls, a direct reported estimate of height was documented in 106. In the remaining 16 cases, an estimate was generated based on a description of the object from which the child had fallen. The estimated reported height of falls ranged from 0.15 to 6 m (median 0.9 m, mean 1.2 m). There was a significant difference in distribution of heights of fall between those with fracture and/or intracranial haemorrhage and those without (mean heights 1.32 m (95% CI 1.03 to 1.60) vs 1.01 m (95% CI 0.79 to 1.23), p<0.0001).
There were three cases with radiologically evident brain parenchymal injury. In the first case, a 5-month-old infant fell approximately 1.5 m from a top bunk onto a hard surface resulting in an occipital skull fracture, posterior fossa subdural haemorrhage (SDH), patchy subarachnoid haemorrhage (SAH) and a cerebellar contusion. There was evidence of SDH in the lumbar spine without associated spinal cord or spinal bony injury on MRI. Following a conservative approach, there were no long-term traumatic complications or neurological sequelae on follow-up MRI and clinical examination at 12 months. In the second case, a 6-month-old was struck by a car and thrown approximately 9 m (30 ft). Initial cranial CT showed multifocal SAH, an occipital skull fracture and posterior fossa SDH. Immediate subsequent MRI showed multifocal brain ischaemic changes and small haemorrhages in keeping with diffuse axonal injury (DAI). Delayed follow-up imaging revealed widespread cerebral atrophy with maturing multifocal ischaemic change. There was clinical evidence of developmental delay and epilepsy. In the third case, a 23-month-old fell out of a first storey window—approximately 3 m in height—landing on a concrete surface. Initial cranial CT showed biparietal skull fractures with underlying SDH and a left parietal cerebral contusion with surrounding low attenuation swelling. MRI within 48 h of injury demonstrated focal left parietal parenchymal swelling and diffusion restriction, along with confirmation of bilateral SDH. Clinical follow-up revealed mild generalised epilepsy but no focal neurological deficit.
Of the 10 patients who underwent MRI of the spine, spinal haemorrhage (subdural haematoma) was seen in only one case—that associated with an extensive occipital fracture and posterior fossa SDH following a blunt impact injury. None of the patients with supratentorial haemorrhage showed evidence of intraspinal haemorrhage.
Reporting of clinical history
In all cases, the primary history was given by the parents of the child (mother, father or both parents together). In nine cases, corroboration was provided by another family member (usually grandparents). In two cases, the injury occurred at a nursery/playgroup where the history was confirmed by staff.
Loss of consciousness was reported fairly consistently between groups (approximately 10%–15%—see table 1). Loss of consciousness followed by a lucid interval and subsequent deterioration was only encountered in the one case of a child with an extradural haemorrhage. The lowest Glasgow Coma Scale (GCS) at the time of scanning was 13 in the group with only scalp swelling or isolated skull fracture/thin SDH. A GCS of 10 or less was only encountered in patients with radiologically evident brain parenchymal injuries, extra-axial haemorrhage with mass effect or post-traumatic seizures.
Head injuries in neonates and infants are common whereas serious injury and neurological sequelae are rare. Warrington et al found that the risk of serious injury in falls under the age of 6 months is less than 1% (21 patients in a study population of 3357).2 More recently, the risk of fatal injury from a short fall has been estimated at being less than one in a million.3 The factors that influence the likelihood and distribution of intracranial injury in children include the height of fall, point of contact with head, nature of the surface of contact and velocity at which the surface is contacted.13 A short or focal point of contact allows a very short period of force with subsequent reduced time of brain deformation and limited potential for developing DAI; more focal brain parenchymal injury relating to the site of fracture may be encountered.14 The dura in infants is tightly adherent to the inner table, explaining the relative rarity of extradural haemorrhage.14 ,15 Diffuse SDH can result from inertial forces alone as occurs with shaking injuries; however, localised SDHs are more commonly associated with impact injury in relation to the site of fracture or impact. When localised, associated brain swelling or death is very rare.16 These pathological and anatomical findings can explain the findings of localised SDH beneath fractures in our study, where AHI was most commonly represented by short falls. Extradural haemorrhage was rare (only one case) and skull fractures, while relatively common, were uncommonly associated with DAI or parenchymal haemorrhage.
Studies have tried to characterise the radiological findings associated with NAHI versus AHI.17 A single focus of intracranial haemorrhage/skull fracture was shown to be more common in AHI in a population of sixty 0–6-year-olds,18 while skull fractures were less frequently encountered in NAHI than in AHI sustained from falls of less than 4 feet/1.2 m in height.19 The appearance of SDH on CT is now known to be unreliable as an indicator of age of haemorrhage.20–22 Diffuse hypoxic-ischaemic injury (HII) and multifocal SDH are commonly seen in cases of NAHI.23–25 In our case series, nearly all SDHs were seen as a thin uniformly hyperdense rim immediately beneath the fracture site (figure 1). Posterior interhemispheric SDH can be seen in cases of accidental trauma (figure 2A,B). Posterior fossa SDH was only seen in association with occipital fracture in two cases, following direct trauma without evidence of distant injury (figure 2D,E).
The height necessary to inflict a serious or even fatal head injury in neonates and infants has been the topic of marked debate in the literature. The contention that low-level falls can prove fatal is commonly ascribed to a study by Chadwick et al,2,6 which reported eight fatal falls, of which seven were from heights of less than 4 feet/1.2 m. However, in all cases, the history provided was inconsistent and did not account for multiple other injuries also reported.1,3 In addition, the study reported the counter-intuitive conclusion that the mortality rate from falls of heights from 4 to 10 feet/1.2–3 m was actually greater than for those from 10 to 45 feet/1.2–14 m. A case series by Plunkett, which did not include any child below 12 months of age, has been criticised for drawing conclusions regarding injury to infants based on cases about older children and for the limitations of postmortem examinations.6 ,27 In addition, the patterns of injuries reported do not replicate the findings most commonly encountered in cases of NAHI. The case of a child dying after evacuation of an SDH, videotaped falling onto her head from a jungle gym, is often cited as the only case of documented death from a short fall in the literature.28 Many other series have not replicated the findings of fatal injuries from short falls, including those where the falls occurred within institutions outwith the care of parents/caregivers.29–34
The range of heights from which infants can fall without sustaining radiologically apparent injury (excepting scalp swelling) was broad in our study, extending from 0.15 to 6 m. The lowest reported fall height to cause radiologically apparent injury (skull fracture) in our series is 50 cm (approximately 2 feet). In all five such cases, the only injury sustained was an undisplaced skull fracture without any associated intracranial haemorrhage. The reported mechanisms were falls onto hard surfaces (wooden floor in two cases, linoleum in one case), fall onto rock and fall from being held after a clash of heads. While in our study there was a significant difference between the mean fall height in those with intracranial injury/skull fracture and those without, a broad overlap exists between the two groups. Higher-level falls (range 1.5–6 m in our study) also resulted in a broad range of outcomes, from no injury/scalp swelling to intracranial haemorrhage and brain parenchymal injury in the case of a fall from a height of 6 m (20 ft) onto a concrete surface. This resulted in only minor neurological sequela (mild epilepsy). The only case in our series that reflects the widespread HII seen in cases of NAHI is the 6-month-old infant thrown 9 m (30 ft) after being hit by a vehicle. The forces generated by this mechanism of injury are clearly considerable with a marked acceleration/deceleration component. The subsequent neurological deficits experienced by this infant mirror those commonly observed in cases of NAHI.
A frequent criticism made in cases of unexplained injury in infancy is that there are no studies that involve cross-sectional imaging in all cases of accidental domestic head injury presenting to hospital. As far as we are aware, this is the first study to do so, however trivial the injury mechanism, but with sufficient clinical features to meet the criteria for cranial CT according to UK guidelines. The results provide an insight into the consequence of such injuries in early life and confirm the findings of earlier studies that indicate the relatively benign outcome from low-level falls. Clearly not all cases of low-level falls/domestic head injuries will present to the medical authorities. However, it is reasonable to assume that the incidence and pattern of any intracranial injury in these cases will not differ significantly from those in whom the caregiver is sufficiently concerned about the injury to seek a medical opinion.
Our study can be criticised for its retrospective design and the reliance of mechanism of injury from often uncorroborated direct caregivers. Indeed the accuracy with which caregivers can estimate the height of falls has been shown to be poor.35 We have tried to include only those cases where either there was a verifiable, corroborated description of the injury mechanism provided by an independent witness or a given mechanism of injury that was considered to be clinically appropriate at presentation and subsequently. This approach reflects the day-to-day experience of radiologists and paediatricians dealing with such clinical workload. The low incidence of radiologically evident intracranial injury and tendency for any intracranial injury (mainly SDH) in our study cohort to be localised to the site of impact suggests that the translational forces associated with low-level/velocity impact trauma are well tolerated in early life. In contrast, in cases of unequivocal, verified NAHI or cases of infants admitted to hospital with unexplained intracranial injury, the more diffuse pattern of intracranial haemorrhage and ischaemic brain damage may indicate that these children were subjected to a different injury mechanism(s) involving inertial forces that result in a more global distribution of intracranial damage.
In the vast majority of cases, AHI in neonates and infants either results in no—or only minor—intracranial injury. While relatively low-level falls can result in skull fracture, in our study, SDH, when seen was usually thin and directly related to the fracture site. In our series, radiologically evident parenchymal or hypoxic-ischaemic injuries approaching those seen in NAHI were very rare and only occurred when associated with significant forces reported by mechanism of injury. The vast majority of cases in our study suffered no long-term documented clinical neurological sequela of injury.
The authors wish to acknowledge Dr Kate Rankin, Emergency Department, Nottingham University Hospitals NHS Trust and Dr Damien Roland, Emergency Department, Leicester Royal Infirmary for help in collating clinical data and radiographic staff at both institutions involved in the study.
Contributors AGT: Data collection and analysis, main manuscript author. SH: Data collection and analysis, manuscript editing. RAD: Image analysis, manuscript editing. TJ: Study design, image analysis, manuscript editing.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.