Child abuse: Evaluation and diagnosis of abusive head trauma in infants and children

INTRODUCTION — The evaluation and diagnosis of abusive head injury in infants and children will be presented here. The epidemiology, mechanisms, and ophthalmologic aspects of abusive head injury in children; management of suspected child abuse; and the initial evaluation of severe traumatic brain injury in children are reviewed separately. (See "Child abuse: Epidemiology, mechanisms, and types of abusive head trauma in infants and children" and "Child abuse: Eye findings in children with abusive head trauma (AHT)" and "Severe traumatic brain injury (TBI) in children: Initial evaluation and management" and "Physical child abuse: Diagnostic evaluation and management".)

BACKGROUND — More than 45 percent of deaths from child abuse occur among children younger than 12 months of age [1]. Abusive head injury is the most common cause of death and long-term disability resulting from physical child abuse. Infants frequently present with nonspecific clinical features without a history of trauma. As a result, as many as 30 percent of children with abusive head injury may be misdiagnosed at the initial evaluation [2,3].

Identification of abusive head injury can be lifesaving [4]. In one chart review describing missed cases of abusive head injury, four of five deaths might have been prevented if the inflicted mechanism had been recognized during previous evaluations for symptoms related to head injury [2].

Cranial injury may be inflicted by blunt force trauma, shaking, or a combination of these and other forces. The constellation of injuries associated with this mechanism has been referred to as "shaken baby syndrome" (SBS), "infant whiplash syndrome," "shaken/impact syndrome," or more simply as "inflicted or abusive head trauma" [5]. For many young infants, crying may be a trigger for the shaking episode, suggesting that the caretakers' response to prolonged crying may be an effective target for prevention strategies [6-8]. (See 'Prevention' below.)

The significance of the diagnosis (with regards to morbidity, child protection, and criminal prosecution) may be similar, regardless of whether injury resulted from direct blows or from shaking [4]. In addition, much of the evidence to support the mechanism of injuries that result from shaking is retrospective and indirect (see "Child abuse: Epidemiology, mechanisms, and types of abusive head trauma in infants and children", section on 'Mechanisms of injury'). For this reason, the term "abusive head injury" is preferred to "shaken baby syndrome" to permit consideration of multiple mechanisms of injury in any child. The generic term "head injury" also serves to distinguish the diagnosis of injury from the investigation of how the injury occurred [4,9,10].

CLINICAL FEATURES — The clinical manifestations of abusive head injury are often nonspecific [4]. Furthermore, a history of trauma usually is not forthcoming, and external signs of injury (such as bruising of the skin) may be minimal or absent.

History — The symptoms for which caregivers most often seek medical attention for children with abusive head injury include seizures, breathing difficulty, apnea, and lifelessness [11-17]. In comparison, in a population-based series describing the clinical features of children with inflicted and unintentional head injury, reasons for seeking care for those with unintentional head injury included local swelling following injury, parental concern for children who were asymptomatic, and lethargy [12].

No history of a traumatic event was offered in 64 to 97 percent of cases of abusive head injury described in several retrospective series [12,13,18,19]. In one of these reports, the absence of a history of trauma had a high specificity (0.97 [95% CI 0.83-1.0]) and high positive predictive value (0.92 [95% CI 0.78-0.98]) for abusive head injury [13]. Other historical features that are typical include attributing the injuries to resuscitative efforts and a changing or developmentally incompatible history [12,13].

Accurately determining when an injury occurred can be difficult because the traumatic event is infrequently disclosed, and initial symptoms may be nonspecific. However, a retrospective review describing cases where perpetrators admitted to abuse noted that 91 percent of children developed symptoms immediately after the injury [20]. Perpetrators who have confessed to shaking infants often report multiple episodes of shaking over weeks to months prior to presentation, which may also influence determination of injury timing [8]. In one small study of 59 children with severe, acute head injury, infants with abusive head trauma (AHT) were more likely to have a delay in seeking care of longer than 30 minutes based on caregiver history despite the presence of severe symptoms [21].

Finally, a history of previous abuse and/or investigation for maltreatment is common among children who have sustained abusive head injury. Retrospective series have reported that over 60 percent of victims of abusive head injury may have a history and/or clinical evidence of previous abuse [11,22,23].

Physical examination — Children with abusive head injury may have physical findings ranging from a normal neurologic examination, to nonspecific signs related to brain injury (such as vomiting and altered mental status), to coma and profoundly unstable vital signs (apnea, bradycardia) requiring resuscitation [4]. This is illustrated in the following reports:

●Two retrospective series reported that occult head injury was diagnosed on screening neuroimaging studies in 37 and 29 percent of children being evaluated for child abuse who had no neurologic symptoms [24,25].

●In a retrospective series describing infants who were evaluated for complaints that were not initially recognized as the result of AHT, 65 percent were described as irritable [2]. An abnormal mental status was reported for 35 percent of children, and 56 percent were vomiting.

●A retrospective report noted that 39 percent of children with abusive head injury required cardiopulmonary resuscitation at initial presentation [12].

The majority of children with abusive head injury initially have an abnormal neurologic examination [11-13]. Many children have no external signs of injury (35 and 40 percent in two reports) [11,12].

Retinal hemorrhages — Retinal hemorrhages are frequently noted in children with abusive head injury (60 to 85 percent in retrospective series) [4,11,12,17,19,26]. The extent of retinal hemorrhages is generally correlated with the severity of intracranial injury and neurologic impairment [26,27]. Among abused, neurologically normal children whose intracranial injuries were diagnosed with screening neuroimaging, retinal hemorrhages were uncommon [24,25].

Although retinal hemorrhages have been reported in other conditions, those associated with abusive head injury are characteristically numerous, involve multiple layers of the retina, and extend beyond the posterior pole to the peripheral retina (picture 1) [28,29] (see 'Differential diagnosis' below). Injury to the eye as the result of abusive head injury is discussed separately. (See "Child abuse: Eye findings in children with abusive head trauma (AHT)".)

Cutaneous bruising — Children with abusive head injury may have cutaneous bruising, particularly around the face [14,30]. Suspicious bruising patterns such as linear marks or bruises involving the ear, neck, or torso in infants and young children also raise concern for inflicted injury among children with a nonspecific presentation. (See "Physical child abuse: Recognition", section on 'Inflicted bruises'.)

However, in one retrospective series describing children with abusive head injury, 54 percent had no bruising noted at initial presentation [11]. Bruising in nonambulatory infants is rare, and any bruise to a young infant should raise concern for either trauma or a bleeding diathesis [31,32].

Associated injuries — Children with abusive head injury frequently have injuries to other organ systems that are typically associated with abuse [17]. In two retrospective series, 22 and 23 percent of children with abusive head injury also had rib fractures [11,12]. Long bone fractures were reported in 23 and 29 percent of cases. In one of the reports, 4 percent of children sustained an abdominal injury [11]. Cervical spinal injury may be more common in children with fatal injuries [33,34] and may be more common than reported in survivors of AHT. (See 'Spine imaging' below.)

EVALUATION — The diagnosis of abusive head injury requires a high index of suspicion. The signs and symptoms may be mild and nonspecific (eg, vomiting, poor feeding, irritability, or lethargy) or severe and even life threatening (eg, apnea or other breathing abnormalities, coma, or seizures) [2,14,35]. Although children who present with serious head injury can usually be readily identified, subtle injuries may be missed. Once the diagnosis is suspected, a careful evaluation must be performed to identify supporting clinical features and exclude other diagnoses.

Laboratory studies — The utility of laboratory studies depends upon the clinical presentation:

●Ill-appearing – Children with abusive head injury are frequently ill appearing with nonspecific clinical features. There is often no history of trauma. The following laboratory studies may provide specific diagnostic information:

•Complete blood count, platelet count, and selected coagulation studies based upon patient presentation (algorithm 1) may indicate anemia caused by intracranial and/or intra-abdominal hemorrhage [36-38] or an underlying bleeding disorder, although coagulopathy (as evidenced by decreased hematocrit, lower platelet count, and mildly prolonged prothrombin time) has been reported in children with parenchymal brain damage as a result of inflicted injury [39]. The probability that a child would have a specific coagulopathy causing intracranial bleeding by condition is shown in the table (table 1). The evaluation for bleeding disorders when child abuse is suspected is discussed in detail separately. (See "Physical child abuse: Diagnostic evaluation and management", section on 'Evaluation for bleeding disorders'.)

•Electrolytes, liver and pancreatic enzyme testing, and a urinalysis are recommended to identify potential injuries to other organ systems and metabolic abnormalities. (See "Physical child abuse: Diagnostic evaluation and management", section on 'Laboratory studies'.)

•When meningitis is a serious consideration, cerebrospinal fluid (CSF) should generally be obtained if lumbar puncture can be safely performed. Xanthochromia and/or red blood cells in CSF may indicate intracranial injury. (See 'Differential diagnosis' below.)

●Well-appearing with concerning findings – Asymptomatic infants with red flags for physical child abuse (table 2 and table 3) also warrant laboratory evaluation as described above.

●Screening for acute intracranial hemorrhage – Among well-appearing infants with symptoms (eg, vomiting or fussiness) that place them at increased risk for abusive head trauma (AHT), biomarkers are being sought to identify those patients with a higher likelihood of acute intracranial hemorrhage who may warrant additional laboratory evaluation, skeletal survey, and neuroimaging [40-43]. As an example, multicenter, prospective validation of a point-of-care test that measured matrix metallopeptidase-p, neuron-specific enolase, vascular cellular adhesion molecule-1, and hemoglobin in the serum of 599 patients (12 percent with acute intracranial hemorrhage) had a sensitivity of 89 percent and a specificity of 48 percent at a specified cutoff identified during derivation [43]. However, measurement of these biomarkers is not widely available, and further study is needed before such testing is part of standard clinical practice.

Imaging — A skeletal survey (table 4 and table 5 and table 6) and neuroimaging are essential components of the evaluation for abusive head trauma.

Skeletal evaluation — We recommend that all infants and young toddlers with suspected abusive head injury receive a skeletal survey (table 6 and table 5) [44].

Children with abusive head injury often sustain other injuries (see 'Associated injuries' above). In addition, identifying injuries that are highly suggestive of child abuse (such as posterior rib fractures and metaphyseal fractures) substantiates the inflicted mechanism for the head injury. (See "Orthopedic aspects of child abuse", section on 'Fracture patterns' and "Orthopedic aspects of child abuse", section on 'Radiographic evaluation'.)

A second skeletal survey performed two to three weeks after the initial study sometimes reveals fractures that were not visualized radiographically at the time of initial hospitalization and is recommended for children with suspected AHT. Skeletal surveys should be interpreted by a pediatric radiologist.

Neuroimaging — In addition to establishing the diagnosis of abusive head trauma, neuroimaging can provide documentation of the extent and some estimation on the timing of injuries. The choice of computed tomography (CT) versus magnetic resonance imaging (MRI) is determined by whether the patient has signs or symptoms of acute head injury and, in asymptomatic patients, whether MRI can be rapidly performed and interpreted by a pediatric radiologist [17,45,46].

Neuroimaging is also indicated in the evaluation of children who are at high risk for child abuse but have a normal neurologic examination and no clinical evidence of head injury. (See "Physical child abuse: Diagnostic evaluation and management", section on 'Intracranial injury'.)

Unstable patients — For hemodynamically or neurologically unstable patients, we recommend unenhanced CT with brain and bone windows as the preferred imaging modality for the initial evaluation of children once the initial trauma resuscitation has occurred [44,47-50]. Whenever possible, three-dimensional skull reconstruction should be added to enhance the detection of skull fractures.

CT is readily available and reliably identifies injuries that require immediate intervention. Subarachnoid hemorrhage, large extra-axial hemorrhage (eg, subdural or epidural hemorrhage), and mass effect are best demonstrated on CT. Other injuries that are frequently identified include cerebral ischemia, cerebral contusion, skull fracture, and scalp hematoma. However, early signs of brain edema or diffuse axonal injury may not be apparent [51-53].

Stable patients — For hemodynamically and neurologically stable patients, we suggest MRI rather than CT for the initial neuroimaging study if MRI can be performed in a timely manner and interpreted promptly by a pediatric neuroradiologist. Fast MRI (shorter, more motion-tolerant MRI sequences designed to be performed without sedation) have been used in some centers as a first-line study to avoid the radiation exposure associated with CT [45,46,54]. For stable children who are undergoing admission for further evaluation, MRI can be deferred to the next day for scheduling purposes if it is not immediately available.

Noncontrast MRI is superior to CT for documenting the pattern, extent, and timing of head injuries and, where available, is preferred to head CT [45,46,55-57]. T1, T2, gradient recalled echo (GRE), susceptibility-weighted, and fluid-attenuated inversion recovery (FLAIR) sequences must be performed in order to characterize the nature of hemorrhages and other fluid collections [46]. MRI may also demonstrate small extra-axial hemorrhages, parenchymal contusions, evidence of shearing injury, and posterior fossa abnormalities missed by CT [15,45-47,52,53]. Diffusion-weighted images (DWI) and apparent diffusion coefficient (ADC) mapping help to identify parenchymal injury in the acutely injured child [58].

MRI has high sensitivity compared with CT for identifying brain injury but has the potential for missing skull fractures and subarachnoid hemorrhage [46,59]. However, MRI may be more sensitive for identifying subdural hematomas than CT at initial presentation. Two observational studies have demonstrated that MRI can be performed rapidly and without sedation in the vast majority of infants and young children [45,46].

Brain MRI should also be done as a follow-up study for abused infants and children who have abnormal initial CT scans, but it is relatively insensitive for the identification of subarachnoid blood and fractures. Sensitivity for restricted diffusion injury increases over the first few days after injury [60]. Consequently, brain MRI is typically obtained two to three days after the injury [47]. Because of the high prevalence of cervical spine injuries in children with AHT, we recommend cervical spine with brain MRI in all victims. (See 'Spine imaging' below.)

Findings associated with abuse — The diagnosis of AHT requires consideration of all clinical findings. While certain patterns of injury on neuroimaging are more consistent with abuse, any finding should be evaluated based upon the history, reported mechanism of injury, and within the context of other injuries to the child. All imaging findings should be verified by a pediatric radiologist. Consultation with a multidisciplinary child abuse team is strongly encouraged to aid in diagnosis and management of children with suspected AHT.

Neuroimaging findings associated with AHT was evaluated in a systematic review of 21 observational studies that analyzed CT findings in 2353 young children hospitalized with closed head injury, including 38 percent with AHT [61]. The following CT findings were significantly associated with AHT compared with unintentional head trauma:

●Subdural hemorrhage (68 versus 23 percent, respectively); additional characteristics of subdural hemorrhage significantly associated with AHT included multiple hemorrhages with differing density, interhemispheric or posterior fossa location, and extension over the convexity of the brain

●Hypoxic-ischemic injury (24 versus 8 percent, respectively)

●Cerebral edema (39 versus 18 percent, respectively)

By contrast, the frequency of subarachnoid hemorrhages was not significantly different following AHT compared with unintentional head trauma. Epidural hemorrhages were significantly associated with unintentional closed head injury rather than AHT (17 versus 4 percent, respectively).

Subcortical abnormalities on head CT may also be a strong marker for child abuse. In one prospective, multicenter observational study of 54 children under three years of age not included in the systematic review described above, a head CT with findings at the subcortical level, including swelling, hypoxic-ischemic changes, brain shift, and hernia was highly associated with abusive head injury (odds ratio [OR] 36; 95% CI 6-209) [62]. Patients with subcortical changes were also significantly more likely to acutely present with seizures, acute encephalopathy, signs of herniation, respiratory distress, or circulatory compromise.

Spine imaging — Given the reported high rate of cervical injuries in children with AHT, we recommend cervical spine MRI concurrent with brain MRI for all victims.

Injury to the cervical spine as the result of abusive head injury is uncommonly recognized clinically but may be more common than reported [63-65]. As an example, in a retrospective study that documented the results of spinal MRI in 78 children sustaining abusive injury, 46 children with accidental injury, and 70 children with nontraumatic conditions, evidence of cervical spine ligamentous injury was found in 78 percent of abused patients compared with 46 percent of those with accidental injury and 1 percent of those with nontraumatic conditions [66]. Spinal MRI of the cervical spine is the best method for detecting these injuries [67].

Additionally, in a study of 47 children diagnosed with AHT, almost two-thirds had a spinal injury; injuries to the thoracic and lumbar spine were documented in addition to the cervical spine even though only the cervical spine was imaged in over 90 percent of patients [68]. The clinical and forensic significance of these injuries has not yet been well defined. While some injuries may suggest a shaking or whiplash mechanism, most do not change clinical management. Based upon this limited evidence, some experts routinely perform whole-spine MRI at the time of brain MRI in children with AHT. Pending further study, we obtain whole spine MRI in selected patients as indicated by their symptoms and severity of brain and cervical spine injury.

Ophthalmologic examination — Children with suspected abusive head injury should have a funduscopic examination, preferably by an ophthalmologist, to identify retinal hemorrhages and other eye injuries. Non-ophthalmologists may have difficulty performing an adequate examination and thereby fail to identify injuries that, although not pathognomonic, suggest inflicted injury [69,70]. (See "Child abuse: Eye findings in children with abusive head trauma (AHT)".)

Wide-field retinal imaging permits photographic documentation of the extent of any retinal hemorrhages present and may be useful for clinical and research purposes but, when performed, should be interpreted by an ophthalmologist [71].

DIAGNOSIS — Establishing an abusive mechanism for any injury is essential for the protection of the child and any other individuals who could be exposed to similar injury. Making the correct diagnosis may be lifesaving. On the other hand, the consequences of misdiagnosis of inflicted injury are serious, including disruption of a family (as the result of placement of children in foster care) and false accusations that could lead to criminal proceedings.

A thorough evaluation by a multidisciplinary team (social worker, nurse, child abuse specialist) should be performed whenever there is a suspicion for physical child abuse. Additional information often can be identified to support or refute the diagnosis. A reasonable medical suspicion that an injury was inflicted is sufficient to make a preliminary diagnosis of abusive head injury and to refer the case to social service and law enforcement agencies for further investigation. (See "Child abuse: Social and medicolegal issues", section on 'Reporting suspected abuse'.)

During the initial evaluation of a child with suspected abusive head injury, the harm of repeated injury caused by missing an abusive injury is more significant than the harm of false accusation. The clinician must maintain a low threshold for suspecting abusive injury and reporting a suspicion to Child Protective Services. The purpose of subsequent investigation by social service and law enforcement agencies is to identify evidence that substantiates or refutes the medical concerns of abuse.

Children with intracranial injury — Evidence suggests that although no single physical finding definitively identifies abusive head trauma (AHT), combinations of clinical features, as noted below, can accurately identify patients with intracranial injury who have a high likelihood of AHT and who warrant a thorough evaluation for child abuse [4,5,72]:

●History and physical findings – In a meta-analysis of high quality studies, each of the clinical findings listed below were strongly associated with AHT [73]:

•Inadequate history (eg, no history of trauma or a history that is inconsistent with the severity of physical findings)

•Apnea or seizures on presentation

•Fractures of the ribs, metaphyseal region, or long bones

•Retinal hemorrhage(s)

•Subdural hemorrhage

•Skull fracture with associated intracranial injury

•Cerebral ischemia on neuroimaging

●Specific injuries – Using only the injuries identified and no other components of the history, an analysis of six studies with a total of 1053 children with intracranial injury estimated the probability for AHT when different combinations of clinical features were observed [74]. The probability for AHT increased with specific findings and when more features were present as follows:

•Intracranial injury alone – Lower probability.

•Intracranial injury and long bone fracture – Moderate probability.

•Intracranial injury and retinal hemorrhage – Higher probability.

•Intracranial injury and rib fracture – Higher probability.

•Intracranial injury and any three of the following: apnea, bruising, long bone fracture, retinal hemorrhage, rib fracture, or seizure – Higher probability; in a subsequent retrospective validation study in a cohort of 198 children (65 with confirmed AHT), the presence of three or more of these features had a sensitivity of 72 percent and a specificity of 86 percent for AHT [75].

●Other findings – Individual observational studies also suggest that AHT is more likely in children with the combination of intracranial injury on neuroimaging and one or more of the following clinical features:

•Past history of child abuse or neglect in the patient or a sibling [76]

•Intra-abdominal injury, especially hollow viscus injury or pancreatic laceration [77]

•Oropharyngeal injuries such as frenulum tears or pharyngeal lacerations [13]

•Additional extracranial injuries that are not explained by the history provided (eg, head and neck bruising or other cutaneous findings) (see "Physical child abuse: Recognition", section on 'Red flag physical findings')

In prospective multicenter validation studies, a previously derived screening rule consisting of any one of the following findings also had high sensitivity but low specificity for child abuse among children three years of age and younger who were admitted for acute symptomatic closed head injury [78,79]:

•Any clinically significant respiratory compromise prior to pediatric intensive care unit (PICU) admission

•Any bruising of the child's ears, neck, or torso

•Any bilateral or interhemispheric subdural hemorrhages

•Any skull fracture other than an isolated, unilateral, nondiastatic, linear, parietal skull fracture

Well-appearing infants — Detection of AHT is challenging in well-appearing infants who typically present with an unrelated complaint and no history of trauma. Infants with injuries that suggest intentional injury (eg, suspicious bruising, burns, oral injury, or fracture) warrant neuroimaging to identify AHT in addition to further evaluation to assess for associated extracranial injuries. (See 'Evaluation' above and "Physical child abuse: Diagnostic evaluation and management", section on 'Evaluation'.)

The following chief complaints should also prompt the clinician to consider AHT in the differential diagnosis [4]:

●Apnea or acute life-threatening event

●Seizure

●Vomiting without diarrhea

●Soft tissue scalp swelling

●Bruising

●Nonspecific neurologic symptoms (eg, lethargy, fussiness, or poor feeding)

Additional findings that suggest trauma as the underlying cause warrant further evaluation with neuroimaging in such patients. As an example, in a prospective, multicenter validation of a scoring system in over 1000 well-appearing infants younger than one year of age (109 with abuse) who presented with the above complaints and had measurement of head circumference and hemoglobin, a score of 2 using the following clinical prediction rule had a sensitivity of 93 percent and a specificity of 53 percent for an abnormality on CT of which 91 percent were traumatic (eg, skull fracture, intracranial hemorrhage, cerebral edema) and 9 percent were atraumatic (eg, hydrocephalus, brain tumors, stroke, or arteriovenous malformation) [38,80]:

●Abnormality on skin examination (bruising or soft tissue swelling) (2 points)

●Three months of age or older (1 point)

●Head circumference >85^th percentile (1 point)

●Serum hemoglobin <11.2 g/dL (1 point)

MRI is preferred to CT of the head in such patients if there is timely availability of the study and interpretation by a pediatric neuroradiologist. (See 'Imaging' above.)

DIFFERENTIAL DIAGNOSIS

Accidental injury — Although many of the clinical features of accidental and abusive head injury are similar, the abusive mechanism is often suggested by a combination of historical features and specific injury patterns [81]:

●A history of a traumatic event has been consistently present among children with accidental head injury in retrospective series that have compared children with inflicted and accidental head injury [12,19,82]. Furthermore, the absence of a history of trauma is highly specific for abusive head injury [13]. (See 'History' above.)

●Retinal hemorrhages have been observed in some children with head injuries from accidental mechanisms [12,19]. The hemorrhages are typically fewer in number and less extensive than those reported with inflicted injuries [28,29]. (See "Child abuse: Eye findings in children with abusive head trauma (AHT)", section on 'Differential diagnosis'.)

●Subdural hematomas occur in association with accidental injury. However, they are more common with abusive mechanisms [61]. In addition, mixed-density subdural hematomas are noted more frequently with inflicted injuries. (See "Child abuse: Epidemiology, mechanisms, and types of abusive head trauma in infants and children", section on 'Subdural hemorrhage' and 'Unstable patients' above.)

Birth trauma — Retinal hemorrhages and subdural hematomas have been noted in asymptomatic newborns. Both occur more commonly with instrumented deliveries [83-85]. These subdural hematomas are often small, located in the posterior fossa or at the tentorium, and are usually resolved by one month of age. (See "Child abuse: Eye findings in children with abusive head trauma (AHT)", section on 'Differential diagnosis' and "Child abuse: Epidemiology, mechanisms, and types of abusive head trauma in infants and children", section on 'Subdural hemorrhage'.)

Among symptomatic infants with subdural hematomas, those with a perinatal etiology often have a history of birth trauma and are younger than those with abusive head injury [86].

Apparent life-threatening event — In two prospective series describing infants who had apparent life-threatening events (ALTEs), 2 percent were diagnosed with an abusive head injury as the cause of the ALTE [35,87]. Children with ALTEs as the result of inflicted head injuries must be urgently identified because injury often occurs repeatedly and can be fatal. (See "Acute events in infancy including brief resolved unexplained event (BRUE)", section on 'Child abuse'.)

Based upon one retrospective observational study, abused children with subdural hematomas who presented with an ALTE are more likely to have extracranial findings of child abuse, including retinoschisis, high-specificity bruising, and internal abdominal injury as well as higher mortality than abused children with subdural hematomas and no ALTE [88]. This study supports the conclusion that findings of abusive head trauma (AHT) are not caused by ALTEs.

Bleeding disorders — Intracranial hemorrhage can occur in children with severe bleeding disorders (such as hemophilia) spontaneously or following an injury. The probability that a child would have a specific coagulopathy causing intracranial bleeding by condition is shown in the table (table 1) [89-91].

Retinal hemorrhages in children with bleeding disorders are usually small in number and are typically confined to the posterior pole, although coagulopathy can lead to more severe retinal bleeding [92,93]. Laboratory abnormalities pinpoint the nature of the bleeding problem (table 7). (See "Approach to the child with bleeding symptoms", section on 'Initial laboratory evaluation' and "Child abuse: Eye findings in children with abusive head trauma (AHT)", section on 'Differential diagnosis'.)

Among boys with hemophilia, intracranial hemorrhage occurs most often in the neonatal period [94]. Intracranial hemorrhage is uncommon in immune thrombocytopenia. (See "Clinical manifestations and diagnosis of hemophilia", section on 'Intracranial bleeding' and "Immune thrombocytopenia (ITP) in children: Clinical features and diagnosis", section on 'Clinical features'.)

Other diagnoses — Glutaric aciduria type 1 is a metabolic condition that shares some clinical features with abusive head injury. Case reports have described children who present with acute subdural hemorrhage or chronic subdural effusions that were mistakenly attributed to child abuse. Glutaric aciduria is diagnosed through urinary and laboratory screening; treatment can prevent further deterioration. (See "Organic acidemias: An overview and specific defects", section on 'Glutaric acidemia type 1'.)

Children with osteogenesis imperfecta have rarely presented with subdural hematomas either spontaneously or as a result of minor trauma in association with fractures and retinal hemorrhages. (See "Differential diagnosis of the orthopedic manifestations of child abuse", section on 'Osteogenesis imperfecta'.)

Children with Menkes disease have presented with subdural collections, acute neurological deterioration, retinal hemorrhage, metaphyseal end plate changes resembling the classic metaphyseal fracture, and other fractures, causing potential confusion with AHT. Cerebral atrophy or excess wormian bones may be a clue, and identification of pili torti under the microscope is virtually diagnostic. Laboratory diagnosis is made by identifying decreased serum copper and ceruloplasmin and confirmed by identifying mutations of the ATP7A gene. (See "Differential diagnosis of the orthopedic manifestations of child abuse", section on 'Copper deficiency'.)

Other conditions that can cause intracranial hemorrhage and/or retinal hemorrhages such as meningitis, cerebral malaria, galactosemia, hypertension, intracranial aneurysms, and hemophagocytic lymphohistiocytosis are rare and can be identified by laboratory and radiologic testing.

MANAGEMENT — The management of infants and children with suspected abusive head trauma (AHT) is similar to that of other patients with suspected inflicted injury. The medical issues are addressed first, followed by additional evaluation for occult injuries or injuries inflicted at an earlier time.

The emergency management of severe traumatic brain injury and intracranial hypertension are discussed in detail separately. (See "Severe traumatic brain injury (TBI) in children: Initial evaluation and management" and "Elevated intracranial pressure (ICP) in children: Management", section on 'Treatment of elevated ICP'.)

Consultation with a multidisciplinary team (eg, social worker, nurse, child abuse specialist) is recommended, if such a team is available [4]. The safety of other children in the home must be ensured by local Child Protective Services, and evidence suggests that other children in the household require medical evaluation, as they are often victims of unreported abuse. (See "Physical child abuse: Diagnostic evaluation and management", section on 'Identifying injuries in other children at risk'.)

Reporting cases of suspected child abuse to Child Protective Services is discussed in detail separately. (See "Child abuse: Social and medicolegal issues", section on 'Reporting suspected abuse'.)

OUTCOME — Mortality rates in series of cases of abusive head injury range from 13 to 23 percent [14,95-97]. In a survey of cases identified between 1994 and 2000 from a traumatic brain injury surveillance system in Colorado, the mortality rate for children younger than three years of age with intentional traumatic brain injury was 17 percent [98]. Several clinical features are correlated with fatal child abuse. As an example, in an observational study of 386 children managed with abusive head trauma (AHT) at multiple children's hospitals, fatal outcomes were significantly associated by adjusted analysis with an initial Glasgow coma score of 3, 4, or 5; retinal hemorrhage; intraparenchymal hemorrhage; or cerebral edema [97].

Among survivors, the initial morbidity from abusive head injury is significant. In one retrospective chart review of 364 cases, death occurred in 19 percent as a direct result of the head injury; among survivors, 55 percent had persistent neurologic deficit and 65 percent had visual impairment [11]. Only 22 percent of the survivors were felt to be "well" at the time of discharge, without identified health or developmental impairment.

Follow-up studies of children with abusive head injury suggest that, although some patients may initially improve, long-term neurocognitive functioning is often poor, and there is a significant risk of long-term disability [51,99-101]. This is illustrated in the following reports:

●Among a group of 13 patients with nonaccidental head injury who were followed over a period of 4 to 14 years (mean seven years, two months), six children initially appeared to have a full recovery [99]. By the end of the study, 11 of the 12 survivors had been diagnosed with abnormalities ranging from psychomotor delay, intellectual disability, and learning disabilities to blindness, seizures, tetraplegia, and hemiparesis.

●In another report, 25 children with abusive head injury were followed for a mean of 59 months with neurologic, behavioral, and social assessments [100]. Among these children, 68 percent had abnormalities noted on follow-up, 36 percent of which were classified as severe. Identified abnormalities included motor (60 percent), visual (48 percent), and speech and language impairment (64 percent); and epilepsy (20 percent). Behavioral disturbances were noted in 52 percent of children. Many of the behavioral problems developed in the second and third year of life. Outcomes correlated with severity of injury but not with age at injury or mechanism of injury.

●An observational study of 39 surviving patients found that 17 had normal development, 17 had mild to moderate developmental delay, and five had severe impairment. Cerebral ischemia on initial head CT or MRI was associated with long-term impairment [51].

Radiologic sequelae of shaking injury include infarction, cerebral atrophy (generalized or asymmetric), encephalomalacia, hydrocephalus, and chronic subdural hemorrhage [52,102-104].

PREVENTION — The implementation of strategies for the prevention of abusive head injury is recommended by the American Academy of Pediatrics (AAP) Committee on Child Abuse and Neglect [105]. Nevertheless, the development of effective strategies is a challenge. Several types of programs have been studied.

Primary prevention efforts, which target a broad population, have included large-scale public education campaigns and hospital-based education programs [106,107]. However, based upon two studies, broad-based applications of such interventions have not been associated with corresponding decreases in hospitalizations for abusive head trauma (AHT) [108,109].

Secondary prevention programs directed toward a specific subset of the population at risk for child maltreatment have demonstrated some success. In one study, prenatal and early childhood home visitation significantly reduced the number of verified reports of child abuse and neglect in the group that received the intervention compared with the group that did not [110]. However, the effectiveness of this type of program in preventing abusive head injury has not been specifically studied.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Child abuse and neglect".)

SUMMARY AND RECOMMENDATIONS

●Background – Abusive head trauma (AHT) is a frequent cause of death and long-term disability in children who have been abused, particularly among children younger than 12 months of age. The diagnosis is often missed at the initial evaluation because clinical features are typically nonspecific, and a history of trauma is often absent. As a result, the child may not receive prompt treatment for the current injury and be at risk for later, potentially fatal, injuries. (See 'Background' above.)

●Clinical features – The signs and symptoms of AHT may be mild and nonspecific (eg, vomiting, poor feeding, irritability, or lethargy) or severe and even life threatening (eg, apnea or other breathing abnormalities, coma, or seizures). The neurologic examination is abnormal for most children with AHT. Retinal hemorrhages are common (though not pathognomonic) and correlate with the severity of injury. Some, but not all, children have cutaneous bruising, particularly around the face. (See 'Physical examination' above.)

●Laboratory evaluation – Laboratory evaluation should include complete blood count with differential, platelet count, and coagulation studies (algorithm 1). In children with suspected abuse, laboratory testing for occult abdominal injury, including liver enzymes, amylase, lipase, and a urinalysis, is also recommended. When meningitis is a serious consideration, cerebrospinal fluid (CSF) should generally be obtained if a lumbar puncture can be safely performed, primarily to exclude infection as a cause for nonspecific symptoms. CSF is typically xanthochromic and/or bloody in children with AHT. (See 'Laboratory studies' above.)

●Imaging – A skeletal survey (table 4 and table 5 and table 6) and neuroimaging are essential components of the evaluation for AHT. For the initial evaluation of hemodynamically or neurologically unstable children, we recommend CT with brain and bone windows as the preferred imaging modality after the initial trauma resuscitation. Whenever possible, three-dimensional skull reconstruction should be added to enhance the detection of skull fractures. (See 'Skeletal evaluation' above and 'Unstable patients' above.)

For stable patients, we suggest MRI rather than CT if MRI can be performed in a timely manner and interpreted promptly by a pediatric neuroradiologist. (See 'Stable patients' above.)

Because of the high prevalence of cervical spine injuries in children diagnosed with AHT, we recommend cervical spine with brain MRI at two to three days after injury in all victims. (See 'Stable patients' above.)

●Ophthalmologic examination – Children with suspected abusive head injury should have a funduscopic examination, preferably by an ophthalmologist and ideally within 24 to 36 hours of presentation, to identify retinal hemorrhages and other eye injuries. (See 'Ophthalmologic examination' above and "Child abuse: Eye findings in children with abusive head trauma (AHT)".)

●Diagnosis – The diagnosis of AHT is aided by the presence of additional clinical features in children with intracranial injury, including a history that is inconsistent with physical findings, retinal hemorrhages, fractures, unexplained bruises, apnea, and seizures. A thorough evaluation by a multidisciplinary team should be performed whenever there is suspicion that an injury is inflicted. Subsequent investigation by social service and law enforcement agencies may substantiate the diagnosis of abusive head injury. (See 'Diagnosis' above.)

●Differential diagnosis – Abusive head injury must be differentiated from unintentional injury. Other diagnoses that must be excluded include birth trauma, bleeding disorders, and other rare conditions such as glutaric aciduria type 1 and other metabolic diseases. (See 'Differential diagnosis' above.)

●Management – In addition to stabilizing the medical condition of the abused child, initial management must also address the safety of other children in the home. (See 'Management' above.)

●Outcome – The morbidity and mortality for abusive head injury is significant. Mortality rates as high as 23 percent have been reported. The majority of surviving children have permanent neurologic impairment. (See 'Outcome' above.)

ACKNOWLEDGMENT — The editorial staff at UpToDate acknowledge Erin Endom, MD, who contributed to earlier versions of this topic review.