Battered Child Syndrome

Battered child syndrome, described by Kempe and coworkers in 1962, was the first child abuse syndrome to become widely recognized, and its description brought the problem of inflicted injury into focus.¹² Battered child syndrome can be seen in infants through adolescents but is most common in children younger than 3 years. Children with this syndrome are brought to medical attention for an unrelated problem or in the setting of a particular acute injury. When evaluated, they are noted to have signs of chronic abuse, which may include poor hygiene, malnutrition, growth retardation, multiple cutaneous bruises of different ages, pattern injuries, burn marks, and skeletal injuries at different stages of healing (Fig. 210-E1).

FIGURE 210-E1 Extensive cutaneous bruising of the trunk in a child with “battered baby syndrome.”

Typically, caretakers blame these visible traumatic injuries on various accidental mechanisms, often of a relatively trivial nature. On questioning, parents may characterize infants as fussy or stubborn and older children as clumsy, hyperactive, or accident prone. Chronically abused children may appear passive and withdrawn but often show strong attachment to the parent, even when the parent is the perpetrator.

Diagnosis of the “classic” battered child is usually straightforward from the history and physical examination. Although readily available computed tomographic scans enable us to rapidly characterize a patient’s head injury, a thorough physical examination of the rest of the body, with removal of the clothes, is the best tool for diagnosing a battered child. The history itself is often sketchy and elusive, and it is not uncommon that the adult accompanying the child is not the patient’s regular or exclusive caretaker. Members of the child protection team can be invaluable at this juncture; pediatricians, social workers, and others experienced in dealing with families in which child abuse is suspected are skilled at interviewing and piecing together the known sequence of events. Their involvement is also particularly useful as the evaluation progresses and the possibility of child abuse is broached with family members, who often respond with adamant denial and even frank hostility. The physician’s role in these exchanges is discussed in more detail later.

Central to Kempe and colleagues’ early description of the battered child was the idea that episodes of physical trauma are recurrent rather than isolated events. This principle has been widely supported by subsequent clinical experience, thus making early diagnosis of the syndrome imperative. However, obvious evidence of chronic abuse may not be readily apparent in all children. Therefore, certain types of injuries that occur with greater frequency in the setting of physical abuse should raise the physician’s level of suspicion for nonaccidental causation. Such injuries include spiral fractures of the humerus, spiral fractures of the femur in infants, metaphyseal fractures in infants, duodenal hematomas, “tin ear,” frenulum tears in nonambulatory infants, immersion burns, patterned bruises, and retinal hemorrhages.^13–16 In addition, children with preexisting disabilities or prematurity may be at particular risk.¹⁷ “Coin rubbing,” seen in Southeast Asian children, and similar therapeutic maneuvers resulting in pattern marks should not be confused with abuse. With respect to neurological trauma, “red flag” injuries include stellate skull fractures, bilateral or multiple skull fractures, and subdural hematomas outside the setting of motor vehicle trauma.^1,18 Additionally, the entire range of soft tissue contusions, cephalohematomas (sometimes from severe hair pulling), skull fractures, and intracranial hemorrhages can occur in the setting of inflicted injury.

A skeletal survey is a mandatory part of the evaluation of suspected nonaccidental injury in infants and young children, and a bone scan may be useful when plain films are equivocal. Anemia, thrombocytopenia, or other hematologic abnormalities are evaluated by standard laboratory tests. Toxicology screening is sometimes unexpectedly positive because of forced ingestion of toxic substances. The question of osteogenesis imperfecta predisposing to multiple fractures is sometimes raised, but in most forms of this condition, patients are distinguished by other radiographic and clinical features (such as blue sclera), a history of consistent and well-described mechanisms of injury, and prompt seeking of appropriate medical care.

It may be useful to obtain a baseline computed tomographic or magnetic resonance imaging (MRI) scan in patients with battered child syndrome, even if an acute head injury is not part of the initial clinical picture. Such a study can both ascertain evidence of previous brain injury and provide a comparison in the event of future injuries.

Shaking-Impact Syndrome

The term shaken baby syndrome was originally coined by Caffey in 1972 to describe infants with acute subdural and subarachnoid hemorrhages, retinal hemorrhages, and periosteal new bone formation at the epiphyseal regions of the long bones.¹⁹ Some authors include infants with more chronic extra-axial collections in this category, but for purposes of this discussion, such cases are considered separately. Although the diagnosis of shaken baby syndrome rests on clinical and radiographic features, the name implies a specific mechanism of injury and was derived in part from the case of a nursemaid who admitted shaking several infants injured in her care in an attempt to burp them.¹⁹ At about the same time that this syndrome was described, clinical and laboratory evidence of the damaging effects of angular acceleration on the brain was being reported.^20–22 Thus, the mechanism of the injuries commonly found in abused infants was postulated to result from “whiplash-shaking” as a form of discipline, and the term shaken baby syndrome became widely accepted as both a diagnosis and a mechanistic description. Support for the validity of the term was found in the observation that many infants with intracranial findings of the syndrome had little if any evidence of blunt impact to the head on initial physical examination. In addition, infants’ relatively large heads, weak neck muscles, and watery brain consistency were thought to render them particularly vulnerable to severe injury from being shaken back and forth by a caretaker.²³ The long bone findings, which can also be seen in accidental injuries in which the limbs are jerked or pulled, were likewise thought to occur as a result of violent shaking. Central to the concept of shaken baby syndrome was the idea that caretakers might inflict these injuries unwittingly in the course of a generally acceptable means of discipline, during choking, or even during play.⁵

Controversy about the mechanism of these injuries arose because of the paucity of a reliable history typically available to the evaluating physician. When a history of trauma is offered, it is usually that of a relatively minor blunt impact, and only rarely is an unsolicited history of shaking obtained. More recently, the term shaken baby syndrome has been questioned because clinical series, autopsies, and biomechanical and radiographic analyses have suggested that many if not most of these infants do, in fact, have evidence of a blunt impact to the head and that the deceleration forces generated by shaking alone are insignificant in comparison to those caused by impact, even when it is against a padded object.^24–26 The frequent lack of dramatic cutaneous bruising can be explained by the dissipation of angular deceleration forces across a relatively wide and soft surface.²⁷ It seems likely both from these studies and from careful questioning of perpetrators that although an infant may be shaken, the final thrust involves the head striking a surface, thereby resulting in the high deceleration forces required to cause subdural hemorrhage and frequently severe parenchymal damage. For these reasons, some authors prefer the term shaking-impact syndrome to distinguish the mechanism of shaking in child abuse from shaking during play, shaking to resuscitate, or other less violent scenarios sometimes postulated as being responsible for injuries.^28,29 In addition, the frequent findings of long bone injuries, rib fractures, cutaneous bruises, skull fractures, subgaleal and subperiosteal hemorrhages, and focal contusions of the brain parenchyma belie a simple nonimpact cause. Still, the question of whether shaking alone is ever sufficient to cause the brain injuries commonly seen in abused infants remains controversial, and battered child syndrome and shaking-impact syndrome result in a spectrum of overlapping injury types and chronicities seen in patients of varying ages.³⁰

Regardless of the exact mechanism of injury, the clinical scenario in shaking-impact syndrome is often remarkably similar from case to case. Affected children are nearly always 2 years or younger, and most are younger than 6 months. They are brought to medical attention because of irritability, poor feeding, or lethargy in mild cases and because of seizures, apnea, or unresponsiveness in more severe cases. The history is often vague, and it is common that those accompanying the child are not the exclusive caretakers. In many cases, no history of trauma is offered, and the diagnosis may come to light when a lumbar puncture done as part of an evaluation for sepsis shows bloody spinal fluid. In other cases, a history of relatively trivial trauma is given. Sometimes, on questioning a history of shaking to resuscitate is obtained.

The child abuse evaluation team, if available, is notified, and specific histories from all caretakers involved with the child should be obtained as soon as possible. Although these types of injuries occur in families of all sorts, they are seen most commonly in more fragmented family situations; typically, the infant has multiple caretakers, the parents are young, resources are limited, or other stressful conditions are present. Drugs or alcohol may be involved. One commonly encountered scenario involves a fussy baby being watched by an inexperienced caretaker, such as the mother’s boyfriend. Starling and associates found that perpetrators were fathers, boyfriends, female babysitters, and mothers, in descending order of frequency.³¹

On physical examination, a range of neurological abnormalities may be found, from mild irritability and lethargy to flaccid coma. Some children with seizures may exhibit “bicycling” movements, which can be mistaken for normal spontaneous activity. Even severely injured young infants often show nonspecific withdrawal to noxious stimulation and may even have spontaneous eye opening, but they can usually be identified by a paucity of normal, spontaneous motor activity and by a distinct lack of crying or vigorous grimacing in response to pain. The fontanelle may be full. Careful inspection frequently reveals mild bruising, most often in the parieto-occipital region or, less commonly, in the frontal area, which may be more apparent after several days. Retinal hemorrhages are typically found. Computed tomography shows subdural or subarachnoid hemorrhage ranging from barely perceptible to sizable collections with a mass effect requiring emergency surgery (Figs. 210-1 and 210-2). The hemorrhage may be unilateral or bilateral and has a particular propensity for the posterior interhemispheric space.³² It may result from an impact to the back of the head, displacement of bone across the lambdoid sutures, and subsequent strain on the underlying venous sinuses and deep veins. MRI is frequently superior to computed tomography in demonstrating small subdural hemorrhages and parenchymal contusions; it is particularly helpful if the diagnosis is equivocal, such as when trauma is denied by the caretakers (Fig. 210-3).³³ MRI is also a useful screening test for arteriovenous malformations or other vascular anomalies that could cause subarachnoid hemorrhage, particularly when all other tests for associated injuries are unrevealing and the cause of the hemorrhage remains unclear. In rare cases, arteriography may be considered to rule out vascular abnormalities when there is no history or radiographic evidence clearly pointing to trauma and the child has an ictal intracranial hemorrhage.³⁴

FIGURE 210-1 Shaking-impact syndrome, bilateral form. There was no history of trauma in this unresponsive 2-month-old infant. A, Acute computed tomographic scan showing posterior interhemispheric blood. B, Severe diffuse brain atrophy seen 2 months after injury. C, Typical long bone injury of the humerus.

FIGURE 210-2 Shaking-impact syndrome with unilateral subdural hematoma. The child reportedly slid down two carpeted steps. A, Acute preoperative computed tomographic scan. B, Progressive brain swelling and unilateral loss of gray-white differentiation. C, Scan 1 year after injury. The child remains severely impaired.

FIGURE 210-3 Magnetic resonance image in an infant with shaking-impact syndrome. Note the thin subdural hemorrhages and focal intraparenchymal contusion.

In severe cases of shaking-impact syndrome, the brain may lose its normal gray-white differentiation and have the appearance of a large unilateral or bilateral supratentorial infarction. This finding may be visible on the initial scan or may develop 1 to 2 days after injury.^35,36 Children with this finding are usually unresponsive on admission and have a dismal prognosis for neurological recovery. The pathophysiology of the so-called black brain seen in these children is incompletely understood but may be due to the synergistic effects of hypoxia, mechanical trauma, and subdural hemorrhage.^37–40 In some children, evidence of a high spinal cord injury can be found, and this may contribute to the apnea and poor outcome seen in some cases.⁴¹ Spinal injuries in child abuse are discussed in more detail later.

Once the acute management issues have been attended to, as outlined later, a diagnostic evaluation for associated injuries and causes should be pursued. A general screening for other occult injuries should be performed, preferably by the pediatric trauma team. Routine laboratory studies for anemia, thrombocytopenia, visceral injury, and coagulopathy are part of the evaluation; it should be kept in mind that coagulopathy can occur as a result of severe brain injury and does not necessarily imply a preexisting condition.⁴²

As in a battered child, a skeletal survey is mandatory, and a bone scan is sometimes helpful in equivocal cases. A repeat skeletal survey in 2 weeks may increase the yield of diagnosed injuries because of more visible changes with healing.⁴³ A formal ophthalmic consultation with mydriatics, when clinically suitable, should be obtained to document retinal hemorrhages. Although such hemorrhages have been reported in 65% to 95% of patients with shaking-impact syndrome, they are not always present and, conversely, may occasionally be found in children with head injuries from accidental causes and, rarely, after resuscitation.^{1,2,15,41,44–48} Nonetheless, the presence of retinal hemorrhages adds greatly to the suspicion of nonaccidental injury, especially when they are bilateral and severe.⁴⁹

No other medical condition fully mimics all the features of full-blown shaking-impact syndrome, but certain features can occur in other conditions. Coagulopathies, vascular anomalies, and anatomic abnormalities such as arachnoid cysts can be associated with subdural hemorrhage.^50–52 Osteogenesis imperfecta can be associated with fractures and, rarely, subdural collections, but not with the other features of shaking-impact syndrome.⁵³ Glutaricaciduria can cause progressive neurological decline and subdural collections, sometimes triggered by a viral illness or injury, but it does not feature bony abnormalities or retinal hemorrhages.⁵⁴ The single most common diagnosis mimicking nonaccidental trauma is accidental injury. In these cases, small epidural hemorrhages or traumatic subarachnoid hemorrhages can be mistaken for subdural bleeding, and unusual subdural hemorrhages can occur when the requisite biomechanics is present in settings not generally associated with this injury; such patients occasionally exhibit retinal hemorrhages as well.^47,55 However, in these cases, the history is clear and consistent, and the clinical status is concordant with the forces involved in the injury, without unexplained skeletal or soft tissue injuries.

In many cases, such as those in which there is no history of trauma but the infant has skull fractures and unexplained long bone trauma, the diagnosis is quite clear. In other cases, despite careful evaluation, the mechanism of injury remains obscure. An algorithm has been developed to match the patient’s injury type with associated findings and best history in an attempt to more objectively classify infant head injuries as accidental or inflicted.² However, until more is understood about the injury thresholds in children of different ages, the pathophysiology of retinal hemorrhages and “black brain,” and the effects of chronic, repeated trauma on the immature nervous system, efforts to assign a definitive mechanism to every suspicious injury will be unsuccessful.

Two additional issues that arise frequently and on which the neurosurgeon may be asked to comment involve the timing of injury and the possibility of multiple, sublethal accidental injuries that might behave synergistically. With respect to the first issue, Willman and coauthors reported on a series of 95 fatal accidental head injuries in children; in all but one patient there was an immediate onset of neurological symptoms and decreased level of consciousness. The one exception was a child with an expanding epidural hematoma.⁵⁶ Although this study included only a small number of infants, it suggests that a prolonged lucid interval is unlikely in a child with a fatal primary brain injury. This conclusion is in accord with data from accidental trauma in adults and from animal models.²⁷ It should also be kept in mind that care must be taken when attempting to base the time that an injury took place on radiographic findings because they may include a spectrum of abnormalities.³⁶ With respect to the possibility of “second impact syndrome” mimicking child abuse, such rare events have been reported in older children and adults and usually involve well-documented concussive sports injuries resulting in acute subdural or subarachnoid bleeding and brain swelling.^57,58 At present, there is no evidence to support the notion that fatal acute traumatic subdural hematoma in otherwise healthy infants occurs from multiple trivial impacts.

Physical Abuse in Older Children

Most physically abused older children brought to medical attention suffer from soft tissue or visceral injuries as a result of direct blows, although intracranial injuries sometimes occur and can be serious or even fatal. The setting is usually that of a biologic or foster family in which deviations from rigid codes of behavior are dealt with by physical punishment and beating, sometimes in an attempt to “save” the child (Fig. 210-4). The parents of both older and younger abused children may have been the victims of child abuse themselves. Occasionally, the perpetrator is psychiatrically impaired, but this is the exception.

FIGURE 210-4 Computed tomographic scan of a 9-year-old boy who was beaten by his foster mother and forced to drink water as a means of discipline. He became unconscious and was found to be profoundly hyponatremic; he succumbed from diffuse brain swelling. Autopsy showed only minor superficial contusions of the brain and subarachnoid hemorrhage.

Evaluation includes a careful history, and the child should be questioned apart from the parent once some degree of trust has been attained. A general trauma evaluation is performed, including routine studies such as urinalysis for hematuria or myoglobinuria. A skeletal survey is generally of limited use in an older child because the typical occult injuries (e.g., metaphyseal and rib fractures) seen in infants do not occur in this age group and remote skeletal injuries have usually healed to the point of being undetectable. A careful history of previous trauma, including fractures, and the physical examination are typically more helpful in detecting suspicious findings in this age group.

Head injuries may include soft tissue lesions; linear, depressed, or basilar fractures; and the range of intracranial lesions seen in trauma in general. Management is guided by the specific injury. Counseling is often in order, and a neuropsychological evaluation may be useful in an older child with acute or chronic brain involvement or behavioral disturbances to help in appropriate school placement.

Acute Subdural Hematoma

Some subdural hematomas resulting from inflicted injury appear identical to those caused by accidental injury and clearly have a mass effect; these hematomas require emergency evacuation in the standard fashion. Even with prompt evacuation, however, changes in the underlying brain often persist or progress and take on the appearance of widespread infarction (see Fig. 210-2A and B). More commonly, subdural and subarachnoid blood is quite diffuse in child abuse injuries and appears as a thin layer without marked compression of the underlying hemisphere. These “smear” collections are generally managed nonoperatively; although aggressive surgical evacuation plus decompression has been reported, this strategy has not been strictly compared with medical management.⁵⁹ Frequently, the processes leading to the widespread parenchymal loss commonly seen in severely affected infants have already been initiated at the time of arrival at the hospital and may be the result of local compression, apnea, hypoxia, mechanical trauma, or seizures.^36,60 Conversely, infants who do not appear critically ill at initial encounter often remain alert and regain a normal level of consciousness quite promptly despite their acute subdural collections and do not eventually suffer large delayed infarctions.

In children in whom gray-white differentiation is lost, brain swelling may be amenable to standard or even extraordinary medical management for increased intracranial pressure, but conventional therapy rarely if ever prevents the swollen brain from progressing to severe atrophy (see Fig. 210-2). In very young infants, brain swelling may not be a life-threatening problem because the skull simply expands to accommodate the swelling; these infants survive, but in a devastated state (see Fig. 210-1B). Because of this dismal outlook, the role of more aggressive measures, including intracranial pressure monitoring, is controversial in this population.^61,62

Chronic Extracerebral Fluid Collection

Large, bilateral, chronic extracerebral fluid collections composed of old blood products or proteinaceous cerebrospinal fluid are sometimes discovered in infants. The term chronic subdural hematoma is often applied to these collections, although the content of the accumulation may vary from thin, watery fluid resembling cerebrospinal fluid to the thick “motor oil” often associated with adult chronic subdural hematomas. Even modern, sophisticated imaging techniques such as MRI may not be able to distinguish fluid in the subdural space from that in the subarachnoid space, and the term chronic extracerebral fluid collection is probably more accurate.⁶³ A definite history of antecedent trauma may not be obtained, and the question of child abuse often arises.

Guthkelch pointed out that sudden acceleration and deceleration of the head, even without direct impact, can result in the tearing of cortical bridging veins in adults.²³