Objective. Advances in radiologic technique have increased the recognition of subdural hematoma. No study to date has addressed the role of child protective investigation into the cause and management of subdural hematoma in children who lack other indicators of abuse.
Methods. Medical records, radiology studies, and social service notes for all infants and children referred for child abuse investigation who had any form of intracranial hemorrhage were reviewed. The study covered the 12 months of 1997. All referrals were to the Suncoast Child Protection Team (St Petersburg, FL).
Results. There were 19 investigations because of subdural hematoma. Eight children had retinal hemorrhage as well as other major findings of trauma, such as bruises and/or fractures; all 8 were victims of child abuse. Two infants had tiny subdurals adjacent to accidental linear skull fractures. Nine infants were investigated for the possibility of abuse that had no findings of trauma apart from the subdural hematoma. These 9 cases form the basis for this study. The age range was 11 days to 15 months. Inflicted cerebral trauma was the medical diagnosis in 8 of the 9 cases; 1 case had a final diagnosis of possible inflicted injury in a high-risk setting.
Conclusions. Infants with subdural hematoma but no other findings of abuse present a difficult challenge to child protection workers. Investigation by a medically oriented team can uncover the circumstances of the trauma in most instances and can usefully direct protective efforts. The high incidence of severe sequelae in infants with inflicted cerebral trauma warrants a vigorous approach.
- CT =
- computed tomography •
- MRI =
- magnetic resonance imaging •
- CSF =
- cerebrospinal fluid
In 1992, Duhaime et al1 proposed an algorithm for the determination of inflicted injury in very young children with intracranial hemorrhage. Those with associated fractures and/or inflicted soft tissue injuries were considered to have presumptive evidence of inflicted injury. Infants with intracranial hemorrhage lacking these findings and also lacking a history of trauma were considered to be suspicious for inflicted injury. Our study looks specifically at children with features that would be classified in the suspicious group in whom subdural hematoma was found on computed tomography (CT) or magnetic resonance imaging (MRI) but who lacked bruises or fractures of abuse. The algorithm of Duhaime et al did not specify the presence or absence of retinal hemorrhages. The subgroup in this study had subdural hemorrhage but no retinal hemorrhage.
The finding of subdural fluid collections in the absence of a specific incident, such as an auto accident or a medical condition, such as meningitis, tumor, or bleeding diathesis raises significant concern that child abuse is the diagnosis. This means that a child protective investigation will normally follow this finding.2 The inquiry is charged with urgency because of the high mortality rate associated with abusive head injury3 and because in some series almost all survivors will ultimately show a consequence of the injury, such as seizure disorder, behavior problem, developmental delay, or learning problem.4
All referrals involving intracranial hemorrhage of any type to the Suncoast Child Protection Team (St Petersburg, FL) in the 12 months of 1997 were reviewed from among more than 400 consultations for alleged physical abuse. The medical charts, child protective medical consultations, case coordinator notes, and interviews with investigators from the Department of Children and Families and from law enforcement departments were analyzed. The geographic area comprises ∼3 million people and is served by a tertiary pediatric referral center, All Children's Hospital.
There were 19 cases of intracranial hemorrhage (subarachnoid, epidural, and subdural) that had a forensic, pediatric consultation by a physician because of possible child abuse of a total of 32 infants and children identified as having some form of intracranial bleed. There were 13 cases of intracranial hemorrhage that were not referred for child protective investigation. Six were postoperative bleeds following neurosurgery; 5 were related to birth trauma. One infant was unrestrained in a motor vehicle accident, and another suffered a fall down stairs in a walker.
Eight of the referred cases were excluded from this study because they had inflicted bruises and/or fractures of abuse and/or retinal hemorrhages. There were several deaths in this group. All 8 had a medical diagnosis of nonaccidental, inflicted cerebral trauma, and protective investigation confirmed the finding of child abuse. Two cases were excluded in whom the intradural bleed was very tiny and located adjacent to a linear skull fracture. Both were the result of accidental falls confirmed by protective investigation.
The study consisted of the remaining 9 infants referred for protective investigation with subdural hemorrhage who lacked external bruises, fractures, and retinal hemorrhages.
A standard protocol was followed in all cases. The infants were examined by a pediatric ophthalmologist for retinal hemorrhage and by a pediatric neurosurgeon for indications for surgery. All infants had CT or MRI studies of the brain, and most had both. A standard radiologic protocol was followed to look for fractures and subgaleal hematomas. All infants were examined in detail by a pediatrician trained in child abuse to uncover any cutaneous manifestations of abuse. Complete blood cell and platelet counts were standard. Most had clotting studies. There were no clinical indicators of bleeding diathesis in any subject before or after the referral. A standard medical history was obtained from caretakers by the examining pediatrician.
In Florida, a child protective team is present in each judicial district. The teams provide psychological evaluations, in-depth psychosocial interviews and other professional services on request. The teams coordinate the professional investigations with those of law enforcement agencies and the Florida Department of Children and Families. In this series, the doctors worked with 4 separate child protection teams on the west coast of Florida. The conclusions are the medical diagnoses rendered by the doctors as team consultants.
The 9 patients included 3 males and 6 females. Four were full term; 4 were moderately premature at 34 to 38 weeks' gestation and 1 was very premature at 28 weeks' gestation. There were 6 singletons and 2 sets of twins (1 set concordant for injury and 1 set discordant). Two infants were Hispanic, 3 black, and 4 white. In 4 families, the parents were 20 years of age or younger.
The infants studied ranged in age from 11 days to 15 months, which is similar to series of inflicted head trauma that show a preponderance of victims under a year of age.5 Three were between 1 and 5 months of age, and 4 were 6 to 9 months of age. Victims of accidental, traumatic brain injury are considerably older on average.5
Inflicted trauma was not suspected on presentation before CT or MRI studies. Of the infants, 2 were studied because of increasing head circumference; 1 had vague neurological movement disorders; 2 had frank seizures; 1 presented with esotropia of sudden onset; 1 had bloody cerebrospinal fluid (CSF) during a work-up for possible neonatal herpes or sepsis; and 1 came to the emergency room with a possible life-threatening event. Only 1 infant had no symptoms and was studied because her twin was found to have a subdural hematoma.
The radiologic findings showed a chronic subdural collection in 2 patients; 5 had subacute subdural hematomas; and 2 had acute subdurals, including the 11-day-old infant. In the majority, there were extra-axial fluid collections on both sides.
Few had additional medical findings at the time of presentation. One patient had preexisting developmental delay and cerebral palsy secondary to extreme prematurity. Two patients, both with chronic, enlarging subdurals, had anemia sufficient to require transfusion. Both of these infants subsequently had neurosurgical intervention.
The investigation of these injured infants was difficult because of the shock and anger raised by the possibility of nonaccidental trauma and also by the long interval between the injury and its discovery in 7 of the 9 cases. All the families and many of the professionals had difficulty understanding the seriousness of an injury that had no visible manifestation.
The profile of this subgroup of infants matches that of other series of children with inflicted head injuries with respect to age (range: 11 days to 15 months of age) and presence of risk factors, such as prematurity, twins, and young and/or stressed parents.6–8It differs in the predominance of female victims.
The 9 infants with subdural hematoma and no retinal hemorrhage could be subdivided into 3 groups:
Group 1 (extrafamilial trauma, child abuse confirmed) had 2 infants who required surgery and who presented with chronic growing subdural hematomas. Both infants, coincidentally, were found on investigation to have been victims of trauma at the hands of female babysitters. Both sitters confessed to using corporal punishment in infants under 6 months of age. One sitter described “spanking” while the other called what she did “shaking,” although she minimized the force when asked to give a demonstration.
Group 2 (intrafamilial trauma, infants inappropriately handled) had 4 infants with subacute subdural hematomas, which were small and did not require surgery. In this group, shaking or other inflicted trauma could not be proven, but the infants were found to be in dangerous situations. Inappropriate handling had been observed.
Group 3 (intrafamilial trauma suspected, high-risk circumstances) had 3 infants in whom the circumstances of injury remained obscure even after extensive investigations. Two had acute subdural collections, and 1 had a subacute subdural hematoma. Investigation did uncover major risk factors for child abuse in all 3 (see Table 1).
Representative Case Reports
Group 1: Case 2
Infant B presented at 9 months of age because of enlarging head circumference. Several weeks before the mother and the pediatrician were puzzled by a number of hard-to-explain symptoms. A series of studies were performed including a cranial ultrasound. Follow-up CT revealed chronic, extensive subdurals bilaterally. The head circumference was ≫95%; she was anemic.
Medical record review showed normal head growth through 4 months of age and abnormal measurements at 7, 8, and 9 months of age. The focus of investigation became the period of 5 to 6 months of age. An incident was found at 6 months of age in which the mother phoned the nurse triage system because her daughter had been sleeping for 17 hours and was not acting herself. Medical findings were nonspecific. Some days later facial bruises were noted when she was picked up from the babysitter. An older sibling stated that the sitter “squeezed” infant B's face.
The sitter had been dismissed and was difficult to locate. When found, she admitted to law enforcement officials that she had shaken the infant at the time of the excessive sleeping incident because her crying disturbed the sitter's own infant.
Group 2: Case 3
Infant J was hospitalized at 15 months of age because of vague neurologic symptoms, such as arching movements of the neck. A diagnosis of encephalitis was entertained before an MRI showed bilateral subacute subdural hematomas. The mother recalled no trauma. The family had been under stress because of the father's death in a firearm accident shortly before infant J's birth.
In intensive investigation, it was found that the mother was still severely depressed. She was attempting to upgrade her employment by returning to school. When competing demands of family and school and grieving became too much, she would isolate herself for long periods leaving the infant in the care of older siblings. The oldest child was 11 years of age (5 ft and 150 lbs). He was also found to be severely depressed. He stated that he had seen his father play with infants by tossing them in the air, and he did the same. On the occasion that the mother saw this happen, she halted the behavior as too rough. What level of trauma the infant was subjected to in her absence is not known.
Both the mother and the older sibling are presently in psychiatric treatment.
Group 3: Case 9
Infant SM was 1 of the set of twins discordant for intracranial hemorrhage. She presented with a seizure at 2 months of age; MRI revealed acute, bilateral moderate subdural hematoma. Delivery at 36 weeks was by cesarean section.
An attorney related to the family was called while the infant was still in the radiograph department. The family subsequently blocked all forms of investigation and ignored the provisions of a relative placement order. Risk factors included a complaint of domestic violence against the mother because of a knife attack on the father; a minor criminal history of the father (disorderly conduct); and a family history of confirmed physical child abuse involving the mother's siblings some years before.
At 4 months of age, a second episode of bleeding was discovered on routine follow-up CT. Custody was changed to a different relative, but the family were no more cooperative.
The case remains before the court. The parents have refused to have court-ordered psychological evaluations. Custody remains officially with a relative.
Shaken baby syndrome or the syndrome of inflicted cerebral trauma is based on findings that include intracranial hemorrhage, retinal hemorrhage, cerebral contusion or other brain tissue injury, and evidence of cerebral trauma, such as an altered state of consciousness. Diffuse axonal injury may often be inferred indirectly, and metaphyseal fractures are frequently found.5 8–10 This series shows that inflicted cerebral trauma may be the cause of subdural hematoma even when other findings are lacking and that “inappropriate handling” of infants, dangerous child care practices, and abuse risk factors are frequently associated with this injury.
Intracranial injury in infants <1 year of age has come to be regarded as less often associated with accidental injury and more often the result of child abuse.2 ,11 Epidural hematoma is the only serious head injury in infants that is most commonly accidental.12 ,13 In this series, maltreatment of one sort or another remains the predominant mode of injury even in the absence of the classic “shaken baby” triad described by Caffey10 of intracranial hemorrhage, retinal hemorrhage and metaphyseal fractures.
Retinal hemorrhage has never been considered a necessary finding to make a diagnosis of either shaken baby syndrome or inflicted cerebral trauma, although it is found in 85% or more of infants with acute, abusive intracranial trauma.2 ,14 The absence of retinal hemorrhage made some workers in this series uneasy about a diagnosis of nonaccidental injury. Because 7 of 9 of these infants had subacute or chronic subdural bleeds, the chance for finding retinal hemorrhage was reduced. Retinal hemorrhage and cutaneous bruises have never been features of chronic subdural hematomas. As in this series, chronic subdurals are a disease of infancy, and only a minority requires neurosurgical intervention.15 ,16
Recognition that chronic or subacute subdural hematomas are frequently associated with child abuse once cases of motor vehicle injury, tumor, shunts, and bleeding diathesis are eliminated has been growing in recent years.2 17–19
Billmire and Myers2 stated: “We believe that the occurrence of intracranial injury in children less than 1 year of age, in the absence of a history of significant accidental trauma such as a motor vehicle accident, constitutes grounds for a child abuse investigation.” This series confirms this recommendation.
Some authors have suggested that because of the large CSF space commonly encountered in infants, these bleeds can sometimes be caused by nonsignificant trauma.20Kleinman,21 however, in his most recent text states, “Given that SDH is rare with minor injury and that prominent CSF spaces are common in normal infants, the view that enlarged CSF spaces predispose to SDH does not have a firm scientific basis… this concept should be viewed with extreme caution.” We would agree with this approach after reviewing this series of cases.
Although the investigations were long and difficult, they were worthwhile from the perspective of child protection. In all 9 instances, conditions were found that needed correction for the wellbeing of the infant. The fact that the trauma was separated from the investigation by several days to several weeks simply made the search for answers different from that of infants with retinal as well as intracranial hemorrhage but not less productive.
There was a marked contrast between infants with subdural hemorrhage alone and the group with retinal hemorrhage and other injuries; the acute, retinal hemorrhage group had a very high morbidity and mortality rate. There was a predominance of male abusers. The subdural-only group was seen typically days or weeks after the event; no boyfriends were involved with the infants in this series. In the place of boyfriends, we found female babysitters and dangerous child care practices.
The difficulties involved in these investigations cannot be underestimated. Because of the allegations and because the initial investigation required out-of-home placement for almost all, there was a tremendous amount of anger and legal maneuvering on the part of the families. The problems were compounded by the length of time between the incident and the discovery of the trauma.
The type of in-depth investigation required to unravel a complex case is not within the reach of individual doctors regardless of their expertise in child abuse. A multidisciplinary child protection team, however, brings together a wide range of resources, which can aid in the medical diagnosis of child maltreatment. Individual cases in this series were elucidated by psychosocial interviews, psychological evaluations, and polygraph interviews by law enforcement agencies, as well as the standard law and child protective investigations. The team coordinators bring this information into focus and correlate the information with the medical findings.
One area that proved crucial in these cases was a complete collection and review of medical records. Plotting a growth chart of the head, for instance, helped pinpoint the likely time of occurrence of head trauma. Reading the original interviews by law enforcement and child protection workers was helpful. Working as a team to collect and interpret the documents was essential. In 1 case, an after-hours call to a nurse triage service proved to be the critical piece of medical information along with a plot of head circumference measurements from birth onward. In a different case, the significant data were contained in a police interview. Although the detective summed up the interview as “she denies trauma,” the child protective services worker noted that in the actual transcript the babysitter had admitted to using corporal punishment on a 5-month-old infant. A child protection team can place this type of data into context with the medical findings to arrive at a diagnosis.
A precise history of the trauma was not usually available. It was the position of the child protection team that these cases represent nonaccidental trauma and that the force involved greatly exceeds reasonable handling. It is not possible to calculate precisely the amount of rotational force needed to produce these injuries, but data from fatal cases suggest that the force is considerable.13 ,14 ,22
We believe that in infants with subdural hematoma who lack an obvious cause a careful investigation is always warranted. Our series found cases of inflicted injury, inappropriate infant handling, and/or high-risk social settings in all infants with intracranial hemorrhage who lacked a medical explanation. The investigations allowed targeting of specific interventions to recommend to the juvenile court to better protect these infants from maltreatment in the future.
- Received September 28, 1998.
- Accepted June 2, 1999.
Reprint requests to (M.W.M.) Suncoast Child Protection Team, Inc, 8800 49th St N, Suite 410, Pinellas Park, FL 33782. E-mail:
- Duhaime AC,
- Alario AJ,
- Lewander WJ
- Billmire ME,
- Myers PA
- ↵Feldman KW. Evaluation of physical abuse: head injuries. In: Helfer ME, Kempe RS, Krugman RD, eds. The Battered Child. 5th ed. Chicago, IL: University of Chicago Press; 1977:194–203
- ↵Reece RM. Child Abuse: Medical Diagnosis and Management. Philadelphia, PA: Lea and Febiger; 1994:8–15
- Krugman RD,
- Bays JA,
- Chadwick DL,
- et al.
- Caffey J
- Harwood-Nash D
- Hobbs CJ
- Helfer RE,
- Slovis TL,
- Black M
- Ewing-Cobbs L,
- Kramer L,
- Prasad M,
- et al.
- ↵Kleinman PK. Head trauma: extra-skeletal trauma and miscellaneous conditions. In: Diagnostic Imaging of Child Abuse. 2nd ed. St Louis, MO: Mosby; 1998:312–313
- Hanigan WC,
- Peterson RA,
- Njus G
- Copyright © 2000 American Academy of Pediatrics