Normal Computerized Tomography of Brain in Children With Shaken Baby Syndrome Yair Morad, MD,a Isaac Avni, MD,a Scott A. Benton, MD, FAAP,b Rachel P. Berger, MD, MPH,c Julie S. Byerley, MD, MPH,d Kathryn Coffman, MD,e Christopher S. Greeley, MD, FAAP,f Edward E. Gustavson, MD,g Carolyn J. Levitt, MD,g Ann Lenane, MD,h Jan Topley, MD,i and Alex V. Levin, MD, MHSc, FRCSCj Objective: To characterize the clinical presentation and clinical course of shaken baby syndrome (SBS) with normal cranial computerized tomography (CT) on admission and to suggest further diagnostic procedures in such circumstances. Methods: Using a worldwide listserv designed to facilitate discussion in the field of child abuse and neglect, we solicited case information for children hospitalized in different medical centers, who were diagnosed with SBS and had a normal CT scan on admission. Results: Nine cases were identified. While all children had an abnormal neurologic examination on admission, eight had a normal CT, and one had “widening of cranial sutures.” In four cases, subdural hemorrhage was diagnosed on magnetic resonance imaging (MRI) 3 to 7 days after admission. Five children had bone fractures. The neurological outcome was normal in four of nine cases. Five children had long-term neurologic damage. The diagnosis of SBS was supported by either perpetrator confession, characteristic evolution of brain abnormalities on CT or MRI, inconsistent or absent explanatory history, and/or other social risk factors. Conclusion: The diagnosis of SBS can be established even when brain CT is normal on admission. The documentation of retinal hemorrhages is of primary importance in establishing the diagnosis of SBS in these cases. (J AAPOS 2004;8:445-450) irst described by Guthkelch,1 and later more specifically defined by Caffey in the 1970’s as “whiplash shaken infant syndrome,”2 the Shaken Baby Syndrome (SBS) is a form of child abuse caused by violent shaking of an infant or young child, usually manifested as a combination of brain, ocular, and skeletal injury. Of this triad, brain injury is considered to be the most frequent, leading to neurologic abnormalities that exist at presentation in virtually all patients. Ocular injury, usually retinal hemorrhages, may be found in 50 to 100% of shaken infants.3,4 Skeletal findings, usually fractures of the long bones and ribs, are reported in 10 to 30% of the cases.5 F From the Department of Ophthalmology, Assaf Harofeh Medical Center, Tel Aviv University, Zrifin, Israel; Department of Pediatrics, Children’s Hospital, New Orleans, Louisiana, USA; Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania, USA; Department of Pediatrics, University of North Carolina School of Medicine, North Carolina, USA; St. Joseph’s Hospital Child Abuse Assessment Center, Phoenix, Arizona, USA; Division of General Pediatrics, Vanderbilt University Medical Center, Nashville, Tennessee, USA; Department of Pediatrics, Children’s Hospital of St. Paul, University of Minnesota, St. Paul, Minnesota, USA; University of Rochester, Department of Emergency Medicine, Rochester, New York, USA; Community Child Health, Pendragon House, Royal Cornwall Hospital, Cornwall, England; Departments of Pediatrics and Ophthalmology, The Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada. Submitted June 18, 2003. Revision accepted June 22, 2004. Reprint requests: Alex V. Levin, MD, MHSc, FRCSC, Department of Ophthalmology, M-158, The Hospital for Sick Children, 555 University Ave., Toronto, Ontario, M5G 1X8, Canada. E-mail: alex.levin@sickkids.ca Copyright © 2004 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/2004/$35.00 ⫹ 0 doi:10.1016/j.jaapos.2004.06.007 Journal of AAPOS Brain injuries are diffuse; in a study of 75 children with SBS, the most common finding was subdural hemorrhage (SDH) (93%), followed by cerebral edema (44%), and subarachnoid hemorrhage (16%).6 Other less frequent findings include parenchymal contusion, epidural hemorrhage, and vascular infarction. Virtually all 75 patients in that study had abnormal cranial computerized tomography (CT) on admission. Similar results were reported by Kivlin and coworkers, in a series of 111 patients.4 It is fairly well accepted that a child with the characteristic ocular and cranial findings or even cranial findings alone can be diagnosed as suffering from SBS. However, cases in which admission CT is normal present a particular diagnostic dilemma. We describe nine cases in which the diagnosis of SBS was made by child abuse experts in such circumstances and comment on the further diagnostic work-up that is necessary in these cases. METHODS Using a worldwide listserv designed to facilitate discussion in the field of child abuse and neglect, we solicited case information for children hospitalized in different medical centers who were diagnosed with SBS despite a normal CT scan on admission. Diagnosis of SBS was based on the presence at least two of the following criteria: (1) abnormal neurological examination; (2) skeletal injury consistent with SBS; (3) retinal hemorrhages; and (4) history of child abuse that included shaking with or without blunt head trauma or an inadequate history to explain the observed October 2004 445 Journal of AAPOS Volume 8 Number 5 October 2004 446 TABLE 1 Clinical and Demographic Details of the Study Group Sex/Age Neurologic Signs No. (months) on Admission 1 M/5 2 F/16 3 M/5 4 F/17 5 6 F/3 M/5 7 M/3 Signs of Trauma on Physical Examination CT MRI Skeletal Survey Normal Normal on admission. Rib fractures Buldging At 5 days, bifrontal fontanelle, subdural fluid that small bruises was present on on his back and first MRI in spine retrospect. “Limp” in the bed, Blood around the Normal At 5 days, small CSF- Normal intensity hygroma child’s mouth, after crying with in the right bruising to ear an “unusual frontal–parietal voice” convexity with small foci of associated acute blood products Normal Normal at 24 hours, Long bone Seizures, profound None 7 days, 6 months and rib lethargy, fractures irregular at 12 days breathing Apathy, decreased None Normal Not done Normal muscle tone Irritability Tender arm Normal Normal at 4 weeks Normal Decreased muscle Widening of Normal 6 days: scalp edema Rib fractures tone, irritable, cranial sutures drowsy Vomiting, irritable Seizures 8 M/2 Fussiness 9 M/5 Difficulty breathing None Normal Normal at 24 hours Bilateral subacute SDH at 7 days Normal Normal at 24 hours Possible healing bite on arm, traumatic lesions around the anus and on the buttocks. Rectal prolapse. None Normal At 3 days, small subdural hemorrhage Long bone and rib fractures Ocular Examination Neurologic Outcome Scattered bilateral Communicating hydrocephalus, mild preretinal and motor developmental RH delay Confluent bilateral Ataxia RH Confluent RH ⫹ VH bilaterally Normal 10 scattered RH in each eye Unilateral VH Confluent RH ⫹ marked papilledema bilaterally Bilateral multilayered RH ⫹ VH necessitating vitrectomy Normal Normal Poor head control, intermittent esotropia, and poor tracking ability Seizure disorder; motor developmental delay of 6 months Normal Normal Multiple RH in right eye, single hemorrhage in the left eye Healing rib fractures and clavicle fracture Bilateral confluent Seizures, swallowing RH difficulties SDH, Subdural hemorrhage; RH, retinal hemorrhage; VH, vitreous hemorrhage. injuries. We excluded children that were believed to suffer from other forms of child abuse without evidence of SBS. All children were evaluated by multidisciplinary child abuse teams at the hospital where the diagnosis was made. RESULTS Nine children fulfilled the inclusion criteria. Ages ranged from 3 to 17 months (average, 6.8 months). Six were males. Case histories are summarized in Table 1; Four children had normal neurological outcome and five children had long-term neurological deficits. All the children with abnormal outcome had abnormal results on MRI scans that were done after admission, while none of the four children with normal outcome had an abnormal MRI scan. All five children with abnormal neurological outcome had severe retinal hemorrhages, sometimes accompanied by vitreous hemorrhage or papilledema. In contrast, three of the four babies with normal neurologic Journal of AAPOS Volume 8 Number 5 October 2004 447 outcome had relatively mild ocular damage. One had few scattered hemorrhages; one had unilateral localized vitreous hemorrhage with no other retinal hemorrhages and the third had only one hemorrhage in one eye and many hemorrhages in the other eye. For none of the babies was a history of violence given on admission. Babies were described to be “sleepy,” “limp,” “choking,” or having seizures. In two cases the perpetrator confessed to abusively shaking the child. Below are two illustrative case reports. Case 1 A 5-month-old male presented to the Emergency Department with a 2-week history of vomiting, occasional episodes of diarrhea, and low-grade fever. On examination he was found to be fussy but consolable, afebrile, and with normal vital signs. His fontanel was full but flat and his tone was decreased. He had a small bruise on his back. A head CT was normal (Figure 1a). He was discharged home. Two days later he returned to the Emergency Department with a bulging fontanel and worsening irritability and was admitted. A repeat head CT was normal. Lumbar puncture showed 588 red blood cells/mm3. MRI of the brain that night was initially read as normal. The following morning rib fractures were noted on radiographs. An ophthalmologist observed scattered bilateral preretinal and retinal hemorrhages in both eyes. Repeat MRI a day later showed small bifrontal subdural fluid collections (Figure 1b). These were subtle, but in retrospect, had been present on the first MRI. The child was routinely cared for by his mother, father, and a nonrelative babysitter. On a follow-up examination two months later, the child was found to have a communicating hydrocephalus and developmental delay. Most of the retinal hemorrhages had resolved except for the remaining evidence of macular hemorrhage. No perpetrator was identified. Case 2 A 16-month-old girl was reported by her mother’s boyfriend to be found “limp” in the bed, with blood around the child’s mouth after crying with an “unusual voice.” On admission to the hospital, physical examination was remarkable only for a small amount of bruising of one ear. No intra-oral injury was noted. Cranial CT was normal (Figure 2a), as were skeletal survey, electro-encephalogram, electrocardiogram, and laboratory examinations (complete blood count, thrombin and pro-thrombin time, serum chemistry). Ophthalmological examination done after 3 days revealed “extensive retinal hemorrhages bilaterally.” MRI done 5 days after admission (Figure 2b) disclosed a small CSF-intensity hygroma in the right frontal–parietal convexity with small foci of associated acute blood products. The child was discharged home. On a follow-up examination one month later she was reported to be “clumsy” and was found to be ataxic. The perpetrator was not identified. FIG 1. (a) Cranial CT of Case 1 done on admission. No abnormalities can be detected. (b) MRI of the brain done in the same child a day later showing small bifrontal subdural fluid collections (arrow). On retrospect these collections were present on MRI done at the night of admission. DISCUSSION Non-accidental head injury is the most common cause of head trauma in infants and young children in the United States.7 Neurologic signs are the most frequent cause for admission of these children.6 However, in the absence of clear evidence for intracranial damage, the diagnosis of SBS may be challenging. In small children, the presenting symptoms of SBS may mimic gastroenteritis, viral illness, encephalitis, and convulsive disorders.8,9 Although detec- 448 Journal of AAPOS Volume 8 Number 5 October 2004 However, reports by Cohen et al (37 children),14 Dias et al (33 infants),15 Kivlin et al (111 children),4 and Morad et al (75 children),6 have failed to find any child with normal CT on admission. With the exception of the paper by Feldman’s group, one must also consider that the other prior reports all come from the earlier days of CT scanning, when subtle signs of intracranial hemorrhage may have been missed. In our cases imaged with CT technology, we suggest other possible explanations. CT Performed in the Hyperacute Phase Because CT is usually readily available, it may be performed in the hyper-acute phase, when clear evidence of intracranial bleeding is not yet apparent. We believe this is demonstrated in our cases 1 and 7, who had a CT and MRI done in the Emergency Department. In these cases, MRI done 5 to 7 days after admission revealed intracranial blood that was not clearly evident in the initial CT and MRI. We believe that during the intervening time blood continued to accumulate until it was demonstrated by MRI. Chabrol and coworkers report a similar chain of events in their description of the role of MRI in detecting children who were abused.16 Giangiacomo and Barkett also reported a child with SBS that had a normal CT scan on admission but with clear evidence of ventricular enlargement and subdural effusion in a CT done 2 weeks later.17 Feldman and coworkers described 34 cases of children diagnosed with SBS,12 7 of whom had normal CT on presentation. Of these, three children who did not improve clinically had another cranial CT later in the course of treatment. These scans disclosed subdural or subarachnoid hemorrhages that were not present on the first CT scan. Subtle Intracranial Damage not Shown on CT FIG 2. (a) Cranial CT of Case 2 done on admission, showing no abnormalities. (b) MRI done 5 days later disclosed small CSFintensity hygroma in the right frontal–parietal convexity with small foci of associated acute blood products (arrows). tion of retinal hemorrhages is extremely helpful, the documentation of intracranial injury is believed by many to be mandatory for establishing the diagnosis. According to the American Academy of Pediatrics, CT imaging without contrast should be performed as part of the initial evaluation for any child with suspected inflicted head injury.10 When the CT scan is interpreted as normal and the only manifestation of child abuse is retinal hemorrhages, the diagnosis of SBS may be questioned or missed. In this study we described nine cases in which a diagnosis of SBS was made despite a reportedly normal head CT on admission. In review of the literature, we have found other similar reports. In 1987, Sinal and Ball found 3/22 (13%) children with SBS to have normal head CT.11 Feldman and coworkers reported a similar rate of 4/34 (11.7%).12 Sato and coworkers had reported a higher rate of 6/19 (31%), three of whom had an abnormal MRI.13 Four of our nine infants had no apparent neurological damage on follow-up examination, and only one needed surgical intervention. Outcome studies of SBS victims suggest that over two-thirds either die or sustain neurologic sequelae.9 It may be that in our group of patients the inflicted damage was not severe enough to cause damage that is recognizable on admission CT. This observation was also made in Feldman’s report in which four children who had normal CT recovered completely.12 However, this is not uniformly the case. Green and coworkers in their series of autopsies done in children who died from SBS report on a child that had only signs of ocular trauma, even at post-mortem examination.18 Injuries Detectable by MRI but not CT Eight of our nine patients had MRI scanning. In cases 1, 3, 7, and 8, an MRI done within 24 hours of admission did not add any clinical information. Although a repeated MRI done several days later was still normal in case 3, in case 7 it revealed subdural hemorrhage, and in case 1, bifrontal subdural fluid were detected. Journal of AAPOS Volume 8 Number 5 October 2004 Sato and coworkers13 compared the usefulness of CT to MRI imaging in child abuse. Of 19 patients studied, 13 had intracranial abnormalities demonstrated by both methods; 3 patients did not have any abnormality in either method and 3 had abnormal MRI imaging but normal CT. They found MRI to be better than CT in detecting SDH, as CT did not detect more than 70% of small or transversely oriented hematomas. Other authors also regard MRI as the method of choice for detection of small subacute and chronic SDH.19 SDH is the most common intracranial finding in SBS.4,6 Small SDH may be missed by CT especially if is thin and lying close against the inner table, above or under the brain, because of volume averaging of the hematoma with the overlying bone.20 SDH may also be obscured by the beam-hardening effect commonly present in the posterior cranial fossa.13,14 In the acute phase, the density of these hematomas is close to that of the CSF or the brain, making them even more difficult to detect, especially in a young infant, which normally has a conspicuous subarachnoid space.20 Another advantage of MRI is the ability to more accurately date blood accumulation. The presence of an old collection of blood may have unique forensic importance in establishing the diagnosis of SBS. It has been recommended that the MRI be performed five to seven days after the injury, since earlier scans may fail to detect acute subarachnoid or subdural hemorrhages.10 MRI is also superior to CT in detecting non-hemorrhagic intra-axial injuries such as cortical contusion or shearing. Diffusion-weighted MRI is a new noninvasive imaging technique that can be performed in addition to conventional MRI. It may be useful in the diagnosis of acute cerebral ischemia and may be more sensitive than conventional MRI in the detection of diffuse axonal injury.21,22 Recently, Biousse et al reported the use of this imaging technique in SBS.23 Apart from intracranial hemorrhage that was seen in all cases, this new MRI technique detected parenchymal lesions, caused by tissue ischemia, which can be missed by conventional CT and MRI. It may be that, in cases such as ours, diffusion-weighted MRI could be useful. Sonography via the anterior fontanelle can also be useful in some cases of SBS. This technique can demonstrate subcortical white matter tears in the frontal and anterior parasagittal regions.24 It may also be helpful in detecting and differentiating subdural collections.25 However, the American Academy of Pediatrics cautioned that because sonography is insensitive in detecting small acute SDHs,26 it should be performed only in conjunction with CT, MRI, or both.10 In reviewing the history given by all the caregivers of the children in our study, it is interesting to note that none reported that the children suffered from any trauma before admission. This phenomenon was noted by others: only 49% of the 75 children reported by Morad et al had any 449 report of trauma before admission,6 and 83% of the children with SBS and SDH reported by Feldman et al were reported to have minor fall or no trauma at all.12 Hettler and Greenes found that, in children with head injury, no report of trauma to be highly specific and predictive for child abuse, up to 100% specificity and positive predictive value in the subgroup of children who had persistent neurologic damage at hospital discharge.28 Papilledema was noted in only one of our patients. This compares with previous reports that also noted papilledema to be uncommon in SBS.6 Previous studies had suggested that a correlation may exist between the severity of the intracranial damage and the extent of retinal hemorrhages.29 Although our study group is small, we have found a similar trend. All the children who had abnormal neurologic outcome had extensive bilateral retinal hemorrhages, while three of the four children with normal neurologic outcome had a milder ocular injury. MRI detected delayed intracranial hemorrhage in four patients, three of whom had an abnormal neurological outcome. In contrast, only one of five of the babies with normal MRI had neurologic damage. Although our sample is small for statistical evaluation, this trend speaks for better prognosis when MRI is normal. What is the preferred mode of action for establishing the diagnosis of SBS when CT is interpreted as normal on admission? CT is only one part of the diagnostic work-up in these children. Abnormal ocular examination, the presence of fractures, or other signs of trauma on physical examination help in establishing the diagnosis of SBS. Evaluation of the CT scan by an experienced pediatric neuroradiologist may detect findings not recognized by those less familiar with the pediatric brain and SBS. If CT is again interpreted as normal, an MRI scan may establish the diagnosis. This ideally should be done on days 5 to 7. Cranial sonography may also add useful information. If the child continues to be neurologically unstable, a repeated CT may show intracranial abnormalities that were not apparent on admission. In conclusion, our report confirms that the diagnosis of SBS can be made in children with normal CT imaging. Negative CT or MRI does not rule out the diagnosis of brain injury caused by SBS, which can occur in the absence of blood on neuroimaging. The reason for the abnormal scan may be the result of timing, misinterpretation by a radiologist, or insensitivity of the CT scan and/or MRI. Our report also emphasizes the importance of diagnosing retinal hemorrhages in cases in which head imaging is normal and the ocular examination in combination with the neurologic examination are the only signs of injury to the child. References 1. Guthkelch AN. Infantile subdural haematoma and its relationship to whiplash injuries. BMJ 1971;2(759):430-1. Journal of AAPOS Volume 8 Number 5 October 2004 450 2. Caffey J. On the theory and practice of shaking infants. Its potential residual effects of permanent brain damage and mental retardation. Am J Dis Child 1972;124(2):161-9. 3. Levin AV. Retinal haemorrhages and child abuse. Vol. 18. London: Churchill Livingston, 2000. p. 151-219. 4. Kivlin JD, Simons KB, Lazoritz S, Ruttum MS. Shaken baby syndrome. Ophthalmology 2000;107(7):1246-54. 5. Kocher MS, Kasser JR. Orthopaedic aspects of child abuse. J Am Acad Orthop Surg 2000;8(1):10-20. 6. Morad Y, Kim YM, Armstrong DC, et al. Correlation between retinal abnormalities and intracranial abnormalities in the shaken baby syndrome. Am J Ophthalmol 2002;134(3):354-9. 7. Childhood Injuries in the United States. Division of Injury Control, Center for Environmental Health and Injury Control, Centers for Disease Control. Am J Dis Child 1990;144(6):627-46. 8. Ludwig S, Warman M. Shaken baby syndrome: a review of 20 cases. Ann Emerg Med 1984;13(2):104-7. 9. King WJ, MacKay M, Sirnick A. Shaken baby syndrome in Canada: clinical characteristics and outcomes of hospital cases. CMAJ 2003; 168(2):155-9. 10. American Academy of Pediatrics. Section on Radiology: Diagnostic imaging of child abuse. Pediatrics 1991;87(2):262-4. 11. Sinal SH, Ball MR. Head trauma due to child abuse: serial computerized tomography in diagnosis and management. South Med J 1987;80(12):1505-12. 12. Feldman KW, Brewer DK, Shaw DW. Evolution of the cranial computed tomography scan in child abuse. Child Abuse Negl 1995; 19(3):307-14. 13. Sato Y, Yuh WT, Smith WL, et al. Head injury in child abuse: evaluation with MR imaging. Radiology 1989;173(3):653-7. 14. Cohen RA, Kaufman RA, Myers PA, Towbin RB. Cranial computed tomography in the abused child with head injury. AJR Am J Roentgenol 1986;146(1):97-102. 15. Dias MS, Backstrom J, Falk M, Li V. Serial radiography in the infant shaken impact syndrome. Pediatr Neurosurg 1998;29(2):77-85. 16. Chabrol B, Decarie JC, Fortin G. The role of cranial MRI in identifying patients suffering from child abuse and presenting with unexplained neurological findings. Child Abuse Negl 1999;23(3): 217-28. 17. Giangiacomo J, Barkett KJ. Ophthalmoscopic findings in occult child abuse. J Pediatr Ophthalmol Strabismus 1985;22(6):234-7. 18. Green M. Ocular and cerebral trauma in non-accidental injury in infancy: underlying mechanisms and implication for paediatric practice. Br J Ophthalmol 1996;80:300-7. 19. Bradley WG Jr. MR appearance of hemorrhage in the brain. Radiology 1993;189(1):15-26. 20. Chen CY, Zimmerman RA, Rorke LB. Neuroimaging in child abuse: a mechanism-based approach. Neuroradiology 1999;41(10):711-22. 21. Cowan FM, Pennock JM, Hanrahan JD, et al. Early detection of cerebral infarction and hypoxic ischemic encephalopathy in neonates using diffusion-weighted magnetic resonance imaging. Neuropediatrics 1994;25(4):172-5. 22. Suh DY, Davis PC, Hopkins KL, et al. Nonaccidental pediatric head injury: diffusion-weighted imaging findings. Neurosurgery 2001;49(2): 309-18; discussion 18-20. 23. Biousse V, Suh DY, Newman NJ, et al. Diffusion-weighted magnetic resonance imaging in Shaken Baby Syndrome. Am J Ophthalmol 2002;133(2):249-55. 24. Jaspan T, Narborough G, Punt JA, Lowe J. Cerebral contusional tears as a marker of child abuse— detection by cranial sonography. Pediatr Radiol 1992;22(4):237-45. 25. Chen CY, Chou TY, Zimmerman RA, et al. Pericerebral fluid collection: differentiation of enlarged subarachnoid spaces from subdural collections with color Doppler US. Radiology 1996;201(2):389-92. 26. Ball WS Jr. Nonaccidental craniocerebral trauma (child abuse): MR imaging. Radiology 1989;173(3):609-10. 27. Feldman KW, Bethel R, Shugerman RP, et al. The cause of infant and toddler subdural hemorrhage: a prospective study. Pediatrics 2001;108(3):636-46. 28. Hettler J, Greenes DS. Can the initial history predict whether a child with a head injury has been abused? Pediatrics 2003;111(3):602-7. 29. Wilkinson WS, Han DP, Rappley MD, Owings CL. Retinal hemorrhage predicts neurologic injury in the shaken baby syndrome. Arch Ophthalmol 1989;107(10):1472-4. An Eye on the Arts – The Arts on the Eye On the eye chart I make out the big E, and from then on it’s guesswork. The doctor procures lens after lens from the cabinets as if they were mounted butterflies and holds them to my face. I see roseate rings of color, rainbow reflections from the thick glass, but try though I may, I can’t climb down those oscillating tunnels of promised clarity. Throughout the exam I am talked about in the third person: “He doesn’t seem to distinguish between lens one and two, let’s not change his prescription for now.” —Stephen Kuusisto (from Planet of the Blind)