This Clinical Report was revised. See https://doi.org/10.1542/peds.2018-1411.
Retinal hemorrhage is an important indicator of possible abusive head trauma, but it is also found in a number of other conditions. Distinguishing the type, number, and pattern of retinal hemorrhages may be helpful in establishing a differential diagnosis. Identification of ocular abnormalities requires a full retinal examination by an ophthalmologist using indirect ophthalmoscopy through a pupil that has been pharmacologically dilated. At autopsy, removal of the eyes and orbital tissues may also reveal abnormalities not discovered before death. In previously well young children who experience unexpected apparent life-threatening events with no obvious cause, children with head trauma that results in significant intracranial hemorrhage and brain injury, victims of abusive head trauma, and children with unexplained death, premortem clinical eye examination and postmortem examination of the eyes and orbits may be helpful in detecting abnormalities that can help establish the underlying etiology.
BACKGROUND
When a previously well child experiences an apparent life-threatening event (ALTE) or unexpected death without obvious cause, pediatricians and other physicians must attempt to identify the etiology. In the case of an ALTE, one should consider diagnoses such as gastroesophageal reflux, seizures, other central nervous system disease, metabolic disease, breath-holding, and abusive head trauma (AHT). Retinal examinations have been used with limited success for screening ALTE victims for possible AHT.1,2 Victims of AHT present to medical care with a wide range of symptoms, from mild irritability and vomiting to unexplained coma or seizures.3 It has been estimated that approximately 4% to 6% of abused children present first to an ophthalmologist,4 and the most common ocular manifestation of abuse is retinal hemorrhage. Some children present with a false history of trauma, and others present with only the symptoms that resulted from their abuse. Unsuspecting physicians misdiagnose the condition of up to one-third of symptomatic victims, depending on their age, severity of symptoms, and family composition.5 When a child dies unexpectedly, considerations include previously undiagnosed or new systemic disease, sudden infant death syndrome, and covert abusive injury.
Retinal hemorrhages have been recognized as a key indicator of abusive head injury for more than 30 years, particularly in association with severe repetitive acceleration-deceleration forces with or without blunt head impact, in children younger than 5 years.6,7 Because retinal hemorrhage rarely results in visual compromise, there may be no external indicators of eye injury unless the vision is significantly impaired by damage to the optic nerve or occipital cortex or there is retinal detachment, significant vitreous hemorrhage, or other severe disruption of the intraocular contents. Both eyes need to have significant visual compromise for a child to become visually symptomatic. Therefore, one cannot rely on ocular signs or symptoms to determine which children might benefit from ophthalmologic consultation and postmortem eye removal. Likewise, retinal examinations should not be limited to patients who are victims of suspected AHT. Searching for retinal hemorrhages as diagnostic criteria for AHT only in infants with suspected abuse creates a selection bias.
Although health care professionals other than ophthalmologists may be skilled at detecting the absence or presence of retinal hemorrhage,8 a full view of the retina and characterization of the number, types, and patterns of the hemorrhages requires consultation by an opthalmologist using indirect ophthalmoscopy, preferably with a dilated pupil. Even when there may be a concern about transiently obliterating pupillary reactivity in the face of a need to monitor neurologic status acutely, techniques such as dilation of 1 eye at a time, use of short-acting mydriatics, and use of a lens that affords some view through an undilated pupil can be employed to allow indirect ophthalmoscopy, preferably within the first 24 hours and ideally within 72 hours after the child's acute presentation. Even if the need for eye examination is realized after 72 hours, ophthalmologic consultation may still be useful to identify persistent abnormalities such as hemorrhages, retinoschisis, and papilledema.
The range of retinal hemorrhage findings in AHT and accidental trauma is broad, and the severity of the findings can be informative of etiology. For example, retinal hemorrhage— predominantly intraretinal, in small numbers, and confined to the posterior pole of the retina—can be seen after significant accidental head injury or in AHT.6,9 More dramatic retinal hemorrhages—multilayered, too numerous to count, and extending to the edge of the retina (ora serrata)—can be seen after AHT, normal birth, fatal extreme accidental injury (such as motor vehicle accidents), and perhaps after fatal head crush injury.6,10,–,13 Ocular fundus examination can also reveal findings of systemic illness that may shed light on the etiology of a child's symptoms, such as a cherry-red spot in metabolic disease, retinal vascular abnormalities in Menkes disease, papilledema, and retinal manifestations of leukemia or bacterial endocarditis. There are no known retinal ocular findings in sudden infant death syndrome, although routine ocular examination has not been common practice in these cases.6,14
Excluding retinal hemorrhages that are associated with vaginal delivery, AHT is the leading cause of retinal hemorrhages in infants. The association of retinal hemorrhage and AHT has been demonstrated repeatedly in clinical studies.15,–,17 Although retinal hemorrhages in AHT can be unilateral or bilateral and vary in degree, the severity of retinal hemorrhage often parallels the severity of brain injury.18,19 Because intraretinal hemorrhages may resolve quickly, a retinal examination is not a substitute for brain imaging when screening physically abused infants who have no neurologic symptoms for AHT.20
The ophthalmologist is in a unique position to detail the hemorrhagic retinopathy. The indirect ophthalmoscope provides a wide and stereoscopic field of view and enables the ophthalmologist to examine the anterior aspects of the retina to the ora serrata, which is not possible using a direct ophthalmoscope even if the pupils are dilated. Eye examination for this purpose should be performed by an ophthalmologist. Attention must be paid to special features, such as the presence of traumatic macular retinoschisis, because these features may have particular diagnostic significance for abuse.
Autopsy is a unique opportunity for examination not only of the eye and its contents but also of the orbital tissues, which may yield findings helpful in differential diagnosis. This is particularly true when a child dies before clinical ophthalmologic consultation can be obtained. Even when premortem ophthalmoscopy is performed, postmortem examination is necessary for viewing the orbital tissues. When possible, examination by a trained ocular pathologist or ophthalmologist with experience in interpreting ocular pathology is ideal. Postmortem eye and orbital tissue examination is another means of documenting retinal hemorrhage and retinoschisis but may also reveal hemosiderin deposition from previous events and orbital findings, such as hemorrhage into the fat, muscles, and cranial nerve sheaths as well as intradural hemorrhage, all of which may have diagnostic significance in identifying abused children.21
One obstacle to postmortem examination of the eyes and orbits has been a societal distaste or resistance that, in some cases, has led to fear among pathologists of legal repercussion. This may reflect a cultural or emotional objection specifically to eye removal. There might be a misconception that eye or orbital removal will alter the appearance of the body postmortem at a funeral viewing when this is not, in fact, the case. Techniques allow for proper removal of the eye and orbits without disfigurement.6 Although consent is not routinely obtained for coroner cases/forensic autopsies, there may be situations or jurisdictions in which specific consent for eye and orbital tissue removal may be sought. If a substitute decision-maker refuses this procedure, it may be necessary to seek legal intervention to allow the procedure to be performed.
STATEMENT OF THE PROBLEM
Although a retinal examination may suggest the etiology for ALTEs and previously unexplained early childhood deaths, premortem clinical ophthalmologic consultation and postmortem removal of the eyes and orbital tissues are not routine practice in some centers. Those who fail to conduct these procedures, particularly when there is no other explanation for the life-threatening event or death, risk losing an important opportunity to gain valuable information. Information gained in such an evaluation might lead to identifying an etiology and, in the case of a surviving child, prevent death by preventing further abuse or recognizing other disease.
In recent years, 3 important trends have emerged in the understanding of hemorrhagic retinopathy in young children. First, there has been a continued and dramatic increase in the evidence supporting the diagnostic specificity of severe hemorrhagic retinopathy as an indicator of AHT, in particular, severe repeated acceleration-deceleration with or without blunt head impact. Extensive literature worldwide, including clinical studies,7 animal models,22 and computer modeling,23,24 well beyond the scope of this report, support this theory. It seems that vitreoretinal traction and orbital injury sustained during the unique repetitive acceleration-deceleration mechanism that distinguishes this form of abuse from single-impact trauma is the critical factor in causing retinal hemorrhage.8 Factors such as hypoxia, anemia, and intracranial pressure may play important secondary roles in modulating the appearance of retinal hemorrhages but do not, in and of themselves, result in such a retinopathy.6 Further research is needed to better define the role of these and other factors as our understanding of the pathophysiology and diagnostic specificity of retinal hemorrhage continues to evolve.
Second, there has been increasing recognition of the importance of detailing the appearance of the hemorrhagic retinopathy rather than simply using the term “retinal hemorrhage” generically. The differential diagnosis of a few intraretinal hemorrhages surrounding the optic nerve is vastly greater than that for “too-numerous-to-count” multilayered retinal hemorrhages that extend to the ora serrata.
Third, there continue to be novel reports of causes of retinal hemorrhage and other ocular findings in young children. These hypothesis-generating observations allow practitioners to broaden and at the same time narrow differential diagnosis. For example, mild posterior pole retinal hemorrhages have been described in osteogenesis imperfecta25 as well as head trauma sufficient to cause epidural hemorrhage,9 and more severe hemorrhage was observed in 2 cases of fatal head crush injury.10,11 Only with more widespread use of ophthalmologic consultation and postmortem ocular and orbital examination can such entities be discovered and fit into the differential diagnostic process appropriately. Photodocumentation, when available, has also proven to be an important ophthalmic procedure for documenting retinal abnormalities for both clinical and educational purposes.
Full ophthalmic examination by an ophthalmologist and/or postmortem eye removal and pathologic examination can be a critical part of the diagnostic evaluation of previously well children who have experienced an unexplained ALTE or death and also for children who have disorders in which there are known associations with ophthalmic findings, particularly AHT.
GUIDANCE FOR PEDIATRICIANS
Ophthalmologic consultation should be part of the diagnostic evaluation of all previously well children younger than 5 years who experience an unexplained death or unexplained ALTE or have a systemic disorder known to have ocular manifestations.
All infants and young children who present with significant intracranial hemorrhage should have an ophthalmologic consultation. Retinal examination is particularly important when there is a suspicion of AHT, but some retinal hemorrhages may occasionally also be found in infants and children with other causes of intracranial hemorrhage.
Ophthalmologic examination should include, whenever possible, full indirect ophthalmoscopic examination through a dilated pupil. Descriptions of findings should be detailed and include the number, type, extent, and pattern of retinal hemorrhages, if present. Retinal abnormalities may be photographed after the clinical examination when a camera is available. When indicated, the examination should include slit-lamp inspection of the anterior segment to identify signs of trauma (eg, hyphema).
Because findings such as retinal hemorrhage may be transient, it is desirable that the ophthalmologic consultation take place preferably within 24 hours after the patient presents for medical care and ideally within 72 hours.
A retinal examination is not an appropriate screening test for brain injury in neurologically asymptomatic potential victims of abuse. Such children should undergo brain imaging as the appropriate screen.
When pharmacologic dilation is felt to be undesirable, as for children with severe unstable central nervous system injury, timely ophthalmologic consultation is still needed. An attempt should still be made to view the retina and optic nerve through the use of direct ophthalmoscopy, small pupil indirect ophthalmoscopic techniques, sequential pharmacologic dilation, and/or fast-acting mydriatics (eg, phenylephrine 2.5%).
When a previously well child younger than 5 years dies without explanation, regardless of whether a premortem retinal examination was conducted, the eyes and orbital tissues should be removed en bloc at autopsy per published techniques.26 When possible, an examination by an ocular pathologist or ophthalmologist with experience in interpreting ocular pathology is preferable.
Postmortem eye removal may not be indicated in children who have clearly died from witnessed severe accidental head trauma or otherwise readily diagnosed systemic medical conditions.
Ophthalmologic examination and/or postmortem eye and orbital tissue removal should be performed in all cases in which a child is alleged to have suffered significant morbidity secondary to a short fall or other minor trauma disproportionate to the clinical injury and consistent with child abuse.
LEAD AUTHORS
Alex, V. Levin, MD, MHSc—Former Section on Child Abuse and Neglect Executive Committee Member
Cindy W. Christian, MD
COMMITTEE ON CHILD ABUSE AND NEGLECT, 2009–2010
Carole Jenny, MD, MBA, Chairperson
Cindy W. Christian, MD
James E. Crawford-Jakublak, MD
Emalee Flaherty, MD
Roberta A. Hibbard, MD
Rich Kaplan, MD
LIAISONS
Harriet MacMillan, MD –American Academy of Child and Adolescent Psychiatry
Janet Saul, PhD – Centers for Disease Control and Prevention
STAFF
Tammy Piazza Hurley
SECTION ON OPHTHALMOLOGY EXECUTIVE COMMITTEE, 2009–2010
Gregg T. Lueder, MD, Chairperson
James B. Ruben, MD, Chairperson-Elect
Richard J. Blocker, MD
David B. Granet, MD
Daniel J. Karr, MD
Sharon S. Lehman, MD
Sebastian J. Troia, MD
LIAISONS
Kyle A. Arnoldi, CO–American Association of Certified Orthoptists
Christie L. Morse, MD–American Academy of Ophthalmology
Michael Xavier Repka, MD–American Association for Pediatric Ophthalmology and Strabismus
The guidance in this report does not indicate an exclusive course of treatment or serve as a standard of medical care. Variations, taking into account individual circumstances, may be appropriate.
All clinical reports from the American Academy of Pediatrics automatically expire 5 years after publication unless reaffirmed, revised, or retired at or before that time.
REFERENCES
Competing Interests
This document is copyrighted and is property of the American Academy of Pediatrics and its Board of Directors. All authors have filed conflict of interest statements with the American Academy of Pediatrics. Any conflicts have been resolved through a process approved by the Board of Directors. The American Academy of Pediatrics has neither solicited nor accepted any commercial involvement in the development of the content of this publication.
Comments
Re: Challenges to The Vitreous Traction Hypothesis Still Go Unanswered
As Dr. Galaznik correctly quoted from our article, the discussion he is requesting is "well beyond the scope of this report".
Conflict of Interest:
None declared
Challenges to The Vitreous Traction Hypothesis Still Go Unanswered
The American Academy Pediatrics Committee on Child Abuse and Neglect (COCAN) recently published a “Clinical Report – The Eye Examination in the Evaluation of Child Abuse”. [1] This 2010 Clinical Report represents the latest opportunity for COCAN authors to address recent challenges to the concept of vitreous traction on the retina as a valid cause of retinal hemorrhages, schisis, or perimacular folds during an inflicted acceleration/deceleration stress. [2] However, on reading it, one finds only a restating of the same assertions, with no adequate support to answer valid critiques.
Close analysis of one passage serves to reveal a major failing of the report: “First, there has been a continued and dramatic increase in the evidence supporting the diagnostic specificity of severe hemorrhagic retinopathy as an indicator of AHT, in particular, severe repeated acceleration- deceleration with or without blunt head impact. Extensive literature worldwide, including clinical studies [3], animal models [4], and computer modeling [5][6] well beyond the scope of this report support this theory. It seems that vitreoretinal traction and orbital injury sustained during the unique repetitive acceleration-deceleration mechanism that distinguishes this form of abuse from single-impact trauma is the critical factor in causing retinal hemorrhage. [7] Factors such as hypoxia, anemia, and intracranial pressure may play important secondary roles in modulating the appearance of retinal hemorrhages but do not, in and of themselves, result in such a retinopathy. [8]” This statement suggests settled science; however, study of the cited references for “the continued and dramatic increase in evidence” betrays an unsettling lapse in rigor.
For clinical studies, the COCAN authors cite Kivlin 2000, which has not gone unchallenged over the past decade. It predates Lantz 2004 [9], Goldsmith 2004 [10], Lueder 2006 [11], Obi 2007 [12], and Emerson 2007. [13] In fact, Kivlin 2008 refutes the “diagnostic specificity” for AHT, reporting documented retinal hemorrhages, schisis, folds, and optic nerve sheath hemorrhages in motor vehicle accidents where there was no reason to suspect “repetitive” acceleration/deceleration, over single-impact trauma. [14]
For animal studies, the COCAN authors cite Wygnanski-Jaffe 2007, “Protective ocular mechanisms in woodpeckers,” which documented no retinal hemorrhages in the woodpeckers. [4] The COCAN authors fail to address the important animal studies specifically cited in the 2007 P3R [2] – Brown 2007 [15] and Binenbaum 2007. [16] In the Binenbaum experiments, 3-5 day old piglets, subjected to >20 - 40 times the rotational accel/decel achievable with abusive shaking, presented no grossly visible retinal hemorrhage, schisis, or folds. Good animal research, including the cited woodpecker study, has failed, thus far, to confirm the vitreous traction hypothesis in appropriate models, but the COCAN Clinical Report implies the opposite.
For computer modeling, the authors cite Cirovic 2005 [5] and Rangarajan 2009. [6] In an oversight, they ignore Cirovic's concluding remarks: “These results are preliminary...The results should therefore not be used in support of proceedings or arguments related to determination of forces in accidental or allegedly abusive, inflicted or accidental trauma in individual human adults, children or infants, or in animals.”[5]
A similar disclaimer accompanied Rangarajan 2009 in a commentary by Dr.'s Margulies, Coats, Christian, Forbes, and Duhaime entitled, “What can we learn from computational model studies of the eye?” [6] It states: “One should be aware that inaccuracy of these inputs will yield fallacious outputs, and the results of the computer model will have little or no similarity to a real-life response. Because the material properties and tissue-tissue interactions of pediatric ocular tissues have not been measured or published, it is currently impossible for a finite element model to mimic the response of the pediatric eye.....” [18]
The 2003 article, “Non-ophthalmologists' accuracy in diagnosing retinal hemorrhages in the shaken baby syndrome” [7] is provided to support that a “unique repetitive acceleration-deceleration mechanism that distinguishes this form of abuse from single impact trauma is the critical factor in causing retinal hemorrhage.” Given that every finding previously alleged as unique to repetitive “acceleration-deceleration” has been documented in single impacts in more recent literature [9][11][12]14], this lone 2003 citation strains under the burden of COCAN’s assertion of uniqueness.
The final sentence of the quoted passage leads to further doubts about COCAN’s engagement of current evidence: “Factors such as hypoxia, anemia, and intracranial pressure may play important secondary roles in modulating the appearance of retinal hemorrhages but do not, in and of themselves, result in such a retinopathy.” [1] They support this with a publication from 2000. [8] Given that standard textbooks of ophthalmology cite anemia, high altitude exposure, and intracranial pressure as primary causes of retinopathy, along with valsalva, and coagulopathy, the authors’ relegation of these factors to secondary modulators seems more presumptive than scientific.
While it is impossible to prove that vitreoretinal traction in an alleged shaking scenario is never a primary cause of grossly visible retinal hemorrhages, schisis, or folds [i.e. prove the null hypothesis], failure to confirm vitreoretinal traction from massive rotational acceleration/deceleration in an appropriate animal model [16] should preclude an AAP Official Clinical Report from asserting that vitreoretinal traction is a valid cause of unique findings with diagnostic implications until evidence-based medical research can confirm its validity.
References: 1.Levin AV, Christian CW, and Committee on Child Abuse and Neglect, Section on Ophthalmology. Clinical Report – The Eye Examination in the Evaluation of Child Abuse. Pediatrics 2010;126(2):376-380.
2.Galaznik JG., Eye Findings and Allegations of Shaking and Non- Accidental Injury [E-letter], Pediatrics (August 8, 2007) http://pediatrics.aappublications.org/cgi/eletters/119/6/1232#13651 (accessed August 25, 2010).
3.Kivlin J, Simons K, Lazoritz S, Ruttum M. Shaken baby syndrome. Ophthalmology. 2000;107(7):1246-1254.
4.Wygnanski-Jaffe T, Murphy C, smith C, et al. Protective ocular mechanisms in woodpeckers. Eye. 2007;21(1):83-89.
5.Cirovic S, Bhola RM, et al., Mechanistic hypothesis for eye injury in infant shaking:: an experimental and computational study. Forens Sci Med Pathol. 2005;1(1):53-59.
6.Rangarajan N, Kamalakkannan SB, Hasija V, et al. Finite element model off ocular injury in shaken baby syndrome. J AAPOS. 2009;13(4):364- 369.
7.Morad Y, Kim YM, Mian M, Huyer D, Capra L, Levin AV. Non- ophthalmologists' accuracy in diagnosing retinal hemorrhages in the shaken baby syndrome. J Pediatr. 2003;142(4):431-434.
8.Levin A. Retinal haemorrhage and child abuse. In:David T, ed. Recent Advances in Paediatrics. London, England: Churchill Livingstone; 2000:151-219.
9.Lantz PE, Sinal SH, Staton CA, Weaver Jr RG. Evidence based case report: Perimacular retinal folds fromm childhood head trauma. BMJ. 2004;328:754-756.
10.Goldsmith W, Plunkett J. Biomechanical analysis of the causese off traumatic brain injury in infants and children. Am J Forensic Med Pathol. 2004;25:89-100.
11.Lueder GT, Turner JW, Paschall R. Perimacular retinal folds simulation nonaccidental injury in an infant. Arch Ophthalmol. 2006;124(12):1782-1783.
12.Watts P, Obi E. Retinal folds and retinoschisis in accidental and non-accidental head injury. Eye. 2008;22:1514-1516.
13.Emerson MV, Jakobs E, Green WR. Ocular Autopsy and Histopathologic Features of Child Abuse. Ophthalmology 2007;114:1384-1394.
14.Kivlin JD, Currie ML, Greenbaum J, Simons KB, Jentzen J. Retinal Hemorrhages in Children Following Fatal Motor Vehicle Crashes. Arch Ophthalmol. 2008;126(6):800-804.
15.Brown S, Levin AV, Ramsy D, Serbanescu I. Natural animal shakinng: A midle for inflicted neurotrauma in children? J AAPOS. 2007;11:85-86.
16.Binenbaum G, Forbes BJ, Raghupathi R, Judkins A, Rorke L, Margulies SS. An animal model to study retinal hemorrhages in nonimpact brain injury. J AAPOS. 2007;11:84-85.
17.Cirovic S, Bhola RM, et al., Mechanistic hypothesis for eye injury in infant shaking:: an experimental and computational study. Forens Sci Med Pathol. 2005;1(1):53-59.
18.Margulies SS, Coats B, Christain C, Forbes B, Duhaime AD. What can we learn from computational study? JAAPOS 2009;13:332.
Conflict of Interest:
I have given testimony in civil and criminal court cases involving allegations of physical abuse of infants and small children.
Reply to Dr. Lantz
We thank Dr Lantz for reminding the readership about this technique. Unfortunately it is inferior to clinical ophthalmoscopy of the live child which should not be deferred in anticipation of autopsy, and also inferior to gross section with photography and histologic examination. The technique mentioned by Dr. Lantz also does not allow visualization of the retrobulbar optic nerve and therefore should not be used as a substitute to proper enucleation of the globe and removal of the orbital contents along with the optic nerve. In addition, the cloudy cornea of the deceased child will undoubtedly make for a hazy view at best in many children.
Conflict of Interest:
None declared
Postmortem Ocular Examination
A forensic pathologist using postmortem monocular indirect ophthalmoscopy (1) can examine of the posterior fundus and portions of the peripheral retina of neonates, infants, and children at autopsy. The nodal angular field of view for a +20D and +30D lens is about 45º and 55º, respectively. Depending on the postmortem interval, corneal clouding (swelling) can cause the fundus to appear hazy; however, the projected aerial image is most often of adequate quality and resolution to detect retinal hemorrhages and folds associated with child abuse, accidents, and natural diseases.
1. Lantz PE, Adams GGW. Postmortem monocular indirect ophthalmoscopy. J Forensic Sci. 2005;50(6):1450-1452.
Conflict of Interest:
None declared