Do retinal hemorrhages show that an infant was abused?
No. Retinal hemorrhages, retinoschisis, macular folds and optic nerve sheath hemorrhages cannot be used to diagnose head trauma, to exclude it, or to identify an intent to harm. They occur in many traumatic and non-traumatic conditions, and the theory linking them to shaking has never been validated. Their presence does not prove abuse; their absence does not exclude it.
Medically reviewed by Evan W. Matshes, MD, FRCPC, D-ABP, Medical Director and Chief Forensic Pathologist · Updated
- Ocular findings cannot diagnose or eliminate inflicted head injury, and they cannot show intent. NAAG does not use them to diagnose or exclude child abuse.
- The belief that retinal hemorrhages mean abuse grew from hypotheses of the 1970s, and was taught as near-certainty before it was ever tested.
- Vitreoretinal traction, the theory behind that belief, was proposed in 1986 and has not been validated. In a 2017 study, 50 piglets subjected to rotations at levels similar to abusive shaking showed no ocular injury.
- Retinal hemorrhages of every size and severity, including retinoschisis and macular folds, are found with raised intracranial pressure, intracranial hemorrhage, bleeding disorders, infection, accidents, natural disease and birth.
- Much of the literature linking the eyes to abuse is circular, and a 2016 Swedish systematic review found insufficient evidence to assess the diagnostic accuracy of the triad.
What is the retina, and what is a retinal hemorrhage?
The retina is a thin layer of brain tissue, about half a millimeter thick, that lines the inside of the back of the eye. It senses light and sends signals to the brain through the optic nerve. When the small retinal blood vessels bleed, they leave specks of blood called retinal hemorrhages. These lie at different depths (layers) of the retina and may extend from the back of the eye (the posterior pole) toward its front edge (the ora serrata).
Related findings include retinoschisis, in which the layers of the retina separate, macular folds, which are folds in the central retina, and optic nerve sheath hemorrhages, which are bleeding around the optic nerve. None of these can be seen by simply looking at a child. They are found by a clinical eye examination, or at autopsy by removing and dissecting the eyes.
Where did the belief that retinal hemorrhages mean abuse come from?
In the early 1970s, the neurosurgeon Norman Guthkelch (1971) and the radiologist John Caffey (1974) proposed that shaking could cause bleeding inside the head and in the retina. Neither claimed that these findings could only be caused by shaking or abuse.
The hypothesis nonetheless hardened into dogma. A 1979 paper described retinal hemorrhage in children under three as “pathognomonic” of the battered child syndrome. By the 1980s, “shaken baby syndrome” was accepted in pediatrics, and by the late 1990s the triad of subdural hemorrhage, retinal hemorrhage and encephalopathy was described as virtually diagnostic, in the absence of severe accidental trauma. Statements of that kind supported expert testimony in criminal and family courts for decades.
Is the vitreoretinal traction theory proven?
No. The theory, proposed in 1986, holds that during shaking the vitreous gel and the retina move at different speeds, so that the vitreous pulls on the retina and tears its vessels. It was adopted by the child abuse and ophthalmology communities without being tested by any validated method. Animal experiments have generally failed to support it: a 2017 study of 50 piglets subjected to rotations at levels similar to abusive shaking reported no ocular injury. In 2001 the National Association of Medical Examiners’ position paper on shaken baby syndrome itself acknowledged that the mechanism by which retinal hemorrhages, retinoschisis and macular folds develop is unknown.
The alternative explanation is much older and well supported. Since the early 1900s, retinal hemorrhage has been recognized as a consequence of raised intracranial pressure and intracranial hemorrhage, and a 1957 study reproduced it experimentally in primates by raising intracranial pressure. A 1958 Mayo Clinic study found retinal or subhyaloid hemorrhage, often severe, in 51 percent of young children with intracranial hemorrhage of any cause. On this view, retinal hemorrhage is a secondary finding: a consequence of raised intracranial pressure, intracranial bleeding, changes in blood flow, bleeding and clotting disorders, lack of oxygen, and time on life support, which can occur whether or not there was trauma.
Are “severe” hemorrhages, retinoschisis or macular folds specific for abuse?
The current version of the hypothesis accepts that mild or moderate hemorrhages have many causes, but holds that “severe” hemorrhages (multilayered, too numerous to count, and extending to the ora serrata), retinoschisis and macular folds are highly specific for abuse. The evidence does not support that refinement:
- Retinal hemorrhages of all sizes, shapes, locations and severity have been reported after accidental trauma, including short falls and crush injuries, and in non-traumatic settings, including severe infection, coagulopathy, ruptured aneurysm, high altitude and normal birth.
- In a general and forensic autopsy service, Lantz and colleagues found retinal hemorrhages of all appearances and severity in about one quarter of both infant and adult cases.
- A 2010 study at a major metropolitan medical examiner’s office, presented by Dr. Matshes, found retinal and optic nerve sheath hemorrhages, including severe ones, across abuse, accidental and intentional blunt trauma, accidental drowning, and natural disease after critical care.
- The belief that macular folds are pathognomonic of shaking rested, as a 2007 study found, on a total of seven cases. Folds and retinoschisis have since been shown to develop over time in hospitalized patients, including cases with no trauma at all.
- A 2017 study concluded that there is no pathognomonic size, distribution or location of retinal hemorrhage in abusive head trauma.
The Royal College of Ophthalmologists has listed other causes including leukemia, hemorrhagic disease of the newborn, sickle cell retinopathy, extracorporeal membrane oxygenation (ECMO), metabolic disease, meningitis, intracranial vascular malformations, protein C deficiency, von Willebrand disease and vaginal delivery.
Why is the supporting research unreliable, and where does the field stand now?
The studies said to validate the eye findings were carried out after shaken baby syndrome was already accepted. Children with subdural and retinal hemorrhages and no history of major trauma were classified as abused because of those findings, and the studies then reported a strong association between those findings and abuse. A hypothesis cannot be validated by a classification that assumes it. A 2016 systematic review by the Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU) found the evidence linking the triad to shaking to be of low quality, and found insufficient evidence on which to assess the triad’s diagnostic accuracy.
Two practical problems add to this. Eyes are usually examined only after a child reaches hospital already compromised, when hypoxia, seizures, raised intracranial pressure, clotting abnormalities, resuscitation and life support may all have contributed, and no method separates their effects. And when ophthalmologists reviewed the same retinal photographs in a blinded study, they varied widely, even in describing what they saw. Ophthalmologists can find and describe retinal hemorrhages; that does not tell them how the hemorrhages arose.
The professions remain divided. In a 2010 survey of National Association of Medical Examiners fellows, about a third did not believe that shaking was a mechanism that could cause lethal injury, and in a 2016 survey only 40.7 percent of forensic pathologists accepted shaken baby syndrome as a valid diagnosis. A 2018 consensus statement from clinical pediatrics, ophthalmology and radiology described severe retinal hemorrhage as specific for abusive head trauma, and drew detailed published rebuttals. The American Academy of Pediatrics’ 2018 clinical report states that although retinal hemorrhages are an important indicator of possible abusive head trauma, “they are also found in other conditions”. At least one trial court, in New Jersey in 2018, found that the vitreoretinal traction theory did not meet the Frye standard of general acceptance. This is general information, not legal advice; admissibility depends on the jurisdiction and the case.
What should a forensic eye examination include?
When an infant or young child dies and the eyes may matter, the examination should include removal of both eyes with their optic nerves, professional photography before and after sectioning, and microscopic examination, including iron stains, by a pathologist experienced in pediatric eyes. The findings must be read against the whole clinical course: the birth history, intracranial pathology, intracranial pressure, oxygenation, coagulation studies, resuscitation and time on life support. On their own, they cannot answer whether the child was injured, how, or by whom.
NAAG provides forensic ophthalmology for formalin-fixed pediatric eyes, independent second-opinion review of ophthalmology records, retinal photographs, autopsy findings and slides, and Deep Dive re-examination of retained tissue. To request a review, use Submit a case.
Further reading. Matshes E, Papetti R. Law, child abuse, and the retina. The Champion (NACDL), November 2018. Matshes E. Retinal and optic nerve sheath hemorrhages are not pathognomonic of abusive head injury. Proc Am Acad Forensic Sci 2010;16:272. SBU. Traumatic shaking: the role of the triad in medical investigations of suspected traumatic shaking. Report 255E, 2016. Christian CW, Levin AV, et al. The eye examination in the evaluation of child abuse. Pediatrics 2018;142(2).
This guide is general medical information. It is not medical or legal advice about any particular child or case.
Frequently asked questions.
Do retinal hemorrhages prove that a baby was shaken or abused?
No. Retinal hemorrhages, including severe ones, occur in many traumatic and non-traumatic conditions. They cannot be used to diagnose physical head trauma, to eliminate it, or to identify an intent to harm.
Does the absence of retinal hemorrhages mean a child was not abused?
No. The absence of ocular hemorrhages should not be used to indicate that an infant was not abused, just as their presence should not be used to suggest that it was.
Can birth cause retinal hemorrhages?
Yes. Retinal hemorrhages are common in newborns after vaginal delivery, both spontaneous and assisted, and are listed among the recognized non-abusive causes.
Can an ophthalmologist tell how retinal hemorrhages were caused?
An ophthalmologist can find and describe them, but there is no scientific basis for determining from the eyes alone whether they were caused by shaking, impact, raised intracranial pressure, disease or the effects of treatment. Blinded studies have shown wide variation between ophthalmologists in describing the same images.
Can the eyes be re-examined in an older case?
Often, if the eyes or their slides were retained. NAAG can examine formalin-fixed pediatric eyes and review existing photographs, slides and records.
Talk to a forensic pathologist
Every case is different. Tell us about yours and we will tell you honestly what an examination or review can answer.
Start with a conversation.
Tell us who you are and, in general terms, what you need. The right member of our team will respond.
San Diego, CA 92121
Please don't email case records or images. The case request form collects what we need to open your matter; we'll arrange secure transfer of records after that.
