Researchers from the multi-institutional Examining Siblings to Recognize Abuse (ExSTRA) research network sought to determine the prevalence of retinal hemorrhages (RH) in children <10 years of age who were being evaluated because of concerns for physical abuse and did not have evidence of traumatic brain injury (TBI). Additionally, the researchers aimed to evaluate the associations of RH in children who had no evidence of TBI, but who had abnormal mental status, facial bruising, and complex or occipital skull fracture. For the study, data collected at 20 participating US centers on all children <10 years of age who underwent subspecialty consultation for physical abuse between January 15, 2010 and April 30, 2011 were reviewed. TBI was defined as any radiological evidence of intracranial trauma, including subdural, subarachnoid, or epidural hemorrhages, brain contusions, and/or brain edema. The following information about RH was collected: presence, number, distribution, number of layers involved, and presence or absence of retinoschesis (splitting of the layers of the retina). RH that were numerous, multilayered, and extended to the periphery of the retina were considered to be “characteristic” for abuse.

Of the 1,122 eligible patients without TBI, 352 (32%) had dedicated examinations for RH, including 198 with complex or occipital skull fracture, altered mental status, or facial bruising. The median age of the 352 study children was 8.5 months (range 2–92 months). Among the 352 children, only 2 (0.6%; 95% CI, 0.1–2.0) had any RH identified. Neither child had “characteristic” RH that were thought to be “forensically significant” or retinoschesis. The authors calculate that it would require approximately 176 dedicated retinal examinations to detect 1 child with RH in this population, and that the RH detected would not likely be forensically significant.

The authors conclude that children being evaluated for physical abuse without TBI on neuroimaging may not require dedicated retinal examinations.