Correlation between retinal abnormalities and intracranial abnormalities in the shaken baby syndrome

PURPOSE: To report correlation between retinal and intracranial abnormalities and to evaluate pathogenesis of retinal hemorrhages in the shaken baby syndrome (SBS). DESIGN: Observational case series. METHODS: Seventy-five children with apparent nonaccidental head trauma consistent with SBS had complete physical examination, complete oplithalmologic examination, neuroimaging by CT or MRI, or both, and skeletal radiographic survey. In this retrospective review, ophthalmoscopic and intracranial abnormalities were correlated. RESULTS: The age of patients ranged from 2 to 48 months (mean - SD, 10.6 +/- 10.4 months). Neuroimaging was abnormal in all 75 cases. Findings included subdural hematoma (70 children, 93%), cerebral edema (33 children, 44%), subarachnoid hemorrhage (12 children, 16%), vascular infarction (nine children, 12%), intraparenchymal blood (six children, 8%), parenchymal contusion (six children, 8%), and epidural hemorrhage (one child, 1%). Sixty-four (64/75, 85%) children had retinal abnormalities, mostly (53/64, 82%) confluent multiple hemorrhages that were subretinal, intraretinal, and preretinal in 47/64 (74%) and bilateral in 52/64 (81%). No association was found between anatomic site (left, right, or bilateral) of intracranial and retinal findings (McNemar test kappa = -0.026-0.106) or between any of the intracranial findings mentioned above and the following retinal findings: normal or abnormal retinal examination, multiple (>10) or few retinal hemorrhages (less than or equal to10), symmetric or asymmetric retinal findings, or retinoschisis (kappa = -0.127-0.104). Signs of possible increased intracranial pressure were not correlated with any retinal abnormality (kappa = -0.030.073). There was no correlation between evidence of impact trauma to the head and retinal hemorrhages (kappa = 0.058). Total Cranial Trauma Score and Total Retinal Hemorrhage Score, both indicating the severity of injury, were correlated (P = .032). CONCLUSIONS: Our study supports previous observations that the severity of retinal and intracranial injury is correlated in SBS. We cannot support the suggestions that in most children with SBS retinal bleeding is caused by sustained elevated intracranial, elevated intrathoracic pressure, direct tracking of blood from the intracranial space, or direct impact trauma. The correlation in severity of both eye and head findings may suggest, however, that retinal abnormalities are the result of mechanical shakingforces. (C) 2002 by Elsevier Science Inc. All rights reserved.