Pediatric abusive head trauma: visual outcomes, evoked potentials, diffusion tensor imaging, and relationships to retinal hemorrhages.

PURPOSE: Function and anatomy of the visual system were evaluated in children with abusive head trauma (AHT). The relationships between retinal hemorrhages at presentation were examined with outcome measures.

METHODS

Retrospective review of data in children with AHT for 1) visual acuity at last follow-up, 2) visual evoked potentials (VEP) after recovery, 3) diffusion metrics of white matter tracts and grey matter within the occipital lobe on diffusion tensor imaging (DTI), and 4) patterns of retinal hemorrhages at presentation. Visual acuity was converted into logarithm of minimum angle of resolution (logMAR) after correction for age. VEPs were also scored by objective signal-to-noise ratio (SNR).

RESULTS

Of 202 AHT victims reviewed, 45 met inclusion criteria. Median logMAR was reduced to 0.8 (approximately 20/125 Snellen equivalent), with 27% having no measurable vision. Thirty-two percent of subjects had no detectable VEP signal. VEPs were significantly reduced in subjects initially presenting with traumatic retinoschisis or hemorrhages involving the macula (p < 0.01). DTI tract volumes were decreased in AHT subjects compared to controls (p < 0.001). DTI metrics were most affected in AHT victims showing macular abnormalities on follow-up ocular examination. However, DTI metrics were not correlated with visual acuity or VEPS. There was large inter-subject variability within each grouping.

DISCUSSION: Mechanisms causing traumatic retinoschisis, or traumatic abnormalities of the macula, are associated with significant long-term visual pathway dysfunction. AHT associated abnormalities of the macula, and visual cortical pathways were more fully captured by VEPs than visual acuity or DTI metrics.