Abstract
Purpose: :
Vigabatrin, an antiepileptic drug, works by increasing the concentration of the neurotransmitter GABA in the brain and retina. Vigabatrin is associated with toxicity in 50% of adults (visual field defects) and in 30% of infants (irreversible reduction in the 30-Hz flicker amplitude). The 30 Hz flicker amplitude is the most sensitive of the ISCEV recommended ERG responses in identifying retinal toxicity due to vigabatrin; yet at the baseline recording (before initiation of vigabatrin therapy), some infants have an abnormal 30-Hz flicker amplitude. The purpose of this study is to determine if seizure-related retinal dysfunction, identified by abnormal baseline ERG, will result in more retinal damage with vigabatrin treatment than no retinal dysfunction at baseline.
Methods: :
At The Hospital for Sick Children (HSC) infants treated with vigabatrin have a baseline ERG before drug treatment with follow up ERGs at 3 to 8 month intervals. This is a retrospective data review of all patients treated with vigabatrin from 1999 to 2011. Infants with a baseline ERG and at least two subsequent ERGs were included in the study. Abnormal flicker baseline was defined as age-corrected ERG amplitude lower than the 2.5th percentile of the lab normal control ERG response for age. Data were not included if it was determined that the baseline recording was corrupted by non-retinal artifact such as Bell’s phenomenon during recording; this was accomplished by excluding data if ERG values were abnormal at baseline but showed significant increases (greater than 0.15 log units) in the subsequent ERG recording. Development of toxicity was compared between infants with normal baselines and infants with abnormal baselines.
Results: :
One hundred and seventy-six records were evaluated. 18 files were excluded because the baseline recording was affected by non-retinal artifact. One hundred and fifty-eight records were analyzed. Of the 23 infants with abnormal baselines, 2 developed ERG toxicity (9%). In the group having normal baseline flicker amplitudes (n=135), 42 developed toxicity (31%). The development of toxicity and baseline abnormality is significant (Fisher’s exact test, p=0.0423), with the group having abnormal baseline recordings showing a lower prevalence of ERG toxicity.
Conclusions: :
Patients with underlying retinal disease are not at greater risk of retinal toxicity. To account for known variability of the 30 Hz ERG, HSC definition of toxicity requires significant (0.2 log unit) reduction of the ERG from baseline on 2 consecutive recordings. The data demonstrate the important of ensuring accuracy of the baseline recording and demonstrates that abnormal baseline response is not associated with increased risk of toxicity.
Keywords: electrophysiology: clinical • drug toxicity/drug effects • neuro-ophthalmology: diagnosis