Anna Brouwer

ERG abnormalities in non-anterior childhood uveitis barrier, caused by inflammation. 23 Although the prolonged cone b-wave and the 30 Hz flicker response indicate inner retinal dysfunction, amplitudes were mostly normal, as were b/a wave ratios. In addition to the abnormal cone ERGs, we also found some abnormalities in the dark-adapted ERG. Although it may be interesting to investigate these abnormalities in more detail, in this study we decided to focus on the more profound and more frequent abnormalities of the cone ERG. Here we saw a consistent and recognizable pattern in the prolonged cone b-wave implicit time and the abnormal 30 Hz flicker response. Our study is the first one that describes ERG changes in childhood uveitis by using an extended ISCEV based protocol, with a greater range of stimulus strengths than the ISCEV standard protocol. The abnormal timing in the cone b-wave, which we found in our study, was most profound at lower stimulus strengths (0.3 cds/m 2 LA and 1.0 cds/m 2 LA), and therefore may not have been discovered by using the standard protocol only. Previous reports on intermediate uveitis and childhood uveitis mostly describe abnormalities in amplitudes but rarely describe implicit times. 24–26 Shamshinova et al. found a subnormal ERG response in 75% of eyes in childhood uveitis, including all anatomic subtypes. ERG abnormalities were more frequently seen when the macula was affected and in non-anterior uveitis. 9 In accordance with our findings, abnormalities of the 30 Hz flicker response in intermediate uveitis have been described. However, abnormal implicit times of the combined rod-cone response have also been reported, whereas we mostly found abnormal cone b-wave implicit times. 24 Other studies on intermediate uveitis mainly describe differences in ERG amplitudes and do not mention implicit times. 25,26 There are several limitations to this study. Due to the retrospective design and limited sample size, weak associations may not have been found. Since pediatric uveitis is not a common entity, we were unable to include more patients. Additionally, we were unable to correct for paired eyes, which would have been preferable since most patients had a bilateral uveitis. We were unable to perform a generalized estimating equation (GEE), since we had a complete separation of data in multiple variables, including CME and the amount of vitreous cells. By using only one eye per patient, we would have discarded almost half of the limited amount of data. Therefore, we decided to perform and present both the analysis with all eyes, and the analysis with only one eye per patient. We did not find statistically significant associations with ERG abnormalities and FA scores in both analyses. We also could not correlate ERG abnormalities to visual field defects, as visual fields were only assessed in a minority of children and were often made a long time before the ERG was performed. 119 5