Anna Brouwer

Yamamoto et al. described that in Vogt-Koyanagi-Harada uveitis the RNFL and the inner retina (RNFL, GCL and IPL combined) were significantly thicker in the active stage of the disease. 20 The correlation between an active uveitis and a thicker RNFL has also been described in several other studies, and even in uveitis eyes with glaucoma. 21–24 In the current study, eyes with a prolonged cone b-IT had also a thicker RNFL compared to eyes with a normal cone b-IT. In our study vitritis and FA score were also correlated to the RNFL (data not shown). The combination of a thinner IPL and a thicker INLwas not found in other studies, because these layers were not analyzed separately. When we analyzed the combined thickness of the IPL and INL, a significant difference was only observed in region 1 (thicker) and region 6 (thinner). In anterior uveitis Lee et al. also found no difference in the combined thickness of the IPL and GCL between the active and inactive phase of uveitis. 24 This highlights that it is important to evaluate the retinal layers individually, because otherwise subtle changes can be overlooked. We saw a clearer association between a prolonged cone b-IT and a thinner OPR. In birdshot uveitis, similar findings were described, i.e. birdshot eyes had a thinner photoreceptor outer segment, which was even more affected in eyes with an abnormal ERG, which may be an indication of a more severe form of birdshot uveitis. 25 A limitation of this study was that no comparison was made in retinal layer thickness between uveitis patients and healthy controls, because we do not have a database with normal values for the thickness of every retinal layer. Non-uveitis eyes of patients with unilateral uveitis are less suitable for comparison, because it may be difficult to be absolutely certain that subtle subclinical changes are absent. However, analysis of these non-uveitis eyes showed that the biggest differences in retinal thickness were between non-uveitis eyes and uveitis eyes with a prolonged cone b-IT. Another limitation of this study was that OCT scans were analyzed using cube data. These have a lower resolution than high definition scans. However, the software program which we used only facilitates the export of thickness analysis data of scans which correspond with a 3D image. As we used a large sample size, we do not expect that a change in thickness would be in favor of one layer or the other, and therefore we expect that these results are representative. Besides the quality of the scans, it is also important to note that the ERG measures the response of the entire retina, whereas the OCT only evaluates the central retina of approximately 6 x 6 mm. This should be kept in mind when imaging results are investigated in relation to functional data. Therefore, the thickening of the INL might be of less consequence on the ERG, because it was only observed in the central regions of the ETDRS grid and the macula contributes only minimally to the full- field ERG. 26 Contrary to the INL, the thinning of the IPL was present in the peripheral 104 Chapter 4