Anna Brouwer

Summary, discussion & future perspectives were borderline normal, i.e. the ERG was within our reference values, but retinal function may have already been somewhat affected by uveitis. Therefore, it would be interesting to investigate three things. First, to investigate if the ERG or the OCT is more sensitive and specific in determining when the retinal function, or retinal structure is abnormal (i.e. are the results of a first test within or outside reference values?). Second, to investigate if the ERG or the OCT is better at monitoring the retinal function, or retinal structure during the disease (i.e. are the results of a second (or third etc.) test significantly different from the first test?). Third, to investigate which of these two tests is best correlated to clinically significant outcomes such as visual function, or the need for systemic treatment. For each of these questions different aspects of a test are important. The first question can be seen as a diagnostic question: is a test abnormal or not? For this question the range of reference values should ideally be small. Regarding the ERG, the implicit times, including the cone b-IT, are probably better suited for this than amplitudes, because the range of implicit times is quite small, whereas the range of amplitudes is quite large. However, if a disease does not affect the implicit times, but only the amplitudes, amplitudes would of course be better for diagnostic purposes. We have no reference values of thicknesses of the INL and IPL on OCT of normal eyes. Therefore, it remains speculative if theOCT would be well suited for this purpose. The second question can be seen as a way to determine how well a test is suited for monitoring. For monitoring it is important that the measurement error is small. A small measurement error makes it easier to determine that a change in a test result is due to an effect of treatment, or disease progression and not due to a measurement error. The cone b-IT may be less ideal for this, because in Chapter 7 we observed that the difference in implicit times between the two DTL electrode positions that we used, was quite large, especially compared to the range of reference values. This implies that the repeatability of the cone b-IT is not very good, which means subtle changes in the cone b-IT and other ERG parameters might be overlooked. Fortunately the variability between the two DTL positions did not differ, so a shift in position would not affect the results of our study. Weather the thickness of the INL and IPL on OCT may be better suited for monitoring than the ERG is yet unknown and should be the subject of further studies. However, retinal thickness analysis of several other retinal layers shows good repeatability, 23,24 so it is probable that this will also be the case for the INL and IPL. 167 8