MfERG implicit time as a predictor of future diabetic retinopathy development

Abstract

New preventative treatments of sight-threatening diabetic retinopathy are under development. Clinical trials of these new therapies would be much strengthened by tools that enable prediction of those eyes and retinal locations most likely to develop retinopathy during the study period. We evaluate the mfERG (multifocal electroretinogram) as a potential tool. Prior studies have demonstrated that mfERG abnormalities are seen in eyes of diabetic subjects before retinopathy is detected. And, in eyes with retinopathy, there is spatial correspondence between abnormal mfERGs and local retinopathy (Fortune et al, 1999). This longitudinal study explores the association between mfERG abnormalities and the occurrence of retinal signs one year later. “Standard” mfERGs were recorded from 11 eyes with mild background retinopathy, 11 eyes of diabetics without retinopathy and 20 age-similar controls. Control data were used to calculate Z-scores for diabetic amplitude and implicit time (IT) data at each of the 103 retinal locations tested. We focus on IT findings. Abnormal mfERG ITs were defined as Z-scores >- 2.0. Analysis I (‘association’) involved identifying sites of new retinopathy development at T1 (1 year follow-up) and then examining the likelihood that those regions had had abnormal mfERGs at T0 (baseline) compared to the likelihood for regions that remained retinopathy free. Areas that developed retinopathy were ∼ 3 times more likely to have IT delays. Analysis II (‘prediction’) classified areas (without retinopathy) on the basis of normal vs. abnormal IT at T0. Areas with abnormal ITs at T0 were 21 times as likely to have had developed retinopathy 1 year later than areas with normal mfERG ITs at T0. These findings suggest mfERG IT recorded under ‘standard’ conditions may be a valuable tool for identifying optimal candidates for clinical trial participation because of their ability to predict retinopathy development. MfERG may also aid clinical management.