Abstract
Abstract: :
Purpose:In diabetic patients with mild (grade <=2a) diabetic retinopathy : 1) to quantify the spatial correspondence of 4 measures of mfERG components to local signs on fundus photography, and 2) to quantify the degree to which various types of diabetic retinal lesions seen on fundus photos are detected by these measures. Methods: MfERGs were recorded at 103 retinal locations in 6 patients with mild diabetic retinopathy, and 11 age-similar control subjects, using a Veris 4 system and standard conditions. At each location in diabetic eyes the Z-score (standard deviation unit based on normal data) of the derived K1 and K2 scalar products (SP), and implicit times (IT) of P1 and N2 were calculated. Fundus photos were graded following a modified Airlie House system. The presence/absence of each lesion type was evaluated within 13 zones symmetrically about fixation. The maximal z-score for implicit time of P1 and N2, and minimal Z score for scalar products were identified for each zone. For each component, a zone was considered abnormal if the maximum scalar product Z score was ≷2, or implicit time Z score was <-2. Spatial coincidence between abnormal mfERG findings and presence of diabetic lesions was examined within 78 zones (6 eyes X 13 zones/eye). Results: K1 SP is insensitive to early diabetic eye disease, showing abnormalities in only 1/6 subjects and no correspondence with fundus findings in this subject. K2 SP identifies abnormal zones in 4/6 eyes, but the spatial association between mfERG and retinopathy did not reach significance (Chi square p<0.08). P1 IT identified at least 1 zone as abnormal in 5/6 eyes, but these were not significantly associated locally with the presence/absence of retinopathy. N2 IT identified multiple zones as abnormal in 5/6 eyes. Further, there is a significant association (Chi square, p<0.033) between local N2 IT changes and fundus abnormalities despite the fact that all of the clinical abnormalities were smaller than 1/2 disc diameter. N2 sensitivity to local fundus signs varied with lesion type being best for edema, soft and hard exudates but less sensitive to milder changes (e.g. 62.5% of microaneurysms were detected). Conclusion: More severe retinal changes (e.g. edema) are most likely to have an associated abnormal mfERG. Among the measures of mfERG, N2 IT shows the best local correspondence to diabetic retinal changes. However, all mfERG measures show abnormalities in regions (and diabetic eyes, data not described) without retinal signs, suggesting sensitivity to pre-clinical changes as well as the ability to predict future disease.
Keywords: 387 diabetes • 395 electroretinography: clinical • 388 diabetic retinopathy