Abstract
Purpose: :
We have developed a model to predict, within one year, the local onset of diabetic edema (DE) in adults with diabetes and diabetic retinopathy (DR).
Methods: :
46 eyes from 23 patients with DR, and no edema, were included in the study. They were followed every six months until either DE developed or this study was concluded. The presence or absence of new DE at the final visit was the outcome measure, and data from the prior visit were used as baseline. A logistic regression model was formulated to assess the relationship between DE development and the following factors: Multifocal electroretinogram (mfERG) implicit time (IT) Z-score, mfERG amplitude (AMP) Z-score, gender, diabetes duration, type of diabetes, blood glucose, HbA1c, age, systolic blood pressure (SBP), diastolic blood pressure, and degree of retinopathy. Thirty five retinal zones were constructed from the 103 mfERG stimulus elements and each was graded for DE. Based on data from 50 control subjects, the maximum IT and minimum AMP Z-score for each zone was calculated. ROC curves from a five-fold cross-validation were used to determine the model’s predictive properties.
Results: :
Edema developed in 10 of the 23 patients (43%), 16 of 46 eyes (35%), and 83 of the 1610 retinal zones (5.2%). In 69% of DE cases, DE qualified as sight-threatening clinically significant macular edema. The mfERG AMP, mfERG IT, SBP, and gender together formed a significant model that is predictive of edema in a given zone. This model has 84 % sensitivity and 75 % specificity for prediction of DE.
Conclusions: :
mfERG, SBP, and gender together are good predictors of local edema onset in patients with DR. Thus the mfERG measure is a candidate for both assessing risk for edema development, and an outcome measure to evaluate novel therapeutics directed at sight-threatening DE.
Keywords: diabetic retinopathy • electroretinography: clinical • diabetes