Purchase this article with an account.
J. Ng, M. A. Bearse, Jr., Y. Han, M. E. Schneck, S. Barez, A. J. Adams; Prediction of Local Diabetic Retinopathy Occurence by Multifocal ERG Delays Over 3 Years. Invest. Ophthalmol. Vis. Sci. 2007;48(13):5008.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Previous 1-year and 2-year studies of the multifocal electroretinogram (mfERG) have shown mfERG implicit time (IT) delays to be a risk factor and significant predictor of the development of nonproliferative diabetic retinopathy (NPDR). Here we investigate the predictive utility of these IT delays over a 3-year period in patients with early NPDR.
A single-center observational study was conducted with 22 diabetics examined at baseline (T0) and for 3 annual follow-ups (T1, T2, T3). Ophthalmic examinations, including 50 deg fundus photos, and mfERG testing were performed at each visit. 35 retinal zones were constructed from the mfERG stimulus array and each zone was assigned the maximum implicit time (IT) Z-score within it based on data from 30 age-similar control subjects. Logistic regression, incorporating generalized estimating equations, was used to investigate the development of "sustained" (evident at more than follow-up) retinopathy (in one randomly chosen eye) in relation to mfERG IT delays and subjects' diabetic characteristics. Empirical ROC curves were used to evaluate the performance of the models. Area under the curve (AUC), sensitivity, and specificity were determined for each curve.
Cumulatively, 44 zones developed retinopathy (predominately microaneurysms/dot hemorrhages) over the course of the study, 12 of which had sustained retinopathy. Transient DR was not predicted well using mfERG delays. In univariate analyses of sustained DR, baseline IT alone was a significant predictor (p = 0.005), with an ROC AUC of 0.80, sensitivity of 75%, and specificity of 74%. Multivariate analyses yielded the following significant predictors: baseline IT (p = 0.04), diabetic duration (p = 0.01) and type (p < 0.01), and blood glucose concentration (p < 0.01). ROC analysis of the multivariate data resulted in an AUC of 0.94, sensitivity of 92%, and specificity of 95%.
The results show a multivariate model incorporating mfERG IT delays accurately predicts the retinal locations of the development of diabetic retinopathy that persists for 2 or more years. Transient DR, of less clinical significance, was not accurately predicted over the 3-year study period. Multifocal ERG delays are promising candidate measures for trials of novel therapeutics directed at preventing or slowing the progression of NPDR.
This PDF is available to Subscribers Only