Patient demographics.

Purpose
Retinal function beyond foveal vision is not routinely examined in the clinical screening and management of diabetic retinopathy although growing evidence suggests it may precede structural changes. In this study we compare optical coherence tomography (OCT) based macular structure with function measured objectively with the ObjectiveFIELD Analyzer (OFA), and with Matrix perimetry. We did that longitudinally in Type 2 diabetes (T2D) patients with mild Diabetic Macular Oedema (DMO) with good vision and a similar number of T2D patients without DMO, to evaluate changes in retinal function more peripherally over the natural course of retinopathy.
Methods
Both eyes of 16 T2D patients (65.0 ± 10.1, 10 females), 10 with baseline DMO, were followed for up longitudinally for 27 months providing 94 data sets. Vasculopathy was assessed by fundus photography. Retinopathy was graded using to Early Treatment of Diabetic Retinopathy Study (ETDRS) guidelines. Posterior-pole OCT quantified a 64-region/eye thickness grid. Retinal function was measured with 10–2 Matrix perimetry, and the FDA-cleared OFA. Two multifocal pupillographic objective perimetry (mfPOP) variants presented 44 stimuli/eye within either the central 30° or 60° of the visual field, providing sensitivities and delays for each test-region. OCT, Matrix and 30° OFA data were mapped to a common 44 region/eye grid allowing change over time to be compared at the same retinal regions.
Results
In eyes that presented with DMO at baseline, mean retinal thickness reduced from 237 ± 25 μm to 234.2 ± 26.7 μm, while the initially non-DMO eyes significantly increased their mean thickness from 250.7 ± 24.4 μm to 255.7 ± 20.6 μm (both p<0.05). Eyes that reduced in retinal thickness over time recovered to more normal OFA sensitivities and delays (all p<0.021). Matrix perimetry quantified fewer regions that changed significantly over the 27 months, mostly presenting in the central 8 degrees.
Conclusions
Changes in retinal function measured by OFA possibly offer greater power to monitor DMO over time than Matrix perimetry data.