Imaging Biomarkers of Mesopic and Dark-Adapted Macular Functions in Eyes With Treatment-Naïve Mild Diabetic Retinopathy

Purpose: To investigate the relationship between imaging biomarkers and mesopic and dark-adapted (ie, scotopic) functions in patients with treatment-naïve mild diabetic retinopathy (DR) and normal visual acuity. Design: Prospective cross-sectional study. Methods: In this study, 60 patients with treatment-naïve mild DR (Early Treatment of Diabetic Retinopathy Study levels 20-35) and 30 healthy control subjects underwent microperimetry, structural optical coherence tomography (OCT), and OCT angiography (OCTA). Results: The foveal mesopic (22.4 ± 4.5 dB and 25.8 ± 2.0 dB, P = .005), parafoveal mesopic (23.2 ± 3.8 and 25.8 ± 1.9, P < .0001), and parafoveal dark-adapted (21.1 ± 2.8 dB and 23.2 ± 1.9 dB, P = .003) sensitivities were reduced in DR eyes. For foveal mesopic sensitivity, the regression analysis showed a significant topographic association with choriocapillaris flow deficits percentage (CC FD%; β = -0.234, P = .046) and ellipsoid zone (EZ) normalized reflectivity (β = 0.282, P = .048). Parafoveal mesopic sensitivity was significantly topographically associated with inner retinal thickness (β = 0.253, P = .035), deep capillary plexus (DCP) vessel length density (VLD; β = 0.542, P = .016), CC FD% (β = -0.312, P = .032), and EZ normalized reflectivity (β = 0.328, P = .031). Similarly, parafoveal dark-adapted sensitivity was topographically associated with inner retinal thickness (β = 0.453, P = .021), DCP VLD (β = 0.370, P = .030), CC FD% (β = -0.282, P = .048), and EZ normalized reflectivity (β = 0.295, P = .042). Conclusions: In treatment-naïve mild DR eyes, both rod and cone functions are affected and they are associated with both DCP and CC flow impairment, which suggests that a macular hypoperfusion at these levels might implicate a reduction in photoreceptor function. Normalized EZ reflectivity may be a valuable structural biomarker for assessing photoreceptor function in DR.