Abstract
Background/Objectives: Diabetic macular edema (DME) is a significant cause of vision loss. The development of peripheral non-perfusion (PNP) might be associated with the natural course, severity, and treatment of DME. The present study seeks to understand the predictive power of central macular changes and clinico-demographic features for PNP in patients with clinically significant DME. Methods: A prospective study using contemporaneous multi-modal retinal imaging was performed. In total, 48 eyes with DME from 33 patients were enrolled. Demographic, clinical history, laboratory measures, ultrawide field photography, fluorescein angiography, optical coherence tomography (OCT), and OCT angiography results were acquired. Anatomic and vascular features of the central macula and peripheral retina were quantified from retinal images. Separate (generalized) linear mixed models were used to assess differences between PNP present and absent groups. Mixed effects logistic regression was used to assess which features have predictive power for PNP. Results: Variables with significant differences between eyes with and without PNP were insulin use (p = 0.0001), PRP treatment (p = 0.0003), and diffuse fluorescein leakage (p = 0.013). Importantly, there were no significant differences for any of the macular vascular metrics including vessel density (p = 0.15) and foveal avascular zone (FAZ) area (p = 0.58 and capillary tortuosity (p = 0.55). Features with significant predictive power (all p < 0.001) were subretinal fluid, FAZ eccentricity, ellipsoid zone disruption, past anti-VEGF therapy, insulin use, and no ischemic heart disease. Conclusions: In the setting of DME, macular vascular changes did not predict the presence of PNP. Therefore, in order to detect peripheral non-perfusion in DME, our results implicate the importance of peripheral retinal vascular imaging.