Clinical and optical coherence tomography angiography-based risk factors for reduced retinal sensitivity in diabetic macular edema

PURPOSE: To compare microperimetric retinal sensitivity (RS) and Optical Coherence tomography angiography (OCTA)-derived perfusion metrics between eyes with diabetic macular edema (DME) and healthy controls, and to explore clinical and imaging parameters associated with lower RS within the diabetic cohort. METHODS: Prospective cross-sectional study of 38 diabetic eyes from 31 patients and 25 control eyes. All participants underwent microperimetry and OCTA on the same day. RS and OCTA-derived vessel metrics were assessed using a standard/general microperimetry pattern and clinician-defined customized subgrids targeting areas of suspected reduced perfusion in the deep vascular complex. Between-group differences were analyzed using linear mixed-effects models with patient identity as a random intercept to account for inter-eye correlation. Within the diabetic cohort, generalized linear mixed-effects models were used to explore parameters associated with RS, including Best-corrected visual acuity (BCVA), anti-vascular endothelial growth factor (VEGF) treatment intensity, persistent intraretinal fluid, previous panretinal photocoagulation, cardial and cerebral comorbidity, Optical Coherence Tomography (OCT)-derived neuroretinal metrics, and OCTA-derived vessel metrics. RESULTS: Compared with healthy controls, diabetic eyes showed significantly lower general macular RS and altered OCTA-derived perfusion metrics, particularly in the deep vascular complex. General macular RS was reduced by - 3.99 dB in diabetic eyes compared with controls (95% CI: -5.51 to - 2.48, p < 0.001). OCTA-derived vessel metrics showed moderate correlations with general RS. Localized OCTA-guided subgrids did not provide additional functional information beyond the standard/general microperimetry pattern. Within the diabetic cohort, worse BCVA, previous panretinal photocoagulation, and persistent intraretinal fluid were associated with lower RS in exploratory mixed-effects models. Cardio- and cerebrovascular comorbidity showed a non-significant tendency toward lower RS. CONCLUSION: Eyes with DME showed reduced retinal sensitivity and altered OCTA perfusion compared with healthy controls. Focal OCTA-guided subgrids did not improve functional assessment beyond the standard/general microperimetry pattern. Reduced general RS in treated DME appears to reflect multifactorial mechanisms, including visual function, residual exudative activity, prior retinal treatment.