Relationship of macular microcirculation and retinal thickness with visual acuity in diabetic macular edema

Purpose: To assess perifoveal capillary blood flow velocity (BFV), capillary occlusion, and retinal thickness at the central fovea in diabetic patients with or without clinically significant macular edema; to examine the relationships of these variables with visual acuity (VA); and to identify their contributions to visual outcome and diabetic macular edema. Design: Comparative cross-sectional prospective study. Participants and Controls: Diabetic patients with clinically significant macular edema (CSME) (n = 22), matched diabetic patients without CSME (n = 22), and healthy volunteers (n = 16). Methods: Capillary BFV was measured by fluorescein angiography using a scanning laser ophthalmoscope and was analyzed by the tracing method. Severity of perifoveal capillary abnormalities was classified by the size and outline of the foveal avascular zone (FAZ) and extent of foveal capillary loss. Macular thickness was measured by optical coherence tomography. Each subject underwent a complete ophthalmic evaluation, and best-corrected VA (BCVA) was converted to the logarithm of the minimum angle of resolution scale. Main Outcome Measures: Relationship of perifoveal capillary BFV, capillary occlusion, and foveal thickness with VA. Results: Best-corrected VA significantly differed among all 3 groups (P<0.0001). Best-corrected VA correlated negatively with BFV (r = -0.644, P<0.0001) among all subjects and positively with retinal thickness at the central fovea in diabetic patients with CSME (r = 0.640, P = 0.0013). There was a positive correlation between BCVA and severity in the size of the FAZ (r = 0.484, P = 0.0015), outline of the FAZ (r = 0.542, P = 0.0004), and extent of foveal capillary loss (r = 0.585, P == 0.0001) among all diabetic subjects. Multiple regression analysis showed that retinal thickness at the central fovea was the only variable that significantly predicted VA (standardized regression coefficient, 0.635; P = 0.0001). Conclusions: Best-corrected VA was associated with perifoveal capillary BFV, severity of perifoveal capillary occlusion, and retinal thickness at the central fovea in diabetic patients, but the greatest contributing factor was only the retinal thickness.