Duration-dependent decoupling of the macular neurovascular unit in type 2 diabetes without retinopathy: A topographic analysis using ultra-widefield OCT angiography.

OBJECTIVE

To characterize duration-dependent topographic macular neurovascular unit (NVU) alterations in type 2 diabetes (T2DM) without diabetic retinopathy (DR) using 26×21 mm ultra-widefield swept-source OCT angiography (UWF-SS-OCTA) and evaluate its diagnostic utility for long-duration eyes.

METHODS

This cross-sectional study included 295 eyes (69 controls; 226 T2DM without DR stratified by duration: <5, 5-10, ≥10 years). The superficial vascular density (SVD), deep vascular density (DVD), superficial microvascular density (SMVD), and nerve-fibre-layer microvascular density (NFMVD) were quantified in the fovea (1 mm), inner (1-12 mm) and outer (12-21 mm) rings and their quadrants. Ganglion cell-inner plexiform layer (GCIPL) thickness was derived from the 26 × 21 mm OCTA scan. ANCOVA adjusted for age, sex, BMI, hypertension, IOP with false discovery rate correction, multivariable regression, and ROC/decision curve analyses were used.

RESULTS

Longer duration was associated with progressive GCIPL thinning and deep-predominant DVD reduction across nearly all macular zones (q < 0.05); the DVD deficit was most pronounced in the outer ring. SVD showed nasal-temporal asymmetry; selective nasal inner loss in ≥10-year group. GCIPL-DVD coupling, present in controls, attenuated in ≥10-year group, indicating NVU decoupling. Duration independently predicted GCIPL (β=-0.32) and DVD-M 12-21 (β=-0.23) after adjustment (P < 0.01). DVD alone discriminated ≥10-year eyes from controls (AUC 0.723, 95% CI 0.63-0.81); addition of GCIPL did not improve AUC (0.724, P > 0.05). Decision curve analysis confirmed net clinical benefit.

CONCLUSIONS

Preclinical NVU decoupling, which is predominant in the deep capillary plexus and nasally asymmetric, already occurs in T2DM without fundoscopic DR and progresses with disease duration. Deep-layer UWF-SS-OCTA metrics may strengthen risk stratification before clinical retinopathy.