Retinal vessel caliber differences between cataract and healthy individuals: a cross-sectional OCT-based analysis

While cataract represents a significant global health burden, evidence directly comparing retinal vessel calibers between cataractous and healthy eyes remains limited. This study characterized retinal vascular morphometry differences between cataractous and healthy eyes, evaluated age- and sex-adjusted associations, separately examined robustness to within-subject clustering, and assessed relationships with cataract severity using the Lens Opacities Classification System III (LOCS III) nuclear opalescence (NO) score. This retrospective cross-sectional study analyzed near-infrared reflectance en face images acquired using a spectral-domain optical coherence tomography (SD-OCT) system from 243 eyes (108 healthy control eyes and 135 cataractous eyes). Eight major retinal vessels ≥ 30 μm were measured: superior and inferior temporal and nasal arterioles and venules. Statistical analyses included independent-samples t-tests and binary logistic regression. Vessel-level comparisons were corrected using the Holm-Bonferroni procedure. The primary multivariable model included age, sex, and superior nasal retinal arteriole (SNRA) caliber; within-subject clustering was assessed separately using generalized estimating equations (GEE). Within cataractous eyes, correlations between the LOCS III nuclear opalescence (NO) score and three arteriolar calibers were additionally corrected using the Holm-Bonferroni procedure. Of 243 eyes, 242 unique participants were identified; one participant contributed both eyes. The cataract group was significantly older than the control group (63.72 ± 10.14 vs. 54.44 ± 8.46 years, p < 0.001). Unadjusted comparisons showed larger STRA, SNRA, and ITRA calibers in cataractous eyes ( p = 0.009, p < 0.001, and p = 0.045, respectively), but only SNRA remained significant after Holm-Bonferroni correction across eight vessel comparisons (adjusted p = 0.007). The SNRA difference was separately robust in a GEE analysis accounting for within-subject clustering (coefficient = 7.73 μm, p < 0.001) and in a one-eye-per-participant sensitivity analysis ( p < 0.001). Venular calibers did not differ significantly. Within cataractous eyes, the LOCS III NO score correlated positively with SNRA caliber (Spearman ρ = 0.548, unadjusted p < 0.001, Holm-adjusted p = 0.003). The weak ITRA correlation was nominal only (Pearson r = 0.175, unadjusted p = 0.042, Holm-adjusted p = 0.084), and STRA was not correlated ( p = 0.374). Sex distribution differed between groups ( p = 0.003). In a multivariable model adjusting simultaneously for age, sex, and SNRA caliber, SNRA remained independently associated with cataract status (adjusted OR = 1.33 per 10-µm increase, 95% CI: 1.09–1.61, p = 0.004), as did age (OR = 1.11 per year, 95% CI: 1.08–1.15, p < 0.001) and female sex (OR = 0.48 vs. male, 95% CI: 0.26–0.86, p = 0.014); model fit improved substantially with adjustment (Nagelkerke R²=0.336). SNRA caliber was independently associated with cataract status after adjustment for age and sex and was separately robust to accounting for within-subject clustering using GEE. SNRA was the only vessel-level association that survived correction for multiple comparisons. Its correlation with the LOCS III NO score remained significant after correction of the three severity correlations, whereas the ITRA finding was nominal and exploratory. The association may reflect a biological vascular response, an optical/measurement effect of lens opacity, or both. Prospective studies with direct control of optical and measurement artifacts are required.