EYS-associated retinal degeneration in South Korea: Genetic, Phenotype, and Natural History.

PURPOSE: To delineate the clinical and genetic spectrum of EYS-associated retinal degeneration (EYS-RD) and to quantify longitudinal progression of retinal structure and visual function.

DESIGN, SETTING, AND

PARTICIPANTS

Multicenter retrospective cohort study at eight tertiary referral centers in South Korea, enrolling 126 patients with biallelic EYS variants.

MAIN OUTCOME MEASURES: Annual progression rates of best-corrected visual acuity (BCVA, logMAR), ellipsoid zone width (EZW,) and area (EZA) on serial spectral-domain OCT, estimated using three-level linear mixed-effects models.

RESULTS

Among 126 patients (63 females, 50.0%; median follow-up, 5.2 years [IQR, 1.2-9.0]), serial OCT was analyzable in 163 eyes of 84 patients, of which 113 eyes had four time points and 152 had three or more, giving 591 EZW observations. In three-level mixed-effects models, EZW declined by 124.1 μm/year (95% CI, 106.4-141.8) and square root-transformed EZA by 0.105 mm/year (95% CI, 0.089-0.120), while BCVA deteriorated by 0.043 logMAR/year (95% CI, 0.025-0.060). Decline was linear over the observed interval: the median within-eye R² for a straight-line fit was 0.900 and a quadratic term for time was not significant (P = .239). A total of 66 unique EYS variants were identified, including 20 novel variants; structural variants occurred in 11 patients (8.7%), including large deletions encompassing non-coding exons 1-2 and a deep-intronic variant (c.1766+2635G>A) activating a cryptic exon. Progression did not differ between patients with biallelic LoF genotypes and carriers of at least one missense allele (time × genotype interaction, P = .923).

CONCLUSIONS

This large multicenter study provides quantitative benchmarks for EZW and EZA progression in EYS-RD, with a relative EZW losses of approximately 4.6% per year and that may serve as reference endpoints for gene therapy trials. Despite linear structural decline, the absence of genotype-phenotype correlation underscores phenotypic heterogeneity and suggests that factors beyond the primary genetic defect contribute to disease variability.