Association Between Retinal Layer Thickness and Hearing Performance in the UK Biobank: A Cross-Sectional Study

Hearing and vision loss are established risk factors for dementia, with growing evidence of shared pathogenic mechanisms. We investigated the association between hearing performance and retinal layer thickness in the UK Biobank.

This cross-sectional study included 32,613 participants (mean age, 56.8 ± 8.14 years; 58.1% women) with bilateral speech reception threshold (SRT) values and optical coherence tomography-derived retinal layer measures. Multivariable linear regression models were fitted with retinal layer thickness as the outcome and the SRT as main exposure, adjusting for age, sex, ethnicity, spherical equivalent, visit type, diabetes, alcohol intake, smoking status, body mass index, education, Townsend deprivation index, and air pollution exposure. Age-equivalent differences were computed as the ratio of the SRT regression coefficient to the age coefficient.

Each 1-dB higher SRT (worse hearing) was associated with a thinner macula (-0.20 µm; P = 2.5 × 10⁻5), ganglion cell-inner plexiform layer (GC-IPL) (-0.09 µm; P = 1.1 × 10⁻5), inner nuclear layer (INL; -0.03 µm; P = 9.7 × 10⁻⁴), INL-external limiting membrane (INL-ELM; -0.0 7 µm; P = 3.2 × 10⁻3), and INL-retinal pigment epithelium (RPE) (-0.07 µm; P = 8.9 × 10⁻3) complexes. No associations were observed for retinal nerve fiber layer (RNFL), RPE, or photoreceptor layers. Participants with impaired hearing (SRT > -5.5 dB; n = 5,475) had a significantly thinner GC-IPL, INL, INL-ELM, INL-RPE, and total macula thickness compared with those with normal hearing (n = 27,138), with the greatest differences in the INL-ELM (-0.24 µm; 11.7 age-equivalent years) and INL-RPE (-0.30 µm; 6.4 years).

Poorer hearing was independently associated with thinner middle and inner retinal layers. These findings may suggest shared neural vulnerability across auditory and visual systems, highlighting their potential as indicators of broader neural health.