The labyrinthine artery supplies the inner ear and functions close to an end artery with limited collateral reserve. Microvascular abnormalities have been implicated in sudden sensorineural hearing loss (SSNHL) and age-related hearing impairment, but direct in vivo assessment of inner-ear microvasculature is not feasible. Current hearing genome-wide association studies (GWAS) are underpowered for vessel-oriented questions. This study evaluates whether retinal arteriolar tortuosity (retina_AT), a measurable retinal microvascular trait, shares genetic architecture with blood pressure, cerebral small vessel disease (cSVD) traits, and hearing phenotypes, and whether it could serve as a candidate genetic window to inner-ear microvasculature.
The authors analysed publicly available GWAS and eQTL summary statistics. They estimated SNP heritability for 19 phenotypes and ran linkage-disequilibrium score regression (LDSC) to compute genome-wide genetic correlations (rg) for 11 prespecified phenotype pairs. Colocalization analysis used an approximate Bayes-factor approach (coloc.abf) between 19 independent retinal loci and systolic blood pressure (SBP). They performed 95% credible-set fine-mapping, UCSC gene annotation, and colocalization with GTEx v8 aortic and tibial artery expression quantitative trait loci (eQTL) data. Colocalization with white-matter hyperintensities (WMH) was assessed for at least one region. The study also included hearing-GWAS power calculations and univariable and multivariable Mendelian randomization (MR) analyses testing effects of blood pressure and resting heart rate on hearing outcomes.
All summary statistics and analysis scripts are publicly archived (sources and accession details reported in S2 Table and at the cited Zenodo repository). No individual-level data were used.
Genome-wide genetic correlation testing across the prespecified pairs identified a single significant result at the genome-wide level: retina_AT with systolic blood pressure (SBP). The reported genetic correlation was rg = 0.10 with strong statistical support (P = 4.2×10 and FDR q = 0.00042 in the abstract text). Other tested pairs, including retina_AT with broadly defined cSVD or hearing phenotypes, did not reach comparable genome-wide significance in the presented analyses.
Colocalization between 19 independent retinal loci and SBP produced heterogeneous results. The strongest retinal locus, at 13q34, did not colocalize with SBP: coloc results favored distinct causal signals (PP.H3 = 0.996), with the association peaks separated by approximately 267 kb and evidence of allelic heterogeneity. In contrast, two secondary retinal loci showed strong colocalization with SBP:
These colocalizing loci suggest distributed sharing of genetic influences on retinal arteriolar tortuosity and blood pressure concentrated at genes linked to vascular tone or related biology.
The study further evaluated colocalization with tissue-matched eQTLs from GTEx v8 in aortic and tibial arteries. For the CSK and PAWR loci, coloc results with artery eQTLs provided moderate support for shared regulatory effects (reported PP.H4 in the range ≈ 0.71–0.75). This lends tissue-relevant, though not definitive, functional support to the interpretation that the same variants may influence gene expression in arterial tissues and the GWAS traits.
At the 13q34 locus, the authors tested colocalization between retina_AT and white-matter hyperintensities (WMH) and observed a well-powered null for sharing in that region (PP.H4 = 0.019). The report notes that only this region and phenotype were tested for WMH colocalization in the manuscript.
On the hearing phenotype side, the genome-wide sharing of blood-pressure magnitude with hearing traits was estimated at approximately rg ≈ 0.10. Power calculations indicated that existing hearing GWAS are underpowered to detect modest genetic sharing: to achieve 80% power to detect rg ≈ 0.10 would require roughly a 1.5-fold larger sample size than currently available; detecting weaker sharing (rg ≈ 0.03) would demand approximately a 15-fold increase in sample size. The authors therefore characterize the observed genetic sharing as a near-powered null for hearing under current sample sizes and exclude rg ≥ 0.12 in their analyses.
Univariable and multivariable Mendelian randomization analyses were performed to test for causal effects of blood pressure and resting heart rate on hearing outcomes using the available summary statistics. These MR analyses did not find robust evidence of a causal effect of blood pressure or heart rate on hearing in the datasets analyzed.
The authors conclude that retinal arteriolar tortuosity shows limited but distributed genetic sharing with blood pressure, localized at vascular-tone–related genes CSK and PAWR, while the strongest retinal locus at 13q34 appears distinct from SBP signal and did not colocalize. No sharing with WMH was observed at 13q34 in the single-region test reported. While retinal microvasculature metrics may provide a candidate genetic window into inner-ear microvasculature, the authors emphasize that this translational analogy remains provisional. Direct testing will require larger, vessel-oriented hearing GWAS and functional studies in inner-ear tissues.
Limitations explicitly reported include reliance on summary-level data and the underpowered nature of current hearing GWAS for detecting modest genetic sharing. The study used publicly available datasets only; no new human participants were involved and no IRB approval was required for the secondary analyses.
Overall, the work proposes a targeted genetic link between retinal microvascular geometry and systemic vascular traits at specific loci and provides a rationale and roadmap for expanding hearing GWAS focused on vascular biology and inner-ear vessel function.