Importance Given that age-related hearing loss is highly prevalent and treatable, understanding its causes may have implications for disease prevention. Objective To investigate whether microvascular retinal signs are associated with age-related hearing loss attributable to a hypothesized underlying shared pathologic entity involving microvascular disease. Design, Setting, and Participants The Atherosclerosis Risk in Communities Neurocognitive Study (ARIC-NCS) is a community-based prospective cohort study of 15 792 men and women aged 45 to 64 years at baseline. The ARIC-NCS participants returned for a fifth clinic visit in 2011-2013 and a sixth clinic visit in 2016-2017. Participants were recruited from 4 US communities (Washington County, Maryland; Forsyth County, North Carolina; Jackson, Mississippi; and Minneapolis suburbs, Minnesota). Participants included a subset of the ARIC-NCS cohort with complete covariate data who underwent retinal fundus photography at visit 5 (2011-2013) and completed hearing assessment at visit 6 (2016-2017) (N = 1458). Overall, 453 participants had diabetes; of those, 68 had retinopathy. Of 1005 participants without diabetes, 42 had retinopathy. Exposures Microvascular retinal signs included retinopathy, arteriovenous (AV) nicking, and generalized arteriolar narrowing measured using the central retinal arteriolar equivalent (CRAE). Main Outcomes and Measures Hearing was measured using the better-hearing ear pure-tone average (PTA) of air conduction speech thresholds (0.5, 1, 2, and 4 kHz). Multivariable-adjusted linear and ordered logistic regression was used to estimate the association between microvascular retinal signs and age-related hearing loss to describe the precision of the estimates and provide a plausible range for the true association. Results After full adjustment among 1458 individuals in the analytic cohort (mean [SD] age, 76.1 [5.0] years [age range, 67-90 years]; 825 women [56.6%]; 285 black [19.5%]), the difference in PTA per dB hearing level in persons with and without retinopathy was 2.21 (95% CI, -0.22 to 4.63), suggesting that retinopathy is associated with poorer hearing, although the width of the 95% CI prevents definitive conclusions about the strength of the observed association. Restricting the analysis to participants without diabetes, the difference in PTA associated with retinopathy was even greater (4.14; 95% CI, 0.10-8.17 dB hearing level), but the large width of the 95% CI prevents definitive conclusions about the association. In analyses quantifying the mean differences in hearing thresholds at individual frequencies by retinopathy status, the estimates trended toward retinopathy being associated, contrary to expectation, with better high-frequency hearing. At 8 kHz, the estimated difference in hearing thresholds in persons with retinopathy vs those without was -4.24 (95% CI, -7.39 to -1.09). Conclusions and Relevance In this population-based study, an association between the presence of microvascular retinal signs and hearing loss was observed, suggesting that retinopathy may have the potential to identify risk for hearing loss in persons without diabetes. The precision of these estimates is low; therefore, additional epidemiologic studies are needed to better define the degree of microvascular contributions to age-related hearing loss.
