Abstract Objective Recently, conductive hearing loss (CHL) has been associated with dementia in two large independent cohorts. However, the disease‐specific association with dementia is unknown. We investigate the association of specific CHL pathologies and dementia in a large national cohort. Study Design Cross‐sectional epidemiologic study. Setting NIH All of Us Research Program. Methods Participants were ≥18 years old (n = 396,194). The exposures were CHL pathologies defined by ICD‐10 code: cholesteatoma (H71.X), tympanic membrane (TM) perforation (H72.X), and otosclerosis (H80.X). The outcome was all‐cause dementia defined by ICD‐10 codes (F01, F03, G30‐32). The odds of dementia in subjects with and without a CHL pathology were assessed with multivariable regression including age, sex, education, race, and ethnicity. Additional analyses included treatments. Results The mean (SD) age was 55 years (±17). After controlling for covariates, the odds of dementia were 1.77 times (1.08‐2.73; P = .015) higher for cholesteatoma compared to those without, 2.09 times (1.68‐2.59; P < .001) higher for TM perforation, and were nonsignificant for otosclerosis. After adding surgical treatment, odds of dementia dropped to 1.40 (0.82‐2.27; P = .198) for cholesteatoma and 2.01 (1.60‐2.50; P < .001) for TM perforation. Conclusions Cholesteatoma and TM perforation were associated with dementia in the All of Us Research Program. Treatment attenuates these relationships. Given the general implausibility for dementia to cause these CHL pathologies, reverse causation—that dementia causes hearing loss—is an unlikely explanation for the association. This study adds to the evidence that cognition is impacted by sensory deprivation but extends to correctable conductive causes.
Powell et al. (2026) studied this question.