AbstractBackgroundTo determine the potential mediating role of loneliness in the relationship between hearing ability and dementia.MethodsDesign: Longitudinal observational study. Setting: English Longitudinal Study of Ageing (ELSA). Participants: Individuals aged 50 and older (N = 4232). Measurements: Self‐reported hearing ability and loneliness were assessed from Wave 2 (2004–2005) to Wave 7 (2014–2015) of ELSA. Dementia cases were ascertained via self‐ or carer‐report or dementia medication at these waves. The medeff command in Stata version 17 was used to do cross‐section mediation analysis between hearing ability, loneliness, and dementia (Waves 3–7). Path‐specific effects proportional (cause‐specific) hazard models were then used to investigate longitudinal mediation (Waves 2–7).ResultsIn cross‐sectional analyses in Wave 7 alone, loneliness only mediated 5.4% of the total effects of limited hearing on dementia (indirect effects = increased risk of 0.06%; 95% CI: 0.002%–0.15%) under limited hearing and 0.04% (95% CI: 0.001%–0.11%) under normal hearing). In longitudinal analyses, there was no statistical evidence of a mediating role for loneliness in explaining the relationship between hearing ability and time‐to‐dementia (indirect effect estimate hazard ratio = 1.01 (95% CI: 0.99–1.05).ConclusionIn this community‐dwelling sample of English adults, there is a lack of evidence that loneliness mediates the relationship between hearing ability and dementia in both cross‐sectional and longitudinal analyses. However, as the number of dementia cases in this cohort was low, replication in other cohorts with larger sample sizes is required to confirm the absence of a mediated effect via loneliness.