Metformin Use and Risk of Incident Hearing Loss Among Participants With Diabetes: A UK Biobank Cohort Study

To investigate the association between baseline metformin use and incident hearing loss among participants with diabetes in the UK Biobank.

This study included participants from the UK Biobank who had been diagnosed with diabetes before baseline and had no evidence of hearing loss (N = 24 981). Metformin use was determined based on self-reported medication use at the baseline assessment. Incident hearing loss was ascertained using a composite definition based on speech-reception-threshold estimate, hearing aid use, cochlear implantation and relevant ICD-10 diagnoses during follow-up. Cox proportional hazards models were used, with analyses stratified by sex and adjusted for demographic, lifestyle, socioeconomic, clinical factors, noise exposure, use of antidiabetic medications and renal function.

During approximately 12 years of follow-up, 725 and 385 incident hearing loss events were observed among male and female participants, respectively. Sex-stratified analyses showed that, among men with diabetes, baseline metformin use was associated with a lower risk of incident hearing loss (fully adjusted model: hazard ratio [95% confidence interval], 0.846 [0.725-0.987], p = 0.0339), and this association remained generally stable across multiple sensitivity analyses. In contrast, among women with diabetes, no statistically significant association was observed between metformin use and the risk of incident hearing loss (fully adjusted model: hazard ratio [95% confidence interval], 0.977 [0.790-1.208], p = 0.8290). Further interaction analysis showed that the interaction term between metformin use and sex was not statistically significant (p for interaction = 0.260).

In participants with diabetes from the UK Biobank, baseline metformin use was associated with a lower risk of incident hearing loss among men, whereas no significant association was observed among women. However, the interaction term between metformin use and sex was not statistically significant; therefore, it cannot yet be concluded that this association differed by sex.