Cross-Sectional Associations Between Hyperglycemia and Hearing Loss at the Year 35 Coronary Artery Risk Development...

Sustained hyperglycemia is associated with pathogenic effects on microvasculature and sensory neurons, and diabetes may also affect the vasculature and neural systems of the inner ear. We examined the association of hyperglycemia with sensorineural hearing loss (HL) at three frequency ranges in a biracial, population-based cohort.

Among 1,930 Black and White men and women aged 53-65 years who were attending the Year 35 examination of the Coronary Artery Risk Development in Young Adults study, we assessed hearing using an automated method for testing auditory sensitivity to obtain pure-tone air-conduction thresholds. Normal hearing (≤ 25 dB) and HL (> 25 dB) were examined across three frequency ranges: 250-2000 Hz (low/middle), 500-4000 Hz (speech), and 4000-8000 Hz (high). Hyperglycemia was categorized as normal fasting glucose (≤ 5.5 mmol/L), prediabetes (5.6-6.9 mmol/L), and Type 2 diabetes (T2D; ≥ 7 mmol/L or taking medication for diabetes). Cross-sectional associations of prediabetes and T2D with any HL were assessed using crude and multivariable-adjusted logistic regression.

Black participants had a higher prevalence of T2D (26.3%) than White participants (13.0%), and men had a higher prevalence of both prediabetes (38.0% vs. 26.0%) and T2D (20.4% vs. 17.8%) than women. HL was rare in the low/middle and speech frequencies, but 38.7% had some HL in the high-frequency range, with 21.6% of the sample considered mild (sound detection thresholds at 26-40 dB) and 10.9% considered moderate (sound detection thresholds at 41-55 dB). Prediabetes was not associated with high-frequency HL, but those with T2D had 50% higher odds (odds ratio [OR] = 1.498, 95% CI [1.173, 1.913]) of HL compared to those without diabetes; this OR increased to 1.657 (95% CI [1.266, 2.170]) after controlling for demographic factors.

In this population-based sample of middle-aged participants, HL was most pronounced in the high-frequency range but was mostly mild. T2D was associated with prevalent HL in both crude and adjusted models. However, prediabetes was not associated with HL, suggesting that interventions to both improve glycemic control and detect HL may provide a window of opportunity for both prevention of progression to T2D and prevention or slower progression of HL.