LAUSR

OBJECTIVE We aimed to investigate whether current levels of total serum bilirubin are associated with different subtypes of sensorineural hearing loss (SNHL) in adolescents. METHODS A set of cross-sectional data from the National Health and Nutrition Examination Survey (NHANES) (2007-2010) was used. A subset of 1404 adolescents was sampled for measurements of total serum bilirubin, tympanometry, and average pure tone threshold at low-frequencies (LPTA: 500, 1000, 2000 Hz) or high-frequencies (HPTA: 3000, 4000, 6000, and 8000 Hz). SNHL was defined as the hearing loss that had type A tympanograms with a peak admittance of 0.3 ml or greater. Associations between serum bilirubin (square-root transformed) and different subtypes of SNHL were evaluated using binary or multinomial logistic regression models with 4-year sampling weights. The bootstrap method was used for estimation of variance and 10-fold cross-validation for assessment of overfitting issue. RESULTS Total serum bilirubin levels were found to be associated with any high-frequency (HPTA>15 dB in at least one ear, adjusted odds-ratio (ORa)(bootstrap 95% confidence interval) = 3.29(1.31-8.19), p = 0.011), but not with any low-frequency (LPTA>15 dB in at least one ear), SNHL in the US adolescents. Furthermore, high-frequency SNHL with HPTA>15 dB in both ears (bilateral) or HPTA≥25 dB in at least one ear, compared to that with HPTA>15 dB in one ear only (unilateral) or HPTA = 15-25 dB in at least one ear, had a stronger association with total serum bilirubin levels (ORa = 5.37(1.27-22.65), p = 0.022 for bilateral; ORa = 2.64(0.84-8.25), p = 0.094 for unilateral; ORa = 5.00(0.95-26.58), p = 0.058 for HPTA≥25 dB in at least one ear; as well as ORa = 3.06(1.15-8.25), p = 0.025 for HPTA = 15-25 dB in at least one ear). No severe overfitting problems were found. CONCLUSION Our findings suggest that current levels of total serum bilirubin may be informative in predicting and/or targeting high-frequency SNHL in the US adolescents.