BackgroundAlthough telomere length has a significant relationship with various age-related diseases, studies on its relationship with hearing status in adults are limited and equivocal. This study investigated the associations between mean telomere length (MTL) and low-, speech-, and high-frequency hearing threshold shifts of adults in the United States. MethodsA total of 2,027 adults, aged 20-69 years, from the National Health and Nutrition Examination Surveys (NHANES, 1999-2002) were included in the analytic sample. The quantitative polymerase chain reaction method was used for the MTL assay, and MTL was expressed using the telomere-to-single copy gene (T/S) ratio. Hearing loss was defined as a pure-tone average (PTA) for the better ear at >= 20 dB HL at frequencies 500, 1,000, 2,000, and 4,000 Hz. Univariate and multivariate linear regression analyses and smooth curve fittings were conducted to evaluate the correlation between MTL and low-, speech-, and high-frequency hearing levels. ResultsThe mean age of the participants was 40.60 +/- 12.76 years, including 952 men (weighted, 48.67%) and 303 (weighted, 12.88%) participants with hearing loss. After adjusting for potential confounders in the multivariate linear regression model, the relationship between MTL and hearing thresholds was not statistically significant. Smooth curve fittings indicated a non-linear relationship between MTL and high-frequency PTA hearing threshold shifts. MTL was inversely related to high-frequency PTA to the turning point (T/S ratio = 0.82) (adjusted beta-21.45, 95% CI -37.28, -5.62; P = 0.008). When the T/S ratio exceeded0.82, MTL was not associated with high-frequency PTA (adjusted beta 0.18, 95% CI -2.21, 2.57; P = 0.8809). ConclusionOur findings revealed that MTL was associated with high-frequency PTA hearing threshold shifts of adults in the United States in a non-linear manner.