A community-based sample of adults born 1946–1964 in the City of Busselton, Western Australia, was enrolled into the Busselton Healthy Ageing Study. Investigators invited 6,690 residents and enrolled 5,107 participants (76.3% response). For this analysis, 3,596 participants had reliable automated audiometry at both baseline (T0, 2010–2015) and the first follow-up (T1, 2016–2021) and were included. Mean baseline age was 57.9 years and 55% were female.
Audiometric outcomes used objective, automated testing (AMTAS controlling a Conera audiometer) performed in sound-proof booths. The study team calculated better-ear four-frequency pure tone averages (4FPTA, 0.5–4 kHz) and better-ear high-frequency pure tone averages (HFPTA, 4 & 8 kHz). WHO categories classified impairment (no impairment <20 dB; mild 20–34 dB; moderate 35–49 dB; and greater severity levels). Only tests graded “good” or “fair” at both time points were analysed. Demographic, environmental, and general health covariates were taken from the baseline visit; only hearing measures were taken from T1.
The cohort showed measurable deterioration in auditory thresholds across the six-year interval. The prevalence of mild-or-greater hearing loss rose from 14% to 26% when assessed by 4FPTA. High-frequency impairment, assessed by HFPTA, increased from 53% to 71%. Threshold worsening occurred in all age bands examined. The dataset indicated greater absolute deterioration at higher frequencies and larger changes in men than women.
These changes were observed across the sampled midlife span and were consistent with aging-related decline, but the magnitude of change differed by sex and frequency band. The authors reported these patterns using linear mixed models to account for repeated measures.
Age and sex remained core determinants: older baseline age and male sex were associated with faster deterioration in both univariate analyses and after adjustment for age and sex where relevant. Ear-related variables also contributed: participants reporting tinnitus or a history of chronic ear infection had more rapid threshold worsening. The study team used standard questionnaire items to record tinnitus and doctor-diagnosed chronic ear infection at baseline.
Education and income were collected and scored, but the article focused analysis on the variables listed above rather than presenting extended socioeconomic subgroup estimates in the current report.
Workplace noise exposure was self-reported. Participants answered whether they had worked in settings where they needed to raise their voice to be heard and, if so, how often. The investigators categorized exposure according to whether participants reported always, frequently, occasionally or never using hearing protection.
Noisy workplace exposure was associated with faster hearing decline, especially among participants who reported inconsistent or absent hearing protection. The team emphasised that protection frequency modified the relationship between occupational noise and threshold change; those always using protection showed smaller deterioration compared with intermittently or never-protected workers.
Baseline metabolic and cardiovascular measures were objective. Blood pressure was measured supine after five minutes’ rest and staged per AHA guidance. Fasting blood samples provided plasma glucose and HbA1c, and ADA criteria determined diabetes and pre-diabetes status. Anthropometry produced BMI and categorized weight status.
After adjusting for age and sex in univariate models, higher BMI, current smoking (pack years considered), hypertension and pre-diabetes were each associated with faster hearing-threshold decline. The investigators reported an association between these broader health profiles and auditory change, suggesting metabolic and vascular factors may influence midlife hearing trajectories. The study did not present intercurrent longitudinal changes in these health measures between T0 and T1 within this article.
Audiometry employed automated quality grading embedded within the testing software; only tests rated “good” or “fair” at both timepoints were included to mitigate measurement error. Trained staff supervised testing but did not perform subjective quality grading. Demographic and health exposures were taken from baseline questionnaires and objective assessments at T0; only hearing was reassessed at T1.
The analytic strategy used linear mixed models to estimate the association between baseline covariates and change in hearing thresholds, accounting for repeated measures and adjusted for age and sex where indicated. The authors reported results from univariate models and described which associations remained significant after adjustment for age and sex.
This analysis used baseline covariates and did not report on longitudinal changes in non-hearing health measures between visits; causality therefore cannot be established from the presented data. Audiometry was automated and quality-graded, but supervision did not include human regrading of test quality; the authors included only tests rated “good” or “fair” which limits but does not eliminate measurement concerns. Noise exposure relied on self-report and protection-use frequency, which are subject to recall bias.
The investigators identified unresolved clinical questions: the causal pathways that link metabolic health, cardiovascular risk profiles, and environmental exposures to hearing trajectories remain to be defined. The article did not report intervention or treatment effects, nor did it provide disaggregated temporal data on changes in metabolic or cardiovascular status between T0 and T1 for this cohort. The authors called for further longitudinal and mechanistic research to clarify these relationships and to examine whether modifying metabolic or occupational exposures alters the course of midlife hearing loss.
Ethics approval was obtained from The University of Western Australia Human Research Ethics Committee. Data access is restricted; de-identified datasets may be requested through the Busselton Population Medical Research Institute following ethics approval and a data access agreement. Funding sources were state and federal grants and private donations; the authors declared no competing interests.