Early Auditory and Psychological Shifts in Newly Enlisted Republic of Korea Air Force Soldiers Exposed to Aircraft Noise: A Pilot Prospective Study

Abstract

Introduction:

Aircraft noise is an occupational and environmental stressor associated with auditory injury and adverse mental health outcomes, but prospective evidence in newly enlisted personnel remains limited. This pilot study examined early changes in hearing thresholds and psychological well-being among Republic of Korea Air Force soldiers.

Methods:

Newly enlisted soldiers were enrolled at baseline (n = 96); 52 completed 8-month follow-up. Air-conduction thresholds were measured at 250 to 8000 Hz. Psychological and sleep outcomes were assessed using the Patient Health Questionnaire-9 (PHQ-9), Beck Depression Inventory (BDI), Center for Epidemiologic Studies Depression Scale-Revised (CESD-R), Beck Anxiety Inventory (BAI), Pittsburgh Sleep Quality Index, and Insomnia Severity Index. Regression models adjusted for noise exposure and shift work evaluated associations between threshold shifts and psychological changes.

Results:

Subclinical threshold shifts of 2 to 3 dB occurred at 4000 Hz in the right ear [False Discovery Rate–adjusted P-values (P_FDR) = 0.007] and left ear (P_FDR = 0.047), and at 250 Hz in the right ear (P_FDR = 0.038). PHQ-9, CESD-R, and BAI scores increased over follow-up (all P_raw < 0.05), although these changes were not significant after FDR adjustment (all P_FDR > 0.05). No incident tinnitus was reported. In adjusted models, 4000 Hz threshold shifts showed borderline associations with increased BDI (β = 0.350, P_FDR = 0.057) and CESD-R scores (β = 0.384, P_FDR = 0.057).

Conclusion:

Short-term aircraft noise exposure was associated with early subclinical auditory shifts and numerical worsening of psychological measures. The 4000 Hz shift may represent a preliminary marker of depressive symptom emergence, supporting further investigation of sensory–psychological interactions in high-noise military environments.

KEY MESSAGES

• (1)Military aircraft noise represents a significant occupational stressor that may affect both auditory and psychological health in newly enlisted personnel.
• (2)This prospective pilot study identified subclinical hearing-threshold shifts accompanied by numerical increases in depression and anxiety scores during the initial 8 months of service at a fighter wing.
• (3)Threshold shifts at 4000 Hz emerged as potential markers associated with increased depressive symptoms after adjustment for occupational noise exposure and shift-work status within this exploratory framework.
• (4)The observed associations between subclinical auditory alterations and mental health indicators support further investigation into integrated monitoring strategies for both sensory and psychological well-being in high-noise occupational settings.

INTRODUCTION

Environmental noise is increasingly recognized as a pervasive public health stressor with substantial implications for quality of life, psychological well-being, and psychiatric morbidity.[1] In particular, high-intensity aircraft noise has been associated with depression, anxiety, and sleep disturbance.[2] A 2020 meta-analysis demonstrated a 12% increase in depression risk for every 10 dB increase in aircraft noise exposure measured by L_den.[3] The psychological burden of chronic noise exposure may extend beyond mood disturbance, as prolonged transportation-related noise exposure has also been associated with increased suicide risk across diverse populations.[4] Complementing these epidemiological findings, experimental animal studies suggest that environmental noise may induce neuroinflammatory and oxidative stress pathways that contribute to anxiety- and depression-like phenotypes.[2]

Although the adverse consequences of occupational noise exposure are well recognized, including evidence that cumulative hearing loss among bus drivers correlates with years of service, investigations evaluating both auditory and non-auditory outcomes have largely relied on cross-sectional study designs.[5] Such approaches are inherently limited in their capacity to characterize the early onset of auditory injury or establish temporal relationships between noise exposure and psychological sequelae. Moreover, much of the existing literature derives from civilian populations residing near military airbases,[6] where interpretation may be complicated by age-related presbycusis in older cohorts. Consequently, prospective longitudinal data examining health trajectories from the onset of intensive aircraft noise exposure remain limited.

Personnel serving in the Republic of Korea Air Force (ROKAF) provide a unique model for addressing these knowledge gaps. This relatively healthy young population undergoes abrupt and mandatory exposure to high-intensity aircraft noise upon enlistment, thereby reducing potential confounding from age-related hearing decline. Although all personnel are exposed to environmental noise levels exceeding 85 weighted equivalent continuous perceived noise level (WECPNL), occupational exposure intensity differs substantially according to military role. This stratified environment permits systematic evaluation of the relationship between varying aircraft noise exposure levels and both auditory and psychological health outcomes.

Accordingly, this prospective study evaluated longitudinal changes in hearing thresholds and mental health indicators, including depression, anxiety, and sleep-related outcomes, among newly enlisted ROKAF soldiers stratified by occupational noise exposure and shift-work status. By examining associations between frequency-specific threshold shifts and psychological symptom scores, this study further explored the potential relationship between subclinical auditory alterations and psychological well-being. These findings provide preliminary evidence to support the development of integrated military health strategies incorporating both hearing conservation and mental health surveillance.

METHODS

Study Design and Ethics

This prospective longitudinal study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines and evaluated potential associations between aircraft noise exposure, hearing thresholds, and psychological health among ROKAF soldiers over an 8-month follow-up period.[7] The study protocol was approved by the Institutional Review Board of the Republic of Korea Air Force Aerospace Medical Center (IRB No. ASMC-23-IRB-010). Written informed consent was obtained from all participants prior to enrollment, and participants were informed of their right to withdraw from the study at any time without penalty.

The study was prospectively registered with the Clinical Research Information Service (CRIS), a primary registry of the World Health Organization International Clinical Trials Registry Platform (Registration No. KCT0010089). In addition, the study protocol and final manuscript underwent review and clearance by the relevant ROKAF authorities to ensure compliance with military information security regulations.

An a priori power analysis was performed using G*Power version 3.1.9.7 (Heinrich Heine University Düsseldorf, Düsseldorf, Germany). Assuming a medium effect size (d = 0.5), α = 0.05, and a statistical power of 0.80, the analysis indicated that a minimum of 34 participants would be required for paired t-test analyses. Among the 96 initially enrolled participants, 52 (54.2%) completed the 8-month follow-up assessment, thereby satisfying the predefined sample size requirement for the planned analyses.

Participant Selection and Eligibility

Eligibility criteria included the following: (1) male soldiers newly enlisted at an ROKAF fighter wing in 2024; (2) age ≥18 years; and (3) provision of voluntary written informed consent. Exclusion criteria were implemented to minimize potential confounding and establish a healthy baseline cohort. These criteria included the following: (1) pre-existing otologic conditions, such as chronic otitis media, or psychiatric disorders, including major depressive disorder, diagnosed prior to enlistment; (2) baseline audiometric thresholds exceeding 25 dB hearing level (HL) at any frequency between 500 and 8000 Hz; and (3) failure to complete follow-up audiometric testing or psychological assessments at 8 months. The follow-up interval was intentionally aligned with mandatory medical examinations required for promotion to corporal. This timing strategy was selected to maximize participant compliance and follow-up completion within the constraints of a conscripted military setting, as these examinations constitute mandatory administrative milestones for all personnel.

Noise Exposure Assessment and Grouping

All participants were stationed at a major Republic of Korea military fighter wing characterized by an average environmental noise level of 85 WECPNL. Residential areas adjacent to the installation are categorized within the 85 to 95 WECPNL range, suggesting that noise intensity within the base may be even greater because of runway proximity. WECPNL is a specialized aircraft noise metric incorporating both flight frequency and nighttime weighting penalties.[8] In contrast to civilian aviation noise, military aircraft noise is characterized by impulsive acoustic patterns and elevated peak intensities.[9] Although L_den (Day-Evening-Night Level) is internationally recognized as the standard environmental noise metric,[10] WECPNL remains the established metric for military aircraft noise assessment in South Korea. Given the uniformly elevated environmental exposure across the installation, participants were stratified according to occupational role rather than geographic exposure. The high-noise group included personnel assigned to aircraft maintenance, propulsion, and ground equipment units. Classification was informed by 8-hour time-weighted average (8-h TWA) monitoring records obtained from 2019 to 2023, which demonstrated a median exposure level of 85.70 dB (interquartile range: 78.40–89.90 dB), with intermittent exposure exceeding 90 dB(A).[11] In contrast, the low-noise group comprised administrative, medical, and support personnel whose 8-h TWA values consistently remained below 90 dB(A) despite exposure to the same high-WECPNL environment.

Audiometric Evaluation

Air-conduction hearing thresholds at 250, 500, 1000, 2000, 3000, 4000, and 8000 Hz were measured using a pure-tone audiometer (MA 42; MAICO Diagnostics, Berlin, Germany). Audiometric evaluations were performed by a board-certified otorhinolaryngologist and technical personnel with at least 3 years of audiometric testing experience. To ensure methodological consistency and reliability, all testing procedures followed the modified ascending method.[12]

All participants underwent pure-tone audiometry at 09:00. Within this military unit, routine flight-related noise-generating activities, including aircraft takeoffs, landings, engine run-ups, and live-fire drills, are scheduled to conclude by 18:00 on the preceding day. Consequently, participants underwent testing after an approximate 15-hour noise-free interval, exceeding the commonly recommended minimum 14-hour quiet period intended to eliminate potential temporary threshold shifts (TTS). Although night-shift duties were not specifically restricted, high-intensity aircraft operations and weapons-related training activities are not conducted during nighttime hours (18:00–09:00), further reducing the likelihood that TTS influenced the measured hearing thresholds.

Psychological and Sleep Measures

Psychological well-being and sleep-related outcomes were evaluated using six validated Korean self-report instruments. Detailed characteristics of each measure, including cultural adaptation procedures, construct focus, scoring structure, and psychometric properties, are summarized below.

Korean Beck Depression Inventory-I (BDI)

The Beck Depression Inventory (BDI) is a self-administered instrument designed to assess the severity of depressive symptomatology. For cultural adaptation, the original scale was translated into Korean and subsequently refined by an English linguistics scholar and a bilingual psychologist. The inventory consists of 21 items rated on a 4-point scale (0–3), yielding a total score ranging from 0 to 63, with higher scores reflecting greater depressive symptom severity. In the present study, a cutoff value of ≥16 was used to identify clinically relevant depressive symptoms. Previous validation studies reported acceptable psychometric performance, including a split-half reliability coefficient of 0.65 and a Cronbach α of 0.78 in the general population, increasing to 0.85 in clinical populations. Two-week test–retest reliability was reported as 0.75. Concurrent validity has also been demonstrated through correlations with other depression scales and through discrimination between depressed and non-depressed populations.[13]

Korean Patient Health Questionnaire-9

The Korean Patient Health Questionnaire-9 (PHQ-9) is a self-report screening tool used to evaluate depressive symptom severity according to the nine diagnostic criteria for major depressive disorder outlined in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV). The Korean version was previously translated and validated in the general population. The instrument consists of nine items scored from 0 to 3, producing a total score ranging from 0 to 27, with higher scores indicating greater depressive symptom burden. A cutoff value of ≥10 was applied in this study. Prior investigations reported strong internal consistency (Cronbach α = 0.882) and construct validity through correlations with the anxiety/depression dimension of the EuroQol Five-Dimension scale.[14]

Korean Center for Epidemiologic Studies Depression Scale-Revised

The Center for Epidemiologic Studies Depression Scale-Revised (CESD-R) assesses depressive symptoms according to the nine core symptom domains of major depressive episodes defined in the DSM-IV. The Korean version underwent prior cultural adaptation and standardization for use in Korean populations. The scale contains 20 items rated on a 5-point Likert scale, with total scores ranging from 0 to 80. Higher scores indicate greater depressive symptom severity, and a cutoff value of ≥13 was used in this study. Previous studies demonstrated excellent internal consistency (Cronbach α = 0.934) and concurrent validity through positive correlations with established depression and anxiety measures.[15]

Korean Beck Anxiety Inventory

The Beck Anxiety Inventory (BAI) evaluates the severity of clinical anxiety symptoms while distinguishing anxiety-related manifestations from depressive symptomatology. The standardized Korean version of the BAI was purchased from and administered through Korea Psychology Co., Ltd. (Daegu, South Korea). The inventory comprises 21 items scored on a 4-point Likert scale (0–3), yielding total scores ranging from 0 to 63, with higher scores representing more severe anxiety symptoms. A cutoff value of ≥16 was used to define clinically significant anxiety symptoms. Previous validation studies demonstrated excellent internal consistency (Cronbach α = 0.96) and concurrent validity through correlations with the BDI-II and Minnesota Multiphasic Personality Inventory-2 scales.[16]

Korean Pittsburgh Sleep Quality Index

The Pittsburgh Sleep Quality Index (PSQI) assesses subjective sleep quality over the preceding month. The Korean version underwent translation and back-translation by sleep specialists, a neuropsychologist, and linguists to ensure cultural adaptation and conceptual equivalence. The instrument consists of 18 items generating seven component scores and a global score ranging from 0 to 21. Higher scores indicate poorer sleep quality, and a cutoff value of ≥6 was applied in this study. Prior validation studies reported good internal consistency (Cronbach α = 0.84) and test–retest reliability (r = 0.65).[17]

Korean Insomnia Severity Index

The Insomnia Severity Index (ISI) evaluates perceived insomnia severity, including difficulties initiating sleep, maintaining sleep, and associated daytime impairment. The Korean version was previously validated through translation and back-translation procedures conducted by sleep specialists and a linguist. The instrument contains seven items rated on a scale from 0 to 4, generating total scores ranging from 0 to 28, with higher scores indicating more severe insomnia symptoms. A cutoff value of ≥15 was used to identify clinically significant insomnia symptoms. Previous studies demonstrated strong internal consistency (Cronbach α = 0.92), test–retest reliability (r = 0.86), and convergent validity with both PSQI scores and polysomnographic parameters.[18] Participants exceeding the predefined cutoff values for each respective instrument were categorized as symptomatic.

Statistical Analysis

All statistical analyses were performed using Python version 3.12.12 (Python Software Foundation, Wilmington, DE, USA). Continuous variables were presented as mean ± standard deviation. Missing data were managed using listwise deletion. To evaluate the potential for attrition bias, baseline demographic and clinical characteristics were compared between participants included in the final analysis and those lost to follow-up.

Distribution normality was assessed using the Shapiro–Wilk test. For variables demonstrating deviation from normality, parametric analyses, including paired t-tests and linear regression models, were nevertheless applied. This approach was considered appropriate given the sample size (n = 52) and the robustness of parametric methods under conditions of moderate non-normality, as supported by the Central Limit Theorem. Longitudinal within-group changes and between-group comparisons of change magnitude were evaluated using paired and independent t-tests, respectively. Fisher’s exact test and McNemar’s test were used for prevalence analyses. Pearson correlation analyses were performed to evaluate associations between hearing-threshold shifts and changes in psychological measures. Variables demonstrating potential statistical associations were subsequently entered into simple and multiple linear regression models adjusted for occupational noise exposure and shift-work status to identify potential predictors of psychological symptom progression. Statistical significance was defined as a two-sided P-value <0.05. To reduce the risk of type I error arising from multiple comparisons across auditory frequencies and psychological scales, the False Discovery Rate (FDR) was controlled using the Benjamini–Hochberg procedure. For findings reaching the initial significance threshold, FDR-adjusted P-values (P_FDR) were reported alongside the corresponding unadjusted P-values to facilitate a more conservative interpretation of the results.

RESULTS

Sample Selection and Participant Attrition

Among the 96 soldiers screened and confirmed eligible at baseline, 52 (54.2%) completed the 8-month follow-up assessment and were included in the final analysis. Forty-four participants (45.8%) were excluded because of loss to follow-up (n = 42) or medical discharge during the study period (n = 2). The two medically discharged participants were excluded for specific clinical reasons: one sustained a traumatic tympanic membrane perforation, whereas the other was discharged following a clinically diagnosed depressive episode. The remaining 42 participants either failed to attend the scheduled follow-up audiometric examination or did not complete the psychological assessment battery [Figure 1].

Figure 1.

Flow diagram of participant recruitment and attrition.

Baseline Characteristics of the Participants

Baseline comparisons between participants included in the final analysis and those lost to follow-up are summarized in Table 1. No statistically significant differences were observed in alcohol consumption (55.0 ± 89.1 vs. 36.7 ± 53.3 g/week; P = 0.234) or smoking history (0.4 ± 1.1 vs. 0.2 ± 0.4 pack-years; P = 0.196). Although the included group was statistically older than the excluded group (20.6 ± 1.7 vs. 20.0 ± 1.1 years; P = 0.035), the absolute difference was small. Differences were identified in both medical and occupational characteristics. Specifically, the included group demonstrated a lower prevalence of allergic rhinitis (11.5% vs. 29.5%; P = 0.039) and shift-work status (34.6% vs. 59.1%; P = 0.024), but a greater proportion of participants classified within the high-noise exposure group (44.2% vs. 15.9%; P = 0.004).

Table 1.

Demographic and Characteristics of Study Participants at Baseline.

Characteristic	Included Participants (n = 52)	Lost to Follow-up (n = 44)	Total (N = 96)	P-Value
Age (years)	20.6 ± 1.7	20.0 ± 1.1	20.3 ± 1.5	0.035
Alcohol intake (g/week)	55.0 ± 89.1	36.7 ± 53.3	46.4 ± 74.5	0.234
Smoking history (pack-years)	0.4 ± 1.1	0.2 ± 0.4	0.3 ± 0.9	0.196
Shift work	18 (34.6%)	26 (59.1%)	44 (45.8%)	0.024
High noise exposure	23 (44.2%)	7 (15.9%)	30 (31.2%)	0.004
Allergic rhinitis	6 (11.5%)	13 (29.5%)	19 (19.8%)	0.039
Sinusitis	2 (3.8%)	4 (9.1%)	6 (6.2%)	0.408

Data are presented as mean ± standard deviation (SD) for continuous variables and number (%) for categorical variables. P-values were calculated using Welch’s t-test for continuous variables and Pearson’s Chi-Square test or Fisher’s exact test for categorical variables.

Longitudinal Changes in Pure-Tone Audiometry Air-Conduction Thresholds

Over the 8-month follow-up period (n = 52), pure-tone audiometry thresholds demonstrated a general upward shift across most tested frequencies in both ears [Figure 2]. The most prominent changes were observed at 4000 Hz, a frequency classically associated with early noise-induced hearing loss (NIHL). In the right ear, the 4000 Hz threshold increased from 3.08 ± 4.97 to 5.69 ± 6.02 dB HL [mean shift = 2.62 dB; 95% confidence interval (CI): 1.13–4.10; raw P = 0.001, P_FDR = 0.015]. Similarly, the left-ear 4000 Hz threshold increased from 3.85 ± 5.91 to 5.87 ± 6.98 dB HL (mean shift = 2.02 dB; 95% CI: 0.38–3.66; raw P = 0.020, P_FDR = 0.072), demonstrating a numerical trend toward elevation that did not remain significant following FDR adjustment. Within the low-frequency range, the right ear also demonstrated a threshold increase at 250 Hz, rising from 3.46 ± 3.51 to 5.19 ± 5.05 dB HL (mean shift = 1.73 dB; 95% CI: 0.56–2.90; raw P = 0.006, P_FDR = 0.038). Threshold changes at other frequencies, including 3000 Hz (P = 0.051) and 250 Hz (P = 0.058) in the left ear and 2000 Hz (P = 0.058) in the right ear, did not reach statistical significance.

Figure 2.

Longitudinal changes in pure-tone audiometry air-conduction thresholds over the 8-month follow-up period. Note: The graphs show the mean hearing threshold for the (A) right ear and (B) left ear. Asterisks indicate frequencies where a statistically difference was observed between baseline and follow-up thresholds based on unadjusted values: ^* P_raw < 0.05, ^** P_raw < 0.01 (paired t-test). Note the zoomed Y-axis scale (−2 to 12 dB HL) to highlight threshold shifts.

Longitudinal Changes in Mental Health and Sleep Measures

Longitudinal assessment of the 52 participants over 8 months demonstrated a general numerical increase across all psychological and sleep-related scales, suggesting a potential decline in psychological well-being and sleep quality over time [Figure 3]. Numerical increases were observed for PHQ-9 scores (0.58 ± 1.18 to 1.77 ± 3.63; mean difference = 1.19; 95% CI: 0.19–2.19; P = 0.024, P_FDR = 0.072), Korean CESD-R (K-CESD-R) scores (0.87 ± 1.77 to 4.27 ± 9.95; mean difference = 3.40; 95% CI: 0.66–6.14; P = 0.018, P_FDR = 0.072), and BAI scores (1.06 ± 2.15 to 2.62 ± 5.84; mean difference = 1.56; 95% CI: 0.06–3.05; P = 0.047, P_FDR = 0.094). Although these measures exhibited upward numerical trends, statistical significance was not maintained following FDR adjustment. No statistically significant longitudinal changes were observed for BDI scores (1.40 ± 2.77 to 2.96 ± 6.67; P = 0.095), ISI-Korean (ISI-K) scores (3.12 ± 2.91 to 4.12 ± 5.16; P = 0.187), or Korean version of PSQI (K-PSQI) scores (3.42 ± 2.01–4.25 ± 3.37; P = 0.100). Prevalence analyses using McNemar’s test demonstrated that, aside from sleep quality impairment, no participants met clinical thresholds for depression, anxiety, or insomnia at baseline. At follow-up, a limited number of newly symptomatic cases emerged, including poor sleep quality (K-PSQI >5: 17.3%–23.1%; P = 0.629), depressive symptoms based on K-CESD-R ≥13 (0.0%–9.6%; P = 0.063), and BDI ≥16 (0.0%–7.7%; P = 0.125). Similarly, the prevalence of PHQ-9 ≥10, BAI ≥16, and ISI-K ≥15 each increased from 0.0% to 5.8% (n = 3 for each comparison; all P = 0.250).

Figure 3.

Changes in mental health scores between baseline and follow-up. Note: Panels (A–F) display boxplots representing the distribution of scores for each scale. The central line represents the median, the box edges represent the interquartile range (IQR), and the whiskers represent the range. Individual data points are overlaid as jittered dots. BAI = Beck Anxiety Inventory, BDI = Beck Depression Inventory, ISI-K = Insomnia Severity Index-Korean, K-CESD-R = Korean Center for Epidemiologic Studies Depression Scale-Revised, K-PSQI = Korean version of Pittsburgh Sleep Quality Index, PHQ-9 = Patient Health Questionnaire-9, ns = not significant. Asterisks indicate scales where a statistically difference was observed between baseline and follow-up scores based on unadjusted values. ^* P_raw < 0.05, (paired t-test).

Comparison of Hearing-Threshold Shifts and Mental Health Changes by Noise Exposure Level

Subgroup analyses were conducted to compare participants in the high-noise exposure group (n = 23) and low-noise exposure group (n = 29) [Figure 4]. The high-noise group demonstrated predominantly high-frequency threshold shifts. In the right ear, thresholds increased at 3000 Hz (2.39 ± 3.65 to 3.91 ± 3.98 dB HL; P = 0.050, P_FDR = 0.174) and 4000 Hz (2.61 ± 3.33 to 7.61 ± 6.19 dB HL; mean shift = 5.00 dB; 95% CI: 2.31–7.69; P = 0.001, P_FDR = 0.020). In the left ear, threshold elevations were also observed at 3000 Hz (2.61 ± 3.33 to 5.65 ± 5.07 dB HL; P = 0.007, P_FDR = 0.051) and 4000 Hz (3.04 ± 4.70 to 6.96 ± 6.35 dB HL; P = 0.019, P_FDR = 0.086). In contrast, the low-noise group did not exhibit substantial high-frequency threshold elevation but demonstrated numerical increases at 250 Hz in both the right ear (3.97 ± 4.09 to 6.55 ± 5.53 dB HL; P = 0.011, P_FDR = 0.153) and left ear (4.48 ± 4.50 to 6.21 ± 4.75 dB HL; P = 0.039, P_FDR = 0.195).

Figure 4.

Comparison of hearing-threshold shifts stratified by noise exposure level. Note: (A,B) High-noise exposure group (n = 23): Personnel working in aircraft maintenance or flight operations. (C, D) Low-noise exposure group (n = 29): Personnel working in administrative or support roles. Asterisks indicate frequencies where a statistically difference was observed between baseline and follow-up thresholds based on unadjusted values. ^* P_raw < 0.05, ^** P_raw < 0.01 (paired t-test). Note the zoomed Y-axis scale (−2 to 12 dB HL) to highlight threshold shifts.

No statistically significant between-group differences were observed in the magnitude of change for psychological or sleep-related measures. Mean PHQ-9 score changes were 1.22 ± 3.22 in the high-noise group and 1.17 ± 4.07 in the low-noise group (P = 0.965). Likewise, between-group differences for BDI (P = 0.367), K-CESD-R (P = 0.393), BAI (P = 0.993), ISI-K (P = 0.514), and K-PSQI (P = 0.387) did not achieve statistical significance. The prevalence of symptomatic cases at follow-up was also comparable between groups (all P > 0.05), including PHQ-9 ≥10 (4.3% vs. 6.9%), K-CESD-R ≥13 (13.0% vs. 6.9%), BDI ≥16 (13.0% vs. 10.3%), BAI ≥8 (13.0% vs. 13.0%), ISI-K ≥8 (21.7% vs. 17.2%), and K-PSQI ≥5 (43.5% vs. 37.9%).

Comparison of Hearing-Threshold Shifts and Mental Health Changes by Shift Work Status

Outcomes were additionally compared between shift workers (n = 18) and nonshift workers (n = 34) [Figure 5]. Shift workers demonstrated predominantly low-frequency threshold shifts. In the right ear, the 250 Hz threshold increased from 4.17 ± 4.62 to 8.61 ± 5.64 dB HL (mean difference = 4.44 dB; 95% CI: 1.95–6.94; P = 0.003, P_FDR = 0.039), whereas the 500 Hz threshold increased from 3.61 ± 3.35 to 5.83 ± 3.09 dB HL (P = 0.028, P_FDR= 0.132). The left-ear 250 Hz threshold also increased from 5.28 ± 5.28 to 8.06 ± 4.89 dB HL (P = 0.028, P_FDR = 0.132). In contrast, nonshift workers demonstrated increases primarily within higher-frequency ranges. In the right ear, thresholds increased at 3000 Hz (2.35 ± 3.53 to 3.68 ± 3.76 dB HL; P = 0.018, P_FDR = 0.062) and 4000 Hz (2.35 ± 3.07 to 6.18 ± 5.78 dB HL; P < 0.001, P_FDR = 0.007). In the left ear, significant threshold elevations were similarly observed at 3000 Hz (P = 0.003, P_FDR = 0.022) and 4000 Hz (P = 0.010, P_FDR = 0.047).

Figure 5.

Comparison of hearing-threshold shifts stratified by shift work status. (a, b) Shift workers (n = 18): personnel engaged in rotating shift duties. (c, d) Nonshift workers (n = 34): personnel working regular daytime hours. Shift workers exhibited statistically threshold shifts in the low-frequency range (250–500 Hz), whereas nonshift workers (including those in high-noise roles) showed shifts primarily in the high-frequency range. Asterisks indicate frequencies where a statistically difference was observed between baseline and follow-up thresholds. ^* P_raw < 0.05, ^** P_raw < 0.01, ^*** P_raw < 0.001 (paired t-test). Note the zoomed Y-axis scale (−2 to 12 dB HL) to highlight threshold shifts.

No statistically significant between-group differences were identified for changes in psychological or sleep-related measures. Mean PHQ-9 score changes were 1.67 ± 5.06 among shift workers and 0.94 ± 2.75 among non-shift workers (P = 0.578). Between-group differences for BDI (P = 0.868), K-CESD-R (P = 0.881), BAI (P = 0.863), ISI-K (P = 1.000), and K-PSQI (P = 0.580) were likewise not statistically significant. The prevalence of symptomatic cases at follow-up also did not differ significantly between groups (all P > 0.05), including PHQ-9 ≥10 (11.1% vs. 2.9%; P = 0.272), BAI ≥8 (11.1% vs. 14.7%), and K-PSQI ≥5 (38.9% vs. 41.2%).

Correlation and Regression Analysis

Correlation analyses demonstrated that threshold shifts at 4000 Hz were associated with increases in depressive symptom measures. In the right ear, 4000 Hz threshold shifts correlated positively with changes in PHQ-9 scores (r = 0.285; 95% CI: 0.012–0.518; P = 0.041), BDI scores (r = 0.392; 95% CI: 0.133–0.601; P = 0.004), and K-CESD-R scores (r = 0.404; 95% CI: 0.147–0.610; P = 0.003). Left-ear 4000 Hz threshold shifts also correlated with changes in K-CESD-R scores (r = 0.310; 95% CI: 0.039–0.539; P = 0.025). No statistically significant correlations were identified between low-to-mid frequency thresholds (250–2000 Hz) and anxiety or sleep-related measures.

Multiple linear regression analyses adjusted for occupational noise exposure and shift-work status demonstrated that mean 4000 Hz threshold shifts were associated with increases in depressive symptom scores [Table 2]. Specifically, each 1 dB increase in the mean 4000 Hz threshold was associated with increases in BDI scores (B = 0.424; 95% CI: 0.073–0.776; β = 0.350; P = 0.019, P_FDR = 0.057) and K-CESD-R scores (B = 0.710; 95% CI: 0.180–1.240; β = 0.384; P = 0.010, P_FDR = 0.057). Following Benjamini–Hochberg FDR adjustment for multiple comparisons, the adjusted P-values for both depressive symptom scales approached, but did not reach, statistical significance (P_FDR = 0.057 for both analyses). ISI scores demonstrated a marginal association with 4000 Hz threshold shifts (B = 0.271; P = 0.071, P_FDR = 0.142), whereas no significant associations were identified for PHQ-9, BAI, or K-PSQI scores in the adjusted models.

Table 2.

Simple and Multiple Linear Regression Analyses of the Association between Threshold Shifts at 4000 Hz and Changes in Mental Health Scores.

Dependent Variable (Δ)	Unadjusted B	Unadjusted P	Adjusted B† (95% CI)	Standardized β	Adjusted P	Adjusted R2	P FDR	Post-hoc Power
Depression
PHQ-9	0.141	0.137	0.166 (−0.037 to 0.369)	0.245	0.106	0.011	0.160	0.066
BDI	0.421	0.012 *	0.424 (0.073–0.776)	0.350	0.019*	0.076	0.057	0.122
K-CESD-R	0.694	0.006**	0.710 (0.180–1.240)	0.384	0.010*	0.098	0.057	0.154
Anxiety
BAI	0.152	0.288	0.176 (−0.134 to 0.486)	0.174	0.259	−0.033	0.31	0.052
Sleep
ISI-K	0.265	0.055	0.271 (−0.024 to 0.566)	0.274	0.071	0.023	0.142	0.073
K-PSQI	0.051	0.582	0.026 (−0.176 to 0.228)	0.040	0.796	−0.045	0.796	0.051

Δ represents the change in score (follow-up minus baseline). The predictor variable is the mean threshold shift at 4000 Hz of both ears; ^† Adjusted for occupational noise exposure group and shift work status. ^* P < 0.05, ^** P < 0.01. B = unstandardized regression coefficient, β = standardized regression coefficient, BAI = Beck Anxiety Inventory, BDI = Beck Depression Inventory, CI = confidence interval, ISI-K = Insomnia Severity Index-Korean, K-CESD-R = Korean Center for Epidemiologic Studies Depression Scale-Revised, K-PSQI = Korean version of Pittsburgh Sleep Quality Index, P_FDR = false discovery rate-adjusted P-value, PHQ-9 = Patient Health Questionnaire-9, R² = coefficient of determination,.

DISCUSSION

This prospective study followed ROKAF soldiers over an 8-month period and identified concurrent early changes in hearing thresholds and mental health indicators. These preliminary findings suggest that exposure to high-intensity aircraft noise during the initial phase of military service may be associated with measurable alterations in both auditory and psychological domains.

Military Aircraft Noise and Hearing Loss

Military aircraft-related NIHL remains an important occupational health concern, as sustained exposure to high-intensity noise is well documented to be associated with clinically meaningful auditory changes.[19] In a study of 1218 ROKAF pilots, high-frequency threshold elevation, particularly at 4000 Hz, progressed with increasing service duration and reached 35.8 dB in the right ear and 41.8 dB in the left ear.[20] These findings are consistent with recent evidence indicating that aircraft noise may adversely affect auditory perception among pilots.[21] Similar risks have been documented internationally. Among Australian fuel tank maintenance workers, hearing thresholds differed by 9.7 to 18 dB across the 500 to 6000 Hz range compared with the general population.[22] Longitudinal data from the United States identified a 14.7% incidence of sensorineural hearing loss among aviators over 14 years.[23] Platform-specific vulnerability has also been reported in French military personnel, with fighter and helicopter crews demonstrating higher thresholds than transport pilots.[24] Ground-based personnel are not exempt from this risk, as noise-induced threshold shifts have also been observed among cryptolinguists.[25]

Within this context, the 2 to 3 dB threshold shifts observed in the present cohort over only 8 months may reflect early subclinical auditory alterations. Such changes may be exacerbated by inconsistent hearing protection use. In a survey of Republic of Korea military personnel, only 18.8% of soldiers reported consistent earplug use, whereas 45.2% reported never using earplugs; nonusers had a 3.41-fold higher risk of hearing loss.[26] The present findings suggest that early subclinical threshold shifts may arise, at least in part, from intermittent compliance with hearing protection protocols, emphasizing the need for more rigorous implementation and monitoring of preventive measures.

Noise Exposure and Mental Health

Prior research supports an association between aircraft noise exposure and psychological distress.[19] In a Korean cross-sectional study of 898 residents living near military airbases, BAI and BDI scores were significantly higher than those of controls; notably, hearing impairment remained associated with anxiety and depression after adjustment for noise intensity.[6] Meta-analytic evidence further supports this relationship, demonstrating a 12% increase in depression risk for each 10 dB increment in L_den (effect size = 1.12; 95% CI: 1.02–1.23).[3] With respect to sleep, a recent population-level assessment around U.S. military airbases estimated that nighttime aircraft noise exposure places a substantial number of residents at risk for adverse sleep outcomes, with thousands predicted to be highly sleep disturbed.[9]

In the present study, however, changes in PSQI and ISI scores did not reach statistical significance over the 8-month follow-up period. Nevertheless, the upward numerical trends observed in these measures may indicate an early cumulative effect on sleep quality that was not fully captured within the current study design. The absence of statistically significant sleep-related findings may be attributable to the modest sample size and relatively short observation period. Longer longitudinal follow-up in a larger cohort may be necessary to more fully characterize the effects of military aircraft noise on sleep architecture during the early phase of service.

Pathophysiological Implications

Bilateral 4000 Hz threshold shifts are characteristic early findings in NIHL.[5,22] Although the 2 to 3 dB shift observed over 8 months was modest, it may represent an early physiological manifestation of cochlear synaptopathy or hidden hearing loss rather than simple measurement variability.[27] Mechanistically, high-intensity noise is hypothesized to promote reactive oxygen species overproduction, potentially leading to mitochondrial dysfunction and apoptosis in outer hair cells, although these biological markers were not directly assessed in this study.[27] In addition, the right-ear 250 Hz shift may relate to the distinctive acoustic profile of military jet noise, which contains substantial low-frequency energy that can be transmitted through bone conduction.[9,28] Systemic stress may also influence endolymphatic homeostasis, paralleling mechanisms implicated in acute low-tone sensorineural hearing loss and suggesting that psychological distress could serve as a metabolic mediator associated with auditory threshold elevation.[27]

Neurobiological and Sleep-Related Pathways

Numerical increases in PHQ-9, CESD-R, and BAI scores may reflect early central neurobiological responses to noise-related stress.[27] Chronic noise exposure has been proposed to induce neuroinflammation and impair synaptic plasticity; potential structural alterations in the hippocampus and amygdala offer a plausible, although theoretical, framework for the mood-related changes observed in this cohort.[2] Furthermore, although changes in sleep measures, including PSQI and ISI scores, were not statistically significant, their upward numerical trends may indicate a potential self-reinforcing pathway between sleep disruption and psychological vulnerability. Subclinical sleep fragmentation may facilitate cerebral oxidative stress through mechanisms involving nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, potentially contributing to progressive reductions in psychological resilience.[2]

Psychosocial Considerations

The observed correlations between 4000 Hz threshold shifts and depressive symptom scores may indicate potential psychosocial consequences of early sensory alterations. Although threshold shifts of 2 to 3 dB are generally considered insufficient to produce overt communication impairment,[11] such subtle auditory changes may still influence psychological well-being through alternative mechanisms, including subclinical cochlear synaptopathy or direct neurobiological responses to acoustic stress. Notably, the absence of newly reported tinnitus cases contrasts with conventional models in which psychological distress is primarily linked to tinnitus burden.[29] This finding may suggest that subclinical auditory threshold shifts themselves could be independently associated with mood-related changes, even in the absence of tinnitus symptoms.

Study Limitations and Strengths

Several limitations should be considered when interpreting the present findings. First, the 8-month follow-up period and 45.8% attrition rate introduce the possibility of selection bias. This attrition appeared to be driven largely by changes in participation behavior during military service progression. Although participants demonstrated high compliance during the recruit phase, willingness to participate in voluntary research procedures appeared to decline following promotion to corporal, with many soldiers prioritizing rest over nonmandatory activities. Baseline differences between completers and noncompleters were observed with respect to noise exposure classification and shift-work status, and this substantial attrition may therefore have contributed to directional bias, potentially leading to overestimation of noise-related effects. Nevertheless, the prospective design conducted within a relatively controlled military environment provides a stronger framework for evaluating longitudinal changes than cross-sectional methodologies.

Second, the use of self-report psychological instruments may have introduced information bias and social desirability bias, particularly within the context of a hierarchical military structure. Third, the observed 2 to 3 dB threshold shifts were subclinical and fall within the range of standard audiometric measurement variability.[30] However, surveillance of these subtle trends within a previously healthy young cohort may provide preliminary insight into the earliest stages of auditory alteration before progression to permanent hearing deterioration. Fourth, the absence of bone-conduction testing and speech audiometry, as well as the possibility of inter-tester variability, should also be acknowledged despite strict adherence to standardized testing protocols.

Fifth, the substantial collinearity between occupational noise exposure and shift-work status complicates interpretation of the independent contribution of each factor. Specifically, the high-noise exposure group consisted predominantly of nonshift workers (n = 23), whereas the low-noise exposure group largely comprised shift workers (n = 18). Consequently, the observed psychological changes may reflect the combined influence of aircraft noise exposure and circadian disruption rather than the isolated effect of either environmental stressor. Future investigations incorporating individual noise dosimetry and more granular occupational exposure assessments will be necessary to disentangle these interacting variables.

Finally, although nonspecific military stressors may also have contributed to psychological outcomes, 4000 Hz threshold shifts demonstrated statistical associations with progression in depressive symptom scores. For each 1 dB increase in the mean 4000 Hz threshold, BDI scores increased by 0.424 points (P = 0.019; adjusted R² = 0.076), whereas K-CESD-R scores increased by 0.710 points (P = 0.010; adjusted R² = 0.098). However, following Benjamini–Hochberg FDR correction for multiple comparisons, these associations did not meet the formal threshold for statistical significance (P_FDR = 0.057 for both scales). Post hoc power analyses demonstrated statistical power values of 0.122 for the BDI model and 0.154 for the CESD-R model, both substantially below the conventional threshold of 0.80. These findings indicate that the study was underpowered to retain significance after stringent correction for multiple testing, particularly given the modest sample size (n = 52).

As an exploratory pilot investigation, the relatively high homogeneity of this military cohort, in which participants shared highly similar living conditions and daily schedules, may partially reduce residual confounding that often complicates observational analyses in general population cohorts. Moreover, the consistency of the observed trend across two independent depression measures suggests the presence of a potentially meaningful pilot signal.[31] Accordingly, the borderline statistical findings may reflect a genuine, albeit preliminary, association between subclinical auditory alterations and emotional well-being rather than a simple type I error. These observations warrant further investigation in larger, adequately powered longitudinal cohorts to clarify the magnitude, reproducibility, and generalizability of these associations.

CONCLUSION

As a prospective pilot cohort investigation, this study suggests that even 8 months of military aircraft noise exposure may be associated with subclinical auditory threshold shifts and modest deterioration in psychological well-being among newly enlisted soldiers. Threshold elevation at 4000 Hz emerged as a preliminary signal potentially associated with increases in depressive symptom scores after adjustment for occupational noise exposure and shift-work status. Given the modest sample size and substantial attrition, these findings should be interpreted as exploratory rather than definitive. Nevertheless, the present results provide an initial framework for understanding the early interplay between sensory alterations and emotional well-being before the development of overt functional impairment. Further large-scale longitudinal investigations are warranted to confirm these associations. In parallel, incorporation of integrated auditory and psychological surveillance into military occupational health programs may provide meaningful benefits for the comprehensive welfare of military personnel.

Availability of data and materials

The datasets generated and/or analyzed during the current study are not publicly available due to military confidentiality and security regulations, but are available from the corresponding author upon reasonable request.

Author contributions

K. Park: Conceptualization, Methodology, Investigation, Formal Analysis, Writing – Original Draft, Writing − Review & Editing. S. Lee: Conceptualization, Methodology, Writing – Review & Editing. H. Kim: Formal Analysis, Writing – Review & Editing. H. Cho: Conceptualization, Methodology, Writing – Review & Editing.

Ethics approval and consent to participate

The study protocol was approved by the Institutional Review Board of the Republic of Korea Air Force Aerospace Medical Center (IRB No. ASMC-23-IRB-010). All participants provided written informed consent before enrollment, with the assurance that they could withdraw from the study at any time without penalty. Furthermore, the study was prospectively registered with the Clinical Research Information Service (CRIS), a primary registry of the WHO International Clinical Trials Registry Platform (Registration No. KCT0010089). Additionally, the study and its resulting manuscript underwent and passed a formal security review by the relevant Republic of Korea Air Force authorities to ensure compliance with military information security protocols.

Conflicts of interest

The authors declare that there are no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding Statement

This study was financially supported by Chonnam National University Hospital Biomedical Research Institute grant (No. BCRI26084).