Objective Hearing loss caused by high levels of noise is a potential occupational health disorder among train drivers around the world. This study aims to investigate the relationship between tunnel driving occupational environment and hearing loss in train drivers, to provide some insights into helping reduce hearing loss among train drivers.
Methods This study analysed cross-sectional data for 1214 train drivers who work at China Railway Guangzhou Group. Health examination was taken by physicians with professional licences, and audiometric testing was performed by health technicians in a sound-isolated room. T/R is defined as the ratio of the length of the tunnels to the length of the railway along drivers’ work routes. Different multivariate models and stratified models were established for sensitivity analysis. A multivariable logistic regression model was used to estimate the ORs of hearing loss associated with tunnel driving occupational environment.
Results The adjusted OR for high-frequency hearing loss in association with the highest T/R levels (30%–45%) versus the lowest T/R levels (<15%) was 3.72 (95% CI 1.43 to 9.69). The corresponding OR for speech-hearing loss was 1.75 (95% CI 0.38 to 8.06). The sensitivity analysis shows our results are suitable for various alternative models.
Conclusions This study found that there was a significant association between tunnel driving occupational environment and hearing loss. Train drivers who work in a higher T/R environment have worse hearing loss.
- health and safety
- occupational health practice
Statistics from Altmetric.com
What is already known about this subject?
Hearing loss is a potential occupational health disorder among train drivers around the world.
Tunnel environments cause higher train interior noise levels and pressure waves, which may influence drivers’ hearing.
What are the new findings?
This is the first study to investigate the associations between tunnel driving occupational environment and train drivers’ hearing loss in China.
A significant association between tunnel driving occupational environment and hearing loss among China’s train drivers was observed in this study.
How might this impact on policy or clinical practice in the foreseeable future?
Railways should apply the system of rotating posts regularly to train drivers to make sure that no train driver works in high T/R routes for a long period of time.
A better hearing health security system for train drivers should be built, especially for train drivers who work in high T/R routes.
Hearing loss is a widespread occupational health disorder around the world, and exposure to noise is the highest risk factor that influences occupational hearing loss.1–3 Some studies show that increased age is the main factor associated with hearing loss.4 5 Other factors such as cigarette smoking, alcohol drinking, hypertension, diabetes mellitus, high cholesterol levels and ototoxic chemical exposure have also been found to influence hearing loss. More recently, genetic variation was also found to be associated with susceptibility to noise-induced hearing loss (NIHL).6–9
For train drivers, noise is a main cause of hearing loss in their workplace.10–12 WHO defined NIHL as a serious public health problem.13 Noise was measured in the cab of a Chinese electric locomotive, where the equivalent continuous A sound level was 88–93 dB(A) and the average was 91 dB(A).14 In Taiwan, the mean value of noise in an electric locomotive driver’s room is 85 dB(A). The noise levels in train drivers’ working environment surpass the standards that were defined by many countries such as the USA.15 16 Furthermore, for safety reasons, train drivers are not allowed to use hearing protection devices against noise while working, such as ear shields and earplugs, because they need to hear and communicate with the radio. These factors increase the risk of NIHL among train drivers.
It has been shown in some studies that train drivers’ hearing loss problems need more attention. In Taiwan, train drivers had worse hearing compared with the general Taiwanese people.17 A sampling survey showed that 15.8% and 3.3% of Iranian train drivers have high-frequency hearing loss and low-frequency hearing loss, respectively.18 Additionally, a cross-sectional study showed that older (≥45 years) Norwegian train and track maintenance workers have a slightly more severe hearing loss compared with reference groups not exposed to noise.10 In contrast, in some studies, the risk of NIHL in train drivers was negligible.11 12
Currently, railways in mountain areas continue to increase. Railway tunnels have become a priority in order to address the diverse terrain in mountains; therefore, there is an increasing number of train drivers operating in railway tunnels. Some studies have found that train interior noise levels are clearly higher in railway tunnels than on the ground. When a train goes through a tunnel, the train is affected by the mixing source because the tunnel functions as a reverberation chamber. Therefore, the train interior noise levels are higher in tunnels than in open-air lines. An experiment on Chinese high-speed trains found that when the train is running at 300 km/hour in a tunnel, the average interior noise level is 8.3 dB(A) higher than that in open air, which is 6.7 dB(A) when the train is running at 350 km/hour.19 The International Union of Railways and Inter City Express Assignment for Technical Design stipulates that the admissible value of train interior noise level in the tunnel is 5 dB(A) higher than that on the ground.20 The higher occupational noise levels result in higher risk for hearing loss. However, as far as the authors know, no studies have focused on the relationship between train drivers’ hearing loss and the tunnel driving occupational environment.
The aim of this study is to analyse hearing loss in Chinese train drivers and examine the association between the tunnel driving occupational environment and the risk of hearing loss. This study used data from the train drivers of China Railway Guangzhou Group (CRGG) who took part in the health examinations and provided self-reports.
This study analysed the train drivers of CRGG. All participants took part in health examinations from which the health information was obtained. Work information (job category and driving environments) was obtained from CRGG. A total of 1614 male train drivers, all of whom drive electric locomotives, participated in the study. All participants are Han Chinese without a family history of ear diseases and got college degree (64%) or bachelor’s degree (36%). The working time of train drivers is 167 hours per month (the Chinese Railway standard). Participants who missed some examinations (n=66), who were trainees (n=236) and who did not have permanent routes (n=98) were excluded from this study. Finally, 1214 participants whose age ranged from 23 to 58 years old were included in the data analysis.
Audiometric testing was performed by health technicians in a sound-proof room using an automated testing mode of the audiometer (Interacoustics Model AD226), a standard supra-aural earphone (TDH 39) and insert earphones (ER-3A). The health technicians were trained and the testing was conducted in accordance with the regulations of Acoustics-Pure tone air conduction threshold audiometry for hearing conservation purposes (Chinese Standards GB7583-87). We computed pure tone average at speech frequencies (0.5, 1 and 2 kHz) and high frequencies (3, 4, 6 and 8 kHz).21 22 Following the standard of evaluation for occupational health testing of train drivers (Chinese Standards TB/T3091-2008), mild hearing loss was defined as an average hearing threshold ≥25 dB for bilateral ears, and moderate hearing loss was defined as an average hearing threshold ≥40 dB.
Definition of tunnel driving occupational environment
Every participant had a fixed work railway route. CRGG provided the work driving routes of the participants. Railway and tunnel data were also obtained from the ‘Locomotive supervision and recording apparatus’ database of CRGG. The total work mileage of all participants was 1949 km, which included 967 tunnels (502 km). T/R is defined as the ratio of the length of the tunnels in drivers’ work routes to the length of the railway in drivers’ work routes. The highest T/R of participants’ work routes was 41.4%, and the lowest T/R was 3.7%. We divided T/R groups into three approximately equal sections by the number of T/R (<15%, 15%–30% and 30%–45%).
Age, cigarette smoking, alcohol drinking, body mass index (BMI), hypertension, diabetes mellitus, fatty liver and nephropathy were considered as potential confounders. In this study, working age was introduced to depict the service year of the participants who have been driving trains. However, it was excluded because there was a direct linear relationship between working age and age. Cigarette smoking and alcohol drinking were defined as dichotomous variables: yes or no. It was reported as ‘yes’ when participants had smoked at least 50 cigarettes in their life.23 24 When participants drank more than a glass of alcohol in 1 week, alcohol drinking was reported as ‘yes’. BMI was calculated by dividing the measured weight in kilograms by the measured height in metres squared, and was divided into three levels: normal (BMI<25), overweight (25≤BMI<30) and obese (BMI≥30). Hypertension was defined as self-reported physician diagnosis, antihypertensive medication, or diastolic blood pressure ≥90 mm Hg or systolic blood pressure ≥140 mm Hg during the health examination. Diabetes mellitus was classified based on fasting glucose ≥7.0 mmol/L during the health examination, self-reported physician diagnosis or self-reported use of medication. Fatty liver was defined as self-reported physician diagnosis or physician diagnosis during the health examination. Renal calculi, hyperuricaemia, high blood urea nitrogen (>0.7.1 mmol/L) and high serum creatinine (>133 μmol/L) during the health examination were considered nephropathy. Total cholesterol (normal: 2.82–5.95 mmol/L; high: >5.95 mmol/L) and triglyceride (normal: 0.45–1.69 mmol/L; high: ≥1.70 mmol/L) during the health examination were also considered as potential confounders. The health examination was performed by physicians with professional titles and were based on the regulations of the Technical Specifications for Occupational Health Surveillance (Chinese Standards GBZ188-2014).
Data were analysed using SPSS V.22.0 for Windows. A multivariable logistic regression model was used to estimate the ORs of hearing loss associated with tunnel driving occupational environment. Speech-frequency hearing loss and high-frequency hearing loss were analysed separately because high-frequency hearing loss is a dangerous factor in human health that has not yet attracted enough attention. A model without adjustment and two adjusted models were developed to assess the influence of potential confounders. Model I was adjusted for age (years) and age2; model II was further adjusted for BMI, hypertension, diabetes mellitus, fatty liver and nephropathy. Age and age2 were fitted to capture the non-linear effect of age. Tests for trends were performed by entering each category of the tunnel driving occupational environment as a continuous variable in the models. P value <0.05 was considered significant.
Different multivariate models and stratified models were implemented to test whether our results are suitable for various alternative modelling.25–27 First, the study examined whether the results differed after further adjustment for total blood cholesterol (normal, high) and triglycerides (normal, high). High blood cholesterol and triglyceride levels can cause inner ear lipid deposition, leading to direct inner ear cell damage which should cause hearing loss. Former research28 proved that cholesterol and triglycerides have a relationship with hearing loss. Second, whether the results differed after excluding adjustment for BMI was examined. Third, age groups were stratified to conduct a sensitivity analysis. Last, different multivariate logistic regression models were built by stratifying every covariate and adjusting for other covariates. It aims to investigate whether the associations between hearing loss and T/R will change in different stratified groups.
Approximately 10.30% and 5.77% of drivers have high-frequency and speech-frequency hearing loss, respectively, which consisted of 42.11% mild hearing loss and 57.89% moderate hearing loss. Hearing loss is made up of 26.90% unilateral hearing loss and 73.10% bilateral hearing loss. The range of working age is 3–37 years, with an average of 17.7 years. Participant characteristics and distributions of tunnel driving occupational environment are presented in table 1. Among all the participants, T/R levels are higher in subjects with hypertension than in those without hypertension. Participants with higher T/R levels are older.
Table 2 shows the logistic regression results for the risk of hearing loss (high-frequency and speech-frequency) in a tunnel driving occupational environment among train drivers. For high-frequency hearing loss, the ORs comparing the highest T/R level and the middle T/R level versus the lowest T/R level were 3.72 (95% CI 1.13 to 9.69) and 2.68 (95% CI 1.04 to 6.88), respectively, in the fully adjusted model (model II). The ORs for doubling the T/R levels are 1.61 (95% CI 1.16 to 2.22). Trend tests for T/R levels are significant in all models (p trend=0.0003, 0.0024, 0.0039). For speech-frequency hearing loss, the fully adjusted ORs comparing the highest T/R level and the middle T/R level versus the lowest T/R level are 1.75 (95% CI 0.38 to 8.06) and 4.65 (95% CI 1.09 to 19.84), respectively (p trend=0.2214).
Table 3 shows the results of the sensitivity analysis. Associations remained unchanged when comparing the original model with the adjusted models. Moreover, age groups (23–35 years, 35–49 years, 50–58 years) were divided to perform stratified sensitivity analysis. Because all of the drivers with speech-frequency hearing loss from the lowest T/R level group were only in the age range of 23–39 years, age-stratified sensitivity analysis was performed only for high-frequency hearing loss. Drivers in the age groups 23–39 and 40–49 years have stronger associations with high-frequency hearing loss, and drivers between the ages of 50 and 58 years show no significant associations (p for interaction by age=0.0113).
Figure 1 shows the results (OR and 95% CI for doubling T/R levels) of stratified analysis by all covariates (except diabetes mellitus because of the small sample size). The ORs are the result of the trend test. In the figure, although each group’s ORs are not exactly the same, we can see that the association between high-frequency hearing loss and T/R is still remarkable. In speech-frequency hearing loss, the ORs in different groups are not much different. The results of the stratified sensitivity analysis show different stratification cannot influence our results.
Figure 2 shows the mean hearing levels at a single frequency for different groups. The mean hearing threshold was defined as the average hearing threshold of both the left and the right ears. The audiogram data have undergone quality control. Irrational data were excluded from the samples, including testing errors (n=4), negative slope of threshold values (n=37), and 40 dB differences or above for bilateral ears at the same frequency (n=7). The mean hearing threshold of the different T/R groups stratified by age is shown in figure 2-1 to 2-3. In the same T/R group, the younger age participants have better hearing, and the ‘notch’ shown in the pure tone audiogram in the 3–6 kHz region (8 kHz thresholds better than those at 3–6 kHz) was more obvious. The mean hearing threshold of different age groups stratified by T/R is shown in figures 2-4-2-6. In the same age group, the participants with lower T/R have better hearing, and the pure tone audiogram shows the characteristics of the ‘notch’ in the 3–6 kHz region were similar.
To our knowledge, this is the first epidemiological study to evaluate the associations between tunnel driving occupational environment and hearing loss among train drivers in China. After adjusting for all covariates, a significant association between hearing loss and tunnel driving occupational environment was found. In addition, the results of the sensitivity analysis performed using different multivariate models and stratified models did not influence the significance of the association.
After adjustment of potential confounding factors, the highest T/R level (30%–45%) and the middle T/R level (15%–30%) presented ORs of 3.72 (95% CI 1.43 to 9.69) and 2.68 (95% CI 1.04 to 6.88) for high-frequency hearing loss compared with the lowest T/R level (<15%). Doubling of T/R levels presented an OR of 1.61 (95% CI 1.16 to 2.22), and the p trend was 0.0039, showing that higher tunnel driving occupational environment levels are significantly associated with high-frequency hearing loss. In speech-frequency hearing loss, compared with the lowest T/R levels, the ORs of the highest T/R levels and the middle T/R levels were 1.75 (95% CI 0.38 to 8.06) and 4.65 (95% CI 1.09 to 19.84), respectively. The OR of the highest T/R level was smaller than that of the middle T/R level. However, in contrast to the lowest T/R level, the ORs of higher T/R levels were obvious. The ORs for speech-frequency hearing loss do not increase with T/R as do the ORs with high-frequency hearing loss. Here are some reasons. First, for train drivers, noise is a main cause of hearing loss in their workplace. Damage from noise often begins at the higher frequencies of 3, 4 or 6 kHz, where the ear is more susceptible to noise.29 Second, the study population is rather young, and exposure to noise often takes a long process to affect speech-frequency hearing loss, so the speech-frequency hearing loss caused by noise is minor in our study. In addition, many cases of speech-frequency hearing loss are due to outer or middle ear pathology, not to industrial noise.
The Chinese Railway, especially the Western railway, is under continuous development, and tunnels have become a priority in order to address the diverse terrain in the mountains. Based on statistical data for 2016, China had 14 100 operating tunnels, with a total mileage of 14 120 km. Among the operating tunnels, there were 102 extra-long tunnels (length >10 km), the total length of which was 1411 km (10% of the total mileage), and 9 super-long tunnels (length >20 km), the total mileage of which was 219 km (1.6% of the total operating mileage). In addition, there were 4240 tunnels under construction with a total mileage of 9300 km. In China, many lines have higher T/R levels than the highest T/R level (30%–45%) in our study, such as the Guiyang–Guangzhou Express Rail Link (T/R=53.2%), Sichuan–Tibet Rail Link (T/R=70.2%) and the Qinling Mountains area of the Xian–Chengdu Rail Link (T/R=93.3%). The results of this study showed that there is an association between high T/R level and hearing loss. These results implied that the actual hearing loss of train drivers in China is more serious than this study has shown.
According to previous research,10–12 noise is the major factor that induces hearing loss among train drivers. Continuous exposure to noise may cause the inner ear to work in an overload state. Furthermore, high-intensity noise may cause mechanical impairment of hair cells in the inner ear. Research has shown the train interior noise levels in tunnels are higher than those in the open field. In Korea, a study on Korea Train Express showed that interior noise increases as much as 10 dB(A) in tunnels compared with the train interior noise levels measured in an open field on a viaduct.30 German trains have reported similar results; interior noise increases approximately 10 dB(A) when a train passes a tunnel with a ballasted track at a speed of 200 km/hour.31 Researchers also found a clear link between high speed and high interior noise levels. Compared with a running speed of 300 km/hour, a speed of 160 km/hour has lower interior noise levels by approximately 15 dB(A),30 except that the noise level in the train cab is more acute than that in other cars.32
The train tunnel pressure wave is another possible reason for the observed association between tunnel driving occupational environment and hearing loss. When the train passes through the tunnel at high speed, a tunnel pressure wave is generated because the air flow is changed by compression waves and expansion waves, which thereby induce a drastic pressure change in the tunnel. Therefore, this induces barometric pressure fluctuations with an amplitude of as much as 5 kPa (the results of vehicle experiments). As the running speed of the train increases, the pressure amplitude inside also rises. Pressure changes can result in different levels of tinnitus for passengers and even rupture the tympanic membrane, which has a serious impact on drivers’ hearing.33–35
In this study, age is potentially an important confounding factor because it is a main risk factor for hearing loss.4 5 This study adjusted for age (continuous covariate) and age2 (continuous covariate). The audiogram in age-related hearing loss (ARHL) is typically downsloping with progressively worsening thresholds in higher frequencies.36 As shown in figure 2, for the 50–58 age group with the highest T/R, the characteristic of pure tone audiogram that 8 kHz thresholds is worse than those at 3–6 kHz, which is similar to ARHL. This characteristic is the result of a synergy between ARHL and NIHL. In contrast, in the young age group, the ‘notch’ is obvious, which is typically created by NIHL in the high frequencies of the audiogram.29 In addition, stratified analysis revealed that the relationship between high-frequency hearing loss and tunnel driving occupational environment only occurred among participants under the age of 50. This observation contrasts with ARHL, which mainly affects older people. Therefore, age is unlikely to confound the association in this study. Besides age, cigarette smoking, alcohol drinking, BMI, hypertension, diabetes mellitus, fatty liver and nephropathy were adjusted in the final models (model I, model II). The associations were not changed in different multivariate models and stratified models for sensitivity analysis.
Some strengths of the study were considered. This study had strong support from CRGG, which provided us credible railway data from China. Moreover, the validity of our study is supported by accurate measurements of health technicians in order to determine hearing loss. Furthermore, for convenience in comparison, all the participants work for the same locomotive depot in CRGG and have similar living environments, without racial differences. In addition, due to examination at entry, all participants had healthy hearing without a family history of ear disease before working. The characteristics of participants can help us explore the intrinsic associations between tunnel driving occupational environment and hearing loss of train drivers.
Despite these strengths, there were some limitations to this study. First, this study is cross-sectional, so it is difficult to establish a causal association between tunnel working occupational environment and hearing loss, but all participants had healthy hearing before exposure due to examination at entry, which decreases confidence in reverse causality to some extent. Second, if there were a few train drivers who changed their work due to serious hearing loss, it might have led to underestimating the severity of the hearing loss. Furthermore, no data were available for ambient noise exposure or for genetic defects. Finally, we cannot ensure that all participants offered accurate self-reports, although we explained the reasons for the study questions.
In summary, this analysis of Chinese train drivers found that there was a significant association between tunnel driving occupational environment and hearing loss. Known risk factors, including various noise exposures and clinical risk factors, were adjusted to prove the risk of hearing loss was independent. Our findings have significant occupational health implications for train drivers. Railways should increase hearing protection of train drivers whose work routes have high T/R levels. If possible, train drivers’ work routes should be changed at regular intervals to prevent train drivers from working in high T/R level work routes all the time. Further studies with more participants and prospective designs will be needed to confirm the causality of the association.
We thank China Railway Guangzhou Group for data support.
Patient consent for publication Not required.
Contributors YP and CF conceived the idea of the study and were responsible for study design and statistical protocol. SY, PX and FW were responsible for collection of participants' data. LH and SP were responsible for data analysis, tables and graphs. All authors contributed to the interpretation of the results and approved the final manuscript.
Funding This work was supported by the National Natural Science Foundation of China (51405517, U1334208), the Natural Science Foundation of Hunan (2015JJ3155), the China Postdoctoral Science Foundation (2015M570691) and the Hu-Xiang Youth Talent Program (2018RS3002).
Competing interests None declared.
Ethics approval Approval for this study was obtained from Xiangya No 2 Hospital of Central South University Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.