Study design and population

This study was conducted on the basis of an occupational cohort of the stone-processing industry in Hualien, Taiwan. Most of the workers in the industry are self-employed or work in small scale enterprises with less than five workers. They are typically insured through the stone labourers’ union in Hualien. Funded by the Taiwan Council of Labor Affairs, we conducted a basic occupational health service programme for the stone workers between 2006 and 2015.12 A cross-sectional study was conducted in the occupational cohort between 2009 and 2010. The establishment of the cohort was described in our previous study.13 In brief, all of the stone workers insured through the stone labourers’ union in Hualien were eligible for the study. We mailed information regarding the study to 1902 members of the stone union. A total of 344 participants chose to participate in the study and received questionnaire survey and health examinations. We made at least three phone calls to each non-respondent to clarify their reasons for not participating. Among the non-respondents, 43% could not be contacted, 30% had conflicts with our scheduled times, 8% were receiving regular health examinations and had no desire to be examined again, 2% could not have an examination due to frailty, 1% believed that they were healthy and did not wish to receive an examination, <1% could not receive an examination due to pregnancy, <1% responded that the examination did not meet their needs, and 17% provided no reason. The reasons for non-participation were not related to existing hearing diseases.13

Walk-through surveys

A total of 47 walk-through surveys of stone-processing workplaces were completed in order to provide a general risk assessment of the industry between 2006 and 2009. The types of jobs, work environments, processing procedures, materials used in each procedure and final products were recorded. On the basis of the walk-though survey, there are seven types of jobs in the stone-processing industry in Hualien, including processing gemstones, crafting large decorations, making vases, making urns, crafting small decorations, making sculptures and manufacturing building materials. The occupational hazards in the stone industry of Hualien include: (1) Safety hazards: the risk of being pulled into the grinding machine, cutting, falls, eye splash injuries; (2) Physical hazards: noise, use of vibrating tools such as drillers, hand-held carving machines or pneumatic hammers; (3) Ergonomic hazards: repetitive movements and lifting heavy stones; (4) Chemical hazards: dusts, oxalic acid, epoxy resin and organic solvents. Epoxy adhesives are commonly used to fix rocks to lathes, to stick different stones together, to caulk cracks, and to coat and strengthen the surfaces of stones when they are carved, shaped, ground or polished. Stone workers use numerous types of epoxy adhesives, and the containers are not labelled. Stone workers generally call them ‘AB glue’ without knowing their components. The mixing and application of epoxy adhesives are usually performed using simple equipment without ventilation. Solvents could be inhaled and affect the workers’ health. However, personal protective respiratory and hearing equipment are seldom used among these stone workers. The reports of the walk-though surveys were discussed in expert committees that included occupational medicine physicians, industrial hygienists, public officers of labour affairs and labour leaders. Representative workplaces for each type of job were selected for subsequent noise measurements and air sampling.

Noise exposure assessment

Between 2006 and 2009, we conducted a total of 40 representative noise assessments in 15 representative workplaces that covered all types of jobs for the study participants (table 1), including one gemstone-processing workplace, one that constructed large decorations, three vase-producing workplaces, two urn-producing workplaces, five that made small decorations, one that focused on sculptures, and two that manufactured building materials (see online supplementary table S1). This study used ANSI S1.25-1991 standard personal noise dose meters (TES-1355, TES Electrical Electronic Corp., Taipei, Taiwan) with the A-scale and slow response mode settings. The A-scale pressure levels of sound across the frequency spectrum (20–200 000 Hz) are similar to the human hearing response.14 The audio dosimeter microphones were clipped to workers’ collars. Doses were calculated on the basis of the actual exposure times at given sound pressure levels divided by the allowable exposure times at those levels. An equivalent sound pressure level in dBA for an 8 h workday was calculated on the basis of the integrated dose obtained from the entire work shift.

Air sampling and analyses of volatile organic compounds

To determine the contents of volatile organic compounds (VOCs) in the stone industry, field ambient air sampling was conducted at 40 workplaces from 20 June 2014 to 6 October 2014. Given the uncertainty regarding the contents of epoxy adhesives used in the stone industry and since the contents of the adhesives used in different types of jobs or procedures may differ, this study used the TO-15 method to measure the air samples from the workplaces rather than traditional industrial hygiene sampling for particular organic compounds. The TO-15 method is a comprehensive method that has been adopted by the USA Environmental Protection Agency (US EPA) as a reference method for measuring mixtures of VOCs. All of the samples were obtained using the Entech Bottle-Vac canister (Entech Instruments Inc., Simi Valley, California, USA), a negative pressure device, with standardised procedures. A vacuum pressure of −30 mm H₂O was reached using a Micro-QT Valve Vacuum Check Gauge before each sample was obtained. A total of 1 L of ambient air was introduced into the evacuated canister using the negative pressure for 30 s. All of the canisters were immediately sent to the Green Energy & Environmental Research Laboratories of Industrial Technology Research Institute (ITRI) for analysis. Gas samples were analysed using an Entech 7500A Robotic Headspace Autosampler attached to an Entech 7150 Air/Headspace Preconcentrator (Entech Instruments Inc., Simi Valley, California, USA) coupled to an Agilent 6890N gas chromatography (GC)/5975C mass spectrometer (MS) (Agilent Technologies, Santa Clara, California, USA). An internal standard that included bromochloromethane, chlorobenzene-d₅ and 1,4-difluorobenzene at 10 ppmv each in humidified zero air was added to each sample and calibration standard to confirm the stability of the MS and to quantify detected compounds. The internal standard was introduced into the trap at the time of collection for all of the calibration, blank and sample analyses. The volumes of the internal standard spiking mixture added to each analysis were the same from run to run. Internal standard responses and retention times were evaluated immediately after data acquisition. If the retention time for any internal standard changed by greater than 20 s from the mean retention time over the initial calibration range or if the area response for any internal standard changed by greater than ±40% between the sample and the most recent valid calibration, the GC/MS system was inspected and corrected. The chromatographic data processing was performed manually by an experienced analyst using Agilent Chemstation Software (Agilent Technologies, Santa Clara, California, USA). Using the high sensitivity GC/MS analysis, hundreds of chemicals may appear in chromatography. In the analysis, mass spectrum peaks that could not be properly identified due to low library matches or a lack of confirmation by the retention time were discarded. We retained only chemicals with quality values greater than 90%. The values of VOCs below the limit of detection (LOD) were replaced with the square root of 2.15 Descriptive statistics, including means, SDs, medians, and minimum and maximum values, were obtained to characterise the distributions of the concentrations of VOCs.

Occupational history

The occupational history questionnaire was designed on the basis of the reports of the walk-through surveys and had been discussed by the expert committee that includes occupational medicine physicians, industrial hygienists, public officers of labour affairs and stone worker leaders. After expert validation, the questionnaire was pretested in 10 stone workers. Any ambiguities were corrected, and a standardised set of instructions was documented. In-person interviews were conducted between September 2009 and September 2010. The participants’ occupational history, including their companies, types of jobs and durations of employment, were verified against labour insurance records. The questionnaire asked about (1) the participant's occupational history of stone working. Workers were asked about the types of jobs and tasks they performed as well as the products they made. Participants were also asked about the years they began and stopped working, their actual cumulative years at specific types of job, average daily working hours, average numbers of workdays per week, subjective noise exposure at work, and use of epoxy adhesives. Additional questions focused on (2) whether they had a history of working in mining, tunnel drilling and construction, sandblasting workshops, cement factories, pulp factories or rock-crushing plants (these are the main noisy industries in Hualien) and (3) each participant's use of protective gear at work, such as masks, ventilation equipment, earplugs and earmuffs.

Outcome measurements, diagnostic criteria and confounding variables

All of the study participants received a health examination in a Hualien medical centre between 2009 and 2010. The participants completed the questionnaire survey first and then received the health examination, which was scheduled 3 months later to prevent recall bias. The health examination included a physical examination conducted by a physician and a pure tone audiogram (PTA) examination. One occupational physician performed the physical examinations and asked about the participants’ medical history. The physician enquired about symptoms, signs, treatments and any hospital visits and examined the patients’ hospital records to confirm the diagnoses of past diseases. Participants with other ear diseases or occupations that might be related to hearing loss other than stone work were excluded, including: (1) those with a medical history of otitis media, congenital or drug-induced hearing loss, or traumatic ear diseases and (2) participants who had worked in other noisy workplaces capable of causing occupational hearing loss, outside of those that processed stone, for more than 10 years.

Participants were prevented from noise exposure for 14 h before the hearing examination. The physician performed otological examinations on all of the participants to rule out any otic diseases. The PTA was obtained using test frequencies of 0.5, 1, 2, 3, 4, 6 and 8 kHz. Tests at each frequency were performed separately in each ear. Audiometric examinations were administered in a qualified audiometric test room in accordance with the national standard of Taiwan, which states that the background noise levels for audiometric tests must be less than 30 dB across the frequency range and that the audiometric tests must be performed by a certified hearing specialist. An occupational physician determined the presence of stone-related noise-induced hearing loss using the following criteria: (1) a confirmed occupational history of stone carving, (2) confirmation that no other work could result in hearing loss except for processing stone, (3) no history of otitis media or congenital or drug-induced hearing loss, (4) the presence of a notch at 4 or 6 kHz and hearing threshold levels at a frequency greater than 25 dB, and (5) symmetrical hearing loss. After excluding participants with asymmetrical hearing losses, hearing threshold levels at each frequency in the right ear were used for the analyses.

Statistical analysis

The sound levels between different types of jobs were compared using one-way analysis of variance (ANOVA) tests. Multivariate logistic regressions were used to adjust for the confounding effects of age. The age-adjusted ORs for hearing threshold levels of greater than 25 dB at each frequency were calculated. Study participants were stratified into three exposure groups: (1) stone workers with epoxy adhesive exposure, (2) stone workers without epoxy adhesive exposure and (3) administrative staff. The administrative staff was used as the reference group to assess the dose–response relationship for hearing losses greater than 25 dB at each frequency. The Cochran-Armitage test was used to test for trends. To explore the interactions between noise and epoxy adhesives on the hearing loss at each frequency, we stratified participants by epoxy adhesive exposure and compared the ORs for hearing losses of greater than 25 dB. If there were no interactions, we would expect the strength of the association to be the same in each stratum. We also used durations of noise exposure of <20 years as a reference to calculate the OR for hearing loss. Statistical calculations were performed using SAS V.9.2 software (SAS Institute Inc., Cary, North Carolina, USA). For all of the analyses, a two-tailed p value <0.05 was considered statistically significant. The study protocol was approved by the Ethics Committee of the Buddhist Tzu Chi General Hospital (IRB098-50), and all of the participants provided written informed consent.