Predictive factors of occupational noise-induced hearing loss in Spanish workers: A prospective study.

The purpose of our study was to identify the main factors associated with objective noise-induced hearing loss (NIHL), as indicated by abnormal audiometric testing, in Spanish workers exposed to occupational noise in the construction industry. We carried out a prospective study in Tenerife, Spain, using 150 employees exposed to occupational noise and 150 age-matched controls who were not working in noisy environments. The variables analyzed included sociodemographic data, noise-related factors, types of hearing protection, self-report hearing loss, and auditory-related symptoms (e.g., tinnitus, vertigo). Workers with pathological audiograms had significantly longer noise-exposure duration (16.2 ± 11.4 years) relative to those with normal audiograms (10.2 ± 7.0 years; t = 3.99, P < 0.001). The vast majority of those who never used hearing protection measures had audiometric abnormalities (94.1%). Additionally, workers using at least one of the protection devices (earplugs or earmuffs) had significantly more audiometric abnormalities than those using both protection measures simultaneously (Chi square = 16.07; P < 0.001). The logistic regression analysis indicates that the use of hearing protection measures [odds ratio (OR) = 12.30, confidence interval (CI) = 4.36-13.81, P < 0.001], and noise-exposure duration (OR = 1.35, CI = 1.08-1.99, P = 0.040) are significant predictors of NIHL. This regression model correctly predicted 78.2% of individuals with pathological audiograms. The combined use of hearing protection measures, in particular earplugs and earmuffs, associates with a lower rate of audiometric abnormalities in subjects with high occupational noise exposure. The use of hearing protection measures at work and noise-exposure duration are best predictive factors of NIHL. Auditory-related symptoms and self-report hearing loss do not represent good indicators of objective NIHL. Routine monitoring of noise levels and hearing status are of great importance as part of effective hearing conservation programs.

Introduction

It is well documented that occupational noise exposure is associated with permanent hearing loss. Noise-induced hearing loss (NIHL) represents a major occupational health hazard worldwide.[1234567] In fact, hearing loss is the third most prevalent condition among older adults, with over 11% of the population being affected.[8] NIHL is particularly common in several sectors with exposure to high noise levels including the construction and manufacturing industries.

NIHL has an insidious onset and may be well advanced by the time that it gives rise to appreciable disability.[9] Pure-tone audiometric testing is used to detect and quantify the degree of NIHL. This provides an objective measure of hearing impairment in individuals exposed to occupational noise.[10] Self-report hearing problems and a physical examination are sometimes used for detection of NIHL at the workplace. However, hearing complaints do not always seem to be associated with early hearing impairment.[11] In addition, although auditory or vestibular symptoms such as tinnitus and vertigo are thought to be related to early NIHL, these symptoms often represent a heterogeneous group of underlying disorders.[12] Furthermore, long-term exposure to noise may cause vestibular symptoms before clinically detectable hearing loss occurs. However, the symptoms are subtle and mostly neglected and do not affect the functional ability of workers.[13] Thus, the use of audiometric testing is of great value for an early diagnosis of occupational hearing loss, especially in high-noise environments.

Although noise controls are regarded as the best strategy for NIHL prevention, many hearing conservation programs rely on the use of hearing protection devices to reduce noise exposure.[13] Nevertheless, hearing conservation programs often fail in the construction industry,[14] and underuse of protection has been partly attributed to inadequate company efforts.[15] In fact, in some workplaces with high noise levels in Spain many workers have been found to be reluctant to use hearing protection devices.[16] Recent evidence raises concern about the effectiveness of hearing protection as a substitute for noise control to prevent occupational NIHL. For instance, Groenewold et al. assessed the association between self-reported hearing protection use at work and incidence of hearing shifts in 19,911 workers over a 5-year period. The measures used suggested a marginally significant difference for hearing shift between workers who reported rare versus frequent use of hearing protection devices.[17]

The purpose of our study was to assess objective occupational NIHL, as measured by audiometric testing, and determine its relation to subjective sensation of hearing loss as well as to auditory and vestibular symptoms in Spanish workers exposed to occupational noise in the construction industry. In addition, we aimed to determine whether the use of hearing protection devices such as earplugs and earmuffs, alone or combined, offer a differential benefit to prevent NIHL in noise exposed workers. We were also interested in identifying main predictive factors of NIHL in this sample controlling for confounding variables.

METHODS

Subjects

We carried out a prospective study using 150 employees exposed to occupational noise (mean age: 40.9 ± 10.3 years) in Tenerife, Canary Islands, Spain, from September 2009 to August 2010, who were seen at the Health Surveillance Center of Adeje district. For comparison, we also recruited 150 subjects (mean age: 41.2 ± 11.4 years) who were not working in noisy environments. The occupational sectors that the workers belonged to are shown in Table 1. Subjects working in the construction industry were included in the exposed group, whereas those working in hotel/restaurant's offices were included in the unexposed group. All subjects signed a written informed consent to participate in the study. We excluded workers with a personal or family history of congenital deafness, ear surgery, prolonged exposure to ototoxic agents (e.g., antituberculosis agents, salicylates, aminoglycoside antibiotics, carbon monoxide, lead, and benzene), a history of hypertension for more than 5 years with poor control or blood pressure values higher than 140/90 mgHg at the time of the assessment. Subjects with a history of poorly controlled diabetes mellitus for more than 5 years, alcoholism, moderate or severe head trauma, mumps and measles, and typhoid fever were also excluded. Based on these criteria, a total of 27 workers were excluded from the study (exposed, n = 15; unexposed, n = 12).

Table 1

Demographic and clinical data in workers exposed and unexposed to occupational noise

Variables	Exposed (n = 150)	unexposed (n = 150)
Age (mean ± SD)	40.9±10.3	41.2±11.4
Gender (F/M)	6/144	79/71
Abnormal audiometry [n (%)]	87 (58)*	31 (20.6)
Tinnitus [n (%)]	16 (10.7)**	4 (2.7)
Vertigo [n (%)]	19 (12.7)	11 (7.3)
Self-report hearing loss	11 (7.3)	0
Alcoholism [n (%)]	62 (41.3)	43 (28.7)
Smoking habit [n (%)]	70 (46.7)	48 (32.0)

*P < 0.001, **P < 0.005

Data collection

For each subject, data of audiometric tests were collected. We also applied a questionnaire with sociodemographic information, smoking and alcohol habit, employment history, current noise exposure, hearing protection use, auditory-related symptoms (e.g., tinnitus, vertigo), and self-assessment of hearing loss. The questionnaire is derived from the protocol for health surveillance of workers exposed to noise of the Spain Ministry of Health.[18]

Audiometric test

Hearing ability was measured using pure-tone audiometry. This test was performed with a standard, calibrated audiometer (Redus 75®, Eymasa, Barcelona, Spain). Air-conduction hearing thresholds were explored at 1,000 Hz, 2,000 Hz, 3,000 Hz, 4,000 Hz, 6,000 Hz, and 8,000 Hz frequencies in both ears, in 5 dB increments. To diagnose NIHL, we used the Klockhoff-modified criteria.[19] Based on these criteria, NIHL was defined as having a history of occupational noise exposure, bilateral hearing impairment, and a threshold level higher than 25 dB at frequencies between 1,000 Hz and 8,000 Hz in the absence of other conditions affecting hearing. Environmental noise levels were not measured.

The self-assessment of hearing loss was determined if the subjects answered affirmatively to three out of the five questions included in the questionnaire recommended by the Protocol for health surveillance of workers exposed to noise of the Spain Ministry of Health.[18] Similar questionnaires have been used in other studies.[20] The questions cover the following areas:

Subjective hearing disturbance,

Difficulties in hearing in a crowd or in a noisy environment,

Need to ask others to repeat frequently during a conversation,

Need to turn the volume up of the TV higher than that others would prefer,

Trouble knowing where sounds are coming from.

Statistical analysis

Descriptive statistics for categorical variables was performed by frequency analysis. To analyze the distribution and percentages for categorical variables in relation to noise exposure (exposed or unexposed) or type of audiometry (normal or pathological), contingency tables were used, and statistical significance was determined with the Chi-square test. The difference in continuous variables (age, duration of noise exposure) across groups was determined by the Student's t-test.

To identify predictors of hearing loss, a logistic regression analysis was conducted, with “audiometry” as the dependent variable. The independent variables included demographic data, factors relevant to hearing loss, and possible confounding variables (e.g., smoking, alcohol consumption). For the regression analysis, continuous variables became ordinal variables (e.g., age groups). OR with 95% CIs were obtained. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) 19 software (IBM, Armonk, NY).

Results

The demographic and clinical data of the workers in our sample are given in Table 1. We noted that there was no significant difference in age across exposed versus unexposed groups. The assessment of audiometric tests revealed a predominance of pathological audiograms in the group of workers exposed to occupational noise compared to those who were unexposed. This difference was statistically significant (Chi square = 43.81, P = 0.00).

The relationship between auditory symptoms (e.g., tinnitus and vertigo) and the exposure to occupational noise was explored. The number of subjects exposed to noise reporting tinnitus (n = 16; 10.7%) was significantly higher compared to those who were unexposed (n = 4; 2.7%) (Chi square = 7.71; P = 0.005). The incidence of vertigo in noise-exposed (n = 19; 12.7%) and unexposed (n = 11; 7.3%) subjects showed no significant between-group difference (Chi square = 2.37; P = 0.124) [Table 1].

The analysis of reports obtained from the self-assessment of hearing loss, all noise unexposed workers reported no abnormality, while 7.3% of workers exposed to occupational noise reported subjective hearing impairment (Chi square = 11.42, P = 0.001). Table 1 shows the percentage of workers reporting alcohol intake and smoking habit in both noise-exposed and unexposed groups.

Workers exposed to noise had a mean of 13.2 ± 9.2 years of exposure. A high percentage of these workers (88.6%) regularly used some form of hearing protection, including earplugs and/or earmuffs. Out of these workers, 35.3% simultaneously used both personal protective devices, while 53.3% used only one of these devices. About 11% of these workers exposed to occupational noise reported that they never used any kind of hearing protection [Figure 1].

Figure 1

Effects of hearing protection devices on pathological audiometry. Distribution of workers exposed to occupational noise according to (a) The use of hearing protection devices and (b) Audiometric results. The bottom panel (table) shows the distribution of workers using hearing protection devices (earplugs, headphones) in isolation or in combination

We analyzed the relationship of noise-related factors and the results of the audiometric test (normal or pathological). We found that workers with pathological audiograms had significantly longer noise-exposure duration (16.2 ± 11.4 years) relative to those with normal audiograms (10.2 ± 7.0 years; t = 3.99, P < 0.001). The vast majority of those who never used hearing protection measures had audiometric abnormalities (94.1%), whereas in those using hearing protection there was a similar distribution of normal (46.6%) and pathological (53.4%) audiograms [Figure 1]. The difference in pathological audiograms between these groups (never vs regular use of protection) was statistically significant (Chi square = 10.26; P = 0.001). An exploratory analysis within workers using hearing protection devices revealed that those using at least one of the devices (earplugs or earmuffs) had a 67.5% of pathological audiometry, which contrasts with those using both measures of protection simultaneously that only had a 32.1% of audiometric alterations (Chi square = 16.07; P < 0.001) [Figure 1].

Logistic regression analysis was performed using audiometry (normal or pathological) in occupational noise-exposed group as the dependent variable. The nominal independent variables (predictors) that were entered into the logistic regression model were gender, hearing protection use, incidence of vertigo or tinnitus, and recreational noise exposure. Continuous independent variables were age, duration of noise exposure, number of cigarettes/day, and the daily amount of alcoholic beverages [Table 2]. This logistic regression model was found to be significant (Chi square = 68.79, df = 8; P < 0.001). The results of the regression analysis revealed that significant predictors of occupational hearing loss in workers exposed to noise were the use of hearing protection measures at work (OR = 12.30, CI = 4.36-13.81, P < 0.001) and the duration of noise exposure (OR = 1.35, CI = 1.08-1.99, P = 0.040). This logistic regression model correctly predicted 78.2% of individuals with pathological audiograms in this sample, indicating a high level of classification.

Table 2

Multivariate logistic regression analysis. Predictive factors of abnormal audiometry in workers exposed to occupational noise

Variables	β	SE	Wald	P	OR	95%; CI
Age (years)	−0.038	0.030	1.666	0.197	0.963	0.908	1.02
Cigarettes/day	0.021	0.021	0.975	0.324	1.021	0.98	1.065
Alcohol/day (g)	0.009	0.007	1.834	0.176	1.009	0.996	1.023
Noise exposure (years)	0.670	0.033	4.210	0.040	1.359	1.087	1.997
Gender	−0.412	1.092	0.143	0.706	0.662	0.078	5.629
Protection measures	2.510	0.485	26.825	0.000	12.306	4.76	13.814
Vertigo	2.045	1.150	3.162	0.075	7.729	0.811	73.62
Recreational noise	−0.186	0.636	0.086	0.770	0.830	0.239	2.888

Discussion

In this study, we compared workers with high occupational noise exposure and those unexposed to identify whether symptoms commonly related to hearing disturbance (subjective sensation of hearing loss, tinnitus, and vertigo) as well as the use of hearing protection devices are indeed closely associated with objective hearing loss, as measured by audiometric test. We also aimed at determining predictive factors of NIHL after controlling for confounding variables such as age, and tobacco or alcohol consumption. We found that, as expected, workers exposed to occupational noise exhibit a significant increase in pathological audiograms, compared with unexposed individuals. Importantly, although subjective symptoms, including self-reported hearing impairment, and tinnitus were particularly present in workers exposed to occupational noise, only the lack of use of hearing protection measures and a long duration of noise exposure emerged as important predictors of NIHL.

It is well established that hearing loss is the most common problem associated with exposure to noise.[21] The National Institute of Health, United States, reported that nearly 20 million workers were regularly exposed to noise, of which 50% (10 million) suffered some hearing damage of different severity.[22] A study conducted by Wu et al., using a system of health surveillance for hearing loss in Taiwan, found a 58% incidence of hearing impairment among workers exposed to noise.[6] Other studies performing pathological audiometry have found that 53-78% of workers exposed to occupational noise have hearing impairment worldwide.[15232425] In a recent study, Money et al. reported that up to 95% of workers exposed to noise had some sort of audiometric alteration.[3] In our study, we found a 73.7% of pathological audiograms in noisy working environments. This incidence is within the range of the literature.[320] Relly et al. suggested that the total number of workers with hearing impairment is sometimes underestimated.[26] However, the progressive introduction of monitoring systems increasingly identifies jobs with hazardous noise levels that require attention for hearing health status of their workers.

Various auditory-related tinnitus are thought to be associated with hearing impairment. In our study, tinnitus was significantly more frequent in noise-exposed workers than in unexposed workers, although its incidence was lower than expected. Other studies of workers exposed to noise indicate that tinnitus is mainly associated with advanced hearing loss due to occupational noise exposure, while the subjects in this study had mild-to-moderate hearing loss.[2728] Because the presence of self-reported tinnitus was assessed using a simple (yes/no) question, we cannot rule out that the way this question was formulated may have influenced the results. Vertigo, however, was found to have no significant difference between groups. Our results suggest that despite a relatively low frequency of these symptoms in workers exposed to occupational noise, tinnitus rather than vertigo might be related to NIHL. However, the regression model revealed no predictive value of these symptoms for the development of NIHL.

Hearing disorders are often associated with objective manifestations, such as audiometric abnormalities as well as with the subjective sensation of hearing loss. However, it is not clear whether subjective manifestations, particularly in individuals regularly exposed to noise, appear early or not, and if they represent a good indicator of an actual hearing impairment. This would significantly help in the prevention of occupational hearing loss. Although several studies on NIHL are based on subjective symptoms of self-reported hearing loss, there is growing evidence of the need for objective measures of hearing damage, such as audiometry, for early and reliable detection of hearing impairment attributable to the work environment, and thus avoid severe and irreparable damage to hearing.[2930] Kerr et al. studied hearing as perceived by the individual and audiometry in 147 construction workers and 150 farmers in order to promote actions that lead to a reduction of NIHL. The sensitivity of perceived hearing loss compared to audiometric damage ranged 0.30-0.72 suggesting that self-reported hearing loss is not consistent with an actual hearing loss measured by audiometry. The poor relationship between loss of perceived and actual audition found suggests that best practice in assessing this dimension of hearing requires the inclusion of audiometric examination as part of a prevention program hearing loss.[23] Of note, although the percentage of pathological audiometry was 58% in our study only 7.3% of this sample reported subjective hearing loss.

It is well-known that there is a causal association between occupational exposure to noise and a permanent hearing loss.[31] The lack of use of hearing protection devices worsens hearing capacity of individuals exposed to noise.[23]

In our study, we evaluated the appropriate use of personal protective equipment in workers exposed to noise. In the noise exposed group, the majority of workers who did not use protective measures had pathological audiograms (94%). Among those who used these measures, there was a similar distribution of normal and pathological audiometry. When analyzing the combined use of protective measures, it was demonstrated that it has greater efficiency than the isolated use of earplugs or earmuffs, as those using both devices simultaneously had only 32% of pathological audiograms while those who opted for either an earplug or an earmuff had 68%.

Most studies agree that the use of hearing protection devices is a key factor for prevention of hearing loss due to occupational noise exposure, and only these devices, and to a lesser extent rehabilitation, can ensure good hearing health in workers exposed to this physical agent.[323334] Pavón's study on the level of environmental noise in opencast mining in the community of Madrid noted the existence of high noise levels in most workplaces. Nevertheless, there were a high number of workers who were reluctant to use hearing protection.[16] To promote safety and health at work should be emphasized in the training of workers. In this sense, Neitzel et al. conducted a study to assess the importance of training in awareness for the use of personal protective equipment to workers in the construction sector. Their results showed that the training of workers exposed to occupational noise is essential to achieve a good level of conservation of hearing health.[35]

A study conducted by Salazar et al. aiming to determine the cochlear function using otoacoustic emissions showed early and subtle damage in the hair cells of the cochlea that did not correlate with changes in conventional audiometry. The study showed that occupational noise significantly decreased the otoacoustic emissions in subjects exposed to noise who used hearing protectors throughout the workday.[36] Taken together, these studies and our results suggest that hearing protection for workers in noisy environments is the best measure and probably the only and most effective measure in preventing occupational deafness, as it has proven to be the main factor for hearing preservation despite the existence of subtle damage to hair cells of the cochlea.

So far, we have shown results of our study that indicate an association between various factors and hearing loss in exposed workers compared to those not exposed to occupational noise. However, it is not entirely clear whether the same association persists when analyzing all factors together. This approach would help to have a better view of which factors are more determinants to be independently associated with increased risk of hearing damage in individuals regularly exposed to noise at work in this sample. For this reason, we decided to carry out a logistic regression analysis, taking as dependent variable the audiometric damage in workers exposed to occupational noise, and evaluate the predictive power of a number of independent variables. The logistic regression model was found to be significant, indicating that they could correctly predict the dependent variable (audiometric damage) in exposed workers. The results indicated that taking all factors into consideration, only a long duration of noise exposure and the lack of use of hearing protection measures emerged as predictive factors of the risk to develop objective hearing impairment, as indicated by the audiometric test. This means that despite the undeniable fact that age-related factors such as presbycusis affect the quality of hearing in subjects working in any kind of environment, sustained noise exposure plays an essential role in NIHL, in addition to a lack of use of protective measures at work. The logistic regression model correctly predicted about 80% of individuals with normal and pathological hearing in the sample, which shows a high level of classification.

Most previous studies proposed that the duration of noise exposure and a bad use of hearing protection devices negatively influenced the audiometric results. Hong et al.'s study assessed the prevalence and characteristics of hearing loss among engineers operating with heavy construction machinery using audiometric tests as part of a protection program. They found that the rate of hearing loss was particularly higher among subjects with more years of work experience in the construction industry. The average use of hearing protection devices was 48% of the time that is required to be used. A significant inverse correlation between the highest frequency (4-6 kHz) hearing loss and use of hearing protection devices (r = –0.134, P < 0.001) was observed.[1] In a study similar to ours, it was found that hearing thresholds among workers of a nuclear center were much higher than those observed in the population of industrial workers with low noise exposure. A review of noise and health indicate that prevention of the effects of occupational noise on health should focus on measures to protect workers rather than on the control of the source of noise.[37]

The objective risk of hearing damage was significantly higher for construction workers compared with controls (OR = 1.6, 95% CI = 1.3-2.1) and increased with the duration of time employment, although the confounding effect of other factors such as age and smoking was not ruled out.[38] Interestingly, Rubak et al. found that workers exposed for more than 20 years at a level of exposure above 85 dB (A) had a threefold increased risk (OR = 3.05, 95% CI = 1.33-6.99) of developing NIHL, which is consistent with our results. Workers who started to work in a noisy environment in the last 10-15 years or workers under 30 years of age showed no increased risk of hearing impairment.[31]

In conclusion, our findings support and extend the following notions:

Workers exposed to occupational noise are particularly prone to objective audiometric damage,

Auditory-related symptoms such as tinnitus are weakly associated to NIHL, and

Self-report hearing loss does not represent a good indicator of an early objective audiometric damage.

In addition, the results of logistic regression analysis indicate that the main predictors of occupational hearing loss were first the use of hearing protection measures at work and second the duration of occupational exposure to noise. Furthermore, the combined use of hearing protection measures, in particular earplugs and earmuffs, are associated with a lower rate of audiometric abnormalities in subjects with high occupational noise exposure. Thus, routine monitoring of noise levels and hearing status in certain populations should be included as part of a program effective hearing conservation.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.