Reproduced with permission
from Acoustics Australia,
26(2), 57-61, 1998.
Occupational Noise-Induced Hearing Loss:
Its Origin, Characterisation and Prevention
Hearing Loss Prevention Research,
National Acoustic Laboratories,
126 Greville Street, Chatswood, NSW 2067.
Permanent hearing loss due to noise exposure constitutes premature aging of the ear caused by depletion of the outer hair cell population. Describing it is complex because many other factors also contribute to this depletion. Managing it is still more difficult because reduction of sound levels reaching the ear is not an adequate strategy by itself. Adequate prevention of any disability is only afforded by predetermination of individual risk coupled with comprehension of its severity. Otoacoustic emission data show that traditional hearing tests have neither given early warning, nor has the terminology ‘mild’ hearing loss indicated that extensive cochlear damage has accumulated.
Keywords: Noise-induced hearing loss, outer hair cells, prevention, early warning, industry standard.
The cause of noise-induced hearing loss is, by definition, over-exposure to loud sound. The condition was first described over a hundred years ago when Dr. Thomas Barr of Glasgow realised that boiler-makers suffered premature loss of hearing. In modern times the condition is regarded as a very complex problem and it costs our country tens of millions of dollars each year in direct compensation costs (Macrae, 1998) and indirect legal costs as well as all the social consequences of poor communication at a personal level.
The primary factor responsible for Noise-Induced Hearing Loss (NIHL) is premature depletion of the three rows of cells in the cochlea called the Outer Hair Cells (OHC). The motor activity of these cells (dubbed the “cochlear amplifier”) is essential to normal hearing. When the OHC are subjected to very loud sounds (120 to 130 dB SPL), the basilar membrane on which they sit can be forced into vibrational amplitudes approaching the size of the cells themselves, causing shear forces rupturing cell membranes; for still louder sounds, producing complete disruption of the surrounding structure. In the mammalian ear new cells do not re-grow — the damage is permanent. Typically the spatial pattern of permanent loss of cells is related to the frequency and level of the sounds. An exposure to third-octave white noise for years will typically result in heavy loss of OHC of up to one tenth the length of the basilar membrane; repetitive impact noise can take out one third the starting population OHC (about 12000 in each ear). This adds to the scattered loss of OHC which occurs with aging beginning from birth, with the cells at the high frequency end being more vulnerable.
Recent research has focussed on the many other mutually potentiating influences (McFadden, 1986a; Morata, 1998) which act upon the ear reducing the population of active OHC. These include hereditary factors (several lines of defective genes are being studied) and the protective presence of melanin in the cochlea (originally assessed using eye colour). Then there are the acquired defects such as due to maternal infection during pregnancy, birth trauma leading to hypoxia, infections, particularly during the first decade of life, plus a whole gamut of toxic influences ranging from heavy-metal poisoning, naturally occurring toxins and commercially-produced chemicals including solvents such as benzine and toluene (Johnson, 1994) antibiotics and loop diuretics. To these we have to add physical injury, due to head impacts and raised barometric pressure. In the past these many effects have been regarded as outside the area of interest. The reason for considering all these “unrelated” effects here is that we now suspect that all these other synergistic factors (let’s lump them together as determining “individual susceptibility”) are swamping the main effect we are trying to measure, confounding attempts to control the rate of accumulation of cochlear damage by setting limits on sound exposure. A second reason the problem is difficult to manage is that we have no way of isolating occupational noise exposure from any other kind of excess sound exposure, e.g. music exposure – it all appears to add up to deplete the OHC population.
2. Characterisation – old and new
Typically the first clinical signs of noise-induced hearing loss are indistinct speech perception, particularly in conditions of raised background noise, while pure tone audiometry first reveals a “noise-notch” at 4 to 6 kHz. It is generally accepted that this dip in sensitivity occurs because the ear canal and drum has a resonance at 3 to 4 kHz emphasising this component of any sound to peak levels at the ear drum of up to +20 dB higher than entering the ear canal and producing a loss of sensitivity at a higher frequency.
By the time a person seeks help for a noise-induced hearing loss the noise-notch may be <25 dB in depth, and the person is accorded typically a 5 percent hearing loss (Macrae, 1998). In traditional compensation parlance the disability is termed “mild” by comparison with possible moderate and severe noise-induced hearing loss. Despite this, it is not the loss of hearing sensitivity which drives sufferers to seek help. Ironically, the most common symptom first presented is the loss of voluntary ability to distinguish between sounds of different location or frequency, particularly under conditions of multiple sources, reverberation or moderately raised background noise. There exist audiometric tests for cochlear selectivity which is essential for voluntary selection (both pure tone masking and speech in noise tests). However, until now this primary and significant form of hearing disability has not only been too time consuming to test, it has been still harder to present in lay or even legal terms.
The inherent difficulty in raising awareness of, and preventing the most common form of hearing loss is describing what the average person wishes they had avoided only after the symptoms of loss of selectivity developed. There is a simple experiment which any person can conduct on themselves which we suspect better describes hearing loss than simply reducing the volume to mimic loss of sensitivity. Turn on the radio onto a talk program and have the volume at normal speaking level. Now try to hold a conversation with someone. Finally imagine the frustration of never being able to turn the radio down in situations of such conflict. Hearing loss is so subtle and so poorly appreciated because the nature of the complaint is qualitatively no different from the normal hearing listener. Quantitatively, however, the presence of competing sound affects the ear much more. We learn from birth to wait for a gap in the conservation before beginning to speak; it is not so much that competing sounds are “masking” what we are trying to “hear”, it is more the case that the ear lacking active OHC processing power and as a result we cannot voluntarily select a particular source of sound and attend only to it. For the person with a problem with selection, if they cannot remove the competing sound, such as in a crowded room, they cannot cope.
The important question investigated at NAL since 1989 is whether the otoacoustic emission technique can provide not just a fast objective measure of hearing ability, but a parameter which better indicates loss of the ability to select in ears with extensive OHC damage. The impetus came from an animal study by Altschuler et al (1992), in which it was shown that while the inner hair cells and just one row of OHC remain intact, hearing sensitivity can remain normal, which suggests that the mammalian ear uses redundancy, or excess numbers of OHC to cope with progressive aging of, and damage to the hearing organ. If such redundancy is demonstrable in humans then a possible correlate may be the net level of motility of the outer hair cells.
A transient otoacoustic emission is the sound re-emitted into the ear canal due to an incident click. Important to this endeavour is the understanding that this stimulus is just large enough to drive all OHC into saturation. The click is supplied preset at 80±1.5 dB SPL peak, and other studies have shown that all active cells in the cochlea should respond at this level and therefore that the net emission should reflect the total number of active OHC. The resulting emission is typically 0 to 10 dB SPL and so signal averaging (sample period of 40 s, duration 20.48 ms) is used to improve the signal to noise ratio by 24 or 30 dB taking about one minute. Also because the click response will be determined by the characteristics of the external ear and middle ear as well, in the standard protocol, a method of differencing is employed such that clicks of two different levels are used and any linear component of the response is subtracted away leaving only the nonlinear response due to the level-dependent change in outer hair cell activity. Also alternate responses are summed into two arrays and the reproducibility between the final averaged waveforms is calculated. If the ear has a fast recovery from the previous click it will respond with high waveform reproducibility (a correlation coefficient of 1.0); if the ear is still recovering it will respond differently and the reproducibility will be lower, towards zero. It turns out that this parameter is a very sensitive measure of net OHC activity of the human ear and can be thought of as speed or “reaction time”. However, being a bounded parameter [-1, 1] and non-normally distributed, the waveform reproducibility is often used to weight the sound level of the emission so that the net response is a sound level. In our experiments we have used a parameter Coherent Emission Strength (CES dB SPL – which is the average sound pressure is multiplied by the square of the reproducibility) to quantify the average reproducible (or coherent) component of the emission sound level.
Figure 1. Comparison of behavioural and objective measures of hearing for 505 ears. The ordinate is a 3 frequency average hearing level (at 1, 2, and 4kHz) as usually plotted in audiograms versus frequency. The abscissa is Coherent Emission Strength (CES dB SPL) – a measurement of the reproducible component of power of the click evoked emission. The heavy line and dots represent the mean value of the hearing thresholds for the appropriate 5 dB band of CES values. The dashed lines represent +/-1 standard deviation about those mean values
By comparing strength of the emission with hearing thresholds for the same frequency range there should be a range of emission strengths over which hearing sensitivity does not change. We subsequently showed (LePage and Murray, 1993) in a study of 505 ears if the strength of the emission is plotted (Figure 1) versus hearing level for the same frequency range (1-4 kHz) it is seen that nearly all cases of hearing loss the emission strength was below a critical value of -3 dB SPL (LePage et al, 1994) or the bottom 20% of the total range of emission strengths. The notable exceptions were cases subsequently confirmed as more central in origin or those from malingerers who did not correctly indicate their true thresholds until advised of the conflict. The range of values of emission strength varies from the highest seen in neonates 38dB SPL down to -12 dB SPL (a range of 50 dB) while test-retest variability is ±4 dB SPL (Murray and LePage, 1997). These results suggest that indeed emission strength may be used as an indicator of redundancy in humans as well and that the rapid, objective, non-invasive click-evoked otoacoustic emission test.
Evidently there is a period of accumulation of latent or subcritical damage during which a person who has had occupational exposure for some years may not be distinguished audiometrically from one who has led a noise-free life. In turn this may explain why in the new standard (AS/NZS1269:1998) emphasis upon monitoring hearing thresholds in occupational workers has been reduced in favour of higher attention to noise-level management. Regular hearing tests not only provide no early warning, they essentially do not measure the parameter which most represents the disability – loss of selection. A 5 percent hearing loss may constitute in excess of 70 percent loss of outer hair cells sensing high frequencies with loss of processing power.
While many hundreds of studies conducted using Transient Evoked Emissions (TEE) have concerned themselves with neonatal screening, Narelle Murray and I have tackled the possibility that TEEs can be used to help quantify the net level of OHC activity in adults, to separate normal aging effect from any accelerated aging effect. Hence we have been questioning why the problem of NIHL is inherently difficult to manage and here too we believe the otoacoustic emission results have shed new light. Borrowing an analogy from signal detection theory, if we regard the effect of loud sound upon hearing as the “signal” we are interested in, and all the other possible causes of hearing loss outside our control as “noise”, the problem is difficult to address because the signal-to-noise ratio may be much poorer than anybody working with NIHL has previously assumed.
Figure 2. A scatterplot of Coherent Emission Strength as a function of age at the time of recording in a population of 2038 people reporting no current hearing problems, left and right ears. The regression lines indicate a slight but significant decline versus age (left below right). The important features are the normally large scatter in values of emission strength and the fact that having low values can occur at any age, reflecting high risk for hearing loss.
Figure 2 shows a scatter plot of CES for teenage and adult subjects between the ages of 10 and 60. We have studied test-retest variability, and in terms of CES, on average, each point on this plot varies only by ±4 dB on test-retest. So whatever uncertainties exist with the TEE technique one thing is quite clear: the range of emission strengths we have measured across the Australian population spans over 40 dB, or five times the test-retest variability, implying that TEE may be giving us a somewhat better than a crude estimation of the total remaining population of OHC, the cells producing the emission.
The other important corollary of this scatter plot, is that because most of the points indicate ears which have no hearing loss (most above the critical value) the TEE technique is potentially providing far more detail about OHC depletion than is pure tone audiometry. If subjects with emission strengths below the critical value are more susceptible to acquiring a hearing loss than those with very high values such as neonates, between 20 and 35 dB SPL then the scatter indicates that many young people are at risk of imminent hearing loss. Indeed we have some work in press (LePage and Murray, 1998) which shows that despite the scatter, there are highly significant effects of certain kinds of noise exposure upon emission strength predominantly in individuals reporting no hearing problems. The sloping lines show the results of a linear regression for left and right ears separately (left below right) and is a highly, significant decline with age. Our current studies also include a cohort in which we are tracking both TEEs and pure tone audiometry for confirmation.
The salient feature of the data in Fig. 2 is that at any particular age, the range of emission strengths is about 80 percent of the total span of 40 dB. These data represent the largest TEE database (2500 people) so far presented in the literature. The high level of scatter implies that there are significant additional sources of variability never previously seen in otoacoustic emission data or alternately discounted. Of immediate concern is that the scatter represents a problem in the measurement technique such as transmission through the middle ear so that the variation is not due to variation in OHC motility (for whatever cause). After nearly a decade of study at NAL we suspect that the scatter in these results irrespective of age is real and not attributable to some form of measurement error or misinterpretation of the origin of the emissions. Our preferred explanation is that the scatter is not of middle ear origin, and even if it were, it is unlikely to account for a variation of in excess of 30 dB in normal listeners when their thresholds do not vary by more than 15dB and that test-retest for the emission level is much less again. The variability is more likely to reflect some individual component of the OHC response such as efferent involvement in the determination of susceptibility or maybe systematic variations in conditions of cochlear regulation. Whatever the case, these factors may affect individual susceptibility by modifying the excess level of OHC motility. Moreover, our study of individual records shows that indeed we have many people in all age ranges whose pure tone audiometry correlates well with the existence of low level emissions.
The interpretation of the scatter (Fig. 2) we are investigating is that it represents high variability in individual susceptibility to hearing loss due to the very many synergistic factors mentioned above. These must be taken into account in any trend analysis in which the independent variable is aging effect, or noise exposure, or effect of toxic substances or head injury and so on. Although our longitudinal epidemiological study has made several assumptions, we have justifiably arrived at the notion of redundancy of OHC function. Since mammalian OHC do not regenerate when permanently damaged it would almost appear that, like many other systems in the human body such as that involved in insulin production, the evolutionary process has arrived at a cochlear structure with considerable excess capacity – very many more OHC than we need to hear normally (or in terms of the cochlear amplifier hypothesis, than we need to maintain adequate gain) so we can afford to lose the greater portion of them before any disability is evident.
The title of the latest Australia/New Zealand Standard AS/NZ1269-1998 has been renamed “Occupational Noise Management” to reflect that more emphasis is being given to reducing sound levels at source and less emphasis given to the monitoring of the onset of hearing loss because the act of monitoring with pure tone audiometry offers no early warning. The rationale is based on the logic that limiting the peak sound levels in the workplace say from LAeq,8h values of 90 dB to 85 dB SPL must limit worker exposure and therefore should produce a reduction in the incidence of NIHL. However, while it is too soon to tell if these latest measures are effective, till recently most efforts to limit sound exposures have not be supported by convincing evidence of a reduction on numbers affected (Royster, 1993). Why? It is simply a problem of more effectively enforcing or motivating employers and workers to conform to guidelines, or is there a more basic reason?
The key to the success of any prevention program is early warning – there are inherent problems in trying to use the same parameter both as a measure of disability and also as a predictor for that disability. Previous standards have specified three basic aims: 1) reduce the level of the noise being produced by machinery or enclose it to keep the sound inside the enclosure, 2) if silencing is not possible to an acceptable level then reduce the level of noise reaching the ear drum with obligatory hearing protection devices (ear muffs or ear plugs) and 3) monitor the hearing levels to identify those at risk for noise damage. Our studies have not only revealed indirect reasons why there has been widespread resistance to reducing sound levels at source, they have also explained why the emphasis upon audiometry component has been dropped in AS1269-1998.
The basic principle which has guided the tradeoff between acceptable sound levels and time of exposure dates from the so-called Equal Energy Hypothesis – a 3 dB increase in sound level equates to halving the maximum duration of exposure, the point of reference now being an LAeq,8h of 85 dB. 88 dB equates to a 4 hour limit and so on, to say, 120 dB at which level the rule limits exposure to only a few seconds. Set in the context of the discussion above, we can see this traditional rule is important for protecting the bulk of the population, but it may do very little for the most susceptible people. Without them being identified and targeted for special attention they will likely still be the first in any program to suffer a hearing loss and so their management program will appear to be ineffective whereas it is only breaking down by failing to detect those most at risk.
Much effort has also been expended on obtaining an adequate method of rating hearing protectors so that the type of device can be matched to the application, not just how its rating must depend upon how they are worn in practice, but taking into account how steeply the rating must be degraded for intermittent use. Because of tremendous variability in real ear attenuation debate continues as to the best method of rating them so that at least most of the population of users has their hearing protected. The predominant rating method in Australia continues as the “SLC80” a nominal so-called “real-world” value of attenuation which derived from the pioneering work of Dick Waugh at NAL. This method of rating is designed to stem hearing loss by protecting the average worker in noise, but our concern here is for workers who may already be at significant risk. This traditional approach may not do much for preserving their hearing.
We are therefore optimistic that the otoacoustic emission approach may be an important adjunct to hearing conservation strategies. Clearly we need to continue to reduce overall rates of accelerated depletion of the OHC population by reducing sound levels, fully realising that irrespective of that measure the most susceptible people will still likely be outside that level of control. Hence we are working towards a new strategy for adoption sometime in the new Millennium. We advocate a two-pronged approach: 1) to reduce sound levels protecting the bulk of the population and 2) introduce the more sensitive method of assessing the level of redundancy in OHC activity providing the capability of using limited resources to target workers most at risk in plenty of time for all concerned to consider all the career choices still available to them.
4. New advances
In addition to these basic strategies some work is being carried out which already has had a profound effect on hearing science, but may soon add to strategies in the field. The first is the notion of fighting fire with fire. A variety of studies have found that the ear is more vulnerable to permanent OHC loss if it goes from quiet into noise (a surprise attack), than if it is first “conditioned” or “toughened” in medium level noise (Canlon et al., 1995). At the Nordic Noise conference in Stockholm in March, Canlon reported that the ear can also be post-conditioned in the same way to reduce the total permanent damage – a sort of “warming down” process for the OHC just like is now recommended form of management of skeletal muscle in athletes.
Perhaps the most important basic discovery presented at Nordic Noise were pictures taken of OHC before and after a loud sound exposure, showing that these cells actually go into a form of cramp or spasm which relaxes only after a period of recovery (Flock et al, 1998). This could be one of the mechanisms responsible for temporary threshold shift. Also, the notion of ear protection coming out of a bottle or a pill is highly revolutionary in an industrial context and will never replace the need for worn hearing protection but two developments show promise. Treating cramping muscles with relaxants is not new, but if the cramping action results in protection, and it occurs because of calcium involvement (e.g. Duan et al, 1998), it is not impossible to conceive this will eventually be manipulated pharmacologically as an adjunct for prevention of hearing loss. A second reason for temporary shift is that excess stimulation poisons the auditory nerve at the point where the fibres connect with the inner hair cells. Puel and Pujol (1998) who showed the effect in the first place have also shown it can be blocked so that noise doesn’t result in a turning off the nerve.
Lastly mention should be made of active hearing protectors which may become cost effective in more general use. Those which show most promise are those in-the-canal active noise reduction devices which constitute a significant improvement over early ear-muff designs. An acoustic signal is introduced into the cavity between the device and the ear drum. Since the volume is much smaller than ear-muff designs cancellation can occur at higher frequencies.
We have shown that the notion of individual susceptibility may be hampering our efforts to show that industrial hearing conservation programs are worthwhile and we should continue to push for reduction of noise levels. However, it is unrealistic to expect to see an effect except in the long term using behavioural measures such as audiometry. Refinement of the new objective techniques such as otoacoustic emissions may provide a better handle on early warning in terms of the notion of assessing cochlear redundancy. If this new approach can eventually be used with more confidence to quantify the population of OHC in any ear, it is possible to conceive it may be used as a general screening tool for early detection such as has been applied to early warning of glaucoma. Finally research into noise-induced hearing loss is leading to some exciting developments both in basic hearing science and in practical field strategies which may eventually substantially change the incidence of premature hearing loss.
Altschuler, RA, Raphael, Y., Prosen, C. A., Dolan, D. F., and Moody, D. B., (1992). Acoustic stimulation and overstimulation in the cochlea: a comparison between basal and apical turns of the cochlea. In: Noise-Induced Hearing Loss, edited by A.L. Dancer, D. Henderson R. J. Salvi and R. P. Hamernik, Beaune, France, Mosby Year Book, 60-72
Canlon, B. and Fransson, A., (1995). Morphological and functional preservation of the outer hair cells from noise trauma by sound conditioning. Hear. Res. 84 (1-2), 112-24.
Duan, M., Canlon, B., and Ernfors, P., (1998). MK-801, an antagonist of the NMDA receptor, can protect the cochlea from noise damage. Nordic Noise PAN 98 Conference Stockholm Sweden.
Flock, Å, Flock, B., Fridberger, A., Scarfone, E., and Ulfendahl, M., (1998). Dynamic change in structure and function of the hearing organ in relation to temporary or permanent sensitivity loss – an electrophysiologic and laser confocal microscopic study. Nordic Noise PAN 98 Conference Stockholm Sweden.
Johnson, A-C., (1994). The ototoxic effect of toluene and the influence of noise, acetyl salicylic acid, or genotype. Scand. Audiol., 23, Suppl 39 (1-40).
LePage, E. L., (1993). Hysteresis in cochlear mechanics and a model for variability in noise-induced hearing loss. In Noise-Induced Hearing Loss edited by A. Dancer, D. Henderson R. J. Salvi and R. P. Hamernik. Mosby Year Book, St.Louis, Chapter 10, 106-115.
LePage, E. L., Murray, N. M., Tran, K., and Harrap, M. J., (1993). The ear as an acoustical generator: otoacoustic emissions and their diagnostic potential. Acoustics Aust., 21 (3), 86-90.
LePage, E. L., Murray, N. M., and Tran, K., (1994). Comparison of otoacoustic emission measures of cochlear damage in the Australian population with hearing loss in the Australian and British populations. In: Better Hearing Australia conference., edited by M. McGrotty, Adelaide, South Australia, 7-11 August, 1994 , Better Hearing Australia.
LePage, E. L. and Murray, N. M., (1993). Click-evoked otoacoustic emissions: comparing emission strengths with pure tone audiometric thresholds. Aust.J.Audiol., 15 , 9-22.
LePage, E. L. and Murray, N.M. (1998). Personal stereo users accumulate latent cochlear damage: A study using click-evoked otoacoustic emissions. Med.J.Aust. (Accepted for publication).
Macrae, J. H., (1998). Workers compensation for industrial deafness. Acoustics Australia, 26, 13-16.
McFadden, D., (1986a). Some issues associated with interactions between ototoxic drugs and exposure to intense sounds. In: Basic and Applied Aspects of Noise-Induced Hearing Loss: Proceedings of a NATO Advanced Studies Institute., edited by R.J. Salvi, D. Henderson R. P. Hamernik V. Colletti, Lucca, Italy, September 23-29, 1985, Plenum Press, 541-550
McFadden, D., (1986b). The curious half-octave shift: evidence for a basalward migration of the traveling-wave envelope with increasing intensity. Basic and Applied Aspects of Noise-Induced Hearing Loss: Proceedings of a NATO Advanced Studies Institute., 295-312.
Morata, T.C., (1998). Assessing occupational hearing loss: beyond noise exposure Scand. Audiol. Suppl. 48, 27 (111-116).
Murray, N. M., LePage, E. L., and Tran, K., (1997). Repeatability of click-evoked otoacoustic emissions. Aust.J.Audiol., 19, 109-118.
Puel, J-L. and Pujol, R., (1998). Perspectives in inner ear pharmacology and clinical implications. Nordic Noise PAN 98 Conference Stockholm Sweden.
Rajan, R., (1995). Involvement of cochlear efferent pathways in protective effects elicited with binaural loud sound exposure in cats. J.Neurophysiol. 74, 582-97.
Royster, J. D. and Royster, L. H., (1993). Significant threshold shift criteria: what works better than OSHA STS? Proc. of the 18th Conference of the National Hearing Conservation Association, Albuquerque, New Mexico. J-Occup-Med. 28, 1055-68.
Standards Australia and Standards New Zealand (1998) Australian/New Zealand Standard AS/NZS 1269. Occupational Noise Management, Standards Australia, Sydney, Australia.