Noise-induced hearing loss includes both temporary (TTS) and permanent (PTS) threshold shifts. Although TTS and PTS have many similarities, their underlying mechanisms are different. Both TTS and PTS are seen in hearing-conservation programs, making it important to consider both when making physiological measurements of inner-ear damage in applied settings. There are many ways that physiological mechanisms could be useful in screening for NIHL. Can normal-hearing and NIHL ears be differentiated from one another? Can the physiological measure be used in place of behavioural hearing-threshold measures of TTS and PTS? Can it be used to indicate sub-clinical damage (i.e., noise-induced permanent alterations to the inner ear without a corresponding hearing decrement)? Can it be used to indicate pre­clinical hearing loss (i.e., the sub-clinical damage eventually turns into hearing loss)? Finally, can the physiological measure be used to predict susceptibility to NIHL? Evoked otoacoustic emissions (EOAEs) depend on normal outer hair cells for their generation. Because this is the site in the inner ear in humans that is most susceptible to noise, there has been considerable interest in the application of EOAEs to NIHL screening. In this review, the application of distortion-product EOAEs (DPOAEs) is considered for this purpose, emphasizing work from our laboratory, but including that of others as well. Wherever possible, we compare the performance of DPOAEs as a screening tool to transient-evoked otoacoustic emissions (TEOAEs). We emphasize the importance of how well DPOAEs perform in screening for NIHL in individuals rather than for groups of people; the importance of using large numbers of subjects; and the importance of longitudinal studies.

Keywords: Distortion-product otoacoustic emission (DPOAE), transient-evoked otoacoustic emission (TEOAE), noise-induced hearing loss (NIHL).

Introduction

2 ) and levels L ₁ and L ₂ dB SPL (L ₁ > L ₂ ). These tones produce a family of DPOAEs in the cochlea; the strongest emission has a frequency of 2f ₁ -f ₂ . The two primary tones are usually chosen with a frequency ratio of f ₂ /f ₁ = 1.2, because this combination produces a relatively high amplitude 2f ₁ -f ₂ DPOAE over a range of primary frequencies and levels. All of the DPOAE studies in this review used a 1.2 frequency ratio of the primary tones. DPOAEs appear to be fairly frequency specific except, perhaps, at high primary stimulus levels (Avan and Bonfils, 1993). For instance, Avan and Bonfils (1993) found frequency specificity for L ₁ =L ₂ =62 dB SPL, but not for L ₁ =L ₂ =72 dB SPL. The source of the DPOAE is primarily at the f ₂ place for moderate level primaries (Brown and Kemp, 1984, Kummer et al., 1995, Brown et al., 1996, Heitmann et al., 1998). There is also a contribution from the 2f ₁ -f ₂ (more apical) frequency place (Kummer et al., 1995, Heitmann et al., 1998, Talmadge et al., 1999, Knight and Kemp, 2000), with the relative magnitude of the 2f ₁ -f ₂ contribution varying across individuals (Dreisbach, 1999, Siegel et al., 2000). DPOAE amplitude does not appear to depend on the basal cochlear state (Avan and Bonfils, 1993).

Although the focus of this review is on DPOAEs, some comparisons to transient-evoked otoacoustic emissions (TEOAEs) are included, particularly if both DPOAE and TEOAEs have been measured in the same subjects. Even though the amplitude of DPOAEs and TEOAEs are correlated (Smurzynski and Kim, 1992, Gorga et al., 1993b, Moulin et al., 1993, Probst and Harris, 1993), their underlying mechanisms may be different (Knight and Kemp, 1999, Shera and Guinan, 1999), and therefore it is interesting to compare their performance in NIHL screening. Like DPOAEs, TEOAEs have good frequency specificity, with the frequency of the TEOAE response matching the frequency place in the cochlea in normal-hearing (Probst et al., 1986, Norton and Neely, 1987) and hearing­impaired ears (Gorga et al., 1993b; Norton, 1993). Both DPOAEs and TEOAEs may be affected by basal as well as on-frequency damage, especially at low stimulus levels (Avan et al., 1993; Avan et al., 1995; Avan et al., 1996; Avan et al., 1997; Dorn et al., 1999). A further complication is that any comparisons between TEOAEs and DPOAEs in a particular experiment may be stimulus dependent and may not generalise to other sets of parameters.

DPOAEs can potentially be used to screen for many NIHL phenomena see [Table - 1]. Detecting existing hearing loss, establishing noise-induced temporary threshold shifts, and establishing noise-induced permanent threshold shifts are also possible audiometrically (and, in fact, whether rightly or wrongly, the audiogram usually is the "gold standard" to which DPOAE performance is compared). Detecting sub­clinical changes, pre-clinical hearing loss, and susceptibility to NIHL cannot, by definition, be assessed audiometrically.

To detect existing NIHL, an individual is categorized as having hearing levels above or below a criterion hearing level based on their DPOAE amplitudes. Next, to detect noise­-induced temporary and permanent changes to the inner ear, DPOAEs are measured before (baseline) and after noise exposure. Changes in DPOAE amplitude mirror changes in hearing thresholds, which are measured audiometrically as TTS and PTS. Detecting sub-clinical inner-ear changes (perhaps sub-clinical damage) and pre­clinical hearing loss with DPOAEs requires measurements more sensitive than the audiogram. We pragmatically define a change in DPOAE amplitude as sub-clinical if the change is greater than that found in a comparable control group and there is no significant threshold shift in the standard audiogram. Pre-clinical loss implies that the sub-clinical change measurable with DPOAEs eventually turns into hearing loss. Lastly, individual differences in NIHL susceptibility may be associated with biological phenomena, which are reflected in DPOAEs. DPOAEs are being investigated for use in all these areas.

Detection of Existing NIHL with DPOAEs

EOAEs are present in essentially all healthy, non-noise-exposed ears (e.g., Kapadia and Lutman, 1997). One of the earliest, most widespread, and most effective clinical uses of EOAEs has been as a screening tool to differentiate normal-hearing from hearing­impaired ears in populations for which audiometric testing is difficult due to age, cognitive ability, or illness. The requisite underlying science was to determine (a) how well normal-hearing and hearing-impaired ears could be differentiated based on EOAE characteristics (particularly response amplitude or a variation of it such as signal-to-noise ratio), and (b) which EOAE stimuli were best at making this determination. Because the science requires both pure-tone threshold and EOAE data, the subject population was not the difficult-to-test one, but rather people who could easily perform audiometric tests.

Gorga and colleagues at the Boys Town National Research Hospital (Omaha, Nebraska, USA) have conducted a series of studies using large numbers of subjects (from pre-school to geriatric in age, with inclusion of all cochlear etiologies) and a rigorous statistical approach (Gorga et al., 1993a; Gorga et al., 1993b; Prieve et al., 1993; Gorga et al., 1994; Gorga et al., 1996; Prieve et al., 1996; Stover et al., 1996; Gorga et al., 1997; Hussain et al., 1998; Gorga et al., 1999). They found that (a) DPOAEs (L ₁ / L ₂ = 65/50 dB SPL with three points per octave) were more accurate than TEOAEs (70 dB pSPL) in separating normal-hearing from hearing-impaired ears (defined as greater than 20 dB HL at 4 kHz), (b) DPOAEs were equivalent to TEOAEs at 2 and 4 kHz, and (c) DPOAEs were inferior to TEOAEs at 1 kHz (Gorga et al., 1993b). In subsequent work with DPOAEs (where L ₁ - L ₂ = 10 dB), they found that optimal L ₂ levels for separating normal-hearing from hearing-impaired ears were in the 50-60 dB SPL range (Stover et al., 1996).

Due to the large individual variations in DPOAE amplitude at each hearing level, there can be substantial categorization error. Gorga et al. (1996) used DPOAE amplitude (L ₁ / L ₂ = 65/55 dB SPL) and DPOAE threshold (i.e., the lowest recordable DPOAE level above the noise floor) to assign ears to one of three categories - normal, impaired, or uncertain. The subjects in their study had hearing levels ranging from 0 to 80 dB HL. The distributions of DPOAE amplitudes (and thresholds) for the normal-hearing and hearing-impaired listeners were used to empirically determine what ranges of DPOAE amplitudes were most likely associated with normal hearing, with hearing impairment, and with normal hearing status. For the category of DPOAE amplitude associated with normal hearing in the 1.0 to 4.0 kHz range, the hearing thresholds ranged from 0 to 40 dB HL; for the category of uncertain hearing status, the hearing thresholds ranged from 0 to 55 dB HL; for the hearing-impaired category, the hearing thresholds ranged from 10 to 80 dB HL (see [Figure 6] in Gorga et al., 1996).

The subjects in the Boys Town studies had a variety of cochlear etiologies. Would focusing entirely on a single etiology, with a more restricted age range, improve the ability to discriminate normal-hearing and hearing­impaired groups? Many have addressed this question, mostly with TEOAEs, and have found that emission amplitude, in general, is not likely to replace audiometric testing. Neither DPOAEs nor TEOAEs were sufficiently good at differentiating normal-hearing people from people with mild to moderate NIHL.

Attias et al. (1998) used DPOAEs (L ₁ =L ₂ =70 dB SPL, and f ₂ at audiometric frequencies) in military personnel to differentiate normal­hearing ( < 20 dB HL) from NIHL ears (ranging in severity from mild to severe). The sensitivity and specificity of the predictions were not strong. Avan et al. (1996) used equilevel primaries at low levels (from 50 to 65 dB SPL; using 1 kHz frequency intervals) in 42 ears from 30 subjects with mild NIHL ( < 40 dB HL), but neither DPOAE amplitude or input-output (I/O) function (DPOAE amplitude as a function of evoking-stimulus level) slope differentiated between the normal-hearing and NIHL ears (>20 dB HL).

Heller (unpublished data) investigated optimal DPOAE parameters to detect typical NIHL audiograms. The purpose was not to differentiate NIHL from normal hearing per se, but rather to choose the most sensitive stimulus levels for monitoring hearing status. For this reason, the normal-hearing group had audiometric thresholds < 5dB HL at 0.5-6.0 kHz, and < 15 dB HL at 8 kHz). The NIHL ears had notches in their audiograms at 3 or 4 kHz that were at least 15 dB worse than their 2 and 8 kHz thresholds. The thresholds at the frequency of the notch ranged from 15-65 dB HL, with 25% of the notches being in the 15-20 dB HL range. L ₁ values of 65, 60, 55, and 50 dB SPL were combined with L ₁ -L ₂ values of 0, 5, 10, 15, 20, and 25 dB, except for the 50/25 dB combination. For each stimulus level, f ₂ ranged from 977 to 4285 Hz in 1/8-octave steps. All attempts to identify a notch and to separate the groups with DPOAEs were not very successful, although many strategies were employed. These strategies included absolute DPOAE amplitude, relative amplitude between low and high frequencies, I/O function slope, area under the I/O function and above the noise floor, and DPOAE threshold.

Compared to DPOAEs, researchers initially had more success using TEOAEs to detect NIHL. Reshef et al. (1993) reported that the spectral width of the TEOAE response was a good predictor of NIHL. Attias et al. (1995) supported these results and reported that if TEOAEs were present, then the hearing level was always less than 20 dB HL, and at frequencies where hearing loss was greater than 20 dB HL, TEOAEs could not be recorded. Prasher and Sulkowski (1999), Lucertini et al. (1996), and Xu et al. (1998) all used TEOAE amplitude to differentiate NIHL from normal-hearing ears and from other etiologies, with various degrees of success. Sliwinska-Kowalska (1998) used both TEOAEs and DPOAEs (70 dB SPL equilevel primaries and I/O functions from 25-70 dB SPL). She described typical TEOAE and DPOAE findings for various severities of NIHL and recommended a qualitative use of EOAEs to substantiate audiological test results and to aid in differential diagnoses among NIHL, pseudohypoacusis, and other cochlear and retrocochlear etiologies.

From these studies, it generally appears that TEOAEs detect NIHL better than DPOAEs. One explanation is that different, level-dependent, mechanisms, which underlie TEOAE and DPOAE production (Shera and Guinan, 1999) are affected differently by noise-induced damage to the inner ear. Depending on which stimulus parameters are chosen, a difference between TEOAE and DPOAE performance may or may not be found. Tinnitus could also confound DPOAE results. If tinnitus is present, DPOAE amplitude in the region of the hearing loss may increase even though hearing sensitivity has decreased (Janssen et al., 1998). Additionally, if DPOAEs are sampled at large frequency intervals corresponding to audiometric frequencies, any given sample could be at the maxima of the DPOAE microstructure, at the minima, or anywhere between. Finally, because there is large individual variation in emission amplitude, DPOAEs and TEOAEs need to be compared in the same NIHL ears before reaching a conclusion about the superiority of one over the other. Indeed, Attias (2000) recently reported that TEOAEs and DPOAEs performed similarly in detecting NIHL.

None of these studies report any better performance for detecting NIHL with EOAEs than found in the Boys Town series of studies (which included all cochlear etiologies). Therefore, the usefulness of screening for NIHL with EOAEs is not apparent for people who are willing and able to complete an audiogram. On the other hand, for people who are motivated to exaggerate a hearing disability or who are unable to do an audiogram, the use of EOAEs for NIHL screening may be a useful tool in the audiologist's arsenal.

In addition, the norms obtained with these types of studies are useful for audiological applications, such as for differential diagnosis as suggested by Sliwinska-Kowalska (1998). Useful data for this purpose would be a compilation of normative data focusing specifically on NIHL as an etiology. These data would be based on a large number of noise­exposed subjects, optimal parameters of DPOAEs and TEOAEs (to differentiate various degrees of hearing loss), and data analyses involving detection statistics. From these data, each individual's EOAE profile (emission amplitude as a function of frequency, for a set of EOAE parameters) could be associated with the most likely audiogram as well as range of possible audiograms. This information would be helpful in assessing the validity of audiograms for individuals with a NIHL etiology.

Detection of TTS with DPOAEs

Another way that DPOAEs might be used clinically is for monitoring noise-induced inner­ear changes. Like TTS and PTS, DPOAEs can show temporary and permanent emission shifts (TES and PES, respectively). TES and PES do not depend on the absolute DPOAE level (as long as it is above the noise floor) because each person can serve as their own control. A baseline DPOAE is measured before the noise exposure, and, after the noise exposure, changes in the DPOAEs are measured. If test-retest reliability is good, which it is (Franklin et al., 1992), and the DPOAEs are sensitive to damage due to traumatic noise exposure, then DPOAEs could be substituted for behavioural audiometric testing.

The relationship between TTS and DPOAE TES has been measured both in laboratory settings, where experimental control of the exposure is possible, and in field settings, which have the advantage of higher noise levels than allowable in a laboratory. In this section we review DPOAEs following noise exposures in laboratory settings, compare them to TEOAEs measured in laboratory settings, and then make the same comparison for field settings.

In the first studies (Sutton et al., 1994, Engdahl and Kemp, 1996), DPOAEs were measured before and after a noise exposure and qualitatively appeared to be similar to hearing. Specifically, Sutton et al. (1994) examined the effect of stimulus level (L ₁ /L ₂ = 60/60, 55/55, 60/35, and 55/30 dB SPL) for DPOAEs following a three-minute noise exposure to a 105 dB SPL tone at 2.8 kHz in one ear from each of 14 subjects. They demonstrated that lower level primary tones, as well as lower f ₂ amplitude relative to f ₁ , increased the sensitivity of detecting post-exposure changes. Four subjects received the same noise exposure a second time with hearing measured instead of DPOAEs, and their TTS recovery functions qualitatively resembled their TES recovery function. Engdahl and Kemp (1996) measured DPOAEs, but no audiometric thresholds, following a ten-minute exposure to a 102 dB SPL, 1/3-octave narrow­band noise centred at 2 kHz. The DPOAE parameters were (a) L ₁ /L ₂ = 60/50 dB SPL with 3 points per octave in 3 subjects, (b) the same levels, but a 28 Hz resolution within a narrow frequency range for 2 subjects, and (c) I/O functions with L ₂ ranging from 50 to 75 dB SPL in 2 subjects. They also noted that the DPOAE recovery functions appeared to be similar to TTS recovery functions in other studies. In agreement with Sutton et al. (1994), they found that the DPOAE amplitude reduction was greatest at low stimulus levels and also noted that the DPOAE amplitude reduction was greatest at the maxima of the DPOAE microstructure (i.e., the sampling of the DPOAE frequency space may influence the size of the shift). Delb et al. (1999) also demonstrated greater test sensitivity with lower DPOAE levels (L ₁ /L ₂ = 60/50 and 60/35 dB SPL) for a one-hour broadband exposure to noise similar to that experienced in a discotheque at 98 dB(A).

Engdahl (1996) and Oeken and Menz (1996) measured hearing threshold and DPOAE amplitude from the same noise exposure, and found a low, or no, correlation between TTS and DPOAE TES. Both compared a single measure of TTS followed by DPOAE measurements shortly after a noise exposure. Engdahl tested eight participants with DPOAE parameters L ₁ /L ₂ = 55/40 dB SPL, sampled at 17 points/octave. The TTS-evoking noise was a 102 dB SPL, 1/3­octave narrow-band noise centred at 2 kHz and presented for 10 minutes. Oeken and Menz tested 102 ears of 59 people with DPOAE parameters L ₁ /L ₂ = 70/65 dB SPL, sampled at 4 points/octave. The TTS-evoking noise was a 90 dB HL white noise, presented for 20 minutes. The low correlations may have been due to the methodology used. In both studies, a single measurement of audiometric threshold was made, then the earphone with the probe used for DPOAE measurements was inserted, and the DPOAE was measured. Oeken and Menz began the audiometric measurement immediately after the noise exposure, but by the time the DPOAE measurement was made the hearing system could have been at a very different stage in recovery (i.e., at a different place along the recovery function). In addition, their DPOAE stimulus levels were quite high. Engdahl used a more elegant approach. He began the first audiometric measurement at 30-60 seconds after the noise exposure and the first DPOAE measurement at two minutes post-exposure. He then attempted to correct for the difference in time of the measurements by fitting an exponential curve to additional DPOAE measurements over a thirty-minute recovery period and extrapolating the DPOAE measurement back to the same time as the hearing threshold measurement. Eight subjects, however, is a small number to study this problem, due to individual differences.

Marshall et al. (1998) found a much stronger relationship between TTS and DPOAE amplitude shifts following a noise exposure (105 dB SPL half-octave narrow-band noise at 1.414 kHz centre frequency) for 14 people, with measurements of TTS and DPOAEs interleaved along the same recovery function. Hearing thresholds were measured with a Bekesy psychophysical procedure at 2 kHz, which was half an octave above the frequency of the noise exposure, and so was where the maximum TTS was expected to occur. DPOAE stimulus levels were L ₁ /L ₂ = 70/60 dB SPL and L ₁ /L ₂ = 65/45 dB SPL. Testing started one minute after the noise exposure ended. All sounds (the pure tones for audiometric thresholds and for DPOAEs, and the narrow-band noise that produced the TTS) were delivered through the same earphone, and were carefully calibrated in each person's ear canal. In a companion study, the same paradigm was used with a different stimulus - TEOAEs at 74 dB pSPL. Eight subjects overlapped both studies, which allowed a better comparison of DPOAEs with TEOAEs.

[Figure - 1] depicts the average amount of shift as a function of recovery time for 14 subjects. The pre-noise-exposure tests are shown at 0 dB HL because the pretest amplitudes were normalized to better show post-noise-exposure shifts. There were two sets of pre-exposure measurements - each set contained a Bekesy threshold measurement and the two DPOAE stimulus levels. The noise-exposure lasted for ten minutes. Beginning at one minute post-exposure, sets of hearing and DPOAE measurements were alternated continuously until 18 minutes post­exposure, and then snapshots of the recovery were taken throughout the remainder of the test session (beginning at 23, 31, 47, 55, and 63 minutes post-exposure). The mean TTS for the group at approximately two minutes post­exposure was 11.6 dB, and the mean DPOAE amplitude decrements were 5.0 dB for the lower level stimulus (L ₁ /L ₂ = 65/45 dB SPL) and 3.2 dB SPL for the higher level stimulus (L ₁ /L ₂ = 70/60 dB SPL). The error bars on the plot are the standard error of the mean (SE _mean ), which were not plotted if smaller than the symbol. The solid, dashed, and dotted lines are linear least-squares fits to the post-exposure data as a function of logarithmic time. Hearing thresholds and DPOAE recovery had the same recovery form (well fitted by a linear shift as a function of logarithmic time), and there was an orderly relationship between TTS and TES along these functions.

In [Figure - 2], the data from [Figure - 1] are re-plotted to show the mean post-exposure Bekesy threshold shift as a function of the average post­exposure DPOAE shift (averaged over a 1/3­octave band centred at 2.002 kHz) at 65/45 and 70/60 dB SPL for each of the points along the recovery function. The largest shifts were measured soon after the noise exposure, and the smaller shifts occurred at longer post-exposure times. The solid and dashed lines are linear least­square fits for the 65/45 and 70/60 dB SPL stimuli, respectively. The correlations were high (-0.97 and -0.99 for the two stimulus levels), and the standard error of estimate (SE _est ) was low (approximately 0.5 dB). Thus, for the group, if the magnitude of the emission shift is known, the magnitude of the hearing-threshold shift can be accurately predicted. Individual recovery functions were similar to the group average function although somewhat more variable (average SE _est = 2.0 dB). A similar result was found for the relationship between Bekesy thresholds and TEOAEs (Marshall and Heller, 1998). It is apparent that the underlying mechanism for TES, for both DPOAEs and TEOAEs, is closely linked to the underlying mechanism for TTS.

Engdahl's (1996) and Oeken and Menz' (1996) correlations were for a single post-exposure measurement. This is more applicable to hearing-conservation programs, where hearing thresholds (or emissions) are measured at a single point in time. Marshall et al. (1998) found a significant correlation (r = 0.58) between TTS and TES for a single point on the recovery function (approximately 2 min post-exposure) for the higher level (L ₁ /L ₂ = 70/60 dB SPL) DPOAE stimulus. This correlation was similar to those found for TEOAEs (r = 0.55 for Attias and Bresloff, 1996, and r = 0.53 for Marshall and Heller, 1998). Inexplicably, the correlation between TTS and TES for the lower level DPOAE (L ₁ /L ₂ = 65/45 dB SPL) was only 0.34.

To further compare DPOAEs and TEOAEs, the spectrum of the average DPOAE shift and of the average TEOAE shift across the eight subjects who were in both the DPOAE (Marshall et al., 1998) and TEOAE (Marshall and Heller, 1998) studies are plotted in [Figure - 3]. The maximum shift for both DPOAEs and TEOAEs was half an octave above the centre frequency of the exposure noise. The spectra for individual subjects were not as orderly, either in relationship of DPOAEs to TEOAEs or in the frequency of maximal shift. For all subjects, the frequency of maximum shift was 0.5-1 octave above the exposure noise, and sometimes was quite broad, spanning as many as three ½-octave bands (in frequency regions including on­frequency to one octave above).

Unlike the results in our laboratory, a study by Vinck et al. (1999) showed different recovery functions for DPOAEs (L ₁ /L ₂ = 65/55 dB SPL with 8 points/octave) and TEOAEs (70 dB pSPL) for a group of ten subjects where the noise exposure was 90 dB SPL broadband noise for one hour. Significant emission amplitude decrements were seen for both TEOAEs and DPOAEs in the same frequency region. The DPOAE recovery function was similar to that seen for other studies. The TEOAE recovery function showed an oscillating recovery pattern at 4 kHz.

There is less control over noise exposures outside the laboratory, but they allow researchers to (a) study more severe exposures than those ethically used in laboratories, (b) observe the development of PTS in individuals, and (c) observe the development of sub-clinical changes and pre-clinical hearing loss in individuals.

Two of the field studies involved discotheque noise exposures. Liebel et al. (1996) measured DPOAEs (L ₁ = L ₂ = 70 dB SPL with 4 points/octave), and TEOAEs (at an unspecified stimulus level) in 92 ears of 46 subjects following a two-hour, 105 dB(A) exposure. The average TTS was broadband, with a maximum TTS >10 dB at 3 and 4 kHz. For the subjects with TTS > 15 dB, the DPOAE amplitude decreased by 8 dB for the ten subjects with TTS at 4 kHz, but by only 1 dB for the nine subjects with TTS at 3 kHz. For subjects with less TTS, DPOAE shifts were not significant. Given the relatively high DPOAE stimulus levels (especially for L ₂ ), the sensitivity of DPOAEs in detecting the TTS was good. Not surprisingly, given their stimulus parameters, TEOAEs were more sensitive indicators of TTS than were DPOAEs. Vinck et al. (1999) measured DPOAEs (L ₁ /L ₂ = 65/55 dB SPL with eight points per octave) and TEOAEs (70 dB pSPL) in one ear from each of eight subjects after five consecutive hours of discotheque exposure (L _eq(5hrs) = 103.5 dB SPL). They found that both DPOAEs and TEOAEs were sensitive to TTS, and, in addition, the recovery patterns of both emission types were similar.

Marshall et al. (2000) conducted a longitudinal study in which subjects were tested annually, up to 4 years. Most of the noise-exposed subjects were military personnel already enrolled in the hearing-conservation program of the U. S. Navy. There was also a quiet control group, which was used to establish the standard error of measurement (SE _meas ) for each EOAE stimulus configuration. A significant emission shift for the noise-exposed group was defined as greater than two SE _meas for the quiet control group. The EOAE test battery contained both DPOAEs and TEOAEs over a range of stimulus levels. Although the primary purpose of the study was to observe the development of PTS, individual cases of TTS also were observed. Both DPOAEs and TEOAEs generally showed significant amplitude decrements in the same frequency region as the TTS.

One of the TTS subjects in this study is shown in [Figure - 4]. [Figure - 4]a shows the audiogram for the right ear of this adult female. In the third year she experienced a 15 dB threshold shift at 4 kHz, relative to her baseline in the first year, from exposure to wood-chipper and chain-saw noise. In the fourth year, the threshold shift had recovered. [Figure - 4]b shows the DPOAEs for the right ear. The DPOAEs paralleled the audiogram. In the third year, the DPOAE amplitude decreased (>2SE _meas ) at 4 kHz, and then returned to baseline in the fourth year. The TEOAEs are shown in [Figure - 4]c. In the third year, there is a decrement not only at 4 kHz, but at lower frequencies too. Note that the TEOAE amplitude did not recover in the fourth year, although the hearing and DPOAEs did recover. We defined this as a sub-clinical change. Because audiometric frequencies above 8 kHz were not measured in this study, there is no way to know whether the TEOAE changes were reflective of very high-frequency threshold shifts (Avan, 1997) or of on-frequency sub-clinical damage.

In summary, DPOAEs are good indicators of TTS, especially if L ₂ is low enough in level. The recovery functions for TTS and DPOAE TES are similar, although the absolute magnitude of TTS from a particular noise exposure is larger than the TES to that same exposure. For individual subjects, the magnitudes of TTS and DPOAE TES usually are related, but the relationship is not perfect. Part of this discrepancy is due to using a behavioural hearing test as the "gold standard," and the other is that DPOAEs and hearing are related, but not identical, processes. Depending on the stimulus configurations, DPOAEs are as sensitive as TEOAEs for measuring TES.

Detection of PTS with DPOAEs

Like TTS, the development of PTS may be studied using DPOAEs. Unlike TTS, however, PTS in humans cannot be induced in the laboratory, so can only be studied in field settings. To develop good screening norms, subjects need to be followed over time to watch the development of PTS, rather than comparing groups of hearing-impaired and normal-hearing subjects. There are few such studies, because of the difficulty in finding suitable populations and because of the time and expense involved in longitudinal designs.

The NSMRL longitudinal study allowed us to observe the development of PTS in individuals. One case of PTS in an adult male is depicted in [Figure - 5]. Despite being occupationally exposed to high levels of noise, the cause of this bilateral loss was most likely from a very high-powered stereo sound system in the subject's car. Only the left ear is shown here, but the right ear showed a very similar loss. His left ear showed a 10 dB broadband significant threshold shift at 2, 3, and 4 kHz in the third year see [Figure - 5]a. Repeat audiograms confirmed the PTS. DPOAE amplitudes see [Figure - 5]b showed a broadband decrease in the third year, but only the shift at 2 kHz was significant (this criterion was based on a single-frequency analysis, so is probably rather strict). [Figure - 5]c shows a broadband TEOAE amplitude shift in the third year, which was significant across all frequencies measured. As with the previous case, the TEOAE amplitude changes at low frequencies are consistent with our definition of sub-clinical damage.

In our longitudinal study, changes in TEOAE amplitude were more likely to be consistent with audiometric PTS than were changes in DPOAE amplitude. Sliwinska-Kowalska (2000) showed results of a longitudinal study measuring hearing, DPOAEs, and TEOAEs in industrial workers. Hearing and DPOAEs did not show decrements over time, but TEOAEs did.

It is premature to conclude that TEOAEs are superior to DPOAEs in detecting PTS and sub­clinical damage. Apparent insensitivity of the DPOAEs could have been due to the parameters of the DPOAE stimuli that were used (an unfortunate fact of life with longitudinal studies is that the science develops while the test paradigm must remain fixed). At least four changes to typical DPOAE test paradigms could result in more sensitive DPOAE results:

(1)In our longitudinal study, we kept the level difference between the two primary tones constant at deltas of 10 and 20 dB as we changed overall level to obtain I/O functions. Kummer et al. (1998) have found that increasing the level difference between the two primary levels as the overall stimulus level is decreased improved the ability of DPOAEs to detect hearing loss. We are currently collecting PTS data at two sites - aircraft carrier personnel before and after a six­month cruise in the Mediterranean, and Marine recruits before and after boot camp - the latter study in collaboration with CDR Keith Wolgemuth and COL Rick Kopke at the Naval Medical Centre San Diego, Dr. Tom Taggart at SUNY-Buffalo, and Dr. Shelley Smith at the University of Nebraska Medical Centre. In those studies, we are using the DPOAE approach suggested by Kummer et al. (1998).

(2) By increasing the sampling density of frequencies throughout the frequency range, the problem of hitting maxima and minima in the DPOAE microstructure is reduced. In our longitudinal study we used only 4 frequencies within each half-octave band. Engdahl and Kemp (1996) used a 28-Hz resolution and showed that it is beneficial to examine the DPOAE microstructure when interpreting results. Unfortunately, this approach takes too much time to be applied to screening applications.

(3) It is important to consider the most appropriate frequency spacing of the two primary tones, which may not be f ₂ /f ₁ = 1.2. Knight and Kemp (2000) have explained why different frequency ratios could be more or less sensitive in detecting emission amplitude shifts. Their work also illustrates that the issue is not to choose between DPOAEs and TEOAEs, but rather to decide what underlying physiological mechanisms are being targeted and to attempt to choose the stimulus parameters accordingly.

(4) Using lower L ₁ /L ₂ stimulus levels (Shera and Guinan, 1999) may increase the contribution of the EOAE mechanism considered to be more sensitive to inner ear changes. In our longitudinal study, we found that although there was a reduction in the number of measurable emissions, we could still capture at least 89% of DPOAEs at 4 kHz (and at least 73% at 2 kHz), even at levels of 55/35 dBSPL. The latest equipment on the market has a much lower noise floor, so it will be possible to measure even more DPOAEs in the future.

Detecting Susceptibility to NIHL with DPOAEs

It would be advantageous to be able to determine a priori who is most at risk for noise-induced TTS and PTS. There are at least five ways that DPOAEs might be used to evaluate susceptibility. All need further investigation.

(1) Is low (or high) DPOAE amplitude indicative of increased (or decreased) NIHL risk? In our TTS studies, high pre-noise­exposure DPOAE, but not TEOAE, amplitude was associated with a decreased risk of TTS. Engdahl (1996), however, did not find this relationship. Perhaps the reason for the difference is that he had fewer subjects. Actually, neither study had enough subjects to make conclusions of this type with any degree of confidence. In our longitudinal study, DPOAE amplitude in the first year was unrelated to PTS risk, but low TEOAE amplitude in the first year was associated with increased risk for developing PTS.

(2) Can a clinical exposure to low-level noise (presented at levels known to not produce TTS in most people) be used to identify those with high susceptibility to TTS? Plinkert et al. (1999) found indications that there were more emission shifts in a group of soldiers (who had previously got TTS from an impulse noise exposure) when challenged with a relatively low noise exposure in the laboratory (which did not give TTS in a group of normal-hearing controls). Their results, however, were preliminary; this line of enquiry should be pursued with a larger group and with the aim of finding the best type and level of noise exposure.

(3) Are sub-clinical DPOAE changes indicative of increased risk for PTS? Sub-clinical signs of noise damage may include both EOAE amplitude decrements and diminished contralateral EOAE suppression (Desai et al., 1999). There are no published studies that have determined whether the development of sub­clinical changes (DPOAE or TEOAE) is associated with increased risk for developing NIHL, but our expectation is that there will be an association.

(4) Is there a characteristic DPOAE amplitude pattern associated with genetic risk for NIHL? Morell et al. (1998) found low DPOAE amplitudes and spectral notching for normal­hearing carriers of the Connexin 26 (GJB2) gene mutation. Liu and Newton (1997) found similar results for Waardenburg's Syndrome. It is possible that gene mutations will cause abnormal DPOAEs and that these abnormal DPOAEs will be diagnostic of predisposition to PTS in otherwise normal-hearing people. This is one of the issues being studied in our Marine Recruit study.

(5) Does efferent strength as measured with DPOAEs predict who is most at risk for developing NIHL? For TTS, Engdahl (1996) found a positive correlation between TTS and the contralateral DPOAE amplitude suppression effect (greater suppression was associated with greater TTS). For PTS, there are no data in humans. The available animal data (Maison and Liberman, 2000) shows that DPOAE adaptation as a measure of efferent strength in guinea pigs was predictive of susceptibility to PTS. In particular, animals with greater efferent strength had less PTS.

Summary of DPOAEs to screen for NIHL

Much more research is needed to determine how well DPOAEs can screen for NIHL. Especially, we need to learn more about how otoacoustic emissions change with the development of PTS in humans. This type of data is difficult and expensive to obtain, but several groups worldwide are now acquiring such data sets, and they will undoubtedly contribute greatly to our knowledge. To successfully use otoacoustic emissions in clinical settings requires researchers to work closely with theorists, and vice-versa. Unfortunately, until the underlying mechanisms for EOAEs are better understood, the only way to determine which protocols are best in detecting NIHL is by empirical test. Likewise, the development of models to explain otoacoustic emissions will be helped by many examples of TTS and PTS development.

Acknowledgements

The NSMRL research in this paper was funded by the Office of Naval Research. The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Navy, Department of Defence, or the United States Government.[61]

 

References

1.	Attias, J. (2000) Detection and diagnosis of NIHL by OAE. In Noise Pollution Health Effects Reduction (NOPHER) 2000 International Symposium on Noise­ Induced Hearing Loss, Cambridge, England.
2.	Attias, J. and Bresloff, I. (1996) Noise induced temporary otoacoustic emission shifts. J Basic Clin Physiol Pharmacol, 7, 221-33.
3.	Attias, J., Bresloff, I., Reshef, I., Horowitz, G. and Furman, V. (1998) Evaluating noise induced hearing loss with distortion product otoacoustic emissions. Br J Audiol, 32, 39-46.
4.	Attias, J., Furst, M., Furman, V., Reshef, I., Horowitz, G. and Bresloff, I. (1995) Noise-induced otoacoustic emission loss with or without hearing loss. Ear Hear, 16, 612-8.
5.	Avan, P. and Bonfils, P. (1993) Frequency specificity of human distortion product otoacoustic emissions. Audiology, 32, 12-26.
6.	Avan, P., Bonfils, P. and Loth, D. (1996) In Scientific Basis of Noise-Induced Hearing Loss (Eds, Axelsson, A., Hellstrom, P. A., Borchgrevink, H., Henderson, D., Hamernik, R. P. and Salvi, R.) George Thieme Verlag, Stuttgart, Thieme, pp. 65-81.
7.	Avan, P., Bonfils, P., Loth, D., Elbez, M. and Erminy, M. (1995) Transient-evoked otoacoustic emissions and high­frequency acoustic trauma in the guinea pig. J Acoust Soc Am, 97, 3012-20.
8.	Avan, P., Bonfils, P., Loth, D. and Wit, H. P. (1993) Temporal patterns of transient-evoked otoacoustic emissions in normal and impaired cochleae. Hear Res, 70, 109-20.
9.	Avan, P., Elbez, M. and Bonfils, P. (1997) Click-evoked otoacoustic emissions and the influence of high-frequency hearing losses in humans. J Acoust Soc Am, 101, 2771-7.
10.	Brown, A. M., Harris, F. P. and Beveridge, H. A. (1996) Two sources of acoustic distortion products from the human cochlea. J Acoust Soc Am, 100, 3260-7.
11.	Brown, A. M. and Kemp, D. T. (1984) Suppressibility of the 2f1-f2 stimulated acoustic emissions in gerbil and man. Hear Res, 13, 29-37.
12.	Delb, W., Hoppe, U., Liebel, J. and Iro, H. (1999) Determination of acute noise effects using distortion product otoacoustic emissions. Scand Audiol, 28, 67-76.
13.	Desai, A., Reed, D., Cheyne, A., Richards, S. and Prasher, D. (1999) Absence of otoacoustic emissions in subjects with normal audiometric thresholds implies exposure to noise. Noise & Health, 2, 58-65.
14.	Dorn, P.A., Piskorski, P., Gorga, M.P., Neely, S.T. and Keefe, D.H. (1999). Predicting audiometric status from distortion product otoacoustic emissions using multivariate analyses. Ear Hear, 20(2), 149-163.
15.	Dreisbach, L. (1999) Characterizing the 2f1-f2 distortion­product otoacoustic emission and its generators measured from 2 to 20 kHz in humans. Unpublished Ph.D. dissertation, Northwestern University.
16.	Engdahl, B. (1996) Effects of noise and exercise on distortion product otoacoustic emissions. Hear Res, 93, 72-82.
17.	Engdahl, B. and Kemp, D. T. (1996) The effect of noise exposure on the details of distortion product otoacoustic emissions in humans. J Acoust Soc Am, 99, 1573-87.
18.	Franklin, D. J., McCoy, M. J., Martin, G. K. and Lonsbury­Martin, B. L. (1992) Test/retest reliability of distortion­product and transiently evoked otoacoustic emissions. Ear Hear, 13, 417-29.
19.	Gorga, M. P., Neely, S. T., Bergman, B., Beauchaine, K. L., Kaminski, J. R., Peters, J. and Jesteadt, W. (1993a) Otoacoustic emissions from normal-hearing and hearing­impaired subjects: distortion product responses. J Acoust Soc Am, 93, 2050-60.
20.	Gorga, M. P., Neely, S. T., Bergman, B. M., Beauchaine, K. L., Kaminski, J. R. and Liu, Z. (1994) Towards understanding the limits of distortion product otoacoustic emission measurements. J Acoust Soc Am, 96, 1494-500.
21.	Gorga, M. P., Neely, S. T., Bergman, B. M., Beauchaine, K. L., Kaminski, J. R., Peters, J., Schulte, L. and Jesteadt, W. (1993b) A comparison of transient-evoked and distortion product otoacoustic emissions in normal-hearing and hearing-impaired subjects. J Acoust Soc Am, 94, 2639­48.
22.	Gorga, M. P., Neely, S. T. and Dorn, P. A. (1999) Distortion product otoacoustic emission test performance for a priori criteria and for multifrequency audiometric standards. Ear Hear, 20, 345-62.
23.	Gorga, M. P., Neely, S. T., Ohlrich, B., Hoover, B., Redner, J. and Peters, J. (1997) From laboratory to clinic: a large scale study of distortion product otoacoustic emissions in ears with normal hearing and ears with hearing loss. Ear Hear, 18, 440-55.
24.	Gorga, M. P., Stover, L., Neely, S. T. and Montoya, D. (1996) The use of cumulative distributions to determine critical values and levels of confidence for clinical distortion product otoacoustic emission measurements. J Acoust Soc Am, 100, 968-77.
25.	Heitmann, J., Waldmann, B., Schnitzler, H., Plinkert, P. K. and Zenner, H. P. (1998) Suppression of distortion product otoacoustic emissions (DPOAE) near 2f1 - f2 removes DP-gram fine structure-Evidence for a secondary generator. J Acoust Soc Am, 103, 1527-1531.
26.	Hussain, D. M., Gorga, M. P., Neely, S. T., Keefe, D. H. and Peters, J. (1998) Transient evoked otoacoustic emissions in patients with normal hearing and in patients with hearing loss. Ear Hear, 19, 434-49.
27.	Janssen, T., Kummer, P. and Arnold, W. (1998) Growth behavior of the 2 f1-f2 distortion product otoacoustic emission in tinnitus. J Acoust Soc Am, 103, 3418-30.
28.	Kapadia, S. and Lutman, M. E. (1997) Are normal hearing thresholds a sufficient condition for click-evoked otoacoustic emissions? J Acoust Soc Am, 101, 3566-7.
29.	Knight, R. D. and Kemp, D. (1999) Relationships between DPOAE and TEOAE amplitude and phase characteristics. J Acoust Soc Am, 106, 1420-1435.
30.	Knight, R. D. and Kemp, D. T. (2000) Indications of different distortion product otoacoustic emission mechanisms from a detailed f1,f2 area study. J Acoust Soc Am, 107, 457-73.
31.	Kummer, P., Janssen, T. and Arnold, W. (1995) Suppression tuning characteristics of the 2 f1-f2 distortion­ product otoacoustic emission in humans. J Acoust Soc Am, 98, 197-210.
32.	Kummer, P., Janssen, T. and Arnold, W. (1998) The level and growth behavior of the 2 f1-f2 distortion product otoacoustic emission and its relationship to auditory sensitivity in normal hearing and cochlear hearing loss. J Acoust Soc Am, 103, 3431-44.
33.	Liebel, J., Delb, W., Andes, C. and Koch, A. (1996) Die Erfassung von Larmschaden bei Besuchern einer Diskothek mit Hilfe der TEOAE und DPOAE. Laryngorhinootologie, 75, 259-264.
34.	Liu, X. Z. and Newton, V. E. (1997) Distortion product emissions in normal-hearing and low-frequency hearing loss carriers of genes for Waardenburg's syndrome. Ann Otol Rhinol Laryngol, 106, 220-5.
35.	Lucertini, M., Bergamaschi, A. and Urbani, L. (1996) Transient evoked otoacoustic emissions in occupational medicine as an auditory screening test for employment. Br J Audiol, 30, 79-88.
36.	Maison, S. F. and Liberman, M. C. (2000) Predicting vulnerability to acoustic injury with a noninvasive assay of olivocochlear reflex strength. J Neurosci, 20, 4701-7.
37.	Marshall, L. and Heller, L. M. (1998) Transient-evoked otoacoustic emissions as a measure of noise-induced threshold shift. J Speech Lang Hear Res, 41, 1319-34.
38.	Marshall, L., Heller, L. M. and Lentz, B. (1998) Distortion-product emissions acoompanying TTS. Assoc. Res.Otolaryngol. Abs. (pp. 150).
39.	Marshall, L., Heller, L. M., Westhusin, L. J. and Lapsley Miller, J. A. (2000) TEOAE/DPOAE changes associated with developing NIHL. Assoc. Res.Otolaryngol. Abs. (pp. 66)
40.	Morell, R. J., Kim, H. J., Hood, L. J., Goforth, L., Friderici, K., Fisher, R., Van Camp, G., Berlin, C. I., Oddoux, C., Ostrer, H., Keats, B. and Friedman, T. B. (1998) Mutations in the connexin 26 gene (GJB2) among Ashkenazi Jews with nonsyndromic recessive deafness. N Engl J Med, 339, 1500-5.
41.	Moulin, A., Collet, L., Veuillet, E. and Morgon, A. (1993) Interrelations between transiently evoked otoacoustic emissions, spontaneous otoacoustic emissions and acoustic distortion products in normally hearing subjects. Hear Res, 65, 216-33.
42.	Norton, S. J. (1993) Application of transient evoked otoacoustic emissions to pediatric populations. Ear Hear, 14, 64-73.
43.	Norton, S. J. and Neely, S. T. (1987) Tone-burst-evoked otoacoustic emissions from normal-hearing subjects. J Acoust Soc Am, 81, 1860-72.
44.	Oeken, J. and Menz, D. (1996) Amplitude changes in distortion products of otoacoustic emissions after acute noise exposure. Laryngorhinootologie, 75, 265-9.
45.	Plinkert, P. K., Hemmert, W., Wagner, W., Just, K. and Zenner, H. P. (1999) Monitoring noise susceptibility: sensitivity of otoacoustic emissions and subjective audiometry. Br J Audiol, 33, 367-82.
46.	Prasher, D.and Sulkowski, W. (1999) The role of otoacoustic emissions in screeening and evaluation of noise damage. Int J Occup Med Environ Health, 12(2), 183-192.
47.	Prieve, B. A., Gorga, M. P. and Neely, S. T. (1996) Click­and tone-burst-evoked otoacoustic emissions in normal­hearing and hearing-impaired ears. J Acoust Soc Am, 99, 3077-86.
48.	Prieve, B. A., Gorga, M. P., Schmidt, A., Neely, S., Peters, J., Schultes, L. and Jesteadt, W. (1993) Analysis of transient-evoked otoacoustic emissions in normal-hearing and hearing-impaired ears. J Acoust Soc Am, 93, 3308-19.
49.	Probst, R., Coats, A. C., Martin, G. K. and Lonsbury­Martin, B. L. (1986) Spontaneous, click-, and toneburst­evoked otoacoustic emissions from normal ears. Hear Res, 21, 261-75.
50.	Probst, R. and Harris, F. P. (1993) Transiently evoked and distortion-product otoacoustic emissions. Comparison of results from normally hearing and hearing-impaired human ears. Arch Otolaryngol Head Neck Surg, 119, 858-­60.
51.	Reshef, I., Attias, J. and Furst, M. (1993) Characteristics of click-evoked otoacoustic emissions in ears with normal hearing and with noise-induced hearing loss. Br J Audiol, 27, 387-95.
52.	Shera, C. A. and Guinan, J. J., Jr. (1999) Evoked otoacoustic emissions arise by two fundamentally different mechanisms: a taxonomy for mammalian OAEs. J Acoust Soc Am, 105, 782-98.
53.	Siegel, J., Borneman, A. and Dreisbach, L. (2000) Suppressor conditions for optimal separation of distortion product otoacoustic emissions. Assoc. Res.Otolaryngol. Abs. (pp. 283).
54.	Sliwinska-Kowalska, M. (1998) The role of evoked and distortion-product otoacoustic emissions in diagnosis of occupational noise-induced hearing loss. Journal of Audiological Medicine, 7, 29-45.
55.	Sliwinska-Kowalska, M. (2000) Industrial hearing loss monitored by otoacoustic emissions. In Noise Pollution Health Effects Reduction (NOPHER) 2000 International Symposium on Noise-Induced Hearing Loss, Cambridge, England.
56.	Smurzynski, J. and Kim, D. O. (1992) Distortion-product and click-evoked otoacoustic emissions of normally­hearing adults. Hear Res, 58, 227-40.
57.	Stover, L., Gorga, M. P., Neely, S. T. and Montoya, D. (1996) Toward optimizing the clinical utility of distortion product otoacoustic emission measurements. J Acoust Soc Am, 100, 956-67.
58.	Sutton, L. A., Lonsbury-Martin, B. L., Martin, G. K. and Whitehead, M. L. (1994) Sensitivity of distortion-product otoacoustic emissions in humans to tonal over-exposure: time course of recovery and effects of lowering L2. Hear Res, 75, 161-74.
59.	Talmadge, C. L., Long, G. R., Tubis, A. and Dhar, S. (1999) Experimental confirmation of the two-source interference model for the fine structure of distortion product otoacoustic emissions. J Acoust Soc Am, 105, 275-­92.
60.	Vinck, B. M., Van Cauwenberge, P. B., Leroy, L. and Corthals, P. (1999) Sensitivity of transient evoked and distortion product otoacoustic emissions to the direct effects of noise on the human cochlea. Audiology, 38, 44­-52.
61.	Xu, Z. M., Van Cauwenberge, P., Vinck, B. and De Vel, E. (1998) Sensitive detection of noise-induced damage in human subjects using transiently evoked otoacoustic emissions. Acta Otorhinolaryngol Belg, 52, 19-24.

---

Correspondence Address:
Lynne Marshall
Naval Submarine Medical Research Laboratory, Groton, CT 06349
USA

Source of Support: None, Conflict of Interest: None

Check

13

PMID: 12678940

Figures

  [Figure - 1], [Figure - 2], [Figure - 3], [Figure - 4], [Figure - 5]
 
 

Tables

  [Table - 1]