Transient Evoked Otoacoustic Emissions and Pseudohypacusis

Transcription

1 A Acad Audiol 6 : (1995) Transient Evoked Otoacoustic Eissions and Pseudohypacusis Frank E. Musiek* Steven P. Bornsteint Willia F. Rintelann$ Abstract The audiologic diagnosis of pseudohypacusis continues to challenge the clinical audiologist. The introduction of otoacoustic eissions (OAEs) to the test repertoire of the audiologist ay prove valuable in the evaluation of pseudo hypacusis. This report highlights five cases in which transient evoked otoacoustic eissions (TEOAEs) were used to cross-check the validity of subjective audiologic thresholds. Relationships of TEOAEs to other objective easures of audiologic thresholds are shown to substantiate the value of the procedure for the diagnosis of pseudohypacusis. Suggestions for the use of OAEs in cases of pseudohypacusis are discussed. Key Words: Malingering, otoacoustic eissions, pseuclohypacusis uring the past 4 years, several procedures have been used to identify pseudohypacusis. The ost obvious routine D audioetric sign of pseudohypacusis is a disagreeent (usually 15 db or ore) between puretone averages and speech reception thresholds, with speech thresholds typically better. Although this does not give an accurate indication of thresholds, it alerts the audiologist to the possibility of pseudohypacusis. The ost widely used special test is the Stenger, either pure tone or speech ; however, this test requires at least a 2-dB difference between ears, does not give precise threshold easures, and has been shown to have an unacceptably high rate of inaccuracy. Other easures include the Bekesy audioetry, the sensorineural acuity level (SAL) test, the Doerfler-Stewart test, the acoustic reflex, the Lobard test, and pure-tone and speech-delayed auditory feedback (Rintelann and Schwan, 1991). 'Section of Otolaryngology and Audiology, Departent of Surgery, Section of Neurology, Departent of Medicine, Dartouth-Hitchcock Medical Center, Lebanon, New Hapshire ; tdepartent of Counication Disorders, School of Health and Huan Services, University of New Hapshire, Durha, New Hapshire ; and tdepartent of Audiology, Wayne State University School of Medicine, Detroit, Michigan Reprint requests : Dr. Frank Musiek, Section of Otolaryngology, Dartouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH Electrophysiologic easures exist that ay help to both identify and quantify pseudohypacusis (Hall, 1992). Traditionally, the ost accurate but infrequently used easure was the slow cortical potential (N,-P2). This test can provide frequency-specific inforation, and the response has been present close to behavioral threshold. The ajor liitation of the test is that it is tieconsuing : approxiately 4 inutes per frequency to adinister, in addition to the tie spent on instruentation setup and application of electrodes. The auditory brainste response (ABR) and the iddle latency response (MLR) tests using click stiuli are also recordable close to behavioral threshold. While the click-evoked ABR is liited to providing inforation about the range of hearing fro 1 to 4 Hz, the MLR can provide frequency-specific results. However, adinistration of the ABR and MLR, like that of the slow cortical potential, is tie-consuing. Another proising clinical tool is the recording of otoacoustic eissions (OAEs) (Robinette, 1992). Otoacoustic eissions are generally subaudible sounds generated by an active response fro the cochlea (ost likely the outer hair cells) to various types of sounds. Otoacoustic eissions provide inforation about a wide range of audioetric frequencies. Both ears can be tested in approxiately 3 inutes. Generally, when using the IL88 or IL92, click-evoked OAEs disappear when the hearing loss for the frequencies of interest reach approxiately 3 db HL, although

2 ournal of the Aerican Acadey of Audiology/Volue 6, Nuber 4, uly 1995 this ust be interpreted cautiously (Kep et al, 199). The priary caveat is that eissions ay give an unacceptably high false-positive identification rate of hearing loss. Therefore, presence of an eission is a good indicator that hearing sensitivity is better than 3 db HL, but absence of an eission cannot rule out noral hearing. We present below five interesting cases of pseudohypacusis. Patient histories and a coparison of ABR and behavioral results to OAEs are presented. These cases illustrate the utility of OAEs and the need for the clinician to be aware of the possible presence of pseudohypacusis. Subjects METHOD We present results of five feale patients (ages 1 to 31 years) who were evaluated in the audiology clinic at the Dartouth-Hitchcock Medical Center (DHMC). None of these patients had a history that would raise one's index of suspicion for pseudohypacusis. Procedure Audiologic evaluations for each of the subjects were conducted in the conventional anner. In addition, iittance easures were derived for all five subjects. Transient evoked otoacoustic eission (TEOAE) tests were then conducted by staff audiologists experienced in TEOAE procedures according to recoendations for OAE easureent outlined by Kep et al (199) for use with the IL88 analyzer. All subjects were tested while seated cofortably in a sound-treated booth. To test TEOAEs, the probe was fitted to each patient's ear by using probe tips of our own design, constructed fro copacted foa or rubber iittance probe tips. Special effort was ade to firly fit the probe deeply into the ear canal. In all subjects, a snug probe fit was aintained throughout the test session as indicated by the derived stability easures. Stability easureents provide a onitor of proper probe fitting by representing the degree of correlation between the initial and final stiulus easureents. They are expressed as percentages. For the five patients in this report, probe stability easures ranged fro 89 percent to 96 percent. The stiuli generated by the IL88 analyzer were the standard 8-icrosecond electric rectangular pulses presented at a rate of 5 per second at a level of 82 ± 2 db peak SPL. The routine nonlinear default ode of presenting stiulus trains was used. In this setting, subsets of four clicks are presented : three are the sae intensity, while one is 9.5 db greater and phase inverted (Kep et al, 199). The stiulus bandwidth is 5 khz, but actual width size varies depending on the probe fitting and the physical properties of each ear. In the present study, a total of 26 presentations of the stiulus train were collected (26 presentations x 4 clicks in each presentation x 4 subaverages collected in each of the two buffers = 28 stiuli). In this default ode of the IL88 analyzer, responses to the biologically generated acoustic stiuli and TEOAE stiuli are collected alternately into two buffers and are averaged. The response is windowed into 2-sec tie fraes. Containation by acoustic stiulus ringing is controlled by eliinating the first 2.5 sec of the response. Responses are filtered by a forwardand-reverse-band-pass filter with 6 and 6 Hz cut-offs and a reverse high-pass filter set at 2 Hz. The two averaged wavefors are superiposed and displayed on the ain response panel. The response spectru is calculated by taking the fast fourier transfor of the two buffers and coputing their cross-power spectru. A separate panel on the IL88 analyzer displays this spectral response as well as the noise spectru in the response. The variable noiserejection syste in the IL88 analyzer was set at 47.3 db. The presence of an OAE response was deterined based on criteria established previously (Vedanta and Musiek, 1991). First, the spectru response level had to be 3 db or ore above the noise level. Second, at least soe portion of the response had to be within 2 Hz of the exained center frequencies. Third, soe portion of the response was required to be at least 2 Hz in width and within the specified frequency region of 1, 2, 3, or 4 khz. ABR easures were obtained using a 1- icrosecond rarefaction click stiulus and insert receivers (ER-3A). A conventional recording ontage was used with the active electrode at the high forehead, the reference electrode at the ipsilateral earlobe, and the ground electrode at the contralateral earlobe. Filter settings were 15 to 3 Hz with a 6 db-per-octave roll-off, and the stiulus repetition rate was generally 15 clicks per second. ABR threshold recordings were ade initially at relatively high stiulus intensities then at decreasing intensities until a repeatable wave V could not be obtained. 294

3 TEOAEs and Pseudohypacusis/Musiek et al (Hz) 8 db nhl ] D OI C 1 2 i b d x ~, ao U Figure 1A Case 1 : audiogra. = right ear ; X = left ear. MLR easures were obtained on soe subjects using a 1-icrosecond click stiulus. A single-channel recording was ade with the active electrode at the vertex, the reference electrode at the ipsilateral earlobe, and the ground electrode at the contralateral earlobe. Ipedance across electrodes was under 5 k dz for all subjects for both the ABR and MLR. Filter settings were 3 to 25 Hz with a 6-dB roll-off, and the stiulus repetition rate was 11.1 clicks per second. Intensity functions were obtained in a anner siilar to those for the ABR. Case 1 RESULTS Case 1 was a 1-year-old feale. She was referred to the Dartouth-Hitchcock Medical 5 db nhl 3 db nhl 2 do nhl 1 sec Figure 1C Case 1 : ABR tracings fro the right ear. Center (DHMC) following ultiple conflicting hearing tests taken at her school. The test results varied, ranging fro borderline noral hearing to a severe hearing loss. At the other's request, repeated tests were adinistered over approxiately 1 year. The patient presented at DHMC with noral hearing sensitivity in the left ear and a sloping ild-to-oderate hearing loss in the right ear through 2 Hz and a severe loss at 4 Hz (Fig. 1A). These results differed fro those of the previous tests, but the audiologist suspected probles with the reliability and validity of the pure-tone results. An otologic evaluation revealed no probles. Interestingly, in contrast to what is usually seen with pseudohypacusis, the pure-tone average (PTA) was significantly better than the speech reception threshold (SRT) (37 db HL vs 53 db HL). A robust TEOAE with an overall level of 11.3 db SPL was found in the right ear (Fig. 1B). The spectru of the TEOAE response and the ABR response wavefor were consistent with noral hearing sensitivity in the 1- to 4-Hz range (Fig. 1C). 1 sec Figure 1B Case 1 : the TEOAE obtained fro the right ear.

4 ournal of the Aerican Acadey of Audiology/ Volue 6, Nuber 4, uly eooo (Hz) ; _ _ C 6 s Case 2 was a 17-year-old feale who was referred to DHMC after coplaining of a progressive hearing loss of 1-year duration in the left ear following adinistration of intravenous antibiotics and general anesthesia during surgery. Pure-tone testing at another facility suggested a oderate-to-profound hearing loss ; however, an SRT of 5 db HL and a speech recognition score of 1 percent raised soe suspicions. The suspected hearing loss in the left ear was first reported after a school hearing screening test and after observations that she had hearing difficulties. Results of agnetic resonance iaging (MRI), electronystagography, glycerol testing, and ultiple blood tests were unrearkable. However, an ABR taken at that facility was reported as "consistent with a oderate-to-severe cochlear hearing loss in the left ear," though threshold ABR testing was not perfored. Figure 2A shows this patient's audiogra obtained at DHMC. Hearing sensitivity for the right ear was within noral liits, but a oderately severe hearing loss with an unusual configuration was found for the left ear. The PTA in this ear (57 db HL) was arkedly elevated relative to the SRT (15 db HL), suggesting hearing sensitivity within noral liits. In so 9 too addition, a speech recognition score of 1 percent at 35 db HL was obtained in the left ear, also consistent with noral hearing sensitivity. The Stenger test was positive for pseudohypacusis for the frequencies 25 to 8 Hz. Figure Figure 2A Case 2 : audiogra. = right ear; X = left ear. 2B shows TEOAEs for the left ear consis- tent with hearing sensitivity within noral liits for 5 Hz through 4 Hz. Figure 2C shows an ABR wavefor intensity function for Case 2 the left ear. A repeatable wave V was elicited at 2 db nhl at 9.45 sec using insert receivers. These results suggest noral hearing sensitivity between 1 and 4 Hz. A P3 also was obtained in this patient at low intensity levels, suggesting noral hearing sensitivity at 1 and 2 Hz. Upon repeated pure-tone testing, volunteered pure-tone thresholds were borderline noral, supporting the electrophysiologic results. However, the patient continued to report hearing difficulties, and it should be reebered that the ABR at another facility indicated a possible oderate sensorineural hearing loss. Therefore, a fluctuating or recovering hearing loss ust be considered, although there is probably a nonorganic coponent. I Figure 2B Case 2: the TEOAE 1 sec obtained fro the left ear.

5 TEOAEs and PseudohypacusisfMusiek et al db nhl 7 db nhl 5 db nhl 5 db nhl 3 db nhl 2 db nhl db nhl 1'sec Figure 2C Case 2 : ABR tracings obtained fro the right ear. The latency of wave V is arked. Note two tie bases (1 and 15 sec). 1 sec Figure 3B Case 3 : the ABR tracing fro the left ear. Top and botto traces have wave V latencies arked. Case 3 Case 3 was a 31-year-old woan with a congenital, profound sensorineural hearing loss in the right ear. Serial audiogras fro a referral center indicated a fluctuating hearing loss in the left ear. Multiple blood tests taken at the referral center were within noral liits with the exception of elevated cholesterol levels. She was given steroid therapy (prednisone) with no iproveent in her hearing. She was then referred to DHMC for further evaluation. Figure 3A shows her voluntary audiogra. Initially, there was good agreeent between the PTA and the SRT in the left ear, consistent with a severe hearing loss (77 db HL vs 7 db HL, respectively). However, it was highly suspicious that her speech recognition score at a +5 db sensation level was 96 percent. A diagnostic (Hz) ABR perfored at that tie indicated a questionable sensorineural hearing loss, although site-of-lesion results did not indicate V111th nerve or low brainste involveent. Due to the diagnosis of hypercholesteroleia, the patient was encouraged to begin a low sodiu and cholesterol diet. Figure 313, which displays the ABR copleted approxiately 3 onths later (with the patient still coplaining of hearing loss), shows a large repeatable wave V at 3 db nhl. In light of these findings, electrophysiologic easureents were expanded to include the MLR. In this patient, all coponents of the MLR (Na, Pa, Nb, Pb) were present and repeatable at noral latencies at 3 db nhl (Fig. 3C). The TEOAEs were then obtained. Figure 3D shows that eissions were present fro slightly above 5 Hz through 4 Hz. The absence of a response at 5 Hz was likely due to excessive noise rather than to lack of an eission at that frequency. The overall echo level of 1.4 db SPL Pa 3 v c 4 d db nhl C ; I E x - X {, 1 Pa Pb 11 Figure 3A Case 3 : audiogra. There was no response at equipent liits for the right ear. Figure 3C Case 3 : the MLR for the left ear. Negative and positive wavefors of the MLR are arked.

6 ournal of the Aerican Acadey of Audiology/Volue 6, Nuber 4, uly 1995 (.5 MPa) db Figure 3D Case 3 : the TEOAE obtained fro the left ear. r V ~~ VIIIV71~ 1 sec and the reproducibility value of 85 percent are consistent with noral hearing sensitivity. Case 4 This 21-year-old woan was referred by another center after subjective coplaints of decreasing hearing ability over a 3-onth period. Prior to referral, audiogras indicated noral hearing sensitivity in the right ear and a ild-to-oderate hearing loss in the left ear. Pure-tone testing revealed hearing sensitivity within noral liits in the right ear, although a slightly rising configuration was noted through 3 Hz. A ild hearing loss was found in the left ear (Fig. 4A). An MRI revealed no abnoralities. In the left ear, there was poor agreeent between the PTA (43 db HL) and the SRT (2 db HL). Furtherore, good-to-excellent speech recognition scores were obtained at 15 db HL (84%) and at 25 db HL (1%). Figure 4B shows the TEOAEs for the left ear. The response spectru and the wavefor show that a cochlear response is present fro 5 Hz through 4 Hz, consistent with noral hearing sensitivity. Furtherore, a robust overall eission of 19.9 db SPL was found, which was greater in the "poorer" ear than in the noral ear, with a reproducibility of 99 percent. Figure 4C shows the ABR, which revealed a repeatable wave V at 2 db nhl with a latency of 8.3 sec using inserts. After being told of discrepancies in the test results, the patient's volunteered thresholds iproved to within noral liits and it was clear that pseudohypacusic behavior had occurred. This supports the fact that TEOAEs were accurate in identifying hearing sensitivity within noral liits. This case illustrates several iportant points. It deonstrates that TEOAEs ay be useful in identifying pseudohypacusis, even when the functional loss is reported to be in the ild hearing-loss range. It is possible that an eission could be present with the initial volunteered thresholds, but this is unlikely because of the large echo level of approxiately 2 db SPL. It is alost ipossible for an echo level of this agnitude to occur if thresholds are as poor as 4 db HL (Collet et al, 1993). It has been shown that as hearing levels becoe poorer, echo sizes becoe saller (Prieve et al, 1993). Case 5 This 15-year-old feale patient was referred to DHMC by an otolaryngologist at another center due to inconsistent audioetric results. The (Hz) P -p / I Q Figure 4A Case 4: audiogra. = right ear; X = left ear.

7 TEOAEs and Pseudohypacusis/Musiek et al Figure 4B Case 4 : the TEOAE obtained fro the left ear. patient had a history of early otitis edia, yringotoies, adenoidectoy early in life, and reports of recurrent tonsillitis and sore throats (treated with aoxicillin) that had increased in frequency during the last several years. In addition, the patient reported recent hearing difficulties. An otologic evaluation revealed no significant findings. Figure 5A shows the audiogra obtained at DHMC, indicating a flat bilateral hearing loss with PTAs of 52 db HL and 47 db HL in the right and left ears, respectively. However, responses to speech recognition tasks were obtained at 45 db HL, resulting in scores of 92 percent and 8 percent in the right and left ears, respectively. These results differed significantly fro those obtained by the referral center. An MRI did not reveal abnoralities. Figures 5B and 5C show the TEOAEs for both ears, indicating cochlear responses just above 5 Hz through 4 Hz that are consistent with noral hearing sensitivity. Cochlear echo levels and reproducibility values were within noral liits. Figure 5D shows repeatable wave Vs for the ABR at 2 db nhl bilaterally, which are in agreeent with eissions results. Figures 6 and 7 show coposite inforation for the cases reported. As a group, these patients' TEOAEs reflect noral findings in regard to overall level and reproducibility. T DISCUSSION hese five subjects deonstrated repeatable TEOAEs, suggesting hearing sensitivity within the noral-to-borderline noral range, despite uch poorer volunteered pure-tone responses. In addition, ABR responses were present at 2 db nhl in four of five of these cases, confiring integrity of the cochlea, the auditory portion of the VIIIth cranial nerve, and the low brain ste through the lateral leniscus. In view of these results and those of Robinette (1992), TEOAEs ay be used as a quick and inexpensive screening tool for suspected pseudohypacusis. However, it is iportant to note that (Hz) c X X-X , b --o-p -xo, f sec Figure 4C Case 4 : the ABR obtained fro the left ear, with the latencies of wave V arked. Figure 5A Case 5: audiogra (8 Hz was not tested). = right ear ; X = left ear.

8 ournal of the Aerican Acadey of Audiology/Volue 6, Nuber 4, uly 1995 Figure 513, 5C Case 5 : the TEOAEs for the right and left ears, respectively. 1 sec 1 sec in soe cases of neural involveent, elevated pure-tone thresholds and noral TEOAEs ay be obtained (Robinette, 1992). RE LE (inserts) 4 d8 nhl 2 db nhl Figure 5D Case 5 : the ABR traces for the right and left ears. These ABR results were obtained with ear inserts (ER-3A). Because OAEs are generated by the cochlea, coproise of the auditory nerve and auditory brainste pathway will not be reflected by OAE results. Hence, the cobination of results that yield noral OAEs and depressed pure-tone thresholds ay not always indicate pseudohypacusis as a proble. The clinician ust be aware that lesions of the auditory nerve, cochlear nucleus, and in soe cases the ore central pathways can result in depressed thresholds, and because the cochlea is not involved, the OAE should be present and noral. If a patient volunteers abnoral thresholds but noral OAEs, the clinician should entertain two possible causes : pseudohypacusis or retrocochlear involveent. If the patient is confronted and urged to respond better but does not, audiologic tests of neural and central integrity should be considered. Otoacoustic eissions ust be used judiciously by the experienced clinician in conjunction

9 TEOAEs and Pseudohypacusis/Musiek et al 6 N 9 O O r a u 7 O O. O Q #1 #2 #3 #4 #5LE #5 RE CASE Figure 6 Coposite of the overall TEOAE level for cases 1 through 5. Line indicates the lowest level of noral based on our previous data (Hurley and Musiek, in press). #1 #2 #3 #4 #5LE #5 RE CASE Figure 7 Coposite of the percentage of reproducibility for cases 1 through 5. Line indicates the lowest percentage for a noral response (Vedanta and Musiek, 1991). with other clinical tools, such as the ABR. In the present study, all noral TEOAEs showed agreeent with ABRs and in soe cases with MLR and P3 easures. In case 5, a full audiogra was not copleted before adinistering an OAE. In this case, pseudohypacusis was considered as a possibility before the pure-tone evaluation was copleted ; thus, a TEOAE test was perfored before the pure-tone evaluation was copleted. This decision was ade because it is easy to perfor this test at the first suspicion of pseudohypacusis, since the adinistration of a TEOAE test takes less than 5 inutes per ear. It is also iportant to note that nothing in any patient's history suggested pseudohypacusis. Pseudohypacusis was suspected when discrepancies eerged in the basic audiologic battery. Although soe of these patients had a history of edical factors that could produce hearing loss, the results strongly suggested pseudohypacusis. This underscores the need for the clinician to be alert to inaccurate test results. The reasons for exhibiting pseudohypacusis are coplex and extend beyond copensation purposes. Of interest is that these subjects were all quite young, with three between the ages of 1 and 17 years. Finally, it ust be noted that TEOAEs ay not be highly frequency-specific with regard to hearing sensitivity (Vedanta and Musiek,1991; Collet et al, 1993 ; Hurley and Musiek, 1994). However, the ajor proble appears to be with false-positive identification of hearing loss rather than with false-negative identification. The TEOAEs are often inaccurate due to the absence of an eission at a certain frequency, even when hearing sensitivity is within noral liits at that frequency (Prieve et al, 1993). Therefore, the presence of an eission obtained with IL88 or IL92 with volunteered thresholds greater than approxiately 3 db HL suggests pseudohypacusis or a neural basis for the hearing loss. REFERENCES Collet L, Levy V, Veuillet E, Truy E, Morgon A. (1993). Click evoked otoacoustic eissions and hearing threshold in sensorineural loss. Ear Hear 14: Hall. (1992). Handbook ofauditory Evoked Potentials. Boston : Allyn and Bacon, Hurley R, Musiek F. (1994). The effectiveness of transient evoked otoacoustic eissions (TEOAEs) in predicting hearing threshold. A Acad Audiol 5: Kep D, Ryan S, Bray P. (199). A guide to effective use of otoacoustic eissions. Ear Hear 11 :935. Prieve B, Gorga M, Schidt A, Nelly S, Peters, Schultes L, estead W (1993). Analysis of transient-evoked otoacoustic eissions in noral hearing and hearing ipaired ears. Acoust Soc A 93 : Rintelann W, Schwan S. (1991). Pseudohypacusis. In : Rintelann W, ed. Hearing Assessent. 2nd ed. Boston : Allyn and Bacon, Robinette M. (1992). Clinical observations with transient evoked otoacoustic eissions with adults. Sein Hear 13 : Vedanta R, Musiek F. (1991). Click evoked otoacoustic eissions in adult subjects : standard indices and testretest reliability. A Otol 12 :