High frequency audiometry in tinnitus patients with normal hearing in conventional audiometry

Context Hearing loss is the most important risk factor of tinnitus, but this relation is not straightforward; some patients with severe tinnitus have normal hearing, whereas many patients with hearing loss do not have tinnitus. Aims The aim was to determine if high frequency audiometry (HFA) may reveal significant differences between normal hearing participants with and without tinnitus. Settings and design This is a case–control study. Participants and methods HFA was done on two groups of participants with normal hearing sensitivity. The first group was composed of 20 adults with tinnitus, whereas the control group was 15 age-matched and sex-matched participants, not suffering from tinnitus. Statistical analysis Data were analyzed using SPSS software package version 20.0. Significance of the results was judged at the 5% level. χ2 with Fisher’s exact as a correction, Kruskal–Wallis, Mann–Whitney, and Pearson’s coefficient tests were used. Results HFA showed no significant difference between the two studied groups. Conclusion Tinnitus in normal hearing participants does not necessarily indicate corresponding damage in the cochlea


Introduction
Tinnitus is the detection of sound without an external source [1]. Most of tinnitus patients display impaired hearing threshold in the pure-tone audiometry (PTA), especially in the high frequency range [2][3][4]. Furthermore, the frequency spectrum of some individual's tinnitus matches the frequency range of the hearing impairment [5,6]. However, some tinnitus patients present with no detectable loss in the frequency range of the conventional PTA (125 Hz-8 KHz) [7].
The human ear has an auditory range that can reach up to 20 000 Hz. Frequencies between 9000 and 20 000 Hz are named extended high frequencies (EHFs) in the international literature [8]. The involvement of EHFs in auditory pathology is diverse. They affect detecting the location of the sound [9] and understanding language, especially in noisy surroundings [10]. They are also associated with agerelated hearing loss, ototoxicity, and acoustic trauma.
It has been thought that a normal PTA does not exclude cochlear damage. Damage of hair cells that code for frequencies above 8 kHz cannot be detected by the conventional audiometry. Tinnitus patients whose audiograms are normal had more frequent cochlear dead regions [11], outer hair cell damage, and impaired hearing thresholds in the EHF region [12], when compared with control groups. In contrast, tinnitus may be induced purely in the central nervous system without damage to peripheral sensory organs [13,14].
In this study, we studied the role of high frequency audiometry (HFA) in the assessment of normal hearing tinnitus patients on conventional audiometry and whether it provides more relevant information about cochlear damage not proved by the conventional audiometry.

Participants
This study was carried on 20 adults with tinnitus aged up to 50 years old with no sex preference and with normal peripheral hearing sensitivity in frequencies 250-8000 Hz. Otologic or neurologic disease, middle ear problems, and patients with occupations with noise hazards were excluded. Fifteen age-matched and sex-matched participants with normal hearing and no tinnitus were involved as control group.

Methods
All participants were subjected to history taking, otoscopic examination, tympanometry to exclude middle ear problems, conventional audiometry (air, bone conduction thresholds, and mid octaves were done).
Thresholds were assessed using the American National Standards Institute (ANSI) approach, which is an ascending technique beginning with an inaudible signal; the level was increased in 5 dB steps till a response occurred. After giving a response, the intensity was decreased by 10 dB, and another ascending series is started. The threshold was the lowest decibel hearing level at which responses occurred in at least 50% of ascending trials [15].
Normal hearing sensitivity was defined as a threshold of 20 dB HL at each frequency examined in the range from 0.25 to 8 kHz. To avoid inclusion of audiograms displaying minor dips, 3 and 6 kHz were also tested. Normal thresholds at EHFs were calculated by using mean +2 SD in the control group. Each age group was calculated separately. Participants were distributed into three age groups from 20 to 30 years, from 31 to 40 years, and from 41 to 50 years.

Pitch matching and loudness matching measurement
The first objective was to determine whether the tinnitus sounds more like pure tone or noise. Narrowband noise centered at pitch match frequency was presented with alternation with the pitch-matched tone and the patient was asked which sounds more like the tinnitus.
The pitch matching procedure is usually a twoalternative forced choice [16]. Two tones were presented to the patient and then asked to choose the one that most closely matched the tinnitus heard. This was continued until the match was made.
Tinnitus is mostly found to be a few decibels above a person's threshold for the frequency being tested [16,17]. For loudness matching, a frequency that was matched to the patient's tinnitus was presented at a level just below threshold and intensity was increased in a 1 dB step until the patient indicated a match [16].

Statistical analysis of the data
Data were analyzed using SPSS software package version 20.0 (SPSS Inc., Chicago, Illinois, USA). Significance of the obtained results was judged at the 5% level [18,19]. For demographic data, we used χ 2 -test for categorical variables; Fisher's exact as a correction for χ 2 when more than 20% of the cells have expected count less than 5; and Student's t-test for normally distributed quantitative variables. For comparing HFA thresholds in different age groups Kruskal-Wallis test was used. For comparing HFA thresholds in control and cases Mann-Whitney test was used. To study the correlation between age and high frequency thresholds Pearson's coefficient was used in cases and control groups.

Results
In the current study, 20 tinnitus patients and 15 controls were enrolled. In the cases group, there were 4 males and 16 females, whereas in the control group 3 were males and 12 were females. Age was distributed into three age groups from 20 to 30 years, from 31 to 40 years, and from 41 to 50 years. In the controls 10 ears were tested in each age group. In the cases, 12 tested ears were in the first group, 14 tested ears in the second age group, and 14 tested ears were in the last age group.
Twelve patients were hearing tinnitus in the form of tones, whereas eight were hearing it as noise. All patients had bilateral tinnitus, six of them complained with right tinnitus more than left, nine patients complained with left more than right tinnitus, and in five patients tinnitus was equal on both ears. Table 1 shows the relationship between age and HF thresholds in the control group showing mean±SD, median, minimum and maximum values.
Normal HFA thresholds were calculated by using mean +2 SD in the control group. Each age group was calculated separately. Normal hearing thresholds in HFA are shown in Table 2.
High frequency audiometry thresholds in cases Table 3 shows comparison between the two studied groups according to high frequency thresholds. This comparison was detailed and classified according to different age groups in Tables 4-6. No significant difference was found between the two groups in terms of mean thresholds across the frequency range from 9 to 16 kHz . Table 7 illustrates the number of nonresponding cases at frequency 14 kHz and 16 kHz reaching the maximum output of the audiometry. In control group, two participants showed no response at frequency of 14 KHz and six at 16 KHz. In cases group, four participants did not respond at 14 kHz and 16 at 16 kHz.
Of the 40 ears tested, only two ears showed high frequency hearing loss at 14 kHz and the remaining were normal in all other frequencies (putting in consideration that four ears out of 40 tested ears gave no response at 14 kHz, and 16 out of 40 tested ears did not give response at 16 kHz up to maximum sound level tested).     Table 3 High frequency audiometry thresholds in the two studied groups   High frequency audiometry in normal hearing tinnitus patients Elmoazen et al 311 Table 5 High frequency audiometry thresholds in the two studied groups at age from 31 to 40 years

Correlation between high frequency audiometry thresholds and age in control and cases
Tables 8 and 9 demonstrate the correlation between age and high frequency thresholds in cases and controls respectively, showing statistically significant positive correlation between age and HFA thresholds in cases starting from frequency 11.2 kHz and in control group starting from frequency 10 kHz.
Pitch matching and loudness matching Table 10 shows the distribution of the studied cases according to pitch matching and loudness matching.
Tinnitus pitch ranged from 1 to 9 kHz with a mean of 3.24 kHz. Loudness matching ranged from 14 dBHL up to 60 dBHL with mean of 31.42 dBHL.

Discussion
The main risk factor of tinnitus is HL [20]. However, this association is not simple or straightforward [21]. Some participants with troublesome tinnitus have audiometrically normal hearing and, conversely, many participants with hearing loss do not report tinnitus [20].
It has been argued that a normal PTA does not necessarily exclude cochlear damage [11]. Thus, the aim of this study was to explore the results of the HFA and see whether it provides additional information in tinnitus patients with normal hearing on conventional audiometry.

High frequency audiometry thresholds in normal participants
Normal HFA thresholds were calculated by using mean +2 SD in the control group. Each age group was calculated separately (from 20 to 30 years, from 31 to 40 years, and from 41 to 50 years).
All participants were able to respond to the maximum sound levels tested up to 12.5 kHz in the EHF range. The number of participants not responding to the maximum sound levels presented above 12.5 kHz increased as the frequency increased, especially in older age groups.
The absence of response to EHF tested in the older age groups is in accordance with other authors' reports with respect to the general tendency of a gradual decrease of hearing sensitivity at higher frequencies and with increasing age [22,23]. The shift occurs first at the highest frequencies and then progresses to lower frequencies as the participants increase in age [8].
The dispersal of the data with increasing frequency demonstrates the great variability of values present in the general population. This could be explained by individual differences in the aging process, dietary quality, and individual nutrient intake. Also environmental factors influence hearing outcomes    like noise exposure, accumulation of ototoxic materials, and the aging process itself [8].
These results were supported by another study that enrolled 645 participants from healthy volunteers. They were divided into seven age groups at 10-year intervals [8]. They showed increase in the hearing threshold as frequencies increased over the conventional and EHF range and some participants started giving no response starting from 11.2 kHz [8].
High frequency audiometry thresholds in tinnitus patients HFA didn't reveal any significant difference in mean thresholds between our group of normal hearing tinnitus patients, compared with a matched group of tinnitus-free controls suggesting that tinnitus with a normal conventional audiogram does not reflect detectable cochlear damage in the EHF range.
Supporting our results, a study done by Barnea included 17 tinnitus patients aged 21-45 years (mean=35 years) with normal hearing and 17 participants as control group based on the mean thresholds across the range from 2 to 8 kHz in each ear. Their results also showed that no significant differences were found between the two groups, in terms of mean thresholds across the frequency range from 9 to 20 kHz [24].
A study by Shim et al. [25] enrolled 18 tinnitus patients, who had a hearing levels less than 25 dB at frequencies of 250-8000 Hz. The HFA was performed, and the mean hearing thresholds at 10, 12, 14, and 16 kHz of each tinnitus ear were compared with those of the 10 age-matched and sex-matched normal ears. In this study,12 had significantly increased hearing thresholds at more than one of the four high frequencies compared with the normal group. When they assessed results according to the frequency, they found that eight patients had decreased hearing ability at 10 kHz, 10 at 12 kHz, eight at 14 kHz, and four at 16 kHz. The high number of abnormal cases compared with our study may be due to their use of the mean as normative value, whereas in the current study we used 2 SD from the mean.
A possible explanation of tinnitus with no hearing loss may be the affection of the central nervous system with no damage to the peripheral sensory organs. In most tinnitus patients, the afferent signals are affected by damage to peripheral sensory organs, and plastic changes might follow in the central auditory pathway, which may induce spontaneous activity.
However, a decrease in afferent acoustic signals is not essential [13,14]. For example, in individuals with somatic tinnitus syndrome, somatic stimuli may stimulate a specific area of the acoustic center. This may cause tinnitus, which occurs regardless of hearing ability [25].
Furthermore, in patients without a decrease in hearing ability, damage to the hair cells in peripheral sensory organs may be mild and biochemical changes preceding structural damage in the hair cells may induce tinnitus [26].Additionally, HFA had employed an evoking stimulus, and tinnitus is argued to be caused by abnormal spontaneous hyperactivity in the auditory pathways. Demonstrable differences between normal listeners with and without tinnitus might be reflected in the spontaneous activity of the auditory pathways [24].

Pitch matching and loudness matching
There was wide range of intersubject variability in tinnitus pitch (1-9 kHz). Tinnitus loudness at the tinnitus pitch frequency was found to have a mean of 31.42 dBSL (ranging from 14 dB up to 60 dBSL). A study done by Barnea [24], found pitch matching in the range from 0.25 to 16 kHz with a mean of 6.8 kHz and loudness matching mean of 15.3 dB SL with a range of 0-45 dB SL. This variability in the tinnitus pitch of the participants with normal hearing sensitivity might partially be caused by the fact that these participants struggle in establishing their tinnitus pitch [24].

Conclusion
The results of this study suggest that tinnitus with a normal conventional audiogram does not necessarily reflect appreciable cochlear damage in the EHFs, which might suggest a further central cause for tinnitus.

Recommendations
Future studies on participants with normal hearing sensitivity, particularly on the spontaneous activity of the auditory pathways, are needed to provide further information about tinnitus in normal listeners.

Financial support and sponsorship
Nil.

Conflicting of interest
There are no conflicts of interest.