INTRODUCTION

Differently from conventional audiometry, which encompasses the frequencies between 0.25 and 8 kHz, the high frequencies audiometry goes from 9 to 16 kHz and, depending on the equipment used, may reach 18 or even 20 kHz.

The idea of analyzing hearing in high frequencies started in the first half of the past century with STRUYCKEN'S monocordion in 1910 (1). The presence of auditory responses in the high frequencies has been mentioned by FLETCHER since 1965, although at this time, we lacked the knowledge about this human capacity (2).

The hair cells responsible for hearing sensitivity in high frequencies are located at the base of the cochlea. These are the cells initially damaged if the individual develops sensorineural hearing loss (1,3).

Besides high frequencies audiometry being important in the early detection of hearing loss, it is also useful for hearing monitoring in cases of ototoxic drug use, for instance to monitor those persons who use chemotherapy with cisplatin, carboplatin, aminoglycosides, and diuretic agents in order to avoid a degenerative process involving the spiral or Corti's organ (4,5).

VASSALLO et al. (3) studied thresholds in the frequencies of 10, 12 and 14 kHz in adults without ear problems or noise exposure, with ages varying between 20 and 50 years, 39 males and 45 females, dividing them in age ranges: from 20 to 29, from 30 to 39 and from 40 to 49 years. The greatest change in thresholds happened in the 14Khz frequency, except in the females from 20 to 29 years old, in which there was no significant hearing reduction in the tested frequencies. Male subjects had worse thresholds than their female counterparts of the same age, in all age ranges testedd.

NORTHERN et al. (5) tested 237 normo-hearing men and women, with three objectives: report the HFA thresholds in function of gender and age, review and compare results from previously held studies, paying special attention to different calibration techniques and, finally, to recommend reference standard values for 0 dBHL, for frequencies from 10,000 to 18,000 Hz. The averages of thresholds obtained in this study can be seen in Chart 1.

The thresholds were obtained in sound pressure level (SPL) measured in high frequency reference coupler. For the 20-29 and 30-39 age groups, thresholds remained relatively stable between 8 and 12 kHz, but dropped quickly for frequencies above 13 kHz. As the frequencies increased, the number of individuals able to respond in each group reduced. Only 66% of the 30-39 years of age group responded in the frequencies above 14,000 Hz.

Studies since 1961, analyzed by the authors presented reasonably agreeing results for 8 to 12 kHz frequencies.

OSTERHAMMEL (6) studied HFA in normal, diabetic and hypercholesterolemic patients of both genders with ages varying from 9 to 84 years. He asked the volunteers about their experience with the high frequencies after the test and some reported much discomfort, almost painful when 18 and 20 kHz were used in the strongest intensity (105 dBSPL). For others it was more of a "feeling" then "hearing". Others had difficulty in distinguishing the sound stimulus from the subjective tinnitus that many had after remaining for some minutes in the acoustically prepared booth.

The author concluded that the loss in conventional audiometry does not necessarily mean similar loss in the high frequencies and that HFA is useful in the investigation of factors that may be important for presbycusis and the early detection of ototoxic effects or injuries caused by environmental factors.
OLIVA et al. (7), aiming at establishing normal hearing parameters in frequencies above 8,000 Hz, carried out a preliminary study, in which they assessed bone conduction from 100 Hz to 20 kHz. Electrodes were placed in both or in only one mastoid, since according to the authors, the electrical signal spreads in any direction as it happens in bone conduction. The result achieved shows that the hearing limit for high frequencies can only be achieved by young people, usually below 20 years of age. This limit was seen in 19.6 kHz; however, the average for this age group is 18.5 kHz. As age increases, this maximum limit reduces, quantified by the authors in about 2,000 Hz per decade of life after 20 years of age. Thus, a 45 year old person will have normal hearing up to 14 kHz, approximately.

DRESCHLER et al. (8) studied the use of HFA in order to monitor and/or early detect ototoxicity (specially cisplatin). The authors noticed that HFA is an important tool to compare the ototoxicity of different treatment protocols. They concluded that the high frequencies audiometry increases cisplatin-induced ototoxicity detection, the 12 and 14 kHz frequencies are specially important, different treatment modalities may be established based on the high frequencies audiometry, the high frequencies injury happens earlier then the lower frequencies and the pre-existing hearing loss changes the pattern, but not the severity of damage caused by ototoxic drugs.
STELMACHOWICZ et al. (9) investigated the auditory thresholds in the frequencies of 8 to 20 kHz in individuals between 10 and 60 years of age, with normal hearing and without past otologic disorders. For those individuals above 30 years of age, normality thresholds related to age and gender were achieved by using CORSO (1963) norms. The authors' goal was to establish normative values for high frequencies and assess the subjects variability in relation to age. As far as age is concerned, significant differences between the groups below 19 years of age, from 20 to 24 and from 25 to 29 years of age were not found. They observed significant changes in the hearing sensitivity in those groups above 30 years of age and in the frequencies above 13kHz comparing the group of 30 to 39 years of age to the younger ones. Their study showed a significant difference on hearing sensitivity for males, which is 4.4 dB lower than that for women. There was no statistically significant difference as to gender related to either frequency or age. They observed that loss of auditory sensitivity in function of age starts early, but does not progress fast. For all groups, the standard deviation increased in function of frequency, reaches its peak value in the15 to 18 kHz range, and that in high frequencies the age-related changes start early but do not progress fast. For all the groups, the standard deviation increases in function of frequency and reaches its peak value between 12 and 16 kHz. The frequency in which the peak occurs is systematically reduced in relation to age, and this is considerably broader and less defined in the older groups. For all age groups, the maximum standard deviation occurred between 70 and 80 dB. The authors suggested that the thresholds obtained in the 10 to 19 years old group (160 individuals) should be used as a reference for HFA normality, since no change in the one year range thresholds were seen. The measures of thresholds they obtained, for the group they considered should be used for reference of normality were: 8 kHz - 28.29 dB; 9 - 30.12; 10 - 30.38; 11 - 34.06; 12 - 36.26; 13 - 40.36; 14 - 44.30; 15 - 50.60; 16 - 61.42; 17 - 69.43; 18 - 82.64; 19 - 96.25; e 20 kHz - 108.96 dB.

ZEIGELBOIM et al. (1) carried out a bibliographic review about the clinical importance of high frequency audiometry and the test variability. The clinical importance was based on the early detection of drug induced ototoxicity, once the cochlea base is the first to be affected by the drug. As to the test variability, when the tone wavelength gets closer to the length of the external acoustic meatus, it favors the creation of stationary waves, thus varying the sound pressure level along the external acoustic meatus. The stationary wave pattern depends on the size of the external ear, the middle ear impedance and the sound source characteristics. A subtle change in the placement of the headphone may also vary the sound pressure over the tympanic membrane.

Due to test variabilities, the authors suggest that the headphone should be calibrated for each individual, at each session, and this may be done through the positioning of a probe microphone in the external acoustic meatus, allowing the measurement of sound pressure close to the ear drum. Due to the lack of official guidelines for the calibration of high frequency equipment, the thresholds must be compared to reference curves related to the age of individuals with normal hearing and no otologic disorder history.

PEDALINI et al. (2) Studied the average of tone thresholds in the HFA in normal 4 to 60 years old individuals, in order to determine the average tone thresholds for high frequencies. The tested frequencies were 10, 12.5, 14 and 16 kHz. For the age range between 30 and 40, 32 individuals were studied, 12 males and 20 females, a total of 59 ears and average age of 36 years. There was hearing loss only in the 16 kHz frequency (the authors considered 25 dBHL as normality threshold) and there was no significant difference between ears, nor between genders. They observed that aging seems to have a critical effect on the tone thresholds in 16 kHz and the normality thresholds used in the conventional tone audiometry can not be the same as those for high frequency audiometry. In the 20-29 and 40-49 year groups, they found a significant difference between the genders (better in females) and reported that there is no literature consensus in these regards, however, it is known that women have better hearing in middle and high frequencies, may be due to some hormonal effect over the blood circulation in the stria vascularis. The authors highlighted that the possible factors related to the lack of studies involving sound perception in the high frequencies are: lack of proper equipment for high frequency audiometry; lack of standardization in relation to what is normal and the lack of knowledge as to the use of this type of assessment in clinical practice.

They added that, besides the physiological characteristics of aging, other histologic changes in ears exposed to noise or ototoxic drugs happen primarily in the cochlea basal turn, and this would justify this special interest for high frequency audiometry, because of the early information provided by this exam.

ZEIGELBOIM et al. (10) assessed the HFA importance in monitoring chronic renal patients who are undergoing conservative treatment. This study was carried out in individuals between 30 and 59 years of age with normal hearing, without past history of ear disorders and under conservative treatment for one year. According to the authors, the ototoxicity is a side effect of the use of certain medication and these effects may involve the cochleovestibular pathways. The study confirmed that the patients presented worse hearing results than the control group, because their high frequency hearing sensitivity was reduced with aging and frequency increas, and the patients got worse in their intra-individual assessment within one year. Among the possible explanations for the hearing loss, they highlighted the renal disease time, use of ototoxic drugs, vascular changes, high serum urea or creatinine levels, high potassium level, age and even unknown causes. They concluded that HFA is important in early detection and monitoring hearing loss, not only in chronic renal patients, but also in patients with otoneurological manifestations.

Due to the lack of HFA normality patterns, and the importance of this exam for early detection and hearing monitoring, the goal of the present study was to trace a high frequencies audiologic profile for 30 to 40 year old individuals with normal hearing, in function of side and gender.


RESULTS

Wilcoxon Test
The hearing thresholds comparisons (dBSPL) in both the right and left ears, for both genders, are presented on Tables 1 and 2.
According to the Wilcoxon test, there was no significant difference as to side in males.
According to the Wilcoxon test, there was a significant difference as to side in females only in the 11,200 Hz frequency. The worst result was seen at the right side.

Mann-Whitney test
Tables 3 and 4 show the comparison of hearing thresholds between the ears, in both genders.
According to the Mann-Whitney test, there was a significant difference between the genders for the right ear, in the frequencies of 10,000 and 16,000 Hz. However, there was no significant difference between the genders for the left ear.

Normality thresholds
Differently from conventional audiometry, we have not set the normality threshold for the high frequencies as of yet. Tables 5 and 6, together with Graphs 1 and 2 show the normal hearing thresholds for both the right and the left ears, respectively, for females.

According to the Wilcoxon test, the right and left female ears were presented separately due to a significant difference in the hearing threshold in the 11,200 Hz frequency.

Table 7 and Graph 3 show the normal thresholds for male hearing.

Since the Wilcoxon test did no show significant difference between the ears for males, and to make statistics calculations for upper and lower limits it is necessary n ³ 30, we decided to group the results of both ears, in such a way that the male sample became n = 44.

DISCUSSION

According to frequency increase there was a decrease in both ears and both genders hearing (Tables 1 and 2), in agreement with the results from NORTHERN et al. (5), who reported that only 66% of the 30 to 39 years of age group responded to frequencies above 14,000 Hz. PEDALINI et al. (2), VASSALLO et al. (3), STELMACHOWICZ et al. (9), ZEIGELBOIM et al. (10) also noticed a decrease in hearing sensitivity in 13,000 Hz, 14,000 Hz and 16,000 Hz frequencies in the same age range studied.

In our study we noticed the difficulty some individuals have in perceiving sound stimulus, specially in 16,000 Hz, and they referred a feeling of an electric current being discharged in their ears. The same fact was noticed by OSTERHAMMEL (6), who reported that in his study, for some individuals, the higher frequencies sound stimuli was more like "feeling a thing" than "hearing it". This is probably due to the fact that higher than 16,000 Hz frequencies are not adequate to differentiate hearing sensitivity alterations in function of age, because there are very few individuals who are apt to answer in this frequency.

There were no significant differences between the ears for males (Table 3). This result agrees with the PEDALINI et al. (2) and ZEIGELBOIM et al. (10) studies. The same can not be said for females, since there was a significant difference in the 11,200 Hz frequency (Table 4), and the left ear presented the best answer. Such result disagrees with ZEIGELBOIM et al. (10), who did not find differences as to the side variable in both genders. Thus, in this study, the gender variable was separated by ear.

Since the tested individuals did not state any preference earwise, the tests were started always in the right ear. It is believed that the left ear has presented better results because it is prepared to receive the sound stimulus. If this had been randomized (L or R), this would probably not have happened if it were "answer learning". If it indeed happened, one could think that the left ear is better.

When we compared each ear hearing threshold according to gender (Tables 5 and 6), we found a difference only in the right ear in the 10,000 and 16,000 Hz frequencies, and the threshold was better for males. Such result disagrees with that from Pedalini et al. (2) who did no find gender differences in the 30 to 39 year old age range, from VASSALLO et al. (3) and STELMACHOWICZ et al. (9), who found better threshold averages in females. The latter thought that the average of threshold differences in relation to gender was 4.4 dB.

It is believed that the average male high frequencies have been better due to social class, because most were from middle class and were not employees nor students from the University of Tuiuti - Paraná. The female group was comprised of professors and secretaries, cooks and janitors who worked in the University.

As one can see from our research, the high frequencies audiometry, is an important resource in early hearing loss detection, in cases of presbycusis or noise (3, 6) and for hearing monitoring during the use of ototoxic drugs (2,3,5,8,9).

Following we present the upper normal limits for right side female ears (Table 5 and Graph 2): 30 dBSPL in 8,000 Hz; 35 in 9,000 Hz; 45 in 10,000 Hz; 50 in 11,200 Hz; 55 in 12,500 Hz; 75 in 14,000 Hz and 95 in 16,000 Hz. For left ear (Table 6 and Graph 3) they were: 35 in 8,000 Hz; 35 in 9,000 Hz; 40 in 10,000 Hz; 45 in 11,200 Hz; 50 in 12,500 Hz; 75 in 14,000 Hz and 95 in 16,000 Hz.
For males, since there were no difference between the ears, we present the upper normality limits for both right and left ear as follows: (Table 7 and Graph 4), 30 in 8,000 Hz; 30 in 9,000 Hz; 35 in 10,000 Hz; 42,5 in 11,200 Hz; 45 in 12,500 Hz; 60 in 14,000 Hz e 85 in 16,000 Hz,
Despite the importance of establishing patterns of normality for age and gender, NORTHERN et al. (5) made provisional recommendations that the thresholds obtained by ZISLIS and FLETCHER (1966) in 5th grade girls to young adult girls, be used as reference because they represent the best hearing values for the 8,000 to 18,000 Hz frequencies, until some official standard be recommended. Agreeing with such recommendation, because they did not consider gender, was OLIVA et al. (7) study, which reported that hearing limits for high frequencies is only achieved by young subjects, usually below 20 years of age; and the STELMACHOWICZ et al. (9) study, that suggested the thresholds obtained for the 10 to 19 years old group be used as a normality reference for high frequency audiometry. Despite this recommendation that young subjects hearing be used as reference normality standard, we could see in this investigation that such value is not normal for individuals between 30 and 40 years old.

CONCLUSIONS
Based on the analysis of the present results, as far as hearing behavior is concerned in the frequencies between 8,000 and 16,000 Hz, in 30 to 40 year old individuals with normal hearing, we conclude that:
a) there was no significant difference in hearing thresholds in relation to the variable side in males in any frequency tested;
b) there was significant difference in relation to the variable side in females only in the 11,200 Hz frequency, and the right ear presented the worst result;
c) there was significant difference in relation to the variable gender in the 10,000 and 16,000 Hz frequencies, only in the right ear, with a better result for males;
d) the upper normality limits for high frequency audiometry in 30 to 40 year old female individuals for the right ear were: 30 dBSPL in 8,000 Hz; 35 in 9,000 Hz; 45 in 10,000 Hz; 50 in 11,200 Hz; 55 in 12,500 Hz;
75 in 14,000 Hz and 95 in 16,000 Hz;
e) the upper normality limits for high frequency audiometry in 30 to 40 year old female individuals for the left ear were: 35 in 8,000 Hz; 35 in 9,000 Hz; 40 in 10,000 Hz;
45 in 11,200 Hz; 50 in 12,500 Hz; 75 in 14,000 Hz and 95 in 16,000 Hz;
f) upper normality limits for high frequency audiometer in 30 to 40 year old male individuals were: 30 in 8,000 Hz; 30 in 9,000 Hz; 35 in 10,000 Hz; 42,5 in 11,200 Hz; 45 in 12,500 Hz; 60 in 14,000 Hz and 85 in 16,000 Hz,

REFERENCES
1. Zeigelboim BS, Fukuda Y, Iorio MCM. Audiometria de alta freqüência. Acta AWHO, 15(3): 155-8, 1996.
2. Pedalini MEB, Sanchez TG, Antonio A, Antonio W, Balbani, A, Hachiya A, et al. Média dos limiares tonais na audiometria de alta freqüência em indivíduos normais de 4 a 60 anos. Pró-Fono, 12(2): 17-20, 2000.
3. Vassallo L, Sataloff J, Menduke H. Air conduction thresholds for high frequencies. J Occup Medicine, 9: 353-7, 1967.
4. Fausti AS, Larson VD, Noffsinger D, Wilson RH, Phillips, DS, Fowler, CG. High - frequency audiometric monitoring strategies for early detection of ototoxicity. Ear Hear, 15: 232-9, 1994.
5. Northern JL, Downs MP, Rudmose W, Glorig A, Fletcher L. Recommended high-frequency audiometric threshold levels (8000-1800 Hz). J Acoust Soc Am, 52:585- 95, 1972.
6. Osterhammel D. High frequency audiometry - clinical aspects. Scand Audiol, 9:249-56, 1980.
7. Oliva JD, Paris MC, Gil MO. Estudio preliminar sobre la audiometria de alta frecuencia. Anales ORL, 14(5): 489-98, 1987.
8. Dreschler WA, Hulst RJAM, Tange RA, Urbanus NAM. The role of high-frequency audiometry in the early detection of ototoxicity. Audiology, 24: 387-95, 1985.
9. Stelmachowicz PG, Beauchaine KA, Kalberer A, Kelly WJ, Jesteadt W. Normative thresholds in the 8-to 20-khz range as a function of age. J Acoust Soc Am, 86(4): 1384-91, 1989.
10. Zeigelboim BS, Mangabeira-Albernaz PL, Fukuda Y. Audiometria de altas frequências: monitoramento auditivo. Compacta, 2(1):21-5, 2001.
11. Jerger J. Clinical experience with impedance audiometry. Arch Otolaryngol., 92:311-24, 1970.