Auris Nasus Larynx 31 (2004) 383–388

www.elsevier.com/locate/anl

Effects of aminoglycoside administration on cochlear elements

in human temporal bones
Takeshi Kusunokia,b, Sebahattin Cureoglua,c, Patricia A. Schachernc,*,
Andre Sampaioa, Hisaki Fukushimaa, Mehmet F. Oktaya,
Michael M. Paparellad
a
International Hearing Foundation, Minneapolis, MN, USA
b
Department of Otolaryngology, Kinki University School of Medicine, Osaka, Japan
c
Department of Otolaryngology, Otitis Media Research Center, University of Minnesota, Minneapolis, MN, USA
d
Minnesota Ear Head and Neck Clinic, Minneapolis, MN, USA
Received 25 November 2003; received in revised form 13 September 2004; accepted 24 September 2004
Available online 13 November 2004

Abstract

Objective: Although there have been numerous reports on the relationship between the period of aminoglycoside administration and cochlear
damage in animals, to date there have been no such studies in humans. The purpose of this study is to observe the early and late cochlear effects
of aminoglycoside administration on hair cells, spiral ganglion cells, stria vascularis, and spiral ligament.
Methods: Specimens were divided into three groups. Group I included ‘‘normal’’ temporal bones with no histopathologic findings of otitis
media and no history of otologic or ototoxic drug administration. Group II consisted of temporal bones that received aminoglycosides within 2
weeks before death and group III of temporal bones that had aminoglycosides from 2 weeks to 6 months prior to death. Patients in groups II
and III received gentamycin, kanamycin or tobramycin. Temporal bones were excluded from groups II and III if patients had a history of
otologic disease or other ototoxic drugs. All temporal bones were examined under light microscopy. Standard cytocochleograms and spiral
ganglion cell reconstructions were done on all temporal bones.
Morphometric measurements of areas of stria vascularis were made in all turns of the cochlea on mid-modiolar sections. Spiral ligament
was divided into four segments according to the locations of different types of fibrocytes. The mean loss of fibrocytes in each segment was
estimated.
Results: The percentages of intact outer hair cells in the basal turn were significantly greater in group I compared to groups II and III. The
mean area of the stria vascularis in the apical turn was significantly less in groups II and III compared to group I.
Conclusion: This study demonstrates that in a short period (within 2 weeks) after aminoglycoside administration, a decrease in hair cells and
in the area of the stria vascularis occurred.
# 2004 Published by Elsevier Ireland Ltd.

Keywords: Ototoxicity of aminoglycoside; Human temporal bone; Hair cell; Spiral ganglion cell; Stria vascularis; Spiral ligament

1. Introduction (guinea pig and Mongolian gerbil). The relationship

between the period of aminoglycoside drug administration
There have been many reports [1–4] of hair cell loss in the and cochlear damage is important to understand the ototoxic
cochlea due to the ototoxicity of aminoglycosides in animals process and to clinically monitor the hearing of the patients
during their administration. Although there have been
numerous studies in animals [5,6], there have been no
* Corresponding author. Present address: Room 226 Lions Research
Building, 20016th St. S.E., Minneapolis, MN 55455, USA.
reports on human temporal bones regarding the relationship
Tel.: +1 612 626 9876; fax: +1 612 626 9871. between the period of aminoglycoside drug administration
E-mail address: [email protected] (P.A. Schachern). and cochlear damage. Therefore, we studied the early and

0385-8146/$ – see front matter # 2004 Published by Elsevier Ireland Ltd.

doi:10.1016/j.anl.2004.09.011
384 T. Kusunoki et al. / Auris Nasus Larynx 31 (2004) 383–388

late cochlear effects of aminoglycoside administration on their location at the interface between sections. For
hair cells, spiral ganglion cells and the lateral wall (stria statistical analysis, one-way analysis of variance (ANOVA)
vascularis and spiral ligament) in the human cochlea. was used to compare the mean number of spiral ganglion
cells in the three groups.

2. Methods 2.4. Stria vascularis

2.1. Subjects Morphometric measurements of area counts of the stria

vascularis were made in all turns of the cochlea at the mid-
This study included 42 temporal bones from 29 subjects modiolar section and the adjacent two sections and the mean
ranging in age from 12 to 78 years. Specimens were divided of the three sections determined. The image was acquired
into three groups. Group I included 16 ‘‘normal’’ temporal with a CCD camera connected to a personal computer. The
bones from 10 patients (mean: 45.5 years, range: 12–78), calibrated image was obtained at a magnification of 80 for
with no histopathologic findings of otitis media and no the stria vascularis. The areas of stria vascularis were
history of otologic disease, diabetes, immune disease or of quantified by determining the areas of their cut surfaces,
receiving ototoxic drugs. Group II consisted of 14 temporal with the aid of a computer. The measurements were made by
bones from 11 patients (mean age: 40.5 years, range: 16–76) using commercially available image analysis software,
that received aminoglycosides within 2 weeks before death Image-Pro@ Plus (Media Cybernetics, Silver Springs, MD;
and group III of 12 temporal bones from 8 patients (mean Version 3.0). Secondary changes such as cystic-like
age: 40.6 years, range: 18–71) that had aminoglycosides structural areas and concretions in the stria vascularis of
from 2 weeks to 6 months before death. Patients in groups II the specimens were excluded from their areas. Statistical
and III received gentamicin, kanamycin, or tobramycin. analysis included one-way analysis of variance (ANOVA) to
Therapeutic doses were as follows: gentamycin, 80 mg  2 compare the mean areas of the stria vascularis in the three
or 3; kanamycin, 500 mg  2 or 3; tobramycin, 80 mg  2 groups. Two temporal bones were excluded because of
or 3 for at least 10 days. Temporal bones were excluded from removal artifacts in the stria vascularis. This study included
groups II and III if patients had a history of otologic disease, 40 temporal bones from 27 subjects ranging in age from 12
immune disease or other ototoxic drugs. All temporal bones to 78 years old. Group I included 16 ‘‘normal’’ temporal
had been removed at autopsy, less than 24 h after death, and bones from 10 patients (mean age: 45.5 years, range: 12–78).
fixed in formalin solution. Each bone was decalcified; Group II consisted of 13 temporal bones from 10 patients
embedded in celloidin, and serially sectioned in the (mean age: 42.1 years, range: 16–76) and group III, 11
horizontal plane at a thickness of 20 mm. Every 10th temporal bones from 7 patients (mean age: 43.3 years, range:
section was stained with hematoxylin and eosin and 18–71).
mounted on a glass slide for light microscopic assessment.
2.5. Spiral ligament
2.2. Hair cells
The spiral ligament was divided into four segments
The cochleae were reconstructed by standard cytoco- according to the appearance of different types of fibrocytes
chleograms according to the methods of Schuknecht and based on the results of previous studies by Spicer and
Gacek [7]. Five temporal bones were excluded because of Schulte [9]. Type I fibrocytes lay circumferentially aligned
removal and compression artifacts in the hair cells. The between the stria vascularis and bone. Type II fibrocytes
percentage of intact outer hair cells was observed in all occupy the superficial inferior spiral ligament between the
available sections and calculated for each turn as a function basilar crest and the stria. Type III fibrocytes are long-
of the total number of outer hair cells possible in that turn. itudinally located in the deepest part of the inferior spiral
This resulted in a percentage for the three turns of each bone. ligament. Type IV fibrocytes lie radially oriented inferior to
To analyze differences among groups, we used Mann– the basilar crest. A rating score was used to evaluate cells in
Whitney’s U-test. the spiral ligament. The score was classified as: 0—within
normal limits (missing less than 1/3 of the fibrocytes); 1—
2.3. Spiral ganglion cells missing 1/3 of the fibrocytes; 2—missing 2/3 of the
fibrocytes; and 3—severe or complete loss at the level of the
Rosenthal’s canal was divided into four segments as mid-modiolar section according to the method of Hequem-
described previously by Otte et al. [8]. Nuclei were counted bourg and Liberman [10]. The average loss of fibrocytes in
in each section using light microscopy. The number of each segment was estimated and evaluated using the above
ganglion cells was determined for each segment and for the scale. To determine if loss of fibrocytes occurred in a
cochlea as a whole by multiplying their summed counts by uniform manner in the cochlea, we examined all available
10 to account for the unmounted sections and by a factor of sections throughout the cochlea in five subjects. Findings
0.9 to account for cells that would be counted because of were compared with those of the mid-modiolar section using
 T. Kusunoki et al. / Auris Nasus Larynx 31 (2004) 383–388 385

Table 1
Spiral ganglion cells and hair cells

Mann–Whitney’s U-test. We did not observe any significant 3. Results

differences in loss of fibrocytes in any turn between the mid-
modiolar section and other sections. One-way analysis of 3.1. Hair cells
variance (ANOVA) was used to compare the mean loss
scores of the spiral ligament in the three groups. Two The assessment of hair cells (Table 1) could not be done
temporal bones were excluded because of removal artifacts in the basal hook region of the cochlea where the supporting
in the spiral ligament. Therefore, this study included 40 cells are superimposed on the hair cells. Inner hair cells of
temporal bones from 27 subjects ranging in age from 12 to groups I–III were mostly intact in all turns. In group I, 1 out
78 years. Group I included 16 ‘‘normal’’ temporal bones of 16 temporal bones from 10 patients had outer hair cell loss
from 10 patients (mean age: 45.5 years, range: 12–78); primarily in the apical turn. In one case of a 40-year-old from
group II consisted of 13 temporal bones from 10 patients group I, missing outer hair cells were seen in over 50% of the
(mean age: 42.1 years, range: 16–76) and group III of 11 apical turn. In group II, 3 out of 11 temporal bones (three
temporal bones from 7 patients (mean age: 43.3 years, temporal bones were excluded from this study because of
range: 18–71). artifacts in hair cells) had outer hair cell damage greater than
386 T. Kusunoki et al. / Auris Nasus Larynx 31 (2004) 383–388

Table 2
Mean of the areas of the stria vascularis

50% in the cochlea. One temporal bone had missing outer There was no relationship between the number of spiral
hair cells in over 50% of the apical turn. In two other ganglion cells and the damage of outer hair cells (Table 1).
temporal bones, missing outer hair cells were observed in
over 50% of the basal turn. Two of three temporal bones with
3.3. Stria vascularis
outer hair cell damage greater than 50% were under 32 years
of age. In group III, 4 of 10 temporal bones (two temporal
The mean areas of the stria vascularis (Table 2) in the
bones were excluded from this study because of artifacts in
apical turns of groups II and III were significantly less than
hair cells) had outer hair cell damage greater than 50% in the
group I (p < 0.05, p < 0.001). There was no significant
cochlea. In three of these, missing outer hair cells were
difference between groups I and II, or III in the middle or
observed in over 50% of the basal turn of the cochlea (Fig.
basal turns (p > 0.05). Some cases of older patients from all
1A and B). In one of them, loss of outer hair cells was
three groups had atrophy of the stria vascularis with
primarily in the apical turn. Two of the four temporal bones
vacuolization, shrinkage of cytoplasm and low cellularity. In
with severe outer hair cell damage in group III were 25 years
some cases, hair cell damage was seen in the same section as
of age.
those with degenerative changes of the stria vascularis and
The percentages of intact outer hair cells in the basal turn
the extent of missing outer hair cells occurred in over 50% of
was significantly greater in group I compared to groups II
the turn. (Fig. 2A and B).
and III (p < 0.05). Although not significant, the incidence of
outer hair cells with damage greater than 50% in group II
27.3% (3/11) and group III 40.0% (4/10) was higher than in 3.4. Spiral ligament
group I; 0.06 % (1/16). The incidence of outer hair cells with
damage greater than 50% in patients 0–32 years of age in In groups I–III, fibrocytes in the spiral ligament tended to
group II was 28.6% (2/7) and in group III—40.0% (2/5). decrease with increasing age. Types II and IV fibrocytes
These were higher than those of group I—0.0% (0/6). were the first to decrease followed by types I and III,
respectively. The most prominent loss of fibrocytes (Table 3)
3.2. Spiral ganglion cells including all types was observed in the apical turn. Loss of
types II and IV occurred in temporal bones from younger
There was no significant difference in the number of patients (Fig. 2A). There was no significant difference in the
spiral ganglion cells among the three groups (p > 0.05). loss scores of fibrocytes among the three groups (p > 0.05).

Table 3
Mean of the loss scores of the fibrocytes
 T. Kusunoki et al. / Auris Nasus Larynx 31 (2004) 383–388 387

Fig. 2. (A) Haematoxylin eosin, original magnification 100. In the left

temporal bone of an 18-year-old from group I, the stria vascularis and inner
and outer hair cells in the apical turn are intact (high power-view (200) of
the square). However, there is severe loss of types I, II and IV fibrocytes. (B)
In the right temporal bone of this 17-year-old from group II, the stria
vascularis in the apical turn is atrophic with vacuolization, shrinkage of
cytoplasm and lower cellularity than the young aged cases in group I (Fig.
2A). Note the apical turn with no outer hair cells (arrow in high power-view
(200) of the square). Haematoxylin eosin, original magnification 100.

Fig. 1. (A) Both outer hair cells and inner hair cells are intact in the basal why outer hair cells of the middle turn were mostly intact in
turn of this left temporal bone from a 25-year-old from group I. Haematox-
ylin eosin, original magnification 400. (B) There are no outer hair cells
both presbycusis and aminoglycoside ototoxicity. It appears
(arrow) in the basal turn of this left cochlea of a 25-year-old from group III. that outer hair cells of the middle turn may be more resistant
Haematoxylin eosin, original magnification 400. to pathological change.
Otte et al. [8] reported the mean number of spiral
ganglion cells as 28,620 at adult normal ears (20–80 years).
4. Discussion Their data is similar to our findings. We found no significant
difference in the number of spiral ganglion cells among our
McFadden et al. [11] reported that age-related cochlear three groups. There was no relationship between the number
hair cell loss in mice advanced from the basal to the apical of spiral ganglion cells and the severity of damage of the hair
turn and there was a greater loss of outer hair cells than cells.
inners. Schuknecht and Gacek [7] reported that in human In our study, group II as well as group III showed a
cochlear pathology in presbycusis, some cases had damage significantly lower percentage of intact outer hair cells than
in the apical turn as well as in the basal turn, but the middle group I. Forge and Fradis [5] reported that changes of hair
turn was intact. In our study, one case that had missing outer cells and stria vascularis in guinea pigs could be observed as
hair cells in over 50% of the apical turn was 40 years old early effects of gentamicin. Moreover, Matz [13] reported
(group I). This could be related to aging. Hair cell loss in that hair cells were primarily affected by aminoglycosides.
aminoglycoside ototoxicity has been reported to be This may suggest that hair cells may be more susceptible to
primarily in the outer hair cells of the basal turn of the ototoxicity than spiral ganglion cells.
cochlea. The loss of outer hair cells in the basal turn was Forge and Fradis [5] reported that at 4 weeks post-
greater than that of the inner hair cells [11–13]. In our study, treatment, there was a highly significant decrease in the
group I did not have any case with missing outer hair cells thickness of the stria vascularis. In our study in the apical
that was greater than 50% of the basal turn, however missing turns, the areas of the stria vascularis were significantly less
outer hair cells of more than 50% of the basal turn were in groups II and III compared to group I, suggesting that the
observed in groups II and III and would seem to be the affect early effects of aminoglycoside ototoxicity may occur in the
of ototoxicity rather than aging. Our study cannot explain apical turn of the stria vascularis. However, there was no
388 T. Kusunoki et al. / Auris Nasus Larynx 31 (2004) 383–388

significant difference in areas of the stria vascularis of the This study demonstrates that even a short period after
other turns among any group. Matz [13] suggested that aminoglycoside administration can promote severe outer
degeneration of the stria vascularis caused by aminoglyco- hair cell loss, and a decrease in the area of the stria
sides starts at the basal turn of the cochlea and progresses vascularis.
toward the apex. Hawke and Jahn [14] reported that
degeneration of the stria vascularis due to aging would also
start at the basal turn and proceed toward the apex. The area
of the stria vascularis in the basal turn of group I showed a References
tendency to decrease due to aging, however, the greater
decrease of the apical turn in groups II and III is probably [1] Forge A. Outer cell loss and supporting cell expansion following
due to ototoxicity of the aminoglycoside. In the mean areas chronic gentamicin treatment. Hearing Res 1985;19:171–82.
[2] Rahoash Y, Altschuler RA. Scar formation drug-induced insult.
of the stria vascularis in the apical turns, the significant Hearing Res 1991;51:173–84.
difference between groups III and I (p < 0.001) was greater [3] McDowell B. Patterns of cochlear degeneration following gentamicin
than between II and I (p < 0.05). This result suggests that the administration in both old and young guinea pigs. Br J Audiol
ototoxic effects to the stria vascularis in the apical turns 1982;16:123–9.
occurs later than 2 weeks after aminoglycoside administra- [4] Wanamaker HH, Gruenwald L, Damm KJ, Ogata Y, Slepecky N.
Dose-related vestibular and cochlear effects of transympanic genta-
tion. Forge and Fradis [5] reported that structural micin. Am J Otol 1998;19:170–9.
abnormalities were observable in both the stria vascularis [5] Forge A, Fradis M. Structural abnormalities in the stria vascularis
and hair cells at the earliest time period after gentamicin following chronic gentamicin treatment. Hearing Res 1985;20:
treatment in albino guinea pigs. 233–44.
An animal study by Hequembourg and Liberman [10] [6] Forge A, Wright A, Davies SJ. Analysis of structural changes in the
stria vascularis following chronic gentamicin treatment. Hearing Res
suggested that degeneration in the spiral ligament, 1987;31:253–66.
particularly the crest of the spiral ligament, might be the [7] Schuknecht HF, Gacek MR. Cochlear pathology in presbycusis. Ann
root cause of age-related hearing loss. Our qualitative Otol Rhinol Laryngol 1993;102:1–16.
analysis in human temporal bones supports their finding. We [8] Otte J, Schuknecht HF, Kerr AG. Ganglion cell populations in normal
observed more loss of types II and IV fibrocytes with aging. and pathological human cochlea. Implications for cochlear implanta-
tion. Laryngoscope 1978;88:1231–46.
This loss was seen in all turns, starting at the apical turn even [9] Spicer SS, Schulte BA. Differentiation of inner ear fibrocytes
in the younger subjects from group I. The most prominent according to their ion transport related activity. Hear Res 1991;56:
loss of fibrocytes in all types was observed in the apical turn. 53–64.
There was no significant difference in the mean number of [10] Hequembourg S, Liberman MC. Spiral ligament pathology: a major
fibrocytes among any of the groups. Therefore it is unlikely aspect of age-related cochlear degeneration in C57BL/6 Mice. JARO
2001;2:118–29.
that ototoxicity was a factor. [11] McFaddwn SL, Ding D, Reaume AG, Flood DG, Salvi RJ. Age-related
We could not discuss the relationship between damage of cochlear hair cell loss is enhanced in mice lacking copper/zinc
the cochlear element and audiometric findings, because most superoxide dismutase. Neurobiol Aging 1999;20:1–8.
subjects did not undergo audiograms prior to death. [12] Hinojosa R, Nelson EG, Lerner SA, Redleaf MI, Schramm DR.
Schuknecht and Gacek [7] reported the concept of four Aminoglycoside: a human temporal bone study. Laryngoscope
2001;111:1797–805.
predominant pathologic types of presbycusis; these being [13] Matz G. Aminoglycoside cochlear toxicity. Otolaryngol Clin North
sensory, neural, strial, and cochlear conductive. If ototoxic Am 1993;26:705–12.
inner ear changes of aminoglycoside administration follow [14] Hawke M, Jahn AF. In: Diseases of the ear, clinical and pathologic
the above concept, the sensory and strial type may overlap. aspects. Niigata: Nishimura Bookstore; 1990. p. 283–9.