2.1

Subjects

The tinnitus group comprised 8 temporal bones from 8 subjects having presbycusis with tinnitus (6 males and 2 females; age range, 64–93 years; mean age ± SD, 79.4 ± 8.7 years; high-tone-loss pattern, 1 subject, and descending pattern, 7 subjects; mean bone-conductance thresholds ± SD, 40.4 ± 15.9 dB).

The control group comprised 8 temporal bones from 8 subjects having presbycusis without tinnitus (5 males and 3 females; age range, 70–88 years; mean age ± SD, 76.1 ± 6.2 years; high-tone-loss pattern, 1 subject, and descending pattern, 7 subjects; mean bone-conductance threshold ± SD, 42.2 ± 16.4 dB).

Autopsy reports of 1000 patients from the temporal bone collection at the University of Minnesota (Minneapolis, MN) were screened to select cases with presbycusis. Patients whose temporal bones were selected for study fulfilled the following criteria for presbycusis: sensorineural hearing losses characterized by insidious onset, bilateral symmetry, progression into old age without clinical evidence of other ear disorders, and age of 60 years or older. Pure tone audiometry had been performed within 24 months of death for all of the selected subjects. On the basis of the medical records and otologic interview sheet, we excluded subjects with history of ear diseases, ototoxic drug use, head and acoustic trauma, central disease, mental disease, and systemic diseases such as diabetes or neurologic disease.

Thereafter, we divided the subjects into presbycusis with tinnitus (8 subjects) and without tinnitus (8 subjects) according to their detailed medical records. We used unilateral temporal bones from the 16 subjects. Cochlear changes were evaluated using light microscopy in both the groups. Temporal bones had previously been removed at autopsy less than 24 hours after death and fixed in a formalin solution. Each bone was decalcified, embedded in celloidin, and serially sectioned in the horizontal plane at a thickness of 20 μ m. Every 10th section was stained with hematoxylin-eosin and mounted on a glass slide for observation under light microscopy.

2.2

Audiometric classification

Patterns of audiometric hearing loss were determined based on bone-conductance thresholds at frequencies of 250, 500, 1000, 2000, 4000, and 8000 Hz. Audiometric patterns with a threshold of less than 25 dB were considered to be normal . The cases were divided into 2 groups based on the audiogram type : patients with an abrupt high tone loss (high-tone-loss pattern) and those with a gradually sloping curve (descending pattern). The high-tone-loss pattern was defined as hearing loss with a threshold of more than 25 dB at 4000 and 8000 Hz and a difference in thresholds of more than 20 dB between 2000 and 4000 Hz. The descending pattern was defined as hearing loss with a threshold of more than 25 dB at 2000, 4000, and 8000 Hz, and a difference in thresholds between 2000 and 4000 Hz with an increase of less than 20 dB. In addition to the audiometric hearing loss patterns, mean bone-conductance thresholds were assessed at 500, 1000, 2000, and 4000 Hz. Mean bone-conductance thresholds were used as a parameter for perceptive hearing loss.

2.3

Criterion of tinnitus

We obtained otologic interview sheet including tinnitus during patients’ lifetime. We therefore selected apparently subjects without tinnitus and excluded subjects with circumstance factors such as noise from subjects with tinnitus. Subjects with tinnitus had experienced tinnitus for the mean duration of 20.3 years (15.7–25.0 years).

2.4

Spiral ganglion cells

Rosenthal’s canal was divided into 4 segments, as described previously , as I (from base to 6 mm), II (6–15 mm), III (15–22 mm), and IV (22 mm to apex). Nucleoli of spiral ganglion cells were counted using light microscopy. The number of ganglion cells was determined for each segment and for the cochlear as a whole by multiplying their summed counts by 10 to account for the unmounted sections and by a factor of 0.9 to account for the cells that would be counted because of their location at the interface between sections .

2.5

Cochlear hair cells

The cochleas were reconstructed by standard cytocochleograms . In each section, the number of cochlear hair cells was assessed. The percentage loss of cochlear hair cells in each turn was counted to compare these 2 groups.

2.6

Stria vascularis and spiral ligament area

Morphometric measurements of the area counts of the stria vascularis and spiral ligament were made in all turns of the cochlea at the midmodiolar level and the 2 adjacent sections. The area of stria vascularis and spiral ligament in each turn was counted as the mean of these 3 sections. The image was acquired with a CCD camera connected to a computer. The calibrated image of the stria vascularis was obtained at a magnification of 200× and that of the spiral ligament was obtained at 40×. The areas of stria vascularis and spiral ligament were quantified by determining the areas of their cut surfaces using a computer. Measurements were made using a commercially available image analysis software Image-Pro Plus (version 3.0; Media Cybernetics, Silver Springs, MD).

2.7

Statistical analysis

Statistical evaluation was performed using the nonparametric Mann-Whitney test. A P value of less than .05 was considered significant.

2.1

Subjects

The tinnitus group comprised 8 temporal bones from 8 subjects having presbycusis with tinnitus (6 males and 2 females; age range, 64–93 years; mean age ± SD, 79.4 ± 8.7 years; high-tone-loss pattern, 1 subject, and descending pattern, 7 subjects; mean bone-conductance thresholds ± SD, 40.4 ± 15.9 dB).

The control group comprised 8 temporal bones from 8 subjects having presbycusis without tinnitus (5 males and 3 females; age range, 70–88 years; mean age ± SD, 76.1 ± 6.2 years; high-tone-loss pattern, 1 subject, and descending pattern, 7 subjects; mean bone-conductance threshold ± SD, 42.2 ± 16.4 dB).

Autopsy reports of 1000 patients from the temporal bone collection at the University of Minnesota (Minneapolis, MN) were screened to select cases with presbycusis. Patients whose temporal bones were selected for study fulfilled the following criteria for presbycusis: sensorineural hearing losses characterized by insidious onset, bilateral symmetry, progression into old age without clinical evidence of other ear disorders, and age of 60 years or older. Pure tone audiometry had been performed within 24 months of death for all of the selected subjects. On the basis of the medical records and otologic interview sheet, we excluded subjects with history of ear diseases, ototoxic drug use, head and acoustic trauma, central disease, mental disease, and systemic diseases such as diabetes or neurologic disease.

Thereafter, we divided the subjects into presbycusis with tinnitus (8 subjects) and without tinnitus (8 subjects) according to their detailed medical records. We used unilateral temporal bones from the 16 subjects. Cochlear changes were evaluated using light microscopy in both the groups. Temporal bones had previously been removed at autopsy less than 24 hours after death and fixed in a formalin solution. Each bone was decalcified, embedded in celloidin, and serially sectioned in the horizontal plane at a thickness of 20 μ m. Every 10th section was stained with hematoxylin-eosin and mounted on a glass slide for observation under light microscopy.

2.2

Audiometric classification

Patterns of audiometric hearing loss were determined based on bone-conductance thresholds at frequencies of 250, 500, 1000, 2000, 4000, and 8000 Hz. Audiometric patterns with a threshold of less than 25 dB were considered to be normal . The cases were divided into 2 groups based on the audiogram type : patients with an abrupt high tone loss (high-tone-loss pattern) and those with a gradually sloping curve (descending pattern). The high-tone-loss pattern was defined as hearing loss with a threshold of more than 25 dB at 4000 and 8000 Hz and a difference in thresholds of more than 20 dB between 2000 and 4000 Hz. The descending pattern was defined as hearing loss with a threshold of more than 25 dB at 2000, 4000, and 8000 Hz, and a difference in thresholds between 2000 and 4000 Hz with an increase of less than 20 dB. In addition to the audiometric hearing loss patterns, mean bone-conductance thresholds were assessed at 500, 1000, 2000, and 4000 Hz. Mean bone-conductance thresholds were used as a parameter for perceptive hearing loss.

2.3

Criterion of tinnitus

We obtained otologic interview sheet including tinnitus during patients’ lifetime. We therefore selected apparently subjects without tinnitus and excluded subjects with circumstance factors such as noise from subjects with tinnitus. Subjects with tinnitus had experienced tinnitus for the mean duration of 20.3 years (15.7–25.0 years).

2.4

Spiral ganglion cells

Rosenthal’s canal was divided into 4 segments, as described previously , as I (from base to 6 mm), II (6–15 mm), III (15–22 mm), and IV (22 mm to apex). Nucleoli of spiral ganglion cells were counted using light microscopy. The number of ganglion cells was determined for each segment and for the cochlear as a whole by multiplying their summed counts by 10 to account for the unmounted sections and by a factor of 0.9 to account for the cells that would be counted because of their location at the interface between sections .

2.5

Cochlear hair cells

The cochleas were reconstructed by standard cytocochleograms . In each section, the number of cochlear hair cells was assessed. The percentage loss of cochlear hair cells in each turn was counted to compare these 2 groups.

2.6

Stria vascularis and spiral ligament area

Morphometric measurements of the area counts of the stria vascularis and spiral ligament were made in all turns of the cochlea at the midmodiolar level and the 2 adjacent sections. The area of stria vascularis and spiral ligament in each turn was counted as the mean of these 3 sections. The image was acquired with a CCD camera connected to a computer. The calibrated image of the stria vascularis was obtained at a magnification of 200× and that of the spiral ligament was obtained at 40×. The areas of stria vascularis and spiral ligament were quantified by determining the areas of their cut surfaces using a computer. Measurements were made using a commercially available image analysis software Image-Pro Plus (version 3.0; Media Cybernetics, Silver Springs, MD).

2.7

Statistical analysis

Statistical evaluation was performed using the nonparametric Mann-Whitney test. A P value of less than .05 was considered significant.