The Laryiigoscope
Lippincott-Raven Publishers, Philadelphia
0 1997 The American Laryngological,
Rhinological and Otological Society, Inc.
Interaction Between Hypertension and
Diabetes Mellitus in the Pathogenesis of
Sensorineural Hearing Loss
Sigsbee W. Duck, MD; Jiri Prazma, MD, PhD; P. Scott Bennett, MD; Harold C. Pillsbury, MD
The purpose of this study is to support the hy-
pothesis that diabetic end-organ damage of the
cochlea is augmented in the setting of hyperten-
sion. A historical perspective reviewing the effects
of diabetes and hypertension as causative factors in
the development of sensorineural hearing loss, as
well as the basic epidemiology and pathophysiol-
ogy of the renal and vascular effects of diabetes and
hypertension, is presented. The results of audio-
logic findings in insulin-dependent diabetic pa-
tients, both normotensive and hypertensive, were
analyzed and correlated with the results of animal
studies to support the hypothesis that sensori-
neural hearing loss in patients and cochlear hair
cell loss in animal studies result from the effects of
hypertension in conjunction with insulin-depen-
dent diabetes mellitus.
Laryngoscope, 107:1696-1605,1997
INTRODUCTION
Over the years in our practice, we have seen di-
abetics with substantial hearing loss and others
with normal hearing. This discrepancy has triggered
our curiosity to look for factors that may be respon-
sible for this difference. Therefore the question was
raised, what is the factor, either alone or in conjunc-
tion with insulin-dependent diabetes mellitus, that
is responsible for the differences in the audiometric
findings of these patients?
Two of the most common chronic medical prob-
lems encountered by physicians are diabetes melli-
tus and hypertension. Conservative estimates of the
prevalence of diabetes mellitus in the general popu-
lation are at least 1%, and hypertension is even
Accepted with Honorable Mention as a Candidate's Thesis by the
Amtxrican Laryngological, Rhinological and Otological Society, Inc.
From the Division of Otolaryngology, School of Medicine, University
o f North Carolina a t Chapel Hill, Chapel Hill, North Carolina.
Send Reprint Requests to Sigsbee W. Duck, MD, Sagebrush Ear,
Nost. and Throat, P.O. Box 2500, Gillette, WY 82717, U S A .
Laryngoscope 107: December 1997
1596
more prevalent, affecting more than 20% of the gen-
eral population of the United States.1 These two con-
ditions, affecting millions of Americans, have been
shown to cause hearing loss in patients,ZJ as well as
cochlear damage in animal studies.435
Diabetes was first associated with hearing loss
by Jordao6 in 1857 and since then has been shown to
cause a slowly progressive, bilaterally symmetric,
sensorineural hearing l0ss.7 Chronic hypertension
has also been implicated in the development of hear-
ing loss.33 Studies have documented that chronic
hypertension potentiates noise-induced decreases in
cochlear function and the development of histologic
cochlear darnage.g.10 A significant relationship
among hypertension, age, and hearing loss in an an-
imal model has also been confirmed.11
We hypothesize that diabetic end-organ damage
of the cochlea is augmented in the setting of hyper-
tension. To our knowledge, no study to date has docu-
mented the effects of concomitant diabetes and hy-
pertension on the inner ear. To examine the effects of
diabetes on the cochlea in the setting of hypertension,
we conducted a prospective analysis of the clinical au-
diologic findings in normotensive and hypertensive
insulin-dependent diabetics and analyzed cochlear
hair cell losses in hypertensive diabetic rats, normo-
tensive diabetic rats, and normotensive nondiabetic
rats.
MATERIALS AND METHODS
Part A
Twenty-two patients with a history of insulin-depen-
dent diabetes mellitus participated in this study. They
were referred by internists for inclusion in this study. Pa-
tients were not considered as hypertensive candidates for
this study unless they had been treated for hypertension
for 2 or more years. Twelve of the patients were being
treated for hypertension for 2 or more years, and 10 had
no history of hypertension. The average age of the pa-
Duck et al.: Pathogenesis of Hearing Loss
tients, the duration of treatment with insulin, the use of
antihypertensive therapy, and other factors contributing
to hearing loss were recorded and evaluated. Patients
with a history of noise exposure (i.e., from hunting or
heavy industrial exposure) and other extrinsic factors
(i.e., aminoglycoside therapy) were excluded. All patients
were evaluated with complete audiometric evaluations in-
cluding airhone, speech reception threshold (SRT) and
speech discrimination scores. The audiologist was blinded
as to the hypertensive status of each of the 22 subjects. If
a significant difference (510 dB) was noted when evaluat-
ing hearing in each ear, the subject was eliminated.
The hearing of diabetic patients was analyzed using
analysis of covariance (ANCOVA), a statistical technique
used to control for the influence of a confounding variable.
Confounding variables occur most often when subjects
cannot be assigned at random to different groups. In this
study hypertensive diabetics were older than normoten-
sive diabetics; because this age factor was not possible to
control, these data were analyzed by ANCOVA.
Part B
This study used a total of 37 animal subjects consist-
ing of 11 LA/N-cp normotensive nondiabetic rats, 13
WKY/N-cp diabetic normotensive rats, and 13 SHRIN-cp
spontaneously hypertensive diabetic rats.
The W - c p normotensive nondiabetic rat was de-
veloped by crossing the Albany/N strain with a hooded
strain.12The WKY/N-cp diabetic normotensive rat was de-
veloped from the Wistar-Kyoto rat strain as an animal
model to study diabetic pathologic features. The SHRIN-cp
spontaneously hypertensive diabetic rats were developed
at the National Institutes of Health (NIH) by establishing
the corpulent cp gene from the obese Koletsky rat strains
into the SHRIN strains.12 The availability of these strains
provides excellent models for the study of the effects of hy-
pertension and diabetes on the inner ear.
All 37 rats were raised in hanging metal cages in a
noise-free environment. The animals were on a light cycle
of 12 hours of light and 12 hours of dark controlled by an
automatic clock. The housing room had a controlled tem-
perature between 21°C and 25°C and humidity between
40% and 50%. This study was performed in accordance
with the Public Health Service (PHS) Policy on Humane
Care and Use of Laboratory Animals.13 The animals were
maintained on diets of 54% starch, 10% casein, 10% lac-
talbumin, 5.9% cellulose, 4% beef tallow, 4% lard, 4% corn
oil, 4%) coconut oil, 1%salt mix, and 1%vitamin mix. They
were sacrificed at 32 to 33 weeks of age. After sacrifice, the
cochleae were fixed in a mixture of Bouin’s solution (a so-
lution of saturated aqueous picric acid, 10% formaldehyde
solution, and 5% acetic acid) and then stored in phos-
phate-buffered 1%paraformaldehyde at a pH of 7.4.
The bony capsule was decalcified with a 10% eth-
lyenediamine-tetraacetic acid (EDTA)in phosphate buffer
at a pH of 7.4. Using a surface preparation technique, the
cochleae were cut into halves along the long axis of the
modiolus. The basilar membranes were removed under a
dissecting microscope by scooping out the stria and basi-
lar membrane with subsequent dissection of the basilar
membrane from the stria, or by dissection of the basilar
membrane from the stria in situ. The dissections were
performed in a phosphate-buffered solution using scoop-
Laryngoscope 107: December 1997
ing and cutting tools fashioned from 30-gauge needles.
The specimens were stained with 0.5% toluidine blue in
0.5% borax and mounted in glycerin on a slide.
Hair cells were counted using light microscopy in
200-ym segments, beginning a t the hook and continuing
toward the apex of the cochlea. The cochlea was divided
into four regions for comparison, including the hook, the
first turn, the second turn, and the third turn. The first
and second turns were further divided into quarter turns
for comparisons between groups. Dividing the cochlea in
this way allowed for a more complete and descriptive
analysis of the distribution of hair cell loss throughout the
cochlea in the different study groups. For each segment
the number of missing hair cells was recorded.
Experimental results collected from the diabetic and
normal rats were analyzed using nonparametric tests. Most
frequently scientific data are analyzed using parametric
tests such as analysis of variance (ANOVA) and Student’s
t-test. However, parametric tests assume a normal distribu-
tion in experimental groups. Inspection of the raw data in-
dicated that the assumption of homogeneity of variance
across treatments, which underlies parametric analysis,
was not met; therefore analysis of effects was undertaken
with nonparametric tests. In this study the distribution be-
tween the samples was relatively large. For example, in hy-
pertensive diabetics when comparing the distribution in the
first turn, third quarter (SD = 7.91) t o the distribution in
the second turn, second quarter (SD = 27.79), the distribu-
tion in the samples was large; therefore statistical evalua-
tion of the data was ranked and then analyzed using a two-
way repeated-measure ANOVA. A P value of less than 0.05
was considered statistically significant.
RESULTS
Part A
All data are presented as a mean, with the stan-
dard error of the mean in parentheses. The average
age of the entire group of patients was 55.7 years
(24.39). The average age of normotensive insulin-
dependent diabetic patients was 47 years (+-5),as
compared with 62.4 years (23.78) of age for hyper-
tensive insulin-dependent diabetic patients. The av-
erage duration of therapy with insulin for normoten-
sive diabetics was 13.4 years (23.72) and for
hypertensive diabetics, 7.3 years (22.35). The average
duration of hypertension requiring treatment was
15.1 years (22.49). The average SRT of normotensive
insulin-dependent diabetics was 11 dB (22.69) and
for hypertensive insulin-dependent diabetics, 26.67
dB (24.48). The average discrimination score of nor-
motensive insulin-dependent diabetics was 98.8%
(20.8%)and for hypertensive insulin-dependent dia-
betics, 95.45% (21.86%). A complete compilation of
the audiometric results with the mean and standard
error of the mean of each particular frequency, SRT
and speech discrimination scores, as well as the fac-
tors of diabetes and duration of hypertension, is
shown in Tables I and 11. The results of the audio-
metric evaluations between the two is demonstrated
in Figure 1. In each frequency there is a significant
increase in the hearing loss between normotensive
Duck et al.: Pathogenesis of Hearing Loss
1597
 TABLE I.
Raw Data, Mean, and Standard Error of Mean of All Studied Diabetic Hypertensive Patients.
Frequency (Hz) Speech Discrimination Duration (y)

Age AVG Blood Hyper-

Patient (y) Sex 250 500 1000 2000 4000 8000 SRT (dB) (Yo) Pressure tension Diabetes

AT 45 M 25 20 7.5 17.5 60 77.5 45 77.5 154/70 5 17

DM 66 M 30 20 22.5 20 42.5 65 55 96 130l70 20 1
BD 56 M 20 20 20 27.5 30 67.5 10 100 130180 20 3
DH 49 M 25 15 2.5 2.5 15 12.5 42.5 96 140183 10 3
IF 73 F 55 45 42.5 55 97.5 100 22.5 94 170190 10 5
DA 69 F 40 40 32.5 22.5 25 35 5 96 140170 15 1
DC 42 F 10 10 2.5 0 2.5 0 35 98 140190 10 1
VB 86 F 35 30 20 20 32.5 55 22.5 100 170188 24 3
MC 63 F 30 25 17.5 20 27.5 40 22.5 94 140190 15 1
RS 69 F 80 75 50 65 85 100 15 100 152160 2 13
CB 66 F 35 45 47.5 35 37.5 52.5 22.5 94 130170 31 20
AGE 65 F 20 15 15 20 35 40 22.5 100 160/80 20 20
Mean 62.42 33.75 30 21.83 25.42 40.83 53.75 26.67 95.45 146.3ff8.4 15.17 7.33
Standard 3.78 5.57 5.56 5.30 5.67 8.3 9.3 4.48 1.86 4.413.1 2.49 2.35
error

SRT = speech reception threshold; AVG = speech recognition if correct response; BP = blood pressure

and hypertensive insulin-dependent diabetics when
age factors are not taken into account. When both
groups are reevaluated taking age factors into ac-
count and reevaluating the audiometric differences
between the two groups (Table III), a statistically sig-
nificant difference (P < 0.05) is still found between
frequency and group in the higher frequencies (4 to 8
kHz).
Part B
The results of this study compare the three
groups of rats: group I, the LA/N-cp normotensive
nondiabetic rats; group 11, the WKY/N-cp diabetic
normotensive rats; and group 111, the SHR/N-cp
spontaneously hypertensive diabetic rats. Outer
hair cell loss is expressed as mean hair cell loss per
cochlear turn and per quarter of the first and second
turns. The hook and third turn were counted and
analyzed as one unit without being divided into
quarters (Fig. 2). The results after summarizing
from quarters in the first and second turns are pre-
sented in Figure 3. In group I (LALN-cp), the hair
cell loss in the hook averaged 0.075 (20.21); cell loss
in the first turn averaged 14.67 (22.04); in the sec-

TABLE II.
Raw Data, Mean, and Standard Error of Mean of All Studied Diabetic Normotensive Patients.
Frequency (Hz) Speech Discrimination Duration (y)
Age AVG Blood
Patient (y) Sex 250 500 1000 2000 4000 8000 SRT(dB) (W Pressure Diabetes
GC 49 F 10 10 5 5 0 2.5 10 100 136184 22
EL 47 F 20 10 10 7.5 10 37.5 32.5 92 120180 5
CR 49 F 20 10 2.5 2.5 10 5 7.5 100 130190 11
EW 53 F 20 15 2.5 7.5 10 15 5 100 120180 2
ZR 70 F 35 35 25 37.5 50 55 5 100 100170 2
vz 37 F 25 15 0 0 0 0 2.5 100 1 18178 24
DS 45 F 30 25 2.5 2.5 2.5 0 10 100 130180 34
ET 62 F 15 15 5 15 27.5 32.5 5 100 134160 17
LC 45 F 15 10 25 5 17.5 10 17.5 100 120180 16
ZT 13 M 10 10 10 5 12.5 7.5 15 96 1OOl74 1
Mean 47 20 15.5 8.75 8.75 14 16.5 11 98.8 120.8177.6 13.4
Standard 5 2.72 2.77 3.05 3.62 5.05 6.26 2.69 0.8 4.2312.72 3.72
error

SRT = speech reception threshold; AVG = speech recognition if correct response; BP = blood pressure.

Laryngoscope 107: December 1997 Duck et al.: Pathogenesis of Hearing Loss

1598
 50
1
-
0 DIABETICS NORMOTENSIVE
DIABETICS HYPERTENSIVE
T I 50 -
I - HSHR
HWKY
-LA
Y
I d
0
40-
40- rr
I 30-
sg $0- f:c
z 20-

20- 8
g 10-
2
10 -
0-

0 ,- I I I I I I I 1 I I I
250 500 lk 2k 4k 8k H 1-1 1-2 1-3 1-4 11-1 11-2 11-3 11-4 Ill
FREQUENCY IN Hz TURNS

Fig. 1. Comparisons of two groups of diabetics: normotensive and
hypertensive. Depicted on horizontal axis are individually tested fre-
quencies of 250 and 500 Hz and 1, 2, 4, and 8 kHz. Bars express
mean hearing loss for tested frequency; vertical bars designate stan-
dard error of mean.
Fig. 2. Mean outer hair cell loss in absolute values in three groups of
rats: normal (control) LNN-cp rats, normotensive diabetic WKY/N-cp
rats, and hypertensive diabetic SHWN-cp rats. H = hook; I = first
turn; II = second turn; 111 = third turn. Numbers 1 to 4 represent the
quarters of each turn; bars designate standard error of mean.

ond turn, the hair cell loss increased to 19.75
(k2.48); and in the third turn, a decrease to 6 ( d . 1 6 )
was noted (Fig. 3). In group I1 (WKYN-cp),hair cell
loss in the hook averaged 4.69 (21.06); loss in the
first turn increased to 11.54 (kl.9); an increase in
the second turn to 28.62 (27.54) was seen; and a de-
crease in the third turn to 1.08 (20.35) was shown.
In group I11 (SHWN-cp),the hair cell loss was 6.17
(k2.36) in the hook, 175.67 (d33.32) in the first
turn, 123.5 (233.25) in the second turn, and 24.33
(28.81) in turn three. Examples of hair cell loss in
the first (LA/N-cp)and third groups (SHWN-cp)are
demonstrated in Figures 4 and 5, respectively.
There was significantly increased hair cell
loss in group I11 compared with groups I and I1
( P < 0.0001).
There was no statistical difference between the
first and second groups. Closer analysis of the dif-
ferences between the turns and the groups revealed
a significant difference between the first turns
(f= 37.32 and P < 0.0001). However, there was no
significant difference when comparing the hook and
third turns between the groups. There was signifi-
cantly increased hair cell loss (P < 0.05) when com-
paring turns I, 11, and I11 (P = 0.0125) and when
comparing the quarter turns of turns I and I1 ( P =
0.00312). As can be seen in Figure 2, there was no
significant hair cell loss within the hook region, turn
111, and the fourth quarter of turn 11.
This study reveals a highly significant outer
hair cell loss in diabetic rats that were also hyper-
tensive. Most of the damage was localized to the
first and second turns of the cochlea.
DISCUSSION
The results clearly support the hypothesis that
insulin-dependent diabetes, in conjunction with hy-
pertension, has a synergistic effect on high-fre-
quency sensorineural hearing loss. In clinical and
animal studies, respectively, there is a significant
high-frequency sensorineural hearing loss in pa-
tients (even when using correction factors for age)
and a definite adverse effect on the cochlea, result-
ing in hair cell loss.

TABLE 111.
Results of Covariance Statistical Evaluation of Two Groups of Patients Tested With Audiometric Hearing Thresholds.
~~ ~

Source df Sum of Squares Mean Square FValue PValue G-G H-F

Group 1 4062.81 1 4062.811 3.524 0.0759

Age 1 7250.155 7250.155 6.289 0.0214
Subject (group) 19 21904.289 1152.857
Frequency 5 38.766 7.753 0.088 0.9940 0.8954 0.9267
Frequency' group 5 1025.085 206.017 2.325 0.0487 0.1 188 0.1072
Frequency' age 5 385.779 77.156 0.875 0.5010 0.4144 0.4312
Frequency' subject (group) 95 8377.277 88.182
'Significant difference between the groups is a frequency group interaction highlighted in boldface

Laryngoscope 107: December 1997 Duck et al.: Pathogenesis of Hearing Loss

1599
 200

180 0-0 SHR

HWKY
L A L A

8 ti0

I I
H
I
I
I
II Ill
TURNS

Fig. 3. Outer hair cell loss in H and turns I. II, and 111. Horizontal bars
express mean of outer hair cell loss in absolute values; vertical bars Fig. 4. Three rows of outer hair cells without any loss in normal (con-
designate standard error of mean. Normal (control), normotensive trol) rats. The inner row of hair cells is out of focus. White horizontal
diabetic, and hypertensive diabetic rats are depicted. bar represents 25 prn.

A thorough review of the epidemiology and
pathophysiology of the association between diabetes
and hypertension helps to better understand the
multisystem vascular and end-organ pathologic fea-
tures associated with a combination of the two dis-
ease processes thereby affecting the cochlea. Dia-
betes is now the most common cause of end-stage
renal failure in patients between the ages of 25 and
65 years in the United States. As a result, more em-
phasis is being placed on evaluating the role of hy-
pertension in association with diabetes and their ef-
fects on renal function.14 Epidemiologic studies of
the association between diabetes and hypertension
have demonstrated a clinical impression that hy-
pertension occurs with more frequency in the dia-
betic patient than the in general population.15 Esti-
mates of the prevalence of hypertension in diabetics
vary between 10% and 80%, according to different
reports. These reports, however, are questioned be-
cause of their methodology.16
even more common in diabetics before the diagnosis
of diabetes, although there were no significantly
convincing data to support this conclusion. In 1981,
Christlieb et a1.21 demonstrated a significant in-
crease in hypertension in juvenile diabetics as they
aged. Barrett-Connor et a1.22 found that there is a
consistent association between diabetes mellitus
and hypertension, with an adjustment for obesity,
reducing its extent but not its prevalence. In con-
trast, Keen et al.23 cited a significant correlation be-
tween glucose intolerance and hypertension that
was independent of age or obesity.
On histologic study it has been shown that the
changes of diabetic nephropathy occur many years
after the onset of microalbuminuria. Control of hy-
pertension has been shown to eliminate the mi-
croalbuminuria and slows the progression of renal
failure.24 It appears that hypertension associated
with diabetes may be caused by metabolic factors

In 1954, Hamilton et al.17 observed that hyper-

tension was not significantly higher in diabetics
than in nondiabetics, except in the 70- to 79-year-old
age group. Keen et a1.18 found no systematic blood
pressure difference between diabetics and controls.
The study on London civil servants showed signifi-
cantly higher blood pressure in diabetics than in
controls.19 This study demonstrated that there was
a correlation between systemic blood pressure, dias-
tolic blood pressure, and glucose, which was inde-
pendent of body mass in men more than 40 years of
age. A 1967 study compared 662 employees of the
Dupont Company with matched nondiabetics (for
age, sex, and employment status) and found that the
prevalence of hypertension (systolic blood pressure
>150 mm Hg and diastolic blood pressure >94 mm
Hg) was higher in the diabetic group.20 Hyperten- Fig. 5. Three rows of outer hair cells with several hair cells (black ar-
sion increased as each category of body weight in- rowheads) in hypertensive diabetic rats. The inner hair cells are out
creased. It was also noted that hypertension was of focus. White horizontal bar represents 25 fim.

Laryngoscope 107: December 1997 Duck et al.: Pathogenesis of Hearing Loss

1600
other than renal factors. Metabolic alterations found
in the adipocytes of diabetic rats include the eleva-
tion of cellular cyclic adenosine monophosphate
(AMP),both basal and norepinephrine stimulated,
and higher beta-adrenergic receptor density.25226 It
has been found that, because of an increased sensi-
tivity to norepinephrine and angiotensin 11, much
less norepinephrine and angiotensin I1 is needed in
diabetics to achieve the same pressor effects in nor-
moglycemic patients.27 Another possible contribut-
ing metabolic factor causing hypertension in stable
nonazotemic diabetics is an increase in sodium ex-
change.28 In the hypertensive diabetic, systolic blood
pressure correlates with sodium exchange, resulting
in hypertension.29 A third factor could lie in genetic
HLA type where the risk of developing insulin-de-
pendent diabetes mellitus for the HLA-identical sib-
ling of an insulin-dependent diabetic patient is ex-
tremely high. Also, siblings with shared HLA
specificities are at a greater risk of developing in-
sulin-dependent diabetes mellitus.30
In benign hypertension there are arteriosclerotic
changes with multiplication of the internal elastic
lamina and thickening of the intima. In arterioles
there is hypertrophy of the media and subsequent
hyalinosis occurs. When combined with diabetes, the
arterial changes are frequently demonstrated in the
preglomerular afferent arterioles. The lumen of the
blood vessels is narrowed, and muscles of the media
are atrophic. This narrowing causes a decrease in
glomerular perfusion and a decreased ability of the
muscle cells to react. The hyalinization in the pre-
glomerular arterioles forms a barrier between the
jwtaglomerular cells and the lumen of the arteriole,
which prevents renin release into circulation, re-
sulting in faulty coordination between the activity of
the juxtaglomerular apparatus and systemic arter-
ial blood pressure.31-33 Studies in diabetic hyperten-
sives have shown that arteriosclerosis was nearly
always present in hypertensive patients.34135 In ad-
dition, arteriosclerosis was also found in a remark-
able percentage of normotensive diabetics. These
studies illustrate that when considering the patho-
physiology of arteriosclerosis in diabetics, hyperten-
sion appears to be one of many contributing fac-
tors .36
Hypertension has been shown to potentiate
damages caused by diabetes. For example, diabetic
microangiopathy of renal cortical capillaries, char-
acterized by basement membrane thickening, was
intensified in the setting of concomitant hyperten-
sion.37 Diabetic nephropathy has been shown to be
accelerated and associated with increased glomeru-
lar basement membrane thickening and proteinuria
in hypertensive diabetic rats compared with nor-
motensive diabetic controls. Hypertension was
found to have an additive effect on glomerular base-
ment membrane thickening, being increased in hy-
pertensive diabetic animals in comparison with
Laryngoscope 107: December 1997
their normotensive counterparts. The finding of a
significant increase in glomerular basement mem-
brane thickening in hypertensive as compared with
normotensive rats suggests that systemic blood
pressure may be an independent determinant of
glomerular basement membrane thickness.38~39
Early diabetic glomerular lesions reveal both
basement membrane thickening and augmentation
of the mesangial matrix.40 Glomerular capillaries
are not normally surrounded circumferentially by
the basement membrane. It can be seen in sections
that the basement membrane of one capillary is
fixed to the mesangium in two places. In a normal
glomerulus these sites are side by side. With mesan-
gial broadening the fixation sites are crowded out,
but there is no reduction of the filtration surface or
marked decrease of the capillary lumen. Therefore
in the early and intermediate stages of glomeru-
losclerosis there is not very much reduction of the
filtration surface and the capillary lumen. Ulti-
mately this process cannot continue, and large nod-
ules form and are flattened, occluding capillaries at
the periphery. The nodules can then be in contact
with and adhere to Bowman’s capsule. On histologic
study, type I diabetics with clinical nephropathy and
hypertension were found to have more frequent ex-
udative lesions, subcapsular fibrosis, and sclerotic
glomeruli. There is also a significant diminution
with regard to open capillaries, endothelial cells,
and Bowman’s urinary space.41
Hostetter et al.42 have conceived the hypothesis
that glomerular hyperfiltration initiates and is re-
sponsible for the progression of diabetic nephropa-
thy. The hyperfiltration is stimulated by features of
the diabetic state (i.e., extracellular fluid volume
expansion that is due to hyperglycemia, renal hy-
pertrophy, increased growth hormone, and glucagon
secretion) or changes in vascular response to vary-
ing amounts of vasoactive hormones. The hyperfil-
tration increases transglomerular protein filtration,
leading to albuminuria and mesangial deposition of
circulating proteins, ultimately resulting in mesan-
gial expansion and glomerulosclerosis. The initial
loss of functioning nephrons exerts a positive feed-
back stimulus to compensatory hyperfiltration in
less affected surviving glomeruli, thus contributing
to their eventual destruction.42
Hypertension often becomes evident during
the development of general glomerulosclerotic le-
sions without response t o diabetes mellitus. Per-
haps this is the result of (or a n indicator of) ad-
vanced glomerular lesions. However, as previously
mentioned, this does not explain the fact that suc-
cessful treatment of hypertension in diabetics with
nephropathy can delay renal insufficiency for
years.43 This forms the view that hypertension plays
an important role in the development of renal
changes in diabetics, accelerating the progression to
renal insufficiency. Hypertension frequently ap-
Duck et al.: Pathogenesis of Hearing Loss
1601
pears only several years after the start of clinical
nephropathy, and thereafter, renal function deterio-
rates at a predictably rapid rate. In diabetics, there
is a slow increase in basement membrane thicken-
ing and matrix augmentation that needs many
years to cause a noticeable diminution of the filtra-
tion surface and total function of glomerular capil-
lary lumina. Only in this advanced stage of
glomerulosclerosis do further glomerular changes
occur (i.e., arteriosclerosis and pericapsular and in-
terstitial fibrosis). These changes play a significant
role that can lead to renoparenchymatous hyperten-
sion. Hypertensive changes may be superimposed
on the diabetic glomerular changes, and further de-
terioration of renal function occurs.44
Numerous studies have documented hearing
losses in diabetics with abnormalities in audiometry
and brainstem auditory evoked responses.2,45-50 Fer-
rer et a1.45 concluded that type I diabetes mellitus
can cause mild sensorineural hearing impairment
that correlates with age and the duration of disease.
Kurien et a1.2 have found that the diabetic popula-
tion has a poorer hearing threshold than the nondi-
abetic population. All age groups of diabetics in
their study showed a high-frequency hearing loss as
compared with the control population. Although
they found no relationship between the duration of
diabetes and hearing loss, poorly controlled and
complicated diabetics have a significant high-fre-
quency hearing loss as compared with patients with
well-controlled and uncomplicated diabetes.2
Long-term insulin-dependent diabetes mellitus
also has an effect on auditory brainstem evoked re-
sponses. Even in the absence of symptoms consis-
tent with diabetic neuropathy, 25% to 40% of dia-
betic patients have been found to have retrocochlear
impairment, as evidenced by the absence of wave I
in auditory brainstem evoked response record-
ings.4"47 This is important for detecting desynchro-
nization of the auditory response that is indicative
of neuropathy in the acoustic nerve.46 Axelsson and
FagerberPH studied three separate age groups be-
tween ages 16 and 50 years (16 to 30 years, 30 to 39
years, and 40 to 50 years) and found that a definite
degree of hearing loss of retrocochlear origin could
be demonstrated in all three age groups (i.e., in
speech, Bekesy, and directional audiometry).48 In
another study, by Huang et a1.,49 Bekesy audiome-
try, impedance audiometry with stapedius reflex
test, speech discrimination and the recording of
electrocochleography, and brainstem auditory
evoked potentials were performed on diabetic pa-
tients and controls and it was found that the dia-
betics were abnormal. The results demonstrated
that the average pure-tone thresholds, the subjec-
tive click threshold, and the action potential reac-
tive threshold, were all significantly elevated in di-
abetics. The average speech discrimination score
was also decreased in diabetics. When compared
Laryngoscope 107: December 1997
1602
with controls, the hearing impairment in diabetics
was characterized by high-frequency sensorineural
hearing loss (after correction for age and sex). Both
cochlear and retrocochlear disease was found to ex-
ist in diabetics with hearing impairment.49
The clinical changes of inner-ear function in di-
abetics seem to become evident as hearing loss of
the perception type with audiometric tracings being
similar to those found in presbycusis (age-condi-
tioned hearing loss).50 PAS-positive thickened capil-
laries were found in the cochlear stria vascularis
that were consistent with diabetic changes found in
other organs (i.e., the kidney and eye.)50 The thick-
ening could potentially affect the function of the
stria vascularis, ultimately resulting in decreased
hearing.50 It is hypothesized that the secretion of en-
dolymph and the building up of the electric potential
necessary for the cochlear microphonic are affected
by the decreased blood f l ~ w . ~ ODepending
-~~ on the
intensity and duration of these changes, it is possi-
ble that the sensory epithelium of the organ of Corti
may ultimately be affected.50
Wackym and Linthicum54 studied the effects of
both longitudinal and radial blood flow impairment
of the cochlea and suggested that diabetic sen-
sorineural hearing loss results from microangio-
pathic involvement of the endolymphatic sac and/or
basilar membrane vessels. An impairment of longitu-
dinal blood flow affects the reabsorption of en-
dolymph and the disposal of high molecular weight
waste products and debris by the endolymphatic
sac.54 Removal of higher molecular weight waste
products from the endolymph may occur by macro-
phagic activity of the endolymphatic sac55,56 or by
free-floating cells that may phagocytize these sub-
stances and then migrate through the epithelium
into the vascular bed.56,57 The effect of narrowed or
absent vascular endothelial tight junctional pores
would limit macrophage or lymphocyte migration
into or out of the endolymphatic sac. These changes
have been illustrated in diabetic microangiopa-
thy.58359 When the ability to effectively remove higher
molecular weight waste products has been jeopar-
dized, a toxic effect may be exerted on the hair cells,
thereby resulting in sensorineural hearing l0ss.64
The impairment of radial flow may theoretically
cause hair cell damage from microangiopathic
changes in the stria vascularis. This is based on the
inability of the thickened strial vasculature to ade-
quately supply the hair cells with the necessary nu-
trients to maintain the ionic character of the en-
dolymph.55 However, this is unlikely to be the
primary mechanism for diabetic sensorineural hear-
ing loss, because diabetics in this study who had
strial microangiopathy did not have a hearing loss.54
Smith et a1.60 have quantitatively demon-
strated basement membrane thickening consistent
with diabetic microangiopathy in the inner ear of
Duck et al.: Pathogenesis of Hearing Loss
insulin-dependent diabetic rats. This finding sug-
gests that inner-ear microangiopathy should be
considered as a possible cause of hearing loss in in-
sulin-dependent diabetic patients. It was also
demonstrated that noise exposure did not signifi-
cantly change basement membrane thickness in ei-
ther diabetic or control animals.60 Spontaneously
diabetic rats have also been shown to have a more
pronounced cochlear hair cell loss when compared
with nondiabetic controls. This loss was found in
the midportion of the cochlea.61 Raynor et al.4 have
also shown that consumption of a high-galactose
diet induces diabetic-like changes in the inner ear.
Hair cell damage in the Raynor study also occurred
primarily in the midportion of the cochlea, with
only a small amount of damage found in the hook
region.4
It has been shown in animal studies that there
is an increase in the permeability of the blood-brain
barrier in adrenaline-induced hypertension in dia-
betic rats.@It has also been shown that early-stage
general vascular disease is exacerbated (i.e., the el-
evation of the extracellular matrix compared with
smooth muscle in arterial walls, marked suben-
dothelial invasion of blood vessels with macrophage
cell types) with combined diabetes and hyperten-
sion."j
It has also been suggested that insulin-depen-
dent diabetes mellitus may increase the hair cell
loss caused by noise overstimulation. There was
found to be an increase in the number of hair cells
lost in insulin-dependent, noise-exposed rats when
compared with rats who had insulin-dependent dia-
betes alone.10
In a comparative study of normotensive and
spontaneously hypertensive rats by Tachibana et
al.6 electrocochleographic studies showed that the
function of the cochlea in the spontaneously hyper-
tensive rats declined with increasing age to a
greater extent than that of normotensive rats. The
primary site of cochlear involvement, as docu-
mented by electron microscopy, was the vascular
strip followed by the organ of Corti.5
Borg and Viberglo state that the increased sus-
ceptibility of spontaneously hypertensive rats to
noise-induced hearing loss may be attributed to the
spontaneous loss of hair cells. Normal, age-related
loss may also explain why old, spontaneously hyper-
tensive rats are somewhat more susceptible to noise
than young rats. Also, the increased susceptibility of
young, spontaneously hypertensive rats may be due
to a smaller number of hair cells at the start of noise
exposure or attributable t o the fact that remaining
hair cells are more susceptible to noise damage. In
older, spontaneously hypertensive rats there may be
a smaller number of hair cells at the start of the
noise exposure. Regardless of the age of the rats, the
pre-noise exposure hair cell abnormality caused by
Laryngoscope 107: December 1997
vascular or other factors is the most likely explana-
tion for the noise susceptibility of spontaneously hy-
pertensive rats.10
Gates et al.3 have shown that in hypertensive
men the age-adjusted high-frequency hearing loss
was significantly greater (better ear, 24.7 vs 22.6
dB hearing loss [HLI, P = 0.01; worse ear, 67.9 vs
59.3 dB HL, P = 0.01) than in normotensive men.
In hypertensive women, age-adjusted low-fre-
quency hearing was significantly worse than in
nonhypertensive women (better ear, 20 vs 18 dB
HL, P = 0.004).3
The combined results of this study support our
hypothesis that diabetic end-organ damage of the
cochlea is intensified in the setting of concomitant
hypertension. This was verified in animal models
by the significant elevations of mean hair cell loss
noted throughout the cochlea in the hypertensive
and diabetic SHR/N-cp rat group when compared
with the diabetic normotensive WKYiN-cp and
nondiabetic normotensive LA/N-cp rat groups.
Clinical studies also show a statistically significant
hearing loss in the higher frequencies in insulin-
dependent diabetic patients with hypertension
when compared with patients with insulin-depen-
dent diabetes alone. The results of this study thus
indicate that either hypertension or a synergistic
effect between hypertension and diabetes has a
more pronounced effect on hair cell loss in rats
than does diabetes alone. The audiometric findings
presented also substantiate the clinical impression
that there is a potentiation of high-frequency hear-
ing loss in patients with hypertension and insulin-
dependent diabetes as opposed to insulin-depen-
dent diabetes alone.
CONCLUSION
The previously described findings support the
hypothesis that diabetic end-organ damage of the
cochlea was intensified in the setting of concomitant
hypertension. This was verified in animal studies by
the significant elevations of mean hair cell loss
noted throughout the cochlea in the hypertensive
and diabetic SHR/N-cp rat group when compared
with the diabetic normotensive WKYN-cp and non-
diabetic normotensive LA/N-cp rat groups. The sta-
tistically significant differences in hearing loss in
insulin-dependent diabetic and hypertensive pa-
tients when compared with the hearing of insulin-
dependent diabetics who are normotensive also sub-
stantiate our hypothesis. It has also been shown in
this study that insulin-dependent diabetic patients
without hypertension have no significant hearing
loss. These results support the hypothesis that hy-
pertension is the essential factor, in conjunction
with insulin-dependent diabetes mellitus, responsi-
ble for high-frequency sensorineural hearing loss in
patients and hair cell loss in laboratory animals.
Duck et al.: Pathogenesis of Hearing Loss
1603
BIBLIOGRAPHY
I . Isselbacher KJ, ed. Harrison’s Principles of Internal Medicine.
New York: McGraw-Hill; 1994:1117,1980.
2. Kurien M, Thomas K, Bhanu TS. Hearing threshold in patients
with diabetes mellitus. J Luryngol Otol 1989; 103: 164-8.
3. Gates G, Cobb JL, DAgostino RB, e t al. The relation of hear-
ing in the elderly to the presence of cardiovascular disease
and cardiovascular risk factors. Arch Otolaryngol Head
Neck Surg 1993;119:156-61.
4. Raynor EM, Robison WG, Garrett CG, e t al. Consumption of a
high-galactose diet induces diabetic-like changes in the in-
ner ear. Otolaryngol Head Neck Surg 1995;113:748-54.
5. Tachibana M, Yamamichi I, Nakae S, et al. The site of involve-
ment of hypertension within the cochlea. Acta Otolaryngol
(Stockhi 1984;97:257-65.
6. Jordao AMD. Consideration sur un cas du diabete. Un Med
Paris 1857;11:447.
7. Jannulis GE, Delijannis G. Diabetes und Gicht als Ursache
van Kochlearisschadigung. Monatsschra Ohrenheilkund
Laringol Rino 1936;70:1504-8.
8. Makishima K. Arteriolar sclerosis as a cause of presbycusis.
Otolaryngology 1978;86:322-6.
9. Axelsson A, Borg E, Hornstrand C. Noise effects on the
cochlear vasculature in normotensive and spontaneously
hypertensive rats. Acta Otolaryngol 1983;96:215-25.
10. Borg E, Viberg A. Age-related hair cell loss in spontaneously
hypertensive and normotensive rats. Hearing Res 1987;30:
111-8.
1 1 . Borg E. Noise-induced hearing loss in normotensive and spon-
taneously hypertensive rats. Hearing Res 1982;8:117-30.
12. Michaelis IV, Carswell N, Hansen CT, et al. A new genetic
model of noninsulin-dependent diabetes and hypertension:
the spontaneous hypertensiveiNIH corpulent rat. In:
Shafrir E, Renold AE,eds. Lessons from Animal Diabetes.
London: John Libbey; 1988:257-63 ( ~ 0 1 2 ) .
13. National Institutes of Health. Office for Protection from Re-
search Risks: Public Health Service Policy on Humane Care
and Use of Laboratory Animals. Bethesda, MD: Office for
Protection from Research Risks, National Institutes of
Health; 1986.
14. Rodby RA, Schwartz MM. Proteinuria, hematuria, hyperten-
sion, and decreased renal function in a patient with dia-
betes for 9 years (clinical conference). Am J Kidney Dis
1992;20:658-67.
15. Fuller JH. Epidemiology of hypertension associated with dia-
betes mellitus. Hypertension 1985;7(suppl 1):3-7.
16. Drury PL. Diabetes and arterial hypertension. Diabetologia
1983;24:1-9.
17. Hamilton M, Pickering GW, Robert JAF, Sowry GSC. The aeti-
ology of essential hypertension. Part I. The arterial pressure
in the general population. Clin Sci 1954;17:11-54.
18. Keen H, Track NS, Sowry GSC. Arterial pressure in clinically
apparent diabetics. Diabetes Metab Rev 1975;1:15944.
19. Jarrett RJ, Keen H, McCartney M, et al. Glucose tolerance and
blood pressure in two population samples: their relation to
diabetes and hypertension. Znt J Epidemiol 1978;7:15-24.
20. Pel1 S, DAlonzo CA. Some aspects of hypertension in diabetes
mellitus. JAMA 1967;202:104-10.
21. Christlieb AR, Warren JH, Krowelski AS, et al. Hypertension,
the major risk in juvenile-onset insulin-dependent diabet-
ics. Diabetes 1981;3O(suppl2):90-6.
22. Barrett-Connor E, Criqui MH, Klauber MR, Holdbrook M. Di-
abetes and hypertension in a community of older adults. Am
J Epidemiol 1981;113:276-84.
23. Keen H, Chloveerakis C, Fuller JH, Jarrett FLJ. The concomi-
tants of raised blood sugar: studies in the newly detected
hyperglycaemics. Part 11. Urinary albumin excretion, blood
pressure and their relation to blood sugar levels. Guy’s Hosp
Report 1969;118:247-54.
24. Mauer SM, Steffes MW, Azar S, et al. The effects of Goldblatt
hypertension on development of the glomerular lesion of di-
abetes mellitus in the rat. Diabetes 1978;27:738-44.
Laryngoscope 107: December 1997
1604
25. Chiappe de Cingolani GE. Cyclic AMP and glucose utilization by
diabetic rat fat cells. Arch Znt Physiol Biochem 1983; 91:l-8.
26. Chiappe de Cingolani GE. Beta-adrenergic receptors and
adenylate cyclase activity in diabetic rat fat cells. Diabetes
1986;35:1229-32.
27. Weidmann P, Beretta-F’iccoli C, Trost BN. Pressor factors and
responsiveness in hypertension accompanying diabetes
mellitus. Hypertension 1985;7(suppl 2):33-42.
28. Beretta-Piccoli C, Weidmann P. Body-sodium blood volume
state in nonazotemic diabetic mellitus. Miner Electrolyte
Metab 1982;7:3642.
29. Weidmann P, Beretta-F’iccoli C, Keusch G, et al. Sodium-vol-
ume factor, cardiovascular reactivity and hypotension mech-
anisms of diuretic therapy i n hypertension associated with
diabetes mellitus. Am J Med 1979;67:779-84.
30. Gorsuch AW, Spencer KM, Lister J, et al. Can future diabetes
be predicted? A study in families of affected children. Dia-
betes 1982;31:862-6.
31. Christleib AR. The hypertension of diabetes. Diabetes Care
1982;5:50-8.
32. Ditscherlein G. Renal histology in hypertensive diabetics. In:
Gries FA, Weidmann P, eds. Diabetes and Hypertension.
Berlin, Heidelberg: Springer Verlag;1988:39.
33. Bader H, Meyer DS. The size of the juxtaglomular apparatus
in diabetic glomerulosclerosis and its correlation with arte-
riosclerosis and arterial hypertension: a morphometric light
microscopic study on human renal biopsies. Clin Nephrol
1977;8:308-11.
34. Ditscherlein G. Renal histopathology in hypertensive diabetic
patients. Hypertension 1985;7(suppl):29-32.
35. Ditscherlein G. Renal histopathology in hypertensive diabetic
patients. Hypertension 1985;7(suppl2):29-32.
36. Ditscherlein G. Renal histology in hypertensive diabetics. In:
Gries FA, Weidman P, eds. Diabetes and Hypertension.
Berlin, Heidelberg: Springer Verlag, 1988:40.
37. Junker U, Jaggi C, Bestetti G, Rossi GL. Basement membrane
of hypothalamus and cortex capillaries from normotensive
and spontaneously hypertensive rats with streptomycin-in-
duced diabetes. Acta Neuropathol 1985;65:202-8.
38. Cooper ME, Allen TJ, O’Brien RC, et al. Effects of genetic hy-
pertension on diabetic nephropathy in the rat: functional and
structural characteristics. J Hypertension 1988;6: 1009-16.
39. Cooper ME, Allen TJ, Jerums G, Doyle AE.Accelerated pro-
gression of the diabetic nephropathy in the spontaneously
hypertensive streptozotocin diabetic rat. Clin Exp Pharma-
col Physiol 1986;13:655-62.
40. Osterby R. Early phases in the development of diabetic
glomerulopathy: a quantitative electron microscopic study.
Acta Med Scand Suppl 1975;574:1-82.
41. Ditscherlein G. Renal histology in hypertensive diabetics. In:
Gries FA, Weidmann P, eds. Diabetes and Hypertension.
Berlin, Heidelberg: Springer Verlag; 1988:38.
42. Hostetter TH, Helmut GR, Brenner BM. The case for in-
trarenal hypertension in the initiation and progression of
diabetic and other glomerulopathies. Am J Med 1982;72:
375-80.
43. Mogensen CE. Antihypertensive treatment inhibiting the pro-
gression of diabetic nephropathy. Acta Endocrinol (Copenh)
1980;94 (Suppl 238):l-82.
44. Ditscherlein G. Renal histology in Hypertensive diabetics. In:
Gries FA, Weidmann P, eds. Diabetes and Hypertension.
Berlin, Heidelberg: Springer Verlag;1988:41.
45. Ferrer J, Biurrun 0, Lorente J, et al. Auditory function in
young patients with type 1 diabetes mellitus. Diabetes Res
Clin Pract 1991;11:17-22.
46. Martini A, Comacchio F, Magnavita V. Auditory brainstem
and middle latency evoked responses in the clinical evalua-
tion of diabetes. Diabet Med 1991;8:S74-7.
47. Parving A, Elberling C, Balle V. Hearing disorders in patients
with insulin-dependent diabetes mellitus. Audiology 1990;
29: 113-21.
48. Axelsson A, Fagerberg S-E. Auditory function in diabetics. Acta
Otolaryngol 1968;66:49-64.
Duck et at.: Pathogenesis of Hearing Loss
49. Huang Y, Pan CY, Gu R, et al. Hearing impairment in diabet-
ics. Chin Med J 1992;105:44-8.
50. Jorgensen MB. The inner ear in diabetes mellitus. Arch Oto-
laryngol 196 1;74:373-8 1.
51. Dohlmann G. Modern view on vestibular physiology. J Laryn-
go1 1959;74:154-61.
52. Omerod FC. The physiology of the endolymph. J Laryngol
1960;74:659-67.
53. Mizukoshi 0. Some metabolic reactions in the cochlea. Jpn J
Otol 1 9 6 0 ; 6 3 ( ~ ~ ~ ~ 1 2 ) : 9 - 1 0 .
54. Wackym PA, Linthicum FH. Diabetes mellitus and hearing
loss: clinical and histopathologic relationships. Am J Otol
1986;7:17&82.
55. Lundquist PG. Aspects on endolymphatic sac morphology and
function. Arch Otolaryngol 1976;212:231-40.
56. Lundquist PG, Kimura R, Wersall J. Experiments in endolym-
phatic circulation.Acta Otoluryngol (Stockhi 1963;188:198-201.
57. Rask-Andersen H, Stahle J. Immunodefense of the inner ear?
Lymphocyte-macrophage interaction in the endolymphatic
sac. Acta Otolaryngol (Stockh) 1980;89:283-94.
58. Williamson JR, Kilo C. Current status of capillary basement
disease in diabetes mellitus. Diabetes 1977;26:66-75.
59. Bloodworth JMB. Diabetic microangiopathy Diabetes 1963;
12:99-114.
60. Smith TL, Raynor EM, Prazma J, et al. Insulin-dependent dia-
betic microangiopathy in the inner ear. Laryngoscope 1995;
105:23640.
Laryngoscope 107: December 1997
61. Triana RJ,Suits GW, Garrison S, et al. Inner ear damage sec-
ondary to diabetes mellitus. Arch Otolaryngol Head Neck
Surg 1991;117:635-40.
62. Oztas B, Kucuk M. Influence of acute arterial hypertension on
blood- brain barrier permeability in streptozotocin-induced
diabetic rats. Neurosci Lett 1995;188:53-6.
63. Todd ME, Song MY, McNeill JH. Coexistence of diabetes and
hypertension results in unique structural alterations in the
renal artery in rats beyond that found with diabetes alone.
Diabetes Clin Pract 1993;19:115-26.
EDITORIAL FOOTNOTE
During the review process several issues were
raised concerning this very interesting report. Read-
ers should note that, with regard to the analysis of
results, the number of subjects is quite small in
terms of reaching meaningful conclusions. All of the
hypertensive patients had been under treatment for
2 or more years (average, 15 years) before the study
began, and their blood pressure was controlled at
the time of the study. The reliability of the analysis
of covariance is not clear in view of the age differ-
ence between the two groups.
Duck et al.: Pathogenesis of Hearing Loss
1605