Download Association between bone mineral density and hearing loss in

The Laryngoscope
C 2012 The American Laryngological,
V
Rhinological and Otological Society, Inc.
Association Between Bone Mineral Density and Hearing Loss in
Osteogenesis Imperfecta
Freya K. R. Swinnen, MS; Els M. R. De Leenheer, MD, PhD; Stefan Goemaere, MD;
Cor W. R. J. Cremers, MD, PhD; Paul J. Coucke, PhD; Ingeborg J. M. Dhooge, MD, PhD
Objectives/Hypothesis: Osteogenesis imperfecta (OI) is a heritable connective tissue disorder, predominantly characterized by bone fragility. In half of the patients, progressive hearing loss develops, which is associated with abnormal bony
changes involving the middle ear ossicles and stapes footplate. In the present study, we investigated whether the development of hearing loss in OI may be related to the overall aberrant bone quality.
Study Design: Observational study.
Methods: Following audiologic evaluation, 56 adult OI patients were classified as presenting normal hearing or conductive/mixed or pure sensorineural hearing loss. Areal bone mineral density (BMD) (aBMD) was measured using lumbar spine
(LS) and whole body (WB) dual X-ray absorptiometry. By means of peripheral computed tomography, volumetric BMD
(vBMD) and morphometric bone parameters were determined at distal and proximal radius, providing separate results for
trabecular and cortical bone. The obtained bone parameters were compared between normal-hearing OI patients and those
with either conductive/mixed or pure sensorineural hearing loss.
Results: Z scores demonstrated decreased LS aBMD, WB aBMD, and trabecular vBMD in OI adults compared to the
healthy population. Patients with conductive/mixed hearing loss had lower trabecular vBMD compared to those with normal
hearing or pure sensorineural loss at both whole-group and between-relatives comparisons.
Conclusions: It is hypothesized that OI patients with lower BMD might be more susceptible to accumulating microfractures, which may interfere with the bone remodeling inhibition pathways in the temporal bone and, therefore, contribute to
stapes footplate fixation and a conductive hearing loss component.
Key Words: Osteogenesis imperfecta, bone mineral density, hearing loss, type I collagen.
Level of Evidence: 2c.
Laryngoscope, 122:401–408, 2012
INTRODUCTION
Osteogenesis Imperfecta (OI) is an uncommon,
hereditary connective tissue disorder that may involve
the musculoskeletal system, skin, sclerae, ear, and cardiovascular system. OI is characterized by a large
phenotypic as well as genotypic heterogeneity. The most
widely used classification of OI was introduced by
Sillence et al.,1 who, on the basis of clinical and radiographic features, distinguished four major types
covering a mild (type I), a lethal (type II), a severe (type
III), and a moderate (type IV) phenotype. More recently,
several attempts have been made to further differentiate
this classification by adding types V, VI, VII, and VIII.2
Still, addition of these uncommon OI types is not supFrom the Department of Otorhinolaryngology (F.K.R.S., E.M.R.D.L.,
Unit for Osteoporosis and Metabolic Bone Diseases (S.G.), and
Center for Medical Genetics (P.J.C.), Ghent University Hospital, Ghent,
Belgium; and FC Donders Institute for Neurosciences (C.W.R.J.C.), Radboud
University Nijmegen Medical Centre, Department of Otorhinolaryngology,
Nijmegen, The Netherlands..
Editor’s Note: This Manuscript was accepted for publication October 3, 2011.
Freya K. R. Swinnen, MS, holds a PhD fellowship of the Research
Foundation Flanders (FWO Vlaanderen), Belgium. The authors have no
other funding, financial relationships, or conflicts of interest to disclose.
Send correspondence to Freya K. R. Swinnen, MS, Department of
Otorhinolaryngology (1P1), Ghent University Hospital, De Pintelaan
185, B-9000 Ghent, Belgium. E-mail: [email protected]
I.J.M.D.),
DOI: 10.1002/lary.22408
Laryngoscope 122: February 2012
ported by all clinicians, because, unlike the initial
classification, their differentiation is based on moleculargenetic rather than clinical grounds.3
The clinical expression may partially be predicted
on the basis of the underlying OI genotype. In 90% of
cases, the genotype is characterized by an autosomaldominant mutation in the COL1A1 or COL1A2 gene,
encoding for two a1(I) and one a2(I) chains of type I collagen, the principal component of the organic matrix of
bone. Mutations in COL1A1 or COL1A2 may hinder the
type I collagen synthesis in an either quantitative or
qualitative way. A mutation causing a quantitative defect
results into the underproduction of a normal type I collagen molecule, whereas a qualitative impairment refers to
the formation and incorporation of structural abnormal
type I collagen.4 The clinical severity of the disease is
partially determined by the whether the nature of the
type I collagen defect is quantitative or qualitative, which
of the chains is affected, and at which position within the
triple helical portion of the chain the mutation occurs.3,4
Although a direct relationship between the collagen
gene mutations and the clinical phenotype remains
largely unclear, the collagen defect causes brittle bones,
the hallmark feature of OI, of which the severity may be
assessed clinically. From a mechanical viewpoint, two
types of bone may be distinguished. Trabecular bone is
prominent in the vertebral bodies and at the end of long
Swinnen et al.: BMD and Hearing Loss in OI
401
bones, whereas cortical bone tissue is mainly found in
the shafts of the long bones. Bone fragility in OI is
determined by decreased bone quantity and altered bone
quality. The amount of bone is commonly evaluated by
means of dual X-ray absorptiometry (DXA) or peripheral
quantitative computed tomography (CT) (pQCT), which
measure bone mineral content and bone mineral density
(BMD), at well-defined sites of the human body. Both
methods have demonstrated a reduced BMD in OI adults
and children.3,5,6 In the OI population, three-dimensional
pQCT measurements of volumetric BMD (vBMD) have
been suggested to be preferable to two-dimensional DXAmeasurements of areal BMD (aBMD).6 PQCT allows
separate measurements of trabecular and cortical compartments and offers insight into geometrical properties
of bone. Fracture rate in OI patients often decreases after
puberty, when the skeleton has attained its adult size.
However, BMD often remains below normal values in
adult life, particularly for trabecular bone.5,6
At the beginning of adulthood, when an improvement in the skeletal condition is often experienced,
another problem may arise. Approximately half of the
OI patients develop a hearing loss. The hearing impairment usually appears as a conductive hearing loss in the
second to fourth decade of life and most often progresses
to a mixed hearing loss thereafter.7,8 A minority of
patients develop a pure sensorineural hearing loss
(SNHL).8 The conductive hearing loss component is
often caused by otosclerosis-like lesions, inducing a
thickened and fixed stapes footplate. However, in comparison with classic otosclerosis, the conductive hearing
impairment in OI is generally characterized by an earlier onset and more often progresses to a mixed hearing
loss, which is associated with a more extensive degree of
pericochlear demineralization on CT.9 In addition, OIrelated conductive hearing loss may be due to ossicular
discontinuity, in particular fractured or atrophic stapes
crura replaced by fibrous threads, whether or not it is in
combination with oval window obliteration.7 The SNHL
component has been suggested to result from several
pathologic conditions with regard to the inner ear structures, such as encroachment of abnormal bone on the
cochlea, hemorrhage into the labyrinth, and atrophy of
the stria vascularis or the cochlear hair cells.10
Like many phenotypic features of OI, the hearing
loss characteristics are heterogeneous, and even relatives with an identical underlying COL1A1 or COL1A2
mutation may radically differ in hearing ability or hearing loss type.11 This variability remains largely unclear.
Because the underlying pathology causing hearing
impairment in OI is apparently tied up with bony
changes in the ossicular chain and the cochlear capsule,
a possible association between hearing performance and
generalized bone disease was investigated.
MATERIALS AND METHODS
Subjects
Patients were recruited at the departments of Otorhinolaryngology, Medical Genetics, Orthopedics & Fysiotherapy, and
Endocrinology & Rheumatology of the Ghent University Hospi-
Laryngoscope 122: February 2012
402
tal (UZ Ghent) and by means of an advertisement in the
periodical of the Belgian patients’ association for OI (Zelfhulp
Osteogenesis Imperfecta). Approval of the study was provided
by the local ethical committee, and, prior to participation, all
patients signed the informed consent form in accordance with
the Declaration of Helsinki. In all patients the diagnosis of OI
was clinically confirmed. The OI type according to the Sillence
classification (I–IV) was assessed by a geneticist.
History
To preclude causes for hearing loss other than OI, medical
and otologic history were documented. Patients were asked for
noise exposition, ear surgery, head injury, and drugs intake, in
addition to fracture history, mobility and sports activities, physical complaints, bisphosphonate (BP) treatment, and family
history of OI.
Audiologic Examination
After micro-otoscopy to evaluate the intactness of the eardrum and ventilation of the middle ear, bilateral tympanograms
were obtained using an 85-dB SPL 226-Hz probe tone, and ipsiand contralateral stapedius reflex thresholds were measured
using 0.5, 1.0, 2.0 kHz and broadband noise stimuli (TympStar;
Grason Stadler Inc., Eden Prairie, MN).
Pure-tone audiometry was performed in a double-walled
sound-attenuated room, applying the modified Hughson-Westlake
method to bilaterally determine air conduction (AC) thresholds
expressed in decibel hearing level (dB HL) at octave frequencies
0.25 to 8.0 kHz and at half-octave frequencies 3.0 and 6.0 kHz, as
well as bone conduction (BC) thresholds at octave frequencies
0.25 to 4.0 kHz and half-octave frequency 3.0 kHz (AC 40 Clinical
Audiometer; Interacoustics, Assens, Denmark). At frequencies
from 0.25 to 4.0 kHz, the air-bone gap (ABG) was calculated by
subtracting the BC from the AC thresholds. Hearing loss was
classified as follows: 1) conductive: BC thresholds < 15 dB HL
and ABG 15 dB averaged over 0.5, 1.0, and 2.0 kHz; 2) pure
sensorineural: AC thresholds 15 dB HL and ABG < 15 dB
averaged over 0.5, 1.0, and 2.0 kHz; 3) high-frequency sensorineural: AC thresholds > 30 dB averaged over 4.0, 6.0, and 8.0 kHz;
and 4) mixed: BC thresholds 15 dB HL and ABG 15 dB HL
averaged over 0.5, 1.0, and 2.0 kHz. Hearing loss was substantiated by comparison with the 95th percentile value for sex- and
age-related hearing thresholds.12
Bone Densitometry
Anthropometric parameters, height and weight, were
measured to the nearest 0.1 cm using a wall-mounted Harpenden stadiometer (Holtain, Crymych, UK) and the nearest 0.1 kg
on a calibrated balance scale in light, indoor clothing without
shoes, respectively. Body mass index was calculated as the ratio
of weight (kilograms) to the squared height (square meters).
By means of DXA at lumbar spine (LS) (L1–L4) and whole
body (WB) level, the aBMD (grams per square centimeter) was
calculated, being the ratio of the bone mineral content (grams)
to the scanned bone area (square centimeters) (Hologic QDR4500A device, software version 11.2.1; Hologic, Bedford, MA).
LS DXA predominantly reflects trabecular aBMD, whereas WB
DXA provides an estimation of cortical bone aBMD.
Three-dimensional, volumetric bone parameters were
determined by pQCT at the proximal radius (66% of radius
length from distal [R-66]) of the dominant forearm, mainly
made up of cortical bone, and at the distal radius (4% of radius
length from distal [R-4]) of the nondominant forearm, predominantly consisting of trabecular bone tissue (XCT2000 scanner,
Swinnen et al.: BMD and Hearing Loss in OI
software version 5.4; Stratec, Pforzheim, Germany). The R-66
pQCT bone density parameter of interest was cortical bone
vBMD (Cort vBMD) (milligrams per cubic millimeter). Additional
parameters at this site were the bone geometry parameters cortical thickness (millimeters), periosteal circumference (millimeters),
and endosteal circumference (millimeters). From R-4 pQCT, trabecular bone vBMD (Trab vBMD) (milligrams per cubic
millimeter) was included for further analyses. Comparisons with
age- and sex-matched reference data enabled the conversion of
absolute values into z scores (number of standard deviations from
the mean) for LS aBMD, WB aBMD, and R-4 Trab vBMD.
Genetic Analysis
Molecular-genetic screening and analysis of mutations in
COL1A1 and COL1A2 were performed at the Center for Medical Genetics from the UZ Ghent in accordance with previously
described procedures.13 Nonsense and frameshift mutations in
COL1A1 or COL1A2 resulting in a premature stop codon were
classified as mutations leading to a quantitative type I collagen
defect. Missense mutations in either of these genes were interpreted as qualitative mutations, as they induce the formation of
structural abnormal type I collagen. To verify whether a splicesite mutation led to a quantitative or qualitative defect, a skin
biopsy was obtained in the proband to perform cDNA analysis
and a COL1A1 null allele test, of which the procedures were
analogous to those described by Symoens et al.14 and Nuytinck
et al.15 A positive COL1A1 null allele test and/or a decreased
migration of type I (pro)collagen on biochemistry pointed toward a quantitative defect of type I collagen synthesis, whereas
splice site mutations associated with a negative COL1A1 null
allele test and/or abnormal biochemical migration patterns of
type I (pro)collagen were considered qualitative mutations.
Statistical Analysis
All data were entered into SPSS for statistical analysis
(SPSS Inc., Chicago, IL). A v2 and Fisher exact test were performed to check associations between categorical variables.
Continuous variables were evaluated for normal distribution by
the Kolmogorov-Smirnov test. To determine whether z scores differed from 0, we used the one-sample t test. The Student t test
was applied for comparison of bone parameters between two
groups. Pearson correlation coefficient (r) was calculated to determine associations between continuous variables. Analyses of
covariance (ANCOVA) were performed to investigate differences
in BMD and geometry parameters as a function of hearing,
employing the Bonferroni adjustment for post hoc multiple comparisons. When the normal distribution was not achieved by the
outcome variable, we used Mann-Whitney and Kruskal-Wallis
tests and Spearman rank correlation coefficient (rs). For betweenrelatives comparisons of bone parameters, paired t tests were
applied. A 5% significance level was used throughout all analyses.
RESULTS
Demographics, Clinical OI Type, and Genotype
Fifty-six adult OI patients (22 male, 34 female)
with a mean age of 43 years (standard deviation: 13.7;
range, 18–80 years) participated in the study. None of
them had been treated with BP intravenously, with oral
BP before the age of 24 years, or with oral BP for a period longer than 4 years. Postmenopausal women (n ¼ 6)
were completely free of hormonal drug administration.
Fifty-two of 56 participants showed a positive family history for OI and originated from 25 different
Laryngoscope 122: February 2012
families from which one to four affected relatives participated in this study. In four participating patients, there
was evidence of an isolated form of OI, which was confirmed genetically.
All patients had been clinically diagnosed as OI
type I (n ¼ 49) or IV (n ¼ 7). Based on the results from
genetic tests, it was possible to differentiate between
patients with quantitatively (n ¼ 45) and qualitatively
(n ¼ 11) impaired type I collagen synthesis. All the
patients with quantitative defects and four of those with
qualitative defects had mutations located in COL1A1.
Only seven patients had a mutation in COL1A2, all of
which induced a qualitative type I collagen defect.
Hearing
Thirty-seven of 56 participants (66%) were diagnosed with hearing loss, which was bilateral in 31
patients. Forty-four of the total number of 112 ears
(39%) demonstrated normal hearing thresholds. The
hearing-impaired ears reflected a conductive, mixed, flat
sensorineural, or high-frequency sensorineural loss or
they had previously undergone stapes surgery. The proportionality of these different types is illustrated in
Figure 1. For further analysis, a classification into three
groups with respect to hearing was introduced. The first
group consisted of ears with normal hearing (NL H).
The second group comprised the ears with conductive
and mixed hearing loss (C/M HL), as well as the ears
that underwent stapes surgery. Finally, the last group
was characterized by pure SNHL, whether or not it was
limited to the highest frequencies. In all patients with
bilateral hearing loss, the same type of hearing loss was
found in both ears. Consequently, each patient could
exhaustively be assigned to the group of NL H, C/M HL,
or SNHL. Six patients had unilateral hearing loss,
which was a C/M HL in five patients and an SNHL in
one patient, and were classified according to their hearing-impaired ear. In summary, 19 patients had NL H
(34%), whereas 29 and eight patients demonstrated C/M
HL (52%) and pure SNHL (14%), respectively.
The occurrence and type of hearing loss as a function of clinical OI type and genotype are presented in
Table I and Table II, respectively. Fisher exact test could
not demonstrate significant associations between the
occurrence or type of hearing loss and the OI type, the
mutated gene, or the type I collagen defect.
Finally, hearing loss occurred in patients with a
positive family history of OI, as well as in patients with
isolated OI. Two of four patients with isolated OI showed
C/M HL. In the 17 families with two or more participating relatives, we noticed intrafamilial variability with
regard to hearing loss.
Associations Between Hearing Loss and Bone
Parameters
Osteosynthetic material biased BMD measurements
in one patient at DXA LS and in five patients at DXA
WB. Consequently, their results were excluded from the
analysis. In addition, pQCT measurements were not
Swinnen et al.: BMD and Hearing Loss in OI
403
Fig. 1. Normal hearing and different
types of hearing loss in 112 ears
from 56 osteogenesis imperfecta
patients. Because a similar underlying pathologic process was suspected, ears that had undergone
stapes surgery and those with conductive or mixed hearing loss (C/M
HL) could be bundled. Ears with
sensorineural hearing loss (SNHL),
whether or not limited to the highest
frequencies, were considered to
form one group.
performed in five patients because forearms had been
fractured in the past 2 years.
Mean values for age and anthropometrics for the
patients belonging to the groups of NL H, C/M HL, and
SNHL, separately, as well as for the whole group may be
consulted in Table III. Results for total fracture number,
BMD, and geometry parameters in the groups of
patients with NL H, C/M HL, and SNHL, as well as for
the whole group of patients, are presented in Table IV.
In the patient population taken as a whole, z scores
were significantly below 0 for LS aBMD (P < .001), WB
aBMD (P < .001), and R-4 Trab vBMD (P < .001). However, when considering the subgroups based on hearing,
z scores remained significantly below 0 in the groups
with NL H and C/M HL but not in the patients with
SNHL, which formed the smallest group.
To determine differences in BMD and bone geometry between the groups, a one-way ANCOVA was
conducted on each bone parameter with appropriate
adjustments. The latter are displayed as covariates in
Table IV and were based on significant correlations of
bone parameters with anthropometrics and age and on
differences in bone parameters according to sex or type I
collagen defect. Comparison of LS and WB aBMD converted into z scores with appropriate adjustments
revealed significantly higher LS and WB aBMD z scores
in the patients with SNHL, compared to those with NL
TABLE I.
Prevalence and Types of Hearing Loss as a Function of Clinical
Osteogenesis Imperfecta Type in 56 Osteogenesis
Imperfecta Adults.
OI Type I
Normal hearing
No. of
Subjects
16
COL1A1
No. of
Subjects
Qualitative
COL1A1
%
No. of
Subjects
COL1A2
%
No. of
Subjects
%
%
%
Normal hearing
14
25
1
2
4
8
29
3
5
Conductive/mixed
26
46
2
4
1
2
5
9
1
2
2
4
45
Sensorineural hearing loss
8
14
404
Quantitative
No. of
Subjects
25
Laryngoscope 122: February 2012
TABLE II.
Prevalence and Types of Hearing Loss as a Function of
Quantitative or Qualitative Type I Collagen Defect in
the COL1A1 or COL1A2 Gene in 56 Adult
Osteogenesis Imperfecta Patients.
OI Type IV
Conductive/mixed hearing loss
OI ¼ osteogenesis imperfecta.
H (P < .01 and P < .05, respectively) or C/M HL (P < .05
for both parameters). At R-4, differences between groups
were observed for Trab vBMD and Trab z score. The
patients with C/M HL reflected lower values than the
groups with NL H (P < .05 for both parameters) and
SNHL (P < .05 for both parameters). At R-66, no differences between hearing groups could be demonstrated for
Cort vBMD or for parameters of bone geometry.
No association was found in the group of patients
with C/M HL or in the patients with SNHL between the
average AC thresholds, BC thresholds, or average ABGs
(0.5, 1.0, 2.0 kHz) and the bone parameters obtained by
DXA and pQCT.
Intrafamilial paired comparisons between relatives
with NL H and C/M HL were executed in eight unrelated families, of which both patients with NL H and C/
M HL participated in the study. Between-relative comparisons for DXA and pQCT z scores are graphically
presented in Figure 2. Paired comparisons of BMD
parameters yielded lower values in patients with C/M
HL compared to their NL H relatives. However, differences were only significant for DXA LS aBMD (P < .05),
LS z score (P < .05), WB z score (P < .05), and R-4 Trab
z score (P < .05).
4
7
—
hearing loss
Sensorineural
hearing loss
Swinnen et al.: BMD and Hearing Loss in OI
TABLE III.
Age and Anthropometrics in Osteogenesis Imperfecta Patients With Normal Hearing, Conductive/Mixed Hearing Loss, Sensorineural Hearing Loss, and in the Whole Group.
NL H
C/M HL
No. F/M
13/6
15/14
Age, yr
35.4 6 8.5
47.2 6 13.8
SNHL
6/2
159.3 6 8.8
34/22
43.6 6 13.7
Height, cm
159.6 6 9.3
Weight, kg
64.0 6 9.5
65.3 6 11.0
75.8 6 15.4
66.4 6 11.7
BMI, kg/m
25.2 6 4.0
24.6 6 3.8
29.8 6 5.4
25.6 6 4.4
Radius length, cm
25.4 6 1.4
26.2 6 2.0
25.4 6 1.6
25.8 6 1.8
2
162.9 6 9.9
49.8 6 15.7
Whole Group
161.3 6 9.5
Data are presented as mean 6 standard deviation.
NL H ¼ normal hearing; C/M HL ¼ conductive/mixed hearing loss; SNHL ¼ sensorineural hearing loss; No. F/M ¼ female-to-male ratio; BMI ¼ body
mass index.
DISCUSSION
Hearing Loss
Sixty-six percent of our patients demonstrated hearing loss, of which the severity varied from mild to
profound and which was bilateral in most cases. Consistent with other studies, mixed hearing loss emerged as
the predominant hearing loss type in the adult OI population.7,8,16 After correction for the physiologically
normal age- and sex-related decline in hearing thresholds, we diagnosed a pure SNHL in a minority of the
adult OI patients, which was in accordance with other
studies.7,8
Although it has often been claimed to occur most often in OI type I, hearing impairment has been reported
in patients with clinical OI types III and IV as well.1,8
Both our OI type I and type IV population demonstrated
hearing loss; however, pure SNHL was not diagnosed in
our type IV patients. Despite the limited number of type
IV patients in the present study, a similar observation
was made by Pillion and Shapiro,16 and in the study by
Kuurila et al.8 less than 4% of type IV patients exhibited
pure SNHL.
Finally, in the population under study, the occurrence and type of hearing loss seemed to be independent
of the mutated gene or the effect of the mutation on type
I collagen synthesis, which was in accordance with the
results of a large Finnish population study.8
Associations Between Hearing Loss and BMD
Compared to the healthy population, our OI
patients had reduced BMD. This was demonstrated by
reduced z scores for aBMD and Trab vBMD, and is in
accordance with results of other studies focusing on
BMD in adult and pediatric OI patients.3,5,6
Furthermore, we encountered variations in BMD as
a function of occurrence and type of hearing loss. Hearing loss presenting as a conductive hearing loss that
progresses to a mixed hearing loss is associated with a
lower trabecular BMD at both between-groups and
between-relatives comparisons. Conversely, a pure
SNHL appears to develop in the OI patients with more
Laryngoscope 122: February 2012
appropriate BMD compared to patients with normal
hearing or C/M HL. However, the relationship between
higher BMD and the development of pure SNHL in OI
relied on a small number of patients.
Before interpreting these findings, we have to admit
that bone metabolism in the normal temporal bone differs from the rest of the skeleton in three major aspects.
First, the ossicles and cochlear capsule attain adult
dimensions and configuration at mid-fetal age, and calcification process is complete at age 1 year, after which
growth, modeling, and remodeling are virtually absent.
In the long bones, growth and modeling continue until
the age of 24 years. Second, the ossicles and bony cochlear capsule are almost completely made up of cortical
bone. The different patterns in BMD according to the
occurrence and type of hearing loss predominantly
account for measurements at sites consisting of mainly
trabecular bone. Unfortunately, because of the partial
volume effect, the R-66 pQCT parameter Cort vBMD has
been reported to be an unreliable parameter when cortical thickness is decreased, as is often the case in OI.17 A
third characteristic unique to the temporal bone is the
endochondral layer of the otic capsule being the sole
region in the human body in which primary trabeculae
persist throughout life.
A conceivable explanation for the development of
otosclerosis-like lesions in the temporal bones of the OI
patients with low BMD seeks supports from the hypothesis that otosclerosis is related to the disruption of the
normal inhibition of bone remodeling in the temporal
bone following the accumulation of intraosseous microfractures and
fatigue
microdamage.18
Recently,
researchers have discovered that the inhibition of bone
remodeling in the normal temporal bone may most likely
be attributed to the action of the antiresorptive cytokine
osteoprotegerin (OPG), which is expressed in extreme
amounts in the inner ear and diffuses centrifugally to
the perilabyrinthine bone and the middle ear through an
extensive network of osteocytic lacunar canaliculi in the
bone of the otic capsule.19 Because of the bone remodeling inhibition, microfractures and fatigue microdamage
are likely to accumulate in the temporal bones. These
Swinnen et al.: BMD and Hearing Loss in OI
405
Laryngoscope 122: February 2012
406
Swinnen et al.: BMD and Hearing Loss in OI
1.26 6 1.10*
n ¼ 27
134 6 40
1.27 6 1.20*
n ¼ 18
157 6 56
39.1 6 4.6
25.6 6 4.1
37.8 6 5.3
24.4 6 5.1
Periosteal
circumference, mm
Endosteal
circumference, mm
25.5 6 3.9
38.7 6 3.9
2.10 6 0.17
1,152 6 44
n¼6
0.70 6 1.28
174 6 42
n¼6
0.47 6 1.38
1.024 6 0.112
n¼7
0.93 6 2.01
0.860 6 0.213
n¼7
17 (3–25)
n¼8
SNHL
25.2 6 4.4
35.6 6 4.7
2.14 6 0.39
1,164 6 47
n ¼ 51
1.61 6 1.16*
147 6 48
n ¼ 51
1.17 6 1.17*
0.991 6 0.083
n ¼ 51
2.09 6 1.38*
0.789 6 0.134
n ¼ 55
15 (7–25)
n ¼ 56
Whole Group
‡
NS
NS
NS
NS
C/M HL < NL H‡; C/M HL < SNHL‡
C/M HL < NL H ; C/M HL < SNHL
‡
NLH < SNHL‡; C/M HL < SNHL‡
NS
NLH < SNHL†; C/M HL < SNHL‡
NS
NS
Between-Groups
Differences
Sex, age, radius length
collagen defect
Sex, age, type I
collagen defect
Sex, weight, type I
Sex, weight, type I
collagen defect
Covariates
Total fracture number is presented as median (interquartile range), as the normal distribution was not achieved. All other data are presented as mean 6 standard deviation.
*The z score was significantly different below 0 at one-sample t test (P < .001).
†
P < .01 at one-way analysis of covariance.
‡
P < .05 at one-way analysis of covariance.
NL H ¼ normal hearing; C/M HL ¼ conductive/mixed hearing loss; SNHL ¼ sensorineural hearing loss ; n ¼ number of subjects; NS ¼ not significant; DXA ¼ dual X-ray absorptiometry; aBMD ¼ areal
bone mineral density; Type I collagen defect ¼ either quantitatively or qualitatively impaired type I collagen synthesis; R-4 ¼ radius measurement site at 4% from distal; pQCT ¼ peripheral computed tomography; vBMD ¼ volumetric bone mineral density; R-66 ¼ radius measurement site at 66% from distal.
2.15 6 0.47
2.12 6 0.32
1,159 6 47
0.987 6 0.072
0.986 6 0.092
1,176 6 48
n ¼ 29
n ¼ 19
n ¼ 27
2.21 6 1.21*
2.35 6 1.20*
n ¼ 18
0.774 6 0.117
0.786 6 0.123
1.93 6 0.90*
n ¼ 29
n ¼ 19
1.43 6 1.32*
15 (6–23)
n ¼ 29
15 (10–23)
n ¼ 19
C/M HL
Cortical thickness, mm
Cortical bone
vBMD, mg/mm3
R-66 pQCT
Trabecular bone z score
vBMD, mg/mm
3
Trabecular bone
R-4 pQCT
z Score
aBMD, g/cm
2
DXA whole body
z Score
aBMD, g/cm
2
DXA lumbar spine
Total fracture no.
Fractures
NL H
TABLE IV.
Total Fracture Number and Bone Densitometry Results in Osteogenesis Imperfecta Patients With Normal Hearing, Conductive/Mixed Hearing Loss, Sensorineural Hearing Loss, and in
the Whole Group.
Fig. 2. Intrafamilial variability in hearing and bone density. Mean between-relative differences in z scores for (A) areal bone mineral density
(aBMD) at lumbar spine (LS) dual X-ray absorptiometry (DXA), (B) whole body (WB) DXA, and (C) trabecular volumetric BMD (vBMD) at 4%
of radius length from distal (R-4) peripheral quantitative computed tomography are displayed in eight families (x axis) of whom at least one
relative had normal hearing (NL H) and one demonstrated conductive/mixed hearing loss (C/M HL). Two NL H data points in (B) and one C/
M HL data point in (C) are missing because of osteosynthetic material and forearm fractures, respectively.
lesions are physiologically normal responses to mechanical load and strains during daily activities and are
constantly removed in other bones by bone remodeling.
Frisch et al.18 have suggested that accumulating microfractures and fatigue microdamage in the temporal
bones may disrupt the OPG signaling pathways and
could offer a pathogenetic factor in otosclerosis, in which
abnormal bone remodeling occurs. Indeed, Proops
et al.20 have demonstrated increases in the numbers of
microfractures in temporal bones with increasing age, as
well as a higher prevalence in otosclerosis. However, it
was not reported whether these microfractures particularly occur at the lateral capsular wall, the site of
predilection for otosclerosis.
Because OI is associated with a higher overall bone
fragility and a much faster accumulation of fatigue damage in cortical bone,2 the accumulation of temporal bone
microfractures in the ossicles and the stapes footplate, acting on the OPG canalicular network, and consequently,
resulting into bone loss and otosclerosis-like foci, sounds
like a very plausible hypothesis for the development of
conductive hearing loss in OI. Still, because the middle
ear ossicles are also relatively metabolically inactive and
dominantly constituted of cortical bone, the association
between lower trabecular BMD and development of middle ear pathology in OI remains ambiguous.
The present findings need to be confirmed in large
OI populations. Moreover, longitudinal follow-up of BMD
and bone geometry parameters in patients with OI from
a young age onwards is recommended, because bone parameters are relatively dynamic and influenced by
growth, drugs, hormonal changes, and physical activity.
In addition, the effects of BP and other drugs on hearing
ability in OI should be investigated. Administration of
BP is the current therapy of preference for increasing
bone mass and BMD in OI children and adults with
high fracture rates and low BMD z scores by inhibition
of the augmented osteoclastic bone resorption activity. If
the development of OI-related conductive hearing loss is
associated with reduced BMD and higher bone turnover
rate, BP administration may have an inhibitory effect on
Laryngoscope 122: February 2012
the development and progression of conductive hearing
loss.
Finally, histomorphometric analyses of temporal
bone structures and in vivo quantitative densitometric
imaging methods directly applied at the level of the temporal bones in OI patients developing different types of
hearing loss will hopefully offer more insight into the
underlying pathophysiology of OI-related hearing loss.
CONCLUSIONS
OI adults with C/M HL, which are most likely due
to otosclerosis-like lesions and fixation of the stapes footplate, were found to reflect a lower trabecular BMD
compared to OI patients with normal hearing or pure
SNHL. Whether this relationship implies that a low
trabecular BMD involves a higher susceptibility to thinning of the middle ear ossicles and accumulation of
microfractures interfering with the temporal bone
remodeling inhibition pathways deserves attention in
future research. Longitudinal follow-up studies evaluating BMD and hearing performance from a young age
onwards, as well as the effects of BP administration in
OI on hearing, should be encouraged.
Acknowledgment
The authors would like to thank all the participating
patients. Furthermore, they would like to express their
gratitude to the employees from the Unit for Osteoporosis
and Metabolic Diseases for their assistance in bone
densitometry.
BIBLIOGRAPHY
1. Sillence DO, Senn A, Danks DM. Genetic heterogeneity in osteogenesis
imperfecta. J Med Genet 1979;16:101–116.
2. Rauch F, Glorieux FH. Osteogenesis imperfecta. Lancet 2004;363:
1377–1385.
3. Rauch F, Lalic L, Roughley P, Glorieux FH. Relationship between genotype
and skeletal phenotype in children and adolescents with osteogenesis
imperfecta. J Bone Miner Res 2010;25:1367–1374.
4. Byers PH, Wallis GA, Willing MC. Osteogenesis imperfecta: translation of
mutation to phenotype. J Med Genet 1991;28:433–442.
5. Gatti D, Colapietro F, Fracassi E, et al. The volumetric bone density and
cortical thickness in adult patients affected by osteogenesis imperfecta.
J Clin Densitom 2003;6:173–177.
Swinnen et al.: BMD and Hearing Loss in OI
407
6. Rauch F, Land C, Cornibert S, Schoenau E, Glorieux FH. High and low
density in the same bone: a study on children and adolescents with mild
osteogenesis imperfecta. Bone 2005;37:634–641.
7. Pedersen U. Hearing loss in patients with osteogenesis imperfecta. A
clinical and audiological study of 201 patients. Scand Audiol 1984;13:
67–74.
8. Kuurila K, Kaitila I, Johansson R, Grenman R. Hearing loss in Finnish
adults with osteogenesis imperfecta: a nationwide survey. Ann Otol Rhinol Laryngol 2002;111:939–946.
9. Tabor EK, Curtin HD, Hirsch BE, May M. Osteogenesis imperfecta tarda:
appearance of the temporal bones at CT. Radiology 1990;175:181–183.
10. Shapiro JR, Pikus A, Weiss G, Rowe DW. Hearing and middle ear function
in osteogenesis imperfecta. JAMA 1982;247:2120–2126.
11. Hartikka H, Kuurila K, Korkko J, et al. Lack of correlation between the
type of COL1A1 or COL1A2 mutation and hearing loss in osteogenesis
imperfecta patients. Hum Mutat 2004;24:147–154.
12. ISO-7029. Acoustics - Statistical Distribution of Hearing Thresholds as a
Function of Age. Geneva, Switzerland: International Organization for
Standardization; 2000.
13. Swinnen FK, De Leenheer EM, Coucke PJ, Cremers CW, Dhooge IJ.
Audiometric, surgical, and genetic findings in 15 ears of patients with
osteogenesis imperfecta. Laryngoscope 2009;119:1171–1179.
Laryngoscope 122: February 2012
408
14. Symoens S, Nuytinck L, Legius E, Malfait F, Coucke PJ, De PA. Met>Val
substitution in a highly conserved region of the pro-alpha1(I) collagen
C-propeptide domain causes alternative splicing and a mild EDS/OI
phenotype. J Med Genet 2004;41:e96.
15. Nuytinck L, Coppin C, De PA. A four base pair insertion polymorphism in
the 30 untranslated region of the COL1A1 gene is highly informative for
null-allele testing in patients with osteogenesis imperfecta type I.
Matrix Biol 1998;16:349–352.
16. Pillion JP, Shapiro J. Audiological findings in osteogenesis imperfecta.
J Am Acad Audiol 2008;19:595–601.
17. Rauch F, Tutlewski B, Schonau E. The bone behind a low areal bone mineral density: peripheral quantitative computed tomographic analysis in
a woman with osteogenesis imperfecta. J Musculoskelet Neuronal Interact 2002;2:306–308.
18. Frisch T, Bretlau P, Sorensen MS. Intravital microlesions in the human
otic capsule. Detection, classification and pathogenetic significance
revisited. ORL J Otorhinolaryngol Relat Spec 2008;70:195–201.
19. Zehnder AF, Kristiansen AG, Adams JC, Merchant SN, McKenna MJ.
Osteoprotegerin in the inner ear may inhibit bone remodeling in the
otic capsule. Laryngoscope 2005;115:172–177.
20. Proops DW, Hawke WM, Berger G. Microfractures of the otic capsule. The
possible role of masticatory stress. J Laryngol Otol 1986;100:749–758.
Swinnen et al.: BMD and Hearing Loss in OI