what-when-how
In Depth Tutorials and Information
to identify significant portions of the bone, leading to
an erroneously low projection area and falsely high cal-
culated areal BMD, as only the more dense bone is mea-
sured. Often it is necessary to manually correct the errors
introduced by the automatic bone detection.
HIGH
AREAL DENSITY
LOW
AREAL DENSITY
PROJECTION
AREA
DE
TERMINANTS OF BMD IN
OI
VOLUMETRIC DENSITY
(IDENTICAL IN THE
TWO BONES)
Changes during Bone Development
Children and adolescents with OI tend to grow more
slowly than their healthy peers.
13
As areal BMD depends
on bone size, slow growth by itself would lead to a
slower accrual of bone mass and density during bone
development. In addition, children with OI have lower
volumetric BMD. A study on patients with moderate
to severe OI found that volumetric BMD at the lumbar
spine was about half of the level found in healthy con-
trols.
14
The deficit is less in patients with mild OI type I,
where volumetric BMD at the spine and in the trabecu-
lar bone of the radius is on average 20% lower than in
controls.
15
Interestingly, both adults and children with
mild OI consistently have higher volumetric BMD at the
shaft of the radius than controls.
15,16
This is explained by
the fact that bone shaft typically is thin and the marrow
cavity is small. Consequently, cortical bone makes up a
larger percentage of the bone's cross-section at the level
of the diaphysis.
15
RADIATION BEAM
LARGE
BONE
SMALL
BONE
FIGURE 24.1
Comparison of volumetric and areal bone den-
sity. The two bones are assumed to have the same volumetric den-
sity, but the larger bone appears denser when projected on an X-ray
screen. The reason is that it absorbs more radiation due to the longer
path length of the radiation beam through the bone.
(From: Rauch F,
Schoenau E. Skeletal development in premature infants: a review of bone
physiology beyond nutritional aspects. Arch Dis Child Fetal Neonatal Ed
2002;
86
: F82-5.)
directly. However, there are a number of approaches that
aim at estimating volumetric BMD from DXA results.
10,11
All of these approaches require assumptions about bone
shape and thus inevitably carry some error. In order to
determine volumetric BMD directly, it is necessary to use
techniques that determine bone volume and bone mineral
content simultaneously, such as computed tomography or
peripheral computed tomography.
1,12
The most widely used site of DXA analysis is the lum-
bar spine. The advantage of this location is that position-
ing is relatively simple in most patients. Total body DXA
scans are also frequently performed. However, only skel-
etal areas that do not contain any metal can be used for
evaluation.
Unfortunately, measurement artifacts are quite fre-
quent in young children with OI. The child may be unco-
operative and move during the study. Therefore parents
are usually present and help to immobilize and comfort
their child. Distraction items such as toys, books and
music are helpful. Shortening the time of scanning proce-
dure is also often an option. Another cause for incorrect
measurements results from the difficulty of positioning
the patient on the measurement table, especially patients
with scoliosis. To avoid this problem, the operator
should check how the spine was positioned on a previ-
ous study by looking at the image obtained then, and try
to place the patient accordingly.
Automatic analysis of DXA scans by computer algo-
rithms may be difficult in children with OI because their
bone density is often very low. The level of X-ray absorp-
tion by such bones may be very close to the detection
limit of the machine. In such cases the analysis may fail
Genotype-Phenotype Correlations
The relationship between genotype and lumbar spine
BMD was assessed in a study on 192 children and ado-
lescents with mutations in the
COL1A1
or
COL1A2
genes.
17
Patients were grouped according to the type
of mutation (haploinsufficiency mutations affecting
the alpha 1 chain, glycine substitutions affecting the
triple helical domain of either the alpha 1 or the alpha
2 chain). Compared to patients with helical mutations,
patients with haploinsufficiency on average were taller
and heavier and had higher lumbar spine areal BMD.
However, the differences were small, given the often
marked differences in the severity of bone fragility
(
Figure 24.2
). After adjustment for age, sex and height
z-scores, the mean lumbar spine areal BMD z-scores
were −4.0 for the haploinsufficiency group and −4.7 for
both helical mutation groups. Thus, areal BMD was very
low in all genotype groups.
This study also showed that average lumbar spine
areal BMD z-score was higher by 0.6 in girls than in
boys. A similar sex-difference was also found in the sub-
group of patients with haploinsufficiency mutations.
Sex-differences persisted after adjustment for height
z-score, suggesting that factors other than bone size
