0021-972X/03/$15.00/0
Printed in U.S.A.
The Journal of Clinical Endocrinology & Metabolism 88(11):5266 –5272
Copyright © 2003 by The Endocrine Society
doi: 10.1210/jc.2003-030432
Cortical Bone Density Is Normal in Prepubertal Children
with Growth Hormone (GH) Deficiency, but Initially
Decreases during GH Replacement due to Early
Bone Remodeling
R. SCHWEIZER, D. D. MARTIN, C. P. SCHWARZE, G. BINDER, A. GEORGIADOU, J. IHLE,
M. B. RANKE
AND
Pediatric Endocrinology Section, University Children’s Hospital, University of Tuebingen D-72076, Germany
Dual energy x-ray absorptiometry (DEXA) has revealed that
GH- deficient adults gain in bone mineral density during GH
therapy. Measurements of volumetric bone density (grams per
cubic centimeter vs. grams per square centimeter) and structure, however, are achieved through peripheral quantitative
computed tomography (pQCT). In 45 prepubertal GH-deficient
children, we studied pQCT measurements before the start and
for 12 months of GH treatment. Serum alkaline phosphatase
(AP), procollagen I carboxyl-terminal propeptide (PICP), and
deoxypyridinoline reflected bone metabolism status. Findings at the start of GH treatment were (mean SD score): bone
area, ⴚ0.44; cortical density, ⴚ0.03; cortical area, ⴚ1.32; cortical thickness, ⴚ1.41; and marrow area, ⴙ0.66. At 12 months,
C
HANGES IN HEIGHT velocity are mainly attributed to
the effectiveness of GH replacement therapy in children with GH deficiency (GHD). In adults with GHD, however, the efficacy of replacement is ascertained by observed
changes in body composition and function. Characteristic of
these adult patients is their low bone mineral density (BMD),
reduced lean body mass, and substantially higher fat mass
(1–3), but, under replacement therapy, a decrease in fat mass
and an increase in muscle mass as well as higher BMD have
consistently been observed (4 –9). Although the effects of GH
on body composition, such as a decrease in skinfold thickness, were described more than a decade ago (10, 11), it is still
not common practice to monitor compartments of the body
during GH therapy in children and adolescents. In view of
the fact that childhood-onset GHD requires life-long medical
care and that distinct effects of GH, such as height increment,
obviously cannot serve as identifiable parameters in adulthood, we support the view that body composition measurements should start at the time GH therapy is initiated in
children and adolescents.
Several investigators have shown that GH therapy has an
effect on bone metabolism in terms of both bone formation
and bone resorption (12–16). There is, however, no consensus
on the biochemical parameters that should be determined,
nor is there clarity about the definitive techniques for inves-
Abbreviations: AP, Alkaline phosphatase; BMD, bone mineral density; ⌬, change in; DEXA, dual energy x-ray absorptiometry; DPD, deoxypyridinoline; GHD, GH deficiency; PICP, procollagen I carboxylterminal propeptide; pQCT, peripheral quantitative computed
tomography; SDS, sd score.
cortical density had fallen to ⴚ0.73 (P < 0.001), whereas cortical area and thickness, and marrow area did not change. AP,
PICP, and deoxypyridinoline increased significantly within
the first 3 months (increase: AP, 66.5 U/liter; PICP, 72 ␮g/liter;
DPD, 11.4 nmol/mmol creatinine). The pQCT showed that cortical density is not reduced in GH-deficient patients. Higher
bone metabolism explains the lower cortical density after GH
therapy commenced. Thus, the manifestation of GH deficiency is evidently similar in children and adults, and pQCT
provides important information in addition to DEXA measurements, as DEXA does not take bone structure into
account. (J Clin Endocrinol Metab 88: 5266 –5272, 2003)
tigating structural bone changes in children that are both safe
and cost-effective as well as physiologically appropriate.
Dual energy x-ray absorptiometry (DEXA) is restricted to
the measurement of a two-dimensional area density (grams
per square centimeter), which is calculated by means of the
x-ray absorption. In contrast, peripheral quantitative computed tomography (pQCT) allows the study of both volumetric bone density (grams per cubic centimeter) and bone
structure (17). This is an important feature, because bone
density alone is not the single parameter that determines
bone stability: in fact, it has been shown that bone structure
is also a relevant factor (18). Our objective was, therefore, to
conduct a longitudinal study of biochemical parameters indicative of bone metabolism and bone structure by applying
pQCT during the first year of GH replacement in prepubertal
children with idiopathic GHD. We assumed that evidence
would be found to corroborate reported findings associating
an increase in bone metabolism with a significant increase in
volumetric bone density during GH treatment.
Subjects and Methods
The investigations were conducted in our unit between March 1999
and April 2001. Of the primary population of 53 prospectively investigated children, 4 were excluded because of organic GHD, and 4 because they developed signs of puberty. The study population, therefore,
consists of 45 (8 female) prepubertal (females, B1; males, G1; testis
volume, ⬍3 ml) children in whom GHD had been recently established
based upon anthropometrical characteristics and biochemical tests (19).
In all children, GH levels (measured by RIA) in 2 standard stimulation
tests (insulin hypoglycemia and arginine) did not exceed 8 ␮g/liter, and
basal levels of IGF-I were below ⫺1.0 sd score (SDS) for age (19). In cases
5266
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
with additional pituitary hormone deficiencies (TSH and ACTH; n ⫽ 3),
replacement of T4 (75 ␮g/m2䡠d, orally) and/or hydrocortisone (10 mg/
m2䡠d) was given at least 3 months before and during GH therapy. In all
patients, magnetic resonance imaging of the hypothalamic-pituitary and
stalk-pituitary region was performed. In 10 patients (including the 3
with additional hormone deficiencies) we observed hypoplasia of the
pituitary gland and/or stalk or an ectopic neuropituitary gland. All
children were treated with a single daily evening injection of authentic
recombinant human GH of various brands. The initial dosage was variable (median dose, 0.028 mg/kg䡠d; range, 0.018 – 0.047 mg/kg䡠d), but
was kept constant during the observational phase. Children with additional chronic disorders and/or dysmorphological features were not
included in the study.
Anthropometrical and biochemical parameters were measured before as well as 3, 6, and 12 months after initiation of GH replacement.
A pQCT scan was performed before as well as 6 and 12 months after GH
treatment started. After overnight fasting, samples of blood and urine
were taken in the morning, between 0900 –1100 h. The study was approved by the ethics committee of the Faculty of Medicine, University
of Tuebingen, and informed written consent was given by the parents.
pQCT
Bone structure and volumetric bone density were measured using a
pQCT device (XCT 2000, Stratec, Inc., Pforzheim, Germany). The scanner
was equipped with a low energy (38 keV) x-ray tube. The radiation dose
for a single scan was approximately 0.3 ␮Sv, with an effective dose for
the forearm of about 0.1 ␮Sv. The radiation source was 45 kV at 150 ␮A.
The machine was calibrated once a week with a standard phantom and
once a month with a cone phantom provided by the manufacturer. The
proximal radius of the nondominant arm was chosen, and cross-sectional measurements were taken at exactly 65% of the ulna length away
from the radius growth plate. For this, the radius growth plate was
precisely located with a scout-view scan. This position of measurement
was chosen because it is the site comprising the biggest muscle area
cross-section for which Neu et al. (20) established age-dependent reference values in 2001 using the same pQCT device for healthy German
children. A relative (65%) distance was chosen because the arm is constantly growing in childhood. This ensured the measurement of the
exactly corresponding site, regardless of changing arm lengths. A 2-mmthick, single tomographic slice was taken at a voxel size of 0.4 mm. Image
processing and calculation of numerical values were performed by
means of the software package supplied by the manufacturer (version
5.4, Stratec, Inc.). The following parameters were determined: 1) total
area (square millimeters), 2) cortical area (square millimeters), 3) marrow area (square millimeters), 4) cortical thickness (millimeters), and 5)
cortical density (milligrams per cubic centimeter) of the radius as well
as muscle area (square millimeters).
The entire cross-sectional area of the radius (total area) and the crosssectional area of cortical bone (cortical area; square millimeters), were
determined by detecting the outer and inner cortical bone contours at
a threshold of 710 mg/cm3. Volumetric cortical BMD (cortical density)
represents the mass of mineral per unit volume of the cortical bone mass
(milligrams per cubic millimeters). Marrow area represents the difference between total area and cortical area. Cortical thickness (millimeters)
was calculated as outer bone radius minus marrow radius [outer bone
radius ⫽ (total area/⌸)1/2 and marrow radius ⫽ (marrow area/⌸)1/2].
This procedure was described by Neu et al. (20). Muscle area was measured at a threshold of 30 –70 mg/cm3. The measurements were transformed into SDSs based on the age-specific references (20). Presuming
height to be an important parameter influencing bone structure, we
decided to establish the SDS by deriving height age, this being the age
obtained by projecting a given patient’s height on to the corresponding
median height of the reference population. To establish the variability
of the measurements, 3 investigators measured the forearm of an adult
volunteer 12 times. The coefficients of variation for total area, cortical
area, marrow area, cortical thickness, cortical density, and muscle area
were 2%, 0.9%, 6%, 1.6%, 0.3%, and 3%, respectively.
J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272 5267
minal propeptide of procollagen I (PICP) as markers of bone formation
and the urinary deoxypyridinoline (DPD) as a marker of bone resorption
(21, 22).
AP
Serum alkaline phosphatase was measured by the p-nitrophenylphosphate color method as provided by Roche (Mannheim, Germany). The measurement requires a sample volume of 150 ␮l serum. The
sensitivity is 5 U/liter, and the intra- and interassay coefficients of
variation were 0.5% and 0.4% at 458 and 579 U/liter, and 2.2% and 2.1%
at 357 and 563 U/liter, respectively (information as given by the producer). For comparison, modified reference data for prepubertal children according to Lockitch et al. (23) were used. Reference values
showed no age dependency for prepubertal children.
PICP
PICP was measured with a sandwich ELISA provided by Metra
Biosystems GmbH (Osnabruck, Germany). The measurement requires a
sample volume of 100 ␮l serum. The sensitivity was 0.2 ␮g/liter, and the
intra- and interassay coefficients of variation were 6.8% and 5.0% at 80.8
and 296.7 ␮g/liter, and 8.8% and 7.8% at 51.1 and 437.9 ␮g/liter, respectively. Our own reference data, established on the basis of 268
healthy children (155 female) between 3 and 18 yr of age, were used to
calculate the SDS. Reference values showed no age dependency for
prepubertal children, and the median and 90% ranges were 274 and
154 – 486 ␮g/liter, respectively (ln mean, 5.61; ln sd, 0.35).
DPD
For the automated chemiluminescence assay employed to measure
DPD (Imulite Pyrilinks-D, DPC Biermann, Bad Nauheim, Germany), a
75-␮l urine volume was used. The measured DPD concentrations were
normalized by means of the urinary creatinine concentrations by calculating a DPD/urinary creatinine ratio in nanomoles per millimoles of
creatinine. The intraassay coefficients of variation in the DPD measurement were 9.3%, 5.9%, and 7.4% at 9.5, 22.3, and 50.1 nmol/mmol
creatinine. The interassay coefficients of variation were 11.2%, 8.4%, and
8.8% at 9.3, 22.9, and 48.8 nmol/mmol creatinine. The sensitivity specified by the manufacturer was 4 nmol/liter (21).
The measured values were compared with the reference values for
prepubertal children reported by Elmlinger et al. (21), which did not
show any age dependency.
Statistical analyses
Statistical analysis was performed using the computed statistics program JMP. Results are expressed as the mean ⫾ sd unless otherwise
specified. The significance of changes was tested using a paired t test.
To study the relationship between serum, urine, and pQCT parameters,
regression analyses (Pearson’s coefficient of variation) were performed.
SDSs were calculated as the patient value minus the mean of the ageand sex-matched reference value divided by the sd of the age- and
sex-matched reference value. ⌬ values (⌬ ⫽ change in) are expressed as
the difference between a value at time t2 minus a value at time t1.
Results
Patient characteristics at the start of GH therapy are shown
in Table 1. At the start of GH treatment, patients were (median and range in parentheses) 7.5 yr old (3.3–14.4 yr) with a
height (SDS) of ⫺2.9 (⫺5.4 to ⫺1.8 sd) and a serum IGF-Ilevel (SDS) of ⫺3.5 (⫺14.7 to ⫺0.8). Height velocity during
the first year of GH therapy was 9.1 cm (5.5–12.4 cm), corresponding to a ⌬ height (SDS) of 0.8 sd (0.2–1.8 sd; see
Table 1).
Biochemical parameters
Biochemical markers
From the large number of potential markers of bone metabolism, we
chose total serum alkaline phosphatase (AP) and serum carboxyl-ter-
At the start of GH treatment bone metabolism parameters
in almost all patients were within the reference range (Fig. 1).
5268
J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272
TABLE 1. Patient characteristics (n ⫽ 45)
Gestational age (wk)
Birth weight (SDS)
Target heightb (SDS)
GH-max in testc (␮g/liter)
(pmol/liter)
IGF-I start of GH (SDS)
Age start of GH (yr)
Bone age start of GHd (yr)
Height start of GH (SDS)
Body mass index start of GH (SDS)
Height velocity 1st yr GH (cm/yr)
⌬ Height 1st yr GH (SDS)
Mean
SD
40a
⫺0.48
⫺0.59
5.30a
0.25
⫺3.50
7.52a
6.00a
⫺2.89
⫺0.51
9.12a
0.77
28 – 42a
0.93
0.68
1.20 – 8.00a
0.06 – 0.372
2.51
3.30 –14.41a
1.70 –12.70a
0.85
0.76
5.53–12.42a
0.46
a
Median, range.
According to Tanner et al. (11).
c
Conversion factor, 0.0465.
d
According to Greulich and Pyle (34).
b
However, in four patients serum AP levels exceeded the 95th
percentile; two patients had PICP serum levels below the 5th,
and 3 had values above the 95th percentile. DPD values were
in the lower normal range, with a median concentration of
23.2 nmol/mmol creatinine. Five patients had DPD levels
below the 5th percentile. Median AP and PICP serum levels
were normal, with concentrations of 293 U/liter and 281
␮g/liter, respectively. Markers of bone metabolism increased significantly after the first 3 months of GH treatment
(see Table 2). Thereafter, no significant further changes
occurred.
Bone density and structure
At the onset of GH therapy. The following measurements are
expressed in terms of median and range (see Table 3 for mean
and sd). Cortical density was normal at 991 mg/cm3 (849 –
1089 mg/cm3), corresponding to 0.03 SDS (⫺2.7 to 2.2). Total
area was only slightly reduced, with an absolute area of 74.9
mm2 (24.0 –133.8), corresponding to ⫺0.43 SDS (⫺3.5 to 1.45).
Cortical area was significantly reduced at 30 mm2 (8 – 62.5),
corresponding to ⫺1.38 SDS (⫺3.3 to 0.6), as was cortical
thickness at 1.12 mm (0.25–1.88), corresponding to ⫺1.32 SDS
(⫺3.6 to 0.4). Bone marrow area was increased at 47.6 mm2
(7.7–97.0), corresponding to ⫹0.51 SDS (⫺1.8 to 4.0). The
SDSs of all parameters were higher when matched with
height age. The total area as well as the marrow area were
above normal, whereas the cortical area and cortical thickness were only slightly reduced (see Table 3 for mean and sd
values). Muscle area proved to be reduced when we compared it with age- as well as height age-matched controls
(Table 3).
During GH therapy. During GH treatment cortical density
decreased significantly: ⌬ cortical density (SDS) ⫽ ⫺0.32 ⫾
0.91 sd (mean ⫾ sd) after 6 months, and ⫺0.70 ⫾ 0.96 sd after
12 months (see also Table 3 and Fig. 2). Cortical area, marrow
area, and cortical thickness did not change significantly (see
Table 3). Total area increased significantly: ⌬ total area
(SDS) ⫽ ⫹0.27 ⫾ 0.72 sd after 6 months, and ⫹0.41 ⫾ 1.00 sd
after 12 months.
No significant correlations were found between changes in
bone metabolism parameters and changes in bone structure.
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
After 3 months of GH therapy, nearly all bone metabolism
parameters had increased (AP at start vs. AP after 3 months:
r ⫽ 0.77; P ⬍ 0.001; PICP at start vs. PICP after 3 months: r ⫽
0.46; P ⬍ 0.001; DPD at start vs. DPD after 3 months: r ⫽ 0.51;
P ⬍ 0.001). The correlation between the changes in AP and
PICP over the first 3 months of GH treatment was highly
significant (r ⫽ 0.47; P ⬍ 0.001). A weak correlation was
found between ⌬AP and ⌬DPD over the 3-month period (r ⫽
0.3; P ⫽ 0.03). ⌬PICP and ⌬DPD showed no correlation (Fig.
3). In the following 9 months, all three parameters showed
no significant change.
Height velocity during the first 6 months of GH treatment
showed a correlation with the change in PICP in the sixth
month of therapy (r ⫽ 0.34; P ⫽ 0.05), but showed a weak
correlation with ⌬AP over the initial 6 months (r ⫽ 0.23; P ⫽
0.06) and no correlation with ⌬DPD during the same period.
We observed a very strong correlation between cortical
area (x-axis) and muscle area (y-axis) before and during GH
treatment. The ratio between cortical area and muscle area
decreased during GH treatment (from 0.030 to 0.026 to 0.025,
at the start and after 6 and 12 months of GH treatment,
respectively; see Fig. 4).
Discussion
In GHD adults, the extent of loss in bone mass and BMD
varies between patients whose disease had an adult onset
and those with a childhood onset (2). The extent of bone mass
loss evidently correlates with the duration of unsubstituted
GHD, and most researchers reported a reduced BMD (1–3).
Another important parameter influencing the loss of bone
mass in adult GHD is the combination with other hormone
deficiencies, such as sex steroids (8). For the determination
of bone mineral status in these studies, DEXA or single photon absorptiometry were used. In one study with GHD children, Boot et al. (4) showed reduced area BMD of the lumbar
spine and total body, as measured by DEXA. With reference
to the interpretation of results in these studies, it must be
remembered that two-dimensional methods such as DEXA
or single photon absorptiometry do not measure a threedimensional volumetric density; instead, only area density is
calculated from the x-ray absorption.
One uniform finding in GHD adults and children is the
increase in the levels of bone metabolism parameters, which
is understood as an expression of increased bone turnover
after the start of GH therapy (9, 12, 16). This is partly the
result of direct GH effects, but is mainly caused by IGF-I,
which has endocrine as well as para- and autocrine activity
(24 –26).
Our present study is the first to examine the relationship
between changes in bone metabolism and direct changes in
bone structure and volumetric cortical density during GH
treatment in children with idiopathic GHD. As could be
anticipated, the markers of bone metabolism increased significantly during the first 3 months of GH treatment. In
contrast to the literature, however, and contrary to our initial
hypothesis, we found a normal cortical density before the
start of therapy, which decreased on treatment. There are two
possible explanations for this finding. Firstly, there could be
a difference in the manifestation of adult GHD compared
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272 5269
FIG. 1. Parameters of bone metabolism
at the start of GH treatment.
TABLE 2. Bone metabolism parameters at start and after 3, 6, and 12 months of GH treatment
AP (U/liter)
Median
Range
AP (␮Kat/liter)d
Median
Range
PICP (␮g/liter)e
Median
Range
DPD (nmol/mmol creatinine)
Median
Range
Start of GH
3 months GH
6 months GH
12 months GH
⌬ 3 months
293
169 –779
381a
211–527
420a,b
189 –788
370.5a,b,c
183– 803
68.5
⫺49 to 369
4884
2817–12,988
6351
3517–13,786
7001
3151–13,136
6176
3051–13,386
1109
⫺817 to 6151
281
80 –769
384a
168 – 828
372a,b
108 –1035
348a,b,c
183– 807
72
⫺238 to 503
23.2
9.2– 40.0
32.1a
7.6 – 65.1
30.1a,b
10.4 –59.3
28.5a,b,c
16.9 – 46.4
11.4
⫺14.9 to 35.2
There was a significant increase in all parameters after 3 months.
Changes from start (P ⬍ 0.001).
b
Changes from 3 months (P ⬎ 0.1).
c
Changes from 6 months (P ⬎ 0.1).
d
Conversion factor, 16.67.
e
No conversion factor available.
a
with childhood GHD. In children, the most important indication of GHD is growth retardation and delayed bone age.
In GHD adults, however, an increase in body fat and a
simultaneous decrease in muscle mass are the most characteristic findings, and are occasionally accompanied by a reduced area BMD. It is likely, however, that the duration of
insufficient GH secretion during childhood is comparatively
too short to affect bone density. It is also relevant to mention
here that 40 – 80% of children with idiopathic isolated GHD
do not prove to have sustained GHD as adults.
Secondly, our study is the first to use a new method for
determining bone density in GHD. The pQCT allows measurement of volumetric bone density as well as determination of bone structure. In comparison with a healthy reference population measured with exactly the same method
(20), the cross-sectional pQCT measurement of the radius in
our patients showed near-normal total area (mean, ⫺0.35
SDS) and increased marrow area (mean, ⫹0.57 SDS), but
decreased cortical area (mean, ⫺1.27 SDS) and cortical thickness (mean, ⫺1.25 SDS) and normal cortical density (mean,
0.03 SDS). The normal findings for total area were unexpected, as the reference group was age-matched. Using stature-matched SDS, we observed that total area was markedly
higher than normal. We assumed that our short patients
would prove to have smaller bones with decreased bone
area, cortical area, and cortical thickness. We thus conclude
that the cortical bones in GHD children who do not receive
GH therapy are wider than normal and have thinner walls
than those of age- and stature-matched healthy children. The
bone marrow enlargement is probably due to the fact that
endocortical bone resorption is more pronounced than pericortical bone deposition. This presumably leads to a compensatory functional stability of bone during the state of
GHD in children. In the light of our findings, which showed
that cortical thickness and area are reduced, whereas cortical
density itself is normal, it becomes possible to explain the
lower area BMD (milligrams per square centimeter) found in
DEXA studies of GHD adults (1–3) and children (4). In the
two-dimensional projection used in DEXA, these changes
lead to a reduction in absorption, which is then calculated as
a reduced area bone density (milligrams per square centimeter) and total bone mass. Our results confirm other published reports on the issue of reduced bone mass. In addition,
our results provide new information, as they demonstrate
that cortical density is normal, whereas cortical proportions
undergo change. Future pQCT studies of adults with GHD
would presumably lead to the same results, but to our knowledge no published reports are yet available. Our findings
5270
J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
TABLE 3. pQCT parameters at the start and after 6 and 12
months of GH treatment
Start of GH
Mean
Cortical density (SDS)
Age
Total area (SDS)
Age
Height age
Marrow area (SDS)
Age
Height age
Cortical area (SDS)
Age
Height age
Cortical thickness
(SDS)
Age
Height age
Muscle area (SDS)
Age
Height age
SD
6 months of
GH
Mean
SD
12 months of
GH
Mean
SD
⫺0.03 1.13 ⫺0.36 0.96a ⫺0.73 0.81b,c
⫺0.44 1.10 ⫺0.18 1.06c ⫺0.03 1.19d,e
0.35 1.33
0.58 1.30a
0.66 1.43a,e
0.84 1.14f
1.42 1.38f
0.94 1.36e,f
1.36 1.43e,f
⫺1.32 0.83 ⫺1.17 0.64f
⫺0.43 1.06 ⫺0.38 0.88f
⫺1.19 0.82e,f
⫺0.40 0.92e,f
⫺1.41 0.98 ⫺1.39 0.78f
⫺0.93 1.04 ⫺1.05 0.93f
⫺1.46 1.04e,f
⫺1.05 1.03e,f
0.66 1.26
1.29 1.43
⫺2.47 0.93 ⫺1.79 0.95g ⫺1.65 1.03g,h
⫺1.59 1.11 ⫺0.89 1.06g ⫺0.70 1.14e,g
P ⬍ 0.05 vs. start.
P ⬍ 0.001 vs. start.
P ⬍ 0.01 vs. 6 months.
d
P ⬍ 0.01 vs. start.
e
P ⬎ 0.1 vs. 6 months.
f
P ⬎ 0.1 vs. start.
g
P ⬍ 0.001 vs. start.
h
P ⬍ 0.05 vs. 6 months.
a
b
c
thus indicate that the manifestation of childhood GHD is not
different from that of adult GHD, except that growth retardation occurs additionally in childhood GHD.
In terms of bone health, both bone stability and solidity are
more important than bone mass. A study by Schönau et al.
(18) illustrated that the solidity of a bone is a function not
only of its density but also of its geometry. Further parameters for bone solidity are trabecular density and alignment
(27, 28), which, however, cannot yet be evaluated in vivo and
thus require a bone biopsy. In summary, bone solidity is
determined by material characteristics, mass (bone mineral
content), architecture and geometry, and the three-dimensional organization of the trabecules. Further research is
needed to establish the significance of each of these parameters for bone fracture prevention. A few studies have suggested that GHD in adults poses a higher risk for fractures
(29). There are, however, no pQCT studies of the bones of
GHD adults, nor is literature available on this subject with
regard to children. It is worth mentioning that none of the
children we studied suffered a bone fracture.
In our cohort of GHD children, cortical density was found
to decrease during the first year of GH therapy. Studies with
DEXA on GHD adults have also shown that bone density
remains unchanged, or even decreases slightly, in the first 6
months of GH therapy (6, 7, 9, 14). Because DEXA was
employed, however, these studies could not clearly determine whether cortical thickness or cortical density was affected by GH therapy.
The decrease in cortical density found in our study could
be explained by the strong increase in bone remodeling and
remodeling space (30). This is reflected in an increase in both
osteoblastic (AP und PICP) and osteoclastic (DPD) param-
FIG. 2. Changes in cortical density during GH treatment.
eters. The rise in DPD in our study was possibly underestimated, because we corrected urinary DPD concentrations
according to urinary creatinine concentrations, which may
increase during GH treatment due to increasing muscle
mass. Further, bones grow rapidly during the first year of GH
therapy (median growth velocity, 9 cm/yr). This may initially led to lower volumetric cortical bone density due to
undermineralized, newly formed bone, associated with the
higher bone turnover. There are no data indicating that the
rate of fractures increases in GHD children during this
catch-up phase; however, there is reportedly a higher incidence of slipped capital femoral epiphysis (31, 32).
Changes in bone metabolism do not correlate with changes
in volumetric bone density and structure. It must be assumed
that the relatively rapid changes in bone metabolism parameters, which are influenced by a multitude of environmental,
behavioral, and nutritional factors, do not directly reflect the
long-term bone changes.
It has been recently suggested that the major factor influencing bone mass is muscle mass (33). We, therefore, tested
the correlation between muscle area and cortical area measured by means of pQCT, and our results showed a strong
relationship for the period before and during GH treatment.
During the course of GH treatment, however, the relation
between muscle area and bone area changes, in that normalization in muscle area is higher than that in cortical area;
therefore, the ratio between cortical area and muscle area
decreased to normal, as in healthy prepubertal children it is
approximately 0.023 (33).
The follow-up of our patients will clarify whether another
increase in cortical density occurs and whether significant
changes in bone structure are observable. The studies with
DEXA show a significant increase in BMD and bone mass
starting after 6 months of GH treatment in GHD adults (7, 8)
and children (4).
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272 5271
FIG. 3. Cross-correlations of changes in
bone metabolism parameters after 3
months of GH treatment (A–C).
FIG. 4. Correlation between muscle
cross-sectional area and cortical area of
the radius before (A) and after 6 (B) and
12 (C) months of GH treatment.
The present study is the first to provide evidence that
cortical bone density does not decrease in untreated GHD
children and that an abnormal bone structure is the likely
cause of the lower x-ray absorption reported in previous
DEXA studies. The same is probably true in the case of GHD
adults. The increase in bone modeling and remodeling together with the catch-up growth result in a decrease in cortical density during the first year of GH therapy in GHD
children. The follow-up of this study population will provide
further data on the influence of long-term GH treatment on
volumetric bone density and bone structure in these children.
Further, the results presented here support the proposal that
the manifestation of GHD is not different in adults compared
with children, and that pQCT is a precise, cheap, and quick
tool that offers important information for the assessment of
changes in bone in addition to DEXA measurements. We,
therefore, recommend that measurements of bone density
and structure should, in the future, be made by means of
pQCT in addition to DEXA.
2.
3.
4.
5.
6.
7.
8.
Acknowledgments
9.
Received March 12, 2003. Accepted August 1, 2003.
Address all correspondence and requests for reprints to: Dr. Roland
Schweizer, Pediatric Endocrinology Section, University Children’s Hospital, Hoppe-Seyler Strasse 1, D-72076 Tuebingen, Germany. E-mail:
[email protected].
This work was supported in part by an educational grant from Pharmacia GmbH (Erlangen, Germany) and Fortüne Forschungsförderung,
a grant from University of Tuebingen.
10.
11.
12.
13.
References
1. Amato G, Carella C, Fazio S, La Montagna, G, Citadini A, Sabatini D,
Marciano-Mone C, Saccá L, Bellastella A 1993 Body composition, bone me-
14.
tabolism, and heart structure and function in growth hormone (GH)-deficient
adults before and after GH replacement therapy at low doses. J Clin Endocrinol
Metab 77:1671–1676
Holmes SJ, Economou G, Whitehouse RW, Adams JE, Shalet SM 1994 Reduced bone mineral density in patients with adult onset growth hormone
deficiency. J Clin Endocrinol Metab 78:669 – 674
Rosen T, Hansson T, Granhed H, Szucs J, Bengtsson BA 1993 Reduced bone
mineral content in adult patients with growth hormone deficiency. Acta Endocrinol (Copenh) 129:201–206
Boot AM, Engels MA, Boerma GJ, Krenning EP, De Muinck-Keizer S 1997
Changes in bone mineral density, body composition, and lipid metabolism
during growth hormone (GH) treatment in children with GH deficiency. J Clin
Endocrinol Metab 82:2423–2428
Janssen YJ, Doornbos J, Roelfsema F 1999 Changes in muscle volume,
strength, and bioenergetics during recombinant human growth hormone (GH)
therapy in adults with GH deficiency. J Clin Endocrinol Metab 84:279 –284
Johannsson G, Rosen T, Bosaeus I, Sjostrom L, Bengtsson BA 1996 Two years
of growth hormone (GH) treatment increases bone mineral content and density
in hypopituitary patients with adult-onset GH deficiency. J Clin Endocrinol
Metab 81:2865–2873
Kann P, Piepkorn B, Schehler B, Andreas J, Lotz J, Prellwitz W, Beyer J 1998
Effect of long-term treatment with GH on bone metabolism, bone mineral
density and bone elasticity in GH-deficient adults. Clin Endocrinol (Oxf)
48:561–568
Kaufman JM, Taelman P, Vermeulen A, Vandeweghe M 1992 Bone mineral
status in growth hormone-deficient males with isolated and multiple pituitary
deficiencies of childhood onset. J Clin Endocrinol Metab 74:118 –123
Valimaki MJ, Salmela PI, Salmi J, Viikari J, Kataja M, Turunen H, Soppi E
1999 Effects of 42 months of GH treatment on bone mineral density and bone
turnover in GH-deficient adults. Eur J Endocrinol 140:545–554
Rutherford OM, Jones DA, Round JM, Preece MA 1989 Changes in skeletal
muscle after discontinuation of growth hormone treatment in young adults
with hypopituitarism. Acta Paediatr Scand 356(Suppl):61– 63
Tanner JM, Whitehouse RH 1967 The effect of human growth hormone on
subcutaneous fat thickness in hyposomatotrophic and panhypopituitary
dwarfs. J Endocrinol 39:263–275
Baroncelli GI, Bertelloni S, Ceccarelli C, Cupelli D, Saggese G 2000 Dynamics of bone turnover in children with GH deficiency treated with GH until
final height. Eur J Endocrinol 142:549 –556
Fujimoto S, Kubo T, Tanaka H, Miura M, Seino Y 1995 Urinary pyridinoline
and deoxypyridinoline in healthy children and in children with growth hormone deficiency. J Clin Endocrinol Metab 80:1922–1928
Hansen TB, Brixen K, Vahl N, Jørgensen JOL, Christiansen JS, Mosekilde
L, Hagen C 1996 Effects of 12 months of growth hormone (GH) treatment on
5272 J Clin Endocrinol Metab, November 2003, 88(11):5266 –5272
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
calciotropic hormones, calcium homeostasis, and bone metabolism in adults
with acquired GH deficiency: a double blind, randomized, placebo-controlled
study. J Clin Endocrinol Metab 81:3352–3359
Rodriguez-Arnao J, James I, Jabbar A, Trainer PJ, Perrett D, Besser GM, Ross
RJM 1998 Serum collagen crosslinks as markers of bone turn-over during GH
replacement therapy in growth hormone deficient adults. Clin Endocrinol
(Oxf) 48:455– 462
Stracke H, Schulz A, Moeller D, Rossol S, Schatz H 1984 Effect of growth
hormone on osteoblasts and demonstration of somatomedin-C/IGF I in bone
organ culture. Acta Endocrinol (Copenh) 107:16 –24
Schönau E 1998 Problems of bone analysis in childhood and adolescence.
Pediatr Nephrol 12:420 – 429
Schönau E, Neu CM, Rauch F, Manz F 2001 The development of bone strength
at the proximal radius during childhood and adolescence. J Clin Endocrinol
Metab 86:613– 618
Ranke MB, Schweizer R, Elmlinger MW, Weber K, Binder G, Schwarze CP,
Wollmann HA 2000 Significance of basal IGF-I, IGFBP-3 and IGFBP-2 measurements in the diagnostics of short stature in children. Horm Res 54:60 – 68
Neu CM, Rauch F, Manz F, Schönau E 2002 Modeling of cross-sectional bone
size, mass and geometry at the proximal radius: a study of normal bone
development using peripheral quantitative computed tomography. Osteoporos Int 12:538 –547
Elmlinger M, Gässler N, Weber M 1999 Detection of free deoxypyridinoline
(DPD) in urines of healthy individuals and patients with osteoporosis and
osteogenesis imperfecta by an automated chemiluminescence immunoassay.
Clin Lab 45:423– 428
Schönau E, Rauch F 1997 Markers of bone and collagen metabolism-problems
and perspectives in paediatrics. Horm Res 48(Suppl 5):50 –59
Lockitch G, Halstead AC, Albersheim S, MacCallum C, Quigley G 1988 Ageand sex-specific pediatric reference intervals for biochemistry analytes as measured with the Ektachem-700 analyzer. Clin Chem 34:1622–1625
Bianda T, Glatz Y, Bouillon R, Froesch ER, Schmid C 1998 Effects of short-
Schweizer et al. • Bone Density Initially Decreases during GH Replacement
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
term insulin-like growth factor-I (IGF-I) or growth hormone (GH) treatment
on bone metabolism and on production of 1,25-dihydroxycholecalciferol in
GH-deficient adults. J Clin Endocrinol Metab 83:81– 87
Langdahl BL, Kassem M, Moller MK, Eriksen EF 1998 The effects of IGF-I and
IGF-II on proliferation and differentiation of human osteoblasts and interactions with growth hormone. Eur J Clin Invest 28:176 –183
Laron Z, Klinger B 2000 Comparison of the growth-promoting effects of
insulin-like growth factor I and growth hormone in the early years of life. Acta
Paediatr 89:38 – 41
Homminga J, Weinans H, Gowin W, Felsenberg D, Huiskes R 2001 Osteoporosis changes the amount of vertebral trabecular bone at risk of fracture but
not the vertebral load distribution. Spine 26:1555–1561
Lin JC, Grampp S, Link T, Kothari M, Newitt DC, Felsenberg D, Majumdar
S 1999 Fractal analysis of proximal femur radiographs: correlation with biomechanical properties and bone mineral density. Osteoporos Int 9:516 –524
Wüster C 2000 Fracture rates in patients with growth hormone deficiency.
Horm Res 54(Suppl 1):31–35
Bex M, Abs R, Maiter D, Beckers A, Lamberigts G, Bouillon R 2002 The effects
of growth hormone replacement therapy on bone metabolism in adult-onset
growth hormone deficiency: a 2-year open randomized controlled multicenter
trial. J Bone Miner Res 17:1081–1094
Clayton PE, Cowell CT 2000 Safety issues in children and adolescents during
growth hormone therapy: a review. Growth Horm IGF Res 10:306 –317
Maneatis T, Baptista J, Connelly K, Blethen S 2000 Growth hormone safety
update from the National Cooperative Growth Study. J Pediatr Endocrinol
Metab 13(Suppl 2):1035–1044
Schönau E, Neu CM, Mokov G, Wassmer G, Manz F 2000 Influence of puberty
on muscle area and cortical bone area of the forearm in boys and girls. J Clin
Endocrinol Metab 85:1095–1098
Greulich WW, Pyle SI 1959 Radiographic atlas of skeletal development of the
hand and wrist. 2nd ed. Stanford, CA: Stanford University Press