ORIGINAL
ARTICLE
E n d o c r i n e
C a r e
Constitutional Thinness: Unusual Human Phenotype
of Low Bone Quality
Bogdan Galusca, Mohamed Zouch, Natacha Germain, Cecile Bossu, Delphine Frere, Francois Lang,
Marie-Helene Lafage-Proust, Thierry Thomas, Laurence Vico, and Bruno Estour
Departments of Endocrinology (B.G., N.G., C.B., B.E.) and Psychiatry (F.L.), and Nuclear Medicine Laboratory (D.F.), Centre Hospitalier
Universitaire Saint Etienne, 42055 Saint Etienne, France; Institut National de la Santé et de la Recherche Médicale U890 (M.Z., M.-H.L.-P., T.T.,
L.V.), F-42023 Saint Etienne, France; and Endocrinology Department (B.G.), Universitatea de Medicina si Farmacie, 700229 Iasi, Romania
Context: Low fat mass and hormonal or nutritional deficiencies are often incriminated in bone loss
related to thinness. Constitutional thinness has been described in young women with low body
mass index (BMI) but close-to-normal body composition, physiological menstruation, no hormonal
abnormalities, and no anorexia nervosa (AN) psychological profile.
Objective: Our objective was to determine whether constitutional thinness is associated with
impaired bone quality.
Design, Setting, and Participants: This was an observational, cross-sectional study on 25 constitutionally
thin and 44 AN young women with similar low BMI (⬍16.5 kg/m2) and 28 age-matched controls.
Main Outcome Measures: Femoral and lumbar spine bone mineral density by dual-energy x-ray absorptiometry, distal tibia and radius bone architecture and breaking strength by three-dimensional
peripheral quantitative computed tomography, and bone turnover markers were determined.
Results: Constitutionally thin subjects displayed a higher percentage of fat mass than AN subjects but
had similar lumbar and femoral bone mineral density, which were significantly lower than in controls
(P ⬍ 0.001). Constitutionally thin subjects displayed more markedly impaired trabecular and cortical
bone parameters in the distal tibia than in the radius. AN bone structure was impaired only in subjects
with a long history of disease. Calculated breaking strength was decreased in constitutional thinness
and long-standing AN in both the radius and the tibia. Bone markers in constitutionally thin subjects
were similar to those of controls. Osteoprotegerin to receptor activator of nuclear factor ␬ B ligand ratio
was higher in constitutionally thin subjects than in controls or AN women.
Conclusions: Young women with constitutional thinness present an unexpectedly high prevalence of
low bone mass (44%) associated with small bone size, overall diminished breaking strength, but normal
bone turnover. Mechanisms related to insufficient skeletal load and/or genetics are proposed to explain
this new phenotype of impaired bone quality. (J Clin Endocrinol Metab 93: 110 –117, 2008)
B
one mass and density are determined by various concurrent
factors such as genetics, age, hormone levels, physical activity, nutrition, body composition, and body weight (1).
In middle-aged and elderly patients, bone loss has been
related to thinness, mainly in a context of low fat mass (FM)
(2). Consumptive (3), infectious (4), digestive (5), or chronic
0021-972X/08/$15.00/0
Abbreviations: AN, Anorexia nervosa; bALP, bone alkaline phosphatase; BMD, bone mineral density; BMI, body mass index; BV/TV, relative bone volume as a percentage of total volume, derived
from trabecular density; CT, constitutional thinness; CTh, absolute thickness of cortical bone; CV,
coefficient of variation; 3D-pQCT, three-dimensional peripheral quantitative computed tomography; D100, mean whole bone (cortical and trabecular) density; Dcomp, bone density of cortical
bone; DHEAS, sulfate salt of dehydroepiandrosterone; Dinn, density of the central part of trabecular
bone;Dmeta,densityofthesubcorticalareaoftrabecularbone;Dtrab,densityofthetrabeculararea
of bone; DXA, dual-energy x-ray absorptiometry; FFM, fat-free mass; FM, fat mass; HA, hydroxyapatite; MOI, moment of inertia; OPG, osteoprotegerin; RANKL, receptor activator of nuclear factor
␬B ligand; sCTX, serum cross-linked C telopeptides of collagen type I; TbN, absolute trabecular
numberpermm3;TbSp,meantrabecularseparation;TRACP5b,tartrate-resistantacidphosphatase
type 5b.
Printed in U.S.A.
Copyright © 2008 by The Endocrine Society
doi: 10.1210/jc.2007-1591 Received July 17, 2007. Accepted October 12, 2007.
First Published Online October 23, 2007
110
jcem.endojournals.org
J Clin Endocrinol Metab. January 2008, 93(1):110 –117
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
diseases or cachexia (6) inducing weight loss has also been
incriminated.
Low bone mass related to low body weight has also been
reported in young women, mainly in a context of anorexia nervosa (AN). More than 50% of young AN women present bone
loss at both trabecular and cortical bone sites (7, 8 ), and fractures are reported to be increased in this population (9). An
intricate network of multiple hormonal and nutritional abnormalities associated with low FM that impair osteoblast and/or
osteoclast activities has been incriminated (7, 10, 11).
In the same age range as AN, a nonpathological state called
constitutional thinness (CT) or leanness has been poorly described (12). These young women are severely thin but continue
to have a close-to-normal FM percentage, normal physiological
menstrual cycles, and no detectable abnormalities of cortisol,
IGF-I, or free T3 secretory patterns (13), and normal energy
metabolism (14). In this context, we wanted to determine
whether constitutional thinness could still represent a phenotype
of impaired bone quality, in terms of bone mass, bone macroarchitecture and microarchitecture, and bone turnover.
Subjects and Methods
This study was approved by the human research ethics committee of
Centre Hospitalier Universitaire, Saint Etienne, France, and all subjects
gave their written informed consent.
Subjects
Three groups of young age-matched Caucasian female subjects
(18 –30 yr) were recruited for the study: 25 constitutionally thin subjects,
44 AN subjects, and 28 controls. The constitutionally thin and AN subjects were matched by body mass index (BMI).
The 25 constitutionally thin subjects were recruited at our outpatient
clinic among the patients evaluated for leanness, referred by internists
after exclusion of celiac disease, infectious diseases, cancer, or other
consumptive diseases. Inclusion criteria were as follows: BMI between
12.0 and 16.5 kg/m2, called “severe underweight” according to the
World Health Organization classification (15), stable throughout the
growth period until the age of 18 (14); presence of physiological menstruation without estrogen-progestin therapy; and desire for weight gain
as the main reason for medical consultation.
All 44 AN subjects displayed active psychiatric illness according to
the criteria of the Diagnostic and Statistical Manual of Mental Disorders
(16). All patients presented restrictive-type AN (no binge/purge crises).
Before disease, the subjects’ BMI ranged between 19.1 and 26.0 kg/m2.
None of these patients used oral contraceptives, and all presented secondary amenorrhea for more than 6 months. Secondary amenorrhea
occurred at least 12 months after the first menstrual period. To evaluate
the effect of duration of undernutrition on bone parameters, two subgroups of AN were distinguished using a cutoff point of 24-month disease duration (17): 23 recently diagnosed patients, 21 patients with a
long history of disease.
A total of 28 medical students were recruited as normal-weight
controls.
In constitutionally thin subjects and controls, all data were collected
during the follicular phase of the menstrual cycle.
None of the subjects included in this study were documented to have
any form of addictive or abusive consumption (alcohol, smoking, drugs,
or physical activity, etc.). None of these subjects were taking calcium,
vitamin supplements, or any other medications.
jcem.endojournals.org
111
Bone mineral density (BMD) and body composition
Femoral neck and lumbar spine (L1–L4) BMD were measured by
dual-energy x-ray absorptiometry (DXA) (Delphi W, Hologic, Inc.,
Waltham, MA). The SD value for lumbar bone density measurement is
0.01 g/cm2 and does not vary with bone density. Z score was calculated
using a manufacturer-supplied reference data set of healthy young adult
female BMD values. According to the International Society for Clinical
Densitometry official position concerning premenopausal women (18),
Z scores less than ⫺2.0 were classified as “below expected range for
age.”
Fat and lean body mass were determined using the same DXA device.
Bone evaluation by three-dimensional peripheral
quantitative computed tomography (3D-pQCT)
Multislice 3D-pQCT was performed on the distal radius and distal
tibia of the nondominant forearm and leg, respectively (XtremeCT;
Scanco Medica AG Bassersdorf, Switzerland). The XtremeCT system
uses a two-dimensional detector array in combination with a 0.08-mm
point-focus x-ray tube, enabling simultaneous acquisition of a stack of
parallel computed tomography slices with a nominal resolution (voxel
size) of 82 ␮m.
Methods used to process computed tomography data have been previously described in detail (19). The following parameters were considered: D100, mean whole bone (cortical and trabecular) density in grams
hydroxyapatite (HA) equivalence per cm3; Dtrab, density of the trabecular area of bone (g HA/cm3); Dmeta, density of the subcortical area of
trabecular bone (g HA/cm3); Dinn, density of the central part of trabecular bone (g HA/cm3); BV/TV, relative bone volume as a percentage of
total volume, derived from trabecular density; TbN, absolute trabecular
number per mm3; mean trabecular thickness (mm); TbSp, mean trabecular separation (mm); Dcomp, bone density of cortical bone (g HA/cm3);
CTh, absolute thickness of cortical bone (mm); and mean cross-sectional
area of slices (mm2). Short-term coefficients of variation (CVs) of densitometric and structural parameters were evaluated from two repeated
measurements on nine women 25– 40 yr old within 1 wk, and ranged
between 0.34 and 4.3%.
Parameters of total and cortical bone strength at both radius and tibia
level were calculated using additional software in accordance with previous descriptions (20 –22): Ixx, Iyy, Imax, Imin ⫺ moment of inertia
(MOI) around x-axis, y-axis, shorter axis, and longer axis, respectively.
Section modulus was calculated (S ⫽ I/C, where I ⫽ MOI about an axis
through the centroid, and C is the distance from the centroid to the
extreme edge of the section), which relates to the breaking strength of the
bone: Ixx/Cy ⫺ MOI around x-axis divided by maximum extent in y
direction; Iyy/Cx; Imax/Cmax ⫺ MOI around new y-axis divided by
maximum extent in x direction; and Imin/Cmin.
Hormonal and bone marker study
Venous blood samples were collected on dry glass tubes containing
EDTA and centrifuged, and plasma was aliquoted and kept frozen at
⫺80 C before the assay. After an overnight fast, blood was obtained at
0800 h for measurement of serum leptin, GH, IGF-I, cortisol, 17␤estradiol, sulfate salt of dehydroepiandrosterone (DHEAS), SHBG, total
testosterone, free T3, PTH, and 25-hydroxy vitamin D3. Samples were
collected every 4 h for a period of 24 h to measure bone markers.
The following techniques were used to measure: leptin, RIA [Nichols
Institute Diagnostics, San Juan Capistrano, CA; manufacturer’s reference range for a normal BMI 3.7–11.1 ␮g/liter (18 –25)]; GH, IRMA
(Beckman Coulter, Inc., Fullerton, CA; manufacturer’s reference level ⬍
5 mU/liter); IGF-I, IRMA (Beckman Coulter; intraassay and interassay
CVs, 7 and 16% respectively, manufacturer’s reference range 107–310
␮g/liter); plasma cortisol, RIA (Immunotech, Marseille, France; intraassay and interassay CVs, 7 and 8%, respectively, detection limit 10
nmol/24 h, normal range 107–310 ng/ml); 17␤-estradiol, RIA (Dia Sorin,
Antony, France; manufacturer’s reference range during the follicular
phase 30 –50 ng/liter); DHEAS, RIA (Beckman Coulter; normal range
85–225 ␮g/dl); SHBG, IRMA (bioMérieux sa, Marcy l’Etoile, France;
112
Galusca et al.
Bone Quality in Constitutional Thinness
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
TABLE 1. Body weight composition, BMD, nutritional and food
intake parameters, and several hormones (mean levels ⫾ SEM)
in AN, CT, and controls
TABLE 2. Linear regression between femoral neck and lumbar
spine BMD and BMI, body composition parameters, leptin
Femoral neck BMD
AN (n ⴝ 44)
Anthropometry and
body composition
Age (yr)
Height (m)
BMI (kg/m2)
FM %
BMD
Femoral neck BMD
(g/cm2)
Lumbar spine BMD
(g/cm2)
Hormonal parameters
Leptin (␮g/liter)
GH (mIU/liter)
IGF-I (␮g/liter)
Cortisol (ng/liter)
17␤-estradiol (ng/
liter)
DHEAS (␮g/liter)
SHBG (nmol/liter)
Free testosterone
index
Free T3 (pmol/liter)
PTH (ng/ml)
25-hydroxy vitamin
D3 (␮g/liter)
Bone markers
Osteocalcin (␮g/ml)
bALP (␮g/ml)
sCTX (pmol/liter)
TRACP 5b (U/liter)
OPG (pmol/liter)
RANKL (pmol/liter)
23.4 ⫾ 1.2
1.62 ⫾ 0.1
15.5 ⫾ 0.1a
9.8 ⫾ 1.1a
CT (n ⴝ 25)
Controls
(n ⴝ 28)
23.1 ⫾ 1.2
1.63 ⫾ 0.02
15.8 ⫾ 0.1a
18.6 ⫾ 0.7a,b
23.9 ⫾ 1.4
1.63 ⫾ 0.01
20.7 ⫾ 0.4
26.3 ⫾ 1.2
0.795 ⫾ 0.03a
0.809 ⫾ 0.02a
0.951 ⫾ 0.02
0.849 ⫾ 0.03
0.873 ⫾ 0.02
a
0.986 ⫾ 0.03
6.0 ⫾ 0.8a,b
4.8 ⫾ 0.6b
295 ⫾ 34b
216 ⫾ 12b
73.1 ⫾ 8.6b
11.2 ⫾ 1.9
4.7 ⫾ 0.6
283 ⫾ 20
266 ⫾ 17
51.6 ⫾ 11.4
a
2.4 ⫾ 0.5a
8.5 ⫾ 0.7a
163 ⫾ 16a
364 ⫾ 31a
14.3 ⫾ 1.4a
196.2 ⫾ 30a
81.4 ⫾ 21.5
0.7 ⫾ 0.1
112.6 ⫾ 19b
82.9 ⫾ 11.7
0.6 ⫾ 0.1
129.4 ⫾ 17
101.5 ⫾ 23.2
0.7 ⫾ 0.1
2.7 ⫾ 0.1a
29.1 ⫾ 2.9
23.0 ⫾ 2.4
3.9 ⫾ 0.1b
35.8 ⫾ 3.6
22.7 ⫾ 4.2
3.5 ⫾ 0.1
34.3 ⫾ 7.1
26.7 ⫾ 3.7
10.3 ⫾ 0.3a
8.1 ⫾ 0.3
9021 ⫾ 1070a
2.55 ⫾ 0.1a
4.8 ⫾ 0.1a
1.18 ⫾ 0.1
16.4 ⫾ 0.6b
7.8 ⫾ 0.5
4345 ⫾ 752b
0.93 ⫾ 0.1b
3.4 ⫾ 0.1b
0.54 ⫾ 0.06a,b
14.9 ⫾ 0.6
8.2 ⫾ 0.5
4073 ⫾ 433
1.15 ⫾ 0.1
3.6 ⫾ 0.1
1.19 ⫾ 0.1
a
Significantly different vs. control group (P ⬍ 0.05).
b
Significantly different vs. AN group (P ⬍ 0.05).
manufacturer’s reference range 30 –100 nmol/liter); testosterone, RIA
[Beckman Coulter; extraction and chromatography, manufacturer’s reference range 7– 65 ng/dl, calculated free testosterone index (⫽ testosterone/SHBG)]; free T3, RIA (Beckman Coulter; manufacturer’s reference
range 2.5–5.8 pmol/liter); intact PTH, Allegro IRMA diagnostic kits
(Nichols Institute Diagnostics; manufacturer’s reference range 10 – 65
ng/ml); and vitamin D (25 OHD3), kit RIA Incstar (Incstar Corp., Stillwater MN; manufacturer’s reference range 10 – 44 ␮g/liter, calcium).
Elsa Ost-Nat (Cis Bio Intl., Gif-sur-Yvette, France), a two-site assay,
was used to quantify intact osteocalcin. Intraassay and interassay CVs
were less than 5% (23, 24). The manufacturer’s mean and reference
range for females 21–30 yr were 13.4 and 7.6 –25.6 ␮g/ml, respectively.
Bone alkaline phosphatase (bALP) was measured using Tandem-R
Ostase two-site Immunoradiometric Assay (Beckman Coulter). The interassay CV was less than 8%. The manufacturer’s mean ⫾ SD for premenopausal women was 8.7 ⫾ 2.7 ␮g/ml.
Serum cross-linked C telopeptide of collagen type I (sCTX) assessment is a competitive immunoassay (CrossLaps-S Elisa; Osteometer A/S,
Rodovre, Denmark). The crosslaps antigen coated on the microwell is a
synthetic peptide with an amino acid sequence specific for the C telopeptide of the ␣1 chain of type 1 collagen (Glu-Lys-Ala-His-Asp-Gly-GlyArg: crosslaps antigen) (25). The intraassay and interassay CVs were less
than 5 and 8.1%, respectively, and the detection limit was 92 pmol/liter.
The manufacturer’s reference range in premenopausal females was 7887579 pmol/liter.
Tartrate-resistant acid phosphatase type 5b (TRACP 5b) activity was
detected using BoneTRAP assay (Immunodiagnostic Systems, Tyne and
BMI
FM %
FFM %
Leptin
Lumbar spine BMD
CT ⴙ
controls
AN ⴙ
controls
CT ⴙ
controls
AN ⴙ
controls
0.56
0.41a
0.38a
0.23a
0.44
0.43
0.33a
0.37
0.53
0.42a
0.42a
0.29a
0.50
0.43
0.37a
0.41
a
When Bonferroni correction was applied, the correlation did not reach the
adjusted ␣-level of significance (P ⬍ 0.002).
Wear, UK). Intraassay and interassay CVs were less than 12.3 and
13.9%, respectively, and the detection limit was 1 U/liter. The manufacturer’s reference range in healthy premenopausal females was 1.03–
4.15 U/liter.
Osteoprotegerin (OPG) was measured by an OPG ELISA kit (Immunodiagnostic Systems). Intraassay and interassay CVs were less than 10
and 7%, respectively, and the detection limit was 0.14 pmol/liter.
Soluble receptor activator of nuclear factor ␬B ligand (RANKL) measurements were performed with a Biomedica Medizinprodukte GmbH
(Vienna, Austria) prototype (ELISA format). Interassay and intraassay
CVs were 7–10% and less than 7%, respectively.
Statistical analysis
All values are presented as mean ⫾ SEM. The principal components analysis multivariate analysis technique using an Orthotran/Varimax transformation was performed to extract and rotate factors to best represent the 18
variables measured or calculated by 3D-pQCT. ANOVA was used to perform a three-group analysis comparing constitutionally thin women with
anorexic women and controls. When ANOVA was significant, post hoc
ANOVA tests (Tukey-Kramer) were performed for comparisons within all
groups. The limit of statistical significance was set at P ⬍ 0.05. Analysis was
performed with and without subject height as covariate for architectural
bone parameters measured by 3D-pQCT. Linear regression was performed
to evaluate several interactions within the following pooled groups: constitutionally thin ⫹ controls, AN ⫹ controls. This pooling was decided because
a nonlinear relationship was found in the overall group for several parameters. Bonferroni’s correction was applied to compensate for multiple tests
per variable (26). All statistical analyses were performed with StatView
software (SAS Institute Inc., Cary, NC).
Results
Clinical and hormonal parameters (Table 1)
Constitutionally thin women with a very low BMI (15.8 ⫾ 0.1
kg/m2) not significantly different from that of AN subjects
(15.5 ⫾ 0.1 kg/m2) were included in this study. In this context of
severe underweight, constitutionally thin subjects displayed a
higher percentage of FM than AN subjects (18.6 ⫾ 0.7 vs. 9.8 ⫾
1.1; P ⬍ 0.001), and implicitly had a lower percentage of fat-free
mass (FFM). Percentage values of FM were significantly lower in
constitutionally thin subjects than in controls. CT leptin levels
were intermediate between AN subjects and controls. No significant difference between constitutionally thin subjects and
controls was observed for the other hormones. GH, IGF-I, cortisol, 17␤-estradiol, free T3, and DHEAS were significantly different in the AN group compared with both constitutionally thin
subjects and controls. No difference in serum calcium levels between the three groups was observed.
ANOVA with repeated measures followed by post hoc tests was used to evaluate intergroup differences. TbTh, Mean trabecular thickness.
a,b
a,b
* P ⬍ 0.05 in: a, long history AN vs. controls; b, AN recently diagnosed vs. long history AN; c, CT vs. controls; d, CT vs. long history AN; and e, CT vs. AN recently diagnosed.
a, c
a,d
a, c
305.9 ⫾ 9.3
167.8 ⫾ 8.2
231.6 ⫾ 8.0
128.1 ⫾ 8.0
0.140 ⫾ 0.006
1.74 ⫾ 0.06
0.088 ⫾ 0.002
0.50 ⫾ 0.02
909.2 ⫾ 6.0
1.17 ⫾ 0.04
276.0 ⫾ 11.5
144.7 ⫾ 9.6
212.6 ⫾ 9.1
100.1 ⫾ 10.7
0.121 ⫾ 0.008
1.46 ⫾ 0.07
0.082 ⫾ 0.004
0.62 ⫾ 0.04
907.9 ⫾ 7.7
1.02 ⫾ 0.04
301.0 ⫾ 10.0
174.2 ⫾ 8.6
238.4 ⫾ 9.6
134.0 ⫾ 8.3
0.145 ⫾ 0.009
1.65 ⫾ 0.08
0.092 ⫾ 0.003
0.53 ⫾ 0.03
894.0 ⫾ 7.6
1.09 ⫾ 0.03
a,b
b
a,b
b
b
a,b
356.7 ⫾ 12.9
153.6 ⫾ 7.1
217.8 ⫾ 7.3
109.1 ⫾ 7.1
0.128 ⫾ 0.006
1.72 ⫾ 0.04
0.078 ⫾ 0.002
0.51 ⫾ 0.01
912.6 ⫾ 6.1
0.93 ⫾ 0.02
355.0 ⫾ 12.8
147.5 ⫾ 9.2
210.5 ⫾ 8.9
104.0 ⫾ 9.8
0.123 ⫾ 0.006
1.60 ⫾ 0.07
0.076 ⫾ 0.004
0.56 ⫾ 0.04
906.0 ⫾ 10.0
0.88 ⫾ 0.03
318.7 ⫾ 13.6
131.4 ⫾ 12.3
190.1 ⫾ 11.7
90.2 ⫾ 12.8
0.110 ⫾ 0.008
1.40 ⫾ 0.09
0.079 ⫾ 0.004
0.81 ⫾ 0.06
908.7 ⫾ 9.7
0.81 ⫾ 0.03
355.7 ⫾ 13.5
161.6 ⫾ 12.3
218.1 ⫾ 11.1
121.2 ⫾ 12.5
0.134 ⫾ 0.009
1.65 ⫾ 0.06
0.082 ⫾ 0.003
0.53 ⫾ 0.04
895.7 ⫾ 10.4
0.91 ⫾ 0.04
D100 (gHA/cm3)
Dtrab (gHA/cm3)
Dmeta (gHA/cm3)
Dinn (gHA/cm3)
BV/TV
TbN
TbTh (mm)
TbSp (mm)
Dcomp (gHA/cm3)
CTh (mm)
Long history
AN
113
BMD (Table 1)
Constitutionally thin subjects had markedly lower BMD than
the control group, in both the femoral neck and lumbar spine
(P ⬍ 0.001). Similar low BMDs were found in lean women with
AN. Up to 44% of constitutionally thin subjects and 50% of AN
patients presented low bone mass (Z score ⬍ ⫺2.0). None of the
controls presented a Z score below the expected range for age.
When pooling constitutionally thin and controls, lumbar
spine and femoral neck BMD were positively correlated with
BMI, but not with body composition parameters (FM and FFM)
or leptin. Inversely, when pooling AN and controls, BMD was
correlated with BMI, FM, and leptin (Table 2).
277.2 ⫾ 12.2
155.7 ⫾ 12.5
217.5 ⫾ 12.2
118.0 ⫾ 12.0
0.129 ⫾ 0.010
1.49 ⫾ 0.07
0.087 ⫾ 0.004
0.61 ⫾ 0.03
884.4 ⫾ 6.6
1.01 ⫾ 0.04
Control
CT
a, c
c,e
c,e
c,e
c,e
a, c
jcem.endojournals.org
AN recently
diagnosed
CT
Control
P value*
AN recently
diagnosed
Long history
AN
Distal tibia
Distal radius
TABLE 3. Trabecular and cortical 3D-pQCT parameters of distal radius and distal tibia (mean levels ⫾ SEM) in subjects with CT, AN patients, and controls
P
value*
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
Bone evaluation by 3D-pQCT (Table 3 and Fig. 1)
Principal components analysis extracted three rotated factors
for both radius and tibia measurements, characterizing: 1) breaking strength (grouping Ixx, Iyy, Imax, Imin, Ixx/Cy, Iyy/Cx,
Imax/Cmax, and Imin/Cmin); 2) trabecular density and structure (grouping D100, Dtrab, Dmeta, Dinn, BV/TV, TbN, and
TbSp); and 3) cortical density and structure (grouping Dcomp
and CTh).
The CT group displayed significantly different trabecular
bone parameters compared with controls in the distal tibia, but
not in the radius: lower bone density (Dtrab), significantly lower
trabecular number (TbN), and larger trabecular separation
(TbSp). Cortical bone in the distal tibia presented reduced thickness (CTh) but normal density (Dcomp) compared with controls.
The distal radius and tibial cross-sectional areas were significantly lower in the constitutionally thin group than in controls or
AN subjects.
Only those AN subjects with a long history of disease displayed impaired cortical and trabecular bone density and microarchitecture. These abnormalities were detected in both the
distal tibia and the radius.
Parameters reflecting the resistance of a bone (total and cortical) to stress (Ixx, Iyy, Imax, Imin, Ixx/Cy, Iyy/Cx, Imax/Cmax,
and Imin/min) were decreased in both the radius and the tibia in
constitutionally thin and long-standing AN subjects.
Adjustment of architectural bone parameters for subject
height did not modify the significance of these differences.
In the AN group, age and disease duration were negatively
correlated with BMD, measured by DXA or 3D-pQCT, and
with some bone structure parameters, whereas no correlation
was found between age and these parameters in the CT group
(Table 4).
Bone remodeling markers
Circadian profiles of the measured bone markers are illustrated in Fig. 2, with mean values presented in Table 1. Osteocalcin, sCTX, and TRACP 5b were not significantly different
between the CT group and controls. Conversely, AN subjects
displayed lower levels of osteocalcin and higher levels of sCTX
and TRACP 5b than controls or constitutionally thin subjects.
bALP was not significantly different between groups.
A strong positive correlation was found between osteocalcin
and sCTX in constitutionally thin subjects (r ⫽ 0.58; P ⬍ 0.001)
114
Galusca et al.
Bone Quality in Constitutional Thinness
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
FIG. 1. 3D-pQCT image reconstruction of distal tibia in controls (A) and constitutionally thin women (B). Lower panel, Cross-sectional area measured by 3D-pQCT and
calculated breaking strength of distal radius and distal tibia in recently diagnosed AN, AN with long history of the disease, constitutionally thin (CT) subjects, and
controls (C). The cross-sectional area (1) and cortical thickness (2) are shown in A. Lower panel, Bars indicated by asterisk (*) are significantly different from controls (Cs)
at P ⬍ 0.05, based on ANOVA and Tukey’s test as post hoc test.
and controls (r ⫽ 0.46; P ⬍ 0.001), whereas no correlation was
observed between these two markers in the AN group.
RANKL was significantly lower in constitutionally thin subjects compared with controls or AN subjects. OPG was significantly higher in AN subjects (Fig. 2), and the OPG to RANKL
ratio was higher in constitutionally thin subjects than in controls
and AN subjects (P ⬍ 0.001).
No significant differences for hormonal parameters and bone
markers were observed between subjects with newly diagnosed
AN and those with long-standing AN.
Discussion
In young women, low bone density related to low body weight
was described only in patients with AN.
Our study evaluated bone characteristics in another population of young women, called constitutional thinness, presenting
the same BMI less than 16.5 kg/m2 with no identified eating
disorder. The nonpathological state of constitutional thinness is
supported by the absence of Diagnostic and Statistical Manual of
Mental Disorders, fourth edition criteria for AN, normal menstruation (27), normal hormonal profile (13), normal levels of sc
adipose tissue (28) and, as we have recently described, a normal
energy metabolism (14). This entity was initially described in
African and Australian aboriginal populations (28, 29), whereas
no estimate of prevalence has been performed in Western populations. No bone quality data have previously been published
on young women with constitutional thinness.
In the present study, in a well-selected population of young
women with constitutional thinness, we report an unexpectedly
high prevalence of low bone mass (44%), characterized by four
aspects: normal bone turnover, high OPG-to-RANKL ratio,
small bone size, and selective bone structural impairment predominantly affecting weight-bearing skeletal regions. In our AN
population, we observed the same prevalence of low bone density but differences in terms of bone turnover, bone size, and bone
architecture. The CT impaired bone quality model has not been
previously reported.
In young women with constitutional thinness, the normal
bone remodeling markers and a strong correlation between bone
formation and bone resorption parameters similar to controls
suggest that the uncoupling of bone turnover observed in AN
(10) does not characterize their low bone density. The coupling
of bone turnover seen in constitutional thinness is in accordance
with the normal hormonal status observed in these young
women.
Over recent years it has become clear that RANKL and OPG
are essential determinants of osteoclast cell biology and bone
resorption. An increase in the OPG-to-RANKL ratio normally
TABLE 4. Correlation between bone parameters and age (in AN and CT groups) and disease duration (AN group)
Age (yr)
CT
AN
AN duration (months)
a
Lumbar spine
BMD
Femoral neck
BMD
Dtrab
(tibia/radius)
TbN
(tibia/radius)
TbSp
(tibia/radius)
⫺0.15a
⫺0.38
⫺0.33
⫺0.11a
⫺0.33
⫺0.36
⫺0.21a/⫺0.23a
⫺0.49/⫺0.45
⫺0.52/⫺0.25a
⫺0.10a/⫺0.15a
⫺0.56/⫺0.50
⫺0.65/⫺0.46
0.27a/0.20a
0.52/0.59
0.62/0.53
When Bonferroni correction was applied, the correlation did not reach the adjusted ␣-level of significance (P ⬍ 0.002).
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
jcem.endojournals.org
115
present study did not demonstrate any difference for these hormones between the constitutionally thin group and age-matched
controls, despite the very low range of body
weight (BMI 12.0 –16.5 kg/m2). It is well
known that low bone density in AN is
strongly related to severe FM reduction and
leptin deficiency (11). In constitutional thinness, leptin is significantly reduced compared with controls and could partly explain the low bone density observed. On the
other hand, leptin interacts with and stimulates GnRH neurons. A leptin threshold
has been proposed as the metabolic gate to
gonadotropin secretion in AN (36). In constitutional thinness, gonadal function is preserved, providing evidence that leptin acts
normally in these subjects despite lower levels than in normal weight subjects. By analogy, we postulate that leptin is able to stimFIG. 2. Circadian profile of plasma osteocalcin (top left panel), sCTX (bottom left panel), OPG (top right
ulate bone formation in constitutional
panel), and RANKL (bottom right panel) in constitutionally thin women (CT) compared with AN subjects
thinness. In the present study, leptin was not
and controls (C). ANOVA with repeated measures followed by post hoc tests was used to evaluate
intergroup differences.
correlated with BMD when constitutionally
thin subjects and controls were pooled.
Some authors have proposed that FFM is a
regulator of bone mass and bone metabolism (37). However, no
inhibits osteoclastogenesis, slows down bone resorption, and
strong correlation was observed in our study between FFM and
reduces bone loss (30). Although interpretation of serum levels
bone parameters after application of the Bonferroni correction.
of these cytokines has not been fully elucidated, the increased
Because no obvious hormonal or body composition causality
OPG-to-RANKL ratio in constitutional thinness could be conwas found, we propose that constitutionally thin bone impairsidered effective because bone resorption markers, sCTX and
ment may be caused by an insufficient skeletal load and/or geTRACP 5b, were situated in the same range as in controls. The
netic factors.
high levels of OPG found in AN patients confirm recently pubThe correlation between BMD and BMI in constitutionally
lished data (31) that are difficult to interpret in a context of
thin women suggests a possible direct causality of low body
increased bone resorption rate. In addition, we report for the first
weight on bone density, structure, or strength, as demonstrated
time that RANKL levels and OPG-to-RANKL ratio are normal
in microgravity models with insufficient skeletal load (38, 39).
in AN subjects.
The effect of mechanical unloading may be cumulative in conBone structure evaluated by 3D-pQCT was altered in constitutionally thin patients because they have a constant low
stitutionally thin subjects mainly in the distal tibia, a bone region
weight during the growth period (12, 14, 40). This hypothesis is
affected by weight bearing. Only one study has previously been
supported by our current findings indicating that the bone strucpublished on AN bone structure using 3D-pQCT and reported
ture of these subjects is mainly impaired in weight-bearing bone
similar bone structural impairments in the distal radius, an apregions. In contrast with constitutional thinness, microgravity
pendicular skeletal site without a weight-bearing effect (32). The
bone deficit models display bone uncoupling (41) and a low OPG
present study confirms these data in both the radius and tibia in
to RANKL ratio (38), suggesting that this hypothesis must be
AN patients, but only in those with long-standing disease. Strong
interpreted cautiously.
correlations with age and disease duration emphasize the imThe low bone mass and overall bone size reduction in conportance of duration of AN on bone.
stitutional thinness associated with normal hormonal parameBone size assessment indicated that constitutionally thin subters led us to hypothesize a genetic causality. We have previously
jects have smaller bones than controls or AN subjects. In accorreported that the families of the constitutionally thin subjects
dance with previous studies (33), AN subjects displayed no modfrequently have low BMI (14), in line with previous studies (28,
ification of appendicular bone size.
29, 40, 42). Therefore, the genetic determinism of severe underThe unusual combination of low bone density and size couweight was postulated. This hypothesis is supported by previous
pled to normal bone turnover found in constitutionally thin
publications (40) suggesting that genes associated with low bone
women needs to be explained. First, low bone density in constimass and severe underweight might be related.
tutional thinness is clearly not related to hormonal influences.
Peak bone mass achieved before the age of 30 yr is one of the
The main hormonal mechanisms identified as being responsible
main determinants of bone mass in the elderly, and the most
for low bone mass in AN involve estrogen deficiency (34), hyimportant period is the bone-building years before the age of 20
percorticism (24, 34), high GH levels, and low IGF-I (35). The
116
Galusca et al.
Bone Quality in Constitutional Thinness
yr (43). The reduced bone breaking strength observed in 20-yrold constitutionally thin subjects, similar to that observed in
long-standing AN, suggests an increased lifetime fracture risk
(44).
In conclusion, young women with constitutional thinness,
associated with low bone mass and normal bone turnover and no
identified hormonal or nutritional causality, represent an unusual model of low peak bone mass and implicitly low bone
quality.
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
18.
19.
20.
21.
22.
Acknowledgments
Address all correspondence and requests for reprints to: Bruno Estour, Endocrinology Department, Centre Hospitalier Universitaire Saint Etienne, 42055
Saint Etienne, Cedex 2, France. E-mail: [email protected].
Disclosure Statement: The authors have nothing to declare.
23.
24.
25.
References
1. Peacock M, Turner CH, Econs MJ, Foroud T 2002 Genetics of osteoporosis.
Endocr Rev 23:303–326
2. Ravn P, Cizza G, Bjarnason NH, Thompson D, Daley M, Wasnich RD, McClung M, Hosking D, Yates AJ, Christiansen C 1999 Low body mass index is
an important risk factor for low bone mass and increased bone loss in early
postmenopausal women. Early Postmenopausal Intervention Cohort (EPIC)
study group. J Bone Miner Res 14:1622–1627
3. Crilly RG, Delaquerriere-Richardson L 1990 Current bone mass and body
weight changes in alcoholic males. Calcif Tissue Int 46:169 –172
4. Thomas J, Doherty SM 2003 HIV infection–a risk factor for osteoporosis. J
Acquir Immune Defic Syndr 33:281–291
5. Andreassen H, Hylander E, Rix M 1999 Gender, age, and body weight are the
major predictive factors for bone mineral density in Crohn’s disease: a casecontrol cross-sectional study of 113 patients. Am J Gastroenterol 94:824 – 828
6. Anker SD, Clark AL, Teixeira MM, Hellewell PG, Coats AJ 1999 Loss of bone
mineral in patients with cachexia due to chronic heart failure. Am J Cardiol
83:612– 615
7. Rigotti NA, Nussbaum SR, Herzog DB, Neer RM 1984 Osteoporosis in
women with anorexia nervosa. N Engl J Med 311:1601–1606
8. Seeman E, Szmukler GI, Formica C, Tsalamandris C, Mestrovic R 1992 Osteoporosis in anorexia nervosa: the influence of peak bone density, bone loss,
oral contraceptive use, and exercise. J Bone Miner Res 7:1467–1474
9. Carmichael KA, Carmichael DH 1995 Bone metabolism and osteopenia in
eating disorders. Medicine (Baltimore) 74:254 –267
10. Galusca B, Bossu C, Germain N, Kadem M, Frere D, Lafage-Proust MH, Lang
F, Estour B 2006 Age-related differences in hormonal and nutritional impact
on lean anorexia nervosa bone turnover uncoupling. Osteoporos Int 17:888 –
896
11. Soyka LA, Grinspoon S, Levitsky LL, Herzog DB, Klibanski A 1999 The effects
of anorexia nervosa on bone metabolism in female adolescents. J Clin Endocrinol Metab 84:4489 – 4496
12. Apfelbaum M, Sachet P 1982 [Constitutional thinness]. Rev Prat 32:245–247
(French)
13. Tolle V, Kadem M, Bluet-Pajot MT, Frere D, Foulon C, Bossu C, Dardennes
R, Mounier C, Zizzari P, Lang F, Epelbaum J, Estour B 2003 Balance in ghrelin
and leptin plasma levels in anorexia nervosa patients and constitutionally thin
women. J Clin Endocrinol Metab 88:109 –116
14. Bossu C, Galusca B, Normand S, Germain N, Collet P, Frere D, Lang F, Laville
M, Estour B 2007 Energy expenditure adjusted for body composition differentiates constitutional thinness from both normal subjects and anorexia nervosa. Am J Physiol Endocrinol Metab 292:E132–E137
15. 1995 Physical status: the use and interpretation of anthropometry. Report of
a WHO Expert Committee. World Health Organ Tech Rep Ser 854:1– 452
16. American Psychiatric Association 1994 Diagnostic and statistical manual
of mental disorders. 4th ed. Washington, DC: American Psychiatric Association
17. Viricel J, Bossu C, Galusca B, Kadem M, Germain N, Nicolau A, Millot L,
Vergely N, Lassandre S, Carrot G, Lang F, Estour B 2005 [Retrospective study
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
of anorexia nervosa: reduced mortality and stable recovery rates]. Presse Med
34(20 Pt 1):1505–1510 (French)
William DL, Robert AA, Ghada El-Hajj F, Anthony H, David LK, Michael M,
Paul DM, Nelson BW 2005 Application of the 1994 WHO Classification to
Populations Other Than Postmenopausal Caucasian Women: The 2005 ISCD
Official Positions. J Clin Densitom 9:22–30
Laib A, Hauselmann HJ, Ruegsegger P 1998 In vivo high resolution 3D-QCT
of the human forearm. Technol Health Care 6:329 –337
Schoenau 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
Siu WS, Qin L, Leung KS 2003 pQCT bone strength index may serve as a better
predictor than bone mineral density for long bone breaking strength. J Bone
Miner Metab 21:316 –322
Bagi CM, Hanson N, Andresen C, Pero R, Lariviere R, Turner CH, Laib A
2006 The use of micro-CT to evaluate cortical bone geometry and strength in
nude rats: correlation with mechanical testing, pQCT and DXA. Bone 38:
136 –144
Garnero P, Gineyts E, Riou JP, Delmas PD 1994 Assessment of bone resorption
with a new marker of collagen degradation in patients with metabolic bone
disease. J Clin Endocrinol Metab 79:780 –785
Vergely N, Lafage-Proust MH, Caillot-Augusseau A, Millot L, Lang F, Estour
B 2002 Hypercorticism blunts circadian variations of osteocalcin regardless of
nutritional status. Bone 30:428 – 435
Bonde M, Qvist P, Fledelius C, Riis BJ, Christiansen C 1995 Applications of
an enzyme immunoassay for a new marker of bone resorption (CrossLaps):
follow-up on hormone replacement therapy and osteoporosis risk assessment.
J Clin Endocrinol Metab 80:864 – 868
Biadi O, Rossini R, Musumeci G, Frediani L, Masullo M, Ramacciotti CE,
Dellosso L, Paoli R, Mariotti R, Cassano GB, Mariani M 2001 Cardiopulmonary exercise test in young women affected by anorexia nervosa. Ital Heart J
2:462– 467
Schneider L, Heard I, Breart G, Henrion R 1979 [Maternal thinness and pregnancy]. Arch Fr Pediatr 36:1068 –1074 (French)
Norgan NG 1994 Interpretation of low body mass indices: Australian aborigines. Am J Phys Anthropol 94:229 –237
Gartner A, Maire B, Kameli Y, Traissac P, Delpeuch F 2001 Body composition
unaltered for African women classified as ‘normal but vulnerable’ by body
mass index and mid-upper-arm-circumference criteria. Eur J Clin Nutr 55:
393–399
Hofbauer LC, Schoppet M 2004 Clinical implications of the osteoprotegerin/
RANKL/RANK system for bone and vascular diseases. JAMA 292:490 – 495
Misra M, Soyka LA, Miller KK, Herzog DB, Grinspoon S, De Chen D,
Neubauer G, Klibanski A 2003 Serum osteoprotegerin in adolescent girls with
anorexia nervosa. J Clin Endocrinol Metab 88:3816 –3822
Milos G, Spindler A, Ruegsegger P, Seifert B, Muhlebach S, Uebelhart D,
Hauselmann HJ 2005 Cortical and trabecular bone density and structure in
anorexia nervosa. Osteoporos Int 16:783–790
Karlsson MK, Weigall SJ, Duan Y, Seeman E 2000 Bone size and volumetric
density in women with anorexia nervosa receiving estrogen replacement therapy and in women recovered from anorexia nervosa. J Clin Endocrinol Metab
85:3177–3182
Biller BM, Saxe V, Herzog DB, Rosenthal DI, Holzman S, Klibanski A 1989
Mechanisms of osteoporosis in adult and adolescent women with anorexia
nervosa. J Clin Endocrinol Metab 68:548 –554
Rappaport R, Prevot C, Czernichow P 1980 Somatomedin activity and growth
hormone secretion. I. Changes related to body weight in anorexia nervosa.
Acta Paediatr Scand 69:37– 41
Holtkamp K, Mika C, Grzella I, Heer M, Pak H, Hebebrand J, HerpertzDahlmann B 2003 Reproductive function during weight gain in anorexia nervosa. Leptin represents a metabolic gate to gonadotropin secretion. J Neural
Transm 110:427– 435
Bakker I, Twisk JW, Van Mechelen W, Kemper HC 2003 Fat-free body mass
is the most important body composition determinant of 10-yr longitudinal
development of lumbar bone in adult men and women. J Clin Endocrinol
Metab 88:2607–2613
Rucci N, Rufo A, Alamanou M, Teti A 2007 Modeled microgravity stimulates
osteoclastogenesis and bone resorption by increasing osteoblast RANKL/OPG
ratio. J Cell Biochem 100:464 – 473
Vico L, Collet P, Guignandon A, Lafage-Proust MH, Thomas T, Rehaillia M,
Alexandre C 2000 Effects of long-term microgravity exposure on cancellous
and cortical weight-bearing bones of cosmonauts. Lancet 355:1607–1611
Bulik CM, Allison DB 2001 The genetic epidemiology of thinness. Obes Rev
2:107–115
Caillot-Augusseau A, Lafage-Proust MH, Soler C, Pernod J, Dubois F, Alex-
J Clin Endocrinol Metab, January 2008, 93(1):110 –117
andre C 1998 Bone formation and resorption biological markers in cosmonauts during and after a 180-day space flight (Euromir 95). Clin Chem 44:
578 –585
42. Parker E, Phillips DI, Cockington RA, Cull C, Poulton J 2005 A common
mitochondrial DNA variant is associated with thinness in mothers and
their 20-yr-old offspring. Am J Physiol Endocrinol Metab 289:E1110 –
E1114
jcem.endojournals.org
117
43. Matkovic V, Jelic T, Wardlaw GM, Ilich JZ, Goel PK, Wright JK, Andon MB,
Smith KT, Heaney RP 1994 Timing of peak bone mass in Caucasian females
and its implication for the prevention of osteoporosis. Inference from a crosssectional model. J Clin Invest 93:799 – 808
44. Lucas AR, Melton 3rd LJ, Crowson CS, O’Fallon WM 1999 Long-term fracture risk among women with anorexia nervosa: a population-based cohort
study. Mayo Clin Proc 74:972–977