Prevalence of Vertebral Fractures in Children with Chronic Rheumatic

Diseases at Risk for Osteopenia

MERANDA NAKHLA, MD, FRCP(C), ROSIE SCUCCIMARRI, MD, FRCP(C), KAREN N. WATANABE DUFFY, MD, FRCP(C),
GALLE CHDEVILLE, MD, FRCP(C), SARAH CAMPILLO, MD, FRCP(C), CIARN M. DUFFY, MB BCH, FRCP(C),
E. MICHEL AZOUZ, MD, FRCP(C), NAZIH SHENOUDA, MD, FRCP(C), ATUL K. SHARMA, MD, FRCP(C), AND
CELIA RODD, MD, FRCP(C)

Objectives To determine the prevalence of and the risk factors for vertebral fractures in a cohort of children with chronic
rheumatic diseases considered at risk for osteopenia.
Study design We conducted a cross-sectional study of patients with chronic rheumatic diseases at the Montreal Childrens
Hospital.
Results Of the 90 study participants (22 boys, 68 girls), 10 boys and 7 girls (19%) were found to have vertebral fractures. These
17 children had a total of 50 fractures, an average of 2.9 per affected child. Fractures in the upper thoracic region (T5-8) accounted
for 55%. Only 56% of all fractures were symptomatic. With multivariate regression, we identified male sex (P < .01), body mass
index z-score (P < .02), and cumulative glucocorticoid dose (P < .01) as significant predictors of the number of vertebral fractures.
Conclusions Our study examined the prevalence of vertebral fractures in a high-risk pediatric population. Nineteen percent
of our cohort had vertebral fractures. Significant risk factors for the development of vertebral fractures include male sex and
cumulative glucocorticoid dose. Better understanding of the extent of the problem in this population will allow us to further
refine screening guidelines and treatment in these patients. (J Pediatr 2009;154:438-43)
steoporosis is a skeletal disorder characterized by low bone mass and deterioration of bone microarchitecture, resulting
in fragility fracture susceptibility. In children, low bone mass is caused by poor
bone accrual and ongoing bone loss caused by a number of factors. These
include inadequate calcium and vitamin D intake, chronic illness, limited weight-bearing
activity, medications, and hormonal imbalances. In adults, age-related fractures associated
From the Divisions of Endocrinology (M.N.,
with osteoporosis are a well-known entity, contributing to significant morbidity.1 In
C.R.), Rheumatology (R.S., K.W.D., G.C.,
children with chronic illnesses, osteoporosis, fragility fractures, and bone pain have
S.C., C.D.), and Nephrology (A.S.), Mon2-4
increasingly been recognized as a source of morbidity. However, the extent of this
treal Childrens Hospital, McGill University
Health Centre, Montreal, Quebec, Canada;
problem is poorly understood. Studies examining vertebral fractures in children have been
and Department of Radiology, Childrens
limited to case reports or have restricted their study population to children treated with
Hospital of Eastern Ontario, Ottawa, On4,5
tario, Canada (E.A., N.S.).
glucocorticoids or with low bone mineral density (BMD). Other studies have relied on
This project was funded by the Montreal
an indirect method of vertebral fracture identification by using dual-energy x-ray absorpChildrens Hospital Research Institute. M.N.
tiometry (DXA) targeting scans, which have subsequently been invalidated by the authors
was funded by the Eli Lilly Canadian Pedi3-7
atric Endocrine Fellowship. The Montreal
themselves. A global assessment of children at risk of vertebral fractures is still lacking.
Childrens Hospital Research Institute did
Moreover, issues such as prevalence, risk factors, and natural history of vertebral fractures
not have a role in the development of the
study design, the collection, analysis, and
in children remain largely unknown.
interpretation of data, the writing of the
The rheumatology population is at particular risk for osteoporosis and fragility
report, or the decision to submit the paper
8-10
fractures because of the chronic inflammatory nature of the disease,
the medications
for publication. The authors declare no po11-13
tential conflicts of interest.
used to control disease activity, including glucocorticoids and methotrexate,
and the
14
Submitted for publication Jun 23, 2008; last
associated reduced physical activity and delayed puberty, which independently contribrevision received Aug 12, 2008; accepted
ute to fracture risk. We sought to determine the prevalence and predictors of vertebral
Sep 9, 2008.
Reprint requests: Meranda Nakhla, Division
fractures in this at-risk population by using plain thoracic and lumbar spine radiographs,

BMC
BMD
BMI
CTD
DXA

438

Bone mineral content

Bone mineral density
Body mass index
Connective tissue disease
Dual-energy x-ray absorptiometry

JDM
JIA
SLE
vBMD

Juvenile dermatomyositis
Juvenile idiopathic arthritis
Systemic lupus erythematosus
Volumetric BMD

of Endocrinology, Montreal Childrens Hospital, 2300 Tupper St, E-315, Montreal, Qc,
H3H 1P3, Canada. E-mail: nakhlam@yahoo.
com.
0022-3476/$ - see front matter
Copyright 2009 Mosby Inc. All rights
reserved.
10.1016/j.jpeds.2008.09.023

which remain the radiographic gold standard.15 We also set

out to determine the predictors of low BMD.

METHODS
In April 2005, we reviewed the charts of all patients
observed in the rheumatology clinic at the Montreal Childrens Hospital and identified 113 individuals who met our
inclusion criteria. From April 2005 to December 2006, we
invited these children and adolescents who fulfilled criteria for
being at risk for osteopenia and who were observed in the
rheumatology clinic at Montreal Childrens Hospital to participate in the study. Children as old as 18 years were included
when they had earlier or current exposure to methotrexate,
corticosteroids, or both and had the following diagnosis:
juvenile idiopathic arthritis (JIA; either polyarthritis, systemic
arthritis, extended oligoarthritis, or psoriatic arthritis with 5
affected joints); connective tissue disease (CTD), including
systemic lupus erythematosus (SLE) and juvenile dermatomyositis (JDM); and systemic vasculitis. Children with preexisting conditions that could affect calcium homeostasis were
excluded, including those with hyperthyroidism, parathyroid
disorders, malignancy, or primary bone disease. Parental informed consent and patients assent were obtained. Study
participants and subjects who refused to participate were
compared for age at diagnosis, duration of disease, and diagnosis.
We reviewed the charts of the study subjects and extracted these characteristics: age at diagnosis, age and duration of disease at time of enrollment, associated diagnoses,
and other medical problems. Total lifetime exposure to oral
and intravenous glucocorticoids was recorded. We converted
the glucocorticoid dosage to prednisone equivalency and also
expressed this as prednisone per kg of body weight. Also, we
recorded the lifetime cumulative methotrexate doses in the
form of mg per m2 of body surface area.
At the time of enrollment, a standardized food frequency questionnaire was administered to patients and their
parents, from which daily calcium (mg) and vitamin D (IU)
intakes were calculated.16 In addition, patients and parents
were asked to complete a questionnaire about their personal
and family history of fractures, osteoporosis, age of pubertal
onset, and use of medications or supplements. At the same
clinic visit, children were examined and Tanner staging was
performed on all study subjects. Heights and weights were
measured with a standard hospital scale and a Harpendon
stadiometer. The heights were measured in triplicate, and the
average was calculated. Heights and weights were converted
to z-scores by using the Centers for Disease Control and
Prevention-National Center for Health Statistics 2000 protocol.17
All patients enrolled underwent plain lateral thoracic
and lumbar spine radiography. These radiographs were read
by 1 of 2 trained radiologists (E.M.A. or N.S.), experts in
reading pediatric vertebral fractures, who were blinded to the
underlying diagnosis in the patients. The senior author reviewed all films. When there was a discrepancy, the films were

Figure. Genant scoring system for vertebral fractures.

re-read by the second pediatric radiologist. Vertebral fractures

were graded on the basis of the semi-quantitative scoring
method of Genant.15 The Genant scoring system assigns a
score for each vertebra from T4 to L4 (Figure). Grade 0
corresponds to a normal vertebral body, grade 1 to a mild
fracture (20%-25% reduction in anterior, middle, and/or posterior height), grade 2 to a moderate fracture (26%-40%
reduction in height), and grade 3 to a severe fracture (40%
reduction in height). A grading of 0.5 was used for minimal
changes of the vertebral height or shape that did not fulfill
criteria for Genant grade 1 and were considered normal
variants. A Genant score 1 was considered to be a definite
compression fracture of the vertebral body.
All study subjects also underwent BMD of the lumbar
spine measured with the Hologic 4500A DXA. Each BMD
was converted to a z-score by using the manufacturers standard pediatric reference range. We also determined the volumetric BMD (vBMD) for subjects 7.5 years old and converted it to a z-score,18 allowing correction of the BMD for
stature. When fractures involved the L2 to L4 region, the
BMD was calculated excluding the affected vertebra(e).
Statistical analyses were performed with SAS software
version 9.1 (SAS Institute, Carey, North Carolina). Two
patients with vertebral fractures were missing dietary assessment of total daily calcium and vitamin D intake, for which
we performed multiple imputations with SAS PROC MI. In
brief, missing values of total daily calcium and vitamin D
intake were imputed (m 10 replicates, method regression). For the imputation, independent predictors were the
same variables used in the Poisson and multivariate regressions described below. Both of these regression techniques
were then applied to each of the imputed datasets, and the
results were formally pooled in SAS PROC MIANALYZE
to provide unbiased parameter estimates that properly reflected the uncertainty associated with missing data.19
The Mann-Whitney U test was used to test significance
for median differences in baseline characteristics and risk
factors between the fracture and non-fracture group. We used
multivariate Poisson count regression analysis to identify pos-

Prevalence of Vertebral Fractures in Children with Chronic Rheumatic Diseases at Risk for Osteopenia

439

Table I. Vertebral fraction population

Subjects
Steroid treated
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Steroid nave
15
16
17

Sex

Diagnosis

M
M
M
F
M
F
M
M
M
M
F
M
F
F

Systemic vasculitis
Systemic vasculitis
Systemic vasculitis
CTD-SLE
CTD-SLE
CTD-JDM
CTD-JDM
JIAsystemic
JIAsystemic
JIAsystemic
JIAsystemic
JIAsystemic
JIApsoriatic
JIApolyarthritis

M
F
F

JIApolyarthritis
JIAextended oligoarthritis
JIAPsoriatic

Number of vertebral
fractures/Pain *

Location of vertebral fractures

2*
3
4
2*
1*
1*
4
3
11*
6
2*
1*
1*
1*

L2; L3
T5; T6; T7
T4:T5; T6; T7
T7; T8
T7
T11
T8; T9; T10; T11
T4; T5; T7
T5; T6; T7; T8; T11; T12; L1; L2; L3; L4; L5
T5:T6; T7; L2; L3:L4
T7:T11
T11
T6
T6

2
4
2

T5; T6
T7; T8; T10; T11
T7; L3

M, Male; F, female.

sible predictive factors for vertebral fractures, which included

age at diagnosis, sex, disease duration, cumulative steroid dose
per kg, cumulative dose of methotrexate per m2, total calcium
and vitamin D intake, including both dietary and supplemental sources, Body mass index (BMI)-z score, and lumbar spine
BMD z-score. The vBMD z-score for subjects 7.5 years old
and bone mineral content (BMC) were also examined as
alternatives to the BMD z-score. Multivariate linear regression was used to identify possible predictors of BMD, vBMD
z-score (subjects 7.5 years old), and BMC. The predictors
included in the model were duration of disease, age at diagnosis,
cumulative prednisone dose in mg/kg, cumulative methotrexate
dose in mg/m2, Tanner stage, and sex.
Ethics approval was granted by the institutional review
board at the Montreal Childrens Hospital.

RESULTS
Ninety-four of the 113 subjects who met the inclusion
criteria agreed to participate in the study; 4 subjects did not
complete the radiography and were excluded from the analyses. Two children had their vitamin D and calcium statuses
imputed and were included in the analyses. The baseline
characteristics between children who declined to participate
or were excluded and children who participated were similar.
At the time of enrollment, the study participants, consisting
of 22 boys and 68 girls, had a median age of 13.1 years (range,
4.3-18.0 years), with a median duration of illness of 4.6 years
(range, 0.4-15.7 years; data not shown). The median age at
diagnosis was 6.0 years (range, 0.9-15.7 years).
Vertebral fractures were detected in 17 patients (19%);
14 patients had been exposed to steroids, and 3 patients had
not been exposed. Of this group, the median age at diagnosis
440

Nakhla et al

was 7.0 years (range, 1.4-14.1 years), with a slight predominance of male patients (10 male; 7 female; Tables I and II). In
the steroid-treated group, 7 patients had JIA (5 systemic
arthritis, 1 polyarthritis arthritis, 1 psoriatic arthritis), 4 patients had CTD (2 SLE and 2 JDM), and 3 patients had
systemic vasculitis (Table I). In the steroid-nave group, all
patients had JIA (1 polyarthritis, 1 extended oligoarthritis,
and 1 psoriatic arthritis). Of the 17 children, there were a total
of 50 fractures, averaging 2.9 fractures per subject. Eightyeight percent of the fractures were mild, with a Genant score
of 1, corresponding to a loss of vertebral height of about 20%
to 25%; 8% of the fractures were moderate, with a score of 2,
a loss of height of 26% to 40%; and 4% of the fractures were
severe, with a loss of height 40%. Thoracic-region fractures
comprised 78% of the vertebral fractures sustained, with the
remaining fractures occurring in the lumbar region. Fifty-five
percent of the fractures were in the upper thoracic region of
T4 to T8. Approximately 14% of the fractures were biconcave, 18% were crush, and 68% were anterior wedge. More
than half the fractures were symptomatic in the location of the
fracture (56%), although the pain was often transient and
mild.
We compared the baseline characteristics and risk factor
variables between patients with and patients without vertebral
fractures (Table II). Except for exposure to long-term glucocorticoids (P .01) and BMI z-score (P .03), there were
no other significant differences seen in risk factors in the 2
groups. In the group without vertebral fractures, 35 patients
(48%) were exposed to glucocorticoids.
We identified male sex, cumulative glucocorticoid dosage per kg of body weight, and BMI z-score as significant
independent predictors of vertebral fracture count (Table III).
The Journal of Pediatrics March 2009

Table II. Baseline characteristics of study population

Characteristics

Vertebral fracture

No vertebral fractures

n
Sex distribution
Median age at diagnosis (range), years
Median duration of illness (range), years
JIA (n, %)
CTD (n, %)
Systemic vasculitis (n, %)
Median z-BMD (95% CI; n 73)
Median z-vBMD (95% CI)
Median height z (95% CI)
Median weight z (95% CI)
Median BMI z (95% CI)*
Median cumulative glucocorticoid dose (95% CI), mg/kg*
Median cumulative methotrexate dose (95% CI), mg/m2
Median total daily vitamin D intake (95% CI), IU
Median total daily calcium intake (95% CI), mg

17
10 M:7 F
7.0 (1.4-14.1)
5.5 (0.6-10.9)
10, 56%
4, 28%
3, 16%
1.40 (2.01-0.31)
0.61 (2.03-0.21)
0.19 (1.36-0.17)
0.48 (0.17-1.17)
1.34 (0.30-1.48)
258.1 (164.4-674.7)
1198.0 (873.6-3967.3)
340 (296-507)
1183 (1068-1468)

73
12 M:61 F
6.0 (0.9-15.7)
4.60 (0.4-15.7)
42, 57%
22, 30%
9, 13%
0.60 (0.87-0.43)
0.17 (0.56-0.07)
0.34 (0.67-0.25)
0.06 (0.16-0.33)
0.34 (0.06-0.55)
0.0 (0.0-116.4)
1584.3 (1693.3-2748.4)
336 (328-451)
1262 (1187-1511)

*P value .05, Mann-Whitney test.

Table III. Multivariate Poisson regression analysis

for characteristics predicting vertebral fracture risk
Characteristic

Odds
ratio

95% CI

P value

Sex (male)
BMI z-score
Duration of disease
Age at diagnosis
BMD z-score
Cumulative glucocorticoid (g/kg)
Cumulative methotrexate (g/m2)
Total daily calcium intake (g)
Total vitamin D intake (1000 IU)

6.04
1.49
1.10
0.98
0.89
4.50
0.84
0.46
2.18

2.85-12.81
1.05-2.09
0.94-1.28
0.89-1.09
0.62-1.30
1.42-14.28
0.68-1.02
0.17-1.25
0.37-12.76

.001*
.02*
.24
.82
.57
.01*
.08
.13
.38

*Statistically significant.

In 3 separate multivariable regression models, the z-scores for

BMD, vBMD (n 73 subjects), or BMC were not found to
increase vertebral fracture risk unless glucocorticoid exposure
was omitted from the model, reflecting a correlation in the 2
predictors (see linear regression results below). Male sex increased the expected number of vertebral fractures by a factor
of 6.04 (95% CI, 2.85-12.81), and each 1 g/kg increase in
cumulative glucocorticoid exposure increased the expected
number of vertebral fractures by a factor of 4.5 (95% CI,
1.42-14.28). Similarly, an increase in BMI z-score of 1 SD
increased the expected fracture count by 1.49 (95% CI, 1.052.09), when these risks are multiplicative odds ratios. The
strength of the association between fracture count and BMI
z-score was considerably weaker (P .08) when vBMD was
included instead of BMD, effectively adjusting the BMD for
body size.
After controlling for all the other factors in multivariate
linear regression, cumulative glucocorticoid dosage, BMI zscore, and duration of disease were significant independent

predictors of reduced BMD z-score (Table IV). These same

factors were also found to be independent predictors of
vBMD z-scores and of BMC (data not shown). The relationship (collinearity) between glucocorticoids and bone mineralization is noteworthy because it suggests an explanation for
the finding that BMD z-score is a significant predictor of
vertebral fractures only when cumulative glucocorticoid dose
is omitted from the model.

DISCUSSION
Literature on the prevalence of vertebral fractures in an
unselected high-risk pediatric population with chronic rheumatic disease is limited; earlier reports restricted their analyses
to patients treated with glucocorticoids or with low BMD or
used an unreliable indirect method of fracture identification.4,5
In our study, we demonstrated a vertebral fracture prevalence of 19% in a high-risk general population of patients
with rheumatic diseases; most fractures occurred in the upper
thoracic region (T4-T8). This is in accordance with earlier
pediatric studies that had found the thoracic region to be the
most commonly affected site, either because of trauma or
fragility fractures.5,6,20 In our population of patients, increased vertebral fracture risk was associated with male sex,
BMI z-score, and lifetime glucocorticoid exposure (mg/kg).
The toxic effects of glucocorticoids on bone metabolism
and calcium homeostasis are well known. A case-control
study examining the effect of prednisolone on vertebral fracture risk in children with juvenile chronic arthritis found that
the mean daily dose of prednisolone (mg/kg) was significantly
higher in those with vertebral fractures compared with those
without fractures.4 However, unlike our study, the mean total
glucocorticoid dose (mg) was not found to be significantly
different in the 2 groups. As was seen in our patients, most

Prevalence of Vertebral Fractures in Children with Chronic Rheumatic Diseases at Risk for Osteopenia

441

Table IV. Predictors of BMD z-scores using multivariate linear regression analysis
Explanatory variables

Sex
Male
Female
Total calcium intake (mg/day)
Total vitamin D intake (IU/day)
Age at diagnosis (years)
Disease duration (years)
Cumulative prednisone dose (mg/kg)
Cumulative methotrexate dose (mg/m2)
BMI z-score

0.10
1.00
1.3 104
3.0 104
1.7 103
0.083
0.0015
6.4 105
0.32

95% CI
0.43-0.63
5.6 104-3.6 104
8.7 104-1.5 103
0.061-0.064
0.0045-0.160
0.002-0.001
1.8 104-5.0 105
0.12-0.52

P value
.71
.56
.60
.96
.04
.001
.10
.01

, Regression coefficients.

fractures were in the upper thoracic region (T6-T8). Another

pediatric study found a lack effect of glucocorticoids on vertebral fracture risk; however, this study was limited by its
small sample size.6 In addition, in adults with rheumatic
disorders, glucocorticoid exposure, as measured by mean daily
dose, has been found to be a risk factor for the development
of vertebral fractures.21,22
Three of the patients who sustained vertebral fractures
had not received steroids. This finding raises the possible role
of inflammatory cytokines in promoting bone fragility.8-10
The presence of tumor necrosis factor and interleukin-1 has
been associated with suppression of osteoblast function and
differentiation, decreasing bone formation.9
We found male sex to be an independent predictor of
vertebral fracture risk, irrespective of diagnosis, increasing the
expected risk of vertebral fractures by a factor of 6.04. Helenius et al also noted an increased male preponderance of
vertebral fractures in children after solid organ transplantation, although the underlying mechanism remains obscure.7
One hypothesis to explain the effect of male sex is that boys
tend to be more involved in physical and risk-taking activities
as compared with girls, potentially increasing their risk of
physical injury.
We did not find methotrexate to have an effect on
vertebral fracture risk or low BMD. Literature on the effect of
methotrexate on BMD and vertebral fractures is conflicting.
A recent study by Reyes et al reported that methotrexate was
an independent risk factor for vertebral fractures in children
with rheumatic disorders in whom glucocorticoids are concomitantly prescribed.5 Conversely, a retrospective cohort
study of subjects with JIA reported a possible protective effect
of disease modifying anti-rheumatic drugs including methotrexate on fracture risk; however, this effect was not significant.23 Larger prospective studies may be required to delineate whether methotrexate is implicated in vertebral fracture
risk or is protective because of its role in controlling inflammation.
We did not find BMD z-score, vBMD z-score, or
BMC to be predictors of vertebral fractures. This is consistent
with findings in a study conducted by Makitie et al, in which
32 patients, with various diagnoses and at risk of having
442

Nakhla et al

secondary osteoporosis, underwent spinal radiographs and

lumbar BMD measurements.3 However, other studies have
demonstrated an association between BMD and vertebral
fracture risk.5 In Makities study, of the 11 patients found to
have vertebral fractures, 3 had normal BMD z-scores, and
after correction for bone size, 8 were found to have normal
BMD scores.3 There are several explanations for the lack of
association between BMD measures and vertebral fracture
risk, including the existence of collinearity between the BMD
z-score and cumulative prednisone dose. As demonstrated in
the linear regression analysis, cumulative glucocorticoid dose
was found to be a predictor of low BMD z-score. With the
inclusion of both independent variables into the Poisson
regression model, it becomes difficult to distinguish their
individual influences on the vertebral fracture count. Also, the
results of BMD z-scores must be interpreted cautiously, because BMD corrects only for the area of bone studied, but not
for bone thickness. Thus, DXA has its limitations in providing information on bone morphology and microarchitecture.
These factors may be affected even with a normal BMD
result.24 Furthermore, a number of clinical variables confound
BMD results, including Tanner stage, skeletal maturation,
height, weight, lean body mass, ethnicity, and bone size.25
Correcting the BMD for height, weight, and BMI z-scores
are strategies that can be used to increase the accuracy of
BMD measurements, and our results suggest that such corrections can explain at least part of the association between
fractures and BMI described in this study and in the literature. Further analyses will be conducted to clarify the clinical
usefulness of these anthropometric corrections.
Our study has several limitations. First, the Genant
scoring method, which has only been standardized in adults,
was used. To date, there are no standardized pediatric assessments of vertebral morphology. Makitie et al devised a pediatric scoring system for the assessment of vertebral morphology.3 However, we did not use this scoring system because it
has yet to be standardized, and the results from Makities
study were released after our study had begun. One future direction will be to use the data collected to compare the
Genant scoring method with the method devised by Makitie
et al. Second, because vertebrae in the mid-thoracic spine
The Journal of Pediatrics March 2009

maybe slightly wedged, these vertebrae maybe misinterpreted

as mild vertebral fractures. In our study subjects, all subjects
who were treated with steroids and had 1 to 2 thoracic
fractures had pain in the area of the fracture. This finding
decreases the likelihood that the radiography readings were
artifacts. Third, we did not request biochemical investigations
because this may have lead to a decreased recruitment rate,
introduced selection bias, and undermined the primary goal of
determining the prevalence of vertebral fractures in this highrisk pediatric population. Fourth, we did not have a disease
activity score that crosses all diseases in our cohort. As a result
of this, cumulative methotrexate and prednisone doses were
used as proxy measures of severity of disease activity.
Our study has identified the prevalence of vertebral
fractures and its associated risk factors in children with rheumatic diseases. Our findings demonstrate that vertebral fractures in children are not uncommon sequelae in this population and stress the need for regular surveillance of vertebral
fractures in patients with rheumatic diseases, particularly because 44% of our fracture group was had no symptoms at the
time of enrollment. This allows for earlier intervention to
decrease the significant morbidity associated with vertebral
fractures. Male sex and cumulative glucocorticoid exposure
are both factors that can aid in delineating patients who are at
particular risk of sustaining a vertebral fracture. Regardless of
BMD, vBMD z-score, or BMC, the presence of vertebral
fractures strongly supports a diagnosis of osteoporosis. Our
findings have identified a high-risk pediatric group and have
contributed to further understanding the determinants of
vertebral compression fractures in children with not only
rheumatic diseases but with other chronic pediatric diseases
that have similar risk factors for osteoporosis.

REFERENCES
1.
Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability
associated with osteoporotic fractures. Osteoporos Int 2006;17:1726-33.
2.
Semeao EJ, Stallings VA, Peck SN, Piccoli DA. Vertebral compression fractures
in pediatric patients with Crohns disease. Gastroenterology 1997;112:1710-3.
3.
Makitie O, Doria AS, Henriques F, Cole WG, Compeyrot S, Silverman E, et al.
Radiographic vertebral morphology: a diagnostic tool in pediatric osteoporosis. J Pediatr
2005;146:395-401.
4.
Varonos S, Ansell BM, Reeve J. Vertebral collapse in juvenile chronic arthritis: its
relationship with glucocorticoid therapy. Calcif Tissue Int 1987;41:75-8.
5.
Reyes ML, Hernandez MI, King A, Vinet AM, Vogel A, Lagomarsino E, et al.
Corticosteroid-induced osteoporosis in children: outcome after two-year follow-up, risk

factors, densitometric predictive cut-off values for vertebral fractures. Clin Exp
Rheumatol 2007;25:329-35.
6.
Valta H, Lahdenne P, Jalanko H, Aalto K, Makitie O. Bone health and growth in
glucocorticoid-treated patients with juvenile idiopathic arthritis. J Rheumatol
2007;34:831-6.
7.
Helenius I, Remes V, Salminen S, Valta H, Makitie O, Holmberg C, et al.
Incidence and predictors of fractures in children after solid organ transplantation: a
5-year prospective, population-based study. J Bone Miner Res 2006;21:380-7.
8.
Manolagas SC, Jilka RL. Bone marrow, cytokines, and bone remodeling. Emerging insights into the pathophysiology of osteoporosis. N Engl J Med 1995;332:305-11.
9.
Canalis E. Effects of tumor necrosis factor on bone formation in vitro. Endocrinology 1987;121:1596-604.
10. Stashenko P, Dewhirst FE, Rooney ML, Desjardins LA, Heeley JD. Interleukin-1
beta is a potent inhibitor of bone formation in vitro. J Bone Miner Res 1987;2:559-65.
11. Kotaniemi A, Savolainen A, Kroger H, Kautiainen H, Isomaki H. Weightbearing physical activity, calcium intake, systemic glucocorticoids, chronic inflammation,
and body constitution as determinants of lumbar and femoral bone mineral in juvenile
chronic arthritis. Scand J Rheumatol 1999;28:19-26.
12. Falcini F, Trapani S, Civinini R, Capone A, Ermini M, Bartolozzi G. The
primary role of steroids on the osteoporosis in juvenile rheumatoid patients evaluated by
dual energy X-ray absorptiometry. J Endocrinol Invest 1996;19:165-9.
13. Trapani S, Civinini R, Ermini M, Paci E, Falcini F. Osteoporosis in juvenile
systemic lupus erythematosus: a longitudinal study on the effect of steroids on bone
mineral density. Rheumatol Int 1998;18:45-9.
14. Soyka LA, Fairfield WP, Klibanski A. Clinical review 117: hormonal determinants and disorders of peak bone mass in children. J Clin Endocrinol Metab
2000;85:3951-63.
15. Genant HK, Wu CY, van Kuijk C, Nevitt MC. Vertebral fracture assessment
using a semiquantitative technique. J Bone Miner Res 1993;8:1137-48.
16. Musgrave KO, Giambalvo L, Leclerc HL, Cook RA, Rosen CJ. Validation of a
quantitative food frequency questionnaire for rapid assessment of dietary calcium intake.
J Am Diet Assoc 1989;89:1484-8.
17. Anonymous. CDC growth charts: United States. Advance Data 2000;314:1-28.
18. Faulkner RA, Bailey DA, Drinkwater DT, McKay HA, Arnold C, Wilkinson
AA. Bone densitometry in Canadian children 8-17 years of age. Calcif Tissue Int
1996;59:344-51.
19. Little RR. Statistical analysis with missing data. 2nd ed. New York: Wiley
Interscience; 2002.
20. Reddy SP, Junewick JJ, Backstrom JW. Distribution of spinal fractures in children:
does age, mechanism of injury, or gender play a significant role? Pediatr Radiol
2003;33:776-81.
21. Peel NF, Moore DJ, Barrington NA, Bax DE, Eastell R. Risk of vertebral fracture
and relationship to bone mineral density in steroid treated rheumatoid arthritis. Ann
Rheum Dis 1995;54:801-6.
22. de Nijs RN, Jacobs JW, Bijlsma JW, Lems WF, Laan RF, Houben HH, et al.
Prevalence of vertebral deformities and symptomatic vertebral fractures in corticosteroid
treated patients with rheumatoid arthritis. Rheumatology (Oxford) 2001;40:1375-83.
23. Burnham JM, Shults J, Weinstein R, Lewis JD, Leonard MB. Childhood onset
arthritis is associated with an increased risk of fracture: a population based study using
the General Practice Research Database. Ann Rheum Dis 2006;65:1074-9.
24. Schonau E. Problems of bone analysis in childhood and adolescence. Pediatr
Nephrol 1998;12:420-9.
25. Gordon CM, Bachrach LK, Carpenter TO, Karsenty G, Rauch F. Bone health in
children and adolescents: a symposium at the annual meeting of the Pediatric Academic
Societies/Lawson Wilkins Pediatric Endocrine Society, May 2003. Curr Probl Pediatr
Adolesc Health Care 2004;34:226-42.

Prevalence of Vertebral Fractures in Children with Chronic Rheumatic Diseases at Risk for Osteopenia

443