вход по аккаунту


Predictors of total body bone mineral density in non-corticosteroid-treated prepubertal children with juvenile rheumatoid arthritis.

код для вставкиСкачать
Vol 40, N o 1 I , November 1997, pp 1Y67-1975
0 1997 Arnerl~dtlCollege ot Rheumdtology
Objective. To determine the extent of significant
osteopenia in prepubertal patients with juvenile rheumatoid arthritis (JRA) not treated with corticosteroids
and to identify variables that are highly related to bone
mineralization in this population.
Methods. In a cross-sectional study, 48 JRA patients and 25 healthy control subjects ages 4.6-11.0
years were evaluated. Total body bone mineral density
(TB BMD) was determined by Hologic dual energy x-ray
absorptiometry. All patients were prepubertal (Tanner
stage 1 or 11) and had never taken corticosteroids. For
comparison, JRA patients were divided into c‘lo~7’
BMD (Z score 5 - 1 ) or “normal” TB BMD (Z score
Results. The overall mean k SD TB BMD scores
did not differ between the JRA subjects (0.75 2 0.06
gm/cm2) and controls (0.73 k 0.07 gm/cm2; P > 0.30).
However, 29.2% of the JRA patients had low TB BMD,
whereas only 16% would be expected to have low TB
BMD based on the standard normal distribution (goodness of fit = 4.84, P = 0.01). The mean Z score for the
JRA patients with low TB BMD was -1.43, and for
those with normal TB BMD, it was 0.32. The JRA
subjects with low TB BMD were significantly younger,
had more active articular disease, greater physical
function limitation, higher erythrocyte sedimentation
rate, higher joint count severity score, lower body mass
index, lower lean body mass, less participation in organized sports, and more protein and vitamin D in their
diet compared with JRA patients with normal TB BMD
(all P < 0.05). Using logistic regression, a model
including age at JRA onset, Juvenile Arthritis Functional Assessment Report (JAFAR) score, triceps skinfold percentiles, percentage US recommended daily
allowance for dietary magnesium intake, and serum
1,25-dihydroxyvitamin D levels was able to accurately
segregate 79.6% of the JRA subjects into either the low
or normal TB BMD groups
= 20.5, P = 0.001).
Conclusion. This study demonstrated that in a
mildly to moderately ill prepubertal JRA population
that had never been exposed to corticosteroids, almost
30% had significantly low TB BMD. The patients with
low TB BMD had more active and severe articular
disease and greater physical function limitation.
Disease-related parameters in JRA appear to exert a
negative effect on bone mineralization even in prepubertal children, which can be demonstrated despite the
exclusion of corticosteroid-treated patients.
Supported in part by a Clinical Science grant and by a
Doctoral Dissertation award for Health Professionals from the Arthritis Foundation, by GCRC grant M01-RR-118084 from the NIH, by the
Children’s Hospital Research Foundation of Cincinnati, by the
Schmidlapp Foundation, and by NIH grants AR-42632 and AR-44059.
Carol J. Henderson, MS, RD, LD, Bonny L. Specker, PhD,
Rosa I. Sierra, MS, Robert W. Wilmott, MD, Daniel J. Lovell, MD,
MPH: Children’s Hospital Medical Center, Cincinnati, Ohio; Gail
Dunkel Cawkwell, MDCM, PhD: All Children’s Hospital, University
of South Florida, St. Petersburg; Barbara N. Campaigne, PhD: American College of Sports Medicine, Indianapolis, Indiana.
Address reprint requests to Daniel J. Lovell, MD, MPH,
Division of Rheumatology, Children’s Hospital Medical Center, 3333
Burnet Avenue, Pavilion 2-129. Cincinnati, OH 45229.
Submitted for publication October 18, 1996; accepted in
revised form June 18, 1997.
Osteoporosis is the exaggerated loss of bone
tissue that leads to reduced bone strength, poor bone
quality, and increased risk of fractures. l t affects more
than 25 million people in the United States, making it
the most common metabolic bone disorder in North
America (1,2). Contributors to healthy bones are hormonal, nutritional, and mechanical factors, and a deficiency in any one may result in bone loss ($4). Only
recently has the prevention of osteoporosis been considered a legitimate domain of pediatrics (5). Almost half
of the adult skeletal mass is formed during the second
decade of life, and calcium accumulation normally tri-
ples during the pubertal spurt (6-9). Maximal deposition
of mineral in the skeleton during childhood is important
to offset subsequent inexorable bone loss that accompanies aging.
Bone loss is common in juvenile rheumatoid
arthritis (JRA), occurring early in the disease even in
children not taking corticosteroids (10-14). Generalized
osteoporosis, pathologic long bone fractures, and vertebral crush fractures have been reported in 15-26% of
JRA patients in older studies utilizing standard radiographic assessment (15,16). Significant osteopenia continues to be a clinical problem in JRA patients. Recent
studies measuring bone mineral density (BMD) in JRA
patients have demonstrated significantly subnormal
BMD levels in both the axial (17) and appendicular
skeleton (18-20); however, few of these studies have
addressed the multifactorial etiology of bone demineralization. Existing studies in JRA have been complicated
by problems in design, such as assessing pre- and
postpubertal patients as one group, by including
corticosteroid-treated patients, and by measuring regional BMD adjacent to inflamed joints (e.g., forearms).
In the present study, non-corticosteroid-treated
prepubertal JRA patients were standardized for total
body (TB) BMD against a cohort of 25 similar age- and
sex-matched, healthy prepubertal controls. On this basis,
a cohort of 14 osteopenic JRA patients were identified,
as well as 34 nonosteopenic JRA patients, who then
could be compared with each other. The objective of
this study was to measure total body bone mineralization, anthropometrics, disease activity, dietary intake,
physical activity, and calcitropic hormones in prepubertal children with JRA who have not received
corticosteroids in order to identify those variables
related to the disease itself that are most predictive of
TB BMD in prepubertal JRA.
Subjects. Forty-eight JRA patients (37 female, 11
male) and 25 healthy controls (14 female, 11 male) were
evaluated in the Clinical Research Center of Children’s Hospital Medical Center (Cincinnati, OH). Subjects were from
similar geographic areas, were ages 4.6-11.0 years, and were
prepubertal (Tanner stage I or 11) (21). All subjects were white.
Study participants were not and had not inhaled, ingested, or
injected corticosteroids and had no past or current chronic
illness except for JRA.
All JRA patients had a definite diagnosis according to
American College of Rheumatology diagnostic criteria (22).
At the time of the study, 21% had active disease, 60% were in
partial remission, 19% were in total remission, 52% were in
Steinbrocker functional class I, and 48% were in functional
class 11. The JRA disease course subtype was systemic in 2
(4.2%), polyarticular in 28 (58.3%), and pauciarticular in 18
(37.5%). Indomethacin, salicylates, and methotrexate (MTX)
were being taken by 1 (2.1%), 5 (10.4%), and 19 (40.4%) of the
JRA patients, respectively.
Informed written consent was obtained from the parents, and assent was obtained from the children. The study was
approved by the Institutional Review Board and the Radiation
Safety Committee.
Study design. This cross-sectional study was performed
from January 1992 to April 1996. Subjects were enrolled
year-round to avoid possible seasonal variations in sun exposure that can contribute to detectable differences in TB BMD.
The same individual (CJH) performed recruitment, instruction, and followup for each component of the study. During the
3 days prior to the study date, including 1 weekend day,
physical activity was measured using 3 standardized methods
simultaneously: the Caltrac accelerometer (Hemokinetics,
Madison, WI), the University of Cincinnati Motion Sensor
(UCMS), and a physical activity diary. A 3-day diet diary was
recorded during the same time. A 12-hour overnight urine
collection was obtained just prior to the study visit and was
analyzed for calcium, creatinine, magnesium, and phosphorus
On the study day, following an overnight fast, height,
weight, anthropometrics, total body Hologic (Waltham, MA)
dual energy x-ray absorptiometry (DXA) scanning, and blood
drawing were performed. Anthropometrics were measured
and assessed according to published methods and were reported as percentiles compared with age- and sex-matched
norms (23,24). A standard rheumatology joint exam was
performed on the JRA patients (25). The Juvenile Arthritis
Functional Assessment Report (JAFAR), a validated questionnaire designed to measure physical function limitation in
routine daily activities in children with JRA, was administered
to the parents of participating JRA patients (26). Assessment
of the Steinbrocker functional class (27) was performed by one
rheumatologist (DJL).
The Caltrac was used to assess truncal movement in
the vertical plane (28,29). Caltrac counts were averaged for the
3 days it was worn. Caltrac standardization methods established in previous investigations (30,31) were modified slightly
for this study.
The UCMS was used to assess lower extremity movement. The UCMS counts once for each movement of 10” or
more from the horizontal plane. Counts for the UCMS were
analyzed each day, then averaged for the 3 days worn.
Physical activity was recorded in a 3-day physical
activity diary, based on a modified 7-day Stanford Physical
Activity Recall Record (32). Parents recorded to the nearest
15-minute interval the number of hours slept and all activities
performed for 3 days. Activities recorded were rated according
to Puhl’s 5-level rating scale (33). The total number of hours
spent performing the recorded activity combined with the total
hours slept was subtracted from 24 hours. A child was assumed
to be engaged in “light activity” for the remaining hours. The
total diary score was averaged for the 3 recording days.
Participation in organized sports during the previous 12
months was determined by parental completion of a
Parents completed the 3-day diet diary, which was
analyzed using the University of Minnesota Nutrition Data
System (version 3.0, 1996). Intakes were compared with agematched US recommended daily allowances (RDAs) for those
having nutrient guidelines.
DXA (model QDR-lOOO/W; Hologic) was used to
determine TB BMD and BMD in the skull, arms, hips, legs,
and trunk (including the rib cage, thoracic and lumbar spine,
and pelvis). In addition, DXA was used to determine lean body
and fat masses. The coefficient of variation for reproducibility
for the actual DXA machine utilized in this study was -0.9%
for TB BMD in children ages 5-10 years (unpublished data).
TB BMD was expressed as the amount of mineral (gm) divided
by the bone area (cm’). The mean +- SD TB BMD did not
differ between boys (0.75 ? 0.06 gmicm”) and girls (0.74 2 0.06
gmicm’), even when adjusted for weight (P > 0.30). Therefore,
TB BMD data from male and female children were combined.
Statistical analysis. The Kolmogorov-Smirnov test
(34) was used to determine the normality of the data before
statistical tests that required a Gaussian distribution were
performed. For non-Gaussian data, the Wilcoxon 2-group rank
sum test was utilized. Tests used included Student’s 2-tailed
t-test, Fisher’s exact test, and the log-likelihood ratio (G’).
Pearson’s product moment correlations were used to compare
independent predictor variables with the outcome variable.
Analyses were performed utilizing PC!INFO statistical software (Retriever Data Systems, Seattle, WA) or SPSS 7.0 for
Windows (SPSS, Chicago, IL).
TB BMD Z scores for JRA subjects were calculated in
the following manner (34): JRA subject’s TB BMD (gmicm’)
- control population’s mean TB BMD value (0.75 gm/
cm2)iTB BMD SD for control population (0.05). The JRA
patients were grouped as “low” TB BMD (TB BMD Z score
5 - 1) and “normal” TB BMD (TB BMD Z score > - 1) for
comparative statistical analysis. This cut point was utilized
based on adult data that indicate an increased fracture risk in
women with BMD >l SD below normal (3,35-37) and the
1994 World Health Organization report which defines osteopenia as TB BMD 2 - 1 to -2.5 SD below normal (38).
Two approaches were used in performing logistic
regression analyses in this project. One approach used principal component analysis (a variable-reduction scheme which
transforms original variables into a new set of uncorrelated
linear combinations). The variables were determined to be
uncorrelated by varimax and oblique rotation. Components
retained for interpretation met these criteria: eigenvalue >1,
communalities 20.70 for original variables comprising each
component, and Cattle’s scree test lying within the curve’s
sharp decent before leveling off accounted for ~ 7 5 %
of the
overall variance, and each factor loading was significant at (Y =
0.01. Logistic regression was performed using the method of
maximum likelihood to determine the binomial outcome variable, low versus normal TB BMD.
The second analytical method for performing logistic
regression analysis utilized 5 recognized factor domains associated with bone mineralization (anthropometrics, diseaserelated, dietary intake, physical activity, and calcitropic
hormonesiminerals). Each of the 45 predictor variables was
placed into 1 of these 5 domains, and the variable within each
factor domain having the highest Pearson correlation with
mean TB BMD Z score for all JRA subjects was selected to
Table 1. Distribution of normal versus low TB BMD in 48 prepubertal patients with JRA”
Observed population
Expected population
*Normal = total body bone mineral density (TB BMD) Z score
>-1.0, n = 34 (70.8% of juvenile rheumatoid arthritis [JRA] patients); low = TB BMD Z score <-l.O, n = 14 (29.2% of JRA
patients). Based on Gaussian distribution, expected populations would
be n = 40 (84% of JRA patients) for normal and n = X (16% of JRA
patients) for low TB BMD. Goodness of fit x’ with continuity
correction = 4.84, P = 0.01.
represent that factor domain. Logistic regression was performed using low versus normal TB BMD as the outcome
variable and the 5 predictor variables in the predictive model.
The control subjects were included in this study
to provide normal standards for the evaluation, since the
overall study objective was to compare the JRA patients
who had low TB BMD with those who had normal TB
BMD. However, to demonstrate how the controls and
JRA subjects compared, the following data are included.
Demographic characteristics (mean 2 SD) for control
subjects compared with the JRA patients, respectively,
were 7.7 2 1.7 and 8.1 t 1.9 years for age, 58.9 ? 26.1
and 44.5 2 29.5 percentile for weight, 55.2 2 23.0 and
45.5 2 28.4 percentile for height, and 63.3 5 20.4 and
58.5 +- 30.3 percentile for body mass index. No difference in the mean ? SD TB BMD was detected between
the control and JRA groups (0.75 ? 0.06 and 0.74 ? 0.07
gm/cm2, respectively, P > 0.30).
Subsequent analyses of data were directed toward determining the role of disease-related variables in
decreased TB BMD in JRA patients. The distribution of
the values showed that a significantly larger-thanexpected proportion of the JRA subjects had low TB
BMD (x2= 4.84, P = 0.01) (Table 1). The mean t SD
TB BMD Z score was - 1.43 ? 0.31 for the low TB BMD
group and 0.32 ? 0.97 for the normal TB BMD group
(P = 0.001). The mean 2 SD TB BMD value was 0.67 2
0.02 gm/cm2 for the low group and 0.77 2 0.06 gm/cm2
for the normal group (P = 0.001).
Table 2 summarizes demographic and disease
characteristics for the low and normal TB BMD JRA
groups. The disease course for the low as compared with
the normal TB BMD group was systemic 7.1% versus
2.9%, polyarticular 78.6% versus 52.9%, and pauciarticular 14.3% versus 44.2% (G2 = 4.4, P = 0.11). In 3
of 5 JRA subjects taking salicylates, the TB BMD was
Table 2. Demographic and disease characteristics of normal versus low TB BMD JRA patient groups4.
Age at study, years
Age at disease onset, years
Disease duration, years
No. of swollen joints
No. of involved joints
Disease severity score?
Duration of AM stiffness, minutes
JAFAR score?
ESR, tnm/hour$
(n = 48)
JRA patients
with normal BMD
(n = 34)
JRA patients
with low BMD
(n = 14)
8.1 5 1.9
4.1 t 1.8
3.8 -t 2.2
5.9 -+ 9.1
8.1 -+ 11.0
15.0 i 27.1
22.6 i- 34.0
7.5 -C 8.9
14.5 t 10.9
8.7 i 1.7
4.5 1 2 . 7
3.9 ? 2.4
4.6 t 8.4
2.5 t 11.0
12.8 1 3 0 . 1
15.7 i 33.3
5.3 t 7.4
12.4 i- 8.7
6.7 5 1.7
3.2 t- 1.8
3.5 i 1.9
9.1 i- 8.7
12.2 ? 10.4
20.4 -+ 17.5
35.6 2 34.2
13.3 t 9.9
19.9 I
normal vs.
low BMD
* Values are the mean i SD. P determined by Student's 2-tailed t-test. TB BMD = total body bone
mineral density; JRA = juvenile rheumatoid arthritis.
t Each joint was evaluated for 4 indices (swelling, tenderness, loss of motion, and deformity) using graded
scales. The severity score is the sum of the scores for all indices for all involved joints.
$ For comparison, published normative data on Juvenile Arthritis Functional Assessment Report
(JAFAR) scores for healthy controls are mean 0.6, median 0, and SD 0.9.
9: Westergren erythrocyte sedimentation rate.
low. Ten of the 19 subjects receiving MTX (52.6%) had
low TB BMD Z scores, and 9 of the 29 subjects not
receiving MTX (31.0%) had low TB BMD Z scores (x2
= 0.94, P = 0.33).
There were several significant disease-related
characteristic differences between the low and normal
TB BMD JRA groups. The mean age at the time of
evaluation for patients with low TB BMD was significantly younger than those with normal TB BMD (P =
0.001) even though both groups were similar in mean
age at disease onset and in duration of disease. Because
of the strong association of BMD with age in children,
TB BMD values were age adjusted; the 2 JRA groups
remained significantly different ( P = 0.01). Significant
differences were also seen in indicators of the extent of
articular involvement. The patients with low TB BMD
had, on average, more swollen joints ( P = 0.04), many
more involved joints over the course of the JRA ( P =
0.003), greater total articular severity score (P = 0.004),
and higher erythrocyte sedimentation rate (ESR) ( P =
0.001). The trend was for those with low TB BMD to
have more morning stiffness, which was in keeping with
the other articular parameters. According to the JAFAR
scores, patients with low TB BMD had significantly
more physical function limitation in routine daily activities ( P = 0.001).
There were also differences in anthropometric
measurements between the 2 groups (Table 3). The low
TB BMD patients were both taller and lighter, resulting
in significantly lower body mass index (P = 0.04). This
difference in body mass index was a reflection of both fat
and lean body components, since the patients with low
Table 3. Anthropometric measurements in normal versus low TB BMD JRA groups''
Height, percentile
Weight, percentile
Body mass index, % i t
Arm muscle circumference,
Triceps skinfold thickness,
Total lean, trunk (gm)
Total fat, trunk (gm)
Total percent lean
Total percent fat
(n = 48)
JRA patients
with normal BMD
(n = 34)
JRA patients
with low BMD
(n = 14)
normal vs.
low BMD
45.5 i 28.4
44.5 2 29.5
58.5 i 30.3
30.5 2 2 8 . 2
44.3 i 29.1
46.1 -+ 30.6
64.1 1 2 8 . 8
44.4 1 2 8 . 1
48.2 i 27.5
40.6 1 2 7 . 4
44.8 2 30.3
27.7 2 2 5 . 4
50.0 f 31.6
55.9 t 32.6
35.9 t 24.7
9,235.6 2 2,309.3
1,807.7 5 1,472.7
73.7 2 6.3
23.1 i 5.1
9,900.0 2 2,3338.3
2,074.7 I
72.7 I
23.6 i- 5.8
7,399.6 11,200.7
1,159.3 1427.2
76.0 1 2 . 6
21.8 -i- 2.4
* Values arc the mean i SD. P determined by Student's 2-tailed t-test. See Table 2 for definitions.
7 Weight (kg)/height' (meters).
Table 4. Physical activity in normal versus low TB BMD JRA groups’
(n = 48)
JRA patients
with normal BMD
(n = 34)
JRA patients
with low BMD
(n = 14)
376.1 t 111.6
384.6 i 117.0
356.6 i 99.3
Caltrac, counts
UCMS, counts
5,305.2 2 2,956.6 5,324.1 i 3,139.3 5,261.8 i 2,599.1
Activity diary
1,304.3 t 130.8
1,310.7 i 141.7
1,289.3 i 103.7
Organized sports participation, c/o
Duration of organized sports
1.3 2 2.0
1.7 i 2.2
0.2 t 0.8
participation, monthsiyear
No. of hours of sleepiday
10.9 t 0.8
10.8 t 0.8
10.9 f 0.8
normal vs.
low BMD
Values are the mean ? SD. Except where noted otherwise, P determined by Student’s 2-tailed t-test.
UCMS = University of Cincinnati Motion Sensor. See Table 2 for other definitions.
1 By Fisher’s exact chi-square test.
TB BMD had lower triceps skinfold (TSF) percentiles
(P = 0.05) and truncal lean body mass (P = 0.001).
Physical activity was quantified in a number of
ways for this study (Table 4). JRA patients with low TB
BMD did not differ significantly from those with normal
TB BMD in truncal movement (Caltrac), lower extremity movement (UCMS), activity diary, and hours of
sleep. However, those with low TB BMD spent fewer
months per year participating in organized sports than
those with normal TB BMD (P = 0.03).
A comprehensive assessment of dietary factors
related to bone mineralization was performed (Table 5).
JRA subjects with low TB BMD ingested significantly
more protein per kilogram of body weight (P = 0.004)
and more vitamin D (P = 0.01) than did subjects with
normal TB BMD. Both groups were similar in all other
dietary parameters. Calcium intake for both groups was
low; 71.4% of those with low TB BMD and 82.4% of
those with normal TB BMD actually ingested less than
the National Institutes of Health (NIH) Consensus
Conference recommended 1,200 mg of calciumiday.
There were no differences seen for sodium, caffeine, and
fiber ingestion.
The laboratory data obtained (Table 6) did not
suggest a difference in bone formation rate or hyperexcretion of calcium in either TB BMD group. Although
serum phosphorus and magnesium levels were significantly higher in the low TB BMD group, the values were
within normal limits. Serum cytokine levels-tumor necrosis factor a, interleukin-1p (IL-lP), IL-6, and IL-8-
Table 5. Dietary-related measurements in normal versus low TB BMD JRA patients*
Protein, gmikg
Fat, gm
Dietary fiber, gm
Calcium, % NIH guidelines?
Dietary calcium, mgS
Magnesium, % RDA
Phosphorus, % RDA
Vitamin D, % RDA
Sodium, mg
Caffeine, mg
(n = 48)
JRA patients
with normal BMD
(n = 34)
JRA patients
with low BMD
(n = 14)
normal vs.
low BMD
1,946 i 407
2.5 ? 0.9
73.9 i 20.6
11.5 i 4.1
83 t 30
992.8 2 355.3
140.7 -t 49.8
136.8 -t 48.4
87.6 t 52.4
2,783 t 878
26.7 i 26.7
1,954 -t 449
2.3 t 0.8
74.2 t 22.6
11.7 -t 4.5
81 t 31
977.8 i 363.9
139.3 -t 56.2
137.0 t 47.0
77.0 t 44.7
2,807 -t 972
29.6 -+ 30.3
1,932 t 336
3.1 i 0.8
73.2 i 15.1
11.0 t 3.3
83 t 30
1,029.0 i 343.6
127.2 -t 53.0
136.4 -t 53.4
113.4 t 61.8
2,725 t 630
20.2 i 14.7
0.0 I
* Values are the mean L SD. P determined by Student’s 2-tailed i-test. RDA = recommended daily
allowance. See Table 2 for other definitions.
i The National Institutes of Health (NIH) guidelines recommend daily dietary calcium intake of 1,200 mg
for a 6-10-year-old child.
$ These means are significantly lower than the NIH guideline of 1,200 mgiday, based on [-tailed t-test of
the observed mean compared with a constant (1,200 mgiday). P values for all JRA patients, those with
normal TB BMD, and those with low TB BMD, respectively, wcre <0.001, <0.001, and 0.04.
Table 6. Laboratory features in normal versus low TB BMD JRA groups”
1,25-dihydroxyvitaminD, pgiml
25-hydroxyvitamin D, nglml
Alkaline phosphatase, unitsiliter
Osteocalcin, ndml
Serum calcium, mgidl
Serum phosphorus, mgidl
Serum magnesium, mgidl
Urinary ca1cium:creatininc
Urinary calciumikgi24 hours
(n = 48)
JRA patients
with normal BMD
(n = 34)
JRA patients
with low BMD
(n = 14)
normal vs.
low BMD
31.8 t 8.0
35.8 -C 11.5
124.0 -C 53.8
13.4 -C 6.8
10.0 t 0.5
4.5 5 0.6
2.0 -t 0.2
0.17 -C 0.10
2.8 t 1.7
30.6 Z 3.2
35.6 Z 11.4
124.5 2 55.4
12.6 t 7.3
9.9 ? 0.5
4.3 i- 0.6
1.9 t 0.2
0.13 t 0.10
2.2 f 1.8
34.7 t 7.1
36.1 t 12.0
122.7 t 52.2
15.2 t 5.4
10.0 t 0.6
4.9 -C 0.4
2.1 i 0.1
0.18 t 0.08
3.3 t- 1.3
* Values are the mean t SD. P determined by Student’s 2-tailed t-test. See Table 2 for definitions.
were measured in a small subset of JRA patients (n =
17). The serum levels of these cytokines did not significantly correlate with the TB BMD and did not differ
between the low and the normal TB BMD groups. More
than 95% of the serum cytokine levels in both groups
were normal.
Regression analysis. Utilizing principal components factor analysis, 14 original variables were retained,
comprising 3 components, each having eigenvalues 2 1,
communalities 20.83, and a scree test which included 3
components (see Patients and Methods). The first component contained only dietary-related variables, the
second component contained anthropometric indices,
and the third component contained disease-related variables, such as the number of involved and swollen joints
and the articular severity score. These 3 principal components accounted for 80.7% of the variance seen in the
actual values for TB BMD ( P < 0.0005). Utilizing
logistic regression, these 3 principal components were
able to accurately segregate 70.8% of the J R A subjects into low versus normal TB BMD groups (x2 =
3.4, P = 0.06).
The second analytical method for trying to understand the relative impact of the variables on TB
BMD utilized 5-factor domains known to be associated
with TB BMD (anthropometrics, disease-related, dietary,
physical activity, and calcitropic hormone/minerals). The
1 predictor variable in each domain most highly correlated with TB BMD was selected to represent that
domain in the subsequent analysis. The 5 domain variables selected and the r values for the univariate correlations with TB BMD were as follows: TSF percentile
(r = 0.56, P = O.Ol), age at disease onset (r = 0.45, P =
0.01), magnesium intake as percentage of RDA (r =
-0.32, P = 0.05), JAFAR score (r = -0.36, P = 0.05),
and serum levels of 1,25-dihydroxyvitamin D (r = -0.48,
P = 0.05). Including all 5 variables in a linear regression
analysis explained 65.6% of the variance in TB BMD
seen in the JRA patients (P < 0.0005). Utilizing logistic
regression analysis with the dichotomous outcome of low
or normal TB BMD, these 5 variables were able to
accurately segregate 79.6% of the JRA subjects by
comparing predicted TB BMD Z scores with the subject’s actual measured TB BMD Z score (?= 20.5, P =
0.001) (Table 7).
This study was designed to comprehensively evaluate factors known to influence bone mineralization in a
very focused population of JRA patients. This approach
was utilized to permit better identification of those
factors influencing bone mineralization that are related
to the JRA and to identify appropriate factors to study
in future intervention trials. The rationale for study
population selection was physiologic. Bone mineralizaTable 7. Predictive model for determining normal versus low TB
Observed TB BMD
Normal TB BMD
* Predicted logistic regression analysis equation:
+ (0.0593)(TSF percentile) + (-0.1854)(JAFAR score)
+ (-0.0087)(magnesium RDA) + (-0.1 127)(1,25-OHvitamin D), where TB
(0.0780)(age at onset)
= total body, BMD = bone mineral density, TSF = triceps skinfold
thickness, JAFAR = Juvenile Arthritis Functional Assessment Report,
and RDA = recommended daily allowance. This model was able to
accurately segrcgate 79.6% of the juvenile rheumatoid arthritis patients into either normal or low TB BMD (x’ = 20.5, P = 0.001), based
on 44 subjects.
tion undergoes significant qualitative and quantitative
changes during puberty compared with prepuberty. This
study sought early indicators of abnormal mineralization
in JRA patients; therefore, all subjects studied were
prepubertal. To avoid the confounding influence of
corticosteroid treatment on bone mineralization, no
study subject was allowed to have ever received corticosteroids. This excluded many patients with more severe
JRA, and the subjects in this study had mild-tomoderate JRA severity. This may explain why this study
disagrees with those of other investigators who demonstrated significantly lower BMD in prepubertal JRA
patients compared with healthy controls (18,39,40).
However, a significant proportion of our JRA study
population had TB BMD that was at least 1 SD below
normal, which permitted univariate and multivariate
comparisons to be performed.
Results from this observational study identified
significant differences between the JRA patients with
low and those with normal TB BMD Z scores. JRA
subjects with low TB BMD were significantly younger,
but even after adjustment for age, significant differences
remained between the JRA groups. Low TB BMD was
significantly associated with a higher level of activity of
joint inflammation (number of swollen joints, ESR) and
overall severity of joint involvement (number of involved
joints, articular severity score). This suggests that one or
more mediators of the inflammatory process may also
negatively influence bone mineralization. Serum levels
of cytokines were not correlated with TB BMD and did
not differ between the two JRA groups. However, these
levels were measured in only a small subset of the JRA
patients, and the vast majority of the levels were normal.
Perhaps the results would have been different if the
cytokines had been measured in a larger sample size, in
joint fluid, or if the study population had had more
active articular disease at the time of study. Surprisingly,
height and weight, well recognized for exerting a large
positive influence on BMD in healthy children (8), did
not differ significantly between the low and normal TB
BMD groups. The body mass index did differ significantly and was lower in the low TB BMD group because
the JRA patients with low TB BMD were both taller and
The importance of physical activity in maintaining bone mass has been demonstrated by studies of
weightlessness and bed rest (41). A strong positive
relationship between physical activity and TB BMD has
been demonstrated in healthy children (42). Although
the JRA subjects participating in this study were uniformly ambulatory and reported participation in regular
daily activities (e.g., full-time school attendees and
Steinbrocker functional class I or 11), those with low TB
BMD had significantly more physical limitation in routine daily activities as measured by the JAFAR and
shorter duration of participation in organized competitive sports. These results suggest that increased physical
activity may prove beneficial to TB BMD in children
with JRA. However, trials would need to be performed
to test this possibility.
In this study, altered TB BMD seemed to be
related to only 2 dietary factors. Ingested animal protein
can significantly increase urinary calcium excretion in
healthy subjects, and the JRA patients with low TB
BMD Z scores ingested significantly greater amounts of
dietary protein that was derived primarily from animal
sources. Vitamin D metabolites enhance calcium absorption, and adequate calcium intake is necessary to
maintain skeletal endowment (43). The percentage
RDA for vitamin D intake was significantly greater in
the low TB BMD group compared with the normal TB
BMD group. This suggests that the ingestion of increased amounts of vitamin D alone is not sufficient to
support normal calcium metabolism in JRA patients.
This is consistent with the results of an earlier study of
BMD and vitamin D supplementation in JRA patients
(10). Perhaps the combination of vitamin D and calcium
supplementation would be more effective in enhancing
bone mineralization.
Based on a 1994 NIH Consensus Report, optimal dietary calcium intake for healthy children ages
6-10 years is thought to be 800-1,200 mg/day (1). The
NIH recommendation was based on results of prospective, randomized, controlled studies and a retrospective analysis of more than 500 individual calcium
balance studies in growing children, which demonstrated clear evidence of threshold behavior for calcium intake on maximal bone mass. In other words,
calcium balance was a linear function of intake up to
the threshold saturation level (- 1,200-1,400 mg dietary calciumiday for children 2-8 years of age).
Intakes above this threshold saturation level did not
produce further enhanced bone accumulation (6).
These data would suggest that the optimal daily
calcium intake recommendation for healthy children
should be at the upper end of the NIH guidelines-1,200
mg/day. The mean dietary calcium intake for both the
control and JRA populations was significantly lower
than this recommended amount. In fact, >70% of the
JRA patients in both the low and normal TB BMD
groups received <1,200 mg/day of dietary calcium. The
actual optimal amount of dietary calcium intake for JRA
subjects may even be higher, given the presence of
several negative influences on bone mineralization in
this chronically ill population. Perhaps more aggressive
calcium supplementation to meet or exceed the NIH
guidelines for healthy children would result in improved
bone mineralization in JRA patients. Excessive dietary
sodium, caffeine, and fiber intake may have a negative
effect on bone mineralization (1,44), but no differences
were seen in the mean intakes for any of these nutrients
in the low and normal TB BMD JRA groups.
Two methods of regression analysis were utilized.
Principal components analysis defined 3 distinct physiologic areas-dietary intake, anthropometric measures,
and articular inflammation measures. The second approach was less mathematically driven but resulted in a
model that predicted a significant proportion of the
variance in TB BMD, both as a continuous and a
dichotomous outcome variable. Both regression approaches support the univariate analyses in highlighting
the multifactorial impact of JRA on bone mineralization.
The effects of puberty, interventions that increase dietary calcium intake, increased physical activity,
and the concurrent evaluation of JRA patients’ ability to
comply with these interventions as a means to enhance
bone mineralization warrant further study. A prospective study following a cohort through puberty and into
young adulthood or a cross-sectional study of postpubertal young adults with JRA is also needed to help answer
questions about the capacity for “catch up” of bone
mineralization and the effect of JRA on peak bone mass.
A recent study by Hopp et a1 suggests that JRA patients
do not actually “catch up” during puberty, but instead,
show a much slower than normal rate of bone mineralization (18).
It is important in studies conducted to evaluate
bone mineralization in J R A to separately analyze
children treated with corticosteroids. The effects of
other medications used in J R A on TB BMD are more
ambiguous. While some evidence suggests that MTX
may cause osteopenia in children treated for malignancies and adults treated for rheumatoid arthritis
(45,46), that effect was not seen in this study. Perhaps
the significant beneficial effect on articular inflammation may actually have a positive effect on bone
mineralization that overrides the demonstrated inhibitory effect of MTX on osteoblasts (47). Further study
in JRA patients of the effect of MTX and other
pharmacologic and nonpharmacologic interventions
on bone mineralization are indicated.
The authors gratefully acknowledge the following individuals for their generous assistance with and contributions
to this project. Many thanks to Drs. Joseph Levinson and
Murray Passo for giving us the opportunity to study their
patients, to Dr. James Heubi, Program Director of the CRC, to
Gemma Uetrecht for her unfailing reliability in conducting the
DXA scans, to Lee Peng, Research Assistant, for serum
cytokine analysis, to Pat Ziegler for calibrating the UCMS
monitors, to Dr. Phil Hashkes, for calibrating the Caltrac
accelerometers, and to Dr. David Glass for his valuable and
much appreciated editorial assistance.
1. NTH Consensus Development Panel on Optimal Calcium Intake:
Optimal calcium intake. JAMA 272: 1942-1948, 1994
2. Nevitt MC: Epidemiology of osteoporosis. Rheum Dis Clin North
Am 20:535-558, 1994
3. Washnich RD, Ross PD, Davis JW, Vogcl JM: A comparison of
single and multi-site BMC measurements for assessment of spine
fracture probability. J Nucl Med 30:1166-1171, 1989
4. Blair SN, Horton E, Leon AS, Lee I-M, Drinkwater BL, Dishman
RK, Mackey M, Kienholz ML: Physical activity, nutrition, and
chronic disease. Med Sci Sports Exerc 28:335-349, 1996
5. Kreipe RE: Bones of today, bones of tomorrow. Am J Dis Child
146:22-25, 1992
6 . Matkovic V: Calcium metabolism and calcium requirements during skeletal modeling and consolidation of bone m
Nutr 54 (suppl 1):245S-260S. 1991
7. Ott SM: Attainment of peak bone mass. J Clin Endocrinol Metab
71:1082A-lO82C, 1990
8. Southard RN, Morris JD, Mahan JD, Hayes JR, Torch MA,
Sommer A, Zipf WB: Bone mass in healthy children: measurement with quantitative DXA. Radiology 179:735-738, 1991
9. De Schepper J, Derde MP, van den Broeck M, Piepsz A, Jonckheer MH: Normative data for lumbar spine bone mineral content
in children: influence of age, height, weight, and pubertal stage.
J Nucl Med 32:216-220, 1991
10. Hillman L, Cassidy JT, Johnson L, Lee D, Allen SH: Vitamin D
metabolism and bone mineralization in children with juvenile
rheumatoid arthritis. J Pediatr 121910-916, 1994
11. Reed AM, Haugen M, Pachman LM, Langman CB: Repair of
osteopenia in children with juvenile rheumatoid arthritis. J Pediatr
122:693-696, 1903
12. Reed A, Haugen M , Pachman LM, Langman CB: 25Hydroxyvitamin D therapy in children with active juvenile rheumatoid arthritis: short term effects on serum osteocalcin levels and
bone mineral density. J Pediatr 119:657-660, 1991
13. Henderson CJ, Lovell DJ: Bone mineral content in juvenile
rheumatoid arthritis-pilot project results. J Rheumatol 18 (suppl
31):1-22, 1991
14. Lovell DJ, Henderson CJ, Specker B, Sierra R, Abrams S, Yergey
A, Vieira N: Relationship of total body bone mineral density and
physical activity in JRA-a controlled study using dual cnergy
x-ray absorptiometry (abstract). Arthritis Rheum 35 (suppl9):S57,
15. Elasser U, Willkins B. Hesp R: Bone rarefraction and crush
fractures in juvenilc chronic arthritis. Arch Dis Child 57377-380,
16. Martel W, Holt JF, Cassidy JT: Radiologic manifestations of
juvenile rheumatoid arthritis. AJR Am J Koentgeiiol 88:400-424,
17. Reed A, Haugen M, Pachman LM, Langman CB: Abnormalities
in serum osteocalcin values in children with chronic rheumatic
diseases. J Pediatr 116574-580, 1990
18. Hopp R, Degan J, Gallagher JS, Cassidy JT: Estimation of bone
mineral density in children with juvenile rheumatoid arthritis.
J Rheumatol 18:1235-1239, 1991
19. Hickman PL, Allen SH, Hillman LS, Cassidy JT: Skeletal density
and bone mineral metabolism in children with juvenilc rheumatoid
arthritis (abstract). Arthritis Rheum 36 (suppl 9):S54, 1993
20. Fantini F, Beltrametti P, Gallazzi M, Gattinara M, Gerloni V,
Murelli M, Parrini M: Evaluation by dual-photon absorptiometry
of bonc mineral loss in rheumatic children on long term treatment
with corticosteroids. Clin Exp Rheumatol 9 (suppl 6):21-28, 1991
21. Tanner JM: Growth at Adolescence. Second edition. Oxford,
Blackwell Scientific, 1962
22. Cassidy JT, Levinson JE, Bass JC, Baum J, Brewer EJ Jr, Fink
CW, Hanson V, Jacobs JC, Masi AT, Schaller JG, Fries JF,
McShane D, Young D: A study of classification criteria for a
diagnosis of juvenile rheumatoid arthritis. Arthritis Rhcum 29:
274-281, 1986
23. Lohman TG, Roche AF, Martorell R: Anthropometric Standardization Reference Manual. Champaign, IL, Human Kinetics
Books, 1988
24. Frisancho RA: New norms of upper limb fat and muscle areas for
assessment of nutritional status. Am J Clin Nutr 34:2540-2545, 1981
25. Brewer EJ, Giannini EH: Standard methodology for segment 1, 11,
and TI1 Pediatric Rheumatology Collaborative Study Group studies. J Rheumatol 9109-113, 1982
26. Howe S, Levinson J, Shear E, Hartncr S, McGirr G, Schulte M,
Lovell D: Development of a disability measurement tool for
juvenile rheumatoid arthritis: the Juvenile Arthritis Functional
Assessment Report for children and their parents. Arthritis
Rheum 34373-880, 1Y91
27. Steinbrocker 0 ,Traeger CH, Batterman RC: Therapeutic criteria
in rheumatoid arthritis. JAMA 140:659-662, 1949
28. Klesges RC, Klesges LM, Swenson AM, Pheley A: An evaluation
of two motion sensors in the prediction of child and adult physical
activity levels. Am J Epidemiol 122:400-410, 1987
29. Klesges LM, Klesges RC: Assessment of physical activity in
children: a comparison of electronic monitoring with behavioral
observation. Med Sci Sports Exerc 19:511-517, 1987
30. Sallis JF, Buono MJ. Roby DC, Nelson JA: The Caltrac accelerometer as a physical activity monitor for school-age children. Mcd
Sci Sports Exerc 22:698-703, 1990
31. Hunter GR, Montoye HJ, Webster JG, Demment R, Ji 11, Ng A:
The validity of a portable accelerometer for estimating energy
expenditure in bicycle riding. J Sports Med Phys Fitness 29:218222, 1989
32. Blair SN, Haskell WL, Ho P, Paffenbarger RS, Vranizan KM,
Farquhar JW, Wood PD: Assessment of habitual physical activity
by a seven-day recall in a community survey and controlled
experiments. Am J Epidemiol 122:794-804, 1985
33. Puhl J, Greaves K, Hoyt M, Baranowski T: Children’s Activity
Rating Scale (CARS): description and calibration. Res Q Exerc
Sport 6196-36, 1990
34. Zar JH: Biostatisticdl Analysir. Second edition. Englewood Cliffs,
NJ, Prcntice-Hall, 1984
35. Johnston CC, Slemenda CW, Melton LJ: Clinical use of bone
densitometry. N Engl J Med 324:1105-1109, 1991
36. Hui SL, Slemenda CW, Johnson CC: Baseline measurement of
bone mass predicts fracturc in whitc women. Ann Intern Med
111:353-36 1, 1989
37. Cummings SR, Black DM, Nevitt MC: Appendicular bone density
and age predict hip fracture in women. JAMA 263:665-668, 1990
38. Alexeeva L, Burkhardt P, Christiansen C, Cooper C. Delmas P;
Johnell 0, Johnston C, Kanis JA, Lips P, Melton LJ, Meunier P,
Seeman E, Stepan J, Tosteson A: Assessment of fracture risk and
its application to screening for postmenopausal osteoporosis.
WHO Tech Rep Ser 834:l-129, 1994
39. Bianchi ML, Bardare M, Caraceni MP, Cohen E, Falvella S: Bone
metabolism in juvenile rheumatoid arthritis. J Bone Miner Res
9: 153-162, 1990
40. Pepmueller PH, Cassidy JT, Allen SII, Hillman LS: Bone mineralization and bone mineral metabolism in children with juvenile
rheumatoid arthritis. Arthritis Rheum 39:746-757, 1996
41. Krolner B, Toft B: Vertebral bone loss: an unheeded side effect of
therapeutic bed rest. Clin Sci 64537-540, 1983
42. Benedict M, Ranz J, Uetrecht G, Specker B: Bone mineral density
in female athletes 7-10 years of age (abstract). Pediatr Res 31:87A,
43. Levenson D1, Bockman RS: A review of calcium preparations.
Nutr Rev 52:221-232, 1994
44. Harris SS: Caffeine and bone loss in healthy postmenopausal
wornen. Am J Clin Nutr 60:573-578, 1994
45. Gilsanz V, Carlson M, Roe TF, Ortega JA: Osteoporosis after
cranial irradiation for acute lymphoblastic leukemia. J Pediatr
1171238-244, 1990
46. Scheven BA, van der Veen MJ, Damen CA, Lafeber F, van Rijn
H, Bijlsma J, Duursma SA: Effects of methotrexate on human
osteoblasts in vitro: modulation by 1,25-dibydroxyvitamin D,.
J Bone Miner Res 10:874-880, 1995
47. May KP, Mercill D, McDermott M T , West SG: The effect of
methotrexate on mouse bone cells in culture. Arthritis Rheum
39:489-494, 1996
Без категории
Размер файла
931 Кб
non, corticosterone, children, density, bones, predictor, tota, arthritis, body, juvenile, prepubertal, treated, mineraly, rheumatoid
Пожаловаться на содержимое документа