1. Top of page
  2. Abstract
  7. Acknowledgements


To characterize local bone geometry, density, and strength, using peripheral quantitative computed tomography (pQCT), compared with general bone characteristics as measured using dual x-ray absorptiometry (DXA), and to assess their relationship to disease-related factors in children with juvenile rheumatoid arthritis (JRA).


Forty-eight children ages 4–18 years with JRA (17 pauciarticular, 23 polyarticular, 8 systemic) were compared with age-matched healthy controls (n = 266). Measurements included cortical and trabecular bone geometry, density, and strength at the distal and midshaft tibia determined by pQCT, and whole-body, lumbar spine, and femoral neck measurements by DXA.


Methotrexate (MTX) was prescribed to 23 of 48 patients (47.9%) and glucocorticoids and MTX were prescribed to 15 of 48 patients (31.3%), with the greatest use in children with systemic JRA. All JRA patients had decreased tibia trabecular bone density, cortical bone size and strength, and muscle mass. Children with systemic JRA had lower femoral neck densities. Systemic JRA was associated with a shorter, less mineralized skeleton, while a narrower, less mineralized skeleton was observed in polyarticular JRA. The tibia diaphysis was narrower with decreased muscle mass, but normal, size-adjusted bone mineral in all subtypes indicated a localized effect of JRA on bone. Patients exposed to glucocorticoids and MTX or to glucocorticoids or MTX alone had greatly reduced trabecular density, cortical bone geometry properties, and bone mineral content, muscle mass, and bone strength.


Children with JRA have decreased skeletal size, muscle mass, trabecular bone density, cortical bone geometry, and strength. Not surprisingly, these bone abnormalities are more pronounced in children with greater disease severity.

Decreased bone mass is commonly found in children and adolescents with juvenile rheumatoid arthritis (JRA) (1–5) and has been observed at the onset of the disease (6) and throughout childhood into adulthood, thus increasing the risk of osteoporosis (1) and fracture (7, 8) later in life. Many factors are thought to contribute to diminished bone mass in children with JRA, including the pathology of the disease itself, decreased physical activity due to pain and reduced joint motility (9), and glucocorticoid treatment for inflammation (6, 10, 11). Early in the course of the disease, JRA often begins as a localized inflammation, which contributes to periarticular osteopenia, a contained loss of bone mineral content (BMC) around the inflamed joint (12). In the later stages of the disease process, in association with chronic inflammation of the affected joints, a generalized reduction in total-body bone mass occurs. Diminished bone mass predominates in children with polyarticular and systemic JRA (8), the more severe of the 3 subtypes.

Treatment with systemic glucocorticoids appears to accelerate bone loss if the cumulative dose is high. However, children with JRA who were never exposed to glucocorticoids also exhibit reduced bone mineralization (5), suggesting that the disease process itself plays a role in bone mass metabolism. Methotrexate (MTX), a disease-modifying antirheumatic drug, is currently prescribed to an estimated 25,000 children in the US, or 39% of patients with JRA (13). Low-dose MTX therapy has been reported to have a sparing effect on whole-body and lumbar spine (L-spine) BMC when used instead of or in addition to systemic glucocorticoids (14–17). In children also treated with glucocorticoids, a slower rate of bone mineral acquisition has been observed.

The majority of studies that have evaluated bone mass in patients with JRA have used dual x-ray absorptiometry (DXA) (5, 7, 8, 10, 11, 17). Although DXA is a practical and common tool for the assessment of bone health, the application of DXA to pediatric populations has several limitations. DXA provides a 2-dimensional reading of the 3-dimensional bone and is unable to distinguish between the cortical and trabecular bone compartments. DXA measures an area of bone (in cm2) and BMC (in mg) and expresses bone mineral density (BMD) in mg/cm2. When comparing 2 bones of the same true density but of different sizes, the larger bone will artificially present a greater BMD value than the smaller bone. This limitation will lead to an overestimation of BMD in smaller youths, a problem that can lead to costly and inaccurate interpretations. These size considerations may be especially important when assessing a population of chronically ill children, such as those with JRA, where the disease has been shown to negatively impact growth (5).

Computed tomography, another method of bone assessment, can provide separate measurements of both cortical and trabecular bone characteristics, including true volumetric BMD (vBMD). Peripheral quantitative computed tomography (pQCT) measures appendicular sites such as the radius or tibia, uses a low local radiation dosage, and is relatively fast and inexpensive. Quantification by pQCT of bone geometry and density properties in children and adolescents with JRA has been limited to the radius (18, 19). We prefer the tibia to the radius for appendicular bone assessment, due to its larger bone size and muscle area and higher rate of bone turnover due to mechanical loading (20). Therefore, this study was undertaken to characterize tibia bone features in patients with JRA as assessed by pQCT in comparison with healthy, age-matched controls, and to identify disease-related factors that predict diminished or altered bone geometry, density, or strength in the 3 JRA subtypes.


  1. Top of page
  2. Abstract
  7. Acknowledgements


Forty-eight patients (11 male, 37 female) from the Intermountain States Database of Childhood Rheumatic Diseases who had been diagnosed as having JRA according to the American College of Rheumatology criteria (21) were included in this cross-sectional study. There were 17 children with pauciarticular JRA (<5 joints affected in the first 6 months of the disease), 23 with polyarticular JRA (≥5 joints affected in the first 6 months of the disease), and 8 with systemic JRA. The controls consisted of 266 healthy, age-matched subjects (123 male, 143 female) living in the same geographic area. Informed consent was obtained from all parents. The study complied with Health Insurance Portability and Accountability Act regulations and was approved by the University of Utah Institutional Review Board for Human Subjects.

Data collection.

Demographic and disease data were extracted from the patients' medical records. These included age at disease onset, sex, disease-related laboratory test results, including the presence of IgM rheumatoid factor (RF), antinuclear antibodies, and HLA–B27, as well as date of JRA onset and course subtypes. In addition, data on duration of disease as well as joint involvement at the time of the pQCT and DXA measurements were extracted from the patients' medical records. MTX and glucocorticoid exposure data were included if the systemic (oral) administration was ≥4 months. Data obtained on patients who received any MTX and glucocorticoids included height, weight, duration of glucocorticoid/MTX therapy (months), weekly/daily dosage, and folic acid supplementation (1 mg/day).

Patients who had missing information regarding disease activity or medication taken, and those with other medical conditions known to affect growth or bone health, such as cystic fibrosis, treatment with oral contraceptives, and/or hormone therapy, were excluded from the study. The disease was considered to be active if the patient was receiving medication and had joint symptoms and synovitis; the disease was considered to be inactive or in remission when there were no reported signs of active synovitis with or without medication. From the medical records, duration of active disease was calculated as time of disease onset to the time of the study.

At the time of the bone measurements, each study participant (or parent) completed a health history questionnaire that included family and personal medical history, medications currently taken, fracture history, and self-reported Tanner growth stage. The self-reported questionnaire of pubertal maturation according to Tanner stage criteria is based on breast/areolar development for girls and testicular development for boys (22). Self-reported Tanner staging may be less accurate than a physical examination, particularly for Tanner stage 3–4 breast or testicular development (23). To ensure consistency for comparison of the JRA cohort with the control population, both groups completed the same questionnaire.

Calcium intake estimates (24) and data on past-year physical activity (25) were also obtained by questionnaire. The latter was used to estimate metabolic hours of weight-bearing physical activity in the previous year. Each weight-bearing activity was assigned a number representing metabolic equivalent (METS). Hours per week of each activity were multiplied by their estimated METS value to obtain METS hours per week (26). Using a stadiometer (Height-Rite, Model 225; Seca, Culver City, CA), height (m) without shoes was measured for each participant, and weight (kg) was measured using a digital scale (Stand on Scale, Model 5002; Scale-Tronix, White Plains, NY). Body mass index (BMI; kg/m2) was calculated for all patients.

Two cross-sectional slices of the nondominant tibia were measured by pQCT (XCT-2000; Stratec Medical Systems/Orthometrix, White Plains, NY) at relative distances of 4% and 66% from the distal tibia growth plate. Dominance and nondominance were determined by asking whether the subject was right- or left-handed. The 4% distal cross-section was used to determine trabecular vBMD (mg/cm3). The 66% distal cross section (diaphysis) was used to determine cortical vBMD and BMC as well as the bone geometry properties: bone cross-sectional area (cortical cross-sectional area, including the marrow area), cortical cross-sectional area (bone cross-sectional area less marrow cross-sectional area), marrow cross-sectional area (bone cross-sectional area less cortical cross-sectional area), and cortical thickness (mm). Muscle cross-sectional area and the polar strength-strain index (pSSI; mm3) were also determined from the 66% distal cross-section.

Muscle cross-sectional area was used to examine the bone–muscle relationship (25), and the pSSI, a modification of the bone strength index (27), was used to assess bone strength. The pSSI is calculated as follows: (cross-sectional moment of inertia, or bone resistance to bending × bone stiffness calculated by cortical vBMD)/maximum BMD. The pSSI provides a method for applying changes in cortical bone content to the clinically significant variable, bone fracture. Analysis parameters and modes were as previously described (28).

In addition, whole-body bone area (cm2), BMC, lean body mass (kg) and percentage body fat, and femoral neck and L-spine bone area, BMC, and areal BMD were determined by DXA (DR4500A; Hologic, Waltham, MA). Femoral neck analysis by DXA is difficult in small subjects; therefore, this analysis was limited to children >8 years of age (29). Height for age, the ratio of whole-body bone area to height for age, and the ratio of whole-body BMC to bone area were assessed to determine whether bone mass was reduced due to short, narrow, or lighter bones, as described by Molgaard et al (30). Femoral neck and L-spine BMC and bone area values were used to determine bone mineral apparent density (BMAD; gm/cm3), as described by Katzman et al (31). The coefficient of variation (CV) for repositioning in adult volunteers using pQCT was <2.5% for trabecular and cortical bone vBMD and <1.0% for BMD measured by DXA in our laboratory. The daily CVs for calibration phantoms were 0.1% and 0.3% for pQCT and DXA, respectively. The same experienced radiology technician performed all measurements at the Bone and Body Composition Laboratory, Center for Pediatric Nutrition Research, University of Utah.

Statistical analysis.

Statistical analysis included an independent t-test to compare differences in demographic variables between the control population and JRA patients both by sex and by JRA subtype at disease onset. Chi-square analysis was used to identify any differences in categorical variables of interest between JRA patients and controls and within JRA subgroups. The pQCT- and DXA-derived bone characteristic values were converted to Z scores using data on healthy controls from the same geographic area (n = 266). Multivariate analysis of covariance and post hoc tests were used to compare means between JRA and control groups, e.g., bone measurements and body composition, using Tanner stage and height Z score as covariates. Bone results are reported as the adjusted mean (95% confidence interval [95% CI]) and were considered statistically significant if the 95% CI was not within the reference range. Correlation analysis was run between selected continuous variables. Stepwise linear regression was performed to identify which variable predicted bone characteristics. Statistical analyses were performed using SPSS-PC+ software (version 13.1; SPSS, Chicago, IL). P values less than 0.05 were considered significant.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Demographic characteristics.

Patients' demographic data by JRA subtype are presented in Table 1. Sex distribution, age, estimated calcium intake (mg/day), and history of previous fractures were similar between JRA patients (all subtypes) and controls. Pubertal maturation was delayed in the pauciarticular and systemic JRA cohorts (P ≤ 0.001). Mean height and BMI Z scores and lean body mass were lower in patients with systemic JRA compared with controls. Patients with systemic JRA also reported significantly less weight-bearing physical activity than patients with pauciarticular or polyarticular JRA or controls (P ≤ 0.01). The percentage of body fat was higher in patients with both the polyarticular and systemic JRA subtypes (P ≤ 0.01). Patients with polyarticular JRA were older at diagnosis and at the time of the study versus patients in the other subtypes (P ≤ 0.01). Neither disease duration nor disease activity at the time of study was different between subtypes.

Table 1. Demographic and clinical characteristics of the patients, by JRA subtype*
 Pauciarticular (n = 17)Polyarticular (n = 23)Systemic (n = 8)
  • *

    Except where indicated otherwise, values are the mean (95% confidence interval). P values were determined by independent t-test or chi-square analysis. JRA = juvenile rheumatoid arthritis; BMI = body mass index; METS = metabolic equivalent.

  • P ≤ 0.001 versus polyarticular JRA and healthy controls.

  • P ≤ 0.01 versus healthy controls.

  • §

    P ≤ 0.01 versus pauciarticular JRA, polyarticular JRA, and healthy controls.

  • P ≤ 0.01 versus systemic JRA.

  • #

    P ≤ 0.01 versus pauciarticular JRA.

  • **

    P ≤ 0.001 versus pauciarticular JRA and polyarticular JRA.

Sex, no. female/male13/420/34/4
Age, years10.8 (8.5, 13.2)12.0 (10.1, 13.8)9.4 (5.9, 12.8)
Tanner stage, %   
Height, Z score−0.3 (−0.8, 0.2)−0.2 (−0.7, 0.3)−1.0 (−2.0, 0.2)
Weight, Z score−0.3 (−0.9, 0.4)−0.2 (−0.7, 0.1)−0.2 (−0.9, 0.6)
BMI, Z score−0.2 (−0.9, 0.5)−0.3 (−0.8, 0.2)−0.6 (−0.4, 1.6)
Lean mass, kg24.3 (19.1, 29.5)26.4 (22.7, 29.6)21.1 (13.0, 29.1)
Body fat, %24.9 (21.3, 28.4)26.2 (21.7, 31.0)26.1 (13.0, 29.1)
Calcium intake, mg/day1,479.1 (1,154.9, 1,804.4)1,154.9 (955.3, 1,354.6)1,207.4 (590.2, 1,824.7)
Physical activity, METS hours/week45.4 (20.9, 70.9)40.7 (23.8, 57.6)26.0 (3.8, 48.2)§
History of fracture, %18.915.014.3
Age at disease onset, years5.1 (3.4, 6.7)8.6 (6.6, 10.5)5.8 (3.3, 6.8)
JRA duration, months59.8 (42.0, 100.8)51.7 (28.8, 61.2)51.8 (20.4, 82.8)
No. of patients with active disease at study initiation11177
Affected joints at study initiation, no.2.1 (1.1, 3.1)8.2 (4.5, 11.8)#9.1 (2.5, 15.8)#
 Exposure, %43.5#87.5**
 Duration of therapy, months7.9 (1.5, 14.2)30.5 (8.5, 52.4)**
 Average dose, mg/kg/day1.0 (−0.8, 2.8)2.9 (−2.6, 8.3)**
 Cumulative dose, gm1.1 (0.0, 2.3)6.6 (2.6, 10.5)**
 Exposure, %58.891.3#100.0
 Duration of therapy, months24.6 (8.0, 41.2)25.0 (16.7, 33.3)29.1 (15.1, 43.1)
 Average dose, mg/m2/week8.4 (4.6, 12.2)12.3 (9.5, 15.0)18.9 (15.3, 22.5)#
 Cumulative dose, gm1.3 (0.4, 2.3)1.9 (0.6, 3.2)2.3 (0.4, 4.2)

Patients with either polyarticular or systemic JRA had a greater number of inflamed joints at the time of the study as well as greater glucocorticoid exposure than patients with pauciarticular JRA (P ≤ 0.01). Seventeen patients (35.4%) had been treated with oral glucocorticoids (prednisone) for a minimum of 4 months, at dosages ranging from 0.08 to 1.00 mg/kg/day. Patients with pauciarticular JRA had either never received glucocorticoids or had not met the criteria for usage. Average cumulative lifetime glucocorticoid dose (mg/kg/day) was significantly higher in the systemic JRA cohort (P ≤ 0.001). Fifteen patients (31.3%) received glucocorticoids and MTX for a minimum of 3 months, and 39 patients (81.3%) received MTX, with a dose range of 5.6–22.8 mg/m2/week. Average MTX dose was significantly higher in the systemic JRA cohort when compared with patients in the pauciarticular JRA cohort (P ≤ 0.01). A significant percentage of patients with polyarticular JRA had a positive RF (43%; P ≤ 0.01), while patients with systemic JRA had a longer duration of glucocorticoid or MTX treatment, and had received a higher average and cumulative glucocorticoid dose (P ≤ 0.001).

All JRA patients were treated with nonsteroidal antiinflammatory drugs and all but 1 patient received folic acid supplementation (1 mg/day) during their MTX treatment. In all, 22 of the patients (45.8%) had received anti–tumor necrosis factor α (anti-TNFα) therapy (5 of 17 pauciarticular, 12 of 23 polyarticular, and 5 of 8 systemic JRA patients). Fifteen patients received etanercept, 1 patient had received infliximab, and 5 had received both etanercept and infliximab (at different times). All patients who received anti-TNFα therapy also received MTX, and 11 (50%) also received glucocorticoids.

DXA and pQCT results.

Femoral neck, L-spine, and tibia trabecular and cortical bone characteristics are presented in Table 2. Tibia bone parameters are grouped by geometry, density, and strength. Results are provided as the adjusted mean of the SD. Femoral neck BMD was lower in patients with polyarticular and systemic JRA versus controls, while BMAD was significantly lower in patients with systemic JRA versus controls. Patients with systemic JRA had the smallest reduction and patients with polyarticular JRA had the largest reduction in muscle cross-sectional area (P ≤ 0.01). L-spine BMD and BMAD values were similar to those in controls and similar among JRA subtypes. Tibia bone cross-sectional area in patients with polyarticular and systemic JRA, and cortical cross-sectional area and thickness in patients with all JRA subtypes, were significantly less than in controls (P ≤ 0.01). Within subtypes, bone cross-sectional area was lower in patients with polyarticular compared with pauciarticular JRA, while cortical thickness was significantly reduced in patients with systemic JRA compared with those with polyarticular JRA (P ≤ 0.01). The lowest cortical BMC was found in patients with polyarticular and systemic JRA (P ≤ 0.01, patients with polyarticular and systemic JRA versus controls, and P ≤ 0.01, patients with polyarticular JRA versus patients with pauciarticular JRA). Trabecular vBMD was greatly reduced in all JRA subtypes (P ≤ 0.001).

Table 2. SD scores of bone characteristics, by JRA subtype*
 Pauciarticular JRAPolyarticular JRASystemic JRA
  • *

    Values are the adjusted mean (95% confidence interval). P values were determined by multivariate analysis of covariance adjusted for height and Tanner stage. N values are patients with pauciarticular juvenile rheumatoid arthritis (JRA)/polyarticular JRA/systemic JRA. BMD = bone mineral density; BMAD = bone mineral apparent density; CSA = cross-sectional area; BMC = bone mineral content; vBMD = volumetric BMD; pSSI = polar strength-strain index.

  • P ≤ 0.01 versus healthy controls.

  • P ≤ 0.01 versus pauciarticular JRA and healthy controls.

  • §

    P ≤ 0.01 versus polyarticular JRA and healthy controls.

  • P ≤ 0.01 versus pauciarticular JRA, systemic JRA, and healthy controls.

Femoral neck (n = 12/19/5)   
 BMD, gm/cm2−0.19 (−0.96, 0.51)−0.71 (−1.25, −0.14)−0.85 (−1.96, 0.26)
 BMAD, gm/cm3−0.65 (−0.38, 0.09)−0.42 (−1.01, 0.18)−1.22 (−2.38, 0.06)
Lumbar spine (n = 17/23/8)   
 BMD, gm/cm2−0.12 (−0.79, 0.54)−0.26 (−0.85, 0.32)−0.54 (−1.52, 0.45)
 BMAD, gm/cm30.08 (−0.39, 0.55)−0.02 (−0.39, 0.42)−0.18 (−0.87, 0.51)
Tibia (n = 17/23/8)   
  Muscle CSA, mm2−0.74 (−1.37, −0.11)−0.87 (−1.43, −0.30)−0.33 (−1.37, 0.70)
  Total CSA, mm2−0.46 (−1.16, 0.24)−1.02 (−1.64, −0.29)−0.74 (−1.89, −0.41)
  Cortical CSA, mm2−0.78 (−1.37, −0.18)−1.16 (−1.69, −0.62)−0.88 (−1.86, 0.10)
  Cortical thickness, mm−0.57 (−1.21, 0.08)−0.68 (−1.26, −0.11)−2.27 (−3.33, −1.32)§
  Marrow area, mm2−0.41 (−1.23, 0.42)−0.43 (−1.17, 0.30)−0.38 (−1.74, 0.97)
  Cortical BMC, mg−0.46 (−1.10, 0.17)−0.93 (−1.50, −0.31)−0.80 (−1.85, 0.34)
  Cortical vBMD, mg/cm31.08 (0.44, 1.72)0.40 (−0.18, 0.97)1.32 (0.15, 2.25)
  Trabecular vBMD, mg/cm3−0.85 (−1.47, −0.22)−1.29 (−1.84, −0.78)−1.34 (−2.36, −0.31)
  pSSI, mm3−0.72 (−1.25, −0.18)−0.96 (−1.47, −0.48)−0.85 (−1.43, 0.33)

Bone abnormalities in the musculoskeletal system.

General bone results are presented in Table 3. Height for age, whole-body bone area to height, and whole-body BMC to bone area were used to assess the entire skeleton to establish whether bones were short, narrow, or lighter (low BMC) (29). Patients with systemic JRA were found to have significant decreases in height Z scores (Table 1), indicating shorter bones, while the skeletons of those in the polyarticular JRA cohort were narrower and lighter versus controls (P ≤ 0.01).

Table 3. General bone findings by JRA subtype*
 Whole-body BA (cm2):height (cm)Whole-body BMC (gm):BA (cm2)
  • *

    Values are the adjusted mean of the SD (95% confidence interval). JRA = juvenile rheumatoid arthritis; BA = bone area; BMC = bone mineral content.

  • P ≤ 0.01 versus healthy controls.

Pauciarticular (n = 17)0.09 (−0.19, 0.27)−0.09 (−0.02, 0.34)
Polyarticular (n = 23)−0.31 (−0.56, −0.05)−0.29 (−0.51, −0.16)
Systemic (n = 8)−0.14 (−0.48, 0.57)−0.22 (−0.60, −0.15)

A 2-step algorithm proposed by Schoenau (32) was used to further delineate whether the bone abnormalities observed in our study population were due to a primary alteration in bone metabolism or were secondary to a loss of muscle mass. Muscle mass is predominantly dependent on height, and bone mineral deposition was directly linked to muscle mass. Figure 1A illustrates the relationship of muscle cross-sectional area to height for all JRA subtypes plotted against controls. Patients with polyarticular JRA were found to have a greatly reduced ratio of muscle cross-sectional area to height versus controls (P = 0.001). The relationship of tibia diaphysis cortical BMC to muscle cross-sectional area (Figure 1B) was similar to that in controls for all JRA subtypes. Comparisons of whole-body lean mass with height and wholebody BMC with lean mass revealed similar relationships (not shown). The presence of a normal ratio of BMC to muscle supports the notion that bone metabolism is normal in children with JRA. Thus, the bone abnormalities found in JRA appear to be the result of low muscle mass.

thumbnail image

Figure 1. Assessment of bone abnormalities in juvenile rheumatoid arthritis (JRA) patients by subtype, versus healthy controls. A, Relationship of muscle cross-sectional area (CSA) to height. B, Relationship of cortical bone mineral content (BMC) to muscle CSA. • = patients with pauciarticular JRA; ▵ = patients with polyarticular JRA; ∗ = patients with systemic JRA. Diagonal line indicates the fit line for controls. Values for muscle CSA to height in the polyarticular JRA cohort were significantly less than values in controls (P = 0.001).

Download figure to PowerPoint

Effect of medication on bone density, geometry, and strength.

The negative impact of long-term glucocorticoid exposure on bone is well documented. Therefore, a secondary analysis was performed with JRA patients grouped by exposure to glucocorticoids. Because all patients with glucocorticoid exposure also received MTX, patients were stratified into 3 groups: those who received glucocorticoids and MTX (n = 16), those who received MTX only (MTX; n = 23), or those who had no exposure to glucocorticoids or MTX (none; n = 9). Bone characteristics by medication exposure group are presented in Table 4.

Table 4. SD scores of bone characteristics, by glucocorticoid and methotrexate (MTX) use*
 Glucocorticoids and MTXMTXNo treatment
  • *

    Values are the adjusted mean (95% confidence interval). P values were determined by multivariate analysis of covariance adjusted for sex, age, lean body mass, and Tanner stage. N values are patients who received glucocorticoids and MTX/MTX only/no exposure to medications. See Table 2 for other definitions.

  • P ≤ 0.01 versus healthy controls.

  • P ≤ 0.01 versus untreated group.

  • §

    P ≤ 0.01 versus glucocorticoids and MTX group, untreated group, and controls.

  • P ≤ 0.01 versus glucocorticoids and MTX group and controls.

Femoral neck (n = 12/15/4)   
 BMD, gm/cm2−0.62 (−1.18, −0.12)−1.07 (−1.69, −0.45)0.66 (−0.24, 1.56)
 BMAD, gm/cm3−0.40 (−0.97, 0.16)−1.19 (−1.89, −0.49)−0.04 (−1.04, 0.97)
Lumbar spine (n = 15/22/9)   
 BMD, gm/cm2−0.49 (−1.02, 0.04)−0.55 (−1.17, 0.06)0.86 (−0.01, 1.71)
 BMAD, gm/cm3−0.04 (−0.35, 0.44)−0.24 (−0.70, 0.22)0.39 (−0.24, 1.02)
Tibia (n = 16/23/9)   
  Muscle CSA, mm2−1.08 (−1.61, −0.54)−0.57 (−1.30, −0.06)−0.35 (−1.11, 0.62)
  Bone CSA, mm2−0.89 (−1.50, 0.28)−0.93 (−1.65, −0.22)−0.16 (−1.14, 0.83)
  Cortical CSA, mm2−1.02 (−1.54, −0.51)−1.18 (−1.79, −0.58)−0.48 (−1.32, 0.35)
  Cortical thickness, mm−0.69 (−1.29, 0.10)−1.37 (−2.08, −0.67)§−0.33 (−1.29, 0.64)
  Marrow area, mm2−0.66 (−1.37, 0.05)−0.29 (−1.13, 0.55)−0.04 (−1.20, 1.11)
  Cortical BMC, mg−0.82 (−1.36, −0.27)−1.00 (−1.64, −0.37)−0.24 (−0.92, 0.84)
  Cortical vBMD, mg/cm30.81 (0.24, 1.39)0.51 (−0.17, 1.18)1.03 (0.17, 2.03)
  Trabecular vBMD, mg/cm3−0.87 (−1.38, −0.36)−1.76 (−2.36, −1.16)−0.61 (−1.43, 0.23)
  pSSI, mm3−1.08 (−1.61, −0.54)−0.57 (−1.20, −0.06)−0.25 (−1.11, 0.62)

Patients exposed to MTX only or to glucocorticoids and MTX had significantly lower femoral neck densities, with the greatest loss seen in the MTX only cohort. Although the MTX group also exhibited the greatest loss in L-spine BMD, L-spine BMAD did not differ from that in controls. The L-spine BMD Z scores for patients with no exposure, however, were significantly higher than those in either the glucocorticoids and MTX group or the MTX only group. The MTX group had the greatest reduction in most parameters of bone geometry and in trabecular vBMD, while the glucocorticoids and MTX group had the greatest reduction in muscle and bone strength. Z scores for bone and cortical cross-sectional area, trabecular vBMD, and bone strength were less than those in healthy controls for all groups, reflecting the effects of other factors on these variables.

An evaluation of the generalized effects of medications on bone (height Z score, and ratios of whole-body bone area to height and BMC to bone area) indicated a shorter, narrower, and therefore lighter skeleton in patients who received glucocorticoids and MTX, while patients who received MTX only had narrower and lighter skeletons (Table 5). Figure 2A shows the relationship of muscle cross-sectional area to height by medication exposure plotted against controls, with patients who received glucocorticoids and MTX exhibiting the greatest loss. The relationship of tibia diaphysis cortical BMC to muscle cross-sectional area, however, was similar to that of controls for all medication classifications (Figure 2B). Thus, bone metabolism appeared to be normal and independent of medication exposure in children with JRA.

Table 5. General bone findings by glucocorticoid and methotrexate (MTX) exposure*
 Glucocorticoids and MTX (n = 16)MTX (n = 23)No. treatment (n = 9)
  • *

    Values are the adjusted mean of the SD (95% confidence interval). BA = bone area; BMC = bone mineral content.

  • P ≤ 0.01 versus healthy controls.

Height, Z score−0.51 (−0.99, −0.04)−0.18 (−0.78, 0.28)0.32 (−0.47, 1.10)
Whole-body BA (cm2):   
 height (cm)−0.44 (−0.68, −0.22)−0.24 (−0.51, 0.03)−0.06 (−0.45, 0.37)
Whole-body BMC (gm):   
 BA (cm2)−0.38 (−0.54, −0.22)−0.50 (−0.68, −0.31)−0.18 (−0.37, −0.15)
thumbnail image

Figure 2. Effect of medications on JRA patients by subtype, versus healthy controls. A, Ratio of muscle CSA to height. B, Ratio of cortical BMC to muscle CSA. ◊ = patients treated with glucocorticoids and methotrexate (MTX); ▪ = patients treated with MTX only; × = no treatment. Diagonal line indicates the fit line for controls. Values for ratio of muscle CSA to height in the glucocorticoid and MTX–treated cohort were significantly less than values in controls (P = 0.001). See Figure 1 for other definitions.

Download figure to PowerPoint

Predictors of bone findings in children with JRA.

Table 6 summarizes the results of the linear regression analyses. Overall muscle mass was the predominant positive predictor of L-spine and femoral neck densities and tibia cortical bone geometry and strength and trabecular density, accounting for 33–96% of the variance in these parameters (P = 0.000). Glucocorticoid exposure (cumulative or average dosage) was the predominant negative predictor of L-spine and femoral neck densities and cortical bone geometry and density, and contributed 33–49% of the variance. The positive relationship between marrow area and average glucocorticoid dosage (r = 0.49, P = 0.001) is consistent with previously reported thinning of the cortical bone associated with long-term glucocorticoid use (32).

Table 6. Predictors of bone findings in children and adolescents with JRA*
  • *

    SEE = standard error of the estimate; LBM = lean body mass; MTX = methotrexate; BA = bone area (see Table 2 for other definitions).

Lumbar spine    
  Tanner stage  0.630.000
  LBM  0.580.000
  Cumulative glucocorticoid dose  −0.330.04
  Tanner stage  0.780.000
  Cumulative glucocorticoid dose  −0.370.02
Femoral neck    
  LBM  0.880.000
  Cumulative glucocorticoid dose  −0.400.008
  Tanner stage  0.590.000
  Cumulative glucocorticoid dose  −0.410.007
  Sex  0.360.02
  Muscle CSA0.69678.880321,008.260.001
   Age  0.820.000
   Remission days  0.340.03
  Bone CSA0.7948.6947398.9430.000
   LBM  0.890.000
  Cortical CSA0.9316.4936325.0950.000
   LBM  0.960.000
   Average glucocorticoid dose  −0.490.001
  Cortical thickness0.790.385831.4050.000
   LBM  0.870.000
   Cumulative glucocorticoid dose  −0.480.001
  Marrow area0.9116.1717313.8990.05
   LBM  −0.950.000
   Average glucocorticoid dose  0.490.001
  Cortical BMC0.9220.6863415.4960.084
   LBM  0.960.000
   Average glucocorticoid dose  −0.430.004
  Cortical vBMD0.4639.997911,032.2710.000
   Tanner stage  0.660.000
   Average glucocorticoid dose  0.470.004
  Trabecular vBMD0.2732.09082192.0520.000
   Joint activity  −0.420.005
   LBM  0.330.03
   LBM  0.960.000
   Average MTX dose  −0.340.03
 Whole-body BA:height ratio0.6016381.2818,859.30.003
  Physical activity  0.550.000
  JRA duration  −0.630.000
  JRA active days  −0.510.001
  Fracture history  −0.320.04
 Whole-body BMC:BA ratio0.930.039220.4200.000
  LBM  0.860.000
  Cumulative glucocorticoid dose  −0.560.000
  Tanner stage  0.490.001
 Muscle CSA:height ratio0.37460.291,654.0020.000
  Age at onset  0.560.000
  JRA remission days  0.330.03
 BMC:muscle CSA ratio0.620.005730.0420.000
  Age at onset  0.630.000
  Average glucocorticoid dose  −0.410.01
  Cumulative MTX dose  0.530.000
  Average MTX dose  −0.330.04

Among the general effects, weight-bearing physical activity was positively related, while JRA duration, active JRA days, and fracture history were negatively related, to skeletal bone size, accounting for 32–63% of the variability. Skeletal mineralization (the ratio of whole-body BMC to bone area) was positively related to muscle mass and pubertal maturation (r = 0.86 and 0.49, respectively, P = 0.001) and inversely related to glucocorticoid exposure (r = −0.56, P = 0.000). Analysis of localized effects revealed that the ratio of muscle cross-sectional area to height was positively predicted by JRA remission days and age at disease onset (r = 0.33 and 0.56, respectively, P < 0.03). The ratio of BMC to muscle cross-sectional area was positively associated with MTX cumulative dose and age at JRA onset (r = 0.53 and 0.63, respectively, P = 0.000) and negatively associated with average MTX and glucocorticoid dose (r = −0.33 and −0.41, respectively, P < 0.04).


  1. Top of page
  2. Abstract
  7. Acknowledgements

To our knowledge, this is the first reported study of tibia bone characteristics measured by pQCT in children and adolescents with JRA. Our findings confirm that children and adolescents with JRA have decreased skeletal size, muscle mass, and trabecular bone density and cortical bone geometry and strength. The incorporation of both pQCT with DXA technologies allowed us to examine and compare both localized and generalized effects of JRA. Moreover, we were able to discern bone density and geometry properties of the peripheral skeleton in children with JRA. Not surprisingly, these bone abnormalities are more pronounced in children with greater disease severity.

Children with JRA have multiple risk factors for loss of muscle and bone mass, including delayed growth and development, malnutrition, decreased weight-bearing activity, inflammation, and glucocorticoid therapy (33). Our JRA cohort was well matched to our control population for age, body size, and calcium intake. A lower self-reported Tanner stage was found in girls with pauciarticular JRA and in boys with systemic JRA, suggesting delayed pubertal maturation, which is not uncommon in children with a chronic illness (34). The effect of pubertal maturation was minimized by including the self-reported Tanner stage in the statistical model.

Physical activity in patients with pauciarticular and polyarticular JRA and calcium intake in patients with all JRA subtypes were similar to findings in healthy controls. Dietary calcium intake and physical activity in the majority of patients (80%) were >800 mg/day and >10 METS hours/week, respectively. At calcium intakes approaching minimum requirements, physical activity has been shown to be a more important predictor of bone density than calcium intake (35). While neither calcium intake nor physical activity was associated with L-spine or femoral neck density in our study population, overall skeletal size was predicted by greater reported physical activity. Although decreased physical activity has been observed in children with mild to moderate JRA (11), only children with systemic JRA in the present study were found to be significantly less active when compared with controls. This difference may be the result of promotion of physical activity for disease management by the health care providers in our rheumatology clinic.

Glucocorticoid exposure differed by JRA subtype and was the predominant negative predictor of generalized and localized bone parameters. Although reported to have a bone-sparing effect, higher MTX dosages were inversely related to bone strength. Although one could speculate that glucocorticoid use is associated with higher MTX dosages, the greatest reduction in bone strength was found in patients who received glucocorticoids plus MTX (Table 4).

DXA has been the predominant technology used to assess bone characteristics in children and adolescents with JRA. Measurement of the L-spine by DXA is commonly used as a surrogate for the evaluation of trabecular bone mass. A number of investigators have reported decreased L-spine BMD in children and adolescents with JRA (2, 3, 5–10). We found no difference in either size- and maturation-adjusted L-spine bone BMD or BMAD Z scores among JRA subtypes or when compared with controls. Femoral neck BMAD values, however, were significantly lower in patients with systemic JRA. The smaller and less mineralized skeletons observed in our polyarticular JRA cohort are consistent with findings in previous studies (15–17).

The use of pQCT technology in our study provided a more comprehensive assessment of the effects of JRA on overall bone size and strength as well as the delineation of trabecular and cortical bone characteristics and their relationship to muscle mass. We found trabecular vBMD to be reduced by ≥0.85 SD by JRA subtype, with the smallest impact observed in patients who had not received either glucocorticoid or MTX therapy. Our finding of diminished trabecular vBMD is consistent with the work of others (18, 19). A cross-sectional study by Roth et al (18) of 57 patients ages 6–23 years with active JRA showed significantly diminished trabecular vBMD measured at the distal radius in patients with polyarticular JRA. A secondary analysis showed decreased trabecular vBMD in the distal radius of patients with JRA not treated with glucocorticoids. Those authors did not state whether the patients not treated with glucocorticoids had received MTX therapy. In our study population, active joint inflammation at the time of study was inversely related, and muscle mass was positively related, to trabecular vBMD. Increased cytokine production, which occurs during active inflammation, is known to increase muscle catabolism and bone resorption (36, 37). Thus, we speculate that increased cytokine production is responsible for trabecular bone loss in children with JRA.

Tibia bone size and strength and cortical bone geometry were reduced in the children with reduced muscle mass. A linear relationship between muscle mass and bone mass is well documented in healthy children and adults (28). In general, the greater the lean mass, the more load the muscle exerts on the bone. The bone then undergoes increased resorption, with osteoblasts stimulated by the increased muscle-induced strain. In this way, the bone compensates for the increased load by thickening over time (20). Cortical vBMD was normal in our study cohort, consistent with the results of others (18, 19). Therefore, cortical bone changes appear to be the result of thinner cortices and not decreased mineralization.

Although the muscle cross-sectional area in the lower leg was reduced in all JRA subtypes, whole-body lean mass was diminished only in patients with systemic JRA. Muscle cross-sectional area loss in our study cohort ranged from −0.46 to −0.89 SD and was not as dramatic as reported by Roth et al (18), who found height-adjusted forearm muscle cross-sectional area measured by pQCT to be reduced by >1.25 SD in 57 patients with JRA ages 6 to 23 years, when compared with healthy controls. This difference may reflect variation in joint inflammation sites and number, greater weight-bearing activity in lower versus upper extremities, or differences in our control population.

Previous studies using DXA have suggested that JRA has a greater negative effect on the appendicular versus axial skeleton, secondary to reduced cortical bone mineralization (2). Our pQCT findings provide additional support for the theory that JRA impacts muscle mass and, in turn, bone size and strength by cortical thinning and not mineral loss. Also, a comparison of our study with others (18, 19) suggests that JRA-related alterations in muscle cross-sectional area and cortical bone geometry and strength differ between upper and lower body appendicular regions. Whether this finding is the result of differing tempos of skeletal growth between these regions due to sex, maturation, weight-bearing activity, or disease-related factors is unknown.

In conclusion, our data demonstrate that children and adolescents with JRA have diminished muscle mass, bone size and strength, and trabecular vBMD and altered cortical bone geometry when compared with healthy controls. Also, pQCT technology appears to be a valuable tool in the assessment of “muscle–bone unit” dynamics and should be considered when monitoring bone-building therapies in children and adolescents with JRA.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Dr. Moyer-Mileur had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Prahalad, Moyer-Mileur.

Acquisition of data. Felin, Prahalad.

Analysis and interpretation of data. Felin, Prahalad, Askew, Moyer-Mileur.

Manuscript preparation. Prahalad, Askew, Moyer-Mileur.

Statistical analysis. Felin, Moyer-Mileur.


  1. Top of page
  2. Abstract
  7. Acknowledgements

We would like to thank John Bohnsack, MD, Marian Szewcyk, PNP, and Jody Quick, BS, for their support and technical assistance, and the families that participated in this project.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • 1
    McDonagh JE. Osteoporosis in juvenile idiopathic arthritis [review]. Curr Opin Rheumatol 2001; 13: 399404.
  • 2
    Pepmueller PH, Cassidy JT, Allen SH, Hillman LS. Bone mineralization and bone mineral metabolism in children with juvenile rheumatoid arthritis. Arthritis Rheum 1996; 39: 73657.
  • 3
    Tortolani PJ, McCarthy EF, Sponseller PD. Bone mineral density deficiency in children [review]. J Am Acad Orthop Surg 2002; 10: 5766.
  • 4
    Masi L, Cimaz R, Simonini G, Bindi G, Stagi S, Gozzini A, et al. Association of low bone mass with vitamin D receptor gene and calcitonin receptor gene polymorphisms in juvenile idiopathic arthritis. J Rheumatol 2002; 29: 222531.
  • 5
    Henderson CJ, Specker BL, Sierra RI, Campaigne BN, Lovell DJ. Total-body bone mineral content in non–glucocorticoid-treated postpubertal females with juvenile idiopathic arthritis. Arthritis Rheum 2000; 43: 53140.
  • 6
    Cetin A, Celiker R, Dincer F, Ariyurek M. Bone mineral density in children with juvenile chronic arthritis. Clin Rheumatol 1998; 17: 5513.
  • 7
    Perez MD, Abrams SA, Loddeke L, Shypailo R, Ellis KJ. Effects of rheumatic disease and corticosteroid treatment on calcium metabolism and bone density in children assessed one year after diagnosis, using stable isotopes and dual energy x-ray absorptiometry [review]. J Rheumatol Suppl 2000; 58: 3843.
  • 8
    Zak M, Hassager C, Lovell DJ, Nielsen S, Henderson CJ, Pedersen FK. Assessment of bone mineral density in adults with a history of juvenile chronic arthritis: a cross-sectional long-term followup study. Arthritis Rheum 1999; 42: 7908.
  • 9
    Lien G, Flato B, Haugen M, Vinje O, Sorskaar D, Dale K, et al. Frequency of osteopenia in adolescents with early-onset juvenile idiopathic arthritis: a long-term outcome study of one hundred five patients. Arthritis Rheum 2003; 48: 221423.
  • 10
    Kotaniemi A, Savolainen A, Kroger H, Kautiainen H, Isomaki H. Development of bone mineral density at the lumbar spine and femoral neck in juvenile chronic arthritis: a prospective one year followup study. J Rheumatol 1998; 25: 24505.
  • 11
    Henderson CJ, Lovell DJ, Specker BL, Campaigne BN. Physical activity in children with juvenile idiopathic arthritis: quantification and evaluation. Arthritis Care Res 1995; 8: 1149.
  • 12
    Butler RC, Davis MW, Worsfold M, Sharp CA. Bone mineral content in patients with rheumatoid arthritis: relationship to low-dose corticosteroid therapy. Br J Rheumatol 1991; 30: 8690.
  • 13
    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: 1659.
  • 14
    Di Munno O, Mazzantini M, Sinigaglia L, Bianchi G, Minisola G, Muratore M, et al. Effect of low dose methotrexate on bone density in women with rheumatoid arthritis: results from a multicenter cross-sectional study. J Rheumatol 2004; 31: 13059.
  • 15
    Bianchi ML, Cimaz R, Galbiati E, Corona F, Cherubini R, Bardare M. Bone mass change during methotrexate treatment in patients with juvenile rheumatoid arthritis. Osteoporos Int 1999; 10: 205.
  • 16
    Girgis SI, Nwokeji A, Shakur BH, Ind PW, Shiner RJ. The effect of the steroid-sparing response to low-dose methotrexate on bone metabolism in glucocorticoid-dependent asthmatics. Clin Chim Acta 2004; 341: 15763.
  • 17
    Tascioglu F, Oner C, Armagan O. The effect of low-dose methotrexate on bone mineral density in patients with early rheumatoid arthritis. Rheumatol Int 2003; 23: 2315.
  • 18
    Roth J, Palm C, Scheunemann I, Ranke MB, Schweizer R, Dannecker GE. Musculoskeletal abnormalities of the forearm in patients with juvenile idiopathic arthritis relate mainly to bone geometry. Arthritis Rheum 2004; 50: 127785.
  • 19
    Lettgen B, Neudorf U, Hosse R, Peters C, Reiners C. Bone density in children and adolescents with rheumatic diseases: preliminary results of selective measurement of trabecular and cortical bone using peripheral computerized tomography. Klin Padiatr 1996; 208: 1147. In German.
  • 20
    Ferretti JL, Capozza RF, Zanchetta JR. Mechanical validation of a tomographic (pQCT) index for noninvasive estimation of rat femur bending strength. Bone 1996; 18: 97102.
  • 21
    Cassidy JT, Levinson JE, Bass JC, Baum J, Brewer EJ Jr, Fink CW, et al. A study of classification criteria for a diagnosis of juvenile rheumatoid arthritis. Arthritis Rheum 1986; 29: 27481.
  • 22
    Taylor SJ, Whincup PH, Hindmarsh PC, Lampe F, Odoki K, Cook DG. Performance of a new pubertal self-assessment questionnaire: a preliminary study. Paediatr Perinat Epidemiol 2001; 15: 8894.
  • 23
    Schlossberger NM, Turner RA, Irwin CE Jr. Validity of self-report of pubertal maturation in early adolescents. J Adolesc Health 1992; 13: 10913.
  • 24
    Rockett HR, Wolf AM, Colditz GA. Development and reproducibility of a food frequency questionnaire to assess diets of older children and adolescents. J Am Diet Assoc 1995; 95: 33640.
  • 25
    Aaron DJ, Kriska AM, Dearwater SR, Cauley JA, Metz KF, LaPorte RE. Reproducibility and validity of an epidemiologic questionnaire to assess past year physical activity in adolescents. Am J Epidemiol 1995; 142: 191201.
  • 26
    Ainsworth BE, Haskell WL, Leon AS, Jacobs DR Jr, Montoye HJ, Sallis JF, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sport Exerc 1993; 25: 7180.
  • 27
    Moyer-Mileur L, Xie B, Ball S, Bainbridge C, Stadler D, Jee WS. Predictors of bone mass by peripheral quantitative computed tomography in early adolescent girls. J Clin Densitom 2001; 4: 31323.
  • 28
    Ferretti JL, Cointry GR, Capozza RF, Frost HM. Bone mass, bone strength, muscle-bone interactions, osteopenias and osteoporoses [review]. Mech Ageing Dev 2003; 124: 26979.
  • 29
    Cratree NJ, Kent K, Zemel BS. Acquisition of DXA in children and adolescents. In: SawyerAJ, BachrachLK, FungEB, editors. Clinical densitometry in growing patients. Totowa (NJ): Humana Press; 2006. p. 7394.
  • 30
    Molgaard C, Thomsen BL, Michaelsen KF. Influence of weight, age and puberty on bone size and bone mineral content in healthy children and adolescents. Acta Paediatr 1998; 87: 4949.
  • 31
    Katzman DK, Bachrach LK, Carter DR, Marcus R. Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 1991; 73: 13329.
  • 32
    Schoenau E. The “functional muscle-bone unit”: a two-step algorithm in pediatric bone disease. Pediatr Nephrol 2005; 20: 3569.
  • 33
    Burnham JM, Leonard MB. Bone disease in pediatric rheumatologic disorders [review]. Curr Rheumatol Rep 2004; 6: 708.
  • 34
    Welten DC, Kemper HC, Post GB, Van Mechelen W, Twisk J, Lips P, et al. Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake. J Bone Miner Res 1994; 9: 108996.
  • 35
    Andreassen H, Hylander E, Rix M. Gender, age, and body weight are the major predictive factors for bone mineral density in Crohn's disease: a case-control cross-sectional study of 113 patients. Am J Gastroenterol 1999; 94: 8248.
    Direct Link:
  • 36
    Gattorno M, Picco P, Vignola S, Stalla F, Buoncompagni A, Pistoia V. Serum interleukin 12 concentration in juvenile chronic arthritis. Ann Rheum Dis 1998; 57: 4258.
  • 37
    Rall LC, Rosen CJ, Dolnikowski G, Hartman WJ, Lundgren N, Abad LW, et al. Protein metabolism in rheumatoid arthritis and aging: effects of muscle strength training and tumor necrosis factor α. Arthritis Rheum 1996; 39: 111524.