RANKL:Osteoprotegerin ratio and bone mineral density in children with untreated juvenile dermatomyositis




To determine bone mineral density (BMD) in patients at the time of diagnosis of juvenile dermatomyositis (DM), to compare the RANKL:osteoprotegerin (OPG) ratio in patients with juvenile DM with that in healthy control subjects, and to evaluate whether BMD is associated with the RANKL:OPG ratio and the duration of untreated disease.


Thirty-seven children with juvenile DM were enrolled. Dual x-ray absorptiometry (DXA) was performed before treatment, and Z scores for the lumbar spine (L1–L4) were determined. The duration of untreated disease was defined as the period of time from the onset of rash or weakness to the time at which DXA was performed. Serum specimens obtained at the time of DXA were analyzed for concentrations of RANKL and OPG, using enzyme-linked immunosorbent assay. The RANKL:OPG ratio was also determined in 44 age-matched healthy control subjects.


At the time of diagnosis of juvenile DM, patients had a significantly increased RANKL:OPG ratio compared with that in healthy children (mean ± SD 2.19 ± 3.03 and 0.13 ± 0.17, respectively; P < 0.0001). In patients with a lumbar spine BMD Z score of −1.5 or lower, the RANKL:OPG ratio was significantly higher than that in patients with a lumbar spine BMD Z score higher than −1.5 (P = 0.038). Lumbar spine BMD Z scores (mean ± SD −0.13 ± 1.19 [range −2.10 to 2.85]) were inversely associated with the duration of untreated disease (R = −0.50, P = 0.003).


Children with juvenile DM have an elevated RANKL:OPG ratio at the time of diagnosis, resulting in expansion of the number of osteoclasts and activation of the bone resorptive function. This may lead to a lack of normal bone mineral accretion and a subsequent reduction in the lumbar spine BMD Z score. Patients with a longer duration of untreated juvenile DM have reduced lumbar spine BMD Z scores. These data suggest that early diagnosis could reduce the likelihood of reduced lumbar spine BMD in these patients by prompting intervention strategies at an early stage.

Children with rheumatic disease have decreased bone mineral density (BMD). Studies of BMD have primarily evaluated patients with juvenile idiopathic arthritis (JIA), and an inverse correlation between higher disease activity and lower BMD has been demonstrated in children with arthritis (1–4). Not surprisingly, patients exposed to corticosteroids were observed to have lower bone density (3, 4), although bone density also remained low in the absence of corticosteroid exposure (2, 5). Recently, a prospective study evaluating children with JIA compared with healthy control subjects demonstrated that not only did children with JIA have decreased bone mineral content at baseline, but that over time the patients had a less-than-expected age-appropriate gain in bone mass (6).

Limited data exist concerning bone metabolism in children with juvenile dermatomyositis (DM), which is the most common pediatric inflammatory myopathy, with an annual incidence in the US of 3.2 cases per 1 million children (7). Juvenile DM is a vasculopathy that produces a characteristic rash and symmetric proximal muscle weakness. Additional clinical criteria for juvenile DM include elevated levels of muscle-derived enzymes, inflammation demonstrated on electromyography, and muscle biopsy revealing specific histologic features (8, 9). A validated Disease Activity Score (DAS) has been established, in which cutaneous criteria are included and muscle involvement is evaluated (10). Recently, the duration of untreated disease has been recognized as an important variable to consider when determining the intensity of therapy at the time of diagnosis of juvenile DM (11).

The mechanisms by which chronic inflammatory disease alter bone mineral metabolism are uncertain. Recently, 3 novel proteins, RANK, RANKL, and osteoprotegerin (OPG), have been implicated in the regulation of bone turnover. All 3 proteins are members of the tumor necrosis factor receptor superfamily. RANK was originally detected on dendritic cells and T cells (12) and subsequently was described on immature osteoclasts (13, 14). This transmembrane receptor has high affinity and specificity for RANKL (12–14). Upon binding of RANKL to RANK, multiple intracellular signaling pathways are activated, including NF-κB and JNK (12–14). The regulation of osteoclast differentiation and function appears to be impacted by these pathways. Two independent groups of investigators identified RANKL as a membrane-bound and soluble protein (15, 16). RANKL messenger RNA (mRNA) is expressed in trabecular bone and in lymphoid tissue active in mediating immune responses (14–17). This protein expands the number of activated osteoclasts in the presence of macrophage colony-stimulating factor (14, 15, 17, 18) and prevents apoptosis of mature osteoclasts (19). OPG is a soluble secreted decoy receptor of RANKL (20, 21). OPG mRNA and protein are found on osteoblastogenic precursors (22). Thus, the effects of RANKL in promoting bone resorption are counterbalanced by the effects of OPG.

The RANKL:OPG ratio in serum, rather than the individual protein concentrations, has been suggested to be the critical factor in determining osteoclastic activation at the level of bone, with higher serum RANKL:OPG ratios being a marker for up-regulation of osteoclastogenesis (22).

The serum RANKL:OPG ratio in children with juvenile DM has not been reported. The purpose of the current study was to investigate this ratio in children with juvenile DM at the time of diagnosis (prior to treatment with corticosteroids), to concurrently determine the BMD status of these patients, and to assess for a correlation between the RANKL:OPG ratio and BMD. We believe that systemic inflammation in juvenile DM may up-regulate osteoclastogenesis and osteoclast-mediated resorption, as reflected in an elevated RANKL:OPG ratio and decreased BMD.


Patient population.

Children in whom juvenile DM was diagnosed by one physician (LMP) at Children's Memorial Hospital Immunology/Rheumatology Clinic between 1996 and 2005 were included. The patients fulfilled the established criteria for juvenile DM, as described by Bohan and Peter (8, 9), and had not received therapy. Age-matched healthy control subjects were recruited from a neighborhood school in the same geographic area. All patients and control subjects in this study signed our institutional review board–approved informed consent forms.

Laboratory evaluation.

Duration of untreated disease was defined as the period of time beginning at the onset of rash or weakness to the date when dual x-ray absorptiometry (DXA) was performed. Disease activity in children with juvenile DM was assessed with a validated DAS (10), an instrument that includes a combination of skin and muscle strength assessments. The DAS provides a numeric value on a scale of 0–20, with higher values being associated with more active disease. Serum levels of the muscle-derived enzymes creatine kinase, lactate dehydrogenase, aldolase, and aspartate aminotransferase/serum glutamic oxaloacetic transaminase were determined by standard laboratory methods.

BMD measurement.

Bone density in patients was measured prior to treatment of juvenile DM. DXA measurements of the lumbar spine (L1–L4) of the patients were performed using a total body scanner (Lunar Prodigy; Lunar, Madison, WI). Appropriate pediatric reference standards (23) were used by one radiologist (RMS) to calculate lumbar spine BMD Z scores. The current standard produces a Z score that adjusts BMD based on a child's age and sex. There is no consensus regarding a definition of osteopenia or osteoporosis in children. For the purposes of this study, we defined low bone density (osteopenia) as a lumbar spine BMD Z score more negative (lower) than −1.5.

RANKL and OPG determination.

Serum samples were collected and stored at −80°C until assayed. In the children with juvenile DM, all sera were obtained on the same day as or within 1 day of the initial DXA measurement. RANKL was measured by a highly sensitive sandwich enzyme-linked immunosorbent assay (ELISA) (sRANKL; Biomedica Gruppe, Vienna, Austria). The detection limit is 0.08 pmoles/liter, with an intraassay (n = 16) coefficient of variation (CV) of 3–5% and an interassay (n = 10) CV of 6–9%. The OPG concentration was determined by a highly sensitive sandwich ELISA (OPG; Biomedica Gruppe). The detection limit is 0.14 pmoles/liter, with an intraassay (n = 16) CV of <10% and an interassay (n = 16) CV of <10%. All assays for RANKL and OPG were performed in duplicate, with results expressed as the mean ± SD. Plates were read using a Sunrise ELISA plate reader (Phoenix Pharmaceuticals, Belmont, CA).

Statistical analysis.

Data were summarized using means and SDs for continuous variables and frequencies for categorical variables. Parametric and nonparametric tests were chosen depending on whether or not the data satisfied normality assumptions. Associations between lumbar spine BMD Z scores and clinical characteristics of the patients with juvenile DM were determined using Spearman's correlation. Wilcoxon's rank test was used to compare the RANKL:OPG ratio between patients with lumbar spine BMD Z scores less than or equal to −1.5 (osteopenia) and those with lumbar spine BMD Z scores greater than −1.5. The concentrations of RANKL and OPG and the RANKL:OPG ratio in patients with juvenile DM versus control subjects were compared using Student's 2-tailed t-tests. Statistical analyses were performed using SAS software, version 9.1 (SAS Institute, Cary, NC). P values less than 0.05 were considered significant.


Demographic and clinical data.

Thirty-seven untreated children with juvenile DM and 44 healthy control subjects participated in the study. The group of patients with juvenile DM consisted of 28 girls and 9 boys (mean ± SD age 6.3 ± 2.4 years). The control group consisted of 23 girls and 21 boys (mean ± SD age 7.0 ± 3.2 years). The demographic features of both groups are shown in Table 1. All of the children with juvenile DM and all of the healthy control subjects were prepubertal. Only 1 child with juvenile DM had calcinosis at the time of diagnosis. The mean ± SD duration of untreated disease in the patients with juvenile DM was 10.5 ± 18.4 months (median 5.4 months, range 0.5–99.7 months). One child had mild disease primarily affecting the skin, which resulted in misdiagnosis for 99.7 months. Patients had a mean ± SD DAS of 11.1 ± 3.5 (median 11, range 4–17). Clinical data for the children with juvenile DM, including anthropometric measurements, are summarized in Table 2.

Table 1. Demographic characteristics of the patients with juvenile dermatomyositis and controls*
CharacteristicPatients (n = 37)Controls (n = 44)
  • *

    Except where indicated otherwise, values are the number (%).

Age, years  
 Mean ± SD6.3 ± 2.47.0 ± 3.2
 Male9 (24)21 (48)
 Female28 (76)23 (52)
 White35 (95)40 (91)
 Nonwhite2 (5)4 (9)
Tanner stage 137 (100)44 (100)
Table 2. Clinical characteristics of the 37 patients with juvenile dermatomyositis*
  • *

    Values are the mean.

  • Assessed in 36 patients.

Duration of untreated disease, months10.5
Total Disease Activity Score (range 0–20)11
Creatine kinase, IU/liter (normal range 27–248)555
Lactate dehydrogenase, units (normal range 162–309)392
Aspartate aminotransferase, IU/liter (normal range 16–52)79
Aldolase, IU/liter (normal range 3.4–8.6)15
Height, percentile46.1
Weight, percentile54.7

BMD status.

Z scores were available for 33 of the 37 children with juvenile DM (Figure 1), while Z scores could not be calculated in the 4 youngest children because of a lack of pediatric reference data for children of this age. The mean ± SD lumbar spine BMD Z score was −0.13 ± 1.19 (range −2.10 to 2.85); 6 of the 33 children (18%) had osteopenia of the lumbar spine. In addition, there was a significant inverse association between duration of untreated disease and the lumbar spine BMD Z score (R = −0.50, P = 0.003) (Figure 2). Among all patients in whom the duration of untreated disease was ≥8 months (n = 8), the mean ± SD lumbar spine BMD Z score was −1.20 ± 0.71.

Figure 1.

Frequency distribution of lumbar bone mineral density (BMD) Z scores at the time of diagnosis of juvenile dermatomyositis, compared with the normal Gaussian distribution curve.

Figure 2.

Relationship between duration of untreated juvenile dermatomyositis (months) and lumbar spine bone mineral density (BMD) Z score (R = −0.5035, P = 0.0028). The linear regression model (solid line) was y = −0.1136x + 0.5964. Diamonds indicate the scores for individual patients (n = 33). Broken lines indicate the 95% confidence intervals.

We observed no significant correlation between the lumbar spine BMD Z score and the DAS. There was also no significant correlation between the level of any of the serum-derived muscle enzymes and the lumbar spine BMD Z score.


The serum RANKL and OPG values and the RANKL:OPG ratios in children with juvenile DM and healthy control subjects are shown in Figures 3 and 4. Serum concentrations of RANKL in the patients with juvenile DM were increased compared with those in healthy age-matched control subjects (28.2 ± 22.8 and 7.9 ± 8.2 pg/ml, respectively; P < 0.001), while the level of serum OPG was decreased in patients (26.5 ± 14.0 and 70.7 ± 23.8 pg/ml, respectively; P < 0.001). Children with juvenile DM had an increase in the RANKL:OPG ratio compared with healthy children (2.19 ± 3.03 and 0.13 ± 0.17, respectively; P < 0.0001). There was no significant difference in the RANKL:OPG ratio between girls and boys, among either patients or control subjects. Among patients with juvenile DM (n = 6) who had osteopenia of the spine (lumbar spine BMD Z score of −1.5 or lower), the RANKL:OPG ratio was significantly higher compared with that in patients with juvenile DM (n = 27) who had a Z score higher than −1.5 (mean ± SD RANKL:OPG 5.3 ± 4.6 and 1.8 ± 2.4, respectively; P = 0.038).

Figure 3.

Serum RANKL and osteoprotegerin (OPG) concentrations in patients with juvenile dermatomyositis (JDM) and healthy control subjects. Patients with juvenile dermatomyositis had significantly elevated levels of RANKL and significantly decreased levels of OPG. Values are the mean and SD.

Figure 4.

Serum RANKL:OPG ratios in patients with juvenile dermatomyositis and healthy control subjects. Patients with juvenile dermatomyositis had a significantly elevated RANKL:OPG ratio. Values are the mean and SD. See Figure 3 for definitions.


We demonstrated that the RANKL:OPG ratio in children with juvenile DM is elevated compared with that in age-matched healthy control subjects, and that this may reflect up-regulated osteoclastogenesis. Furthermore, an increased RANKL:OPG ratio correlated with reduced lumbar spine bone density, with the higher ratios seen in patients with the lower lumbar spine BMD Z scores. A recent study demonstrated expression of RANK and RANKL in synovial cells from children with JIA, and RANKL mRNA was detected in peripheral blood mononuclear cells (PBMCs) from patients but not in PBMCs from healthy controls (24). Increased RANK and RANKL expression was considered to be the mechanism for the development of bone erosions and osteoporosis.

Taken together with results from the current study, mounting evidence suggests that the RANKL:OPG ratio reflects the critical components of the pathway resulting in the decreased bone density encountered in children with rheumatic disease. In contrast, another study in children with JIA demonstrated decreased serum levels of RANKL and elevated serum levels of OPG (25). It was speculated that such findings were attributable to a protective mechanism to prevent bone destruction. However, the patients were receiving a variety of treatment regimens and had a varied clinical spectrum of JIA, which may have altered the protein concentrations. Thus, it is difficult to compare those findings with the findings in our population of untreated patients with juvenile DM.

Although several studies have investigated bone status in children with rheumatic disease, only one previous study addressed BMD exclusively in children with juvenile DM (26). That study demonstrated that osteopenia (BMD Z score less than or equal to −1.5) is encountered more frequently than expected in these patients. Six (18%) of 34 children in the current study had a Z score of −1.5 or lower; based on a normal distribution for this Z score, the expected frequency is 6.8%. A matter of concern for both parents and physicians is that the treatment of choice for juvenile DM is corticosteroids, which are known to reduce BMD. In our study, 20 (61%) of 33 patients had lumbar spine BMD Z scores <0 at baseline, and corticosteroid therapy has the potential to further diminish BMD.

Not surprisingly, we found that children with a longer duration of untreated disease, which translates to prolonged periods of inflammation, had lower lumbar spine BMD. This observation may explain partially why we did not see an association of BMD with the DAS. We suspect that children with more severe disease and a higher DAS come to clinical attention rapidly, and BMD changes are not detected by DXA. Children with a longer duration of untreated disease experienced several more months of osteoclast-mediated bone resorption, which is detected by DXA. It remains unknown whether children with an elevated DAS at the time of diagnosis and a shorter duration of untreated disease have subsequent deterioration in BMD status.

The importance of identifying children with decreased bone density is that such children have the potential for an increased risk of fracture. Although this risk has not been defined as fully in children as in adults (27), an increased risk of fractures has been demonstrated in pediatric rheumatic disease (28, 29). A study of children with systemic JIA showed that almost one-fourth of patients sustained at least 1 fracture, and half of these were vertebral compression fractures (30). We suggest that supplementation with calcium and vitamin D should be considered at the time of diagnosis in children with juvenile DM, to prevent further bone loss associated with both the inflammatory process itself as well as corticosteroid exposure (1).

A limitation of this study is the cross-sectional design, because only patients who were first seen at the time of diagnosis were included. We utilized this cohort in a longitudinal analysis and demonstrated an association of an elevated RANKL:OPG ratio with decreased BMD (Rouster-Stevens KA, et al: unpublished observations). In pediatric BMD analysis, the Z score is adjusted for age and sex. Height and weight may also affect BMD. However, in our patient population, the Z scores for height and weight approximated 0 and likely had minimal influence on BMD results. Additionally, dietary intake of calcium and vitamin D, both of which impact BMD, was not considered in the analysis. It is difficult to accurately ascertain nutrient intake, particularly in children; therefore, this was not included in the BMD assessment. Physical activity may also alter BMD; however, all patients in this study had diminished functional status related to the weakness that is characteristic of juvenile DM.

In conclusion, we have documented that at the time of diagnosis of juvenile DM, untreated patients have an elevated RANKL:OPG ratio compared with that in healthy children, and the elevated ratio correlates with lower BMD at that time. Approximately 1 of 5 patients have osteopenia of the lumbar spine prior to receiving corticosteroids. Patients with a longer duration of untreated disease are more likely to have decreased lumbar BMD. Alterations in the OPG/RANKL/RANK pathway provide a mechanism for the diminished bone mineral accretion that may occur in children with juvenile DM. Our results indicate that one possible mechanism of increased bone resorption, leading to lowered BMD, is the result of an elevated RANKL:OPG ratio. Parents of children with juvenile DM should be encouraged to provide sufficient dietary vitamin D and calcium as key components of their child's care.


Dr. Pachman 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. Rouster-Stevens, Langman, Pachman.

Acquisition of data. Rouster-Stevens, Langman, Price, Shore, Abbott, Pachman.

Analysis and interpretation of data. Rouster-Stevens, Langman, Shore, Seshadri, Pachman.

Manuscript preparation. Rouster-Stevens, Langman, Seshadri, Pachman.

Statistical analysis. Rouster-Stevens, Langman, Seshadri.