Bone mineral density in juvenile dermatomyositis: Assessment using dual x-ray absorptiometry




To determine the bone mineral density (BMD) status of our juvenile dermatomyositis (DM) population and to compare the frequency of osteopenia in patients with active disease requiring corticosteroids with that in patients with inactive disease who are not receiving corticosteroids.


Medical charts of all children diagnosed as having juvenile DM at our institution between 1989 and 1999 were reviewed for demographic and clinical data, including disease activity and duration of corticosteroid therapy. BMD measurements of the lumbar spine (L1–L4) were performed using dual x-ray absorptiometry (DXA). Z scores were calculated from the BMD data for comparison with published normative data.


A total of 15 patients were assessed: 10 with active disease, and 5 with inactive disease who had not taken corticosteroids for an average of 6.0 years (range 3.4–8.1 years). Baseline BMD measurements demonstrated osteopenia or frank osteoporosis in the majority of patients, including 6 of the 10 patients with active disease and 4 of the 5 patients whose disease was in remission. Fourteen patients had serial BMD measurements. Persistent or worsening osteopenia was documented in all patients who had ongoing active disease, except for 3 patients who had been treated with bisphosphonates because of vertebral compression fractures.


Osteopenia is common in patients with juvenile DM, and it usually worsens with ongoing disease. It can persist for many years after the disease enters remission. Bisphosphonates appeared to beneficially affect bone mineralization in our patients. Treatment to prevent the long-term complications of osteoporosis in patients with juvenile DM should be considered and requires further study.

Juvenile dermatomyositis (DM) is a systemic inflammatory disorder characterized by proximal muscle weakness and a typical rash. Although treatment with corticosteroids has improved the outlook for children with juvenile DM, considerable morbidity is still encountered (1). Significant osteoporosis of the vertebral bodies can occur, and not uncommonly, this leads to compression fractures (2, 3). The fact that many patients have active disease and receive treatment in the late prepubertal and early pubertal periods, when bone formation is most rapid, has raised concerns that mineral loss may lead to long-term osteoporosis and subsequent complications in adulthood (2).

The development of dual x-ray absorptiometry (DXA), which enables rapid and precise assessment of bone density using low-dose radiation (4), has facilitated the study of osteoporosis in patients with chronic rheumatic diseases. Children with juvenile rheumatoid arthritis (JRA) have been shown to have reduced axial bone density, with evidence of low bone turnover and decreased bone formation (5). However, relatively little has been published on bone mineral abnormalities in children with juvenile DM. Reduced serum osteocalcin levels have been reported in a small number of patients with active disease (6). Perez et al (2), using calcium isotopes, found a net decrease in calcium retention in patients with juvenile DM, particularly those receiving corticosteroids.

The aims of this study were to determine the frequency and severity of osteopenia in our juvenile DM population and to compare bone mineral density (BMD) in patients with active disease requiring corticosteroids with that in patients with inactive disease who were not receiving corticosteroids.


Study patients.

All patients with a diagnosis of juvenile DM who were receiving followup care between 1989 and 1999 at the Izaak Walton Killam (IWK) Health Centre, a tertiary pediatric referral facility, were eligible for inclusion in the study. Medical charts were reviewed, and the following parameters were extracted: sex, age, race, clinical presentation, duration of disease, medication use (past and present), timing and duration of corticosteroid therapy, occurrence of vertebral fractures, and presence of calcinosis.

Twenty consecutive children met the eligibility criteria. Of these, 3 were lost to followup, and 2 others from outside the province who had inactive disease at the last assessment were noncompliant with followup. The remaining 15 patients were enrolled in the study. All 15 patients fulfilled the criteria for juvenile DM established by Bohan and Peter (7). At the time of their baseline BMD measurement, none of the patients had another disease and none were receiving medications known to affect bone metabolism, other than corticosteroids. Because of the theoretical risks of increasing the likelihood of calcinosis, none of the patients had received supplementary vitamin D or calcium (8).

Following enrollment, a pediatric rheumatologist (BL) assessed all of the patients. Height, weight, Tanner stage, and presence of calcinosis were determined. Disease activity was assessed according to a published protocol (6), based on the presence of muscle weakness, skin vasculitis, degree of impairment in ambulation, and use of corticosteroid medication.

This study was approved by the IWK Health Centre Research Ethics Committee. Informed consent was obtained from all patients or their parents.

BMD measurement.

All BMD measurements were obtained with the same Hologic QDR 2000 DXA instrument (Hologic, Bedford, MA). In pubertal and postpubertal female patients, BMD measurements were performed within the first 10 days of the start of menses to prevent radiation exposure to a fetus in the event of an unexpected pregnancy. All BMD measurements were performed on the lumbar spine (L1–L4). This site was chosen because of its pathophysiologic relevance to vertebral compression fractures and because a high-quality age-related normative database is available for the pediatric population (9). Z scores were calculated from the BMD measurements and were used to compare our study population with healthy controls, as well as for subgroup analysis within our study population.

There are currently no consensus criteria for defining osteopenia or osteoporosis on the basis of BMD measurements in children. For the purpose of this study, the following values were used: normal = BMD Z score greater than –1, osteopenia = BMD Z score between –1 and –2.5, and osteoporosis = BMD Z score less than –2.5. These values were chosen to maintain consistency with adult data demonstrating an increased fracture risk in women with a BMD >1 SD below normal (10), as well as with World Health Organization standards (11).

In all but 1 of the 15 participants, serial BMD determinations were available to provide information on the bone mineralization trend. This quantity was defined as the difference between the final BMD Z score determination and its baseline value.

Subgroup definitions.

Patients were organized into inactive or active clinical disease subgroups. Those with inactive disease (n = 5) were not receiving corticosteroids. Those with active disease (n = 10) were receiving corticosteroids.

The active disease group was further subdivided according to bisphosphonate treatment (after developing vertebral compression fractures). Those in the active/untreated disease group had active clinical disease and were receiving corticosteroids, but not bisphosphonates (n = 6). Those in the active/treated disease group had active clinical disease and were receiving corticosteroids and bisphosphonates (n = 3).

Statistical analysis.

In general, data are expressed as the mean ± SD. Two-tailed t-tests were used for comparisons of group means. It was not possible to perform multiple regression analysis to identify independent predictors of bone density because of the relatively small sample size. Statistical analyses were performed using SPSS software, version 9.0 (SPSS, Chicago, IL), and Epi Info 2000 software, version 1.1 (Centers for Disease Control and Prevention, Atlanta, GA). An alpha value of 0.05 was selected as the limit of statistical significance.


Demographic and clinical data.

A total of 15 patients, all white, participated in the study. Nine were female and 6 were male, with a mean age at diagnosis of 7.9 years (range 3.6–14.2 years). The demographic and clinical features of the patients with active and inactive juvenile DM are shown in Table 1.

Table 1. Demographic and clinical features of patients with inactive and active juvenile DM*
Juvenile DM subgroup, patient/sex/age at diagnosis, yearsDuration of active disease before baseline BMD, yearsDisease severityDisease activity at BMD assessmentSteroid treatment at baseline BMDMedications during the disease courseVertebral fractureCalcinosis
BaselineFollowupDuration of treatment, yearsDuration off treatment, years
  • *

    DM = dermatomyositis; BMD = bone mineral density; IVIG = intravenous immunoglobulin; AZA = azathioprine; CYC = cyclophosphamide; MTX = methotrexate; CSA = cyclosporin A; HCQ = hydroxychloroquine; NA = not available.

  • Active disease was defined as a value ≥3 and inactive disease as ≤2 (see ref.6).

  • Patient continued to take steroids.

Inactive disease         
 1/F/14.01.2Severe001.27.7SteroidsT2, L1None
 3/F/5.23.3Moderate003.38.1Steroids, IVIGNoneNone
 5/F/8.24.5Moderate004.45.9Steroids, AZA, CYCNoneNone
Active disease         
 6/M/5.13.3Moderate432.8Steroids, IVIG, MTXNoneNone
 7/M/3.63.4Moderate431.3Steroids, MTX, CYC, CSANoneNone
 8/F/10.50.3Mild430Steroids, MTXNoneNone
 9/F/6.70.2Severe530.04Steroids, IVIG, MTXNoneNone
 10/F/5.04.2Moderate332.22 (restarted)Steroids, IVIG, MTX, HCQNoneMild
 11/F/7.04.1Moderate423.8Steroids, IVIG, MTXNoneNone
 12/M/14.25.6Moderate4NA4.5Steroids, IVIG, MTXT4None
 13/F/9.40.6Severe540.3Steroids, IVIG, MTX, CYCL1–L5None
 14/M/4.71.4Moderate531.2Steroids, IVIG, MTX, CSAT7, T9None
 15/M/10.31.0Moderate440.3Steroids, MTXT9, L1None

Twelve patients had a chronic continuous course of juvenile DM. Three patients had a monocyclic course. One patient developed mild superficial calcinosis (<5 mm2) on her right elbow. Ten patients had active disease at the time of their baseline BMD assessment, while 5 had inactive disease and were not taking corticosteroids. Patients with active disease had been treated with corticosteroids for an average of 1.6 years (range 0–4.5) prior to their baseline BMD measurement, compared with 2.9 years (range 1.2–4.4) for patients with inactive disease. The significant difference in the mean number of years of corticosteroid treatment prior to baseline BMD measurements in the 2 groups was due to 4 patients with active disease whose baseline BMD measurements were performed within 4 months of starting corticosteroid treatment. Patients with inactive disease had not received steroids for an average of 6.0 years (range 3.4–8.1) at the time of their initial BMD assessment and were taking no medications for their juvenile DM. Other prescribed medications in patients with active disease included methotrexate, cyclosporin A, cyclophosphamide, intravenous immunoglobulin, and hydroxychloroquine, either alone or in various combinations.

Baseline mineralization status.

Baseline BMD measurements were performed at a mean age of 12.2 years (range 4.8–22.9 years) (Table 2). Osteopenia or osteoporosis was present in the majority of patients (6 of the 10 patients with active disease; 4 of the 5 patients with inactive disease). The mean ± SD baseline Z score was –1.54 ± 1.23 (range 0.23 to –4.01) in the active disease group and –1.22 ± 0.23 (range –0.84 to –1.43) in the inactive disease group. This difference was not statistically significant (P = 0.58 by unpaired t-test).

Table 2. Anthropometric data and lumbar spine BMD data obtained from measurements at baseline and the last followup in patients with inactive and active juvenile DM*
Juvenile DM subgroup, patientAge, yearsWeight, kgHeight, cmTanner stageBMD L1–L4, gm/cm2Z score, L1–L4
  • *

    BMD = bone mineral density; DM = dermatomyositis; NA = not available.

  • Disease became inactive between baseline and followup BMD measurements.

  • Treatment with bisphosphonates between baseline and followup BMD measurements. Patient 13 received oral alendronate; patients 14 and 15 received intravenous pamidronate.

Inactive disease            
Active disease            

Followup mineralization data.

During the course of their illness, 5 patients developed symptomatic vertebral compression fractures, which were confirmed radiologically. BMD data were not available on 2 of these patients at the time of fracture, but in the other 3 patients, the mean Z score was –2.83 (range –1.62 to –4.01). These latter 3 patients were treated with bisphosphonates. One patient (patient 13) received oral alendronate and 2 patients (patients 14 and 15) received intravenous pamidronate because of concerns regarding the risk of esophageal irritation.

Fourteen patients had serial BMD measurements, with the most recent followup studies performed an average of 28.2 months (range 14.4–48.0) after the baseline study. Of these 14 patients, the disease remained inactive in the 5 members of the inactive disease group, and the patients were taking no medications at the last followup. Of the 9 patients in the active disease group who had serial BMD measurements, all but 1 patient continued to have active disease requiring corticosteroid treatment at the last followup. Six of these 9 patients had received no treatment to prevent or correct osteopenia between BMD measurements. One patient in the inactive disease group, 3 patients in the active/untreated disease group, and 2 patients in the active/treated disease group went through puberty between their baseline and last followup BMD studies.

Anthropometric and lumbar spine BMD data at baseline and at the last followup are summarized in Table 2. The mean ± SD Z score at followup for patients in the inactive disease group was –1.00 ± 0.34. Four of the 5 patients in this group either had osteopenia or were borderline for osteopenia (Z score –0.94). All were female patients who had reached maturity prior to their baseline BMD measurement. The only male patient in the inactive disease group went through puberty between his baseline and followup BMD assessments. He achieved a BMD value within normal limits, with a Z score of –0.52.

The mean ± SD Z score at followup for patients in the active/untreated disease group was –1.60 ± 0.63 (range –0.98 to –2.8). All but 1 patient in this group had either osteopenia or frank osteoporosis. The remaining patient had a borderline value (Z score –0.98). Followup Z scores were lower in all patients in this group compared with their baseline values, except for the 1 patient (patient 11) whose disease became inactive prior to the repeat BMD measurement (Figure 1).

Figure 1.

Serial Z scores for bone mineral density (BMD) of L1–L4 in patients with active juvenile dermatomyositis who were treated or were not treated with bisphosphonates between their BMD measurements at baseline and the last followup.

The mean ± SD Z score at followup for patients in the active/treated disease group was –0.82 ± 1.05 (range –0.16 to –2.03). In all 3 patients, Z scores improved compared with baseline (Figure 1). Two of these patients had followup Z scores in the normal range, while 1 patient had a Z score reflecting osteopenia. Two patients, including the patient with persistent osteopenia, went through puberty between BMD measurements, while 1 patient (patient 14) remained prepubertal at the time of his last followup BMD measurement.

The changes in BMD Z scores between baseline and last followup in patients in the active/treated and active/untreated disease groups are illustrated in Figure 1. In patients with active juvenile DM, bisphosphonate therapy was associated with a significantly improved mineralization trend (difference in Z scores at baseline versus last followup P < 0.001, by paired t-test) compared with those not receiving bisphosphonates.


The data presented herein show that significant osteopenia is common in patients with juvenile DM, even in those with inactive disease. However, patients with active disease receiving corticosteroids are most severely affected. Serial BMD measurements in this group demonstrate that osteopenia persists and often worsens with ongoing active disease, in the absence of treatment for osteoporosis. Among patients with ongoing active disease, only those treated with bisphosphonates had improvements in the BMD Z scores compared with baseline. These findings are consistent with previous reports of an association between disease activity and a significant reduction in BMD in children with chronic rheumatic diseases (2, 12). The findings are also consistent with a report of improved vertebral BMD in 38 children with rheumatic diseases, including 6 with juvenile DM, who were treated with alendronate because of significant osteopenia (13).

An important observation in this study was the presence of osteopenia at the baseline determination in the majority of patients (4 of 5) with inactive disease despite a period of 3–8 years since corticosteroids had last been taken. These 4 patients had gone through puberty prior to their baseline BMD measurement. Peak bone mass is generally believed to be reached by late adolescence or early adulthood (9, 14). Patients with inactive juvenile DM with persistent osteopenia may be at increased risk of developing osteoporosis later in life due to lower bone reserves to withstand the subsequent age-related bone resorption. Our findings are in contrast to the results of a study of patients with a history of JRA who were found to attain normal peak bone mass at the lumber spine, despite the fact that most of the patients had active disease during puberty (15). A possible explanation for the difference in findings is that either the juvenile DM disease process itself or the higher corticosteroid doses that are required results in more profound osteopenia than is found in JRA, and it may therefore take longer to correct this condition.

The small number of subjects in this study limits our ability to generalize our findings. However, the results suggest that longitudinal measurements of BMD may provide valuable information in patients with juvenile DM, even in those with inactive disease, that may influence decisions about the necessity for, and optimal timing of, medications to reduce the risk of osteoporosis. Bisphosphonates appear to have a beneficial effect on bone mineralization. However, despite increasing data on the safety of their use in children (13), there are still concerns about possible long-term effects on a fetus in a woman who has ever received bisphosphonates. Further study of the use of bisphosphonates in children with juvenile DM is needed. Aggressive therapy to control active disease, potentially including newer therapies, may also reduce the long-term risks of osteopenia.


We would like to thank Dr. A. Ross for expert assistance with BMD measurements, Dr. L. Dodds and Ms Colleen O'Connell for assistance with statistics, and Ms Peggy Brinton for secretarial support.