Long-term muscular outcome and predisposing and prognostic factors in juvenile dermatomyositis: A case–control study




To compare muscle strength, physical health, and HLA–DRB1 allele carriage frequencies in patients with longstanding juvenile dermatomyositis (DM) with that of controls, and to determine the presence of and risk factors for muscle weakness and magnetic resonance imaging (MRI)–detected muscle damage in juvenile DM patients.


Fifty-nine patients with juvenile DM examined a median of 16.8 years (range 2.0–38.1 years) after disease onset were compared with 59 age- and sex-matched controls. Muscle strength/endurance was measured by manual muscle testing (MMT) and the Childhood Myositis Assessment Scale (CMAS); health status was measured by the Short Form 36. HLA–DRB1 alleles were determined by sequencing in patients and 898 healthy controls. In patients, disease activity/damage was measured by the Disease Activity Score (DAS), Myositis Damage Index (MDI), Health Assessment Questionnaire/Childhood Health Assessment Questionnaire, and MRI scans of the thigh muscles. Early disease characteristics were obtained by chart review.


Patients had lower muscle strength/endurance (P < 0.001 for both) and physical health (P = 0.014) and increased HLA–DRB1*0301 (P = 0.01) and DRB1*1401 (P = 0.003) compared with controls. In patients, persistent muscle weakness was found in 42% with MMT (score <78) and in 31% with the CMAS (score <48), whereas MRI-detected muscle damage was found in 52%. Muscle weakness and MRI-detected muscle damage were predicted by MDI muscle damage and a high DAS 1 year postdiagnosis.


A median of 16.8 years after disease onset, juvenile DM patients were weaker than the controls; muscle weakness/reduced endurance was found in 31–42% of patients and MRI-detected muscular damage was found in 52% of patients. The outcomes were predicted by high disease activity and muscle damage present 1 year postdiagnosis.


Juvenile dermatomyositis (DM) is a rare, vasculopathic, autoimmune disease of childhood, characterized by proximal symmetric muscle weakness in addition to characteristic skin lesions. Sustained muscle weakness can be caused by chronic muscle inflammation, inactivity, or be the consequence of muscle damage (atrophy, fatty infiltration, and/or calcinosis).

Muscle weakness in patients with longstanding juvenile DM has previously been found in ∼35% of patients in two series of 12 and 17 patients after a mean of 18 years of followup (using a 0–5 muscle grading system; method not specified) (1, 2). Self-reported muscle weakness has more recently been described in 23% of 65 patients after a median of 7.2 years of disease duration (3). Validated methods for measuring muscle strength and endurance, manual muscle testing (MMT) and the Childhood Myositis Assessment Scale (CMAS), are now recommended as outcome measures in clinical juvenile DM trials (4, 5). Recently, a low MMT score and a low CMAS score were found in 41% and 53% of juvenile DM patients, respectively, in a large, multinational, multicenter study after ∼7 years of followup time (6). However, none of these studies have included controls.

Magnetic resonance imaging (MRI) of the thigh muscles is known to discriminate between disease activity and damage, and therefore likely have a role in imaging long-term complications of juvenile DM (7–13). Fatty infiltration, atrophy, and calcinosis have previously been reported in 3 small series (n = 4–19 patients) after 4–18 years of followup (1, 14, 15). Muscle damage shown by MRI (atrophy and fatty infiltration) has recently been investigated in 34 patients (part of a bigger cohort) 6.8 years after diagnosis in relation to organ damage, but the relationship between MRI findings and muscle strength was not investigated (16).

Alleles at the 8.1 ancestral haplotype (HLA–B*08/DRB1*0301/DQA1*0501) have earlier been described as a strong immunogenetic predisposing factor in juvenile DM (17). Although less common than in adults, myositis-specific antibodies (MSAs) have been found in juvenile DM patients (17–20), but their occurrence in long-term juvenile DM has not been described.

Therefore, the aims of the present study were to compare muscle strength, physical health, and HLA–DRB1 allele carriage frequencies in patients with longstanding juvenile DM with that of controls from the general population, and to explore the presence of and early risk factors for muscle weakness and MRI-detected muscle damage. Second, we wanted to explore how the outcome measures in juvenile DM patients correlated with patient characteristics and other measures of disease activity and damage at followup.


Patients and matched controls.

Inclusion criteria consisted of: disease onset before age 18 years, age ≥6 years at followup, minimum 24 months of disease duration, and a probable or definitive diagnosis of DM according to the criteria by Bohan and Peter (21). A retrospective inception cohort of Norwegian patients with juvenile DM diagnosed between January 1970 and June 2006 was identified (as previously described in detail) (22). Sixty-six patients fulfilled the inclusion criteria and 4 were diseased; the remaining 62 patients could all be tracked (due to the Norwegian population register), and 59 (95%) participated in the study.

Sex- and age-matched controls from the general population (living in Oslo or the neighboring county of Akershus) were randomly selected from the National Population Register. Exclusion criteria consisted of: mobility problems, inflammatory rheumatic disease, other autoimmune conditions treated with immunosuppressive agents, and heart or lung disease (except for mild asthma). An invitation to participate was sent to 243 individuals, of whom 69 accepted. Five responders were excluded because of the presence of exclusion criteria, and 5 additional responders were excluded because we identified 2 matched controls for 5 patients (the last responder to answer was excluded).

Informed consent was obtained from all of the patients and controls (and their parents if age <16 years), according to the Declaration of Helsinki. The study was approved by the Regional Ethics Committee for Medical Research.

Data collection and clinical measures.

All of the patients and matched controls were examined during a 1–1.5-day program between September 2005 and May 2009 at Oslo University Hospital. This included clinical examination by a single physician (HS) and measures of muscle strength and endurance (CMAS and MMT); blood samples included erythrocyte sedimentation rate (ESR), creatinine level, and muscle enzymes (creatine kinase, lactate dehydrogenase, aspartate aminotransferase, and alanine aminotransferase).

Health status was measured by the Norwegian version of Short Form 36 health survey (SF-36), version 1.0 (23, 24), in participants ages ≥12 years (n = 51). The SF-36 measures 2 distinct components: the physical component summary (PCS) scale and the mental component summary (MCS) scale; low scores indicate poor health status. In all of the participants (or their parents) ages ≥12 years, exercise inducing sweating or breathlessness in the last year was registered. This included exercise frequency (categories: 0 = never, 1 = less than once a month, 2 = once a month, 3 = once a week, 4 = 2–3 times a week, 5 = 4–6 times a week, and 6 = every day) and average weekly hours of exercise (categories: 0 = none, 1 = 30 minutes, 2 = 1 hour, 3 = 2–3 hours, 4 = 4–6 hours, and 5 = ≥7 hours).

In the patients with juvenile DM, disease activity was measured by the Disease Activity Score (DAS) (4, 25). The DAS consists of 2 components, DAS skin (range 0–9) and DAS muscle (range 0–11); a low score represents low disease activity. Cumulative organ damage was measured by the Myositis Damage index (MDI; range 0–40, where 0 = no damage) (26). Damage was defined in 11 organ systems, including MDI muscle damage (muscle atrophy, muscle weakness due to damage or muscle dysfunction). The DAS was also calculated from the time of diagnosis, and the DAS and MDI were calculated 1 year postdiagnosis (by chart review) (22). The Health Assessment Questionnaire (HAQ) (27) and the Childhood Health Assessment Questionnaire (C-HAQ) (28, 29) were used to measure physical function in patients ages ≥18 years (n = 39) and ages <18 years (n = 20), respectively (range 0–3, where 0 = no difficulty with daily activities). Information of immunosuppressive medication was obtained from the medical records. Followup time was defined as the time from disease onset to the followup examination.


MMT was used as a quantitative measure of muscle strength in all of the patients and controls (performed by HS). MMT-8 (the 8-muscles version applied on the right side) using a 0–10 scale was chosen (where 0 = no muscle contraction and 10 = full muscle strength) (5). MMT-8 has the same validity and reliability as a group of 24 muscles tested bilaterally (30). The following muscle groups were tested: neck flexors, shoulder abduction, elbow flexion, hip extension, hip abduction, knee extension, wrist extension, and ankle dorsiflexion. Muscle weakness according to MMT was defined as: less than the mean − 2 SDs, calculated from the MMT scores of the matched controls.


The CMAS, a performance-based instrument developed to evaluate muscle strength, physical function, and endurance in children with juvenile DM (31), was completed in all of the patients and 56 controls (by 2 experienced physiotherapists blinded for disease variables, including the MMT scores). The CMAS has a very good interrater reliability, and was validated for use in patients with juvenile DM ages <18 years (32), but has not previously been used in adult juvenile DM patients. The CMAS consists of 14 ordinal items of proximal and axial muscle performance; the scores range from 0–52, where 52 = maximal muscle performance. Muscle weakness/reduced endurance according to the CMAS was defined as: less than the mean − 2 SDs, calculated from the CMAS scores of the matched controls.

MRI of the thigh muscle.

MRI was performed using a 1.5T or 1.0T scanner (Siemens) in 58 patients at followup. Proximal thigh muscles were examined by a coronal T1-weighted spin-echo sequence and coronal and axial STIR sequences. Thirty-five patients were examined with an additional axial T1-weighted spin-echo sequence.

The MRI scans were scored independently by 2 experienced radiologists (EM and EK) who were blinded for clinical information; in cases of disagreement, consensus was made. Muscle, fascia, subcutaneous, and cutaneous layers were scored in both thighs. Edema and calcinosis were scored in all layers. Atrophy, defined as volume loss, was scored in muscle and subcutis. Fatty infiltration was scored in muscle. For all findings, we used a yes/no score. Muscle edema and muscle fatty infiltration were also scored using a 0–4 scale (10, 33, 34). Muscle fatty infiltration was scored as follows: grade 0 = normal, grade 1 = fatty streaks, grade 2 = muscle greater than fat, grade 3 = muscle equal to fat, and grade 4 = muscle less than fat. Tumoral, linear, and/or speckled areas with low signal on both T1-weighted spin-echo sequences and STIR sequences were interpreted as fibrosis and/or calcinosis. Despite the lack of radiographic confirmation of density, we chose to name these findings as calcinosis.

MRI-detected muscle damage (yes/no) was defined as at least one of the following: calcinosis in the muscle or fascia, muscle atrophy, or muscle fatty infiltration. Edema in muscle or fascia was interpreted as possible inflammatory disease activity.

HLA–DRB1 alleles and autoantibodies.

DNA was extracted from peripheral blood samples from all of the patients and DRB1 genotyping was performed by sequencing using the BigDye Terminator, version 3.1 (Applied Biosystems), chemistry on an ABI3730 DNA Analyzer (Applied Biosystems), followed by allele assignment using AssignSBT, version 3.2.7 (Conexio Genomics). DRB1 data from 898 healthy individuals from the Norwegian Bone Marrow Registry (genotyped by the same method) were used as a control group.

In patients, antinuclear antibodies (ANAs) were analyzed with indirect immunofluorescence using HEp-2 or HEp-2000 cells and fluorescein-labeled anti-human IgG (light and heavy chain) in a routine diagnostic setup. Positive samples were titrated and fluorescence patterns were interpreted. Also, sera (drawn at the followup visit) were tested for MSAs and myositis-associated antibodies (MAAs) with a commercial assay Euroline Myositis Profile (Euroimmune) according to the manufacturer's instructions.

Statistical analysis.

Differences between patients and matched controls were tested by the paired-sample t-test (normally distributed continuous variables), Wilcoxon's rank sum test (continuous non–normally distributed or ordinal variables), or McNemar's test (binary data). Differences between patient groups were tested by the independent-sample t-test, the Mann-Whitney U test, or the chi-square test, as appropriate. Correlations were determined by the Spearman's correlation coefficient (rs).

In order to identify possible early risk factors for an unfavorable muscular outcome (low MMT score, low CMAS score, and MRI-detected muscle damage), logistic regression analyses were performed on the relationship between the outcome variables and patient characteristics/disease variables assessed at diagnosis and 1 year postdiagnosis. Variables with P values less than 0.15 were included in the multivariate model, and highly correlated independent variables were avoided (r >0.7). Possible predictors were chosen by backward elimination after controlling for age and sex. Data on the strength of the associations were given as odds ratios with 95% confidence intervals. A P value less than 0.05 was considered as statistically significant. SPSS, version 16.0 (SPSS), was used for statistical analyses.


Characteristics of juvenile DM patients and controls.

Fifty-nine patients with juvenile DM (assessed a median of 16.8 years after disease onset) and 59 age- and sex-matched controls were included in the study (Table 1). Of the patients and controls, 20 (34%) were ages 6–17 years and 39 (66%) were ages ≥18 years; 36 (61%) were females. According to the diagnostic criteria by Bohan and Peter, 53 (90%) of the patients with juvenile DM were found to have definite DM and 6 (10%) were found to have probable DM. The patients exercised less than the controls, but no differences were found with regard to any other characteristics (Table 1).

Table 1. Characteristics of patients with juvenile DM and matched controls*
 Juvenile DM patientsMatched controls
  • *

    Values are the median (range) unless otherwise indicated. N = 59 for patients and controls unless otherwise stated. DM = dermatomyositis; NA = not assessed.

  • Scored in patients and controls ages ≥12 years (n = 51).

  • P < 0.05.

Females, no. (%)36 (61)36 (61)
European origin, no. (%)57 (97)57 (97)
Age at symptom onset, mean ± SD years8.3 ± 4.5NA
Age at diagnosis, mean ± SD years9.1 ± 4.5NA
Followup time, mean ± SD years16.9 ± 10.6NA
Age at followup, years21.5 (6.7–55.4)21.6 (6.2–55.4)
Height, cm168.0 (121.9–200.0)168.7 (122.0–191.9)
Weight, kg62.8 (23.3–115.1)63.0 (23.4–109.2)
Body mass index, kg/m221.5 (14.1–33.7)22.0 (13.3–33.9)
Hours of physical activity, category3 (0–5)4 (0–5)
Frequencies of physical activity, category3 (0–6)4 (0–6)

Muscle involvement, health status, and use of medication.

Twenty-five patients (42%) versus 2 controls (3%) had muscle weakness according to MMT (score <78), and 18 patients (31%) versus 3 controls (5%) had reduced muscle strength/endurance according to the CMAS (score <48). When using a more arbitrary cutoff for muscle weakness (MMT score <80 and CMAS score <52), 42 patients (71%) versus 7 controls (12%) and 42 patients (71%) versus 16 controls (29%) were found to have low scores, respectively. The median MMT score was 78 (range 58–80) for juvenile DM patients compared with 80 (range 76–80) for the controls (P < 0.001). The median CMAS score for juvenile DM patients was 50 (range 23–52) versus 52 (range 45–52) for the controls (P < 0.001) (Figures 1A and B). The median SF-36 PCS score for the patients was 53.9 (range 26.9–61.5) versus 56.9 (range 32.1–63.7) for the controls (P = 0.014) (Figure 1C). No difference was found in the SF-36 MCS score, any of the measured muscle enzymes, or the ESR between patients and controls (data not shown). However, the creatinine level was lower in patients compared with controls (Figure 1D).

Figure 1.

Measures for muscle strength (A and B), physical health (C), and creatinine level (D) assessed at followup in patients with juvenile dermatomyositis and age- and sex-matched controls (n = 59 unless otherwise stated). Data are shown as box plots, where the boxes show the 25th to 75th quartiles and the lines inside the boxes show the median. B, n = 56 for controls, C, n = 51 for patients and controls. MMT = manual muscle testing; CMAS = Childhood Myositis Assessment Scale; SF 36-PCS = Short Form 36 health survey physical component summary.

When comparing the part of the cohort diagnosed after 1990 with their respective controls (n = 30), the MMT and CMAS score were significantly lower in the patients (P < 0.001 and 0.002, respectively); the SF-36 PCS scores and creatinine level were also nominally lower, albeit not reaching statistical significance (P < 0.291 and 0.274, respectively). The MMT score was lower in the 17 patients (29%) receiving an immunosuppressive agent at followup compared with the 42 patients (71%) who did not (median 75, range 58–80 versus median 79, range 66–80; P = 0.001), whereas the CMAS score was not significantly reduced (data not shown).

Genetic markers and autoantibodies.

The distribution of carrier frequencies of HLA–DRB1 alleles in patients with juvenile DM and healthy controls is shown in Table 2. HLA–DRB1*0301 and HLA–DRB1*1401 were carried more often by patients than controls.

Table 2. Carrier frequency of selected HLA–DRB1 alleles in patients with juvenile DM and controls*
 Juvenile DM patients (n = 59)Healthy controls (n = 898)OR (95% CI) versus controls
  • *

    Values are the number (percentage) of patients and controls positive for each allele unless otherwise indicated. The DRB1 phenotypes shown were selected based on associations seen in the present and/or previous studies. DM = dermatomyositis; OR = odds ratio; 95% CI = 95% confidence interval.

  • P ≤ 0.05.

  • P ≤ 0.01.

  • §

    P ≤ 0.005.

DRB1*018 (14)203 (23)0.56 (0.27–1.16)
DRB1*01016 (10)170 (19)0.52 (0.23–1.16)
DRB1*01022 (3)8 (1)4.55 (1.20–17.28)
DRB1*01030 (0)19 (2)0.38 (0.05–2.87)
DRB1*030123 (39)219 (24)1.99 (1.17–3.40)
DRB1*07016 (10)149 (17)0.61 (0.27–1.36)
DRB1*12012 (3)47 (5)0.78 (0.23–2.58)
DRB1*14016 (10)28 (3)3.71 (1.55–8.86)§
DRB1*150114 (24)225 (25)0.95 (0.52–1.74)

Positive MSAs at followup were found in 2 patients with anti–signal recognition particle and anti–Jo-1. Positive MAAs were found in 3 patients: anti-Ku (1 positive) and anti–Ro 52 (2 positive). ANAs were detected at a titer of ≥40 in 24 patients (41%), whereas 15 patients (25%) had a titer of ≥160.

MRI findings.

MRI-detected muscle damage was found in 30 patients (52%) and muscle edema interpreted as most likely inflammatory disease activity was found in 5 patients (9%; 4 of 5 of the patients with probable inflammatory activity also had muscle damage). The muscle fatty infiltration score (median 0, range 0–3) showed a weak correlation with body mass index (rs = 0.279, P = 0.034). The median muscle edema score was 0 (range 0–2). The MRI findings in different soft tissue layers are shown in Table 3.

Table 3. MRI findings in 58 patients with juvenile DM a median of 16.8 years after disease onset*
 MuscleFasciaSubcutisCutisAny layer
  • *

    Values are the number (percentage) of the total. MRI = magnetic resonance imaging; DM = dermatomyositis; NA = not assessed.

Edema5 (9)0 (0)2 (3)0 (0)6 (10)
Atrophy17 (29)NA11 (19)NA21 (36)
Fatty infiltration25 (43)NANANA25 (43)
Calcinosis9 (16)11 (19)4 (7)3 (5)14 (24)

Relationship between the outcome measures, disease variables assessed at followup, and disease duration.

Correlations between the outcome measures and other disease variables are shown in Table 4. MMT and CMAS scores at followup correlated moderately (rs = 0.522, P < 0.001). No correlations were found between the outcome variables and any muscle enzyme (data not shown). When analyzing patients diagnosed after 1990 separately, the CMAS and MMT still correlated significantly with the SF-36 PCS and the HAQ/C-HAQ (data not shown). The CMAS also correlated moderately with MMT, DAS muscle, and the MDI when analyzing the patients ages >18 years separately (data not shown).

Table 4. Correlations between muscular outcome and other disease variables in patients with juvenile DM after a median of 16.8 years*
Disease variable (possible range)MRI-detected muscle damageCMASMMT
  • *

    Values are the Spearman's correlation coefficient. N = 59 unless otherwise stated. DM = dermatomyositis; MRI = magnetic resonance imaging; CMAS = Childhood Myositis Assessment Scale; MMT = manual muscle testing; DAS = Disease Activity Score; MDI = Myositis Damage Index; HAQ = Health Assessment Questionnaire; C-HAQ = Childhood Health Assessment Questionnaire; SF-36 = Short Form 36 health survey; PCS = physical component summary; MCS = mental component summary.

  • P < 0.01.

  • P ≤ 0.001.

  • §

    P < 0.05.

CMAS (0–52)−0.40
MMT (0–80)−0.110.52
DAS muscle (0–11)0.43−0.71−0.74
DAS skin (0–9)0.13−0.10−0.08
MDI total (0–40)0.71−0.42−0.16
MDI muscle score (0–3)0.60−0.50−0.41
HAQ/C-HAQ score (0–3)0.31§−0.39−0.36
SF-36 PCS (0–100; n = 51)−0.240.520.44
SF-36 MCS (0–100; n = 51)0.15−0.08−0.16
Serum creatinine level−0.180.380.54
Erythrocyte sedimentation rate0.18−0.44−0.28§
Cumulative prednisolone at followup0.44−0.28§−0.11
Hours of physical activity, category (0–5)−

In patients with a followup time of 2.0–9.9 years, 10.0–19.9 years, 20.0–29.9 years, and 30.0–40.0 years, MRI-detected damage was found in 5 (22%) of 23, 9 (69%) of 13, 10 (71%) of 14, and 6 (67%) of 9 patients, respectively, a low CMAS score was found in 6 (26%) of 23, 2 (15%) of 13, 5 (36%) of 14, and 5 (56%) of 9 patients, respectively, and a low MMT score was found in 11 (48%) of 23, 7 (54%) of 13, 5 (36%) of 14, and 2 (22%) of 9 patients, respectively. MRI-detected muscle damage was more often found in patients diagnosed before than after 1990 (71% versus 33%; P = 0.004).

Early predictors for muscle weakness and MRI-detected muscle damage.

A low MMT score (<78) after a median of 16.8 years was predicted by MDI muscle damage 1 year postdiagnosis, and a low CMAS score (<48) after a median of 16.8 years was predicted by a high DAS muscle 1 year postdiagnosis (Table 5). MRI-detected muscle damage was predicted by MDI muscle damage and a high DAS total 1 year postdiagnosis. We were not able to identify any predictors assessed at the time of diagnosis for any of the outcomes (data not shown).

Table 5. Early predictors of unfavorable muscular outcome after a median of 16.8 years in 59 patients with juvenile DM*
Outcome variables (controlling factors/early disease variables)Univariate analysesMultivariate analyses
OR (95% CI)POR (95% CI)P
  • *

    Result of the logistic regression analyses. Variables associated with the outcome variables in the univariate analyses (P < 0.15) were tested in the multivariate analyses, except that only one DAS score was chosen (due to high intercorrelation between DAS total, DAS skin, and DAS muscle). DM = dermatomyositis; OR = odds ratio; 95% CI = 95% confidence interval; MMT = manual muscle testing; MDI = Myositis Damage Index; DAS = Disease Activity Score (range 0–20); CMAS = Childhood Myositis Assessment Scale; MRI = magnetic resonance imaging.

  • The following variables were tested univariately: age at diagnosis; duration from symptom to diagnosis; HLA–DRB1 alleles; initial prednisolone dose <1 mg/kg; cumulative prednisolone (gm) at 1 year/body weight (kg); started antiinflammatory treatment and started methotrexate treatment during the first year postdiagnosis; DAS skin, DAS muscle, and DAS total at 1 year; MDI muscle damage at 1 year; muscle enzymes; and positive antinuclear antibodies and erythrocyte sedimentation rate at 1 year. If not listed in the table, these variables did not correlate univariately with the outcomes (P > 0.15).

  • In the multivariate analyses, sex and age at followup were used as controlling factors, and other variables were entered by backward elimination.

Low MMT (score <78)    
 Female sex6.65 (1.88–23.57)0.0038.28 (1.77–38.70)0.007
 Age at followup, years0.97 (0.92–1.01)0.1270.92 (0.86–0.99)0.020
 MDI muscle damage 1 year postdiagnosis, no/yes5.06 (1.55–16.46)0.00724.09 (3.13–185.64)0.002
 DAS muscle score 1 year postdiagnosis1.24 (0.98–1.58)0.078  
 Antiinflammatory medication started within the first year2.25 (0.78–6.51)0.134  
 Cumulative prednisolone at 1 year (gm)/kg bodyweight1.01 (1.00–1.01)0.110  
Low CMAS (score <48)    
 Female sex1.42 (0.44–4.52)0.5561.54 (0.41–5.88)0.524
 Age at followup, years1.04 (0.99–1.09)0.0761.02 (0.97–1.08)0.512
 MDI muscle damage 1 year postdiagnosis, no/yes4.44 (1.36–14.58)0.014  
 DAS muscle score 1 year postdiagnosis1.62 (1.21–2.16)0.0011.56 (1.15–2.12)0.004
 DAS total score 1 year1.24 (1.05–1.45)0.010  
MRI-detected muscle damage    
 Female sex0.54 (0.19–1.58)0.2610.49 (0.97–1.11)0.257
 Age at followup, years1.08 (1.02–1.13)0.0051.04 (0.19–1.05)0.236
 MDI muscle damage 1 year postdiagnosis, no/yes14.86 (3.00–74.12)0.0017.80 (1.31–46.52)0.024
 DAS muscle score 1 year postdiagnosis1.70 (1.23–2.36)0.001  
 DAS skin score 1 year postdiagnosis1.55 (1.16–2.07)0.003  
 DAS total score 1 year postdiagnosis1.41 (1.16–1.71)< 0.0011.25 (1.01–1.53)0.039
 Initial prednisolone dose <1 mg/kg, no/yes2.35 (0.82–6.78)0.113  
 Duration of symptom to diagnosis1.81 (0.83–3.94)0.133  
 HLA–DRB1*01 positive8.22 (0.94–71.82)0.057  


In our study, the patients with juvenile DM were weaker, with lower endurance and physical health, than matched controls from the general population. After a median of 16.8 years, muscle weakness was found in 31–42% of the patients, and MRI-detected muscle damage was found in 52%. Sustained muscle and skin disease activity and early muscle damage 1 year postdiagnosis predicted muscle weakness and MRI-detected muscle damage at followup. To our knowledge, this is the first controlled study to investigate long-term muscular outcome and prognostic factors in juvenile DM.

We described earlier the representativeness of our cohort (22), which we believe contains the majority of Norwegian patients with juvenile DM diagnosed between 1970 and 2006. Our cohort is comparable to other hospital- or registry-based cohorts with regard to female predominance, age at diagnosis, medication, and muscle weakness at disease onset (11, 35–37).

Our matched controls were randomly selected from the Norwegian population register, which is a strength of our study. Patients with juvenile DM were weaker than the controls and had poorer self-reported physical health. Patients diagnosed after 1990 were also weaker than their respective controls, but no significant difference was found with regard to physical health. Creatinine level was lower in the patients; a low creatinine level has been associated with muscle damage (5). The patients exercised less than the controls; however, the level of weekly exercise did not correlate with muscle strength in the patients. Furthermore, muscle strength showed a positive correlation with physical health (SF-36 PCS) and a negative correlation with the HAQ/C-HAQ in the patients, and also when analyzing the patients diagnosed after 1990 separately. This supports the fact that muscle weakness has an impact on experienced physical health and disability in this patient group. The examiners were not blinded to which individuals were patients or controls, which could have influenced the results, with a tendency to give the patients lower scores. On the other hand, patients often cooperate to a higher degree than controls in testing situations (38). Taken together, we therefore believe our results are valid.

In our study, 31–42% of the patients had muscle weakness and reduced endurance as assessed by MMT (score <78) and the CMAS (score <48). Our cutoff levels for muscle weakness were relatively low, but were based on the scores in matched controls. To our knowledge, there are no established cutoff levels for muscle weakness using these tools. In a newly published uncontrolled, multinational, multicenter study of 373 patients with juvenile DM examined after 7.7 years of followup, reduced muscle strength was found in 41% (defined as an MMT score <80) and 53% (defined as a CMAS score <52) of patients (6). Using these cutoffs, 71% of our patients had reduced muscle strength for MMT and the CMAS. The discrepancy between these studies might be explained by differences in the application of the methods, since most patients only had moderate weakness. However, in our study, the CMAS and MMT were scored by different examiners blinded for each others' scores; by both methods, our patients were weaker than recently reported by Ravelli et al (6). Furthermore, MMT and the CMAS correlated moderately to strongly with each other and with other measures for disease activity and organ damage; the CMAS also correlated with MRI-detected muscle damage. Taken together, we believe this supports our findings. The CMAS has not previously been applied in adults with juvenile-onset DM. However, the CMAS also correlated with other disease measurers in the patients ages >18 years, supporting the use of the tool also in this patient group.

MRI changes compatible with muscle damage were found in ∼50% of the patients, and correlated strongly with cumulative organ damage. Comparable correlation between the T1-weighted score (calcinosis not included) and muscle damage in patients with juvenile DM has recently been found (16). More patients diagnosed before than after 1990 had changes compatible with muscle damage shown by MRI; we have earlier shown that the patients diagnosed before 1990 were treated less aggressively, but with a higher cumulative dose of prednisolone (22). The muscle damage shown by MRI may partly be explained by steroids, supported by the positive correlation between cumulative prednisolone dose and signs of MRI-detected muscle damage in the present study. However, a high cumulative prednisolone dose is probably a result of high disease activity over time. Weekly exercise did not correlate with muscle damage on MRI in our study. It is previously known that muscle atrophy can be the result of chronic muscle inflammation, steroid therapy, or inactivity (13, 39). We did not perform MRI scans in our controls, but in previous reports, focal intramuscular changes or fatty replacement has not been described in the normal population (including overweight people) (8, 40, 41).

The number of patients with MSAs was lower than reported in other studies (17–20). Since MSAs/MAAs for most patients were analyzed in blood samples drawn many years after disease onset, we may have underestimated the true frequency of these antibodies. It has been shown that anti–Jo-1 might disappear after prolonged remission (42, 43). Alternatively, the lower frequency of MSAs could be due to lower sensitivity of the dot-blot method as compared with immunoprecipitation technology (19). Given the low frequency of MSAs and MAAs in our study, we did not analyze the correlation between these antibodies and the outcomes.

A clear association between juvenile DM and the carriage of particular HLA–DRB1 alleles was observed. In accordance with previous publications (17, 44), the frequency of patients with HLA–DRB1*0301 was increased and supports this allele as a predisposing genetic factor. Whether DRB1*0301 is directly involved or whether it marks another causal variant on the autoimmune ancestral haplotype AH8.1 remains to be settled. Interestingly, we found a novel association with DRB1*1401, which suggests that this allele also acts as a predisposing genetic factor for juvenile DM. However, this result needs to be confirmed in larger data sets. The previously proposed protective HLA alleles DRB1*01 and DRB1*07 (44) also occurred at a reduced frequency among our patient population, albeit not reaching statistical significance. This could be due to limited power and also possibly the discrepancy in the direction of association for different DRB1*01 subtypes. In contrast, DRB1*15, also suggested to confer protection (44), was equally frequently present in our cases and controls. Surprisingly, the reported protective factor HLA–DRB1*01 was univariately correlated with MRI-detected muscle damage at followup, but did not reach statistical significance in the multivariate analyses.

We have earlier identified high disease activity and organ damage present at 6 months postdiagnosis as predictors of cumulative organ damage at followup in the same cohort (22). We now show that early muscle damage and sustained disease activity 1 year postdiagnosis predict muscle weakness and MRI-detected muscle damage at followup. These findings again highlight the impact sustained early disease activity and early damage have on unfavorable outcomes in juvenile DM (6, 45). A low starting dose of prednisolone and a long duration from symptom to diagnosis correlated with MRI-detected muscle damage in the univariate but not in the multivariate analysis; we might have been underpowered to detect such associations. Our study is also limited by the retrospective application of data assessment from early disease course in addition to the cross-sectional design, with a lack of longitudinal assessments of the outcome measures.

In conclusion, patients with juvenile DM after a median of 16.8 years of followup were weaker and had lower endurance and physical health than the matched controls. Muscle weakness and MRI-detected muscle damage at followup were predicted by muscle damage and muscle disease activity 1 year postdiagnosis. These results show that juvenile DM in our cohort has an impact on muscle strength many years after diagnosis, and also in patients diagnosed after 1990. Hopefully, the muscular outcome in juvenile DM will improve with advances in therapy.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Sanner 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 conception and design. Sanner, Gran, Flatø.

Acquisition of data. Sanner, Kirkus, Merckoll, Tollisen, Røisland, Lie, Taraldsrud, Gran, Flatø.

Analysis and interpretation of data. Sanner, Kirkus, Merckoll, Lie, Taraldsrud, Flatø.


We thank Are Hugo Pripp for statistical advice, Siri Tennebø Flåm for DRB1 genotyping, and the Norwegian Bone Marrow Registry for making control frequencies for DRB1 available.