History of infection before the onset of juvenile dermatomyositis: Results from the National Institute of Arthritis and Musculoskeletal and Skin Diseases Research Registry




To obtain data concerning a history of infection occurring in the 3 months before recognition of the typical weakness and rash associated with juvenile dermatomyositis (JDM).


Parents or caretakers of children within 6 months of JDM diagnosis were interviewed by the registry study nurse concerning their child's symptoms, environment, family background, and illness history. Physician medical records were reviewed, confirming the JDM diagnosis.


Children for which both a parent interview and physician medical records at diagnosis were available (n = 286) were included. Diagnoses were as follows: definite/probable JDM (n = 234, 82%), possible JDM (n = 43, 15%), or rash only (n = 9, 3%). The group was predominantly white (71%) and had a girl:boy ratio of 2:1. Although the mean age at onset was 6.7 years for girls and 7.3 years for boys, 25% of the children were ≤4 years old at disease onset. In the 3 months before onset, 57% of the children had respiratory complaints, 30% had gastrointestinal symptoms, and 63% of children with these symptoms of infection were given antibiotics.


This study provides evidence that JDM affects young children. The symptoms of the typical rash and weakness often follow a history of respiratory or gastrointestinal complaints. These data suggest that the response to an infectious process may be implicated in JDM disease pathogenesis.


Juvenile dermatomyositis (JDM), an often devastating systemic vasculopathy with primary disease expression in skin and muscle, is the most common of the pediatric inflammatory myopathies (1). The 2 major components of this illness are rash and symmetric proximal muscle weakness. Children with JDM are easily identified by the characteristic periorbital heliotrope rash, giving a raccoon-like appearance, which may also cross the bridge of the nose. Other localized erythematous cutaneous signs include papules or “alligator skin” over the metacarpal and proximal phalangeal joints, knees, and elbows, but the often-pruitic rash can involve the limbs and trunk as well. The second major symptom of JDM, symmetric proximal muscle weakness, can be insidious in onset and is frequently recognized only when the child has loss of functional ability, such as climbing into bed or out of a car. Either symptom brings the child to medical attention, but as many as 4–5 physicians may be consulted before the correct diagnosis is achieved and the appropriate medical therapy initiated (2).

Few studies have characterized the 3-month time frame preceding the recognition of rash or weakness, although there is limited evidence that an infectious event may be associated with JDM (3, 4). A national case-control study (1989–1992) compared data from directed interview of the parents of children with newly diagnosed JDM with similar data from controls, who were age-matched playmates or children with juvenile rheumatoid arthritis residing in the same region during a comparable time period (5). More JDM children (63%) had symptoms of an illness in the 3 months before the recognition of definite symptoms of JDM compared with 41% of control children during the same time period (5). However, the type of specific illness, such as the association of fever with gastrointestinal or respiratory symptoms, was not characterized in that study (5). Studies of infectious agents, including group A beta hemolytic streptococci (6, 7) and picornavirus (3, 8), by testing tissue from newly diagnosed JDM patients have been inconclusive. Studies of JDM sera for antibody to enterovirus, including coxsackievirus B (CVB), have been inconsistent as well (4, 5, 8). In contrast, gene expression profile data from untreated JDM muscle documented an intense interferon α/β (IFN-α/β)–induced response, compatible with an immune response to a microbial antigen (9). Finally, clusters of cases of JDM have been reported in brief, also suggesting possible temporal and environmental components in the pathogenesis of JDM (10–12).

The purpose of this study was to obtain information by telephone interview of parents and caretakers of patients with newly diagnosed JDM concerning the child's symptoms in the 3-month time period prior to the first definite JDM complaint of rash or weakness.


Patient recruitment.

Children newly diagnosed with JDM who lived in the continental USA were recruited for this investigation from 1994 through 1999. The study was publicized at scientific and medical meetings as well as in medical journals. Registry participation in the last 2 years of the study (1998–1999) was also facilitated by a Web site, which provided access to the consent form required for study entry. Patients were referred from 45 states; 398 cases were referred by physicians and 52 cases by parents (either by hotline, Web page, or word of mouth). The referring physicians included pediatric rheumatologists (87%), neurologists (6%), adult rheumatologists (4%), pediatricians (1%), and dermatologists and other specialists (0.1%).

Study population.

Of the 450 patient referrals, there were 323 families with children who were newly diagnosed with JDM (within 6 months of diagnosis) and who were willing to participate in the study. Once the signed informed consent was returned, the registry nurse project manager contacted the parents and scheduled a 1-hour telephone interview. The interview format had been field tested and validated prior to implementation, and was conducted within 4–12 months of the diagnosis date. A copy of the physician medical record at diagnosis was also requested to provide medical information concerning the diagnostic criteria of Bohan and Peter (13), to confirm patient eligibility for this study. The data for this report were considered only if both family interview and physician records at diagnosis had been obtained. Racial background of the child was assigned by parent self classification, which included the race and country of origin of the parents and grandparents, when known. When the parents were unable or reluctant to designate race, the assignment of race was derived from the matching physician chart review obtained at diagnosis. Identification of the first definite JDM symptom (rash and/or weakness) by the caretaker was designated as the date of disease onset. The diagnosis date was defined as the day when a physician first provided the classification “juvenile dermatomyositis.” The children were divided into 2 groups—≥6 years of age at disease onset and <6 years of age—to determine if there were possible differences in symptoms between the 2 groups that may be associated with maturation. For example, circulating lymphocyte subsets and cytokine production (14), as well as immunoglobulin levels, including Gm allotypes (15), vary with the age of the child and appear to stabilize at ∼6 years of age.

Questionnaire content.

Prior to their scheduled interview, the families of children with JDM were requested to gather available information (medical bills, immunization records, travel records) concerning the period of time that was of interest. The registry nurse used an interview form consisting of 100 questions (see Appendix A, available at the Arthritis Care & Research Web site at http://www.interscience.wiley.com/jpages/0004–3591:1/suppmat/index.html), which included 6 major topics: an account of the course of JDM up to the time of the interview, history before JDM, ethnic background, family background (i.e., ethnicity, structure, socioeconomic status, and history of family illness), the child's medical background, and environmental exposures (i.e., stress, infectious agents, and toxins). A central concern was the child's symptoms, including functional status at disease onset and diagnosis, as well as the date and type of medical therapy initiated.

Data collection and processing.

The interview data were entered into a Paradox database system (Corel Corporation, Ottawa, Ontario, Canada), reviewed, confirmed, and then analyzed by the Methodology and Data Management Core Unit (Northwestern University Feinberg School of Medicine, Division of Rheumatology) in the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)-sponsored Multidisciplinary Clinical Research Center. Comparison of the response from the subgroup of parents interviewed within 6 months of the JDM diagnosis with the response from the group of parents who were interviewed 1 year after the JDM diagnosis yielded similar data, suggesting minimal recall bias. The interview results were pooled for final analysis.

Data analysis.

Data were examined using descriptive methods. Chi-square and t-tests or Fisher's exact tests were utilized for comparisons of variables between groups. Statistical analysis was performed using SAS version 8.2 (SAS Institute, Cary, NC). Because of the descriptive nature of this study, no correction was made for the number of comparisons tested. Analysis of the association of animal exposure with symptoms of JDM used chi-square analysis (3 × 2).



A total of 323 children's families were enrolled in the NIAMS registry and interviewed, but medical records were not obtained for 37 children. Therefore, 286 patients were included in this report. Data from the physician records provided evidence that the patient population met the criteria of Bohan and Peter for the diagnosis of JDM (13): 234 (82%) of 286 of the enrolled children fulfilled the criteria for definite or probable JDM. Nine children (3%) had rash only and 43 (15%) were classified as having possible JDM (rash plus one other criterion, usually proximal muscle weakness).

The demographics of the children with newly diagnosed JDM are consistent with previous studies (Table 1). The majority were white (71%); 12% were Hispanic and 9% were African American (a racial composition similar to that of the overall US population). There were fewer children of American Indian (n = 3) or Asian (n = 2) origin. More than twice as many girls had JDM than boys (ratio 2:1). The mean age at disease onset ranged from 6.6 years (white girls) to 8.1 years (African American girls). Figure 1 shows the distribution by the children's age (in years) at which the parents noted the first JDM symptom for both boys and girls. Fifty percent of the children had their first JDM symptom by 6 years of age. Of note, 45 of the 194 girls (23%) were ≤4 years old (21 of whom [11%] were ≤2 years of age), whereas 29 of the 92 boys (31.5%) were ≤4 years old (8 of whom [9%] were ≤2 years of age), confirming the involvement of very young children in 25% of the entire group.

Table 1. Demographics of juvenile dermatomyositis study population (n = 286)
Demographicsn (%)MaleAge at onset, mean ± SD boysFemaleAge at onset, mean ± SD girlsAge at onset, mean ± SD total
  • *

    Consists of 3 American Indian/Alaskan, 2 Asian/Polynesian, and 17 mixed or unknown.

White203 (71)637.3 ± 4.11406.6 ± 3.46.8 ± 3.7
Hispanic35 (12)147.3 ± 3.6217.0 ± 4.37.1 ± 4.0
African American26 (9)117.6 ± 3.6158.1 ± 4.67.9 ± 4.2
Other*22 (8)47.9 ± 3.8186.4 ± 2.56.7 ± 3.7
Totals28692 194 6.9 ± 3.7
Figure 1.

Age at disease onset of juvenile dermatomyositis (JDM) in 286 children from the JDM National Institute of Arthritis and Musculoskeletal and Skin Diseases Registry.

History of symptoms reported in the 3 months before rash or weakness.

The majority of children, 159 of 286 (56%), had ≥1 of the following constitutional symptoms: fever, weight loss, fatigue, and headache (Table 2). When the group was divided into children <6 years of age compared with those ≥6 years, fever was more frequently reported for the younger group (57% versus 31%; P < 0.0001), whereas headache was a more common complaint in the older children (31% versus 16%; P = 0.005).

Table 2. Family-reported symptoms in children 3 months prior to onset of the rash or weakness of juvenile dermatomyositis
 n (%)<6 years of age, n (%)≥6 years of age, n (%)P
  • *

    Fever, fatigue, headache, or weight loss.

  • Cough, sore throat, rhinorrhea, cold, flu, lower respiratory illness (pneumonia, asthma, bronchitis), earache, otitis media, or strep throat.

  • Vomiting, diarrhea, nausea/upset stomach, or abdominal pain.

  • §

    Difficulty swallowing, difficulty speaking, weakness, muscle pain, arthritis, tightness of joints, or stiffness.

  • Change in color, paronychia, lumps, rash, finger pits, or pain with cold.

History of antecedent illness286129157 
Constitutional*159 (56)78 (60)81 (51)0.13
 Fever121 (42)73 (57)48 (31)< 0.0001
 Headache69 (24)21 (16)48 (31)0.005
All respiratory162 (57)83 (64)79 (50)0.02
 Earache43 (15)28 (22)15 (9)0.004
 Rhinorrhea119 (42)63 (49)56 (36)0.02
 Sore throat91 (32)34 (26)57 (36)0.07
 Lower respiratory/illness38 (13)22 (17)16 (10)0.09
Gastrointestinal87 (30)43 (33)44 (28)0.33
JDM related, - possible    
 Musculoskeletal§90 (31)36 (28)54 (34)0.24
  Muscle pain49 (17)14 (11)35 (22)0.01
  Stiffness36 (13)10 (8)26 (17)0.02
 Skin45 (16)23 (18)22 (14)0.38

More than half of the JDM patients (162 of 286; 57%) had a history of respiratory symptoms, which were more common in the younger group (64% versus 50%; P = 0.02). Upper respiratory symptoms, including earache and rhinorrhea, were more frequently reported for children younger than age 6 (P = 0.004 and 0.02, respectively). Gastrointestinal symptoms, including vomiting, diarrhea, nausea, or abdominal pain, were found in 87 (30%) of 286 of the children, regardless of age. Older children more frequently reported symptoms that could also be attributed to the myopathic process of JDM itself, such as generalized musculoskeletal complaints or muscle pain and stiffness (P = 0.01 and 0.02, respectively).

Use of medications in the 3 months before rash or weakness.

More than 60% of the entire group of children in this study were given medications for their symptoms, but there was no difference in the frequency of the use of antibiotics, nonsteroidal antiinflammatory drugs (NSAIDs), or cough medicine by age (data not shown).

Antibiotics were taken with similar frequency by children with constitutional and respiratory complaints (100 of 159; 63%) and with gastrointestinal complaints (56 of 87; 64%). Children with musculoskeletal symptoms were given antibiotics as well (60 of 90; 67%). Parents gave cough medicine to 91 (57%) of 162 children who had respiratory symptoms before the onset of JDM, and other prescription drugs were given to ∼20% of children with any of these complaints. In contrast, NSAIDs were given to only 17 (11%) of 159 children with constitutional symptoms and 15 (9%) of 162 children with any respiratory problem.

Environmental factors.

The only positive association of history of environmental exposure with the onset of JDM that was identified is shown in Table 3. Children with a history of contact with sick animals had more constitutional and gastrointestinal symptoms than children who had JDM without sick-animal contact. The types of animal contacts were primarily domestic in nature (dogs, cats) and were not different between the 2 groups. No differences were seen in season of the year for either the report of symptoms of infection or recognition of the JDM rash and weakness. Furthermore, no significant differences were obtained using a 4-month moving window in each year of the study, or with data analysis by year of disease onset or by Centers for Disease Control and Prevention epidemiology region. Other factors pertinent to the child's residence (such as size of community, residence age, relationship to bodies of water) showed no trends for constitutional, respiratory, or gastrointestinal symptoms ascertained by parent report. The symptoms of children living within 5 miles of open water (87%), with or without the inclusion of swimming pools, were not significantly different from children who did not have this exposure. The parent's response to questions concerning possible chemical exposure did not show an association with either the onset of JDM or symptoms of infection at the beginning of the illness. Other factors not found to be significant were maternal age, the mode of feeding (breast or bottle) in the neonatal period, or presence of maternal breast implants.

Table 3. Children with juvenile dermatomyositis: animal exposure at disease onset (n = 286)
HistoryNo exposure to domestic, farm, or wild animals n = 50Exposure to healthy animals n = 82Exposure to sick animals n = 54P
  • *

    Fever, fatigue, headache, or weight loss.

  • Vomiting, diarrhea, nausea/upset stomach, or abdominal pain.

  • Fever, cough, sore throat, runny nose, headache, tiredness, nausea, vomiting, diarrhea, rash, earache, cold, flu, pneumonia, asthma, bronchitis, weight loss, weakness, muscle pain, arthritis, abdominal pain, blood in stool, difficulty swallowing, difficulty speaking, lumps or calcifications, shortness of breath, tightness, pitting of fingers, coldness of hands, change in color of hands or feet, stiffness, paronychia, urinary tract infections, otitis media, strep throat, or other.

Constitutional, n (%)*23 (46)98 (54)38 (70)0.032
Gastrointestinal, n (%)13 (26)50 (28)24 (44)0.044
Any antecedent illness, n (%)29 (58)113 (62)43 (80)0.033


JDM is a rare disease, with an incidence of 2.1/million children/year (16). There are few descriptions of the patient before the recognition of JDM rash and weakness.

The data in this study document, for the first time, show the occurrence of early symptoms of JDM in very young children: 25% were ≤4 years old. The JDM complaints often followed a history of symptoms compatible with an infectious process, which the parents reported, and involved the respiratory system in >50% of cases and gastrointestinal tract in ∼33%. Not unexpectedly, infectious symptoms were more frequent in children ≤6 years old. Physician prescription of antibiotics to 63% of children with infection-related symptoms in the 3 months prior to the diagnosis of JDM suggests that their illness was concerning to medical personnel, but control data are lacking. Although the present study was not case-control in design, the historical data suggest that symptoms associated with an infection precede the onset of JDM.

The demographic features of the NIAMS registry JDM patient group are consistent with previous descriptions of smaller studies of children diagnosed with JDM living in the Western world, with a predominance of female and white children (2, 17, 18), and differ elsewhere (19). It is known that JDM is a disease of young children (20). However, Figure 1 shows that 25% of the patients are diagnosed with JDM at age 4 or younger, which had not been previously documented. This observation is of importance because most of the examination tools used to assess disease activity and disease progression in JDM (Childhood Myositis Assessment Scale, Childhood Health Assessment Questionnaire) have been validated for children older than 4 years of age (21, 22). All of the children in this study were classified by the Bohan and Peter criteria (13) and met the diagnosis of JDM; only 9% had no muscle involvement at diagnosis. The frequency of a persistent amyopathic form of JDM is controversial (23), because most, but not all, children with amyopathic JDM lacking muscle involvement at diagnosis develop myositis later during the disease course (24).

Several lines of evidence from case studies and molecular investigations suggest that an immune response to antigen stimulation may play a role in the pathogenesis of JDM. A previous national case-control investigation (1989–1992) documented that age- and region-matched controls had fewer symptoms of infection in the 3 months prior to interview (44%) than the children who developed JDM (64%) in the 3 months before rash or weakness onset (P = 0.03). However, the number of children in that study (n = 80) was not sufficient to identify the type of infection, thus providing the incentive for this investigation (5).

Infectious agents associated with the clinical onset of inflammatory myositis in both adults and children, with or without the rash of dermatomyositis, include parvovirus B19 (25) and Borelia burgdorferi (26). There are reports of increased antibody titers to hepatitis C virus (27) or hepatitis B (28) in patients with polymyositis/dermatomyositis, as well as documentation of chronic staphlococcal infection followed by the symptoms of JDM (29). Other infectious agents that have been considered in the pathogenesis of JDM include influenza, parainfluenzae, herpes simplex, and Toxoplasma gondii (30, 31), but serologic confirmation is lacking for these agents (5). Human T lymphotropic virus type I has also been implicated in a few children (32–34).

The infectious agents most frequently identified in JDM patients are CVB (4, 35) or enterovirus (5, 36), which often is associated with both upper respiratory and gastrointestinal symptoms (the most common complaints in this study). Neutralizing and complement-fixing antibody titers to coxsackievirus B2 and B4 were increased in JDM sera compared with titers from age- and geographic-matched control samples, suggesting previous exposure (4). The serologic evidence of CVB infection is not consistent, however, because a national case-control study documented elevated complement fixing enteroviral titers in JDM sera compared with their age-matched controls, but none were directed against specific CVB antigens (5).

Group A beta hemolytic streptococcus infection can not only precipitate symptoms of dermatomyositis but also can incite clinical recurrence of symptomatic disease (7). A streptococcal M protein peptide (aa 367-375) and its myosin homolog (aa 114-122) selectively stimulated JDM lymphocytes (cytotoxic response, increased lymphocyte proliferation, and tumor necrosis factor α production) when JDM peripheral blood mononuclear cells were compared with controls (6); stimulation that is compatible with immune activation (37). In addition, these streptococcal peptides also elicited limited T cell receptor Vβ usage (12, 15, 18) using a T cell capture analysis (6).

Further evidence for an activated immune response compatible with a preceding infection in JDM is obtained from gene expression profiles of muscle of untreated children with JDM, which show an increase in IFN-α/β inducible immune genes (9). Both viral and microbial antigens are associated with induction of IFN-α/β (38), adding to the evidence that a possible infectious agent may be central to the pathophysiology of JDM.

In conclusion, a history of complaints compatible with an enteroviral infection (including upper respiratory or gastrointestinal symptoms) occurred in the 3 months prior to the diagnosis of JDM for some, but not all, children. It remains to be seen if these specific symptoms of infection in children before they are diagnosed with JDM are associated with a particular type of disease course or outcome.


This study could not have been performed without the careful work of Anh Chung. The data could not have been amassed without the efforts of the referring clinicians. We thank the following participants who referred patients to the NIAMS JDM registry for their dedication and contributions: Drs. Leslie Abramson, Balu Athreya, Susan Ballinger, Bram Bernstein, Bradley Bloom, Jonathon Bohnsack, Elizabeth Candell, James Cassidy, Joe Cole, Chester Fink, Robert Fuhlbrigge, Harry Gewanter, Tom Griffin, Hillary Haftel, Jean Hayward, Michael Henrickson, Gloria Higgins, Roger Hollister, Donna Hummell, Masanori Igarashi, James Jarvis, Rita Jerath, Mary Jones, Larry Jung, Sedney Kahn, Gregory Keenen, Yukiko Kimura, Alexander Lawton, Thomas Lehman, Bob Leschner, Carol Lindsley, Betty Lowe, Katherine Madson, Alan Martin, Richard Mier, Paula Morris, Arthur Newman, Jonathon Nocton, Katherine O'Neil, Barbara Ostrov, Murray Passo, Linda Ray, Robert Rennebohm, Deborah Rothman, Kenneth Schikler, Peggy Schlesinger, Robert Sheets, David Sherry, Richard Silver, Charles Spencer, Lenny Stein, Robert Sundell, Ilona Szer, Carol Wallace, and Dowain Wright.