Seasonal influence on the onset of idiopathic inflammatory myopathies in serologically defined groups




To assess possible seasonal patterns in the onset of polymyositis (PM) and dermatomyositis (DM).


The study group comprised 503 patients who met the criteria for probable or definite PM or DM and for whom detailed data on the time of myositis onset were available. Statistical analyses were performed using a Poisson model that assessed associations of ethnicity, sex, autoantibody presence, and month of onset of muscle weakness.


There were no significant seasonal patterns of disease onset in myositis patients as a whole or in the total PM or DM populations. Significant seasonal associations were present, however, in the serologically defined groups. In the 131 patients with antisynthetase autoantibodies who were categorized as non-black, myositis onset peaked in March–April (P = 0.03). Among the antisynthetase-positive patients, the association was predominantly in those with PM (n = 85; P = 0.05) and in men (n = 51; P = 0.042). Patients with anti–signal recognition particle autoantibodies, however, did not have a significant seasonal onset, which is in contrast to previous findings. Patients without myositis-specific autoantibodies showed a significant peak in summer, with myositis onset in June–July (n = 252; P = 0.03); this seasonal association was significant in women (n = 182; P = 0.005), whereas there was no seasonal pattern in men (P = 0.9).


These findings, in conjunction with other data, suggest that diverse environmental agents, acting upon different immunogenetic backgrounds, result in distinct immune responses and clinical syndromes in the idiopathic inflammatory myopathies. Our results emphasize the importance of considering more homogeneous disease groups, based on clinicopathologic features, immune responses, ethnicity, and sex, when attempting to decipher the pathogeneses of autoimmune disorders.

Idiopathic inflammatory myopathy (IIM) constitutes a diverse group of syndromes that share the characteristic of weakness due to chronic inflammation in the muscle as their common feature (1). The most frequent forms, polymyositis (PM) and dermatomyositis (DM), are traditionally diagnosed by the presence of proximal muscle weakness, elevations in the serum levels of muscle-associated enzymes, characteristic rashes in the case of DM, typical myopathic changes on electromyography, and chronic mononuclear cell infiltration with evidence of degeneration and regeneration of myocytes on muscle biopsy (2). Although the causes of these disorders remain unknown, many lines of evidence suggest that they are the result of environmental exposures that induce immune activation in genetically susceptible individuals, resulting in chronic tissue inflammation (1).

Case reports, case series, and epidemiologic investigations have suggested that either infectious or noninfectious agents could play a role in the pathogenesis of these increasingly recognized autoimmune syndromes (1, 3). Elevated titers of antibodies to certain viruses in some juvenile and adult patients with myositis (4, 5), as well as the presence of viral RNA in muscle biopsy specimens (6), have been reported, but not all studies support these findings (7, 8). In case reports and case series, other viruses (hepatitis B and C, human immunodeficiency virus, human T lymphotropic virus type 1, and influenza A and B) (9–13), parasites (toxoplasma, nematodes, and Wuchereria bancrofti) (14–16), and bacteria (group A streptococci) (17, 18) have been reported to be associated with myositis. Moreover, exposures to a number of noninfectious agents, including drugs (especially D-penicillamine and lipid-lowering agents) (19–21), vaccines (22, 23), medical devices (collagen and silicone implants) (24, 25), and ultraviolet radiation (26, 27), have been proposed as possible disease triggers. Finally, several animal models of myositis that are induced by viruses, drugs, or parasites have been developed, which add to the growing body of evidence for the likely role of environmental agents in the pathogenesis of IIM (28).

There are reports of spatial clustering (29) and seasonal associations with myositis onset or increased disease activity (30–32). Such clustering by place or time suggests a common infectious and/or noninfectious source, if bias can be excluded. These reports in small series of IIM patients encouraged us to undertake an independent study of a larger number of patients from multiple centers in an attempt to better understand the possible roles of clinicopathologic features, immune responses, ethnicity, and sex in the seasonal onset of myositis.


Study participants

A cross-sectional retrospective study of the time of onset of myositis was conducted in PM and DM patients from participating referral centers. Consecutive patients who met the criteria for definite or probable PM or DM (2) and for whom data were available regarding the presence of myositis-specific autoantibodies and the time of myositis onset, as defined by time of first muscle weakness, were studied. These criteria for PM or DM require the exclusion of infectious, metabolic, dystrophic, and other myopathies. Inclusion body myopathies were also excluded from this study due to their uncertain time of disease onset. The participating centers were located in Bethesda, Maryland (n = 229), Pittsburgh, Pennsylvania (n = 77), Houston, Texas (n = 35), Oklahoma City, Oklahoma (n = 7), and Aachen, Germany (n = 105). Another group of myositis patients (n = 50), recruited from a number of other centers in the US and who served as controls for the Scleroderma Criteria Cooperative Study (33), were also included in our study.

Data collection. Data were collected on a total of 503 patients (268 with PM and 235 with DM; 146 men and 357 women; 366 white, 109 black, and 28 other races), including demographic information and the date of first muscle weakness, which ranged from 1959 to 2000. For the purpose of the present analysis, patients were grouped as black or non-black. The non-black group comprised all races not self-categorized as African American (i.e., white, Hispanic, Asian, and others). Myositis-specific autoantibodies were determined by standard RNA and protein immunoprecipitation assays (34). Autoantibody tests for all patients from the US were performed in one laboratory (The Myositis Testing Laboratory, Oklahoma Medical Research Foundation) using validated methods. Autoantibody tests for the German subjects were done in Aachen, Germany, with the use of comparable methods.

Statistical analysis

Data representing the months of onset of muscle weakness were analyzed by applying a Poisson regression model. The regression approximated that described by Bacchetti (35), except that no corrections were used for the numbers of days or work days in the month. Season is inherently a circular variable, since December is adjacent to January, and a method for detecting seasonality needs to allow for smooth departures from the null hypothesis of flat incidence across the year. The model that we applied is based on the assumption of a sinusoidal relationship of the logarithm of the expected count and the onset of muscle weakness in a given month, with a trigonometric function (shifted sinusoidal function) of the angle corresponding to the middle of the month. Under the null hypothesis, in which it is assumed that symptom onset occurs regardless of the season, the data should be distributed approximately uniformly through the months of the year, i.e., as a multinomial with 12 cells and approximately equal probabilities. For subgroups of patients showing evidence of a seasonal pattern of onset, the maximum time of year at risk was estimated as the peak in the estimated sinusoidal curve, i.e., the time corresponding to the arc tangent of the a coefficient in relation to the b coefficient, where a and b are the estimated coefficients of the sine and the cosine, respectively.


The characteristics of the population evaluated in this study are summarized in Table 1. There were no significant seasonal patterns of myositis onset in the myositis patients as a whole or in the myositis subgroups defined as male, female, non-black, black, or those with PM or DM. Onset of myositis, however, tended to be in March–April in the 131 non-black patients with antisynthetase autoantibodies (P = 0.03) (Figure 1, top). When these patients were assessed by clinical group, seasonality of onset was prominent in the patients with PM but not in those with DM (P = 0.05 versus P = 0.36) (Table 2). Of interest, black patients with antisynthetase autoantibodies did not show a seasonal pattern of onset (P = 0.4). However, there was a difference in seasonality when antisynthetase-positive patients were assessed by sex; the 51 male patients with antisynthetase antibodies showed significant seasonality in disease onset as compared with the 122 female patients with these antibodies (P = 0.042 and P = 0.7, respectively).

Table 1. Clinical, demographic, and serologic information on the myositis patients in the study*
CategoryNo. of patientsNo. male/ no. femaleNo. white/black/ Hispanic/Asian/other
  • *

    MSA = myositis-specific autoantibodies; tRNA = transfer RNA.

Clinical group   
Serologic group   
  Antisynthetase autoantibodies173122/51124/42/1/4/2
   Histidyl-tRNA (anti–Jo-1)13294/38101/27/1/1/2
  Anti–signal recognition particle2922/79/19/1/0/0
Figure 1.

Significant seasonal associations with month of onset of muscle weakness among selected groups of myositis patients. Seasons of weakness onset are indicated at the outer edge of the circle, and the number of patients in whom onset occurred in any given month is depicted by the length of the shaded cone beside each month (0.7 mm per patient). Non-black patients with antisynthetase antibodies had a significant peak in myositis onset in March (n = 21) and April (n = 17) (top). In contrast, patients without myositis-specific autoantibodies had a significant peak in onset in June (n = 30) and July (n = 34) (bottom) (see Table 2 for details).

Table 2. Seasonal associations with myositis onset in the different myositis groups*
GroupTotal no. of patientsSeasonal associationP
  • *

    MSA = myositis-specific autoantibodies; PM = polymyositis; DM = dermatomyositis.

  • Non-blacks were all races not self-categorized as African American (i.e., white, Hispanic, Asian, and other).

    Clinical groups   
   Clinical groups   
     Clinical groups   
   Clinical groups   
   Clinical groups   

In contrast, among the 252 patients without myositis-specific autoantibodies, onset of myositis tended to be in the summer (June–July) (P = 0.03) (Table 2 and Figure 1, bottom). Among patients without myositis-specific autoantibodies, seasonality could be attributed primarily to patients with DM (P = 0.04) rather than to those with PM (P = 0.4). Of interest, female patients without myositis-specific autoantibodies showed strong seasonality of myositis onset (P = 0.005), but male patients did not (P = 0.9). Patients with anti–histidyl transfer RNA synthetase (anti–Jo-1) autoantibodies did not demonstrate significant seasonality of onset (P = 0.08) in the present study; similar results showing a lack of association with seasonality were obtained with anti–signal recognition particle (anti-SRP) or anti–Mi-2 autoantibodies (P > 0.15 for both), although the numbers of patients with anti-SRP or anti–Mi-2 were smaller.


The myositis syndromes can be divided into homogeneous groups of patients who differ according to genetic background, clinical signs and symptoms, therapeutic responses, and prognoses based on the presence or absence of myositis-specific autoantibodies (34). Herein we present evidence that a seasonal pattern of disease onset is prominent in myositis patients when defined by the presence or absence of certain myositis-specific autoantibodies, and we show that some of these differences are further related to sex, ethnicity, or clinical features.

Our initial report indicated a nonrandom seasonal onset of muscle weakness in a small number of patients with anti–Jo-1 or anti-SRP autoantibodies who were evaluated at a single institution (31). Our current independent, multicenter study of a larger number of subjects showed, as before, a seasonal association, with a peak in March–April, in patients with anti–Jo-1, but this association did not reach significance (P = 0.08). A stronger association was seen, however, in non-black men with antisynthetase autoantibodies. In this study, patients with anti-SRP autoantibodies did not show a significant seasonal onset in the fall; however, a strong association of myositis onset in the summer was found in female patients without myositis-specific autoantibodies. The discrepancy between the results in this and prior studies might be due to a number of factors, including the different populations being studied.

Seasonal associations in adult PM or DM have not been studied extensively, and the few investigations published often have yielded different findings. Onset of myositis in the spring was found in 12 (45%) of 26 PM/DM patients in Greece (32). Although a similar onset in spring was reported in 55% of children in a study of 124 childhood-onset cases from 1 US study (36), another US study demonstrated DM onset between June and January in 11 of 12 children (4). This latter finding was supported by an association of onset of juvenile DM with the summer and winter months in 48 cases from the United Kingdom and Ireland (37). In the only study to assess seasonal associations with increases in myositis disease activity, relapses involving weakness, both muscular and cutaneous, occurred more often during the summer and spring in Australian patients with DM, but no significant seasonal trends were found in PM patients (30).

The March–April peak in myositis onset in the non-black male patients with antisynthetase autoantibodies and the strong association of myositis onset in the summer that was found in our female patients without myositis-specific autoantibodies might be due to a greater exposure to, or to the greater effects of, selected spring–summer environmental agents in these subjects compared with other groups. Possible environmental factors that could be consistent with such seasonal onsets include ultraviolet radiation or infectious agents, for which growing evidence implies a possible association with myositis (1, 26, 27). It is possible that men and women, as well as different ethnic groups, either have different environmental exposures or may respond differently to the same exposures at different times of the year. Furthermore, naturally occurring daily and seasonal cycles of T lymphocytes, natural killer T cells, and other immune functions have been reported and may have evolved to allow individuals to anticipate and to cope with seasonal variations in immunologically challenging conditions (38).

Although attempts were made to avoid possible confounding in all aspects of the study, there are limitations to the approaches used in this investigation. For example, given the rarity of myositis, we studied consecutive patients from a number of referral centers in different geographic regions and in whom disease onset spanned several decades. Although this was necessary to obtain adequate power for our study, it also allowed for possible differential environmental effects from different locations and in different years. Furthermore, because our study was retrospective, we did not have complete data on possible environmental risk factors, including infectious or ultraviolet exposures, that could help explain the observed seasonal associations. Another limitation of this study is that, despite our efforts, we still had relatively small numbers of subjects in the subgroups, which limited our power to address second-order questions, such as whether the seasonal patterns were different in different regions of the country.

These findings of seasonal patterns of myositis onset in different groups, in conjunction with other data, suggest that a variety of environmental agents, acting upon different genetic backgrounds, result in distinct immune responses and clinical syndromes in IIM. Given the increasing evidence of heterogeneity among many autoimmune disorders as they are defined today, and the possibility of varying gene–environment interactions in different subsets of subjects (1, 39), more homogeneous disease groups that are based on clinicopathologic features, genetics, immune responses, ethnicity, and sex should be defined in future attempts to decipher the pathogeneses of autoimmune disorders.


We thank Drs. Gabor Illei and Steven Bauer for their useful comments on the manuscript, and Dr. Richard Leff for his vision in initiating this line of research.