1. Top of page
  2. Abstract
  8. Acknowledgements


Spontaneous pneumomediastinum is a rare complication of dermatomyositis (DM) and polymyositis (PM). The aim of this study was to characterize this complication and determine its prognostic factors.


We retrospectively collected a multicenter series of PM/DM cases complicated by pneumomediastinum. We analyzed all published cases and combined those that were exploitable with ours for an investigation of the factors associated with poor survival.


We collected 11 PM/DM cases complicated by interstitial lung disease and pneumomediastinum. Five of the 9 DM patients had clinically amyopathic DM without muscle weakness and high serum creatine kinase levels. The outcome was favorable in 7 of these patients and 6 had no sequelae. In total, ∼25% of our patients of the 21 analyzable cases studied died within 1 month. With a median followup of 240 days, the cumulative estimated Kaplan-Meier survival rate was 64% at 1 year and 55% at 2 years. Poor survival was associated with absence of muscle weakness (P = 0.02), initial low vital capacity (P = 0.006), and initial low carbon monoxide diffusion capacity (P = 0.04).


In this first large series of patients with connective tissue disease complicated by pneumomediastinum to be reported, most patients had DM and half amyopathic DM, as in previous reports. Pneumomediastinum may occur before DM diagnosis and may thus reveal DM with minimal or no muscle involvement. Death was associated with an absence of muscle weakness and severe pulmonary involvement before the onset of pneumomediastinum. Corticosteroids and immunosuppressive therapy can result in complete recovery, as in half our cases.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Dermatomyositis (DM) and polymyositis (PM) are systemic inflammatory disorders affecting skeletal muscles and other organs, including the lungs in particular. In 10–22% of cases, pulmonary signs are related to aspiration pneumonia or respiratory muscle weakness (1, 2). Interstitial lung disease (ILD) is present in 10–43% of cases (1–4) and is associated with a poor outcome (3). The mechanisms of deterioration involved in severe forms of ILD in PM/DM patients have been little investigated.

Pneumomediastinum is the presence of free air around mediastinal structures. It mostly occurs after injury, acute mediastinitis, or assisted ventilation. However, pneumomediastinum may also arise spontaneously, in the absence of any obvious cause. Spontaneous pneumomediastinum has been reported as a rare complication of various types of ILD, including interstitial pneumonia associated with connective tissue diseases. Inflammatory myositis is the most frequent connective tissue disease associated with pneumomediastinum. Forty-two cases have been reported in journals published in English or French, to our knowledge (5–26). Pneumomediastinum is often fatal in this setting (8). Among recent series of PM/DM patients with ILD (2–4, 15, 27–31), only 3 reports mentioned pneumomediastinum (3, 4, 15). A prevalence of 8.3% has been reported for pneumomediastinum among patients with PM/DM (15). In a recent retrospective cohort study of 70 DM/PM patients with ILD, pneumomediastinum or pneumothorax was reported in 8.6% patients, and the 6 patients with pneumomediastinum or pneumothorax died (4). The authors concluded that spontaneous pneumomediastinum or pneumothorax is a severe complication in DM/PM patients. However, the disease mechanisms, outcome, and prognostic factors of pneumomediastinum associated with connective tissue diseases have not been evaluated.

We report a series of 11 new cases of PM/DM-associated pneumomediastinum and an exhaustive literature review, with the aim of determining the clinical, radiographic, and histologic characteristics of ILD in these patients, and variables predictive of the outcome of pneumomediastinum in patients with PM/DM.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Case recruitment.

A retrospective national multicenter study was initiated by the Club Rhumatismes and Inflammation (CRI), a section of the French Society of Rheumatology. A nationwide survey was carried out to identify cases of pneumomediastinum associated with PM/DM in French rheumatology and internal medicine departments. Participating physicians were asked to complete a structured questionnaire, including details of the characteristics of the myositis and respiratory disease.

Diagnostic criteria.

The diagnostic criteria used for PM/DM were those of Bohan and Peter, including symmetric muscle weakness, high serum muscle enzyme levels, myopathic changes on electromyography (EMG), typical histologic findings on muscle biopsy, and characteristics of dermatologic manifestations (heliotrope rash, facial erythema and edema, periungual erythema, and Göttron's papules) (32). Clinically amyopathic DM was diagnosed on the basis of characteristic cutaneous signs of DM with no clinical or biologic evidence of muscle disease (33, 34). Cases of iatrogenic or accidental pneumomediastinum were not included.

Data collection.

A database was created that included clinical history, physical examination, laboratory test results, radiographic findings, pulmonary and muscle biopsy findings, and EMG.

Pulmonary function tests.

Pulmonary function tests (PFTs) including vital capacity (VC), total lung capacity (TLC), and carbon monoxide diffusion capacity across the lung were measured by spirometry, using the single-breath method. Data were expressed as percentages of predicted values. Arterial oxygen tension (PaO2) was determined at rest.

High-resolution computed tomography of the lungs.

Abnormalities consistent with ILD were evaluated on high-resolution computed tomography (HRCT) of the lungs. These abnormalities included lung parenchymal nodules, linear opacities, irregular interfaces between the peripheral pleura and aerated lung parenchyma, ground-glass opacities, honeycombing, and traction bronchiectasis.

Bronchoalveolar lavage.

PM/DM patients underwent bronchoalveolar lavage (BAL) if the treating physician considered this procedure to be indicated. Patients were considered to have lymphocytosis if lymphocyte counts exceeded 15%. Neutrophils were considered to be present in excess if they accounted for more than 4% of the count.

Lung biopsy.

When considered necessary, a lung biopsy was performed. The specimen obtained was examined by a pathologist with expertise in ILD. Histologic results were used to obtain a diagnosis according to the American Thoracic/European Respiratory Society international multidisciplinary consensus classification of idiopathic interstitial pneumonia (35).

Three of our cases have previously been reported as a letter or case report in French medical journals (36–38).

ILD courses.

The course of disease in each patient with pneumomediastinum was assessed based on clinical signs, radiographic images, and PFT abnormalities at the most recent followup visit. Remission was defined as the complete resolution of pulmonary symptoms, with the normalization of PFT abnormalities (2). Improvement was defined as an improvement in pulmonary signs and pulmonary PFT on treatment, but without a return to normal values. Deterioration was defined as a worsening of PFT results or radiographic assessments. Both survival and status at the last visit were determined.

Literature analysis.

Published cases of pneumomediastinum associated with connective tissue diseases were identified by a computerized Medline search of articles published from 1966 to 2007. The key words used, both alone and in association with each other, were dermatomyositis, polymyositis, amyopathic dermatomyositis, connective tissue disease, pneumomediastinum, interstitial lung disease. Only publications in English or French were selected for review. The references of the articles obtained were examined to identify additional reports. Only cases sufficiently detailed for individual analysis were included.

Survival and statistical analysis of prognostic factors.

Continuous data were expressed as the mean ± SD or median (interquartile range). Categorical data were expressed as numbers per group (percentage). We analyzed the characteristics of respiratory and muscular disease in our patients and in the published cases identified. Cases were considered for inclusion in the prognostic analysis only if the duration of followup was determined and status provided for the most recent examination. For patients for whom no duration of followup was indicated, the authors were contacted up to 3 times by e-mail and mail to obtain additional information. Once our data and those of exploitable published cases had been compiled, survival analysis was carried out to identify prognostic factors for all collected data. The Kaplan-Meier method was used to compute survival distributions, and differences were assessed using the log rank test. A P value less than 0.05 was considered significant. All statistical evaluations were performed with SAS software 9.1 (SAS Institute, Cary, NC).


  1. Top of page
  2. Abstract
  8. Acknowledgements

Characteristics of the study patients.

We collected 11 cases of pneumomediastinum associated with DM (8 cases), PM (2 cases), and sclerodermatomyositis (1 case). Seven of the patients were female and 4 were male, and the mean age of the patients was 42 years (range 21–75). The clinical, pulmonary, and muscular characteristics of our 11 patients are summarized in Tables 1 and 2. All of our patients had ILD. Six of the 11 patients in our series had typical muscular findings, including muscular pain and proximal weakness, whereas 5 presented no muscle-associated clinical or biologic abnormality. Patients 4 and 7 had isolated EMG myopathic changes and patient 8 had isolated inflammatory muscular changes at biopsy examination despite no clinical or biologic evidence of muscle involvement. Thus, 5 of the 9 DM patients (55.6%) had clinically amyopathic DM. No associated cancer was found. Initial clinical manifestations were polyarthritis in 7 cases, muscular pain in 3 cases, and dyspnea in 1 case. Myositis was diagnosed a mean of 5 months after the first symptoms of the connective tissue disease (range 1.5–14 months). Myalgia was present at the onset of the disease in 6 cases, appeared secondarily (4 months after the first symptoms) in 1 case, and was absent in 4 cases. Muscular weakness was noted in 6 cases. Dyspnea was present initially in 2 cases, appeared secondarily in 7 cases, and was absent in 2 cases. Weight loss was observed in 6 of the 11 patients (mean 6.8 kg, range 0–22), arthralgia in 7 patients, and arthritis in 4 patients. Cutaneous findings included Göttron's papules in 10 cases, periorbital rash in 5, erythema over the metacarpophalangeal joints in 5, and cutaneous vasculitis of the lower limbs in 2. Raynaud's phenomenon was present in 6 patients.

Table 1. Characteristics of our 11 patients with polymyositis (PM)/dermatomyositis (DM) complicated by interstitial lung disease and pneumomediastinum (PNM)*
Patient/sexDiagnosisMuscle weaknessCreatine kinase (normal)HRCT chest findingsPaO2 at rest before PNM onset, mm HgVital capacity before PNM onset, %Total lung capacity before PNM onset, %Carbon monoxide diffusion capacity before PNM onset, %BAL, cells/mm3Lung histologic pattern (ATS/ERS)Time between first symptoms of myositis and PNM, monthsAssociated pneumo-thoraxProgression
  • *

    HRCT = high-resolution computed tomography; PaO2 = arterial oxygen tension; BAL = bronchoalveolar lavage; ATS/ERS = American Thoracic Society/European Respiratory Society international multidisciplinary consensus classification of the idiopathic interstitial pneumonia (35); NSIP = nonspecific interstitial pneumonia; DAD = diffuse alveolar damage; OP = organizing pneumonia.

1/MDMNGround glass opacities77796778370,000; 34% lymphocytesNSIP, DAD, and OP1.5Resolution
2/FDMNGround glass opacities, subpleural blebs82536498,000; 22% lymphocytesRespiratory bronchiolitis4+Lost to followup
3/MPM+23NGround glass opacities, honeycomb cysts, paracardiac and subpleural blebs80476937Normal5+Died 8 months after PNM
4/MDM-NGround glass opacities, paracardiac blebs741001007255,000; 39% lymphocytes; 27% PMN6Resolution
5/FDM+5NDiffuse opacities predominant in the basal area8080757029Resolution
6/FDM+5NDiffuse opacities predominant in the basal area958575853Resolution
7/FDMNGround glass opacities, paracardiac blebs65618162150,000; 43% lymphocytes4Died 9 months after PNM
8/FDMNGround glass opacities, paracardiac blebs, honeycomb cysts643883380,000; 23% lymphocytesUsual interstitial pneumonia7+Died 2 months after PNM
9/FPM+17NGround glass opacities, paracardiac blebs76827280Normal18Resolution
10/MDM+NHoneycomb cysts, paracardiac blebs60657073280,000; 78% lymphocytes19+Resolution
11/FScleroDM+1.5NGround glass opacities, pericardial effusion9054576410Improvement
Table 2. Clinical characteristics of our 11 patients and of 42 analyzable patients with polymyositis/dermatomyositis complicated by PNM from published studies*
 Our seriesPublished cases (5–26)Total
No.MeanNo. of exploitable casesMeanNo. of exploitable casesMean
  • *

    Including 16 patients published in a Japanese publication for whom data were summarized in English by Matsuda et al (8). See Table 1 for definitions.

Age, years114242395340.5
Sex, male/female114/74222/204326/27
Cutaneous vasculitis (%)2/111810/263812/3732
Arthritis (%)7/116310/234317/3450
Muscle weakness (%)6/115412/235218/3453
Normal level of creatine kinase (%)5/1145.517/345022/4548
Erythrocyte sedimentation rate (mm/hour)11619442052.5
Pneumothorax (%)4/11365/21249/3228
Vital capacity before PNM onset (% predicted)11676611764
Total lung capacity before PNM onset (% predicted)10754661472
Carbon monoxide diffusion capacity before PNM onset (% predicted)10689601964
PaO2 before PNM onset (mm Hg)11769692072.5
Interval between first symptom of myositis and PNM (months)1110178289
Corticosteroids (%)11/1110023/2310034/34100
High-dose corticosteroid infusions (%)7/1163.67/2330.514/3441
Immunosuppressive agents (%)8/1172.722/4153.630/5257.6
Mortality rates (%)3/112718/424321/5339.5

It should be stressed that pneumomediastinum occurred before diagnosis of the connective tissue disease in 2 patients with amyopathic DM (patients 1 and 2). The mean time lag between the onset of the first symptoms of myositis and pneumomediastinum was 10 months (range 1.5–29). A triggering factor was suggested by the clinical history in 5 cases: a forced cough (2 cases), PFT (1 case), infectious pneumonitis (1 case), and a forced walk (1 case). In 3 patients, pneumomediastinum was discovered on a systematic chest radiograph or HRCT of the lungs (patients 6, 9, and 11). All other patients presented an association of increased dyspnea, chest pain, and subcutaneous emphysema. The predominant signs and symptoms at onset of pneumomediastinum were increased dyspnea (4 cases), subcutaneous emphysema (2 cases), cervical pain (1 case), and cough (1 case). Pneumomediastinum occurred before treatment in 2 cases and, in the other cases, a mean of 6 months (range 2 days to 22 months) after treatment initiation. Associated pneumothorax occurred in 4 patients (Table 1) and was unilateral in 2 cases and bilateral in 1 case. Pneumomediastinum relapsed 3–4 times after a mean of 2.5 months (range 1–6) in 3 patients.

Biologic findings.

Erythrocyte sedimentation rate (ESR) was high in 78% of cases, with a mean ESR of 61 mm/hour (range 45–90). Serum creatine kinase concentration was normal in 6 cases and high in 5 cases, at a mean of 10 times the normal level (10N, range 1.5–23) (Tables 1 and 2). Antinuclear antibodies were present in 4 patients. Anti–double-stranded DNA, anti-Ro/SSa, and anti-Sm antibodies were detected in 1 patient (patient 11); anti-RNP antibodies were present in another (patient 9). Two patients had antinuclear antibodies without detected specificity. None of these patients had clinical manifestations of lupus, Sjögren's syndrome, or mixed connective tissue disease. Antisynthetase antibody was present in 1 case and displayed anti–lysyl/aspartyl specificity.

Muscle investigations.

EMG showed myopathic changes at the time of diagnosis in 7 cases, with no such changes observed in 3 cases (patients 1, 2, and 8). In 1 case, myopathic changes were observed later in the course of the disease. Muscle biopsy was performed in 8 patients, and showed typical inflammatory changes in 5 (patients 5, 6, 8, 9, and 10) and no change or nonspecific changes in 3 (patients 1, 2, and 4).

Pulmonary investigations.

Before the onset of pneumomediastinum, arterial blood gas measurements in room air showed an average PaO2 of 76 mm Hg (range 60–95). PFT showed a restrictive defect in 10 cases, with a mean VC of 67% (range 38–100%) and TLC of 75% (range 57–100%) (Tables 1 and 2). A decrease in carbon monoxide diffusion capacity was observed in 7 cases (range 37–85%) and isolated in 1 case. BAL was performed in 8 patients (Table 1). The differential cell pattern was lymphocytic in 5 patients and mixed (lymphocytic and neutrophilic) in 1 patient. It was normal in 2 cases. Chest radiographs and HRCT showed diffuse infiltrative opacities predominantly in the basal area in all cases (Table 1). HRCT showed a pattern typical of usual interstitial pneumonitis (UIP), with basal or subpleural honeycombing and/or traction bronchectasis in only 3 cases. However, the marked lymphocytosis observed on BAL (patient 10) in 1 of these cases is not typical of UIP (35), and this patient had a favorable outcome with clinical and functional resolution, consistent with the course of nonspecific interstitial pneumonia (NSIP). In most cases, HRCT findings were suggestive of NSIP, with predominantly bilateral patches of ground-glass opacities without honeycombing and consolidation. However, an alternative histologic pattern was found on biopsy in 3 of these 8 cases. According to the American Thoracic Society/European Respiratory Society international multidisciplinary classification of idiopathic interstitial pneumonia (35), one of these biopsy samples was classified as UIP and another as respiratory bronchiolitis pattern. The third case was an unclassifiable interstitial pneumonia with a mixed pattern of NSIP, diffuse alveolar damage (DAD), and organizing pneumonia. Most of the other cases had a favorable course, compatible with NSIP. Paracardiac or subpleural blebs were found in 7 cases.

Treatment and outcome.

All of our patients received oral corticosteroids (prednisone 1 mg/kg/day) and 7 received initial high-dose methylprednisolone infusions. Gammaglobulin infusions were administered to 5 patients. Intravenous cyclophosphamide was administered (700 mg/m2) monthly in 5 cases for a mean duration of 8 months (range 2–18). Azathioprine (2 mg/kg/day) was administered to 4 patients for a mean duration of 17 months (range 2–36). Methotrexate and cyclosporin A were administered in 1 patient each. Pneumomediastinum was treated expectantly in all cases. Therapy with thoracic drainage followed by pleurectomy was performed in 2 cases of associated pneumothorax, whereas 2 other patients were treated expectantly with bed rest. With a median followup of 2 years (range 14 months to 8 years) after the onset of pneumomediastinum, outcome was favorable in 7 of our patients, 3 died, and 1 was lost to followup. Six patients had no sequelae (Table 1). These patients corresponded globally to those with an initially less severe respiratory defect on PFT. With a median followup of 5.4 years (range 6 months to 8 years), subsequent PFTs were entirely normal in 4 cases and normal HRCT results for the lungs were obtained during followup in 2 cases. One patient displayed an improvement (patient 11). Three patients died of respiratory failure, and 2 died 8 and 9 months, respectively, after the onset of pneumomediastinum (Table 1). These patients died a mean of 8 months (range 4–11) after ILD diagnosis. One of them died of a respiratory failure with suffocating pneumothorax in the context of Pseudomas aeruginosa pulmonary infection.

Literature review.

Including our 11 cases, to our knowledge, 62 cases of connective tissue diseases associated with pneumomediastinum have been reported in publications in English and French (Table 3). Forty-nine (79.0%) of these patients had DM and 3 (4.8%) had PM (20). Of the 42 exploitable DM cases identified from previous studies (5–26), including our 8 cases, 22 (52%) had amyopathic DM. The other diagnoses were lupus (3 cases) (39–41), rheumatoid arthritis (3 cases) (42–44), and systemic sclerosis (3 cases) (45–47). In addition to our 11 cases, the literature review provided us with 42 other cases of PM/DM for which analyses of clinical and treatment characteristics were possible, including 16 case reports published in Japanese but summarized in English by Matsuda et al (8). Table 2 presents the available data for these 42 cases of PM/DM and our 11 cases.

Table 3. Connective tissue diseases associated with pneumomediastinum in our 11 cases and in published cases*
 Our series (n = 11)Published cases (5–26, 39–47) (n = 51)Total, no. (%) (n = 62)
  • *

    Including 16 case reports published in Japanese but reviewed by Matsuda et al (8).

Dermatomyositis84149 (79.0)
Polymyositis213 (4.8)
Sclerodermatomyositis101 (1.6)
Systemic lupus erythematosus033 (4.8)
Rheumatoid arthritis033 (4.8)
Systemic sclerosis033 (4.8)

Survival analysis.

For the survival analysis, we included 10 of our 11 cases for whom information about followup was available, 20 published cases for whom followup was detailed (5–23), and 1 case for whom information about followup was obtained by contacting the author (25). In total, 31 patients were included in this analysis, 11 of whom had died and 20 of whom had favorable outcome. The overall mortality rate was 34.4%, with ∼25% of patients dying during the first month of respiratory distress. With a median followup of 240 days, the cumulative estimated Kaplan-Meier survival rate was 64% at 1 year and 55% at 2 years. The survival curve for these patients is shown in Figure 1.

thumbnail image

Figure 1. Kaplan-Meier probability of survival for 31 analyzable polymyositis/dermatomyositis patients with pneumomediastinum (5–23), including 10 of our cases.

Download figure to PowerPoint

Analysis of prognosis.

Statistical analysis was carried out on the 31 exploitable published cases, including our observations, to search for prognostic factors associated with a poor outcome (Table 4). The factors associated with poor survival were an absence of muscle weakness (P = 0.02) and an initial decrease in vital capacity (P = 0.006) or carbon monoxide diffusion capacity (P = 0.04) before the onset of pneumomediastinum (Figure 2). The use of high-dose methylprednisolone infusions tended to be associated with a poor outcome (P = 0.09). The administration of immunosuppressive drugs was of no prognostic value. Indeed, 11 of the 20 patients (55.0%) in the survival group and 5 of the 11 patients (45.5%) who died had received immunosuppressive treatment before or at the onset of pneumomediastinum (P = 0.59) (Table 4).

Table 4. Comparison between patients who died and survivors in 31 analyzable PM/DM cases with PNM from previous publications (5–23), including 10 of our cases*
 Fatal outcome (n = 11)Survivors (n = 20)P (log rank test)
  • *

    Values are the number; median (interquartile range) or number (percentage) unless otherwise indicated. CTD = connective tissue disease; see Table 1 for additional definitions.

Age, years11; 41.0 (23.0–57.0)20; 38.0 (26.5–45.5)0.8347
Sex, male/female5/612/80.4653
Cutaneous vasculitis1 (9.1)7 (35.0)0.3764
Arthritis6 (54.6)9 (45.0)0.6416
Muscle weakness3 (27.3)15 (75.0)0.0185
Creatine kinase levels (x times normal values)11; 1.0 (1.0–1.5)20; 1.5 (1.0–1.5)0.9121
PaO2 before PNM onset (mm Hg)8; 64.5 (62.0–80.0)11; 77.0 (74.0–90.0)0.3094
Pneumothorax5 (45.5)3 (15.0)0.1740
Erythrocyte sedimentation rate, mm/hour5; 45.0 (43.0–48.0)10; 37.0 (14.0–51.0)0.2923
C-reactive protein level, mg/liter5; 1.1 (5.0–17.0)8; 0.6 (5.5–29.0)0.5510
Vital capacity before PNM onset (% predicted)5; 47.0 (45.0–60.0)10; 76.0 (65.0–82.0)0.0060
Carbon monoxide diffusion capacity before PNM onset (% predicted)3; 40.0 (37.0–62.0)14; 70.0 (60.0–78.0)0.0356
Time between first CTD symptoms and diagnosis of PM/DM (days)11; 90.0 (30.0–120.0)14; 90.0 (30.0–210.0)0.2264
Time between first CTD symptoms and PNM (days)11; 150.0 (72.0–180.0)13; 240.0 (105.0–450.0)0.1903
Time between first treatment of CTD and PNM (months)11; 51.0 (30.0–120.0)19; 90.0 (2.0–210.0)0.4051
High-dose corticosteroid infusions7 (63.6)5 (25.0)0.0940
Immunosuppressants before or at PNM onset5 (45.5)11 (55.0)0.5949
thumbnail image

Figure 2. Kaplan-Meier probability of survival for 31 analyzable polymyositis/dermatomyositis patients with pneumomediastinum (5–23), including 10 of our cases, as a function of A, the presence or absence of muscle weakness, B, initial vital capacity, or C, initial carbon monoxide diffusion capacity.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  8. Acknowledgements

This is the first large series of patients with connective tissue disease complicated by pneumomediastinum to be reported. As in our cases, 80% of previously reported patients had DM, and of these half had amyopathic DM. Two patients in our series had pneumomediastinum before amyopathic DM diagnosis and the absence of muscle involvement is likely to delay diagnosis in most cases. Thus, in a clinical setting of connective tissue disease, pneumomediastinum should lead to consideration of a diagnosis of DM, even in the absence of typical muscle involvement.

By contrast, with the high prevalence of amyopathic DM in patients with PM/DM associated with pneumomediastinum, this clinical presentation generally accounts for 5–11% of DM cases. Amyopathic DM is a rare but distinct form of DM. Patients display the hallmark cutaneous signs of DM, but without skeletal muscle involvement. ILD has been reported in 13% of amyopathic DM patients (48). Rapidly progressive ILD unresponsive to high-dose corticosteroids and poor prognosis have been reported in these DM patients with typical cutaneous manifestations but little, if any, muscle involvement (4, 14, 29). The mechanisms of respiratory degradation have not been evaluated.

We collected 11 cases of PM/DM-associated pneumomediastinum from 9 rheumatology or internal medicine departments out of a total of almost 500 cases of myositis. The prevalence of pneumomediastinum in myositis may therefore be estimated at ∼2.2%. DM and PM are twice as frequent in women as in men, but myositis complicated by pneumomediastinum seems to be more frequent in men. Indeed, the sex ratio (F:M) for the 42 exploitable cases, including our 11 observations, was 1.04. All our patients with pneumomediastinum had underlying ILD and, in most cases, pneumomediastinum occurred within a year of the initial onset of signs and symptoms of PM/DM. This finding is consistent with previous data for PM/DM and associated ILD, showing that, in ∼40% of cases, ILD preceded PM/DM, and that the 2 conditions occurred simultaneously in another 40% of cases (2).

Patient with ILD, such as idiopathic pulmonary fibrosis, are at risk of developing pneumomediastinum. Retrospective studies have estimated that 5–14.7% of patients with idiopathic pulmonary fibrosis have pneumomediastinum detectable by a systematic computed tomography (CT) scan (49, 50). In the absence of pulmonary disease, spontaneous mediastinal emphysema is generally a benign condition that can be treated expectantly. By considering our cases and other exploitable cases from previous publications, we found that the overall mortality rate was high, with one-quarter of patients dying during the first month. The low cumulative estimated Kaplan-Meier survival rates of 64% at 1 year and 55% at 2 years confirm the poor prognosis of pneumomediastinum in patients with PM/DM (3, 8). In our series, poor survival was associated with an absence of muscle weakness, an initially low VC, and an initially low carbon monoxide diffusion capacity. Decreases in VC and carbon monoxide diffusion capacity before the onset of pneumomediastinum reflect more severe interstitial pneumonitis, with the occurrence of pneumomediastinum worsening the patient's already poor respiratory condition. These factors associated with a poor prognosis are consistent with those previously reported in patients with ILD-associated PM/DM: DM (4, 31), amyopathic DM (3, 4), an initial forced VC ≤60% (3), a Hamman-Rich–like presentation (2, 3), and a neutrophilic BAL (2). Despite the association of pneumomediastinum with a high rate of mortality within the first few weeks, the survivors in our series had a good prognosis. ILD resolved in almost half of survivors, with complete remission.

The subclassification of ILD on the basis of lung histology may be a better predictor of survival than radiograph or clinical presentation (2, 27), but the relationship between histologic pattern and prognosis is less clear for patients with autoimmune diseases than for those with idiopathic ILD. The NSIP pattern observed most frequently in our cases is consistent with previous reports showing NSIP to be the histologic pattern of ILD most frequently encountered in PM/DM patients (28, 30, 31). Patients with a PM/DM-associated NSIP pattern follow a course similar to patients with idiopathic NSIP, who have a survival rate of 60% at 5 years (28). PM/DM patients with chronic organized pneumonia have a more favorable prognosis (21, 27) than patients with DAD and UIP, which have been identified as the histologic features most strongly associated with rapidly progressive ILD (3, 21, 27, 29). Unlike the NSIP pattern, these patterns do not seem to be particularly associated with PM/DM-associated pneumomediastinum.

In our analysis of 52 patients with PM/DM complicated by ILD-associated pneumomediastinum, including our cases, autoantibodies to aminoacyl–transfer RNA synthetases were found in only 2 cases, and in neither case were these antibodies directed against PL-7. In PM/DM patients, anti–Jo-1 antibodies are found in approximately 15–25% of patients, and even more frequently in cases with ILD, whereas other types of antisynthetase antibodies are found in only ∼3% of patients (51). The usual absence of myositis-specific antibodies in amyopathic DM patients with ILD contrasts with the situation in classic DM with lung disease. Yamasaki et al recently reported a higher (17%) prevalence of anti-threonyl (PL-7) antibodies in Japanese PM/DM patients with milder muscle involvement than in patients with anti–Jo-1 antibodies, but only 16% of these patients had amyopathic DM (51). Further research should focus on the search for associated antibodies in amyopathic DM and PM/DM ILD patients with pneumomediastinum, to facilitate the earlier diagnosis of this severe entity.

Pneumomediastinum usually results from rupture of the alveoli due to a marked increase in intraalveolar pressure. In severe ILD, subpleural or paracardiac blebs may form due to the distortion of lung architecture, and the rupture of paracardiac blebs may also lead to air leakage in the mediastinum. Six of the 11 patients in our series had paracardiac blebs on CT scan, whereas very few such patients have been reported in previous studies (21). However, these abnormalities may have been underreported. The development of pneumothorax, pneumomediastinum, or both probably depends on the location of the rupture (peripheral for pneumothorax and adjacent to vessels for pneumomediastinum) with the air spreading in the peribronchial and perivascular spaces in the mediastinum for the latter (7). This common mechanism may account for the coexistence of pneumomediastinum and pneumothorax in 4 of our patients, as also reported in patients with idiopathic pulmonary fibrosis. An increase in intraalveolar pressure may have been a triggering factor in 5 of our cases. The 3 or 4 relapses of pneumomediastinum encountered in 3 of our patients probably resulted from additional alveolar or paracardiac bleb rupture in a context of severe or progressive ILD. Alternatively, DM-associated vascular disease may cause necrosis of the bronchial wall, leading to pneumomediastinum (15). Kono et al found an association between cutaneous vasculitis and pneumomediastinum (15). However, in our series and the published cases considered in this analysis, cutaneous vasculitis was mentioned for only 12 (32%) of the 37 patients that could be studied. It has also been suggested that weakening of the interstitial tissue of the lungs by corticosteroids might also favor the development of pneumomediastinum (14). However, in 2 of our patients and 1 other reported case (15), pneumomediastinum occurred before the administration of corticosteroids, which represents the first-line treatment for inflammatory myositis.

The optimal treatment for patients with ILD associated with PM/DM remains to be determined. Most clinicians treat symptomatic patients with corticosteroids, with the addition of azathioprine, methotrexate, cyclosporine A, or cyclophosphamide in cases of fulminant respiratory disease or failure to respond to corticosteroids (21). High-dose pulses of intravenous methylprednisolone have also been used. Intravenous Ig therapy, which may be beneficial in patients with severe myositis, has not been shown to be effective against ILD.

All of the patients in our series and in published cases were treated with corticosteroids. We cannot entirely rule out deleterious effects of high-dose pulses of corticosteroids, but patients with more severe or progressive ILD were offered such treatment in our study, and the observed trend toward lower levels of survival may reflect only the presence of more severe and/or progressive ILD in the patients treated. No significant effect of immunosuppressive drugs was demonstrated in this study, probably due to the limited number of patients reported to date and available for analysis. However, such treatment seems likely to be beneficial.

Our study has several limitations. The flaws inherent to retrospective studies include a lack of exhaustivity in record keeping and data acquisition. However, the incidence of pneumomediastinum in PM/DM is low, and this is the first series to be studied extensively. The small number of patients included in our study may also be seen as a limitation. However, this study includes more than one-quarter of the previously reported cases of PM/DM-associated pneumomediastinum, all of which were included in our analysis. This is also the first study to review all published cases of connective tissue diseases complicated by pneumomediastinum. The number of deaths, even with the compilation of previous published cases, was too small for multivariate regression analysis. However, we believe that the results of our prognostic study are nonetheless valid and that such analysis would not change the main conclusions.

In summary, pneumomediastinum is a rare and severe complication of ILD in myositis or, more rarely, other connective tissue diseases. It occurs more frequently in patients with amyopathic DM and may occur before diagnosis of this condition. The rapidly progressive course of the pneumopathy probably favors the rupture of alveoli or of constituted paracardiac blebs. An absence of muscle weakness and initial severity of the ILD are the major factors associated with a poor prognosis. Corticosteroids and immunosuppressive treatment should be introduced rapidly, particularly if factors associated with a poor prognosis are present, and may make it possible to achieve a favorable outcome without sequelae, as observed in half of our cases.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Dr. Puéchal had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Le Goff, Sibilia, Ravaud, Puéchal.

Acquisition of data. Le Goff, Chérin, Cantagrel, Gayraud, Hachulla, Laborde, Papo, Sibilia, Zabraniecki, Puéchal.

Analysis and interpretation of data. Le Goff, Chérin, Gayraud, Hachulla, Sibilia, Ravaud, Puéchal.

Manuscript preparation. Le Goff, Chérin, Sibilia, Ravaud, Puéchal.

Statistical analysis. Le Goff, Ravaud, Puéchal.


  1. Top of page
  2. Abstract
  8. Acknowledgements

We thank Dr. Roseli S. Isfer from Sao Paulo, Brazil, and Professor Aidaoui from Constantine, Algeria, for providing additional data for published observations, the CRI of the French Society of Rheumatology for the use of facilities for conducting the survey, and Dr. Frédérique Capron, MD, for reviewing the lung biopsy of patient 1. We also thank Bruno Crestani, MD, PhD, for critically reviewing the manuscript.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  • 1
    Dickey BF, Myers AR. Pulmonary disease in polymyositis/dermatomyositis. Semin Arthritis Rheum 1984; 14: 6076.
  • 2
    Marie I, Hachulla E, Cherin P, Dominique S, Hatron PY, Hellot MF, et al. Interstitial lung disease in polymyositis and dermatomyositis. Arthritis Rheum 2002; 47: 61422.
  • 3
    Kang EH, Lee EB, Shin KC, Im CH, Chung DH, Han SK, et al. Interstitial lung disease in patients with polymyositis, dermatomyositis and amyopathic dermatomyositis. Rheumatology (Oxford) 2005; 44: 12826.
  • 4
    Ye S, Chen XX, Lu XY, Wu MF, Deng Y, Huang WQ, et al. Adult clinically amyopathic dermatomyositis with rapid progressive interstitial lung disease: a retrospective cohort study. Clin Rheumatol 2007; 26: 164754.
  • 5
    Bradley JD. Spontaneous pneumomediastinum in adult dermatomyositis. Ann Rheum Dis 1986; 45: 7802.
  • 6
    Carmody E, McNicholl J, Chadwick G, Bresnihan B, Fitzgerald MX. Prolonged spontaneous pneumomediastinum in adult dermatomyositis [letter]. Ann Rheum Dis 1987; 46: 566.
  • 7
    Cicuttini FM, Fraser KJ. Recurrent pneumomediastinum in adult dermatomyositis. J Rheumatol 1989; 16: 3846.
  • 8
    Matsuda Y, Tomii M, Kashiwazaki S. Fatal pneumomediastinum in dermatomyositis without creatine kinase elevation. Intern Med 1993; 32: 6437.
  • 9
    Ogawa M, Watanabe R, Ueda S, Ohto M, Imai T. Mediastinal emphysema developing after a pulmonary function test in a patient with dermatomyositis. J Rheumatol 1994; 21: 1756.
  • 10
    De Toro Santos FJ, Verea-Hernando H, Montero C, Blanco-Aparicio M, Torres Lanzas J, Pombo Felipe F. Chronic pneumomediastinum and subcutaneous emphysema: association with dermatomyositis. Respiration 1995; 62: 536.
  • 11
    Nagai Y, Ishikawa O, Miyachi Y. Pneumomediastinum and subcutaneous emphysema associated with fatal interstitial pneumonia in dermatomyositis. J Dermatol 1997; 24: 4824.
  • 12
    Santiago MB, Chalhoub M, Pereira ST. Amyopathic dermatomyositis complicated by interstitial pulmonary disease and pneumomediastinum [letter]. J Rheumatol 1998; 25: 20423.
  • 13
    Jansen TL, Barrera P, van Engelen BG, Cox N, Laan RF, van de Putte LB. Dermatomyositis with subclinical myositis and spontaneous pneumomediastinum with pneumothorax: case report and review of the literature. Clin Exp Rheumatol 1998; 16: 7335.
  • 14
    Yamanishi Y, Maeda H, Konishi F, Hiyama K, Yamana S, Ishioka S, et al. Dermatomyositis associated with rapidly progressive fatal interstitial pneumonitis and pneumomediastinum. Scand J Rheumatol 1999; 28: 5861.
  • 15
    Kono H, Inokuma S, Nakayama H, Suzuki M. Pneumomediastinum in dermatomyositis: association with cutaneous vasculopathy. Ann Rheum Dis 2000; 59: 3726.
  • 16
    Korkmaz C, Ozkan R, Akay M, Hakan T. Pneumomediastinum and subcutaneous emphysema associated with dermatomyositis. Rheumatology (Oxford) 2001; 40: 4768.
  • 17
    Barvaux VA, Van Mullem X, Pieters TH, Houssiau FA. Persistent pneumomediastinum and dermatomyositis: a case report and review of the literature. Clin Rheumatol 2001; 20: 35961.
  • 18
    Kuroda T, Morikawa H, Satou T, Tanabe Y, Murakami S, Ito S, et al. A case of dermatomyositis complicated with pneumomediastinum successfully treated with cyclosporin A. Clin Rheumatol 2003; 22: 458.
  • 19
    Saint-Georges F, Mulliez P, Lebas D, Modiano P, Bisch D. Pneumomediastinum complicating the anti-synthetase syndrome. Rev Mal Respir 2005; 22: 10314. In French.
  • 20
    Maruoka H, Honda S, Takeo M, Kitazato H, Hanai N, Ayukawa R, et al. A case of polymyositis complicated with interstitial pneumonitis and pneumomediastinum. Mod Rheumatol 2006; 16: 557.
  • 21
    Powell C, Kendall B, Wernick R, Heffner JE. A 34-year-old man with amyopathic dermatomyositis and rapidly progressive dyspnea with facial swelling: diagnosis. Pneumomediastinum and subcutaneous emphysema secondary to amyopathic dermatomyositis-associated interstitial lung disease. Chest 2007; 132: 17103.
  • 22
    De Souza Neves F, Shinjo SK, Carvalho JF, Levy-Neto M, Borges CT. Spontaneous pneumomediastinum and dermatomyositis may be a not so rare association: report of a case and review of the literature. Clin Rheumatol 2007; 26: 1057.
  • 23
    Terao M, Ozawa K, Inui S, Murota H, Yokomi A, Itami S. A case of dermatomyositis complicated with pneumomediastinum. Mod Rheumatol 2007; 17: 1569.
  • 24
    M'Raihi ML, Dakhlia MS, Chebbi ML, Hamza M, el Gharbi B. Pneumomediastinum in dermatopolymyositis. Rev Mal Respir 1989; 6: 5557. In French.
  • 25
    Isfer RS, Nakao SK, Cernea SS. Dermatomyositis associated with pneumomediastinum and subcutaneous cellular tissue emphysema [letter]. Br J Rheumatol 1994; 33: 1092.
  • 26
    Jang KA, Kim SH, Choi JH, Sung KJ, Moon KC, Koh JK. Subcutaneous emphysema with spontaneous pneumomediastinum and pneumothorax in adult dermatomyositis. J Dermatol 1999; 26: 1257.
  • 27
    Tazelaar HD, Viggiano RW, Pickersgill J, Colby TV. Interstitial lung disease in polymyositis and dermatomyositis: clinical features and prognosis as correlated with histologic findings. Am Rev Respir Dis 1990; 141: 72733.
  • 28
    Douglas WW, Tazelaar HD, Hartman TE, Hartman RP, Decker PA, Schroeder DR, et al. Polymyositis-dermatomyositis-associated interstitial lung disease. Am J Respir Crit Care Med 2001; 164: 11825.
  • 29
    Lee CS, Chen TL, Tzen CY, Lin FJ, Peng MJ, Wu CL, et al. Idiopathic inflammatory myopathy with diffuse alveolar damage. Clin Rheumatol 2002; 21: 3916.
  • 30
    Cottin V, Thivolet-Bejui F, Reynaud-Gaubert M, Cadranel J, Delaval P, Ternamian PJ, et al. Interstitial lung disease in amyopathic dermatomyositis, dermatomyositis and polymyositis. Eur Respir J 2003; 22: 24550.
  • 31
    Fujisawa T, Suda T, Nakamura Y, Enomoto N, Ide K, Toyoshima M, et al. Differences in clinical features and prognosis of interstitial lung diseases between polymyositis and dermatomyositis. J Rheumatol 2005; 32: 5864.
  • 32
    Bohan A, Peter JB. Polymyositis and dermatomyositis (first of two parts). N Engl J Med 1975; 292: 3447.
  • 33
    Pearson CM. Polymyositis and dermatomyositis. In: McCartyDJ, editor. Arthritis and allied conditions. Philadelphia: Lea & Febiger; 1979. p. 74261.
  • 34
    Sontheimer RD. Would a new name hasten the acceptance of amyopathic dermatomyositis (dermatomyositis siné myositis) as a distinctive subset within the idiopathic inflammatory dermatomyopathies spectrum of clinical illness? J Am Acad Dermatol 2002; 46: 62636.
  • 35
    American Thoracic Society and European Respiratory Society. American Thoracic Society/European Respiratory Society international multidisciplinary consensus classification of the idiopathic interstitial pneumonias: this joint statement of the American Thoracic Society (ATS), and the European Respiratory Society (ERS) was adopted by the ATS board of directors, June 2001 and by the ERS Executive Committee, June 2001 [published erratum appears in Am J Respir Crit Care Med 2002;166:426]. Am J Respir Crit Care Med 2002; 165: 277304.
  • 36
    Gayraud M, Lhote F, Valeyre D, Kemeny JL, Khellaf M, Battesti JP. Pneumothorax and pneumomediastinum associated with dermatomyositis [letter]. Ann Med Interne (Paris) 1989; 140: 4901. In French.
  • 37
    Bousquet E, Rolland Y, Bon E, Cantagrel A, Tubery M, Murris M, et al. Pneumomediastinum, a rare complication of dermatomyositis: report of a case. Rev Rhum Engl Ed 1996; 63: 4579.
  • 38
    De Maupeou F, Chiavassa-Gandois H, Giron J, Zabraniecki L, Sans N, Fournie B, et al. Spontaneous pneumomediastinum in dermatomyositis: a pathophysiological approach using imaging. Rev Mal Respir 2003; 20: 9658. In French.
  • 39
    Richards AJ, Talbot IC, Swinson DR, Hamilton EB. Diffuse pulmonary fibrosis and bilateral pneumothoraces in systemic lupus erythematosus. Postgrad Med J 1975; 51: 8515.
  • 40
    Masuda A, Tsushima T, Shizume K, Mochizuki T, Isono K, Demura H, et al. Recurrent pneumothoraces and mediastinal emphysema in systemic lupus erythematosus. J Rheumatol 1990; 17: 5448.
  • 41
    Paira SO, Roverano S. Bilateral pneumothorax and mediastinal emphysema in systemic lupus erythematosus. Clin Rheumatol 1992; 11: 5713.
  • 42
    Patel A, Kesler B, Wise RA. Persistent pneumomediastinum in interstitial fibrosis associated with rheumatoid arthritis: treatment with high-concentration oxygen. Chest 2000; 117: 180913.
  • 43
    Dikensoy O, Bayram N, Bingol A, Filiz A. Bronchiolitis obliterans in a case of juvenile rheumatoid arthritis presented with pneumomediastinum. Respiration 2002; 69: 1002.
  • 44
    Kobayashi T, Satoh K, Ohkawa M, Satoh A. Multiple rheumatoid nodules with rapid thin-walled cavity formation producing pneumothorax. J Thorac Imaging 2005; 20: 479.
  • 45
    Mohammad A, Boon Low T, O'Dwyer D, McElvaney G, Kearns G. Spontaneous pneumo-mediastinum in systemic sclerosis: a case report. Rheumatology (Oxford) 2007; 46: 13767.
  • 46
    Teixeira Moreira Almeida Mdo S, Dias LT, Fernandes SJ, Almeida JV. Spontaneous pneumomediastinum and subcutaneous emphysema in systemic sclerosis. Rheumatol Int 2007; 27: 6757.
  • 47
    Jun JB, Song SY. The development of pneumomediastinum after pulmonary function testing in a patient with systemic sclerosis [letter]. Rheumatol Int 2007; 27: 10978.
  • 48
    Gerami P, Schope JM, McDonald L, Walling HW, Sontheimer RD. A systematic review of adult-onset clinically amyopathic dermatomyositis (dermatomyositis siné myositis): a missing link within the spectrum of the idiopathic inflammatory myopathies. J Am Acad Dermatol 2006; 54: 597613.
  • 49
    Fujiwara T. Pneumomediastinum in pulmonary fibrosis: detection by computed tomography. Chest 1993; 104: 446.
  • 50
    Franquet T, Gimenez A, Torrubia S, Sabate JM, Rodriguez-Arias JM. Spontaneous pneumothorax and pneumomediastinum in IPF. Eur Radiol 2000; 10: 10813.
  • 51
    Yamasaki Y, Yamada H, Nozaki T, Akaogi J, Nichols C, Lyons R, et al. Unusually high frequency of autoantibodies to PL-7 associated with milder muscle disease in Japanese patients with polymyositis/dermatomyositis. Arthritis Rheum 2006; 54: 20049.