Abstract
Background
Spontaneous pneumomediastinum and/or pneumothorax (SPM/P) have been Linked to unfavorable outcome in anti-melanoma differentiation-associated gene 5 positive dermatomyositis (MDA5+DM) patients. This study attempted to investigate the prognostic value of SPM/P and identify the predisposing factors of SPM/P in patients with interstitial lung disease (ILD) complicated with MDA5+DM.
Methods
This study was conducted in a large inception cohort of MDA5+DM-ILD (from 2014 to 2023) at Shanghai Renji Hospital. Baseline demographic data, pulmonary function tests, high-resolution CT imaging, laboratory parameters and survival status within 12 months were collected and compared between survivors and deceased patients, as well as between patients with and without SPM/P. Kaplan–Meier curves were plotted and Cox regression models were applied to identify prognostic factors. Furthermore, logistic regression analysis was employed to determine the predictors of SPM/P development.
Results
In a cohort of 523 MDA5+DM-ILD patients, 92 (17.6%) developed SPM/P. Of those, 63 SPM/P cases occurred within 3 months since ILD onset and had a significantly higher 12-month mortality than those without SPM/P (82.5% versus. 35.4%, p < 0.001). Multivariable Cox regression model identified SPM/P occurred within 3 months as an independent risk factor for survival (HR 1.59, 95% CI 1.09–2.34, P = 0.018). Predisposing factors for SPM/P included severe restrictive ventilation dysfunction, male sex, and CMV viremia.
Conclusions
SPM/P occurred within 3 months since ILD onset was identified as a poor prognostic factor in MDA5+DM-ILD. A refined FVC%-based staging system combining SPM/P was proposed for better risk stratification and to gauge clinical trial design.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12931-025-03347-x.
Keywords: MDA5+ DM-ILD, Spontaneous pneumomediastinum, Pneumothorax, Interstitial lung disease, Dermatomyositis
Introduction
Anti-melanoma differentiation–associated gene 5 dermatomyositis (MDA5+DM) is a specific subtype of idiopathic inflammatory myopathy characterized by prominent interstitial lung disease (ILD) [1]. MDA5+DM-ILD frequently manifests as a rapidly progressive clinical course within 3 months, inducing a high six-month mortality reaching up to 50% [2]. In alignment with assessments for progressive pulmonary fibrosis (PPF), two domains have been implemented in the evaluation and risk stratification of MDA5+DM-ILD, i.e., pulmonary function tests (PFTs) and high-resolution computed tomography (HRCT) measurements. For instance, our data demonstrated that baseline forced vital capacity (FVC) could effectively stratify MDA5+DM-ILD into three risk stages for prognosis: Stage1 (FVC% ≥ 50% predicted, low to medium risk), Stage2 (FVC% < 50% predicted, high risk), and Stage3 (unable to perform PFTs, very high risk), with estimated six-month mortalities of 15%, 45%, and 90%, respectively [3]. In the meantime, machine-learning-based HRCT analyses suggested that the extent of the lung parenchymal involvement, along with the texture of the lesions, mainly ground-glass opacity (GGO) and consolidation but not fibrotic components, are of high predictive value for prognosis [4, 5]. The attempt for these staging assessments is critical. For instance, the evidence of current treatment protocols, such as ‘triple therapy’ (with an average baseline FVC% of 78% [6]) or JAK inhibitor (with an average baseline FVC% of 73% [7]) studies, are largely based on those Stage1 (low to medium risk) population. For those with more advanced-stage disease, Stage3 (extremely high risk) in particular, no pharmacological treatment has been proven to be beneficial and lung transplantation is likely to be the last resort. Furthermore, the staging system is helpful in terms of upgrading the management strategy, i.e., to capture the disease in early stage as soon as possible and initiate treatment accordingly before too late.
Nevertheless, spontaneous pneumomediastinum and/or pneumothorax (SPM/P)—a severe structural complication in MDA5+DM-ILD characterized by mediastinal gas detection through HRCT or X-ray—has not been included (as “area of interest”) in machine-learning-based image analysis studies, despite its clinical significance in MDA5+DM patients [4, 5]. As a breaching of the pulmonary integrity, SPM/P has been frequently reported in MDA5+DM-ILD and linked with high mortality [8–14]. The only exceptional study reported by Jin et al. suggested that SPM/P (n = 36) does not exacerbate the already shattered outcome [15].
Therefore, we conducted the current retrospective study based on a large inception cohort of MDA5+DM-ILD patients to investigate three pertinent issues: (1) whether SPM/P indeed increases the mortality risk, (2) what are the predisposing factors of SPM/P, and (3) how the refined FVC%-based staging system by incorporating SPM/P performs in predicting the outcome of MDA5+DM-ILD.
Patients and methods
Study cohort
This retrospective study was conducted in an inception cohort at Renji Hospital, Shanghai, a major tertiary referral center for rheumatic diseases. Patients fulfilling the following criteria were included: initial diagnosis of dermatomyositis based on the Bohan and Peter criteria [16, 17] or clinically amyopathic dermatomyositis (CADM) defined by Sontheimer [18], subsequently confirmed according to the 239th European Neuromuscular Centre (ENMC) criteria for dermatomyositis [19], with anti-MDA5 antibody, with ILD diagnosed by pulmonary HRCT, and within 3-month interval between ILD diagnosis and hospital admission.
The study protocol was approved by the Ethics Committees of Renji Hospital (approval number: LY2023-284-C). Written informed consents were obtained from all patients.
Data collection
Baseline was defined as the time of first admission to our centre. Baseline demographic information, clinical manifestations, laboratory tests, pulmonary function tests (PFTs), infection events and survival status within 12 months were collected. Serologically, anti-MDA5 antibodies were detected using EUROLINE Autoimmune Inflammatory Myopathies 16 Ag (IgG), which were also quantitatively confirmed by ELISA with methodology provided previously [3].
The diagnosis of SPM/P was independently verified by two blinded investigators to ensure diagnostic accuracy and minimize potential bias. According to a previous report by Kaneki T et al. [20], based on the severity of imaging manifestations, SPM cases were additionally classified into three grades (Grade 1–3). Grade 1 SPM is defined as free air confined to the superior portion of the mediastinum only; grade 2 SPM is defined as free air extending from the superior portion of the mediastinum to the middle portion of the mediastinum; grade 3 SPM is defined as free air extending from the superior to the inferior portion of the mediastinum.
Routine PFTs or bedside spirometry were conducted at baseline. Patients who were too severe to perform bedside spirometry were classified as'unable to perform PFTs' [3]. Rapidly progressive interstitial lung disease (RPILD) was defined as the presence of each of the following three conditions within 3 months since the onset of respiratory symptoms: 1) progressive worsening of dyspnea, 2) acute alveolar exudative lesions (i.e., ground-glass opacities, consolidations) on pulmonary HRCT, 3) hypoxemia (PaO2/FiO2 < 300 mmHg based on arterial blood gas analysis) or severe restrictive ventilation dysfunction (FVC < 50% pred or unable to perform PFTs) [21]. The optimal cut-off values for continuous variables including Lactate dehydrogenase(LDH) and serum ferritin(SF) were in line with our previous validated data [3], i.e., 400 U/L for LDH, 2000 ng/ml for SF.
Infection events comprehensively evaluated by the treating physicians occurred prior to SPM/P or within the first 3 months since ILD onset in the non-SPM/P group were recorded. Assessments were based on a combination of clinical symptoms, imaging findings, and definitive pathogenic evidences, including qualified sputum cultures, blood cultures, bronchoalveolar lavage fluid cultures, viral DNA blood tests. The diagnosis of Aspergillus infections required the presence of relevant clinical symptoms and imaging features (e.g., cavitary lesions), along with either a positive immunological response to galactomannan Aspergillus antigen (GM) assay or pathogenic evidence of Aspergillus infection (microscopy or culture from lower respiratory tract samples) [22]. Cytomegalovirus (CMV) viremia was defined as a blood CMV-DNA concentration ≥ 1000 copies per mL, as determined by polymerase chain reaction [23, 24].
Study design and statistical analysis
To assess potentially different influences upon survival, patients with SPM/P were divided into two groups based on the time of SPM/P occurrence: (1) within three months of ILD diagnosis and (2) beyond three months [25].
Summary statistics were expressed as numbers with proportions for categorical variables and median with interquartile range (IQR) for continuous variables. Categorical variables were compared using Fisher’ s exact test or Pearson Chi-square test, while continuous variables were compared using Mann–Whitney U-test, as appropriate. Kaplan–Meier curves were plotted to compare the cumulative survival of patients with or without SPM/P. All candidate parameters with P-values < 0.1 between survivors and deceased in the univariable analysis were incorporated in the multivariable Cox regression for survival. Independent prognostic factors were determined by the final multivariable Cox regression model, with hazard ratios (HRs) and 95% Confidence Interval (CIs) calculated and presented in the forest map. To determine the predisposing factors of SPM/P, variables clinically related to SPM/P occurrence and those with P-values < 0.1 in the univariable analyses were included in multivariable logistic models. Based on the number of events, five factors were selected for the optimal multivariable model using the backward stepwise algorithm with the minimal Akaike Information Criterion (AIC) value. C-statistics were reported to reflect the discriminative performance of logistic models.
No imputation method was conducted to address missing data. Parameters with > 20% missing data were excluded from the multivariable analysis. All statistical analyses were conducted in IBM SPSS Statistics 26 or R version 4.3.1. Two-sided P-value < 0.05 was considered statistically significant.
Results
Study population and overall survival
The typical diagnostic sequence observed in MDA5+ DM patients and temporal distribution of SPM/P occurrence are illustrated in Fig. 1A, highlighting that most SPM/P events happened within the first 3 months of disease. The characteristic disease progression pattern observed in MDA5+ DM patients is depicted in Fig. 1A. The horizontal axis represents disease duration (months), with Month 0 (baseline) defined as the time of first hospital visit. The vertical axis shows the event numbers of SPM/P. Patient mortality was then tracked over the following 12-month period from baseline (Fig. 1A). Between April 2014 and April 2023, a total of 523 eligible patients were enrolled in this study, including 92 patients with SPM/P and 431 patients without SPM/P. Patients were stratified to three groups: (1) without SPM/P, (2) SPM/P onset ≤ 3 months, and (3) SPM/P onset > 3 months. Kaplan–Meier curve analysis shows that patients who developed SPM/P beyond 3 months since baseline (n = 29, 12-month mortality = 37.9%) had a comparable survival with those without SPM/P (n = 431, 12-month mortality = 35.4%) (Fig. 1B). Subsequent survival analyses thus focused exclusively on the 63 patients who developed SPM/P within 3 months (12-month mortality = 82.5%) and 431 patients without SPM/P (Dotted box in Fig. 1C).
Fig. 1.
A Typical diagnostic sequence of MDA5+DM and temporal distribution of SPM/P occurrence. The horizontal axis represents disease duration (months), with baseline defined as the time of first hospitalization. The vertical axis shows the event numbers of SPM/P. Survival outcomes were assessed over 12 months since baseline. B MDA5+DM patients were stratified to three groups: (1) without SPM/P, (2) SPM/P onset ≤ 3 months, and (3) SPM/P onset > 3 months. Kaplan–Meier curve analysis shows overall survival condition among MDA5+ DM-ILD patients with or without SPM/P, indicating that patients who developed SPM/P beyond 3 months had comparable survival with those without SPM/P. C Flow chart of the inception cohort of MDA5+ DM-ILD. Patient stratification: Individuals were first categorized by SPM/P occurrence. Those with SPM/P were further subdivided based on time of onset (≤ 3 vs. > 3 months). Subsequent analyses primarily focused on early-onset cases (≤ 3 months)
Baseline characteristics of deceased (n = 205) and survivors (n = 289) are summarized and compared in Table 1. The median age at admission was significantly higher in deceased (56.0 years versus. 49.0 years; P < 0.001). Male patients were more frequently observed in the deceased group (40.9% versus. 30.1%; P = 0.012). Survivors had a longer disease duration since the first dermatomyositis symptom (3 months versus. 2 months; P < 0.001).
Table 1.
Baseline characteristics between survivors and the deceased among MDA5+ DM-ILD patients
| Characteristics | Survivors n = 289 |
Deceased n = 205 |
P-value |
|---|---|---|---|
| Demographic information | |||
| Male, n(%) | 87(30.1) | 84(40.9) | 0.012 |
| Age on admission, y.o | 49.0(42.0–56.0) | 56.0(49.0–63.5) | < 0.001 |
| Disease duration since first symptom of DM, months | 3.0(2.0–4.0) | 2.0(1.0–3.0) | < 0.001 |
| Smoking history, n(%) | 31(10.7) | 27(13.1) | 0.406 |
| Clinical manifestations | |||
| SPM/P,n(%) | 11(3.8) | 52(25.3) | < 0.001 |
| RPILD,n(%) | 68(23.5) | 172(83.9) | < 0.001 |
| Fever,n(%) | 151(52.2) | 127(61.9) | 0.032 |
| Dysphagia,n(%) | 31(10.7) | 31(15.1) | 0.146 |
| Arthralgia,n(%) | 141(48.7) | 66(32.1) | < 0.001 |
| Myalgia,n(%) | 93(32.1) | 36(17.5) | < 0.001 |
| Muscle weakness,n(%) | 153(52.9) | 95(46.3) | 0.148 |
| Skin ulcer,n(%) | 54(18.6) | 38(18.5) | 0.967 |
| Gottorn'sign,n(%) | 220(76.1) | 154(75.1) | 0.798 |
| Heliotrope rash,n(%) | 226(78.2) | 164(80.0) | 0.629 |
| Periungual erythema,n(%) | 66(22.8) | 48(23.4) | 0.881 |
| Mechanic hands,n(%) | 97(33.5) | 61(29.7) | 0.371 |
| Laboratory tests | |||
| SF(ng/mL) | 725.9(318.6–1400.0) | 1,500.0(851.9–2,733.9) | < 0.001 |
| SF > 2000 ng/mL,n(%) | 28(9.6) | 74(36.0) | < 0.001 |
| LDH(U/L) | 314.0(238.5–409.5) | 358.0(268.5–482.0) | < 0.001 |
| LDH > 400U/L,n(%) | 31(10.7) | 89(43.4) | < 0.001 |
| ALT(IU/L) | 55.0(33.0–101.0) | 63.0(34.5–112.0) | 0.141 |
| CRP elevation,n(%) | 56(19.4) | 86(41.9) | < 0.001 |
| Anti-Ro52 antibody positive,n(%) | 180(62.2) | 149(72.6) | 0.016 |
| Lymphocyte count < 0.8*109/L,n(%) | 136(47.0) | 158(77.0) | < 0.001 |
| Pulmonary function | |||
| FVC_based stage,n(%) | < 0.001 | ||
| FVC% predicted ≥ 50% (Stage1) | 230(79.5) | 50(24.3) | |
| FVC% predicted < 50% (Stage2) | 53(18.3) | 52(25.3) | |
| Unable to perform PFTs (Stage3) | 6(2.0) | 103(50.2) | |
| Infectiona | |||
| Bacterial,n(%) | 30(10.3) | 50(24.3) | < 0.001 |
| Aspergillus,n(%) | 17(5.8) | 27(13.1) | < 0.001 |
| PJP,n(%) | 2(0.6) | 15(7.3) | < 0.001 |
| CMV,n(%) | 39(13.4) | 39(19.0) | 0.072 |
Bold values indicate significance. DM: dermatomyositis; SPM/P: spontaneous mediastinal emphysema/pneumothorax; CRP elevation was defined as C-reactive protein > 8 mg/L. FVC-Stage1 was defined as FVC% predicted ≥ 50%, FVC-Stage2 was defined as FVC% predicted < 50%, FVC-Stage3 was defined as unable to perform PFTs
RPILD rapid-progressive interstitial lung disease, SF serum ferritin, LDH lactate dehydrogenase, ALT alanine Aminotransferase, PJP Pneumocystis jirovecii pneumonia, CMV cytomegalovirus
ainfections occurred before the development of SPM/P or occurred within the first 3 months of ILD were counted. Continuous data are presented as median (first quartile, third quartile), and binary data are presented as n/N (%) of the patients, where N is the total number of patients with available data
Notably, SPM/P occurred significantly more frequently in the deceased than survivors (25.3% versus. 3.8%, P < 0.001). The incidence of RPILD was significantly higher in the deceased than the survivors (83.9% versus. 23.5%, P < 0.001). As expected, pulmonary function parameters were significantly worse in the deceased group on admission. Specifically, more patients were categorized to Stage2 (FVC% predicted < 50%) and Stage3 (unable to perform PFTs) based on PFTs at baseline in deceased than survivors (25.3% versus. 18.3% and 50.2% versus. 2.0%, respectively, P < 0.001). Fever was significantly more common among deceased (61.9% versus. 52.2%, P = 0.032) while arthralgia and myalgia were less frequently reported in deceased (32.1% versus. 48.7%, P < 0.001; 17.5% versus. 32.1%, P < 0.001, respectively). Laboratory analyses demonstrated significantly higher levels of SF and LDH in deceased compared to survivors (SF: 1500.0 versus. 725.9 ng/mL, P < 0.001; LDH: 358.0 versus. 314.0 U/L, P < 0.001). Additionally, the deceased group showed significantly higher percentages of elevated C-reactive protein (CRP), anti-Ro52 antibody positivity, and lymphocyte counts < 0.8 × 109/L (41.9% versus. 19.4%, P < 0.001; 72.6% versus. 62.2%, P = 0.016; and 77.0% versus. 47.0%, P < 0.001, respectively). Furthermore, deceased had significantly higher incidences of bacterial, Aspergillus, and Pneumocystis jirovecii pneumonia (PJP) infections within the first 3-month of ILD.
Prognostic factors in MDA5+DM-ILD
The multivariable Cox regression model identified predictors for all-cause mortality in MDA5+DM-ILD patients, as summarized in Fig. 2. SPM/P occurred within 3 months since ILD onset emerged as an independent risk factor for survival (HR 1.59, 95% CI: 1.09–2.34, P = 0.018). Moreover, RPILD was also identified as an independent risk factor for survival (HR 4.04, 95% CI: 2.38–6.87, P < 0.001). A particularly noteworthy finding was that patients classified as FVC-Stage 3 (unable to perform PFTs) had a significantly worse outcome compared to those classified as FVC-Stage 1 (FVC% ≥ 50% predicted) (HR 4.86, 95% CI: 2.87–8.21, P < 0.001). In addition, older age (HR 1.06, 95% CI: 1.04–1.07, P < 0.001), LDH > 400U/L (HR 1.89, 95% CI: 1.36–2.63, P < 0.001) and a shorter disease duration since first symptom of dermatomyositis were also identified as poor prognostic factors for MDA5+DM-ILD.
Fig. 2.
Predictors for all-cause mortality in MDA5+ DM-ILD patients determined by multivariable cox analysis. Bold values indicate significance. MDA5, anti-melanoma differentiation-associated gene 5; DM, dermatomyositis; ILD, interstitial lung disease; SPM/P, spontaneous pneumomediastinum/pneumothorax; RPILD, rapid-progressive interstitial lung disease; CRP, C-reactive protein; LDH, lactate dehydrogenase; SF, serum ferritin; PJP, Pneumocystis jirovecii pneumonia; CMV, cytomegalovirus
Given that potential collinearity could exist between FVC%-based staging and RPILD, RPILD was excluded from the multivariable Cox regression to test the robustness of the primary findings (Supplementary Fig. 1). In this sensitivity analysis, SPM/P occurred within 3 months of ILD was consistently associated with a significantly higher mortality (HR 1.59, 95% CI: 1.08–2.33, P = 0.018).
Baseline comparisons between the SPM/P group and the non-SPM/P group
Baseline characteristics of patients in the SPM/P group were compared with those in the non-SPM/P group, as summarized in Table 2. Significantly more patients in the SPM/P group were classified as FVC-Stage3 (57.1% versus. 16.9%, P < 0.001), whereas less were classified as FVC-Stage1 (14.3% versus. 62.8%, P < 0.001). The incidence of RPILD was significantly higher in the SPM/P group than the non-SPM/P group (87.3% versus. 42.9%, P < 0.001). More male patients were observed in the SPM/P group (53.9% versus. 31.7%, P < 0.001). Conversely, myalgia was less frequently reported in patients with SPM/P (15.8% versus. 27.6%, P = 0.048). Levels of SF and LDH in the SPM/P group were significantly higher than the non-SPM/P group (1515.0 ng/mL versus. 970.8 ng/mL, P = 0.001, and 449.0U/L versus. 319.0U/L, P < 0.001, respectively). Additionally, CRP elevation was also more common in the SPM/P group (49.2% versus. 25.7%, P < 0.001). Infection events occurred significantly more frequently in the SPM/P group. In specific, bacterial, Aspergillus infections and CMV viremia occurred more frequently in the SPM/P group (28.5% versus. 14.3%, P = 0.004, 19.0% versus. 7.4%, P = 0.002, and 28.5% versus. 13.9%, P = 0.003, respectively). In addition, the SPM/P group had a trend-toward higher incidence of PJP infection (7.9% versus. 2.7%, P = 0.053). Treatment information including maximum dosage of prednisone and immunosuppressants were presented in Supplementary Table 1.
Table 2.
Baseline characteristics of MDA5+ DM-ILD patients without and with SPM/P occurred within the first 3 months of ILD
| Characteristics | Without SPM/P n = 431 |
SPM/P n = 63 |
P-value |
|---|---|---|---|
| Demographic information | |||
| Male,n(%) | 137(31.7) | 34(53.9) | < 0.001 |
| Age on admission,y.o | 52(44–59) | 51(46–59) | 0.708 |
| Disease duration since first symptom of DM, months | 2.0(2.0–3.0) | 3.0(2.0–4.0) | 0.220 |
| Smoking history,n(%) | 48(11.1) | 10(15.8) | 0.275 |
| Clinical manifestations | |||
| RPILD,n(%) | 185(42.9) | 55(87.3) | < 0.001 |
| Fever,n(%) | 244(56.6) | 34(53.9) | 0.693 |
| Dysphagia,n(%) | 51(11.8) | 11(17.4) | 0.208 |
| Arthralgia,n(%) | 185(42.9) | 22(34.9) | 0.229 |
| Myalgia,n(%) | 119(27.6) | 10(15.8) | 0.048 |
| Muscle weakness,n(%) | 221(51.2) | 27(42.8) | 0.212 |
| Skin ulcer,n(%) | 80(18.5) | 12(19.0) | 0.926 |
| Gottorn'sign,n(%) | 323(74.9) | 51(80.9) | 0.299 |
| Heliotrope rash,n(%) | 338(78.4) | 52(82.5) | 0.454 |
| Periungual erythema,n(%) | 99(22.9) | 15(23.8) | 0.883 |
| Mechanic hands,n(%) | 143(33.1) | 15(23.8) | 0.136 |
| Laboratory tests | |||
| SF(ng/mL) | 970.8(390.3–1765.0) | 1,515.0(689.9–2764.3) | 0.001 |
| SF > 2000 ng/mL,n(%) | 76(17.6) | 26(41.2) | < 0.001 |
| LDH(U/L) | 319.0(246.0–424.0) | 449.0(298.0–597.0) | < 0.001 |
| LDH > 400U/L,n(%) | 90(20.8) | 30(47.6) | < 0.001 |
| ALT(IU/L) | 58(34–106) | 58(31–97) | 0.656 |
| Anti-MDA5 titer(RU/ml) | 160.0(119.4–191.3) | 150(122.4–188.8) | 0.593 |
| CRP elevation,n(%) | 111(25.7) | 31(49.2) | < 0.001 |
| Anti-Ro52 antibody positive,n(%) | 287(66.5) | 42(66.6) | 0.990 |
| Lymphocyte count < 0.8*109/L,n(%) | 250(58.0) | 44(69.8) | 0.074 |
| Pulmonary function | |||
| FVC_based stage,n(%) | < 0.001 | ||
| FVC% predicted ≥ 50% (Stage1) | 271(62.8) | 9(14.3) | |
| FVC% predicted < 50% (Stage2) | 87(20.1) | 18(28.5) | |
| Unable to perform PFTs (Stage3) | 73(16.9) | 36(57.1) | |
| Infection | |||
| Bacterial,n(%) | 62(14.3) | 18(28.5) | 0.004 |
| Aspergillus,n(%) | 32(7.4) | 12(19.0) | 0.002 |
| PJP,n(%) | 12(2.7) | 5(7.9) | 0.053 |
| CMV,n(%) | 60(13.9) | 18(28.5) | 0.003 |
Bold values indicate significance. MDA5,anti-melanoma differentiation-associated gene 5; CRP elevation was defined as C-reactive protein > 8 mg/L
DM dermatomyositis, ILD interstitial lung disease, SPM/P spontaneous pneumomediastinum/pneumothorax, RPILD rapid-progressive interstitial lung disease, FVC-Stage1 was defined as FVC% predicted ≥ 50%, FVC-Stage2 was defined as FVC% predicted < 50%, FVC-Stage3 was defined as unable to perform PFTs. RPILD rapid-progressive interstitial lung disease, LDH lactate dehydrogenase, SF serum ferritin, PJP Pneumocystis jirovecii pneumonia, CMV cytomegalovirus
Predisposing factors for SPM/P occurrence
All variables with P-value < 0.1 between the SPM/P and non-SPM/P groups in the univariable analysis, including demographic characteristics, serum biomarkers, and respiratory physiological parameters potentially associated with SPM/P occurrence, were incorporated into the multivariable logistic regression model. The optimal model selected based on the minimum AIC value by the backward stepwise algorithm was presented in Supplementary Table 2. The final multivariable logistic regression model, as presented in Table 3, identified three predisposing factors for SPM/P occurrence within 3 months in MDA5+DM-ILD: male sex (odds ratio [OR] 2.57, 95% CI 1.42–4.66, P = 0.002), severe restrictive ventilation dysfunction (FVC-Stage2: OR 5.23, 95% CI 2.19–12.45, P < 0.001; FVC-Stage3: OR 11.79, 95% CI 5.15–26.99, P < 0.001; both compared to FVC-Stage1), and CMV viremia (OR 2.32, 95% CI 1.16–4.63, P = 0.017). The C-statistics of the final multivariable logistic model was 0.83 (95% CI: 0.78–0.88).
Table 3.
Baseline factors associated with the occurrence of SPM/P within the first 3 months of ILD determined by multivariable logistic regression
| Characteristics | Level | OR (univariable) with 95%CI | P-value | OR (multivariable) with 95%CI | P-value |
|---|---|---|---|---|---|
| Age | 1.00 (0.98–1.02) | 0.856 | |||
| Male | 2.52 (1.47–4.30) | < 0.001 | 2.57 (1.42–4.66) | 0.002 | |
| Myalgia | 0.49 (0.24–1.00) | 0.051 | |||
| FVC-based stage | |||||
| FVC% predicted ≥ 50% | Stage1 | ||||
| FVC% predicted < 50% | Stage2 | 6.23 (2.70–14.37) | < 0.001 | 5.23 (2.19–12.45) | < 0.001 |
| Unable to perform PFTs | Stage3 | 14.85 (6.84–32.23) | < 0.001 | 11.79 (5.15–26.99) | < 0.001 |
| RPILD | 9.14 (4.25–19.66) | < 0.001 | |||
| SF > 2000 ng/mL | 3.28 (1.88–5.74) | < 0.001 | |||
| LDH > 400U/L | 5.38 (2.98–9.68) | < 0.001 | 1.65(0.88–3.12) | 0.120 | |
| CRP elevation | 2.79 (1.63–4.79) | < 0.001 | |||
| Lymphocyte count < 0.8*109/L | 1.68 (0.95–2.97) | 0.076 | |||
| Bacterial infection | 2.38 (1.29–4.38) | 0.005 | |||
| Aspergillus infection | 2.93 (1.42–6.06) | 0.004 | 1.97 (0.88–4.42) | 0.098 | |
| PJP infection | 3.01 (1.02–8.85) | 0.045 | |||
| CMV viremia | 2.43 (1.32–4.47) | 0.004 | 2.32 (1.16–4.63) | 0.017 | |
Bold values indicate significance. CRP elevation was defined as CRP > 8 mg/L; CI confidence interval. FVC-Stage1 was defined as FVC% predicted ≥ 50%, FVC-Stage2 was defined as FVC% predicted < 50%, FVC-Stage3 was defined as unable to perform PFTs
RPILD rapid-progressive interstitial lung disease, LDH lactate dehydrogenase, SF serum ferritin, PJP Pneumocystis jirovecii pneumonia, CMV cytomegalovirus
Prognostic factors in MDA5+DM-ILD with SPM/P
Baseline characteristics between survivors (n = 11) and deceased (n = 52) among patients with SPM/P occurred within 3 months were compared in Supplementary Table 3. Deceased had significantly shorter disease duration since first symptom of dermatomyositis than survivors (2.0 versus. 3.0 months, P = 0.044). Significantly more patients suffered concomitant pneumothorax in the deceased group compared to survivors (53.8% versus. 18.2%, P = 0.031). More patients received drainage treatment in the deceased than survivors (34.6% versus. 9.1%, P < 0.001). Consistent with clinical expectations, a significantly higher percentage of deceased were classified as FVC-Stage3 compared to survivors (65.3% versus. 18.1%, P < 0.001). Not surprisingly, deceased had significantly higher levels of SF and LDH.
We additionally evaluate whether the severity of SPM (based on radiological grades) and drainage treatment would influence the prognosis in patients with SPM/P. As showed in Supplementary Fig. 2 A, patients with widest extent of SPM (Grade 3) had numerically higher mortality, although non-significantly different from those with Grade 1 and 2 SPM (log-rank P = 0.27). A total of 19 (30.2%) patients with SPM/P received the drainage treatment. However, survival rates did not differ significantly between patients who received drainage and those who did not (log-rank P = 0.20), both extremely low (Supplementary Fig. 2B). Finally, we performed a multivariable Cox regression analysis to determine the predictors for all-cause mortality in MDA5+DM-ILD patients with SPM/P (Supplementary Table 4). RPILD was the strongest prognostic factor in this subpopulation, following by age. Neither the severity of SPM nor drainage showed significant association with the prognosis.
Mortality risk stratified by FVC-Stage and the occurrence of SPM/P
All-cause mortalities of patients stratified by the occurrence of SPM alone (n = 33), pneumothorax alone (n = 11), or both (n = 19) within 3 months, were summarized in the Venn diagram (Fig. 3A). Overall strikingly high mortalities were observed. Specifically, patients with SPM alone had a mortality of 72.7%, while complicated pneumothorax alone would push up the mortality to 100.0%. The mortality risk matrix delineates the prognostic value of combining FVC%-based staging and SPM/P occurrence (Fig. 3B), categorizing patients into three distinct risk groups: low-to-medium risk (green, mortality ~ 15%), high risk (yellow, mortality ~ 50%), and extremely high risk (red, mortality > 80%). This novel stratification provides a comprehensive visual representation of mortality risks associated with different clinical scenarios. Overall, the presence of SPM/P significantly elevated the risk profile, particularly when pneumothorax occurred; with the exception of FVC-Stage3, where the mortality was already “saturated” (beyond 90%) irrespective of the existence of SPM/P.
Fig. 3.
A The Venn diagram illustrates the distribution of structural complications in the cohort: 33 patients developed spontaneous pneumomediastinum (SPM) alone, 11 developed pneumothorax alone, and 19 patients experienced both complications concurrently. Different mortalities were observed between groups. B Mortality risk was stratified into a 3 × 3 matrix (nine-grid) based on three-category FVC-Stage (horizontal axis): FVC-Stage1 (FVC% ≥ 50% predicted), FVC-Stage2 (FVC% < 50% predicted), FVC-Stage3 (unable to perform PFTs) and the presence/absence of structural complications (vertical axis): without SPM/P, SPM alone, and SPM with pneumothorax. The occurrence of SPM/Pneumothorax could directly elevate the mortality risk to a higher stage. C A proposed prognostic staging system by incorporating pulmonary physiology parameters (baseline FVC% and PaO2/FiO2), structural changes (extent of GGO and consolidation on HRCT, the occurrence of SPM/Pneumothorax), age and prognostic biomarkers (lymphocyte, LDH and serum ferritin) in MDA5+DM-ILD. Patients could be stratified to three risk stages with distinct estimated 12-month mortalities: Stage 1 (≈15%), Stage 2 (≈50%), and Stage 3 (≈90%). A gradient color scale (green-orange-red) indicates a gradually increasing mortality risk from left to right. Darker color stands for higher mortality risk. The occurrence of SPM/Pneumothorax could directly elevate the mortality risk to a higher stage. Older age, lymphocytopenia, elevated LDH and SF levels additionally increase the mortality risk
A refined prognostic staging tool proposed for MDA5+DM-ILD was thus established (Fig. 3C), integrating several key clinical dimensions: pulmonary physiology (based on FVC% three-categorical Stage and its surrogate: PaO2/FiO2) [3, 4], pulmonary structural changes (extent of GGO and consolidation on HRCT, the occurrence of SPM/Pneumothorax) [5, 9, 26], age, and key biomarkers [3, 21, 27–30]. In specific, PaO2/FiO2 < 300 mmHg has been identified as an independent prognostic factor in MDA5+DM-ILD in our previous work [3]. Besides, in accordance to the Berlin definition for ARDS [31], PaO2/FiO2 < 200 mmHg was employed to define a highest-risk population (Stage 3) in the staging system. The prognostic thresholds of the extent of GGO and consolidation on HRCT, the two most significant radiological lesions, i.e., 25%/50%, were derived from validated studies [5, 9, 26]. In addition, Age over 60 years old., lymphocyte < 800 × 10⁶/L, LDH > 400U/L and SF > 2000 ng/ml were all identified as poor prognostic factors of MDA5+DM-ILD in previous studies [3, 21, 30]. Patients could be stratified to three risk stages with distinct estimated 12-month mortalities: Stage 1 (≈15%), Stage 2 (≈50%), and Stage 3 (≈90%) in this staging system. A gradient color scale (green-orange-red) indicates a gradually increasing mortality risk from left to right. This comprehensive staging system, presented in Fig. 3C, employs a risk coding system that correlates with estimated 12-month mortalities, providing a clinically valuable prognostic assessment framework.
Discussion
In the current study, we systemically analyzed SPM/P, a unique and not uncommon pulmonary structural damage, in a large single-centre MDA5+DM-ILD inception cohort. The incidence of SPM/P in this cohort is 17.6%, which is at the upper margin of previous reports (5% to 15.5%) [21, 32–34]. The analysis revealed that the prognosis of SPM/P is time-sensitive, particularly when occurring within the initial 3-month period following MDA5+DM-ILD diagnosis (based on the first abnormal HRCT). This early-onset complication demonstrated a strong association with mortality, resulting in a strikingly low 12-month overall survival rate of merely 17.5%.
Furthermore, the majority of SPM/P cases (68.5%) occurred during the acute phase of the disease, representing the most critical subgroup. In contrast, patients who developed SPM/P beyond 3 months demonstrated mortalities comparable to patients without SPM/P, suggesting a distinct prognostic pattern for late-onset cases. This finding provides a plausible explanation for the “no-difference” studies [8, 21], if the cohort enriched more cases who developed SPM/P in subacute or chronic phases. Similarly, Le Goff [32] reported a 12-month survival rate of 36% which is higher than our finding –the discrepancy may well be attributed to the dilution effect of including SPM/P beyond 3 months since the diagnosis of ILD. Our subsequent analyses thus focused on those SPM/P cases within 3-month, in order to align with previously established risk stratification modules, which were all derived from patients’ baseline data with disease duration less than 3 months.
Next, we took a step back to assess all possible risk factors from baseline to predict 12-month mortality through multivariable cox analysis. It is confirmatory that FVC%-based staging is the most prominent contributor. In addition, SPM/P also plays a role with a HR of 1.59. Other risk factors, such as older age, higher LDH level have been previously reported [15, 31, 32, 35, 36]. Of interest, the time interval between the first symptom of dermatomyositis and baseline was associated with outcome – the longer interval the better survival. For example, a patient with a 6-month duration of cutaneous DM lesions, although with a ILD diagnosis within 3-month by definition, he/she might have a better chance for survival due to a not-so-rapidly-progressive nature of the disease.
Since SPM/P indeed contributes to the mortality, we proceeded to address its predisposing factors. Step-wise multivariable logistic regression suggested that again, FVC%-based advanced stages were highly relevant. This observation supports the speculation that the severity of ILD is what really matters; while the presence of SPM/P is the epiphenomenon of significant pulmonary structural damage [37]. Of note, male sex (irrelevant to smoking status) is associated with SPM/P, which has been replicated previously [12, 38]. Concomitant infections, such as Aspergillus and CMV, have been linked to SPM/P [15]. In our center, ganciclovir is routinely initiated when CMV-DNA levels exceed 103copies/mL and continued until CMV-DNA is no longer detectable in the blood. Similarly, confirmed Aspergillus infections are treated with voriconazole, with therapy guided by clinical symptoms and laboratory indicators. Additionally, sulfamethoxazole-trimethoprim (SMZ-TMP) is administered for the prophylaxis and treatment of PJP. According to our data, CMV viremia was associated with the occurrence of SPM/P within the first 3 months of ILD, although it’s probably secondary to intensive immunosuppressive treatment. This finding thus suggested that monitoring and treating CMV infection should be considered as a routine practice in the management of MDA5+DM-ILD to prevent SPM/P. However, CMV viremia was not significantly associated with overall survival.
The predictors for all-cause mortality in MDA5+DM-ILD patients with SPM/P were further explored. RPILD was demonstrated to be the strongest poor prognostic factor in this subpopulation, following by age. Neither the severity of SPM (based on radiological grades) nor drainage showed significant association with the prognosis based on our data. However, this finding should be interpreted cautiously due to relatively limited sample size. In addition, there might be a potential treatment-by-indication error for evaluating the impact of drainage. In clinical practice, patients with more severe SPM/P were prone to receive drainage, which could explain the extremely high mortality in those who received drainage. Future prospective studies with larger sample size and more rigorous design are expected to explore the accurate association between radiological severity of SPM, drainage treatment and overall survival in MDA5+DM-ILD.
It’s noteworthy that patients with pneumothorax, concomitant with SPM or not, had an even worse outcome with extremely high mortality (overall exceeding 90%), which could be explained by the fact that pneumothorax will incur more pronounced ventilation/perfusion imbalance and often be accompanied with extensive interstitial lesions. Our data convincingly support that pneumothorax should be regarded as a red-flag sign together with SPM in MDA5+DM-ILD, particularly within the first 3 months.
Finally, our findings demonstrate that the FVC%-based staging system remains the most robust risk stratification tool in MDA5+DM-ILD. However, the presence of SPM/P provides additional prognostic value. A risk matrix combined this observation with FVC%-based staging indeed enhanced the prediction accuracy. In other words, when SPM/P occurred, the patient’s initial FVC%-based stage should be escalated to a higher risk level. In our practice, SPM/P especially with the presence of pneumothorax, is considered as a “red flag” for initiating emergency lung transplantation evaluation. A refined MDA5+DM-ILD staging tool with a combination of data-driven items and expert-opinion has thus been proposed.
Our study has several limitations. First, as a retrospective study, missing data is an inherent issue, particularly in real-world cohorts. Parameters with > 20% missing values (e.g., KL-6, PaO2/FiO2) were excluded from analysis, potentially overlooking some prognostic factors. Second, the study was limited to Chinese population, so caution is needed when extrapolating conclusions to other ethnic groups. Third, only hospitalized patients with MDA5+DM-ILD were included, introducing severe-patient-enrollment bias typical of a tertiary referral center. Fourth, potential anti-MDA5 syndrome may be overlooked; thus, prompt screening for anti-MDA5 antibodies is crucial, especially in patients with suspected autoimmune-related interstitial pneumonia or unexplained acute respiratory distress syndrome. Lastly, due to treatment-by-indication bias in real-world studies, the impact of immunosuppressive treatments on survival and the development of SPM/P was not assessed.
In conclusion, our study substantiates the role of SPM/P occurring within the first 3-month of ILD serves as a poor prognostic factor in MDA5+DM patients. Furthermore, incorporating SPM/P as an indicator of severe pulmonary structural damage improves the predictive performance of pulmonary function (FVC%)-based staging system. Further investigation is essential to validate these findings and to foster optimal treatment strategies tailored upon different stages of the disease.
Supplementary Information
Acknowledgements
We thank the Rheumatology-intensive care unit staff and the lung transplantation team (lead by Prof. Chang Chen) for their vital support.
Clinical trial number
Not applicable.
Authors’ contributions
J.J. and W.W wrote the main manuscript text, W.X. prepared figures,H.W. and X.W. collected data, Q.F and S.Y. the proofreading of articles. All authors reviewed the manuscript.
Funding
This work was supported by the Shanghai Science and Technology Committee Rising-Star Program (24YF2725100) and the Science and Technology Commission of Shanghai Municipal Jiading District (JDKW-2023–0021).
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study was conducted in accordance with Good Clinical Practice, including the Declaration of Helsinki. The study protocol and informed consent form was approved by the Ethics Committees of Renji Hospital (approval number: LY2023-284-C). All patients provided written informed consent before screening.
Consent for publication
Written informed consents were obtained from all patients.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Wanlong Wu and Shuang Ye are joint senior authors and contributed equally to this work.
Contributor Information
Wanlong Wu, Email: yczxwwl@163.com.
Shuang Ye, Email: ye_shuang2000@163.com.
References
- 1.Selva-O’Callaghan A, Pinal-Fernandez I, Trallero-Araguás E, Milisenda JC, Grau-Junyent JM, Mammen AL. Classification and management of adult inflammatory myopathies. Lancet Neurol. 2018;17(9):816–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Wu W, Guo L, Fu Y, Wang K, Zhang D, Xu W, et al. Interstitial lung disease in anti-MDA5 positive dermatomyositis. Clinic Rev Allerg Immunol. 2021;60(2):293–304. [DOI] [PubMed] [Google Scholar]
- 3.Wu W, Xu W, Sun W, Zhang D, Zhao J, Luo Q, et al. Forced vital capacity predicts the survival of interstitial lung disease in anti-MDA5 positive dermatomyositis: a multi-centre cohort study. Rheumatology. 2021;61(1):230–9. [DOI] [PubMed] [Google Scholar]
- 4.Xu W, Wu W, Zheng Y, Chen Z, Tao X, Zhang D, et al. A computed tomography radiomics-based prediction model on interstitial lung disease in anti-MDA5-positive dermatomyositis. Front Med. 2021;8:768052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Xu W, Wu W, Zhang D, Chen Z, Tao X, Zhao J, et al. A novel CT scoring method predicts the prognosis of interstitial lung disease associated with anti-MDA5 positive dermatomyositis. Sci Rep. 2021;11(1):17070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tsuji H, Nakashima R, Hosono Y, Imura Y, Yagita M, Yoshifuji H, et al. Multicenter prospective study of the efficacy and safety of combined immunosuppressive therapy with high-dose glucocorticoid, Tacrolimus, and Cyclophosphamide in interstitial lung diseases accompanied by anti-Melanoma Differentiation–Associated gene 5–positive dermatomyositis. Arthritis Rheumatol. 2020;72(3):488–98. [DOI] [PubMed] [Google Scholar]
- 7.Chen Z, Wang X, Ye S. Tofacitinib in amyopathic dermatomyositis-associated interstitial lung disease. N Engl J Med. 2019;381(3):291–3. [DOI] [PubMed] [Google Scholar]
- 8.Arango Guerra P, Monsalve Yepes S, Chavarriaga Restrepo A, Velasquez Franco CJ. Spontaneous pneumomediastinum: An extra muscular manifestation of anti-MDA5 dermatomyositis. Report of 2 cases. Reumatología Clínica (English Edition). 2024;20(4):218–22. [DOI] [PubMed] [Google Scholar]
- 9.Yamaguchi K, Yamaguchi A, Itai M, Kashiwagi C, Takehara K, Aoki S, et al. Clinical features of patients with anti-melanoma differentiation-associated gene-5 antibody-positive dermatomyositis complicated by spontaneous pneumomediastinum. Clin Rheumatol. 2019;38(12):3443–50. [DOI] [PubMed] [Google Scholar]
- 10.Fujiwara T. Pneumomediastinum in pulmonary fibrosis. Chest. 1993;104(1):44–6. [DOI] [PubMed] [Google Scholar]
- 11.Abe K, Furuta S, Kobayashi Y, Sugiyama T, Kagami SI, Nakagomi D, et al. Prognosis of spontaneous pneumomediastinum occurring in dermatomyositis or polymyositis patients with interstitial lung disease according to antimelanoma differentiation-associated gene 5 antibody status: a retrospective cohort study. RMD Open. 2023;9(1):e002770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Subki AH, Almani IM, Albeity A, Aljabri BK, Alsolaimani R, Halabi H. Spontaneous Pneumomediastinum and Subcutaneous Emphysema in Dermatomyositis: A Case Series and Literature Review. JIR. 2023;16:1431–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yashiro M, Asano T, Sato S, Kobayashi H, Watanabe H, Miyata M, et al. Anti-MDA5 antibody-positive hypomyopathic dermatomyositis complicated with pneumomediastinum. Fukushima J Med Sci. 2018;64(2):89–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Okabayashi H, Nakashima S, Fujino K, et al. Tension Pneumomediastinum in Anti-MDA5 Antibody-positive Dermatomyositis-associated Interstitial Lung Disease: A Case Report and Literature Review. Intern Med. Published online 2024:3418–3423. 10.2169/internalmedicine.3418-23 [DOI] [PMC free article] [PubMed]
- 15.Jin Q, Lin S, Chen X, Xu Y, Tian X, He L, et al. Spontaneous pneumomediastinum in anti-MDA5-positive dermatomyositis: prevalence, risk factors, and prognosis. Semin Arthritis Rheum. 2024;65:152352. [DOI] [PubMed] [Google Scholar]
- 16.Bohan A, Peter JB. Polymyositis and dermatomyositis: (first of two parts). N Engl J Med. 1975;292(7):344–7. [DOI] [PubMed] [Google Scholar]
- 17.Bohan A, Peter JB. Polymyositis and dermatomyositis: (second of two parts). N Engl J Med. 1975;292(8):403–7. [DOI] [PubMed] [Google Scholar]
- 18.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(4):626–36. [DOI] [PubMed] [Google Scholar]
- 19.Mammen AL, Allenbach Y, Stenzel W, Benveniste O, Allenbach Y, Benveniste O, et al. 239th ENMC international workshop: classification of dermatomyositis, Amsterdam, the Netherlands, 14–16 december 2018. Neuromuscul Disord. 2020;30(1):70–92. [DOI] [PubMed] [Google Scholar]
- 20.Kaneki T, Kubo K, Kawashima A, Koizumi T, Sekiguchi M, Sone S. Spontaneous pneumomediastinum in 33 patients: yield of chest computed tomography for the diagnosis of the mild type. Respiration. 2000;67(4):408–11. [DOI] [PubMed] [Google Scholar]
- 21.Wu W, Guo B, Sun W, Chen D, Xu W, Chen Z, et al. Effectiveness and safety of tofacitinib versus calcineurin inhibitor in interstitial lung disease secondary to anti-MDA5-positive dermatomyositis: a multi-centre cohort study. Eur Respir J. 2025;30:2401488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Denning DW, Cadranel J, Beigelman-Aubry C, Ader F, Chakrabarti A, Blot S, et al. Chronic pulmonary aspergillosis: rationale and clinical guidelines for diagnosis and management. Eur Respir J. 2016;47(1):45–68. [DOI] [PubMed] [Google Scholar]
- 23.García-Bustos V, Salavert M, Blanes R, Cabañero D, Blanes M. Current management of CMV infection in cancer patients (solid tumors). Epidemiology and therapeutic strategies. Rev Esp Quimioter. 2022;35(Suppl3):74–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bruminhent J, Autto S, Rotjanapan P, Ngarmjanyaporn P, Bushyakanist A, Kirdlarp S, et al. A prospective study of cytomegalovirus-specific cell-mediated immune monitoring and cytomegalovirus infection in patients with active systemic lupus erythematosus receiving immunosuppressants. Open Forum Infect Dis. 2021;8(6):ofab248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Li C, Liang M, Jiang H, Zhao J, Wu C, Wang Q, et al. The long-term prognosis of pneumomediastinum associated with dermatomyositis: a two-centre retrospective cohort study. Rheumatology. 2021;60(5):2288–95. [DOI] [PubMed] [Google Scholar]
- 26.Fujiki Y, Kotani T, Isoda K, Ishida T, Shoda T, Yoshida S, et al. Evaluation of clinical prognostic factors for interstitial pneumonia in anti-MDA5 antibody-positive dermatomyositis patients. Mod Rheumatol. 2018;28(1):133–40. [DOI] [PubMed] [Google Scholar]
- 27.Sugiyama Y, Yoshimi R, Tamura M, Takeno M, Kunishita Y, Kishimoto D, et al. The predictive prognostic factors for polymyositis/dermatomyositis-associated interstitial lung disease. Arthritis Res Ther. 2018;20(1):7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lian X, Zou J, Guo Q, Chen S, Lu L, Wang R, et al. Mortality risk prediction in amyopathic dermatomyositis associated with interstitial lung disease. Chest. 2020;158(4):1535–45. [DOI] [PubMed] [Google Scholar]
- 29.Lv X, Jin Y, Zhang D, Li Y, Fu Y, Wang S, et al. Low circulating monocytes is in parallel with lymphopenia which predicts poor outcome in anti-melanoma differentiation-associated gene 5 antibody-positive dermatomyositis-associated interstitial lung disease. Front Med. 2022;8:808875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.He S, Zhou Y, Fan C, Ma J, Chen Y, Wu W, et al. Differences in sex- and age-associated mortality in patients with anti-MDA5-positive dermatomyositis. Mod Rheumatol. 2023;33(5):975–81. [DOI] [PubMed] [Google Scholar]
- 31.Acute Respiratory Distress Syndrome: The Berlin Definition. JAMA. 2012;307(23). Available from: http://jama.jamanetwork.com/article.aspx?doi=10.1001/jama.2012.5669. Cited 2025 Aug 11. [DOI] [PubMed]
- 32.Le Goff B, Chérin P, Cantagrel A, Gayraud M, Hachulla E, Laborde F, et al. Pneumomediastinum in interstitial lung disease associated with dermatomyositis and polymyositis. Arthritis Rheum. 2008;61(1):108–18. [DOI] [PubMed] [Google Scholar]
- 33.Zhou M, Ye Y, Yan N, Lian X, Bao C, Guo Q. Noninvasive positive pressure ventilator deteriorates the outcome of pneumomediastinum in anti-MDA5 antibody-positive clinically amyopathic dermatomyositis. Clin Rheumatol. 2020;39(6):1919–27. [DOI] [PubMed] [Google Scholar]
- 34.Kono H. Pneumomediastinum in dermatomyositis: association with cutaneous vasculopathy. Ann Rheum Dis. 2000;59(5):372–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Xia J, Gu S, Li M, Liu D, Huang X, Yi L, et al. Spontaneous breathing in patients with severe acute respiratory distress syndrome receiving prolonged extracorporeal membrane oxygenation. BMC Pulm Med. 2019;19(1):237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Okamoto S, Tsuboi H, Noma H, Tabuchi D, Sugita T, Nishiyama T, et al. Predictive factors for pneumomediastinum during management of connective tissue disease-related interstitial lung disease: a retrospective study. Intern Med. 2021;60(18):2887–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yoshida K, Kurosaka D, Kingetsu I, Hirai K, Yamada A. Pneumomediastinum in dermatomyositis itself is not a poor prognostic factor: report of a case and review of the literature. Rheumatol Int. 2008;28(9):913–7. [DOI] [PubMed] [Google Scholar]
- 38.Wang L, Lv C, You H, Xu L, Yuan F, Li J, et al. Rapidly progressive interstitial lung disease risk prediction in anti-MDA5 positive dermatomyositis: the CROSS model. Front Immunol. 2024;15:1286973. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.