In this single-center case–control study, we collected clinical data from 78 patients with DM in our department. We compared the DM patients with ILD with those without ILD to find the similarities and differences in clinical characteristics between the two groups.

Our study showed that the mean age of the overall patients with DM was 55.9 years. Among them, male patients accounted for 37%, while patients with CADM accounted for 32%. These data are similar to those reported by Bendewald et al. [10]. Gottron's papules and heliotrope rash were the most common skin lesions, affecting 64% and 65% of patients, respectively. In the comparison of the two groups of patients, we found that DM patients with ILD were older, with higher rates of Gottron's papules, mechanic’s hands, CADM, myocardial involvement and death, and lower rates of muscle weakness and heliotrope rash. These characteristics may indicate potential risk factors for such patients. Cao et al. found that DM patients with ulcerative Gottron's papules have an increased risk of ILD [9], which is similar to our findings.

A large number of previous studies have shown that DM is associated with a higher risk of malignancy [11,12,13,14,15,16,17]. Population-based retrospective studies have shown that the prevalence of DM with malignancy is about 20% [18,19,20,21]. Interestingly, DM patients associated with ILD have a significantly reduced risk of malignancy [22, 23]. In this study, we also observed that all four patients with malignancy were DM patients without ILD. However, the protective mechanism of ILD against malignancy remains unclear. Future research is expected to reveal its mechanism.

According to the current evidence, systemic glucocorticoids are still the basis for initial treatment of dermatomyositis related muscle diseases [24]. The initial dose of prednisone is usually 0.5 mg/kg/day. For patients with severe muscle involvement or progressive ILD, 500-1000 mg/day × 3 days of methylprednisolone pulse therapy can be considered [25, 26]. In addition, antimalarial drugs, MMF, methotrexate and other immunosuppressive drugs can be considered according to the different conditions of patients [27], while immunoglobulin can be used to treat refractory DM after the failure of other drugs [28]. Notably, MMF is currently considered the preferred first-line agent for DM with ILD [27]. In this study, most patients used glucocorticoids and immunosuppressants, which is consistent with the current recommendation. In the comparison between the two groups of patients, the five patients who died were all DM patients with ILD. This suggests that DM patients with ILD may have more serious disease and need more active immunosuppressive therapy.

In terms of autoantibodies, we found that the positive rate of ANA was the highest in both groups. In addition, the positive rates of anti-SSA/Ro52 and anti-MDA5 antibodies increased significantly in patients with ILD. Many studies have shown that anti-MDA5 DM usually present with CADM, which is significantly associated with ILD and RP-ILD [4, 29, 30]. In this study, we also found that the positive rate of anti-MDA5 antibody and the rate of CADM were higher in DM patients with ILD, which was consistent with previous study. Recently, Xu et al. found that anti-SSA/Ro52 antibody is prevalent in anti-MDA5 DM, and is associated with poor prognosis [31], which is consistent with our results. Therefore, CADM patients with positive anti-SSA/Ro52 and anti-MDA5 antibodies are associated with more aggressive clinical manifestations [31].

In addition, we also found that DM patients with ILD had higher positive rates of anti-SSA/Ro60, anti-JO-1, and anti-EJ antibodies. Anti-aminoacyl-transfer RNA synthetase (ARS) antibodies includes anti-Jo-1, anti-PL7, anti-PL12, anti-EJ, anti-OJ, anti-KS, anti-Zo, and anti-Ha/YRS antibodies. Patients with positive antibodies of these antibodies are called anti-synthase antibody syndrome, which is a special phenotype of DM [32]. Patients with anti-synthetase syndrome are usually associated with symptoms such as fever, ILD, arthritis, Raynaud's phenomenon, Gottron's papules, and mechanic's hands. In our study, the positive rates of anti-JO-1 and anti-EJ antibodies were higher in patients with ILD, which is consistent with the clinical features of anti-synthetase syndrome.

In terms of serological indexes of the two groups, we found that ALB and PNI were significantly reduced in patients with ILD, which may be related to the severity of the disease and long-term malnutrition. In addition, we also found that the increase of serum ferritin was a valuable feature of patients with ILD, while the two groups were similar in ALT, AST, LDH, α-HBDH, CK, ESR, CRP, NEUT, LYM and PLT. In recent years, elevated serum ferritin has been considered as one of the risk factors for poor prognosis in anti-MDA5 DM [33, 34]. High levels of serum ferritin are associated with macrophage activation syndrome, which is commonly used to refer to a secondary hemophagocytosis observed in rheumatic diseases [35, 36]. The mechanism mainly includes lack of regulation of T lymphocytes and excessive cytokine (such as TNF-α, IL-β, IL-6, IL-18, etc.), leading to the activation of macrophages [11, 36]. However, the role of serum ferritin in the pathogenesis of DM remains unclear. Recently, a study developed and validated a scoring model called FLAIR to predict mortality in CADM-ILD patients [37], in which serum ferritin ≥ 636 ng/ml was a risk factor. In our study, four of the five patients who died with ILD had serum ferritin > 2000 ng/ml, and one patient had serum ferritin of 752 ng/ml. This further supports the importance of serum ferritin in predicting poor outcomes. However, serum ferritin lacks specificity in diagnosis.

Our multivariate logistic regression showed that old age, Gottron’s papules and anti-SSA/Ro52 were independent risk factors for ILD in DM. This result was consistent with previous studies [9, 31]. Some variables that may have clinical significance, such as malignancy, myocardial involvement, ALB, have no statistical significance in multivariate logistic regression. This may be related to the small sample size.

We also focused on the data of five patients who died and found that they had some common clinical features. All the five patients had ILD and old age, with high positive rates of anti-MDA5 and anti-SSA/Ro52 antibodies, and significantly increased serum ferritin. This is basically consistent with previous studies [31, 37]. In addition, we also found that these five patients had a high positive rate of Epstein–Barr virus infection. Viruses may play a role in triggering immune activation in DM [38]. However, some scholars tried to isolate the specific virus from DM muscle tissue without success [39]. Future research may further reveal the relationship between viruses and DM.

Those patients with DM specific skin lesions but without myopathy are called CADM, which is currently considered as a subtype of DM [2]. CADM is usually diagnosed after 6 months of absence of myopathy [40]. In this study, we found that the percentage of muscle weakness in patients with CADM was significantly lower, which was consistent with the characteristics of the disease itself. In addition, we also found that patients with CADM had a lower rate of previous treatment, a shorter course of disease, a higher rate of RP-ILD, and a higher positive rate of anti-MDA5 antibody. The characteristics of low rate of previous treatment and short course of disease suggest that patients with CADM are more acute and severe than those with classical DM. This verified that anti-MDA5 DM usually presented with CADM and had a higher rate of RP-ILD [4, 29]. These findings suggest that clinicians should pay more attention to patients with CADM.

There are some limitations to our study. First, the total number of patients included in this study is small. Secondly, the generality of the results may be limited by the fact that this was a single-center study conducted in our hospital. Finally, some patients with ILD were severely ill and unable to perform lung function test, and most of our patients were not tested for lung function due to their poor compliance or physician omissions in ordering test. Therefore, the results of lung function test are not presented in our article. Despite these limitations, our study still revealed some distinct clinical features of DM patients with ILD, which has important guiding significance in clinical practice.