To the best of our knowledge, although this is a single-center retrospective cohort study with a small sample size, this is the first study to systematically explore the clinical characteristics and prognosis of IIM patients with fever in the initial stage. Previous studies on fever and disease prognosis mainly focused on traumatic brain injury [12] and malignant tumors [13]. However, fever was also a general problem for hospitalized patients with IIM. The differential workup of fever remains a special challenge for clinicians, particularly in the context of the initial diagnosis of IIM and also in the course of the disease in IIM patients. In the present study, the fever group showed no response to antimicrobial treatment or no evidence of pathogenic origin, including clinical signs and etiologic tests. Body temperature was restored to a normal level after they were treated with glucocorticoid and/or IS. Due to this fact, the initial fever of IIM patients was mainly attributed to the disease activity. Our study revealed that patients with fever at initial diagnosis have unique clinical features and fever was a risk predictor for poor outcomes of IIM.

Fever was common (35.4%) in our cohort, similar to the incidence that 34.9% (67/192) of feverish patients with IIM in Zhejiang, China. Although the IIM type and MSA (ASS and anti-MDA5) were not significantly different between the two groups, our cohort is consistent with previous findings [14] showing that fever seemed more common in CADM, ASS, and anti-MDA5(+) IIM. One study [15] reported that ~ 60% of ASS experienced one or more febrile episodes during an average follow-up of 5 years, which is higher than in our cohort, but in the present study, we focused on the initial stages of febrile episodes.

As a frequent extra-muscular manifestation, ILD leads to increased mortality in patients with IIM [2, 3]. The frequency of ILD in patients with IIM has been widely reported (8.6–85.6%), and a meta-analysis of 23 studies revealed that 834/2079 patients with IIM (40.1%) had ILD [16]. In the present cohort, We identified a slightly higher rate of ILD, especially in the fever group, than that has been reported in previous studies. The underlying mechanism between fever and ILD in IIM is not clear, maybe suggest both represent components of profound systemic inflammation. Previous studies also reported that the prevalence of MH ranged from 5 to 56% in given IIM populations [17]. Importantly, MH has been already a risk factor for ILD [18]. Consistent with ILD, MH occurred more frequently in the fever group [19]. In addition to the above, the level of muscle enzyme in the fever group was much lower than in the control group. A growing body of evidence suggests that the clinical manifestations of IIM are a spectrum in which at one end, patients have prominent muscle symptoms, higher CK levels, and less extra muscular involvement, whereas, at the other end, patients have fewer muscular symptoms, modest increases in CK levels, and more skin and pulmonary involvement [20]. Our study shows patients with fever are more likely at this “less muscular” end, exhibiting lower CK levels and an increased risk of ILD.

We also found the fever group has higher levels of CRP and ESR, but lower albumin, suggesting a hyperinflammatory state in these patients. Among these serological parameters, the ratio of ESR to CRP was used for distinguishing flare and infection in SLE feverish patients [21]. Recently, it has been proposed that hypoalbuminemia and elevated CRP values independently predicted 30-day mortality [22].

Our study assayed the prognosis of fever in the initial stage from multiple perspectives, including the incidence of AE-ILD, relapse, and all-cause mortality. In our cohort, 53.6% of the fever group exhibited AE-ILD, remarkably much more than 13.7% of the non-fever group. The increased all-cause mortality rate was also observed in the fever group. Since age at onset and treatments were different between these two groups as previously described, forward logistic regression multivariable analysis was used to provide further evidence that fever in the initial stage was the strongest independent factor in poor prognosis, both AE-ILD and all-cause mortality. The result was consistent with a previous report by Chanyuan Wu et al. showing that fever was more common in the in-hospital death group [23].

There were several limitations of this study. The major limitations of this study were the retrospective single-center design and the limited number of patients. Larger, independent, multicenter studies ideally covering different ethnic populations are mandatory to thoroughly evaluate the prognostic value of fever in the initial stage. Second, fever is a common symptom of many rheumatic diseases. Clinically, it is difficult to identify the cause, although feverish patients with infection are excluded based on clinical manifestations and etiological examination. Third, Although these patients were treated with glucocorticoid combined with IS, the specific treatment regimens (e.g. the glucocorticoid protocols and IS treatments) were not consistent.