Prevalence, Risk Factors, and Mortality of Invasive Pulmonary Aspergillosis in Patients with Anti-MDA5+ Dermatomyositis: A Retrospective Study in China

Introduction

Aspergillus species are ubiquitous fungi that cause a broad spectrum of diseases in humans. One of the most severe forms is invasive aspergillosis, which primarily affects the lungs and is known as invasive pulmonary aspergillosis (IPA).¹ IPA is encountered frequently in individuals with compromised immune systems, particularly those with severe or prolonged neutropenia, deficiencies in cell-mediated immunity, hematologic malignancies, recipients of hematopoietic stem cell or solid organ transplants, and patients undergoing chemotherapy.^2,3 Globally, the number of reported cases of invasive aspergillus is greater than 300,000 annually,⁴ with a high mortality rate ranging from 30 to 80%.^5–8 This trend is expected to continue due to the widespread use of immunosuppressants and biologics, which increase the risk of fungal infections.

Anti-melanoma differentiation-associated gene 5 positive dermatomyositis (anti-MDA5+ DM) is a distinct subtype of idiopathic inflammatory myopathy (IIM), with a particularly high prevalence in East Asia. It is characterized by a hallmark rash of dermatomyositis, absent or minimal muscle involvement, and prominent interstitial lung disease (ILD).⁹ Recently, the occurrence of opportunistic infections, particularly fungal infections, has been reported to be prevalent in patients with anti-MDA5+ DM, and represents an additional risk factor for mortality in addition to rapidly progressive ILD (RPILD).^10,11 However, the occurrence of IPA in anti-MDA5+ DM remains poorly understood, with only a small number of published studies available on this subject. Baray et al¹² reported a case of an anti-MDA5+ DM patient who developed IPA and experienced rapid progression to respiratory failure, ultimately resulting in death. The absence of large-scale clinical studies has limited the understanding of the diagnosis and treatment of IPA in patients with anti-MDA5+ DM. Therefore, the aim of the current study was to investigate the prevalence, risk factors, and prognosis of IPA in anti-MDA5+ DM patients treated at our center, with the objectives of achieving early identification of high-risk individuals, timely intervention, and enhancement of patient outcomes.

Methods

Study Population

A retrospective analysis was carried out on patients diagnosed with anti-MDA5+ DM between January 2016 and March 2023 at the Rheumatology Department of the China Japan Friendship Hospital. The diagnosis of IIM was based on the Bohan and Peter criteria¹³ and subsequently re-validated retrospectively by two experienced rheumatologists, in accordance with either the 2017 EULAR/ACR IIM classification criteria¹⁴ or the 2018 ENMC DM criteria.¹⁵ The myositis-specific antibody (MSA) profile featuring the anti-MDA5 antibody was detected via immunoblotting, according to the manufacturer’s instructions (Euroimmun, Lübeck, Germany). Patients with a concurrent presence of positive anti-MDA5 antibodies and other MSA were excluded from the study. Additionally, individuals diagnosed with other connective tissue diseases or those with concurrent malignant tumors were also excluded. The research protocol was granted approval by the Ethics Committee of the China Japan Friendship Hospital (reference number 2022-KY-156), and the study was conducted in accordance with the Declaration of Helsinki, 2000. Considering the retrospective nature of this investigation, patient consent was waived because their privacy was preserved and medical care was not impacted.

Patients who had previously undergone lower respiratory tract specimen examinations, such as deep sputum, bronchoalveolar lavage fluid (BALF), or lung tissue analyses, to diagnose pulmonary fungal infection were selected for subsequent investigation. These patients were classified according to the European Organization for the Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) criteria, which included either proven, probable, or possible IPA.¹⁶ “Proven” cases required histopathological or cytopathological confirmation, while “probable” cases required the presence of a host factor usually immune suppression, clinical features such as fever unresponsive to antibacterial treatment, and mycologic detection. Patients who met the criteria for a host factor and had clinical characteristics without mycological evidence were classified as “possible” cases. The diagnostic criteria for aspergillus colonization were based on the proposals of Soontrapa et al¹⁷ that included the following aspects: 1. Pathological study compatible with other causes; 2. No evidence of positive galactomannan (GM) and no compatible imaging; and 3. The patient survived for at least 30 days without antifungal treatment. The IPA-positive (IPA+) group only included cases diagnosed with proven or probable IPA, whereas other cases were categorized as the IPA-negative (IPA-) group.

Clinical Data

All patient-related demographic information, laboratory test results, and details of treatment were documented comprehensively. Laboratory data were acquired concomitantly with the examination of lower respiratory tract specimens, encompassing assessments of neutrophil, lymphocyte, CD4+ T-cell, and CD8+ T-cell counts, as well as the erythrocyte sedimentation rate (ESR) and serum levels of albumin, lactate dehydrogenase (LDH), C-reactive protein (CRP), and ferritin. Additional tests were conducted to screen for possible fungal infections, which included serum (1,3)-beta-D-glucan (BDG) level (reference range: < 10 pg/mL), serum GM level (reference range: < 0.5), BALF GM level, tracheoscopy involving the detection of leukoplakia or ulcers on the wall of the airway, and chest imaging that involved the evaluation of ground-glass opacity (GGO), consolidations, cavities, nodules, halo signs, pleural effusion and bronchiectasis.¹⁸ The presence of ILD was assessed by means of chest computed tomography (CT), while RPILD was defined using previously published established criteria.¹⁹

Comprehensive documentation of the patient’s treatment history within the three-month period preceding etiological testing was conducted, which included information on the dose of glucocorticoid (GC) administered on a daily basis (expressed as prednisone equivalents), pulse methylprednisolone therapy (PMT), immunosuppressive agents, and intravenous immunoglobulin (IVIG).

Follow-Up Study

The duration of the follow-up period was determined by measuring the time elapsed from the onset of symptoms to the patient’s death or last investigation/visit. All follow-up evaluations were concluded by March 2023, with the date of the last follow-up or death being documented meticulously for each patient. A 12-month interval was chosen as the reference point for survival analysis, with patients whose follow-up duration fell short of 12 months being categorized as censoring values.

Statistical Analysis

The statistical analysis was carried out using IBM SPSS software (version 23.0, Armonk, NY, USA). The data were expressed as mean ± SD, median (interquartile range), or number (percentage) where appropriate. Comparisons between groups were conducted using independent-sample t-tests or Mann–Whitney U-tests for continuous variables and Chi-square tests or Fisher’s exact tests for categorical variables. Logistic regression was performed in the univariate and multivariate analyses to identify risk factors for the occurrence of IPA and to calculate the odds ratios (OR). The optimal cut-off value was determined by receiver operating characteristic (ROC) analysis, with the Kaplan-Meier method with Log rank testing used to evaluate survival differences. For all the analyses, a two-sided p-value < 0.05 was regarded as statistically significant.

Results

Microbiology

Of the 415 patients diagnosed with anti-MDA5+ DM, 28 (2 proven cases and 26 probable cases) were found to have IPA, representing a prevalence rate of 6.7% (Table 1). The distribution of aspergillus species isolated from the respiratory specimens of patients with IPA is shown in Figure S1. The most prevalent aspergillus species identified was A. fumigatus, accounting for 37.9% of cases, followed by A. flavus 20.7%, A. terreus 10.3%, and A. tubinggensis and A. niger at 6.9% each. In addition, A. versicolor, A. retardis, A. Fischeri, A. nidulans, and A. oryzae were each isolated in only a single patient (3.4%).

Table 1 Aspergillus Species Identified from Respiratory Specimens in Patients with Anti-MDA5+ Dermatomyositis

Patient Characteristics

Of the 415 patients diagnosed with anti-MDA5+ DM, 126 had a lower respiratory tract specimen examined at our center and were included for further analysis. In the study cohort, 28 patients were classified as the IPA+ group, while the remaining 98 patients did not demonstrate any clinical evidence of IPA and were categorized as the IPA- group. A comprehensive analysis comparing the clinical characteristics of these two groups is summarized in Table 2.

Table 2 Baseline and Clinical Characteristics of Anti-MDA5+ DM Patients with and without IPA

This analysis showed that there were no significant differences in demographic characteristics, comorbidities, clinical symptoms, and treatment regimens between the IPA+ and the IPA- groups. The mean age and gender distribution in the IPA+ group were similar to those of the IPA- group (50.5 ± 12.6 vs 48.1 ± 11.4 years, P = 0.337; 50% vs 59.2%, P = 0.386, respectively). However, compared to the IPA- group, the IPA+ group had a significant decrease in lymphocyte count, particularly the CD4+ T-cell count (249.0 × 10⁶/L vs 339.0 × 10⁶/L, P = 0.020), as well as reduced levels of serum albumin (32.3 vs 35.2 g/L, P = 0.021), and an increase serum ferritin levels (1432.3 vs 673.5 ng/mL, P = 0.009).

Different Detection methods for IPA Infection

A comparative analysis of the serum levels of BDG and GM, and tracheoscopy findings in patients diagnosed with anti-MDA5+ DM showed no significant variations between the IPA+ and IPA- groups (Table S1). In the IPA+ group, the positive rates of serum BDG and GM levels and the incidence of leukoplakia and ulcers observed by tracheoscopy were very low, with values of 14.3%, 7.1%, 18.2%, and 4.5%, respectively. In contrast, the levels of GM in BALF were significantly higher in the IPA+ group than that measured in the IPA- group (0.75 vs 0.16, P < 0.001). In addition, the results of the radiological examinations showed that GGO, consolidation, halo sign, cavity, pleural effusion, and other manifestations were similar between the IPA+ and IPA- groups, with the exception of pulmonary nodules being more frequently observed in the IPA+ group (22.2% vs 6.2%, P = 0.022). Representative radiological images are shown in Figure 1.

Figure 1 Representative radiological images of anti-MDA5+ DM patients with IPA. (A) Axial CT image showing a cavitary lesion in a patient with IPA (arrow). (B) Axial CT image of a patient with IPA demonstrating the presence of diffuse micronodules throughout both lungs (arrows). (C) A patient diagnosed with IPA with bilateral pulmonary consolidation and diffuse ground-glass opacities. (D) A patient diagnosed with IPA with bilateral subpleural and bronchial vascular bundle consolidation (arrow).

Risk Factors for IPA Occurrence in Anti-MDA5+ DM

We employed a binary logistic model regression to determine the risk factors associated with the occurrence of IPA (Table S2). Univariate analysis showed that increased levels of serum LDH (OR = 1.003, P = 0.044), serum ferritin (OR = 1.000, P = 0.037), BALF GM (OR = 2.490, P = 0.004), and the presence of pulmonary nodules (OR = 4.333, P = 0.019) were significant risk factors for the occurrence of IPA. On the other hand, an increased CD4+ T-cell count (OR = 0.997, P = 0.034) and higher serum albumin level (OR = 0.889, P = 0.018) were shown to be important protective factors. However, multivariate analysis showed that increased BALF GM levels were the sole independent risk factor after adjusting for confounding variables (adjusted OR = 2.191, P = 0.029). We then performed a ROC analysis to identify the optimal BALF GM level for detecting IPA and demonstrated the cut-off value was 0.585, with the area under the curve (AUC) being 0.779 (Figure S2). As a consequence, a BALF GM value exceeding 0.585 was considered positive, with a sensitivity and specificity of 60.0% and 87.5%, respectively.

Antifungal Treatments and Outcomes

None of the patients in our study had received prophylactic antifungal therapy prior to their diagnosis of IPA. Instead, voriconazole-based antifungal therapy was initiated immediately upon diagnosis in all the patients. Two patients received caspofungin in combination with voriconazole, while one patient received concurrent administration of amphotericin B and voriconazole.

Regarding prognosis, our study showed a 12-month mortality rate of 25.0% (7/28) in the IPA+ group, which was higher than the 15.3% (15/98) observed in the IPA- group. However, this difference between the two groups was not statistically significant. Further analysis of the clinical characteristics of survivors and non-survivors in the IPA+ group showed that non-survivors had a longer interval between disease onset and IPA diagnosis (4.0 vs 1.0 month, P = 0.008), higher incidence of RPILD (100.0% vs 42.9%, P = 0.010), lower lymphocyte count (450 ×10⁶/L vs 840 ×10⁶/L, P = 0.020), more co-infections with Pneumocystis jirovecii (57.1% vs 14.3%, P = 0.043), and a greater number of disease-related complications such as respiratory failure (85.7% vs 33.3%, P = 0.029), ICU admissions (57.1% vs 4.8%, P = 0.008), and mechanical ventilation (57.1% vs 9.5%, P = 0.021) compared to that observed in the survivors (Table 3). The survival curves are shown in Figure 2, with further Supplementary Information available in Table S3.

Table 3 Comparison of Clinical Characteristics Between Surviving and Non-Surviving Patients with Anti-MDA5+ DM and IPA

Figure 2 Kaplan-Meier curves of anti-MDA5+ DM patients with and without IPA. (A) Patients with IPA had a lower survival rate than those without IPA, although this difference was not statistically significant (P = 0.1611). (B) Patients with RPILD in the IPA+ group had a significantly lower survival rate than those in the IPA- group (P = 0.0090). (C) Patients with a CD4+ T-cell count < 200×106/L had lower survival rates compared to those with a CD4+ T-cell count ≥ 200×106/L (P = 0.0211). (D) Patients coinfected with Pneumocystis jirovecii had lower survival rates compared to those without this infection (P = 0.0280). Abbreviations: RPILD, rapidly progressive interstitial lung disease; PJ, Pneumocystis jirovecii.

Discussion

Although the occurrence of IPA in rheumatic diseases has not been studied extensively, its high fatality rate highlights the need for heightened awareness.²⁰ Our investigations showed that the prevalence of IPA in anti-MDA5+ DM was 6.7%, with A. fumigatus being the predominant aspergillus species. While there were no pathognomonic clinical features observed following IPA infection in anti-MDA5+ DM patients, identifying BALF GM in this subgroup has some potential significance as an independent predictor of IPA. Notably, patients with an established IPA diagnosis and co-occurrence of RPILD, decreased lymphocyte count, or a Pneumocystis jirovecii co-infection tended to have worse outcomes. Therefore, early recognition of patients at a higher risk of developing IPA, prompt diagnosis, and timely initiation of antifungal therapy are vital to improve prognosis and ultimately enhance the care of patients with anti-MDA5+ DM.

Individuals with rheumatic diseases exhibit a unique susceptibility to infectious complications, attributable, in part, to inherent immune dysfunction associated with the disease and partially to the administration of immunosuppressive therapies. In our previous investigation, we reported that patients with IIM had an infection frequency of 27.6% and an opportunistic infection rate of 10.4%.¹⁰ Notably, patients with anti-MDA5+ DM, a subtype of IIM, have been shown to be particularly susceptible to diverse infections.¹¹ However, there are only limited reports on the prevalence of IPA in anti-MDA5+ DM. Our study determined that the prevalence of IPA in patients with anti-MDA5+ DM was 6.7%, roughly equivalent to the incidence observed in patients with hematologic malignancies.²¹ Consistent with several published studies,^17,22 we identified A. fumigatus as the predominant aspergillus species in anti-MDA5+ DM patients. However, this finding was different from reports in African nations, where A. flavus and A. niger are the most prevalent aspergillus species.²³ This discrepancy may stem from variations in environmental conditions, geography, climate, and other factors across different regions.

As a consequence of its nonspecific clinical manifestations, IPA presents considerable clinical challenges for early diagnosis in patients with anti-MDA5+ DM. Identification of the risk factors for the occurrence of IPA is therefore crucial. Host factors listed in the EORTC/MSG criteria traditionally serve as classic risk factors for IPA.¹⁶ Emerging research has also identified additional non-classical risk factors such as critical illness, chronic lung disease, biologics, small molecule kinase inhibitors, and viral pneumonia.^3,24 Importantly, patients with anti-MDA5+ DM often exhibit dysregulated immunity in the initial stages of the disease. Our study demonstrated specifically that the IPA+ group had a more substantial reduction in lymphocyte count than the IPA- group, particularly for CD4+ T-cells. Cramer et al confirmed the involvement of CD4+ T-cells in the pathogenesis of IPA by demonstrating that fungus-specific CD4+ T-cells had a significant protective effect against IPA.²⁵ Moreover, Camargo et al proposed that impaired T cell responsiveness and depletion of naive CD4+ T-cells in patients with hematologic malignancies following chemotherapy or hematopoietic stem cell transplantation may be crucial risk factors for the development of IPA.²⁶ However, the precise pathophysiological mechanisms underlying the role of CD4+ T-cells in defending against IPA remain to be fully elucidated.

The timely identification of patients with suspected IPA is therefore imperative, with reliable non-invasive detection methods playing a pivotal role in achieving this goal. Nevertheless, presently available clinical detection methods have certain limitations. Although serum and BALF GM level are recommended as markers for the diagnosis of IPA,²⁷ prior investigations have indicated that serum GM levels have a sensitivity range of 41% to 78%, while BALF GM has a higher sensitivity range of 60%-100%.²⁸ However, our study showed that serum GM level had a low sensitivity in patients with anti-MDA5+ DM, making it unsuitable for IPA screening. Only an increased GM level in BALF qualified as an independent risk factor for the occurrence of IPA. In addition, tracheoscopy and radiological findings do not have sufficient specificity to identify anti-MDA+ DM patients with IPA. The detection of more pulmonary nodules on chest CT scans we observed in the IPA+ group compared to that in the IPA- group was limited by a low sensitivity, precluding their use in IPA screening, but warranting their use for disease surveillance during follow-up. Therefore, we consider that the development of more accurate, effective, and non-invasive IPA screening methods is imperative.

The predominant therapeutic approach for IPA advocates voriconazole monotherapy as the primary treatment modality, with isavuconazole and posaconazole recommended as alternative options in cases where voriconazole is contraindicated. Liposomal amphotericin is proposed as an alternative regimen for patients receiving mold-active azole therapy or experiencing intolerance to voriconazole. Furthermore, echinocandins are acknowledged as promising choices for second-line or salvage therapy.²⁹ Despite adherence to standard treatment protocols, our patients continue to exhibit high mortality rates within the IPA+ cohort.^5–8 In the current study, the IPA+ group had a higher 12-month mortality rate (25%) compared to that of the IPA- group (15.3%), although this difference was statistically insignificant. We speculate that this lack of significance may be attributable to the small sample size in our study. Non-survivors in the IPA+ group had a longer interval between disease onset and IPA diagnosis, suggesting the possibility of delayed diagnosis and treatment. Furthermore, non-survivors in this group had a higher prevalence of RPILD, with severe lymphocyte depletion and co-infection with Pneumocystis jirovecii. As a consequence, clinicians should exercise utmost caution and closely monitor anti-MDA5+ DM patients who present with these unfavorable prognostic factors, and promptly initiate antifungal therapy and other supportive measures.

This study had several limitations that should be acknowledged. Firstly, the study was a single-center retrospective design, and therefore intrinsic bias could not be avoided completely. Secondly, only patients suspected of having fungal infections underwent fungus-related assessments, potentially leading to an underestimation of the prevalence of IPA in anti-MDA5+ DM patients. Also, the challenge of breaking down fungal cell walls may have exacerbated this underestimation by decreasing the detection rate. Thirdly, as lung tissue biopsy is invasive and challenging to implement widely in anti-MDA5+ DM patients, most patients with IPA were classified as probable rather than proven cases. Finally, although the use of antibiotics may correlate with IPA incidence, we did not analyze this association because of the complex nature of their use in anti-MDA5+ DM patients.

Conclusions

While IPA is not rare in patients with anti-MDA5+ DM, early diagnosis remains a challenge due to the limited clinical specificity of symptoms. For patients suspected of having IPA, detecting GM in BALF is valuable for making a timely diagnosis, as this increase is an independent risk factor for IPA occurrence. Furthermore, co-occurrence of RPILD, decreased lymphocyte counts, and co-infection with Pneumocystis jirovecii may serve as potential adverse prognostic factors, with early antifungal therapy in patients with these risk factors possibly improving their prognosis.