A retrospective multicenter cohort study was conducted in the Department of Respiratory and Critical Care Medicine of Affiliated Hospital of Yangzhou University, Jiangdu People’s Hospital, Nanjing Drum Tower Hospital Group Yizheng Hospital and Gaoyou People’s Hospital between January 2024 and May 2025. The study was approved by the local ethics committees (2025-YKL02-(K01)).

The patient inclusion criteria were as follows: (1) patients with age ≥18 years; (2) patients with a positive influenza polymerase chain reaction (PCR) from nasopharyngeal swab or bronchoalveolar lavage fluid (BALF); (3) patients had pulmonary infiltrates on chest imaging. The diagnostic criteria for invasive pulmonary aspergillosis (IPA) were based on the EORTC/MSGERC (2020), IDSA (2016), and FUNDICU (2024) guidelines (Bassetti et al., 2024; Donnelly et al., 2020; Patterson et al., 2016). Proven IPA needs a positive Aspergillus in sterile body fluids or tissues. The probable IPA requires the combination of (1) host factors; (2) clinical symptoms; (3) chest CT showing lesions, infiltrative shadows, or cavities and (4) microbiological evidence, such as positive for Aspergillus culture or PCR, single Galactomannan (GM) test ≥1.0, or single serum GM ≥0.7 with BALF GM ≥0.8. The IAPA patients were reviewed and consensus was achieved by two senior pulmonologists according to the host risk factors, clinical symptoms, chest imaging, laboratory test, and response to treatment. Patients were excluded if the clinical information was insufficient. Non-IAPA were the patients with influenza but without pulmonary aspergillosis.

Respiratory co-infection was defined as (1) detection of clinically relevant bacterial or fungal pathogen in BALF or positive blood culture which was treated using antibiotics or antifungals, respectively; (2) a positive detect for Cryptococcal capsular polysaccharide in blood, or positive GM in serum or BALF.

All statistical analysis was carried out with SPSS 26.0 software (Chicago, USA). The numerical measurements were given as mean ± standard deviation. Receiver operating characteristic (ROC) curve analysis was performed and the areas under the working characteristic curve (AUC) values were displayed, including a 95% CI. An independent sample t-test was employed to analyze the differences between groups and the Chi-square test was used to compare counting data. Odds ratio (OR) was calculated for the baseline data that differed between the two groups. Two-tailed probability values of < 0.05 were considered statistically significant. The pie chart and bar chart were made in Excel spreadsheet (version 2010).

This study enrolled 140 patients which were stratified into two groups: Non-IAPA (n=98) and IAPA (n=42). Baseline characteristics and clinical features between groups were systematically compared, and no significant differences between both groups were observed in age, sex, influenza subtype, or most comorbidities, except for diabetes ( Table 1 ). ​​Diabetes was significantly more prevalent in IAPA patients (35.7%, 15/42) than that in Non-IAPA patients (17.3%, 17/98; p = 0.018, OR=2.65). Other comorbidities, including ​​COPD​​, ​​cardiovascular disease, ​​chronic kidney disease, and ​​solid cancer​​, showed no significant differences between groups. Rare conditions such as ​​hematologic malignancy​​ and ​​hepatitis were similarly distributed. ​​Lymphopenia​​ at influenza diagnosis was remarkably higher in IAPA patients (66.7%, 28/42) compared to Non-IAPA patients (39.8%, 39/98; p = 0.004, OR=3.03). ​​​​Chemotherapy​​ exposure was rare and limited to Non-IAPA patients. Recent ​​corticosteroid use​​ (within 28 days) was 4.8% (2/42) of IAPA patients and 3.1% (3/98) of Non-IAPA patients. Neutropenia​​ was rare in both groups. ​​CURB-65 scores​​, indicating pneumonia severity, were comparable between groups. Influenza A​​ was the dominant subtype in both groups, while ​​Influenza B​​ was less common. The baseline data indicated that Diabetes​​ and ​​lymphopenia​​ emerged as significant risk factors for IAPA, with Diabetes prevalence doubling in the IAPA group.

Non-specific respiratory symptoms such as cough, sputum production, fever, hemoptysis, chest pain, or dyspnea were present to varying degrees in both IAPA and Non-IAPA patients and did not have differential significance (Toda et al., 2021). It is worth noting that most IAPA patients in the present study do not show improvement in their respiratory symptoms before receiving antifungal treatment, or their symptoms worsen after a period of improvement. In this study, five to seven days after admission, patients who still had worsening respiratory symptoms with or without fever received chest CT examination timely. If IAPA or other con-infection was suspected, serum 1,3-β-D-glucan (G) test, GM test and bronchoscopy examination were performed. The obtained BALF samples were immediately subjected to microbiology rapid on-site evaluation (M-ROSE), including Gram staining, fungal smear, and acid fast staining. Culture, GM test and targeted next generation sequencing (tNGS) in BALF were also performed. If fungal staining of BALF reveals mold or GM test was positive and the probable IPA diagnosis was reached after evaluation, immediate antifungal treatment was initiated. If Aspergillus is detected in BALF samples through tNGS test or identified as Aspergillus after culture, and the probable IPA diagnosis was reached after evaluation, antifungal treatment will also be initiated. The criteria for elevated GM in serum or BALF: single sample ≥ 1.0, or serum ≥0.7 and BALF ≥0.8 (Donnelly et al., 2020).

IPA often appears as wedge-shaped lung consolidation, multiple small nodules, masses, and multiple small patches on chest CT scans, and the most distinctive features are wedge-shaped consolidation shadows and small nodular edges with “halo sign”. These radiographic features are frequently nonspecific and can change as the disease progresses (Alexander et al., 2021). In this study, there was no difference in the distribution of lung lobes (unilateral, bilateral) between IAPA patients and the control group. The number of nodules and cavities in the IAPA group was significantly higher than that in the Non-IAPA group, although there were also more halo signs in IAPA patients, the difference was not statistically significant ( Table 2 ; Figure 1 ). It is worth noting that the lung imaging of the same IAPA patient may have different manifestations at different stages of the disease course. As shown in Figures 2A–C , an IAPA patient’s right lung lesion initially presented as ground glass opacities, then developed into subpleural wedge-shaped opacities, and later became solid nodular opacities. In addition, there may not be only one chest CT manifestation in the same IAPA patient, and multiple imaging forms often coexist, such as cavities and consolidation shadows ( Figure 2D ), consolidation shadows with or without inflation sign ( Figure 2E ). When influenza patients present with unexplained imaging manifestations of viral infection in the lungs, timely and comprehensive examinations should be conducted to determine whether there is IPA, in order to avoid missed diagnosis.

Early and rapid diagnosis of IPA has always been a goal pursued by clinical workers, but so far, there has been no breakthrough due to operability and technical difficulties. In this study, we used a microbiological rapid on-site evaluation (M-ROSE) method to perform wet slide examination of lower respiratory tract specimens, which can quickly provide some special and important information, especially for the rapid diagnosis of lower respiratory tract fungal infections (invasive aspergillosis, mucormycosis, etc.). The presence of septate hyphae resembling antlers suggests an aspergillosis infection ( Figure 3 ), while the absence of branches or few branching aseptate hyphae may indicate a mucormycosis infection. Table 2 shows that the probability of detecting hyphae using M-ROSE technology in IAPA patients reached 35.7% (15/42), significantly higher than that in the Non-IAPA group. Moreover, these patients who were detected hyphae in BALF under the microscope and were preliminarily judged to be Aspergillus based on morphology were ultimately confirmed by culture or tNGS results, demonstrating extremely high specificity. Therefore, based on clinical symptoms and chest imaging, patients suspected of having IPA should initiate antifungal treatment if their BALF was highly suspected of having Aspergillus through M-ROSE by experienced laboratory experts. And then further evaluation should be conducted based on tNGS, culture or histopathology results.

The GM test plays a critical role in the diagnosis of IPA (Patterson et al., 2016; Sarrafzadeh et al., 2010). In immunocompromised patients, serum GM test demonstrates moderate sensitivity (69.2%) and specificity (72.2%) for IPA (Sarrafzadeh et al., 2010). A previous study showed that BALF GM test outperforms serum GM, with sensitivity reaching 91.7% and specificity 92.5 (Wu et al., 2021). In the present study, the sensitivity and specificity of serum GM for IAPA diagnosis were 35.7% (15/42) and 93.9% (92/98), respectively. As for BALF GM, the sensitivity and specificity were 71.4% (30/42) and 91.8% (90/98), respectively ( Table 2 ; Figure 1 ). These results showed that BALF GM test significantly increased diagnostic sensitivity of IAPA.

Microbial culture is crucial for diagnosing pathogens, however, the positive rate of Aspergillus culture is relatively low. Niu et al. reported that in patients with acute exacerbation of chronic obstructive pulmonary disease complicated with IPA, the positive rate of Aspergillus culture was only 29.0% (Niu et al., 2024), while it was even lower to 9.5% (4/42) in IAPA patients in the present study ( Table 2 ; Figure 1 ).

In the EORTC/MSGERC (2020) guideline, PCR technology was first recommended for the diagnosis of IPA (Donnelly et al., 2020). With the advancement of gene sequencing technology, many detection techniques based on PCR technology such as tNGS have made significant progress. The tNGS can detect dozens to hundreds of known pathogenic microorganisms in test samples through the combination of ultra-multiplex PCR amplification and high-throughput sequencing techniques (Wang et al., 2025). It showed in the present study that the sensitivity and specificity of tNGS for IAPA diagnosis were 100% (42/42) and 86.7% (85/98), respectively ( Table 2 ; Figure 1 ).

Previous studies showed that co-infections were common in patients with influenza (Al-Dorzi et al., 2024; Fu et al., 2025). Similar to previous studies, bacterial co-infection accounts for the vast majority of co-infected pathogens in patients with influenza in this study. Bacterial co-infection​​ rates were numerically higher in IAPA patients (59.5%, 25/42) but did not reach statistical significance compared to Non-IAPA patients (49.0%, 48/98; p = 0.252). Different from those studies, we reported that atypical pathogens​​ (e.g., Mycoplasma, Chlamydia, Legionella) were frequently identified in the present study and they were more in IAPA patients (16.7%, 7/42) than those in Non-IAPA patients (5.1%, 5/98; p = 0.025). ​​Mycobacterium tuberculosis, were relatively rare and there was no significant difference between the two groups. Rare co-infections, such as ​​pulmonary mucormycosis​​ and Pneumocystis jirovecii, showed no significant differences. ​​Unknown pathogens​​ were less common in IAPA patients (14.3%, 6/42) than in Non-IAPA patients (30.6%, 30/98; p = 0.043) ( Table 3 ; Figure 4 ).

For the treatment of IAPA, voriconazole was chosen as first-line drug (78.6%). For patients with severe liver or kidney dysfunction or intolerance, isavuconazole was given for treatment. In the IDSA (2016) guideline, the treatment duration for IPA should be continued for a minimum of 6–12 weeks, depending on the degree and duration of immunosuppression, site of disease, and evidence of disease improvement (Patterson et al., 2016). However, the average antifungal treatment duration (32.2 ± 4.8 d) in the present study was obviously shorter than that mentioned in the IDSA guideline ( Table 4 ). These results suggest that IAPA patients improved faster than those with hematological disorders (Lanternier et al., 2020), and therefore the treatment duration for IAPA patients was shorter. As shown in Table 4 , stay in hospital and 60-d mortality in IAPA group were greatly higher than those in Non-IAPA group, which suggested that co-infected with aspergillosis significantly increased stay in hospital and mortality.

Neutrophil-to-lymphocyte ratio (NLR) is calculated using the absolute counts of neutrophils and lymphocytes obtained from routine complete blood counts. It is an inflammatory index reflecting systemic inflammatory cascades, which was considered to be a novel biomarker in patients with inflammatory diseases (Wu et al., 2024). A previous study showed that NLR is an important indicator of systemic inflammation (Zahorec, 2001). Whether NLR could predict IAPA prognosis remains unclear. As shown in Table 4 , the average NLR in IAPA group was significantly higher than that in Non-IAPA patients. The ROC curve analysis result showed that the AUC was 0.95 (95% CI, 0.87–1.0), which indicated a high diagnostic accuracy ( Figure 5 ). According to the optimal Youden index in ROC curve (Yang et al., 2025), the optimal cutoff value of NLR was 17.70, corresponding to a sensitivity of 0.87 and a specificity of 0.97. These results indicated that NLR has significant value in predicting the prognosis of IAPA.