The human respiratory tract is exposed to various microorganisms present in the ambient atmosphere including fungi. In most individuals, immune function can prevent fungal growth and invasion, but a growing number of immunocompromised patients and the widespread use of antibiotics has led to an increase in the incidence of pulmonary fungal infections (Limper et al., 2011; Denning and Chakrabarti, 2017; Li et al., 2021). Generally, fungal infections are more common in the elderly. However, one report provided evidence that the incidence of fungal infections in hospitalized patients aged 14–30 years has shown a clear upward trend from 2013 to 2019 (Li et al., 2021). Invasive fungal infections are associated with high mortality, with mortality rates as high as 67% being reported among patients with acute infections (Chen et al., 2001). Clearly, pulmonary fungal infections represent a significant clinical and financial burden to the medical community.

To date, few studies have reported the incidence of pulmonary co-infection of fungi and bacterium or the species of co-infecting bacteria. One report described fungal and bacterial co-infection in burn wounds (de Macedo and Santos, 2005). Presently, the clinical treatment for pulmonary fungal infections usually consists of combining antifungal and antibacterial drugs (Guan et al., 2020). In the absence of a complete understanding of fungal and bacterial co-infection, empirical addition of antibacterial drugs could enhance the risk of drug resistance. Furthermore, the non-specific symptoms of pulmonary co-infections, the limitation of diagnostic methods, and the long, time-consuming traditional detection methods with poor detection rates have hampered efforts to distinguish fungal and bacterial co-infection from other pathogenic infections.

Given the limitations of current pathogenic diagnosis, a rapid and accurate alternative method is imperative. Metagenomic next-generation sequencing (mNGS), also known as high-throughput sequencing, is a promising technique that can be used for rapid identification of infectious pathogens without the need for culture, and with greater sensitivity than traditional culture methods (Gu et al., 2019; Wang et al., 2019). Moreover, mNGS furthers the concept of “precision diagnosis and treatment,” as it can be used for multidisciplinary infection diagnosis. According to one report on peripheral pulmonary infectious lesions, the pathogen detection rate via mNGS is nearly 89%, which is significantly higher than that via traditional detection methods (Huang et al., 2020). Identifying pathogens at an early stage may change the outcome of the disease, but traditional methods based on microbiological and biochemical features are time-consuming and have low detection rates (Petrucelli et al., 2020). Therefore, herein, we used mNGS technology to identify pathogens from lung biopsy samples of patients with fungal infections, or bacterial and fungal co-infections, comparing results to those obtained by traditional methods, and assessing the type and prevalence of the species found to better understand and treat these infections.

Current evidence on the incidence of pulmonary fungal infections accompanied by other pathogenic infections is sparse, with most studies focusing on the relationship between Aspergillus and Pseudomonas aeruginosa infection (Briard et al., 2015; Briard et al., 2016). Our results showed that 40.3% of patients with pulmonary fungal infections had fungal and bacterial co-infections. The clinical manifestations of fungal and bacterial co-infections are more serious than those of fungal infections alone. Here, we only studied the co-existence of bacteria and fungi; however, we do not know which microorganism appeared first in the pulmonary infections. Some studies have reported that A. fumigatus colonization is preceded by fungal infection in patients with chronic obstructive pulmonary diseases (COPD) and cystic fibrosis (Borman et al., 2010; Gago et al., 2019).

Microorganisms do not exist alone, and invariably exhibit direct or indirect communication among themselves (Kolwijck and van de Veerdonk, 2014). Compared with a single pathogen infection, a differential immune response is mounted when patients are infected with multiple pathogens, and this may lead to different clinical outcomes. At present, relatively few studies have focused on the interaction between fungi and bacteria, especially in the respiratory system. Prior studies were mostly performed in vitro and in animal models, and the results typically demonstrated that bacteria had an inhibitory effect on fungi (Ferreira et al., 2015; Sass et al., 2018). For example, in a murine model of pulmonary disease, immunosuppressed mice with A. fumigatus and P. aeruginosa co-infection had a higher survival rate than animals infected with A. fumigatus alone. However, our result shows that the proportion of patients with bacterial-fungal co-infection is as high as 40.3% among patients with pulmonary fungal infections, which may suggest that interspecies microbial interactions have a positive correlation.

In our study, among the eight types of fungus isolated, Aspergillus, Pneumocystis, and Rhizopus were the most prevalent. However, results of another study analyzed the pathogens in 1,644 in-patients with pulmonary fungal infection and found Aspergillus, Cryptococcus, and Talaromyces marneffei as the three most common fungi (Li et al., 2021). Comparing their study with ours, only the Aspergillus was commonly observed in patients with bacterial and fungal co-infections, which may be ascribed to the different geographical distribution and the considerably larger number of cases in the earlier study. For example, Talaromyces is epidemic in mountainous/tropical regions in south east Asia, which is likely not present in patients accepted at the location examined in this study presuming no such travel history.

Fungal infections commonly occur in immunocompromised individuals. However, in our study, a large proportion (19.3%) of patients with pulmonary fungal infections had normal immune function. An increasing number of reports show that immunocompetent patients suffer from fungal infections, probably due to environmental exposures, genetic factors, or structural pulmonary risk factors (Denning and Chakrabarti, 2017; Latgé and Chamilos, 2019). Immunocompromised status was nevertheless a significant risk factor for co-infection of bacteria and fungi, which is consistent with previous studies (Zhou et al., 2020). The presence of pulmonary cavity lesions also increased susceptibility to co-infection. Pulmonary cavitation is a relatively serious pulmonary affliction, which may cause massive pulmonary hemorrhage and affect the ability of drugs to reach the diseased site, leading to ineffective treatment (AlShanafey et al., 2019). Antibiotic treatment has been shown to lead to fungal dominance (Jain et al., 2021), however, it is not identified as a risk factor in this study, which is likely due to the patient’s short-term treatment. CT cavitary images may be used as the standard for comparing the severity of infection in cases of fungal infection to those with fungal and bacterial co-infection. In our study, the proportion of pulmonary cavitation was significantly higher in the co-infection group than in the fungal infection group. Hence, we concluded that the combined bacterial infection may aggravate the disease, which is consistent with current findings in humans. A retrospective cohort study showed that compared with P. aeruginosa infection alone, A. fumigatus and P. aeruginosa co-infection caused more rapid decline in patients with lung function and worsened clinical outcomes (Amin et al., 2010). Another study suggested that S. maltophilia and P. aeruginosa infections are associated with a higher probability of concurrent Aspergillus infection, which exacerbates clinical manifestations (Granchelli et al., 2018). These findings explain, to some extent, the diversity of interactions and clinical effects observed in fungal and bacterial co-infections. Similarly, the diversity of interactions between microorganisms also partially explains why some clinical treatments against pathogenic bacteria fail. In our study, the appearance of pulmonary cavitation was not common. There are numerous unknown factors surrounding these observations, and further research is required to fully understand the mechanisms at play.

We did not find a difference in the distribution of bacterial species in people with normal and defective immune functions. However, since we only studied the distribution of bacterial species in the short term after the use of antibacterial drugs, we do not know the long-term effects. For fungal and bacterial co-infections, 17 types of bacteria were identified, almost all of which were hospital-associated pathogens. P. aeruginosa, the most commonly isolated bacterial species in this study, is a gram-negative opportunistic pathogen existing in diverse environmental settings that harbors multi-drug resistance and poses serious therapeutic challenges (Lister et al., 2009; Leroy et al., 2021). Interestingly, four strains of Mycobacterium tuberculosis were isolated in pulmonary bacterial and fungal co-infection, of which three were co-infected with Aspergillus and one with Rhizopus. A study once reported 8 cases of Cryptococcus and Mycobacterium tuberculosis co-infections, and most of them occurred in the brain (Chen et al., 2016). Based on statistical analysis and the description of the bacteria likely to co-infect with fungi in patients with pulmonary fungal infection, our findings will be useful for future clinical management of pulmonary fungal diseases.

The identification of co-infection of fungi and bacteria had a notable impact on clinical treatment. Fungal infections are treated with only antifungal treatment, while antibacterial and antifungal treatment is offered when patients are co-infected. In our study, a large proportion (66.4%) of patients had received empirical antibiotic treatment before their pathogenic test results were available. After the etiological results became known, 40.5% of the previously used empirical antibacterial drugs were discontinued or changed, since patients without bacterial co-infection may have other diseases, necessitating the continued use of antibacterial drugs. Anti-fungal or antibiotic treatment can skew the composition of the species that colonize the host, which consequently show less diversity. It is quite common for immunocompromised patients to be prescribed preventive antibiotic use.

mNGS is an unbiased approach for sequence identification of pathogenic microorganisms. One study revealed that mNGS is advantageous for the evaluation of fungal infections and suggested that mNGS combined with smear analyses should be used as a routine diagnostic tool for identifying invasive fungal infections (Li et al., 2018). We used mNGS combined with conventional laboratory-based diagnostic testing to gain more accurate and comprehensive information related to co-infecting microbes in patients with pulmonary fungal infections. mNGS offers the important advantage of detecting infectious pathogens, since it is less affected by prior antibiotic exposure.

Anti-fungal or antibiotic treatment can change the composition and diversity of the microbial species present in the body. As we only obtained each patient’s medication history for one month before diagnosis, we cannot be sure of the impact their earlier medication history might have had on the test results.

The incidence of fungal and bacterial co-infections is considerable in our mNGS-based study. Our research results should lead to more rational and precise anti-infective treatments, especially for patients who are difficult to diagnose by conventional methods, thereby having a positive impact on assessment of risk and clinical outcomes in these cases.

The data presented in the study are deposited in the NCBI and EMBL-EBI repository, accession number PRJNA773581 and PRJEB48166, respectively.

The studies involving human participants were reviewed and approved by the Ethics Review Committee of Tianjin Medical University General Hospital. The patients/participants provided their written informed consent to participate in this study.

YW and JF contributed to the research design and revision of the manuscript. ZZ drafted the research protocol, analyzed the results, and drafted the manuscript. JS conducted data analysis and assisted in writing the manuscript. CY and LY conducted data acquisition and analysis. JC and XL contributed to the data analysis and data deposition in an acceptable repository. All authors approved the submitted version and agreed to be responsible for all aspects.

This research was supported by grants from National Science and Technology Major Project of China (No.2018ZX10305409-001-001), National Natural Science Foundation of China (81970083, 81270144, 81570084 and 30800507 to JF), and the National Key Technology R&D Program, China (2015BAI12B00 to JF).

Authors JC and XL were employed by company Guangzhou Sagene Biotechnology Company, Limited.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.