The frequency of community patients with an immunocompromised host has increased dramatically over decades. Despite significant advances in the prevention, diagnosis, and treatment of infection in the immunocompromised host, it remains a major cause of morbidity, increased length of stay, increased total costs, and mortality. Intensive care mortality rates are significantly higher among immunocompromised hosts. The superimposition of the compromised host defenses and critical illness make the detection and management of infections in such patients more difficult but crucial toward salvaging the patient outcome (Shorr et al., 2004; Linden, 2009). Thus, a rapid and accurate pathogen diagnosis is essential to guide anti-infective treatment and improve the prognosis (Cecconi et al., 2018; Fernando et al., 2021). A variety of microbes, including Pneumocystis jirovecii and their mixed infections, are important pathogens responsible for severe pneumonia in immunosuppressed patients (Cilloniz et al., 2019). However, the current diagnostic efficiency of conventional methods is merely 30%–40% (Musher et al., 2013; Jain et al., 2015). To improve the efficacy of pathogen diagnosis using quick and accurate test methods is of great importance to the precise anti-infective treatment.

An emerging trend of the increasing incidence of P. jirovecii pneumonia (PJP) has been exhibited recently. This may be due to the increase of various immunodeficient populations [especially in non-human immunodeficiency virus (HIV) immunodeficient patients] (Roux et al., 2014), as well as the improved efficiency of the state-of-the-art technology, metagenomic next-generation sequencing (mNGS), to detect the pathogen (Linden, 2009; Harpaz et al., 2016; Cilloniz et al., 2019; Tsang et al., 2021). It was reported that PJP in non-HIV immunosuppressed patients was characterized by more insidious onset, more rapid progression, and more common concomitant mixed infection with other pathogens (Armstrong-James et al., 2014; Rego de Figueiredo et al., 2019). In addition, the condition was more likely to progress into acute progressive respiratory failure within a short period of time (Kotani et al., 2017). Therefore, early, rapid, and accurate pathogen diagnosis is very important not only to guide the treatment of PJP and mixed infections but also to improve the prognosis.

The state-of-the-art technology, mNGS, can theoretically detect the nucleic acids of all pathogens in the specimens in a rapid, comprehensive, and unbiased manner (Chiu and Miller, 2019). Previous studies have reported that mNGS is able to detect pathogens in a variety of specimens including bronchoalveolar lavage fluid (BALF) (Wilson et al., 2019; Chen et al., 2021; Gu et al., 2021). This technology is characterized by high sensitivity, high accuracy, and short detection time (Chiu and Miller, 2019; Gu et al., 2021). It has significant advantages for diagnosing pathogens responsible for mixed infections in immunosuppressed patients, severe infections, and rare and new pathogen infections (Chiu and Miller, 2019). However, the efficacy of BALF-mNGS in the etiological diagnosis of non-HIV immunosuppressed patients with PJP and mixed infections is unknown.

This study focused on PJP and mixed infections as important types of severe infections in non-HIV-infected immunosuppressed patients. BALF-mNGS was performed to detect PJ and the pathogens of mixed infections. Pathogen diagnostic efficacy and time consumption were analyzed and compared between BALF-mNGS and conventional methods. In addition, the value of BALF-mNGS in the accurate etiological diagnosis of severe pneumonia in non-HIV-infected immunosuppressed patients was also evaluated.

This study complied with the 1964 “Declaration of Helsinki” and its subsequent amendments. This study was approved by the ethics committees of each study institution. Each patient or legal representative provided written informed consent at the time of registration.

With the increase of immunosuppressed population, severe pneumonia in non-HIV immunosuppressed patients has become a significant part of respiratory intensive care unit (RICU) patients, and the resulting sepsis and multiple organ failure are the important causes of death (Shorr et al., 2004; Ramirez et al., 2020). There are often rare pathogen, conditional pathogen, and multiple pathogen mixed infections due to the impact of underlying diseases and immune function. These infections are characterized by atypical clinical symptoms, rapid disease progression, and high mortality (Linden, 2009; Konduri, 2016). Additionally, these patients often use a variety of anti-infective drugs during the course of disease, resulting in a low pathogen detection rate of conventional methods. Therefore, rapid and accurate pathogen diagnosis has become a clinical challenge. Fortunately, we have this new state-of-the-art technology of mNGS for pathogen identification in severe lung infections. In this study, BALF-mNGS was used in the etiological diagnosis of non-HIV immunosuppressed patients with severe pneumonia admitted to the RICU of 7 hospitals.

Our previous study showed that ~20% patients with severe pneumonia admitted to the RICU were immunosuppressed patients, and P. jirovecii was found in 45% of these patients (Wu et al., 2020). In this study, we also found P. jirovecii as a major infectious pathogen in immunosuppressed patients, indicating that PJP has become an emerging trend in RICU patients with pneumonia these years. So, dynamically monitoring CD4+ lymphocytes in peripheral blood is necessary for immunosuppressed patients. For immunosuppressed patients with CD4+ lymphocytes <200/μl (<14%), accompanied by fever, shortness of breath, and diffuse ground-glass-like exudates in lungs, P. jirovecii and other suspected pathogens should be detected as quickly as possible. The rapid and accurate diagnosis of causative pathogens can help clinicians propose antibiotic treatments, improving the patients’ prognosis.

The ratio of hematological malignancies and transplantation was meaningfully higher in patients with PJP and mixed infection. In 4–12 months after transplantation, the combined use of multiple immunosuppressants was a high-risk factor for PJP. We also found that PJP patients were more susceptible to hypoxic respiratory failure, which was characterized by more severe hypoxia, a higher possibility of receiving invasive mechanical ventilation, and higher susceptibility to shock. This suggested that BALF-mNGS was more urgently needed in these kinds of patients to help guide the subsequent anti-infective treatment.

P. jirovecii is an ascomycete, mainly colonizing on the human type I alveolar epithelium (Alanio et al., 2016). It only causes disease in the immunosuppressed population, leading to serous exudative inflammation in the lungs (Lee and Chuang, 2018). The typical lung imaging features of PJP are characterized by diffuse interstitial exudation in both lungs (Alanio et al., 2016; Mu et al., 2016). However, similar imaging findings may be observed in other lung diseases such as lung GVHD, the pulmonary invasion of connective tissue disease, and viral pneumonia (Gotway et al., 2005). Conventional etiological diagnostic methods include microscopic examination, PCR, serum, and/or BALF fungal G test plus serum LDH (Esteves et al., 2015). However, there are some limitations, such as the difficulty of P. jirovecii culture in vitro, low positive rate of microscopic examination, and high false-positive rate of fungal G tests (Calderon et al., 2010; Huang et al., 2011; Esteves et al., 2014). qPCR has been recommended for the diagnosis of PJP in the European Conference on Infections in Leukaemia (ECIL) guidelines, but it is hard for qPCR to identify all pathogens in immunosuppressed patients with PJP mixed infections, especially for rare and opportunistic pathogens. In addition, there are no commercial qPCR kits for P. jirovecii in China, and the cut-off value remains unclear. Nevertheless, qPCR was also performed in some cases of this study, the results were shown in Table S7 and the Supplementary Material . It should be noted that PJP patients with immunosuppression are prone to mixed infections, which made the accurate diagnosis more difficult. A rapid and accurate pathogen diagnosis is needed.

In this study, mNGS showed a higher overall positive detection rate for various pathogens than conventional methods in both groups, especially for PJP mixed with rare and atypical pathogens. This has preliminary illustrated the advantages of BALF-mNGS for rapid pathogen detection in patients with PJP-mixed infections. The 59 severe pneumonia patients with positive BALF-mNGS but negative conventional test results mainly manifested underlying diseases as hematological malignancies with agranulocytosis, within 8 months after hematopoietic stem cell transplantation and within 8 months after solid organ transplantation. This highly suggested that these patients were a population of special concerns and needed early BALF-mNGS diagnosis.

Different from previous studies (Li et al., 2018; Wu et al., 2020), in this study, the vast majority of the observation group were mixed infections of P. jirovecii with other pathogens (85.8%), while most patients in the control group also manifested as mixed infections of multiple pathogens (76.9%), suggesting that immunosuppressed patients with severe pneumonia were more prone to mixed infections. To improve the pathogen detection rate, BALF-mNGS should be used as soon as possible despite conventional methods.

Compared with other common pathogens of lung infections such as Aspergillus, Cryptococcus, and M. tuberculosis, the detection rate of P. jirovecii using BALF-mNGS was significantly higher. In the same BALF sample, the number of P. jirovecii sequences was often higher than other intracellular pathogens. The reason might be that the pathogenic site of P. jirovecii is in alveolar epithelial cells, mainly causing diffuse alveolar injuries. Accordingly, conventional cell wall disruption can meet the testing requirements. In addition, given the small amount of sputum in PJP patients, the positive detection rate was higher with lavage fluid than with sputum and blood (Bateman et al., 2020). It suggests that BALF-mNGS can be fully applied in the pathogen diagnosis of PJP. For intracellular pathogens, the positive detection rate is relatively low; therefore, deep wall disruption may be necessary. When intracellular bacterial coinfection is considered, the depth of the wall disruption of the samples should be appropriately increased to improve the overall positive detection rate. Concomitant PCR is also needed to further verify mNGS results. One feasible subplan is to select the tissue samples of the corresponding lesions for mNGS detection with reference to lung HRCT characteristics. Since PJP is prone to the complication of a viral infection, an mNGS RNA detection process is needed. The simultaneous application of mNGS and conventional methods can not only increase the positive rate of pathogen diagnosis but also improve the specificity. In addition, it is necessary to make a comprehensive judgment based on a patient’s clinical characteristics, underlying diseases, imaging characteristics, and background medications to improve the diagnostic accuracy.

Like other highly sensitive test methods, mNGS can also produce false-positive results. The nucleic acid sequences of dead, contaminating, and even locally colonized pathogenic microorganisms are all detectable, sometimes even with a high level of sequence numbers. Sputum specimens and inappropriate lavage specimens are susceptible to the interference of oral and airway-colonizing bacteria, which might increase the background noise and even cause false-positive results. Factors such as sampling sites, standardized sterile operation, specimen submission, and the depth of subsequent wall disruption can also affect the results. Therefore, diagnostic accuracy can further be improved through close cooperation between clinicians and testing microbiologists, through simultaneous mNGS detection and conventional testing of the same specimen.

The pre-evaluation of possible pathogens to carry out targeted detection is the guiding principle for most commonly used conventional detection methods, but it is highly susceptible to the subjective empirical influence of clinical and laboratory personnel, with a narrow pathogen spectrum, leading to a high rate of missed diagnosis. It also requires multiple submissions, a higher specimen demand, and longer detection time. In this study, we found good agreement between BALF-mNGS results and patients’ clinical characteristics in the etiological diagnosis of PJP. The detection spectrum was wider, and the detection time was significantly shorter than conventional methods. It is able to greatly shorten the etiological diagnosis time and provide strong support for rapid and accurate anti-infective treatment.

There were certain limitations for this study. The number of patients enrolled in this study was limited. A prospective study with a larger number of patients would be needed to support the use of mNGS in the early diagnosis of PJP in immunosuppressed patients to further guide the clinical diagnostic standards.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: http://ngdc.cncb.ac.cn, PRJCA006572.

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000 (Cilloniz et al., 2019). The patients/participants provided their written informed consent to participate in this study.

YS and QL designed the paper and guided the implementation. HS and FW drafted the manuscript. HS, FW, MZ, XX, ML, WG, XW, HH, QW, and GY collected the cases and summarized the data. ZL and HX processed the data and did the statistical analysis. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

This study was supported by The Outstanding Clinical Discipline Project of Shanghai Pudong (PWYgy2018-6) and the Science and Technology Project of Xi'an (No. 21RGSF0013).

ZL and HX are employed by Hugobiotech Co., Ltd.

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.