ABSTRACT
The MASS cohort comprises 2000 ICU patients with severe pneumonia, covering community-acquired pneumonia, hospital-acquired pneumonia, and ventilator-associated pneumonia, sourced from 19 hospitals across 10 cities in three provinces. A wide array of samples including bronchoalveolar lavage fluid, sputum, feces, and whole blood are longitudinally collected throughout patients' ICU stays. The cohort study seeks to uncover the dynamics of lung and gut microbiomes and their associations with severe pneumonia and host susceptibility, integrating deep metagenomics and transcriptomics with detailed clinical data.
ABSTRACT
BACKGROUND: For decades, the incidence and clinical characteristics of Pneumocystis jirovecii colonization in patients with severe pneumonia was unclear. RESEARCH QUESTION: What are the clinical features and outcomes associated with P jirovecii colonization in individuals diagnosed with severe pneumonia? STUDY DESIGN AND METHODS: In this multicenter, retrospective, matched study, patients with severe pneumonia who underwent BAL clinical metagenomics from 2019 to 2023 in the ICUs of 17 medical centers were enrolled. Patients were diagnosed based on clinical metagenomics, pulmonary CT scans, and clinical presentations. Clinical data were collected retrospectively, and according to propensity score matching and Cox multivariate regression analysis, the prognosis of patients with P jirovecii colonization was compared with that of patients who were P jirovecii-negative. RESULTS: A total of 40% of P jirovecii-positive patients are considered to have P jirovecii colonization. The P jirovecii colonization group had a higher proportion of patients with immunosuppression and a lower lymphocyte count than the P jirovecii-negative group. More frequent detection of cytomegalovirus, Epstein-Barr virus, human herpesvirus-6B, human herpesvirus-7, and torque teno virus in the lungs was associated with P jirovecii colonization than with P jirovecii negativity. By constructing two cohorts through propensity score matching, we incorporated codetected microorganisms and clinical features into a Cox proportional hazards model and revealed that P jirovecii colonization was an independent risk factor for mortality in patients with severe pneumonia. According to sensitivity analyses, which included or excluded codetected microorganisms, and patients not receiving trimethoprim-sulfamethoxazole treatment, similar conclusions were reached. INTERPRETATION: Immunosuppression and a reduced lymphocyte count were identified as risk factors for P jirovecii colonization in patients with non-Pneumocystis pneumonia. More frequent detection of various viruses was observed in patients colonized with P jirovecii, and P jirovecii colonization was associated with an increased 28-day mortality in patients with severe pneumonia.