Patients with severe community-acquired pneumonia who carried a heavy burden of Pseudomonas aeruginosa in their lungs were markedly more likely to die within four weeks, according to a multicenter retrospective study published in Annals of Clinical Microbiology and Antimicrobials. The researchers used metagenomic next-generation sequencing, or mNGS, not merely to detect the bacterium but to quantify it, and found that a semi-quantitative abundance metric tracked a stepwise ecological collapse of the lung microbiome alongside worsening clinical outcomes.
The study drew on 130 adults with severe community-acquired pneumonia caused by Pseudomonas aeruginosa, treated at five hospitals between September 2021 and June 2025. Of 148 patients initially screened, seven under 18 years of age and eleven with incomplete clinical data were excluded. Diagnosis followed the ATS/IDSA criteria for severe pneumonia, and identification of the pathogen relied on conventional microbiology, including culture of bronchoalveolar lavage fluid or sputum, with mNGS used for additional detection and quantification rather than as the sole basis for etiological diagnosis.
To quantify the bacterium, the team sequenced BALF samples at a central laboratory and normalized read counts to a metric called reads per ten million, or RPTM, for each detected organism. Laboratory personnel performing the sequencing and primary microbiome analyses were blinded to patients’ clinical outcomes. Patients were then stratified into three groups: low abundance below 10,000 RPTM (61 patients), medium abundance between 10,000 and 100,000 (30 patients), and high abundance above 100,000 (39 patients).
The microbiome findings showed a clear gradient. Alpha diversity, measured with Shannon and Simpson indices, declined progressively as Pseudomonas abundance rose, with all comparisons reaching statistical significance. Beta diversity analysis based on Bray-Curtis dissimilarity revealed distinct community structures across the three groups, with principal coordinate analysis showing clear separation supported by PERMANOVA (R2 = 0.343, P < 0.001) and ANOSIM (R2 = 0.393, P < 0.001). In the high abundance group, Pseudomonas aeruginosa accounted for up to 95.99 percent of the total microbial abundance, a near-complete takeover of the lung ecosystem. Linear discriminant analysis effect size identified Staphylococcus aureus and Candida albicans as enriched in the low abundance group, while Pseudomonas was enriched in the high abundance group.
Clinical outcomes worsened in parallel. Markers of inflammation and severity, including white blood cell count and procalcitonin, were significantly higher in the high abundance group (both P < 0.001), and duration of mechanical ventilation and length of hospital stay both increased with abundance (both P < 0.001). The 28-day mortality rate climbed from 21.31 percent in the low abundance group to 43.33 percent in the medium group and 58.97 percent in the high group (P = 0.001). Spearman correlation analysis linked Pseudomonas RPTM positively with mechanical ventilation duration, hospital stay, APACHE II score, and white blood cell level.
The authors also introduced treatment effectiveness, defined as achievement of clinical stability within 72 hours of therapy, as an intermediate variable visualized in a Sankey diagram. As abundance increased from low to high, the proportion of patients achieving effective treatment fell while treatment failure rose, and among patients with ineffective treatment, mortality increased progressively across abundance strata.
For prognosis, ROC analysis showed that Pseudomonas RPTM had moderate predictive value for 28-day mortality, with an area under the curve of 0.761. The reported optimal cutoff was 41,122; the abstract lists sensitivity of 69.40 percent and specificity of 75.30 percent, while the results text reports 69.6 percent and 70.2 percent. Grey zone analysis defined an inconclusive range: values below 2,287 were associated with roughly 90 percent survival, whereas values above 220,339 were associated with roughly 90 percent mortality. In univariate analysis, APACHE II score, immunocompromised status, and Pseudomonas RPTM were associated with 28-day mortality, but in multivariable logistic regression only RPTM remained independently associated, with an odds ratio of 2.219 (95 percent confidence interval 1.509 to 3.262, P < 0.001); immunocompromised status showed only a non-significant trend (OR 2.287, P = 0.098).
The authors interpret these associations cautiously. RPTM is a semi-quantitative estimate of relative abundance and does not directly measure absolute bacterial load, so the findings should be read in terms of relative microbial abundance rather than absolute pathogen burden. They also note that Pseudomonas aeruginosa shows substantial genomic and phenotypic heterogeneity, with different strains carrying varied virulence determinants, plasmids, and antimicrobial resistance genes, meaning abundance alone cannot capture strain-specific biology that may influence prognosis.
Several further limitations temper the conclusions. The retrospective design may introduce selection bias and residual confounding, the sample size was relatively limited, and the RPTM cutoff and grey zone values require external validation. Sequencing was performed at a single time point, so dynamic changes in microbial abundance over the course of illness were not assessed. Missing data were handled with multiple imputation across ten imputed datasets, and the authors state the proportion of missing data was low and did not involve key variables.
Within those constraints, the study suggests that quantitative mNGS metrics may extend beyond pathogen identification toward prognostic interpretation, potentially complementing established severity scores such as APACHE II and SOFA for early risk stratification in severe pneumonia. The authors emphasize that prospective multicenter studies are needed to validate RPTM-based risk stratification before any routine clinical implementation, and that integrating quantitative sequencing with genomic characterization of resistance and virulence determinants may yield a more complete picture of pathogen behavior in future work.
Severe community-acquired pneumonia remains a leading cause of mortality in critically ill patients worldwide, and despite advances in antimicrobial therapy and supportive care, outcomes for those admitted to intensive care units remain suboptimal. Early identification of the causative pathogen and assessment of disease severity are considered critical for guiding appropriate treatment and improving prognosis, yet most conventional diagnostic approaches characterize pathogens only qualitatively. This limitation restricts their ability to reflect pathogen abundance and its relationship with disease progression, which is precisely the gap the present study sought to address.
Pseudomonas aeruginosa is a well-recognized opportunistic pathogen in this setting, distinguished by intrinsic and acquired resistance to multiple antibiotics, rapid adaptation to the host environment, and an ability to dominate the pulmonary microbiota under pathological conditions. Previous work has linked infection with this organism to increased disease severity, prolonged hospitalization, and higher mortality in critically ill patients, making it a natural candidate for quantitative rather than purely qualitative assessment.
The lung microbiome is increasingly viewed as a dynamic ecosystem with an important role in respiratory health and disease. Disruption of microbial diversity, termed dysbiosis, has been observed in patients with severe pneumonia and is often accompanied by overgrowth of dominant pathogens. Prior studies have suggested that reduced microbial diversity is linked to worse outcomes in critically ill populations, providing the conceptual backdrop for examining whether a single quantitative sequencing metric could capture both ecological disruption and clinical trajectory.
Metagenomic next-generation sequencing has emerged as a powerful tool for comprehensive characterization of respiratory pathogens and the lung microbiome, and it has been increasingly applied in community-acquired pneumonia, particularly severe cases, where it has been reported to improve pathogen detection compared with conventional microbiological methods. Beyond identification, the platform yields semi-quantitative metrics such as RPTM that can serve as indicators of microbial abundance in clinical samples. The authors’ own previous work observed that mNGS-derived RPTM was positively associated with length of hospital stay in patients with pulmonary mycobacterial diseases, with concurrent changes in lung microbiome diversity, suggesting its potential to reflect pathogen abundance, microbial community alterations, and disease severity. The clinical relevance of such quantitative metrics in severe pneumonia due to Pseudomonas aeruginosa, however, had remained poorly understood.
The technical pipeline underlying the abundance estimates deserves note. Bronchoalveolar lavage fluid samples of 1.5 to 3 milliliters were obtained via bronchoscopy, centrifuged, and subjected to nucleic acid extraction using commercial kits at the central laboratory. Sequencing libraries were prepared with a tagmentation kit and run on an Illumina NextSeq 550Dx platform, generating no fewer than 20 million reads per sample. Raw reads underwent quality control with Trimmomatic, host-derived sequences were removed by alignment to the human reference genome GRCh37/hg19 using Bowtie2, and the remaining non-human reads were taxonomically classified with Kraken2 against a curated microbial reference database comprising more than 24,000 pathogenic organisms. Read counts for each detected organism were then normalized to RPTM for downstream analysis.
The statistical framework was similarly detailed. Alpha diversity was evaluated with both the Shannon index, which reflects species richness and evenness, and the Simpson index, which places greater emphasis on community dominance, providing complementary perspectives on microbial diversity. Alpha diversity indices were compared between groups using the Wilcoxon rank-sum test, beta diversity was assessed on Bray-Curtis dissimilarity, and differences in community structure were tested with permutational multivariate analysis of variance and analysis of similarities. Differentially abundant taxa were identified using linear discriminant analysis effect size with an LDA score above 2.0 and a P value of 0.05 or below as thresholds. Associations between microbial features and clinical outcomes were evaluated with Spearman’s rank correlation, and grey zone analysis was applied to define the range of inconclusive cutoff values around the mortality threshold. All analyses were performed in R version 4.4.1, missing values were handled through multiple imputation generating ten imputed datasets, and all statistical tests were two-sided.
The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Quzhou People’s Hospital, with the requirement for written informed consent waived owing to the retrospective design and use of anonymized data. Disease severity was assessed with the APACHE-II score, and clinical outcomes captured included duration of mechanical ventilation, length of hospital stay, and 28-day mortality. Immunocompromised status was classified according to underlying conditions or treatments known to impair immune function, including hematologic malignancies, solid organ transplantation, long-term corticosteroid use, and immunosuppressive therapy, and was included as a clinically relevant variable based on prior evidence of its association with poor outcomes in severe community-acquired pneumonia.
The raw sequence data have been deposited in NCBI under BioProject number PRJNA 1432697, and additional deidentified data are available from the corresponding author upon reasonable request. The article is published open access in Annals of Clinical Microbiology and Antimicrobials, volume 25, article number 33.
Subject of Research: Quantitative metagenomic sequencing of Pseudomonas aeruginosa burden and lung microbiome disruption in severe community-acquired pneumonia
Article Title: Metagenomic-based quantification of Pseudomonas aeruginosa burden links microbiome collapse to mortality in severe community-acquired pneumonia
Article References: Zhao, J., Chen, X., Wang, X., Cui, Y., Zhuge, J., Zhang, Y., Zhang, L., Yan, Y., Fang, H., Hua, Z., & Li, G. (2026). Metagenomic-based quantification of Pseudomonas aeruginosa burden links microbiome collapse to mortality in severe community-acquired pneumonia. Annals of Clinical Microbiology and Antimicrobials, 25(1), Article 33. https://doi.org/10.1186/s12941-026-00880-5
Image Credits: AI Generated
DOI: 10.1186/s12941-026-00880-5
Keywords: Pseudomonas aeruginosa, severe community-acquired pneumonia, metagenomic next-generation sequencing, lung microbiome, RPTM, 28-day mortality, microbial diversity, dysbiosis, risk stratification, BALF, APACHE II, retrospective study
News Source: Morgan Morrow. (October 29, 2026). Research report: Metagenomic-based quantification of Pseudomonas aeruginosa burden links microbiome collapse to mortality in severe community-acquired pneumonia. Scienmag.
