When Aspergillus strikes, the mouth tells a different story in COPD flare-ups

When a patient with chronic obstructive pulmonary disease suddenly deteriorates, doctors usually blame bacteria. But a new study from China suggests that the mouth may reveal which pathogen is actually driving the crisis, and that the culprit leaves a surprisingly specific fingerprint in the oral microbiome and its metabolic output.

Acute exacerbations of COPD, known as AECOPD, are episodes in which symptoms abruptly worsen, often triggered by infection. These events are major drivers of disease progression, hospitalization, and death. While bacterial infections are well-established triggers, fungal pathogens such as Aspergillus and Candida have increasingly been recognized as underappreciated contributors, particularly in patients exposed to corticosteroids, broad-spectrum antibiotics, or recurrent hospital stays. Yet whether different fungal statuses leave distinct traces in the host-microbial ecosystem beyond the lower airways has remained poorly understood.

Now a team led by Yanmei Feng and colleagues at Chongqing University Central Hospital has taken a step toward answering that question. In a prospective cohort study running from August 2024 to July 2025, they enrolled 18 patients with AECOPD and stratified them into three groups of six: non-fungal exacerbators, Aspergillus-associated AECOPD, and Candida-positive AECOPD. Fungal status was determined through an integrated assessment of clinical manifestations, respiratory fungal culture, serum fungal biomarkers such as galactomannan and beta-D-glucan, molecular testing where available, chest CT findings, and histopathology where available. Simple fungal isolation from respiratory specimens, without compatible clinical and radiological evidence, was not considered sufficient to establish invasive fungal infection.

The researchers collected oral samples before treatment and, where available, after treatment. They then combined 16S rRNA gene sequencing with untargeted LC-MS metabolomics to characterize oral microbial and metabolic profiles across the three groups. The primary analyses compared pretreatment microbial and metabolic profiles across fungal-status groups, while longitudinal pre/post-treatment analyses were secondary and exploratory, performed only in groups with sufficient paired samples.

The results were striking. At pretreatment, the Aspergillus-associated group showed significantly reduced alpha diversity, meaning lower microbial richness and diversity, and distinct microbial community separation from non-fungal exacerbators, with a Bray-Curtis R of 0.378 and a P value of 0.009. In contrast, the Candida-positive group showed no significant global shift in community composition. The oral microbiome was predominantly composed of Bacillota, Actinomycetota, Pseudomonadota, and Bacteroidota, with common oral taxa such as Streptococcus, Rothia, Veillonella, Actinomyces, and Prevotella prevalent across groups. Exploratory LEfSe analysis identified subgroup-specific candidate taxonomic signatures: Lautropia, Tannerella, Prevotella pallens, Capnocytophaga, and Johnsonella were associated with the Aspergillus group, while Actinomyces and Leptotrichia were linked to the Candida group.

Metabolomic profiling revealed even more pronounced baseline differences in the Aspergillus-associated group, with 11 upregulated and 90 downregulated metabolites compared with the non-fungal group. The differential metabolites were predominantly mapped to sphingolipid metabolism, purine metabolism, one-carbon metabolism, and cAMP/cGMP signaling pathways. These pathways are biologically plausible mediators of fungal-host interactions, given that sphingolipid signaling is central to membrane integrity and immune activation, while purine and one-carbon metabolism are involved in nucleotide synthesis and methylation reactions that can shape inflammatory responses.

In secondary exploratory longitudinal analyses, metabolic shifts were observed in the non-fungal group, with 4 upregulated and 78 downregulated metabolites, and in the Candida-positive group, with 14 upregulated and 8 downregulated metabolites. No formal longitudinal analysis was performed in the Aspergillus-associated group because only two post-treatment samples were available, largely due to the prolonged voriconazole-based antifungal treatment course of 4 to 6 weeks and post-discharge follow-up constraints. The Candida-positive group received caspofungin and/or fluconazole for 7 to 14 days, allowing more complete paired sampling.

Integrated microbiome-metabolomics analysis using O2PLS identified a dominant shared latent component differentiating pretreatment Aspergillus-associated AECOPD from non-fungal AECOPD, with an R2X of 0.612 and an R2Y of 0.476. Microbial features contributing to this shared component included Veillonella, Porphyromonas, and Clostridia_UCG-014, while contributing metabolite features included propionate, succinate, and pyruvate. These joint loading patterns represent feature contributions to the shared cross-omics covariance structure rather than direct pairwise taxon-metabolite associations. Procrustes analysis did not support global concordance between microbial community structure and metabolomic profiles, suggesting that the shared cross-omics signal is concentrated within a limited subset of features rather than reflected by ecosystem-wide alignment.

The authors emphasize that these findings support the presence of a distinct oral microbial-metabolic phenotype associated with Aspergillus status but do not establish the longitudinal treatment response of Aspergillus-associated AECOPD. They also caution that the observed differences should be interpreted as differences in microbial-metabolic phenotype rather than evidence that Aspergillus is more prevalent or etiologically more important than Candida in AECOPD. The biological contrast may reflect the distinct properties of these fungi: Aspergillus species are filamentous fungi capable of extensive tissue invasion and potent immune activation, whereas Candida species typically behave as commensal yeasts with opportunistic pathogenicity under conditions of immune imbalance.

Several limitations warrant attention. The sample size was modest, particularly after subgroup stratification, which may limit statistical power and generalizability. Although the researchers applied an integrated clinical, microbiological, and radiological assessment to minimize the inclusion of simple fungal colonization, histopathological confirmation was unavailable in most patients, so residual misclassification between colonization and clinically relevant fungal infection cannot be completely excluded. The study also did not directly characterize the oral mycobiome, limiting the ability to delineate cross-kingdom interactions. Paired lower-airway samples were not collected concurrently, so the present study cannot determine whether the observed oral alterations are directly reflected in the lower airways. Moreover, multiple metabolite-wise comparisons were performed using nominal P values without formal false discovery rate correction, so the identified metabolic features should be regarded as exploratory and require validation in larger independent cohorts. The clinical relevance of these signatures also remains to be established, because the study did not evaluate associations with outcomes such as hospitalization duration, mechanical ventilation, readmission, or longitudinal lung-function decline.

Despite these caveats, the study provides candidate signatures for future validation in larger longitudinal cohorts. Given the non-invasive and repeatable nature of oral sampling, these signatures may provide a useful basis for future biomarker development. The oral-lung axis, with microaspiration and shared mucosal immune signaling as potential mechanisms, offers a biologically plausible framework linking oral microbial ecology to lower airway disease. As the authors conclude, these exploratory candidate signatures support further investigation of the oral microenvironment as a clinically accessible compartment in AECOPD, while paired oral and lower-airway studies are needed to determine whether they reliably reflect lower-airway microbial and inflammatory alterations.

Subject of Research: Oral microbiome and metabolomics in fungal-associated COPD exacerbations

Article Title: Oral microbial and metabolic alterations in acute exacerbations of COPD across fungal status

Article References: Feng, Y., Tang, F., Ma, Q., Ao, Z., Xiong, Q., Qin, W., Yang, Y., Tao, S., & Guo, R. (2026). Oral microbial and metabolic alterations in acute exacerbations of COPD across fungal status. Frontiers in Cellular and Infection Microbiology, 16, Article 1843568. https://doi.org/10.3389/fcimb.2026.1843568

Image Credits: AI Generated

DOI: 10.3389/fcimb.2026.1843568

Keywords: AECOPD, Aspergillus, Candida, oral microbiome, metabolomics, COPD exacerbations, microbial dysbiosis, oral-lung axis, Oral, microbial, metabolic, alterations