Introduction
Pneumonia is common and potentially fatal, with CAP causing high hospitalizations and significant mortality. Its economic burden is substantial, and HAP/VAP further increase mortality, hospital stay, and cost. Multiple pathogens can cause pneumonia, but most cases lack an identified cause, and pathogen type does not predict severity.
Diagnosis is challenging due to the absence of a standardized definition and reliance on imaging plus clinical symptoms. Severe CAP criteria are subjective. Existing biomarkers and scoring systems (PSI, CURB-65, procalcitonin) have limitations in accuracy and reliability.
These challenges drive interest in new diagnostic tools. Breath-based metabolomics offers a real-time, noninvasive method for detecting biomarkers to aid diagnosis, therapy monitoring, and outcome prediction. A PubMed search identified relevant observational, in vitro, and animal studies up to July 2025.
Methods
Exhaled breath analysis has long been used in clinical care, such as CO₂ monitoring for ventilation and nitric oxide for asthma. These markers, along with others, are being explored for CAP but need further research.
Breath contains hundreds of metabolites including gases, VOCs, and nonvolatile compounds offering insight into physiological and biochemical processes. Samples can be collected from spontaneously breathing or mechanically ventilated patients, either through mouthpieces/masks or via ventilator circuits.
Collection methods vary: VOCs are best captured as exhaled breath vapor (EBV) and analyzed with GC/MS or PTR-MS after preconcentration (e.g., SPME, NTD). Larger molecules require exhaled breath condensate (EBC), obtained through cooling devices such as K-tube, ECoScreen, TurboDECCS, or RTube, followed by LC/MS analysis.
Advanced devices like EBCare smart mask and TreSenso Tech’s portable VOC analyzer are nearing clinical application, offering real-time biomarker detection for respiratory infections and conditions.
Identifying specific VOC markers through in vitro approaches
In vitro research shows that different bacteria produce distinct VOC patterns, but only a few metabolites are unique to specific pathogens. Variations in culture conditions and bacterial growth lead to inconsistent results, so VOC ratios are often more useful than single markers.
However, many VOCs identified in vitro do not appear reliably in patients due to low levels or host-related factors. Several studies found mismatches between lab and clinical findings, though one larger study replicated most VOCs in patient samples.
Only one study showed strong accuracy for identifying A. baumannii both in vitro and in vivo, but even there the VOC profiles differed. These discrepancies highlight the complexity of host–pathogen interactions and suggest that in vitro biomarkers may not directly translate to real-world infections.
Making clinical diagnosis more precise
Exhaled breath can help distinguish pneumonia from other acute cardiorespiratory conditions, enabling faster diagnosis and treatment. Studies show elevated exhaled NO in ventilated pneumonia patients and multibiomarker breath profiles that differentiate CAP from heart failure, asthma, and COPD with moderate-to-high accuracy (AUC up to 0.91). Breath analysis has also detected ARDS earlier than clinical signs, highlighting its potential as a noninvasive diagnostic tool.
Understanding clinical trajectory
Breath metabolites offer a noninvasive way to monitor host response and treatment in respiratory infections. Longitudinal studies in severe COVID-19 and ARDS show that breath profiles reflect disease severity, clinical trajectory, and recovery. In CAP, markers of oxidative stress, inflammation, and immune response—such as exhaled NO, H₂O₂, TBARs, and hepatocyte growth factor (HGF)—change with treatment, decreasing as patients recover. These findings suggest that exhaled breath, especially condensate, provides a real-time window into lung repair and immune response, potentially offering insights not captured by systemic measurements.
Phenotyping patients with pneumonia
Exhaled breath may help phenotype pneumonia and reveal mechanisms behind susceptibility and variable outcomes. Studies in obstructive lung disease show breath profiles reflect airway inflammation, cell type, steroid responsiveness, and predict exacerbations, suggesting similar insights could apply to pneumonia.
EBC protein biomarkers could differentiate pneumonia subtypes and predict prognosis. For example, cytokine levels in ventilated or CAP patients correlated with disease severity, sepsis, and mortality, highlighting dysregulated immune responses not always seen in blood. Validated EBC analysis could complement current approaches to identify inflammatory subtypes and guide treatment.
Understanding lung immunity and risk of pneumonia
Breath analysis has the potential to assess an individual’s risk for pneumonia, much like cholesterol and blood pressure predict heart attacks. Exhaled metabolites reflect the state of lung defenses, including macrophages, T lymphocytes, and epithelial cells, and vary with age, comorbidities, environmental exposures, or protective interventions like vaccination. Distinct VOC profiles can indicate immune status and lung health, offering a noninvasive approach to predict susceptibility, guide prevention, and prioritize at-risk individuals for early interventions.
Conclusion
Progress in breath biomarker research for pneumonia is slow due to technological constraints, lack of standardized collection methods, environmental and patient-related confounders, and low protein concentrations in EBC. Small sample sizes and overlapping biomarkers with other lung diseases limit interpretability.
Standardized protocols, improved detection technologies, and large multicenter studies are needed. Once validated, exhaled breath analysis could offer a simple, noninvasive way for early diagnosis, risk stratification, and better understanding of pneumonia biology, enabling targeted therapies.
