Don't Hold Your Breath! The Revealing Science of Breath Analysis

The real challenge of breath analysis lies in accurately sifting through this complex "breathprint." The low concentrations of many VOCs, coupled with their potential structural similarities, make their distinct identification and quantification a demanding task. Furthermore, the ever-present risk of contamination from ambient air, the hard-to-standardize sampling environment, or even the materials used in collection devices necessitates stellar experimental design and execution (Westphal et al., 2023). The ultimate goal is to confidently identify true biomarkers: specific VOCs, or distinct patterns of VOCs, that reliably and accurately indicate a particular physiological or pathological condition, while effectively filtering out confounding signals or "noise."

Before the secrets within our breath can be deciphered, the breath itself must be collected. As it turns out, how this collection is performed is of immense importance, significantly impacting the quality and reliability of the results. The field of breath analysis employs a variety of sampling strategies, broadly categorized into two main approaches: direct (online) analysis, where breath is analyzed in real-time, and indirect (offline) analysis, where breath is collected and stored for later examination in a laboratory. Each approach has its own set of methodologies, advantages, and challenges. The lack of universally standardized procedures for breath collection has been a notable hurdle in translating promising research findings into routine clinical tests, making the choice and execution of sampling methods a critical consideration.

Online Breath Analysis: Real-Time Insights

Direct, or online, breath analysis involves the patient exhaling directly into an analytical instrument, providing immediate or near real-time results. This approach offers the advantage of speed and minimizes issues related to sample storage and degradation. Several sophisticated mass spectrometry techniques are well-suited for this purpose. Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) is known for its rapid analysis and direct quantification capabilities without prior chromatographic separation (Alenezy et al., 2025). Similarly, Proton Transfer Reaction Mass Spectrometry (PTR-MS) allows for real-time monitoring of VOCs, using soft ionization to detect a wide array of compounds (Alenezy et al., 2025). Another powerful technique is Secondary Electrospray Ionization Mass Spectrometry (SESI-MS), which ionizes compounds directly from the breath at atmospheric pressure. SESI-MS is often coupled with high-resolution mass spectrometry (HRMS), enabling the sensitive detection of a broad spectrum of VOCs (Bajo-Fernández et al., 2024; Wüthrich et al., 2024). However, a challenge with direct infusion techniques like SESI-HRMS is feature annotation, as the absence of a pre-separation step like chromatography can make it difficult to differentiate isomers or deal with isobaric interferences (Wüthrich et al., 2024). While these online methods are incredibly fast and provide immediate data, a potential drawback can be the logistics for large-scale, multi-center studies if the instruments are not easily portable or if there is no option for sample storage and re-analysis.

Offline Breath Analysis: Capturing Samples for Later Investigation

Indirect, or offline, breath analysis involves collecting breath samples that are then stored and transported to a laboratory for subsequent, more detailed examination. This approach is often more practical and scalable for large clinical studies or when access to specialized analytical instruments is centralized.

One of the most traditional methods for offline collection is the use of breath bags, often made from inert materials like Tedlar® or Teflon® (Hagens et al., 2021; Alenezy et al., 2025). Patients exhale into these bags, typically filling them with "whole breath," which is a mixture of air from the entire respiratory tract. The VOCs captured in the bag are then usually transferred onto a sorbent tube for pre-concentration before analysis by techniques such as Thermal Desorption-Gas Chromatography-Mass Spectrometry (TD-GC/MS) (McCartney et al., 2022). While widely used, breath bags can present challenges, including potential background contamination from VOCs emitted by the bag material itself, or changes in the sample composition during storage and transport due to permeability or condensation (Westphal et al., 2023; Bajo-Fernández et al., 2024).

To overcome some limitations of bags, direct sorbent tube collection methods have gained prominence. Devices such as the Mistral or ReCIVA samplers, as discussed by Di Gilio et al. (2020), or specialized breath gas samplers (Hagens et al., 2021) enable breath to be drawn directly through sorbent tubes. This process captures and concentrates VOCs on the sorbent material immediately, minimizing storage issues associated with bags and reducing the risk of sample degradation or contamination during transit.

When collecting breath, it is not just what is collected, but also which part of the breath is targeted. Alveolar breath, also known as end-tidal breath, represents the air from the deepest part of the lungs. This fraction is considered to be richer in endogenous VOCs that have exchanged from the bloodstream and is theoretically less contaminated by ambient air present in the upper airways or mouth. Some sophisticated samplers are specifically designed to selectively capture this alveolar fraction, often employing CO2 sensors to monitor the exhalation profile or precise volume control mechanisms to isolate the last portion of the exhaled air (Di Gilio et al., 2020; Bajo-Fernández et al., 2024). The study by Di Gilio et al. (2020) specifically explored the differences observed when sampling whole breath versus alveolar breath.

Another consideration in breath sampling is the "clean air" conundrum. Some collection devices incorporate a system to supply the patient with purified or filtered air to inhale before and during the sampling process. The rationale is to "wash out" exogenous VOCs from the lungs that may have originated from the immediate room air, thereby aiming for a cleaner sample of endogenous compounds. However, the universal efficacy of this approach is still a subject of research. The absorption, distribution, and elimination kinetics of VOCs in the body are complex and highly dependent on the physicochemical properties of each compound (such as volatility and lipophilicity). For certain VOCs, particularly those that are fat-soluble or those resulting from long-term environmental exposure, a brief washout period may not be sufficient to clear them from the system. Indeed, some research, including the work by Di Gilio et al. (2020), suggests that attempting a lung washout might even introduce an additional source of variability into the breath profile. Despite these ongoing discussions, there is broad consensus among researchers on the critical importance of always collecting and analyzing an ambient air sample concurrently with the breath sample (McCartney et al., 2022). This allows for a proper assessment of potential contributions from the immediate environment to the patient's breathprint and helps to distinguish endogenous compounds from exogenous artifacts.

Finally, Exhaled Breath Condensate (EBC) is another offline collection method, though it targets a different fraction of breath components. EBC collection involves cooling the exhaled air, often using a cold trap (e.g., cooled to -78°C with dry ice/isopropanol, as by Wüthrich et al., 2024), causing water vapor and aerosols from the airway lining fluid to condense into a liquid sample. This liquid can then be analyzed for a variety of biomarkers, including non-volatile or less volatile compounds, proteins, and DNA, often using techniques different from those employed for gaseous VOCs, such as liquid chromatography-mass spectrometry (LC-MS) or GC-MS after appropriate extraction (Khoubnasabjafari et al., 2022; Wüthrich et al., 2024).

A variety of pre-concentration and preparation techniques are employed in breathomics. Sorbent-based trapping followed by Thermal Desorption (TD) is a cornerstone for the analysis of gaseous VOCs. In this approach, VOCs from breath, whether collected directly onto a tube or transferred from a collection bag, are adsorbed onto sorbent tubes. These tubes are carefully packed with one or more sorbent materials chosen for their affinity towards the target analytes and their ability to release them upon heating. Common sorbent choices include porous polymers like Tenax® TA (often favored for its hydrophobicity, making it suitable for humid breath samples, and its effectiveness for a broad range of VOCs, typically C6-C30), graphitized carbon blacks (e.g., Carbopack™, Carbotrap™), and carbon molecular sieves (e.g., Carboxen™), which are generally more suited for trapping very volatile compounds (Westphal et al., 2023; Lawal et al., 2017; Bajo-Fernández et al., 2024; Alenezy et al., 2025). Multi-bed sorbent tubes, containing sequential layers of different sorbents (e.g., Tenax GR and Carbograph 5TD as used by Hagens et al., 2021, or Carbosieve S III/Tenax® TA as by Wüthrich et al., 2024 for GC-MS of EBC volatiles), are frequently used to capture a wider volatility range of VOCs. However, careful selection and packing are needed to avoid issues like analyte breakthrough or irreversible adsorption, and to ensure consistent performance (Lawal et al., 2017; Bajo-Fernández et al., 2024). 

Solid Phase Microextraction (SPME) is another widely adopted, solvent-free pre-concentration technique. It utilizes a fused-silica fiber coated with a thin layer of a sorbent material (such as Polydimethylsiloxane (PDMS), Carboxen (Car), Divinylbenzene (DVB), or various combinations like Car/PDMS or DVB/Car/PDMS). The fiber is exposed to the breath sample, either by insertion into the headspace above a collected sample (e.g., in a bag or vial) or by direct exposure to the exhaled air stream. VOCs partition from the gas phase and adsorb/absorb onto the fiber coating until an equilibrium is approached. The fiber, now carrying the concentrated analytes, is then retracted and directly inserted into the hot injection port of a GC, where the VOCs are thermally desorbed for analysis. SPME is valued for its simplicity, minimal sample volume requirements, and the availability of various fiber coatings catering to different analyte polarities and volatilities (Bajo-Fernández et al., 2024; Khoubnasabjafari et al., 2022). On-fiber derivatization is also possible to enhance the detection of specific compound classes.

Needle Trap Devices (NTDs) represent a hybrid approach, aiming to combine the ease of use of SPME with the larger sampling capacity of sorbent tubes. These devices consist of a needle packed with a small amount of sorbent material. Breath is drawn through the needle, allowing VOCs to be trapped. Subsequently, the trapped analytes are thermally desorbed into a GC, similar to SPME fibers (Bajo-Fernández et al., 2024; Lawal et al., 2017).

For Exhaled Breath Condensate (EBC), which is a liquid matrix, different sample preparation strategies are often required. These might include Solid Phase Extraction (SPE) cartridges for clean-up and concentration of non-volatile or semi-volatile analytes, removing interfering substances from the EBC matrix (Khoubnasabjafari et al., 2022). Other common steps for EBC can involve protein precipitation or liquid-liquid extraction (LLE) to isolate specific metabolite classes before analysis by techniques such as LC-MS or Nuclear Magnetic Resonance (NMR) spectroscopy (Khoubnasabjafari et al., 2022; Wüthrich et al., 2024). For instance, Wüthrich et al. (2024) employed dynamic headspace vacuum in-tube extraction (DHS-V-ITEX) for analysis of volatile compounds from EBC prior to GC-MS, a novel application for this matrix, and simple filtration for LC-MS analysis of heavier, more polar compounds.

In some analytical schemes, particularly for LC-MS/MS, chemical derivatization is employed. This involves reacting target VOCs (e.g., carbonyl compounds like aldehydes and ketones) with a specific reagent (e.g., 2,4-dinitrophenylhydrazine, DNPH) to form derivatives. These derivatives often exhibit improved chromatographic properties, ionization efficiency, or detection sensitivity compared to the parent compounds, making them more amenable to LC-MS/MS analysis (Sani et al., 2023).

Throughout all these preparation steps, rigorous Quality Control (QC) is indispensable to ensure data reliability. This involves the consistent analysis of various types of blanks (e.g., sorbent tube blanks, system blanks, field blanks, ambient air blanks) to identify and subtract background signals and potential contaminants originating from sampling materials, the analytical system, or the environment (Westphal et al., 2023; McCartney et al., 2022). The use of internal standards, such as stable isotope-labeled analogues of target analytes (e.g., deuterated compounds like chlorobenzene-d5 or naphthalene-d8, as used by McCartney et al., 2022), is critical. These are typically added to the sample or sorbent tube at an early stage (e.g., before storage or immediately before analysis) to correct for variations in analyte recovery during sample preparation, desorption efficiency, injection variability, and instrument response drifts (Bajo-Fernández et al., 2024; Westphal et al., 2023). Regular analysis of calibration standards, standard mixtures (like Grob mix for GC-MS performance monitoring), or certified reference materials is also crucial for instrument calibration, performance verification, and ensuring the accuracy of quantification (McCartney et al., 2022; Hagens et al., 2021).

The demand for high throughput, consistency, and reduced human error in clinical breath analysis strongly advocates for automation in sample preparation and introduction. Automated systems for TD, SPME, SPE, and liquid handling for derivatization significantly reduce operator variability, increase sample processing capacity, enhance reproducibility and precision, and contribute to the overall standardization of the analytical workflow. These are critical factors for validating biomarkers, enabling inter-laboratory comparisons, and ultimately, for the successful translation of breath tests into routine clinical practice. The development of integrated and automated analytical systems, such as the TD-SIFT-MS instrument (Belluomo et al., 2024) or Point-of-Care (POC) devices like the octane breath test developed by Hagens et al. (2021), further highlights the field's progression towards more streamlined, robust, and clinically applicable methodologies.

The same study intriguingly proposed that breath 3-hydroxy-2-butanone (acetoin) levels might be associated with the sugar metabolism of specific oral bacteria, finding a correlation between these microbes and lung cancer, which underscores how breath VOCs can mirror complex systemic and microbial interactions.

Beyond lung cancer, the diagnostic potential of breath analysis is being explored for other malignancies, such as oesophagogastric cancer and mesothelioma, as well as various other respiratory conditions like Acute Respiratory Distress Syndrome (ARDS), for which octane has been investigated as a biomarker using a point-of-care breath test (Hagens et al., 2021). Furthermore, breath tests hold promise for monitoring disease progression and treatment efficacy, providing a dynamic and non-invasive assessment tool. The COVID-19 pandemic also spurred research into breath analysis for detecting viral infections, with studies like McCartney et al. (2022) showing that the VOC signature can differ even between viral variants (Delta vs. Omicron), highlighting the sensitivity of breathprints to specific pathological states and the importance of considering such variations in diagnostic model development. As highlighted by the work of Sani et al. (2023), breath VOCs can also offer valuable clues about underlying metabolic pathway alterations; for example, 2-pentanone levels might reflect changes in lipid metabolism in cancer patients. Alenezy et al. (2025) provide an overview of various biomarkers like acetone for diabetes, nitric oxide for asthma, and ammonia for kidney disease, further illustrating the broad diagnostic potential. While the array of potential applications is vast and promising, the field is actively working to overcome challenges, primarily the need for standardization, to translate these research findings into reliable, everyday clinical diagnostics.

For breath analysis to become a routine diagnostic tool, the methods used, from the moment a patient gives a sample to the final data interpretation, must be standardized, validated, and highly reproducible. This means:

• Standardized Collection Protocols: Which sampling method? Whole or alveolar breath? What about ambient air controls and minimizing pre-analytical errors (like diet or medication effects)? Di Gilio et al. (2020) compared different sampling devices (Tedlar bags, Mistral, ReCIVA) and highlighted variability, emphasizing the need to understand each method's characteristics and potential biases, such as the impact of clean air supplies or device-specific flow rates.

• Validated Analytical Methods: Ensuring that the techniques used to measure VOCs are accurate, precise, sensitive, and specific, as demonstrated by Belluomo et al. (2024) for their TD-SIFT-MS method for 21 target VOCs, and by Hagens et al. (2021) for their POC octane breath test.

• Robust Data Analysis: Developing reliable ways to process complex data and identify true biomarkers (with sample sizes that provide enough power to do so).

• Inter-Laboratory Comparability: Results from one lab need to be comparable to another. Ideally, automated methods should be shared easily.

This is where automation platforms play a pivotal role. By automating critical steps like thermal desorption, sample preparation, and injection, systems like the PAL System can significantly reduce variability, improve precision, encourage sharing methods and enable the high throughput necessary for large-scale clinical validation studies and eventual routine use.

1. Alenezy, E. K., Kandjani, A. E., Shaibani, M., Trinchi, A., Bhargava, S. K., Ippolito, S. J., & Sabri, Y. (2025). Human breath analysis; Clinical application and measurement: An overview. Biosensors and Bioelectronics, 278, 117094.

2. Bajo-Fernández, M., Souza-Silva, E. A., Barbas, C., Rey-Stolle, M. F., & Garcia, A. (2024). GC-MS-based metabolomics of volatile organic compounds in exhaled breath: applications in health and disease. A review. Frontiers in Molecular Biosciences, 10, 1295955.

3. Belluomo, I., Whitlock, S. E., Myridakis, A., Parker, A. G., Converso, V., Perkins, M. J., Langford, V. S., Španěl, P., & Hanna, G. B. (2024). Combining Thermal Desorption with Selected Ion Flow Tube Mass Spectrometry for Analyses of Breath Volatile Organic Compounds. Analytical Chemistry, 96(4), 1397–1401.

4. Bruderer, T., Gaisl, T., Gaugg, M. T., Nowak, N., Streckenbach, B., Müller, S., Moeller, A., Kohler, M., & Zenobi, R. (2019). On-line analysis of exhaled breath focus review. Chemical Reviews, 119(19), 10803–10828.

5. Di Gilio, A., Palmisani, J., Ventrella, G., Facchini, L., Catino, A., Varesano, N., Pizzutilo, P., Galetta, D., Borelli, M., Barbieri, P., Licen, S., & de Gennaro, G. (2020). Breath Analysis: Comparison among Methodological Approaches for Breath Sampling. Molecules, 25(24), 5823.

6. Hagens, L. A., Verschueren, A. R. M., Lammers, A., Heijnen, N. F. L., Smit, M. R., Nijsen, T. M. E., Geven, I., Schultz, M. J., Bergmans, D. C. J. J., Schnabel, R. M., & Bos, L. D. J. (2021). Development and validation of a point-of-care breath test for octane detection. Analyst, 146(14), 4605–4614.

7. Khoubnasabjafari, M., Afshar Mogaddam, M. R., Rahimpour, E., Soleymani, J., Saei, A. A., & Jouyban, A. (2022). Breathomics: Review of Sample Collection and Analysis, Data Modeling and Clinical Applications. Critical Reviews in Analytical Chemistry, 52(7), 1461–1487.

8. Lawal, O., Ahmed, W. M., Nijsen, T. M. E., Goodacre, R., & Fowler, S. J. (2017). Exhaled breath analysis: a review of ‘breath-taking’ methods for off-line analysis. Metabolomics, 13, 110.

9. McCartney, M. M., Borras, E., Rojas, D. E., Hicks, T. L., Hamera, K. L., Tran, N. K., Tham, T., Juarez, M. M., Lopez, E., Kenyon, N. J., & Davis, C. E. (2022). Predominant SARS-CoV-2 variant impacts accuracy when screening for infection using exhaled breath vapor. Communications Medicine, 2, 158.

10. Miekisch, W., Schubert, J. K., & Noeldge-Schomburg, G. F. E. (2004). Diagnostic potential of breath analysis—focus on volatile organic compounds. Clinica Chimica Acta, 347(1-2), 25–39.

11. Pauling, L., Robinson, A. B., Teranishi, R., & Cary, P. (1971). Quantitative analysis of urine vapor and breath by gas-liquid partition chromatography. Proceedings of the National Academy of Sciences of the United States of America, 68(10), 2374–2376.

12. Sani, S. N., Zhou, W., Ismail, B. B., Zhang, Y., Chen, Z., Zhang, B., Bao, C., Zhang, H., & Wang, X. (2023). LC-MS/MS Based Volatile Organic Compound Biomarkers Analysis for Early Detection of Lung Cancer. Cancers, 15(4), 1186.

13. Westphal, K., Dudzik, D., Waszczuk-Jankowska, M., Graff, B., Narkiewicz, K., & Markuszewski, M. J. (2023). Common Strategies and Factors Affecting Off-Line Breath Sampling and Volatile Organic Compounds Analysis Using Thermal Desorption-Gas Chromatography-Mass Spectrometry (TD-GC-MS). Metabolites, 13(1), 8.

14. Wüthrich, C., Vadakkechira, A., Fuchsmann, P., Wacker, S., Zenobi, R., & Giannoukos, S. (2024). Comparative analysis of feature annotation methods for SESI-HRMS in exhaled breath analysis. Journal of Chromatography A, 1734, 465296.