Exhaled Breath Condensates: Analyzing the Expiratory Plume

The mixture of gases and droplets released during exhalation can be hydrodynamically designated as a “plume.” Traditionally, pulmonologists focused on respiratory gases in these plumes, but recent identification of biomarkers in exhaled breath condensates (EBCs) has initiated vigorous efforts to harness this simple, noninvasive approach for diagnostic purposes (>700 citations in PubMed since 2000).

A typical condenser removes about half of the approximately 45 μl of liquid water that can theoretically be extracted from each liter of saturated exhaled air. Nearly all of this water evaporates from lung surfaces, but approximately 1/10,000 of the condensate (∼4.5 nl) represents droplets released from the airway lining fluid (ALF). The assumption that these droplets are released from airways seems justified because alveolar convection is probably negligible and oral contributions can be minimized. ALF droplets can be generated by detachment from airways or rupture of bubbles from previously occluded airways (1), perhaps associated with audible rales.

Passage of exhaled air through the mouth at a site near the condenser makes it difficult to avoid contamination of EBC with salivary fluid. To ensure that saliva represents less than 10% of ALF volume, less than 10 nl of saliva should be collected in each milliliter of EBC, a daunting requirement. However, this was successfully accomplished in 36 of 40 samples when the mouthpiece was separated from the condenser, as judged by a sufficiently sensitive amylase assay (2).

Nonvolatile biomarkers (e.g., cytokines, ions, and urea) found in EBCs are exclusively derived from ALF droplets (Figure 1) (3–8). Dilution (D) of ALF by water vapor is generally quite variable, but D can be estimated with nonvolatile “dilutional” standard indicators (S), selected to have similar concentrations in ALF and plasma, for example, urea, cations, or conductivity following lyophilization: [nonvolatile]_ALF = D × [nonvolatile]_EBC, where D = [S]_ALF/[S]_EBC = [S]_plasma/[S]_EBC.

Figure 1.

Contributions of diffusion and convection to formation of the expiratory plume. Rapid diffusion of water (evaporation) occurs from the fluid lining the surfaces of the airways (e.g., bronchi), the airspaces (e.g., alveoli), and mouth into the expiratory flow. Exhaled water vapor remains in the gas phase until it is cooled and captured as droplets in the condenser. Droplet formation within the lungs is largely confined to the airways, where turbulence is encountered. These droplets contain airway lining fluid and are the source of virtually all of the nonvolatile biomarkers, ions, and dilutional indicators that reach the condensate. They are presumably formed by convective detachment from the surface fluid on the airway surfaces or by expulsion of aqueous films spanning airway openings (perhaps causing audible rales). Because the rate of droplet formation is so miniscule compared with the movement of gases from the lungs, transport of nonvolatile constituents is generally much slower than diffusive transport of volatile molecules. Most of the NH₃ and many other gases (e.g., from bacterial metabolism) diffuse into the exhaled stream from the mouth, nose, and upper gastrointestinal tract. High concentrations of these volatile molecules are much more readily maintained at sites (e.g., oral cavity) relatively remote from exchange vessels. Diffusion of respiratory gases occurs primarily between the gas, tissues, and blood in the alveoli, but the site of diffusion of more soluble vapors (such as acetone) is typically greater in the airways.

Lyophilization is used to extract water and volatiles (e.g., NH₃ and CO₂) from ice at temperatures below −50°C, and should not be confused with centrifugal evaporation at warmer temperatures, which may denature proteins. The EBC samples can be concentrated by reconstituting the dried samples in smaller volumes of pure water. D must be determined before conclusions can be reached about concentrations of nonvolatiles in ALF from EBC measurements. For example, increases in EBC nonvolatile concentrations may reflect collection of more ALF droplets rather than increased ALF concentrations, unless it can be shown that EBC dilution remains relatively unchanged. The reliability of specific dilutional indicators can be checked by showing that alternative markers yield similar values of dilution (3) and by documenting that incorporation of dilutional indicators reduces data variability (5–8).

Volatile biomarkers and water directly diffuse as gases into the expiratory airflow from fluid covering both airspaces and airways (epithelial lining fluid). Volatile solutes with tissue–air partition coefficients above 100 (e.g., acetone) are lost primarily from airways, whereas those with lower coefficients are mostly lost from alveoli (9). Correction for dilution of droplets by water vapor is unnecessary if most of the biomarker is gaseous. However, gas phase measurements may be preferable to EBC studies because the distribution of volatiles between exhaled gas, lung fluid, and condensates can be quite variable. Droplet formation can be minimized by keeping exhaled air warm and measuring partial pressures in the gas phase (e.g., with gas chromatography/mass spectrometry). Volatile biomarkers diffuse more rapidly through lipid membranes separating gas, tissue, and blood compartments of the lungs than nonvolatile indicators and are consequently much more likely to be influenced by plasma concentrations and extrapulmonary inflammation. Because they can also diffuse directly from the oral, nasal, and gastrointestinal tracts into the exhaled breath, increases in EBC concentrations of volatile markers may be unrelated to pulmonary disorders.

Diffusion and convection play additive roles in the transport of partially ionized acids and bases in exhaled air. However, diffusion of uncharged molecules of many volatile acids and bases overwhelms transport of ions, which is limited by the small volume of airway droplets in the EBC. For example, though concentrations of NH₃ are less than 1% those of NH₄⁺ in saliva, diffusion of gaseous NH₃ from saliva accounts for most of the NH₄⁺ found in the EBC (3, 10), and this transport promotes alkalinization of EBC. High salivary concentrations of NH₄⁺ and NH₃ are maintained by bacterial degradation of urea in a pool of fluid relatively remote from the circulation. Any NH₃ that is inhaled from the mouth or produced locally in the lungs should rapidly diffuse into the pulmonary and bronchial microvasculature (11).

Early studies suggested that EBC acidification is characteristic of asthma, indicating airway inflammation (12). The fundamental error in most EBC pH studies is the failure to measure the buffering capacities (β) of the ALF and EBC. These are essential for calculating ALF pH. NH₄⁺ is overwhelmingly the most abundant cation in EBC, and it is associated with nearly equivalent concentrations of HCO₃⁻ in samples exposed to room air or 5% CO₂ (3, 10, 13). Addition of weak or strong nonvolatile acids from the ALF generally has comparatively modest effects on EBC pH, which are difficult to detect. EBCs remain relatively alkaline unless acid concentrations in EBCs approach those of NH₄⁺ (14), which seldom occurs in oral EBCs_. A report that EBC pH is not influenced by NH₄⁺ concentrations (15) appears untenable and was not confirmed in subsequent studies (16, 17). Controversy persists regarding the best PCO₂ at which pH should be measured, but conventional flushing with argon does not remove all of the CO₂, fails to raise pH of EBC to expected levels, and can remove variable amounts of water and other acid–base pairs (10, 16). These problems may explain why EBC acidification was not confirmed in a recent large, multicenter study of individuals with asthma (18) and suggest that conventional studies of EBC pH should be abandoned.

In conclusion, EBC studies are best used for investigations of nonvolatile constituents, for example, electrolytes, nongaseous neutral molecules (e.g., sugars and urea), and macromolecules (e.g., cytokines and nucleic acids) in the ALF. In theory, this vast family of potential biomarkers can also be recovered by impaction, bubbling through water, or filtration. The initial enthusiasm regarding EBCs dimmed somewhat when the marked dilution and variability of EBC concentrations were fully appreciated. Although estimates of dilution may reduce variability found in EBC data, progress will require more sensitive and reproducible assays of all indicators. Nevertheless, it would be difficult to exaggerate the importance of noninvasively measuring airway biomarkers, and investigators should be encouraged to perfect analytical and collection techniques.