Review of Non-Invasive Detection of SARS-CoV-2 and Other Respiratory Pathogens in Exhaled Breath Condensate

Abstract

In 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged to cause high viral infectivity and severe respiratory illness in humans (COVID-19). Worldwide, limited pandemic mitigation strategies, including lack of diagnostic test availability, resulted in COVID-19 overrunning health systems and spreading throughout the global population. Currently, proximal respiratory tract specimens such as nasopharyngeal swabs are used to diagnose COVID-19 because of their relative ease of collection and applicability in large scale screening. However, localization of SARS-CoV-2 in the distal respiratory tract is associated with more severe infection and symptoms. Exhaled breath condensate (EBC) is a sample matrix comprising aerosolized droplets originating from alveolar lining fluid that are further diluted in the distal and then proximal respiratory tract and collected via condensation during tidal breathing. The COVID-19 pandemic has resulted in recent resurgence of interest in EBC collection as an alternative, non-invasive sampling method for the staging and accurate detection of SARS-CoV-2 infections. Herein, we review the potential utility of EBC collection for detection of SARS-CoV-2 and other respiratory infections. While much remains to be discovered in fundamental EBC physiology, pathogen-airway interactions, and optimal sampling protocols, EBC, combined with emerging detection methods, presents a promising non-invasive sample matrix for detection of SARS-CoV-2.

Introduction

SARS-CoV-2 has been detected in both proximal respiratory tract (PRT) and the distal respiratory tract (DRT) of infected individuals. PRT specimens, which include saliva, nasal, anterior nares swabs (NS), oropharyngeal (OP), and nasopharyngeal swabs (NP), are widely utilized in most of the COVID-19 diagnosis and screening. While these samples are relatively easy to acquire, viral load fluctuations in the PRT during different infection stages cause false negative detections; which is a significant concern.^{1 2 3 4 5 6} For example, Green et al. found that 18.6 % of 2,413 patients with an initial COVID-19 NP negative PCR result, but with high suspicion of disease, yielded positive results when a retest was performed on new samples collected on the same or following day.⁷ These results indicate that approximately 450 subjects with false negative results could have unknowingly continued to spread the infection. In contrast, DRT specimens, such as sputum, tracheal aspirates, and bronchoalveolar lavage fluid (BAL), have been shown to be more sensitive for detection of SARS-CoV-2 infection, but these require significantly more invasive collection processes.^{8 9 10} Thus, a truly representative specimen that meets relevant criteria: (1) non-invasive sampling; and (2) allows for accurate diagnosis of COVID-19 and related illnesses, is of particular interest.

Because exhaled breath condensate (EBC) is simple to collect and the specimen is representative of the entire respiratory tract, EBC could be a model specimen providing critical information on virus activity.^{11 12 13 14} The promising status of breath profiling in disease diagnosis, especially in respiratory medicine, has grown as the field of modern chemistry continues to advance. Kuban and Foret summarized the historical contributions of notable scientists and inventors who pioneered rapid growth in the field of exhaled breath.¹⁵ As noted by Gould et al.in a 2020 review of exhaled breath analyses for viral detection, EBC is understudied compared to the more extensively evaluated volatile organic compounds (VOCs).¹⁶ Nevertheless, EBC samples contain numerous of non-volatile compounds including proteins, small molecules, and viral nucleic acids.^{15 17 18 19}

EBC comprises exhaled breath aerosols (EBA) from the airway lining fluid (ALF) which are produced during tidal or normal quiet breathing as well as water vapor and dissolved components. EBA accounts for 0.1% of the total EBC volume, and in normal tidal breathing, the size distribution of EBA particles fall under 5 μm.^{20 21 22 23} The large surface area of the pulmonary capillary bed enables the exchange of gases and secreted by-products of cellular metabolism between plasma and respiratory structures of the DRT.²⁴ Evaporation events cause the air surrounding the lung surfaces to be saturated with water vapor. The resultant moist air plays a critical role in maintaining the humidity of the pulmonary environment, preventing damage to the lung epithelial lining, and stabilizing gas exchange.²⁵ The solute component of EBC contains a medley of prospective biomarkers of various lung pathologies and closely correlates with aerosolized particles of the ALF.²⁰

The dilution of EBC during exhalation is presumed to result in high intra subject variability of collected EBC volume and solute concentration (Figure 1). Thus, a dilution correction becomes necessary for quantification of biomolecules in EBC. While EBA is collected via specially designed filters for particle impaction, EBC is commonly collected via simple condensation onto cooled surfaces. ^{23 26} As a non-invasive sampling method, EBC, offers a safe way for sampling ALF from intensive care patients requiring mechanical ventilation assistance.²⁷ These advantages make EBC suitable for longitudinal studies.

Figure 1.

Formation of respiratory droplets in respiratory tract regions. The ALF houses biomarkers of interest which can undergo aerosolization. Water vapor produced by evaporation events in ALF dilutes biomarker-containing droplets as they pass from DRT to PRT via exhaled air.

The objective of this review is to evaluate the utility of EBC for capturing and detecting respiratory viruses; with a particular focus on SARS-CoV-2. The scope will cover the theoretical underpinning of SARS-CoV-2 localization in the lungs and consider the potential for the non-invasive detection of SARS-CoV-2 in EBC.

Characteristics of SARS-CoV-2 interaction with humans

The RNA genome of SARS-CoV-2 extends nearly 30 kilobase pairs in size, which is typical of most B lineage ß-CoVs, and estimates of the virus size range between 65 to 125 nm in diameter.^{28 29 30 31} The genome encodes four structural proteins Spike (S) surface glycoprotein, Membrane (M), Envelope (E), and nucleocapsid (N) proteins that are important for host entry, viral assembly and spread within the body.³²

SARS-CoV-2 recognizes human cell membrane-localized zinc metalloenzyme and carboxypeptidase, human angiotensin-converting enzyme 2 (hACE2), as the primary functional host cell entry receptor.^{33 34 35 36 37 38 39} SARS-CoV-2 surface-expressed S glycoprotein mediates host cell recognition and adhesion to hACE2 target cells (Figure 2).

Figure 2.

Graphical illustration of SARS-CoV-2 structure and host cell attachment and entry model. Virus recognizes and enters cell by forming “hACE2/S” interactions on the host cell membrane. Reprinted from Naqvi et al. with permission.⁴⁰

Alveolar epithelial cells have the highest expression of hACE2 receptor, to which SARS-CoV-2 binds.^{39 41 42 43 44} No detectable epithelial cell surface expression of hACE2 receptors are found in the oral mucosa, nasal spaces, or nasopharynx.⁴¹ However, upward synchronized movement of ciliated lung epithelia during mucociliary clearance from the DRT are likely to cause detectable SARS-CoV-2 in the PRT.⁴⁵

Airway dispersion of SARS-CoV-2 in the respiratory tract

The respiratory tract begins from PRT, comprising the nares and buccal cavity, extending distally to the pharynx, larynx, and trachea. The DRT comprises the lower trachea bifurcating into left and right bronchi, one bronchus per lung. Each bronchus further divides into smaller bronchioles, continuing downwards through the terminal bronchioles and alveolar ducts until reaching the dead-end spaces of the alveolar sacs located deep within the lung.^{46 47 48 49} During pulmonary ventilation, the varying luminal diameters of PRT structures control the rate at which inspired air reaches the alveoli region of the DRT, where gas exchange occurs. Consequently, the processed air is then passed out as exhaust into the exterior environment. The opening and closing of the bronchioles regulate the amount of air that the lungs receive. Several mathematical models based on bronchial morphometric inputs have been built to simulate lung function. Modeling the relationship between SARS-CoV-2 particles and the external physical forces (inertia-induced impaction, Brownian motion induced-diffusion, gravitational sedimentation and interception) associated with particle deposition within the respiratory tract is pertinent to predicting respiratory sites with the highest contact with the virus (Figure 3).⁴⁶ Thus, to accurately predict the viral particle dispersion trend, parameters such as respiratory tract geometry, airflow dynamics, lung capacity, breathing rate, aerodynamic properties of SARS-CoV-2 particles, and subject’s health status must be taken into account.⁴⁶ Diffusion is the key transport phenomenon for particles below 0.5 μm which undergo Brownian motion. Consequently, as the size of particles and airflow rate decreases as in normal breathing, Brownian motion increases and is, therefore, an important mechanism for the deposition of particles in the lower lungs and alveolar regions.⁴⁶ Madas et al. applied a “stochastic lung deposition model” to probe exposure to cough-induced SARS-CoV-2 aerosolization and, consequently reporting a direct PRT colonization pattern. Their model appears to contradict reports of DRT serving as the initial site of infection.^{50 51} The model output, however, is backed by available clinical data on PRT symptoms appearing early in COVID-19 infection.⁵² The observed dispersion pattern could occur as a result of inertial impaction of cough-generated SARS-CoV-2 microdroplets (greater than or equal to 5 μm) in the PRT upon exposure, especially at the point of bifurcation of the trachea into the left and right bronchi.⁵³ Given conflicting reports of the virus localization, it is clear that SARS-CoV-2 could potentially be localized in multiple anatomical spaces within the respiratory tract and therefore, sampling of the entire dimension of the human respiratory tract is important for the accurate diagnosis of COVID-19.

Figure 3.

Distributional mechanics of aerosol particle deposition in the respiratory tract. Particles are subject to one or more of the specified dispersion forces based on their size distribution, air flow rate and location in the pulmonary region. Reproduced from Verma et al. with permission.⁴⁶

EBC production

A precise mechanism in which non-volatile solutes are aerosolized within the respiratory tract from the ALF has not been fully elucidated.¹⁵ The entire pulmonary surface is lined by a layer of ALF (approximately 50 μm thick, though ALF is thinner around the alveolar region) from which particles present in the lungs are released into the airway space by aerosolization and subsequently exhaled. Respiratory fluid droplets were previously hypothesized to be formed by the shedding of the pulmonary fluid-layered walls during normal breathing as a result of a turbulent flow of air in the respiratory tract.^{23 54 55} The turbulence model however, conflicts with experimental observations summarized in Rosias’ 2012 review of EBC collection and analysis. ²³ Micron-size droplets may be formed in large airways via turbulent energy from certain exhalation maneuvers such as coughing, singing, speaking and other activities that engage the vocal folds. The significant increase in exhalation flow rate should show a similar increase in the concentration of aerosolized particles being generated according to the model. However, experimental data showed that droplet production was observed to be much higher during inhalation than exhalation. In contrast, Hagen-Poiseuille’s equation, shows an inverse relationship between airflow resistance and airway diameter such that individual small airways with decreasing luminal diameter would thus exhibit greater resistance to airflow due to higher turbulence. However, the flow of air in the bronchioles and alveoli remains laminar because of the parallel branching of numerous bronchioles which collectively increases the surface area with a concurrent reduction in total airflow resistance.⁵⁶ The turbulence model thus, limits respiratory droplet formation during tidal breathing to PRT e.g., trachea, where turbulence-induced aerosolization is most likely.⁵⁷ Given these inconsistencies, alternative models have moved away from focusing on turbulence-induced particle shedding.

The bronchiole fluid film burst model (BFFB) (Figure 4), was proposed in 2009 by Johnson and Morawska.⁵⁸ The model directly associates aerosolization with bronchiolar closure and reopening during tidal breathing. Here, exhalation-induced closure of the fluid-layered walls of bronchioles contact each other forming a fluid-filled film. Subsequent reopening of the bronchioles during inhalation disrupts the integrity of the film, forcing it to burst open and release micron-size droplets that are then exhaled.^{58 59} The BFFB model is consistent with the following findings: (1) in deep exhalation, the resultant reduction in bronchial luminal diameter will enhance fluid film formation and thus, increase aerosol generation during reopening; (2) in deep exhalation, residual volumes in the alveolar region can be accessed which will improve aerosolization yield;²³ (3) evidence of the presence of alveoli surfactant molecules in EBC indicating a possible alveolar origin of respiratory fluid droplets. While respiratory droplet formation during tidal breathing is most probable via the BFFB model, this model does not account for droplet formation under similar breathing conditions in larger proximal airways where a greater exhalation force would, in theory, be required to generate fluid films.⁵⁹ Ultimately, a more idealized model would combine aspects of both the turbulence and the BFFB models in the generation of respiratory droplets. The turbulence model lacks aerosolization due to surface tension while BFFB focuses on the DRT generation and does not account for possible droplet formation in the proximal ends of the respiratory tract. As EBC is sampled from the entire respiratory tract, the development of a model that accounts for a continuum of droplet formation from the DRT through the PRT is an important area of future research in the field.

Figure 4.

Theoretical illustration of the BFFB model of respiratory fluid droplet formation. During exhalation, the bronchioles contracts thus, creating a fluid filled film that ruptures upon inhalation. Reproduced from Kubáň et al. with permission.¹⁵

EBC collection

Together, the American Thoracic Society and European Respiratory Society Task Force have established best practices and methodologies for the collection of EBC which was efficiently summarized in a review by Horváth et al.¹⁷ It is important to note that the task force provides basic recommendations and that parameters such as collection temperature, device type, material composition of collection tube, and duration of collection, might vary according to the physical and chemical properties of the bio-analyte under investigation.

Both commercial and non-commercial EBC collection devices rely on the basic principle of condensation which involves the rapid cooling of exhalates at very low temperatures (5°C – 10°C) onto inert surfaces where the liquid portion of the exhalate condenses.⁶⁰ The inert surface of the collection tubes ensures no reactivity with the biomolecules of interest, thus maintaining their chemical integrity. Ten (10) minutes of tidal breathing in adults and 20 minutes in children generally yields as much as 1 to 3 mL of liquid condensate.^{21 60} A 2013 review by Ahmadzi et al. further investigated the extent to which commercially available EBC collection devices and parameters influence EBC biomarker profiles.⁶¹ Some commercially available collection apparatuses are portable and can be used in both inpatient and outpatient care. The RTube for example, comes with a one-time use propylene tube that serves the function of a condenser and a collector while low temperatures are maintained by the cooled aluminum sleeve. The device also employs a multi-purpose exhalation valve that can also be used as plunger to push fluid from the walls of the condenser.²³ Further details on commercially available EBC collection devices and methodologies can be found in the review by Rosias et al.²³

In addition to commercial devices, EBC has been successfully collected using non-commercial collection apparatuses of varying designs. Standard design (Figure 5) consists of a long tubing or collection duct lined in the interior with Teflon and a double-walled glass condensing system. The mid end of the duct is immersed in an ice containing bucket such that, the subject breathes out into one end of the tubing while the distal end empties into a collection tube. ^{21 27 60 62}

Figure 5.

A traditional in-house EBC collection device setup. The exhaled breath is warm and moist and once in contact with ice cold surfaces of the collection duct, a temperature gradient is generated causing condensation. Reproduced from Tankasala et al. with permission.⁶⁰

Detecting respiratory pathogens in EBC

In the last decade, multiple research groups have isolated the genetic material of respiratory pathogens in exhaled breath. Some studies used specially designed face masks to collect aerosolized virus particles via impaction from EBA,^{63 64 65} while others utilized in-house designed or commercially available EBC collection devices to collect samples for analysis.^{66 67} There have only been a handful of studies focused on detecting the presence or absence of respiratory viruses/pathogens in EBC or EBA (Table 1).^{63 64 65 66 67} More so, due to the recent emergence of COVID-19, it is unsurprising that, at the time of writing this review, few studies addressing the detection of COVID-19 in EBC had been published.^{4 68 69 70} Many of these studies demonstrated the presence of SARS-CoV-2 in EBC and EBA, though the positive detection rate in EBC and EBA compared to various reference sampling methods was highly inconstant. In the future, comparing EBC and EBA collection to multiple more established sampling matrices such as saliva, OP, and NP swabs within the same study would enable a more detailed assessment of the viability of EBC and EBA sampling for SARS-CoV-2 detection. In particular, quantifying SARS-CoV-2 viral load in EBC samples compared to viral loads in these more established sample matrices could elucidate important aspects of disease severity and pathological outcomes.

Table 1:

Summary of literatures on respiratory pathogen detection in EBC

Pathogen(s)	EBC/EBA capture method	Detection Method	No. of subjects	% Positive by EBC/EBA	% Positive by reference sample	Citation
SARS-CoV-2	RTube	RT-PCR	40	93.5	51.6, NP	Ryan et al., 2021 4
SARS-CoV-2	In-house EBA/EBC collection systems	RT-qPCR	21	25	79.2, Sputum 78.9, Feces	Feng et al., 202171
SARS-CoV-2	RTube, RTubevent device	RT-PCR	48	31.3	N/A	Sawano et al., 202172
SARS-CoV-2	Electret air filter-based device	RT-qPCR	15	70	100, OP	Malik et al., 202173
SARS-CoV-2	BioScreen II device	RT-PCR	14	22*	28, Throat swab	Zhou et al., 202170
SARS-CoV-2	Swirling Aerosol Collection (SAC) device	RT-PCR	39	11.1	37, Throat swab	Li et al., 2021 14
Influenza Virus	In-house EBC collection device	aqf-PCR	30	73.3	100, Nasal swab	Li et al., 2021 74
SARS-CoV-2	BioScreen	RT-PCR	35	16.7	5.4, Surface swab 3.8, Air samples	Ma et al., 2020 69
M. tuberculosis	In-house Respiratory Aerosol Sampling Chamber (RASC)	MS	36	NA	N/A, Sputum with GeneXpert	Chen et al., 2020 75
14 common PRT viruses	ECoVent/ RTube	RT-PCR/Gene sequencing	102	7	46.8, NP	Houspie et al., 2011 66
Human papilloma virus (HPV)	EcoScreen	INFINITI HPV-QUAD assay/gene sequencing	157	17.7	19.7, Bronchial brushing/neo-plastic lung tissue	Carpagnano et al., 2011 76
Influenza virus	RTube	RT-PCR	19	5.3	66.7, NP	St. George et al., 2010 67
Rhinovirus, influenza, parainfluenza 3, human metapneumovirus	Collection mask	RT-PCR	50	36.4	63.6, Nasal mucus	Stelzer-Braid et al., 2009 65
Influenza Virus	Teflon filter	qPCR	13	33	100, Nasal swab	Fabian et al., 2008 64
Rhinovirus, respiratory syncytial virus, influenza virus, parainfluenza (1–3), human metapneumovirus	Electret-based Aerosol capture mask	PCR	20	66.7	100, Nasal mucus	Huynh et al., 2008 63

^*22.2% collection failure rate.

^aquantitative fluorescence PCR

Of particular interest is a study by Ryan et al. (late 2020) that showed evidence of SARS-CoV-2 detection in EBC compared to NP.⁴ By utilizing reverse transcription polymerase chain reaction (RT-PCR), they reported a 93.5% virus detection rate in EBC samples of patients with pre-confirmed COVID-19 cases and exhibiting symptoms. In contrast, NP swabs showed only 51.6% detection. Motivated by the higher false negative detection rates of SARS-CoV-2 in commonly used NP specimens, this study strongly establishes EBC as an alternative specimen for diagnosing COVID-19 cases.

In PRT viral infections, NP swab sensitivity is superior to EBC sampling. Houspie et al. conducted a screen of 14 different circulating PRT viruses in EBC samples of 102 symptomatic volunteers by employing nucleic acid extraction techniques followed by RT-PCR analysis. Their result showed a 6-fold decrease in EBC detection rates for Rhinovirus, Human Respiratory Syncytial Virus B, Influenza A and B compared to the control NP specimens.⁶⁶ In another study by St. George et al., influenza virus was detected by nucleic acid-based approach in one out of EBC samples collected from 19 symptomatic volunteers.⁶⁷ In both studies, subjects underwent similar EBC collection procedures; greater than 1 mL volume collected within 10 minutes of tidal breathing through the mouth using commercial RTube condensers. Given that the strength of any diagnostic assay is measured by their specificity/sensitivity values, it is possible that data presented by the different groups may have been limited by the sensitivity window of the developed assay. However, even after prior EBC sample enrichment, NP specimens collected in parallel to EBC demonstrated much higher virus detection rates. Importantly, considering that samples (EBC and NP) to be compared represent different areas of the respiratory tract, individual sample positivity rates may depend on the severity of the infection and the pathogen of interest. For instance, higher viral detection in NP is more likely to be seen amongst patients experiencing PRT viral infections, such as influenza.

In addition to direct molecular detection of respiratory pathogens via EBC, researchers have shown signature-based detection of bacterial pathogens using mass spectrometry. Chen et al. have shown that mass spectrometry can be used to detect molecular signatures of Mycobacterium tuberculosis (Mtb) in exhaled breath particles of EBA.⁷⁵ They extracted lipid content from exhaled breath particles for a signature-based detection of subjects infected with Mtb compared to those without TB. The detection hinges, however, on extensive sample preparation of the lipid content from the EBA.

While lipid sample preparation remains a challenge for moving many mass spectrometry techniques to the clinic, emerging methods to integrate and obviate extensive sample preparation are underway. Chen et al. describes an octadecyl (C18) bonded resin-based capture and concentration of proteins and high molecular weight species from DRT during EBC sampling.⁷⁷ The ability to efficiently capture and identify proteins and other high molecular weight species shed from deep lung tissue is a valuable resource in pathogen detection. In a different approach, Xie et al. use multiple reaction monitoring profiling (MRM-profiling) to distinguish lipid classes with similar reaction profiles and were able to distinguish differences between various E. coli strains without having to extract and identify each specific lipid.⁷⁸ The use of MRM-profiling requires significantly less lipid sample preparation and may allow pathogen detection directly from EBC samples in the future.

Estimate of virus (SARS-CoV-2) yield in EBC of infected persons

We utilize a single compartment mathematical model originally developed by Riediker et al. to calculate the expected amount of SARS-CoV-2 viral load present per sample of EBC.⁷⁹ The model was originally developed to simulate an enclosed space where virus emission patterns in the EBA mimicked those of normal breathers and cough emitters. Each category is further subdivided into low-to-moderate-to-high emitters. The model assumes that respiratory droplet formation is by bronchiole reopening and that a correlation exists between EBC and sputum samples. Their establishment of a mass-balance equation predicted a cumulative total viral load emission in subject’s exhalates (microdroplet size up to 20 μm in diameter) at 4.9 × 10⁻⁶ viral copies/mL for a moderate emitter and 6.37 × 10⁻¹viral copies/mL for a high emitter (Table 2).

Table 2:

Calculation of the expected concentration of SARS-CoV-2 viral load in exhaled microdroplets entering EBC under normal breathing conditions.

EBC	Moderate Emitter	High Emitter
*Expected virus yield per tidal volume	2.45 × 10−3 virions/breath	3.19 × 102 virions/breath
Total breaths over (10 – 15) min of EBC collection at median RR of 16 breaths/min 80	≥ 160 breaths	≥ 160 breaths
Expected volume of EBC collected from 10 – 15 min of normal breathing	≥ 1mL	≥ 1mL
Exhalate viral load in EBC per mL	≥ 3.92 × 10−1 copies/mL	≥ 5.10 × 104 copies/mL

^*Viral load estimations are based off of the single-compartment model by Riediker et al.,⁷⁹ tidal volume of 500 mL,⁸¹ and respiratory rate (RR) = 12 to 20 breaths/min.⁸⁰

Since EBC is collected under tidal breathing conditions, our estimate leveraging these volumes is based on the normal breathing virus emitters exhibiting minor or no symptoms and, therefore, these individuals may or may not know of their infection status. These individuals thus increase the probability of viral transmission in their surrounding community.^{82 83} We estimate viral load concentration in EBC using the model output bound by the following assumptions: (1) virus particles in EBA and EBC are similar regardless of phase transformation in the latter; (2) a similar breathing pattern is observed for both moderate and high normal breathing emitters regardless of distinct stages of infection; (3) 10 – 15 minute collections will yield between 1 – 3 mL of EBC at a tidal volume of 500 mL; (4) high efficiency EBC collection means no significant viral particle losses. It is important to note that these predictions are valid within the boundaries of the stated assumptions. In reality, while there is a high chance of multiple localizations of SARS-CoV-2 in the different respiratory anatomical regions, the efficiency of ALF particle aerosolization is nonuniform and constantly affected by airway particle location, volatility, solubility, and molecular polarity.⁸⁴

The viral load of the high emitters, exhaling 5.10 × 10⁴ copies/mL, falls well above the limits of detection for most RT-PCR assays (100 – 8000 copies/mL).^{85 86 87 88 89 90 91} However, reliable detection of SARS-CoV-2 in EBC samples of moderate emitters, exhaling approximately 0.39 viral copies/mL, will likely require nucleic acid extraction and purification techniques to enrich samples by at least 300x prior to amplification and detection by RT-PCR. While the lack of direct detection of virus in the sample reduces the utility of EBC collection as a point-of-care molecular diagnostic tool, EBC collection remains a valuable laboratory testing matrix. Further, detection of host-response signatures to SARS-CoV-2 remains an active area of research that does not require direct viral presence in the sample.

Challenges in EBC matrix sampling

Despite the long history of EBC collection, with the very first EBC paper published in 1979 by Larson et al., and the innumerable potential applications in respiratory medicine, research in this area has progressed rather slowly.⁹² Defining the absolute concentrations of individual solutes in EBC relative to ALF remains a difficult challenge in the standardization of EBC sampling.^{20 21} Until a gold-standard of validation is proposed, reliable quantification of EBC biomarker levels poses a significant challenge.

Particles of varying droplet size emanating from the pulmonary mucosa are continuously diluted with water vapor, including vapor formed in PRT ventilated regions that do not engage in perfusion as they exit the respiratory tract. The variability in water vapor dilution causes a wide distribution in the levels of EBC biomarkers observed between collection.^{15 25 20 60 93 94 95} Hence, it is unclear if reported changes in the concentrations of EBC components reflect the rate of endogenous production of a biomarker or mere oscillations in droplet formation. By calculating a dilution normalization term that abides by the assumption that there are equal concentrations of reference parameters in both the ALF and plasma, the effect of water vapor variability on EBC biomarker sampling can be minimized. Effros and colleagues have devised strategies for estimating the dilution factors based on Na⁺ and K⁺ ion profiling, comparing conductivity of lyophilized EBC, and quantification of total protein and urea.⁹⁶ A more recent approach devised by Tankasala et al. utilize a temperature-driven process for the selective collection of DRT-enriched EBC with little to no contributions from PRT air. Breath that is exhaled is sorted through the actuation of a two-way valve as the temperature of the expired air reaches a pre-set threshold percentage.⁹⁷

Finally, ALF dilution during exhalation results in extremely low concentrations of non-volatile solutes. These concentrations are near, or sometimes below, the lower limit of detection for many diagnostic assays. For example, a 1 mL volume of EBC only contains about 0.1μL of relevant ALF solutes.⁹⁸ Even exogenous entities such as viral pathogens, can be highly diluted in EBC and require extremely sensitive detection. However, this same dilution and contamination of ALF from secretions in the PRT, may further enrich EBC with SARS-CoV-2 viral particles, proteins, and DNA from secretions of the nose, and mouth.

Conclusion

EBC involves the rapid condensation and collection of respiratory droplets of ALF containing particles of interest produced under quiet exhalation through the mouth. The precise origin of these droplets is still unknown, although, the BFFB mechanism is largely referenced because it closely captures biomarker aerosolization within the DRT. EBC has been shown to contain respiratory viruses including SARS-CoV-2, HPV, rhinovirus, human respiratory syncytial virus B, influenza A and B. The attractiveness of EBC specimen for the detection of COVID-19 hinges on (1) EBC collection as a safe, non-invasive collection method providing an opportunity for continuous sampling of the pulmonary system; (2) the evidence of viral spread through respiratory droplets; (3) the correlation with DRT specimens in terms of biochemical composition and origin; (4) the advantage of sampling the entire respiratory tract.

There is limited research on respiratory virus detection in EBC and EBA and there remains insufficient body of work in this field. While most studies used molecular detection methods on EBC samples, the potential for signature-based detection using spectrometry is an emerging area of interest. Due to the low concentrations of pathogen biomarkers in the EBC, host biomarker detection of non-volatile inflammatory markers/immune defense cells (cytokines, chemokines, T-lymphocytes) may also be of interest.^{99 100} Alternatively, nucleic acid concentration and extraction will enable more sensitive detection of pathogens within the EBC compared to direct viral measurement alone. Ultimately, EBC is an understudied, yet promising, sampling matrix which can provide relevant insights into an individual’s respiratory health. However, much more fundamental experimental research, modeling of key interactions, and translational clinical work may be required to fully uncover the potential of EBC for respiratory infection staging and diagnosis.