Collection of exhaled breath condensates is a novel and non-invasive method for obtaining samples of the upper respiratory tract. The collection of EBC is easy to perform and can be conducted in a home environment. This method is much more agreeable for the patient when compared to the unpleasant and invasive collection of nasal swabs, BAL, aspirates, etc. This aspect renders the method very attractive for routine laboratory diagnostics of viral infections. Most studies that perform breath analyses for viral detection use modified face masks, with a removable central region in electret or a removable Teflon filter on which exhaled particles impact [12–14]. With the RTube™ collection device, aerosolized particles of the airway lining fluid are precipitated into a condensate when the breath is cooled which serves as an immediate starting point for molecular testing.
Until now, this is the study with the largest subset of volunteers that investigated EBC as a specimen for the detection of respiratory viruses. Previous studies reported the inclusion of a limited subset of participants and investigated the presence of a limited number of viruses in the breath samples. The study performed by Fabian and colleagues, included 12 volunteers . Huynh and co-workers recruited 9 volunteers for exhaled breath sampling . In the study by Stelzer-Braid et al., 50 EBCs were analysed  and St-George et al. report the participation of 12 adults . These studies have focused on the detection of InfA and -B, PIV1-3, HRSV and HMPV, while we have screened the samples for a panel of 14 commonly circulating respiratory viruses. Based on the analysis of 99 EBCs (3 EBCs were excluded), our results support the exhalation of RV and InfB in 7% of our samples. Since many of the volunteers had already been experiencing symptoms for 1 to 7 days, we initially presumed that they were already recovering from the infection and were no longer exhaling the virus. For common cold infections it is suggested that a person may already be infectious for 1 or 2 days before experiencing any symptoms. However, in a second part of our study we started collecting EBCs in parallel with nasal swabs from patients presenting themselves to their medical doctor, 1 to 3 days after onset of symptoms. Only for 1 condensate the same pathogen was detected in both the EBC and the NS. The detection rate for respiratory viral pathogens in the NS was 46.8% which is much higher than the 7% detection rate in the EBCs. The low detection of virus positive condensates can therefore not be attributed to the fact that volunteers were no longer infectious. The discrepant detection rate between samples may also be explained by different severity of respiratory infection, since comparator samples were of different parts of the respiratory tract. Patients that delivered a positive NS may have possibly suffered from an upper airway infection whereas EBC positive volunteers may have experienced a more advanced, lower respiratory tract infection. However, the effect of nasal inhalation on EBC collection, guiding formed particles in the upper respiratory tract to the lower compartments, in stead of oral inhalation was not investigated. Patients with positive EBC samples were experiencing symptoms for maximum two days at the time of collection. However, this was not different for 7 patients with positive NS. Six patients that provided positive NS were experiencing symptoms for a longer period at the time of collection (Table 3). In the group of volunteers that provided an EBC negative or EBC and NS negative sample, the manifestation of symptoms were reported ranging from 1 day to more than two weeks. When reported symptoms were compared between EBC positive patients (7) and NS positive patients (15), 27% and 33% in the positive NS group experienced shivering and muscle pain whereas this symptom was not indicated by any patient of the EBC positive group. In all groups fever, headache, watering eyes, stuffed nose, frequent sneezing, sore throat and coughing were reported.
Volunteers were not diagnosed with other pathogens before participation in the study. Since we did not test these samples for other than viral pathogens, we can not exclude the possibility that some of the negative NS are positive for bacteria or other pathogens causing respiratory illness. Recently, one study reported a detection rate of 5% for influenza in EBC . This is in the same range of the detection rate that we report for respiratory viruses in general. Other studies with a limited number of patients, describe a markedly higher sensitivity of 33 to 36% [12–14] but the higher percentage may be due to the low number of participants subjects were included . Remarkably, the studies reporting this higher detection rate used collections masks, while the study using the RTube™ reported comparable findings. Face masks consist of electret which trap viruses based on permanently charged fibres . In addition, the Teflon filter has 2 μm pores which will retain all larger particles. Possibly, the lower detection rate can partly be explained by the fact that the RTube™ is manufactured in polypropylene and does not possess a virus attracting and filtering feature like the aforementioned materials.
The qRT-PCR developed by Lu and coworkers for the detection of RV, did not allow the assessment of the viral load present in the EBC samples . Also for 4 NS, the viral titer remained undetermined, probably due to the limited sensitivity of the assay. For diagnosis, more sensitive methods might be necessary to detect respiratory viruses present in EBC since it is unpredictable how diluted the viral particles in the specimen are. Recently, nested qRT-PCR assays have been developed to allow a more sensitive detection of viruses in aerosols .
Also person-dependent factors, such as the number of particles produced, the exhaled volume and the age of the patient, have been suggested to play an important role for exhalation of viral particles. The participants that were recruited in the study of Fabian and coworkers were 12 years of age and older . For hospitalized children a much higher rate of virus positive samples is reported . In our study, the majority of volunteers were between 20 and 30 years old. Only two children less than 10 years and 3 elderly people (> 70 years) were included. One of the children tested positive for InfA in the NS, but the infection was not confirmed in the EBC.
For influenza, an exhaled generation rate of <3.2 to 20 influenza RNA copies per minute was predicted by quantifying the virus aerosols that impacted on a removable Teflon filter of a collection mask . We used the RTube™ in combination with the ECoVent, that allowed the registration of additional ventilation parameters such as breathing frequency and exhaled volume. In this way, when the number of RNA copies in the EBC is quantified, the amount of viral particles that are exhaled per litre or per minute can be estimated. Unfortunately, we were not able to predict a virus generation rate for InfB since viral load remained undetermined.
Although an inventive, new and promising method, EBC collected by the RTube™ does not appear to be appropriate for diagnosis of respiratory infections. Nonetheless, this method may provide an alternative for current sample procurement for epidemiological studies of circulating viruses. This technique also confirms the observation that viruses are able to disseminate through normal breathing, particularly RV.
In addition, EBC collection from patients during respiratory infections may be further investigated for biomarker patterns. In calves that were experimentally infected with bovine RSV, an increase in leukotriene B4, indicating oxidative stress, was observed. This increased level was also associated with the development of bronchial hyperresponsiveness . In humans, a transiently elevated H2O2 level was observed during common cold infection. This marker returned to baseline values when volunteers recovered from infection. H2O2 has also been recognized as an interesting marker in asthma, where it is associated with chronic lower airway inflammation . In InfA infected volunteers, an increased CO level was observed during upper respiratory infection. This observation might imply that CO is an indicator of airway inflammation or represents one of the host defence mechanisms against viral infection . Therefore, a better identification of the biomarker signature in condensates of individuals experiencing a viral infection might imply interesting findings towards the identification of markers reflecting inflammation or antiviral protection. This may contribute to the biomarker profiles established for diseases like asthma and COPD, for which viral infections are suggested to trigger or exacerbate symptoms .