Exhaled breath condensate (EBC) may be collected from breath saturated with water vapour, via a cooling system. EBC contains volatile and non-volatile substances, which can be used to monitor infl ammatory lung diseases. These include endogenous substances, such as hydrogen peroxide, thiobarbituric acid-reactive substances, isoprostanes, prostaglandins, leukotrienes and nitrite/nitrate as the end products of NO metabolism. These non-volatile mediators in exhaled breath can be used as markers of oxidative stress and infl ammation to investigate, and perhaps monitor, lung diseases . In contrast to bronchoalveolar lavage, collecting exhaled breath condensate is a totally non-invasive and a safe way to assess breath constituents and it can be repeated within a short period of time, even in asthmatic children. It does not infl uence airway function, nor induce infl ammation, unlike bronchoalveolar lavage or sputum induction . Scheideler et al.  reported that the origin of the proteins detected in breath condensate are partially from the naso-oropharyngeal tract and partially from lower regions of the airways.
Collection of exhaled breath condensate
Exhaled breath condensate collection can be performed using either commercially available equipment or simple cooling devices with glass condensing chambers whereby an inner chamber is suspended in a larger glass chamber and cooled by means of ice. Subjects are instructed to take normal tidal breaths, exhaling through a cooling system to form a condensate. In addition to simple glass collection devices, successful collection has been reported using a variety of devices, such as Tefl on-lined tubing in an ice-fi lled bucket and a commercially available condenser (Erich Jaeger GmbH, Hoechberg, Germany) .
During collection, nasal secretions, saliva and sputum may contaminate EBC. The following modifi cations may help to exclude nasal contamination: (1) inhalation and exhalation without a nose-clip; and (2) exhalation against a resistance to ensure velum closure and minimize nasal contamination . To prevent salivary contamination, subjects should rinse their mouth before collection and keep the mouth dry by periodically swallowing their saliva. Measuring amylase concentrations in samples can exclude signifi cant salivary contamination . A number of factors affect EBC volume during collection which include the temperature and humidity of exhaled air. Water loss in the expired breath is not great except under extreme conditions of exercise at altitudes when minute ventilation is very high [83,84]. We have demonstrated that EBC volume is proportioned to tidal and minute volume and also showed that deep breathing to vital capacity can give an signifi cantly larger yield of EBC than tidal breathing . Other studies have shown that controlling breathing by using visual cues can improve reproducibility of EBC volume, but nonetheless poor repeatability of infl ammatory markers may be seen .
Markers in exhaled breath condensate
Nitric oxide is difficult to measure because it is a free radical which reacts rapidly with oxygen, superoxide, water, thiols, amines, and lipids to form products with biochemical activities ranging from bronchodilation to cytotoxicity . Nitrates and nitrites are products of nitric oxide metabolism, which can be detected in EBC. Nitrite is rapidly oxidised to nitrate in aqueous solutions. Moshage et al. reported that in whole blood >95% nitrite is very rapidly oxidized to nitrate within 1 hour , thus a single measurement of plasma nitrite alone is probably meaningless.
Production of nitric oxide is generally increased in infl ammatory diseases, including asthma. Nitrite/nitrate levels in EBC are raised in asthmatic patients , which may be clinically useful in the management of infl ammatory lung diseases. Corradi et al. measured oxides of nitrogen (NOx) in EBC and showed that NOx values in smokers, asthmatic and community-acquired pneumonia patients were significantly higher than normal controls and decreased during the recovery phase of pneumonia [90,91].
We have compared total NOx levels in EBC and plasma among smoking, ex-smoking and non-smoking subjects. The total NOx concentration in samples of exhaled breath condensate and plasma was quantifi ed as nitrite by the method of nitrate reductase and the reaction of nitrite with 2,3-diaminonaphthalene (DAN) . We demonstrated that while plasma NOx concentrations in smokers are signifi - cantly higher than either non-smokers or ex-smokers, there was no correlation between plasma and EBC NOx levels . This indicates a disparity between NOx in EBC and plasma, and the complexity of NO regulation is further illustrated by the fact that exhaled NO decreases after smoking, while NOx in EBC appears to increase [75,76].
Vascular endothelial and smooth muscle cells, together with infl ammatory cells, especially eosinophils, produce superoxide anions (O2 –) and release several reactive oxygen- and nitrogen-derived species such as NO [94,95]. Superoxide anion radicals and nitric oxide react rapidly to form the peroxynitrite anion (ONOO–), a potent oxidant capable of damaging lipids and proteins in biological membranes, and which can also cause airway hyperresponsiveness and enhancement of infl ammatory cell recruitment [95–97]. Increased peroxynitrite together with elevated expression of iNOS have been observed in asthmatic patients compared with control subjects . Peroxynitrite is able to add a nitro group to the 3-position adjacent to the hydroxyl group of tyrosine to form nitrotyrosine .
Abundant expression of iNOS with the formation of nitrotyrosine in airway epithelium and infl ammatory cells has been observed in bronchial tissue from asthmatic patients compared with little or no nitrotyrosine in normal airway epithelium. A signifi cant inverse correlation was demonstrated between the nitrotyrosine in infl ammatory and epithelial cells with pulmonary function in patients with asthma . This suggests that nitrotyrosine is a marker of oxidative stress in asthmatic airways, owing to the signifi cant correlation between nitrotyrosine in EBC and exhaled NO.
Glucocorticosteroid treatment reduces the formation of nitrotyrosine in asthmatic patients [94,95]. Elevated nitrotyrosine in EBC has been reported in patients with cystic fi - brosis and increased nitrotyrosine together with ecNOS and iNOS expression in the pulmonary lesions of human tuberculosis has been reported by Choi et al. [99,100].
Glutathione (GSH) is a low molecular weight and endogenous thiol in human lung and the reaction of nitric oxide and glutathione can form nitrosothiol, which is detectable in EBC . Corradi et al.  noted high levels of nitrosothiols in EBC of smokers and patients with severe asthma, chronic obstructive pulmonary disease (COPD) and cystic fi brosis (CF). A positive correlation between nitrosothiol values and smoking history in current smokers was also found. Nitrosothiols were not found to be elevated in mild asthmatic patients, and it was suggested that this may be helpful to classify asthmatic severity .
Other non-nitrogenous markers
Hydrogen ions, measured as the pH value of EBC, is another simple, reproducible and useful marker for airway infl ammation. Kostikas et al. reported that a signifi cantly lower EBC pH value is present in patients with COPD and bronchiectasis compared with either asthmatic patients or control subjects . Patients with moderate asthma had signifi cantly lower pH values compared with mild asthmatic patients, and EBC pH is lower in cystic fi brosis children than in healthy control subjects, in addition, the EBC pH of CF patients with an exacerbation was signifi cantly lower than that of stable patients with CF [102,103].
van Beurden et al. showed that hydrogen peroxide (H2O2) was a reproducible marker in EBC which can remain stable for a period of up to 40 days of frozen at –70°C . Hydrogen peroxide in EBC was reported to be increased during the common cold, and returned to normal within 2 weeks of recovery in otherwise healthy subjects, thus this marker appears to refl ect upper respiratory tract infl ammation . Expired breath condensate H2O2 is not elevated in patients with cystic fi brosis although others have indicated that the variation in expiratory fl ow rates may affect and limit the usefulness of exhaled hydrogen peroxide as a marker of airway infl ammation [106,107]. A recent study showed that mean concentration of H2O2 was significantly elevated in patients with COPD compared to control subjects, and H2O2 levels in patients with severe and moderate COPD were signifi cantly higher than those with mild COPD .