Before the recent publication of reports linking cyclo-oxygenase-2 (COX-2) inhibitors with myocardial infarction,¹ there was strong evidence that this popular group of analgesics does not have the respiratory complications of non-specific anti-inflammatory drugs (NSAIDS).² These side effects of the NSAIDs are relatively common (10–20% of asthmatics) and potentially very serious. Patients experience increased inflammation of the sinuses, nasal polyposis, and severe and potentially fatal airway obstruction. It was postulated that NSAIDs reduce production of prostaglandins that protect the airways and increase synthesis of cysteinyl leukotrienes such as leukotriene B₄ (LTB₄) which aggravate inflammation. The clinical observation that the COX-2 drugs are safe in this respect suggests that inhibition of the COX-1 enzymes is the culprit. It appears that asthmatics can use these very effective analgesics with relatively little risk of exacerbations.

Regardless of the eventual application or modification of NSAIDs and COX-2 inhibitors, widespread use of these drugs has taught us much about the role of the COX enzymes in the lung and other organs. The study by Montuschi and colleagues³ in this issue of Thorax serves two important functions: it supports the role of COX-1 activity in protecting the airways and it highlights the potential usefulness of exhaled breath condensate (EBC) as a simple non-invasive method for following inflammatory events in the lungs and the effect of treatment on these disorders.

EBC samples are collected by cooling expired air and collecting the water which condenses on the surfaces of the collecting device. Although this approach avoids the risks, discomfort, and expense of alternative procedures such as sputum induction, bronchoscopy and bronchial biopsy, it is associated with some rather formidable challenges of its own. Concentrations of all non-volatile solutes, including electrolytes and inflammatory markers, are very low and difficult to measure. This reflects the fact that more than 99.9% of the EBC is composed of water vapour (a gas) which only becomes a liquid when cooled in the condenser.⁴ Aside from the technical problems associated with measuring low solute concentrations, it cannot be assumed that changes in EBC concentrations provide reliable information regarding events in the airways. Increases in the concentrations of LTB₄ after NSAIDs could indicate either an increase in the concentrations in respiratory secretions or simply an increase in the number of respiratory droplets generated in the lungs that were then incorporated in the EBC.

There are two ways to compensate for changes in the dilution of respiratory droplets by water vapour. One approach is to monitor the EBC and plasma concentrations of a dilutional reference indicator such as urea or conductivity (these measurements must be conducted in samples which have been freeze dried to remove NH₄⁺ and HCO₃⁻).⁴ Montuschi et al have used an alternative approach. They have compared the EBC concentrations of a variety of markers and found that NSAIDs increase LTB₄ and decrease PGE₂ concentrations, a change that could not be attributed to an effect of dilution.

There is one puzzling aspect of this report which remains unexplained. Values of LTB₄ in this and a previous study conducted by some of the authors of the present report were about one sixth of the bronchoalveolar lavage (BAL) fluid concentrations in normal subjects and nearly the same as BAL fluid concentrations in the presence of pulmonary inflammation.⁵ Studies with dilutional indicators suggest that EBC concentrations of non-volatile indicators should be about 1% of those found in BAL fluid samples, which might be below the lower limits of their assays. This discrepancy highlights the importance of developing ultrasensitive assays of inflammatory mediators, and the usefulness of dilutional markers which permit calculation of the absolute rather than the relative concentrations of mediators in the respiratory fluid that lines the airways.