Bacterial colonization as a potential source of nosocomial respiratory infections in spirometers.

We read with interest the article by BURGOS et al. [1], which is a welcome addition to the literature on the potential problems of cross-infection of patients from respiratory testing equipment. Although few water-sealed spirometers are now used in the UK, the data provide valuable information on the potential types of microorganisms encountered during testing in patients without clinically apparent respiratory infection. We have two comments on this study. Firstly, the method of sampling from both devices may not provide the true picture of contamination. It is likely that, by sampling from discrete areas using swabs, colonies of organisms may have been missed completely. An alternative method may be to wash the entire area under investigation with a broth solution and then to culture samples of this broth accordingly [2]. This may well have provided a better picture of overall contamination of the tubing and spirometer bell. In our study [2], we recovered 82% of organisms from the spirometer tubing, whereas in the study by BURGOS et al. [1] only 13% of organisms were recovered from the pneumotachograph system. Secondly, BURGOS et al. [1] suggest that bacterial and antiviral filters may be a solution to preventing contamination of respiratory equipment, and readers may be lulled into thinking that this simple solution will solve cross-contamination problems. However, several studies have shown that, in practice, the filters currently available are not as efficient as expected. Although KIRK et al. [3] stated that the Pall Pf305 filter had a bacterial efficiency of 99.9%, this has subsequently been demonstrated to be incorrect both by ourselves [4] and by CLAYTON et al. [5] for the Pall and other filters. A more accurate figure is 66% efficiency [4]. The differences may be explained by the experimental methods used, in particular the assumptions made in calculating filter efficiency and in the flow rates at which testing took place. There are few studies that have used flow rates comparable to the peak flow rates encountered in routine lung function testing. Much of the early bench testing used flow rates of 50–100 L·min-1, which clearly do not reflect the conditions during actual measurements of peak expiratory flow (PEF) or forced vital capacity (FVC). Would the authors be willing, therefore, to indicate how effective bacterial filters are in protecting respiratory equipment from contamination and patients from possible cross-infection? In the light of our experience, we are cautious about the value of using filters routinely until a filter is produced for which convincing evidence of efficacy is provided.