Variation in nebulizer output and weight output from Mefar dosimeter: implications for multicentre studies.

We wish to comment upon the paper of CHINN et al. [1] in which both weight and aerosol output data were given for 193 Mefar nebulizers from 27 centres involved in the European Community Respiratory Health Survey (ECRHS). A single centre in Turin determined nebulizer output using the sodium fluoride (NaF) tracer method described by DENNIS et al. [2]. The weight output was given as 11 mg·s-1 with a coefficient of variation of 5.8%. Aerosol output (AO), as a percentage of weight output gave a mean of 43% (i.e. mean AO of approximately 4.7 mg·s-1) with a between-centre range of 40–52% (AO range 4.4–5.7 mg·s-1). We regard these data from the single Turin centre with some concern, in view of our experience with similar Mefar nebulizers, which is supported by published data by DENNIS et al. [3], suggesting a considerably higher AO. They found a similar weight output to that given by CHINN et al. [1] but a considerable betweenbatch variation in Mefar nebulizer output, with mean aerosol output between batches ranging from 5.7–10.6 mg·s-1 (AO 51–75%). We recently re-analysed the outputs of a batch of Mefar nebulizers used in the ECRHS. We found a mean AO of 6.7 mg·s-1 (AO 61.4%). A batch of unused Mefar nebulizers analysed at the same time, had an AO of 7.1 mg·s-1 (AO 65.3%). Furthermore, we recently supervised the calibration of a batch of 10 Mefar nebulizers by the NaF method by a third centre (Freeman Hospital, Newcastle Upon Tyne; unpublished data) and obtained a mean AO of 6.9 mg·s-1. It seems unlikely that the aerosol outputs from five batches of Mefar nebulizers, as measured by DENNIS et al. [3] and ourselves, would differ so much from the outputs reported by CHINN et al. [1]. It would also seem unlikely that a modern welldesigned nebulizer system would have such a poor AO (43%) as that described by CHINN et al. [1]. One has to wonder whether this represents some systematic methodological artefact and it is also our experience that NaF nebulizer calibration is technically difficult even with extensive experience. Could we suggest that a number of batches of nebulizers used in the ECRHS survey be sent to ourselves and others with substantial experience of measuring aerosol output, so that the data presented by CHINN et al. [1] can be corroborated? It is vital to the credibility of the airway reactivity measurements made in the ECRHS survey that true aerosol output be known, and that potential between-batch variability is fully established and quantified.