Issue link: http://www.e-digitaleditions.com/i/830814
Inhalation JUNE 2017 23 dimensions: 1.5 mm x 12 mm, see the inset in Figure 1) assists in damping residual flows. To increase the total number of imaged particles per experiment, the syringe pump then withdraws a small amount of aerosol that replaces the particles within the chamber with a new set of particles that have not yet been imaged. In this manner, at least three different sets of particles can be examined from a single aerosol sample. After the measurement, the solenoid valve opens for approximately 10 seconds to clean the system and the particles are collected by the HEPA filter (Figure 1A). Proof-of-concept experiments The results of APSD measurements from two com- mercial drug/inhaler combinations are presented in Figure 2 alongside published benchmark experiments obtained using a CI or the APS. The corresponding MMADs and geometric standard deviations (GSDs) are summarized in Table 1 for quick comparison. While the APSD measurements obtained with the new system for a Symbicort ® Turbuhaler ® (AstraZeneca, Cambridge, UK) inhaler agree very well with results from the Andersen Cascade Impactor (ACI) 11 and the APS, 12 an overestimation of MMAD is observed for measurements from a Seretide ® Diskus ® (GlaxoSmithKline, Brentford, UK) com- pared to the NGI 13 and the APS. 12 Note in particular the higher estimation of particle density in the size range of 4.46-8.06 µm (Figure 2A). Nevertheless, such discrepancies do not exceed differences previ- ously reported between APS and CI measurements. 3 While discrepancies between these measurements and the literature may result from biases associated with the traditional methods, sources of variability related to limitations of preliminary experiments with the sys- tem should also be considered. For example, the use of a straight tube as an induction port, rather than an L-shaped throat, can lead to biased results because the bent geometry is known to alter particle size distribu- tions through particle deposition and agglomerate breakup. 14 In addition, different batches of inhalers have been used, compared to published data. Another source of bias in these measurements may stem from a lower detection rate for smaller particles, a limitation that may be overcome by introducing a proper cali- bration scheme by measuring monodispersed particles of known size and density. Note also the higher vari- ability in the measurements for larger particles (com- pare Figures 1A and 1B). This may result from the lower number of particles measured for larger particles compared to smaller ones. For example, only 405 par- ticles were measured, on average, in a single measure- ment for a Seretide Diskus compared to 2,870 parti- cles for a Symbicort Turbuhaler. Obtaining more accurate measurements across the larger particle range would require increasing the number of measured particles, for example, by adding a second imaging system with a larger FoV. Finally, one should recall that the initial assumptions of spherical particles and constant density may add to observed discrepancies between these measure- ments and CI-based data. Despite these limitations, promising results have been obtained using this initial prototype. Specifically, aerodynamic particle size measurements with the device seem to be most accurate in the particle size range of 1-5 µm. It is expected these measurements APSD measurements obtained using the prototype device and published data from CI and APS measurements for two commercial DPIs (see text for references) Figure 2 Aerodynamic diameter (µm) Aerodynamic diameter (µm) Seretide Diskus NGI APS Prototype 0.55-0.94 0.94-1.66 1.66-2.82 2.82-4.46 4.46-8.06 8.06-12.0 0-0.4 0.4-0.7 0.7-1.1 1.1-2.1 2.1-3.3 3.3-4.7 4.7-5.8 5.8-9.0 9.0-10 ACI APS Prototype Symbicort Turbuhaler Mass fraction (%) A 60 45 30 15 0 B