Evaluation of probability density functions to approximate particle size distributions of representative pharmaceutical aerosols

The purpose of this work was to evaluate the application of probability den sity functions (PDFs) and curve-fitting methods to approximate particle siz e distributions emitted from four pharmaceutical aerosol systems characteri zed using standard methods. The aerosols were produced by a nebulizer, pres surized metered dose inhaler (pMDI), dry powder inhaler (DPI) and nasal spr ay. PDFs selected for analysis were (i)log-normal, (ii) upper-limit, (iii) Nukiyama-Tanasawa (iv) Rosin-Rammler and (v) modified Rosin-Rammler. Two cu rve-fitting methods were used to estimate the adjustable parameters of the PDFs: linear least-squares fit of the cumulative distribution function (Met hod A) and non-linear least-squares fit of the probability density function (Method B). Large truncation of the pMDI and DPI particle size distributio ns obtained by cascade impaction resulted in poor fits of the PDFs. The neb ulizer and nasal spray were not affected by truncation and were well repres ented by the Rosin-Rammler and log-normal PDFs, respectively (Method B only ). Probability distribution functions (Method B) were fitted without bias f rom linear coordinate transformation and produced significantly better fits than Method A for each aerosol system (p < 0.05). Considerable caution mus t be used when estimating representative parameters from cumulative or prob ability distributions. (C) 2000 Elsevier Science Ltd. All rights reserved.