A first-principles model for prediction of product dose uniformity based on drug substance particle size distribution.

The unit dose uniformity (UDU) of low-dose drug products can be affected by active pharmaceutical ingredient (API) particle size. UDU relative standard deviation increases as the fraction of large API particles increases and/or as the unit dose decreases. Control of API particle size has traditionally been based on the empirical relationship of d(90) and/or d(50) statistics to drug product uniformity. Several articles have been written that have identified a theoretical relationship between these particle size statistics, dose, and the probability of meeting US Pharmacopeial UDU testing criteria (Huang CY, Ku S. 2010. Int J Pharm 383:70-80; Rohrs B, Amidon G, Meury R, Secreast P, King H, Skoug C. 2006. J Pharm Sci 95(5):1049-1059; Huang CY, Ku S. 2010. J Pharm Sci 99:4351-4362; Yalkowsky SH, Bolton S. 1990. Pharm Res 7(9):962-966). However, these theoretical relationships assume a fixed shape for the API particle size distribution (PSD, i.e., lognormal) and do not account for changes in the distribution shape. A more rigorous method for predicting the effect of a given PSD on UDU is to evaluate the contribution of individual particle size bins on UDU variability. The latter approach is taken in this work, and the derivation reveals that the individual contribution of particles size bins can be expressed completely in terms of a single-particle-size statistic, D[6,3]. D[6,3] is therefore a valid predictor of UDU, regardless of the shape of the PSD (e.g., multimodal) and can form the basis of a particle size control strategy for low-dose drug products.