In VitroTests for Aerosol Deposition II: IVIVCs for Different Dry Powder Inhalers in Normal Adults

Abstract

A new in vitro test method for dry powder inhalers (DPIs) was recently found to be predictive of the published in vivo results for Budelin Novolizer. The present study was intended to assess the method's robustness by evaluating correlations between average drug deposition in vitro and in vivo from five different DPIs. In vitro drug deposition from five marketed DPIs was assessed in a realistic physical airway model of a "medium" sized adult in an experimental setup that allowed deposition to be characterized regionally for carefully selected simulated air flow rate versus time profiles. The DPIs studied were Spiriva(®) HandiHaler(®), Relenza(®) Diskhaler(®), Salbutamol Easyhaler(®), Pulmicort(®) Turbuhaler(®), and Foradil(®) Aerolizer(®). In vitro regional deposition results were compared with those reported in the literature in order to create in vitro-in vivo correlations (IVIVCs) for each inhaler. Mean percent total lung deposition (TLD ± SD) in vitro for Spiriva HandiHaler, Relenza Diskhaler, Salbutamol Easyhaler, Pulmicort Turbuhaler, and Foradil Aerolizer were 17.3 ± 1.2, 22.6 ± 1.1, 29.0 ± 1.1, 28.0 ± 3.0, and 21.7 ± 1.2, respectively. These results showed excellent agreement with reported in vivo values, with absolute prediction errors in TLD of ≤ 2% for all DPIs except Relenza Diskhaler. Similarly, in vitro mouth-throat and device deposition results were stoichiometrically comparable to those reported in vivo for all DPIs except Relenza Diskhaler and Turbuhaler. Inspection of the scintigraphy studies for Relenza Diskhaler and Turbohaler revealed possible problems with powder labeling and result interpretation in their in vivo clinical assessments. A characteristic physical airway model representing a medium-sized adult, when coupled to carefully chosen characteristic inhalation maneuvers used in the clinic, produced results that correlated with regional drug deposition estimates from scintigraphy across a group of different DPIs.