DEVELOPMENT AND USE OF AN IN-VITRO SYSTEM TO EVALUATE INHALER DEVICES

An in vitro system was developed to better emulate particle deposition in the respiratory tract. Inhalers were connected to a glass throat (BP 1988 Appendix XVII C) or to a silicone throat that exactly duplicated the surface geometry of the oral and pharyngeal cavity. This throat was created from a direct impression of the mouth and CAT scans of a patient's head and neck and could be separated into three parts. This allowed deposition patterns to be observed. Adapters were fabricated so that the outlet of either throat could be connected to a collection filter unit or any one of several sizing instruments. A mass flow meter enabled airflow to be monitored. Airflow was produced from a vacuum pump or by human inhalation. A Rotahaler was tested using capsules containing 20 mg of spray dried mannitol:sorbitol:carboxyfluorescein (CF) 10:1:0.01 of 3.4 mu m MMAD. The vacuum pump was set at 30, 60 or 120 1/min air flow for 4 s. The surface of the glass or silicone throat was left 'dry' or was coated with a polyethylene glycol mixture to better represent the 'wet' surface of the oral cavity. Powder in the device, filter unit and throat(s) was quantified by assay of the CF using spectrofluorimetry. The recovery of the weighed dose was 98.6 +/- 8.9% (n = 83). The dose emerging from the inhaler was dependent on the flow rate (Anova, p < 0.01) and was 1.0 +/- 0.2 mg (30 1/min, n = 25), 5.1 +/- 0.5 mg (60 1/min, n = 30) and 6.2 +/- 0.6 mg (120 1/min, n = 28). However, the percentage of this dispensed dose recovered from the filter unit (lung) was independent of the flow rate and only varied with the type and condition of the throat used (Anova, p < 0.05). The mass deposition in the throats was ranked: glass-dry < silicone-dry < glass-wet < silicone-wet. The results indicate that the use of a wet artificial throat, modeled on human anatomy, will provide a more conservative estimate of lung deposition compared to a glass throat when used with a dry powder inhaler.