To understand the fate of aerosols delivered by high-flow nasal cannula using continuous nebulization, an open-source anatomical model was developed and validated with a modified real-time gamma ratemeter technique. Mass balance defined circuit losses. Responsiveness to infusion rate and device technology were tested. A nasal airway cast derived from a computed tomography scan was converted to a 3-dimensional-printed head and face structure connected to a piston ventilator (breathing frequency 30 breaths/min, tidal volume 750 mL, duty cycle 0.50). For mass balance experiments, saline mixed with Technetium-99m was infused for 1 h. Aerosol delivery was measured using a gamma ratemeter oriented to an inhaled mass filter at the hypopharynx of the model. Background and dead-space effects were minimized. All components were imaged by scintigraphy. Continuous nebulization was tested at infusion rates of 10-40 mL/h with gas flow of 60 L/min using a breath-enhanced jet nebulizer (BEJN), and a vibrating mesh nebulizer. Drug delivery rates were defined by the slope of ratemeter counts/min (CPM/min) versus time (min). The major source of aerosol loss was at the nasal interface (∼25%). Significant differences in deposition on circuit components were seen between nebulizers. The nebulizer residual was higher for BEJN (P = .006), and circuit losses, including the humidifier, were higher for vibrating mesh nebulizer (P = .006). There were no differences in delivery to the filter and head model. For 60 L/min gas flow, as infusion pump flow was increased, the rate of aerosol delivery (CPM/min) increased, for BEJN from 338 to 8,111; for vibrating mesh nebulizer, maximum delivery was 2,828. The model defined sites of aerosol losses during continuous nebulization and provided a realistic in vitro system for testing aerosol delivery during continuous nebulization. Real-time analysis can quantify effects of multiple changes in variables (nebulizer technology, infusion rate, gas flow, and ventilation) during a given experiment.
Read full abstract