Performance of low air volume dry powder inhalers (LV-DPI) when aerosolizing excipient enhanced growth (EEG) surfactant powder formulations

Abstract

Efficient delivery of dry powder aerosols dispersed with low volumes of air is challenging. This study aims to develop an efficient dry powder inhaler (DPI) capable of delivering spray-dried Survanta-EEG powders (3–10 mg) with a low volume (3 mL) of dispersion air. A series of iterative design modifications were made to a base low air volume actuated DPI. The modifications included the replacement of the original capsule chamber with an integral dose containment chamber, alteration of the entrainment air flow path through the device (from single-sided (SS) to straight through (ST)), change in the number of air inlet holes (from one to three), varying the outlet delivery tube length (45, 55 and 90 mm) and internal diameter (0.60, 0.89 and 1.17 mm). The modified devices were evaluated by determining the influence of the modifications and powder fill mass on aerosol performance of spray-dried Survanta-EEG powders. The optimal DPI was also evaluated for its ability to aerosolize a micronized powder. The optimized dose containment unit DPI had a 0.21 mL powder chamber, ST airflow path, three-0.60 mm air inlet holes, and 90 mm outlet delivery tube with 0.89 mm internal diameter. The powder dispersion characteristics of the optimal device was independent of fill mass with good powder emptying in one 3 mL actuation. At 10 mg fill mass, this device had an emitted mass of 5.3 mg with an aerosol Dv50 of 2.7 µm. After three 3 mL actuations, >85% of the spray-dried powder was emitted from the device. The emitted mass of the optimal device with micronized albuterol sulfate was >72% of the nominal fill mass of 10 mg in one 3 mL actuation. Design optimization produced a DPI capable of efficient performance with a dispersion air volume of 3 mL to aerosolize Survanta-EEG powders.

Introduction

Dry powder inhalers use the patients’ inspiratory effort to disperse and deliver the powder aerosol to the lungs (1). These devices operate using large inspiratory volumes (>250 mL) to aerosolize the powder formulation. There are a number of delivery scenarios that would require much lower dispersion air volumes to be employed, for example, delivery of powder aerosols to small animal models and the delivery of surfactant powder aerosols to neonates. Efficient delivery of powder aerosols to small animals faces challenges since the powder needs to be dispersed with a very low volume of dispersion air as the animals have very small tidal volumes (e.g., 2–3 mL for rats) (2–4). Similar challenges exist for premature neonates who have tidal volumes in the range of 5–8 mL/kg (5). Active dry powder inhalers (DPIs) are preferred to deliver aerosols to animal models as they are free from dependence on inspiratory effort and can be translatable in delivery to pediatric patients requiring low inhalation flow rates (6,7). There are a number of approaches that are currently employed to address these delivery scenarios.

For small animal powder aerosol delivery, the PennCentury™ DP-4M dry powder insufflator (Penn-Century, Inc, Philadelphia, PA) is the most commonly used device designed to deliver dry powder aerosols (1–2 mg) to small animal models using a low volume (2 mL) of dispersion air (8). However, this device is no longer available for purchase. There are a range of lung delivery efficiencies in animals that have been reported for this device dependent upon the formulation and other factors (9,10). In addition to the PennCentury™ insufflator, a considerable number of custom devices have been reported to deliver dry powder aerosols with a low volume of dispersion air. Qiu et al. reported a custom-made device able to disperse 1 mg dry powder formulation to mice with 0.6 mL of dispersion air (11). However, the delivery efficiency of this device was not reported and the performance of the device with a larger mass of powder (>1 mg) was not confirmed. Chaurasiya et al. also designed and validated a simple custom-made device to deliver dry powder aerosol to mice (12). This device used a cannula tube for intubation and drug loading which was connected to a syringe for powder dispersion. The delivery efficiency of the device was 92% when dispersing 1 mg of spray-dried powder with 1 mL of dispersion air. Durham et al. developed a dosator that aerosolized 10 mg of spray-dried powder using a 5 mL air volume and demonstrated its use in guinea pigs (13). Tonnis et al. reported an aerosol generator device which was capable of delivering spray-dried powder in mice with a gentle push of 3 mL air in one minute (10). The estimated emitted dose from the device was 0.5 mg with fine particle fraction (<5 µm, % of emitted dose) of 49%. However, the aerosol generator was relatively large and required a large mass (~180 mg) of powder to be loaded. Sinha and Mukherjee used a device which consisted of a dry powder inhaler and a nose-only inhalation chamber to deliver dry powder aerosols to mice (14). Although this device was able to efficiently fluidize the powder bed (respirable fraction <5 µm: ~50%) and uniformly distribute the drug in different inhalation ports of the chamber (mass median aerodynamic diameter, MMAD: ~4.6 µm), the device is complicated and requires a large mass (>50 mg) of powder to be loaded. Okamoto et al. developed a device capable of delivering ~5 mg of spray-dried powder in rats with 2 mL of dispersion air (15). This device was used by Jalalipour et al., to deliver dry powders to rats (16).

The use of aerosolized surfactant formulations could be an alternative method of surfactant replacement therapy (SRT) administration to eliminate the problems related to the currently used liquid instillation approach (17,18). Surfactant replacement therapy, which involves invasive intubation and intratracheal bolus instillation of a liquid surfactant formulation, is the first-line treatment for preterm infants with respiratory distress syndrome (19). SRT has been used as an effective therapy since the early 1990s (20). However, multiple concerns including rapid fluctuations in hemodynamics and cerebral perfusion (21), non-uniform distribution of the surfactant within the lung airways, requirement of a high dosage regimen (17), increased risk of bradycardia, hypoxia and hypotension (22) after liquid instillation have motivated the search for an aerosol delivery alternative.

Among the available aerosol generation and delivery techniques, dry powder surfactant formulations and their delivery during non-invasive ventilation could be advantageous over the current method of SRT liquid instillation since it avoids the need for invasive intubation and does not require the patient to be taken off ventilation to be treated (23). In addition, dry powders are more stable than the liquid formulations and have the potential to achieve higher lung delivery efficiency compared with nebulized medications (24–25). However, a combination of both the drug formulation and delivery device determine the effective delivery of inhaled medications including dry powder aerosol to the lungs (26).

Several formulation approaches have been previously discussed to improve dry powder aerosol delivery to the lungs (27,28). Excipient enhanced growth (EEG) is a well-studied formulation approach which has been adapted for use with both dry powder and liquid formulations for efficient aerosol generation and delivery to the lungs (29,30). The EEG approach can be useful in developing a dry powder aerosol formulation of surfactant which can help to treat preterm infants with respiratory distress syndrome. In this approach, powder formulations around 1 µm in size are aerosolized to minimize upper airway deposition losses and hygroscopic growth then takes place in the airways that promotes deposition and minimizes exhalation losses of the submicrometer aerosol.

Several devices have been reported to deliver dry powder surfactant formulations (31–33). Pohlmann et al. developed a reservoir-based device that can store and deliver spray-dried surfactant powder using pressure pulses consisting of 10–30 mL of air (31). Ruppert et al. reported a dry powder aerosolizer device to deliver micronized surfactant dry powder formulation to rabbits and mice models (32). Although the device generated aerosols with 85% of particles less than 5 µm, the delivery efficiency is dependent on the pressure used to aerosolize the powder. Walther et al. developed a capsule-based low flow inhaler device to deliver spray-dried surfactant powders to rabbits and preterm lambs (33). This device can be used as a stand-alone device or be incorporated into a ventilatory system like nasal continuous positive airway pressure (nCPAP).

The purpose of this study was to develop a highly efficient DPI device capable of delivering (3–10 mg) spray-dried Survanta-EEG powders with a very low volume (3 mL) of dispersion air using up to 3 actuations. For this purpose, we modified the DPI originally developed by Farkas et al. (34) which used multiple actuations of 10 mL volumes of dispersion air. High delivery efficiency was established based on an emitted mass sufficient to theoretically elicit a positive effect in a rat surfactant depletion model.

We have developed an excipient enhanced growth (EEG)-based surfactant powder using a naturally derived surfactant in a commercial liquid surfactant replacement product Survanta® that contains both SP-B and SP-C. For Survanta-EEG powder, 3–10 mg is the expected delivered dose to assess therapeutic efficacy and safety in small animal models. To minimize the amount of time an animal spends without assisted ventilation, delivery of a maximum dose in the least number of actuations with a low volume of dispersion air is preferred. The dispersion efficiency of the optimal device was assessed by aerosol characterization of the Survanta-EEG powders. The performance of the optimal low volume DPI was also evaluated with micronized powder to broaden its applicability to disperse powders prepared by methods other than spray drying.

Materials and methods

Chemicals

Survanta® intratracheal suspension was purchased from Cardinal Health, Inc. (Greensboro, NC). Sodium chloride and ethanol were purchased from Fisher Scientific Co. (Hanover Park, IL). L-leucine was purchased from Sigma-Aldrich Chemical Co. (St Louis, MO). Pearlitol® PF-mannitol was donated by Roquette Pharma (Lestrem, France). Micronized albuterol sulfate (AS) (USP grade) was purchased from Letco Medical, LLC (Decatur, AL). Freshly collected deionized water was used throughout the study.

Preparation of Survanta-EEG powders

Survanta-EEG powders were prepared by spray drying of the feed dispersions containing Survanta®, hygroscopic excipients (mannitol and sodium chloride) and a dispersion enhancer (l-leucine) at a ratio of 40:30:10:20% w/w using the Buchi Nano Spray Dryer B-90 HP (Büchi Labortechnik AG, Flawil, Switzerland). The feed dispersions were prepared with 0.125% w/v solids by addition of all the formulation components to 5% v/v ethanol in water followed by sonication in a heated water bath (Fisher Scientific™ CPXH, Hanover Park, IL). The prepared feed dispersions were spray-dried with the spray dryer in an open mode configuration using the medium nozzle and the following operating conditions: inlet temperature of 70 °C, pump speed of 3%, spray intensity of 80%, and inlet gas flow of 120 L/min. These operating conditions resulted in an outlet temperature of 37–41°C and drying chamber pressure of 40–42 mbar. The spray-dried powders were collected from the electrostatic precipitator into glass vials and stored in a desiccator (0% RH) in the fridge (2–8 °C) when not in use.

Particle size distribution

A Sympatec HELOS laser diffraction apparatus equipped with RODOS/M disperser and ASPIROS sample feeder (Sympatec GmbH, Clausthal-Zellerfeld, Germany) was used to determine the primary particle size distributions of the spray-dried Survanta-EEG powders. Powder samples placed into the ASPIROS sample feeder were dispersed in the laser beam at 1.0 and 4.5 bar dispersion pressure using the RODOS/M disperser. Low dispersion pressure (1.0 bar) was used to get the idea about powder dispersion behavior and higher dispersion pressure (4.5 bar) was used to get the primary particle size of the powder particles. Particle size distributions (Dv10, Dv50 and Dv90) of the powders were calculated using Fraunhofer theory in the WINDOX 5.0 software (Sympatec GmbH, Clausthal-Zellerfeld, Germany). The Sympatec software was used to calculate particle fractions less than 1 and 5 µm, respectively. A total of ten batches of spray-dried powders were used in this study.

Low air volume dry powder inhaler (LV-DPI)

The initial device design was based on a low air volume dry powder inhaler (LV-DPI) developed by Farkas et al. (34). The base device consists of two halves that when assembled, pierce a size 0 capsule housed in the device chamber. The capsule is pierced by two hollow capillaries: an inlet capillary with an internal diameter (ID) of 0.60 mm and an outlet capillary with an ID of 0.89 mm. Dispersion air is introduced into the device using a disposable syringe via a standard luer lock connection to the inlet. The LV-DPI device was able to efficiently aerosolize and deliver a powder aerosol with 10 mL of dispersion air. In one version of this previously developed LV-DPI, both the inlet and outlet capillaries were configured on the same side of the capsule (single-sided (SS) design), which provided good device emptying and aerosol characteristics with an emitted dose of 85%, MMAD of 2.1 µm and fine particle fraction less than 5 µm (FPF_<5µm) of 89% (34).

Modification of LV-DPI

The SS LV-DPI was initially modified with an extended length outlet delivery tube of 90 mm (Fig. 1a) based on the delivery tube of the PennCentury™ Dry Powder Insufflator (Penn-Century, Inc., Philadelphia, PA) to deliver Survanta-EEG powders using a low volume of dispersion air (3 mL). The intention of using the extended delivery tube was to develop a device for use with oral-tracheal intubation of small animal models. A series of in vitro iterative design modifications to the LV-DPI were investigated in order to maximize the aerosol delivery efficiency (emitted dose) while maintaining good aerosol performance for the Survanta-EEG powders using a 3 mL air volume suitable for small animals and eventually neonatal administration.

Fig. 1.

Axial cross-section of assembled containment unit DPIs: (a) SS with 0.68 mL powder chamber, (b) SS with 0.21 mL powder chamber, (c) ST with 90 mm delivery tube, (d) open powder chamber view of studied air inlets and (e) delivery tube internal diameters; SS: single-sided, ST: straight through.

Powder chamber volume

The original capsule-based device was redesigned to eliminate the need for a capsule (Fig. 1a). This allowed the removal of the protruding inlet and outlet capillaries for capsule piercing. The formulation was loaded directly into an integral dose containment unit with inlet and outlet channels as shown in Fig. 1b. The dose containment chamber was reduced to 30% of its original 0.68 mL volume resulting in a volume of 0.21 mL to account for the lower dispersion air volume of 3 mL, which is approximately 30% of the original volume of 10 mL used for LV-DPI (34).

Airflow pathway

The airflow pathway into and out of the dose containment unit was varied to investigate its effect on aerosol performance. A straight through (ST) airflow path (Fig. 1c) was compared to the SS airflow path (Fig. 1b). In the ST airflow path, air enters the chamber from one side and aerosol exits out of the opposite end of the dose containment unit (Fig. 1c).

Air inlet holes

The air inlet holes in the ST airflow pathway were further modified by replacing the original single air inlet hole (0.6 mm diameter) with three 0.60 mm air inlet holes (Fig. 1d).

Delivery tube internal diameter

The internal diameter of the delivery tube was varied using 0.60 mm and 1.17 mm diameters for comparison with the original 0.89 mm tube ID (Fig. 1e).

Delivery tube length

The length of the delivery tube, as measured from the outer o-ring of the device (Fig. 1a), was varied using 45 mm and 55 mm lengths for comparison with the 90 mm tube length.

Design and construction of the LV-DPI devices

Autodesk Inventor was used to create the designs of the inhalers. The designs were then exported as .STL files to be prototyped. All the DPIs were built by rapid prototyping using a Stratasys Object24 3D Printer (Stratasys Ltd., Eden Prairie, MN) with a rigid opaque photopolymer, VeroWhitePlus (RGD835). The inlet delivery tubes were fabricated as channels in the DPI body with a VeroWhitePlus (RGD835) contact surface. The outlet delivery tubes were custom cut from stainless steel (SAE 304) tubing and secured in place with water resistant epoxy. After printing, the devices were fully dried before use.

Aerosol performance testing using the spray-dried Survanta-EEG formulation

The aerosol performance of the modified LV-DPIs were initially investigated using a fill mass of 3 mg of spray-dried Survanta-EEG formulation.

Determination of emitted mass

Gravimetric analysis was used to determine the emitted mass of the spray-dried Survanta-EEG powders from the modified LV-DPIs to determine the device emptying capacity. Pre-weighed powder samples were manually loaded into the dose containment unit of the modified LV-DPIs. The assembled device including the outlet tube and loaded with powder was weighed to determine the pre-actuation weight. A 5 mL syringe filled with 3 mL of room air was connected to the luer-lock inlet of the device. The device was actuated by manual compression of the syringe. If multiple actuations were employed, the syringe was disconnected, re-filled with air and the actuation process repeated. Finally, the syringe was disconnected and the device was reweighed (post-actuation weight). The emitted mass of the powder was determined from the difference between the pre-actuation weight and post-actuation weight of the device. The percentage of powder emitted mass was calculated as a percentage of the initial loaded mass.

Aerosol characterization

The particle size distribution of the spray-dried Survanta-EEG powder aerosols exiting the modified DPIs were characterized using a Malvern Spraytec® equipped with the RT Sizer Software (Malvern Instruments, Ltd., Worcestershire, UK). Pre-weighed powder samples loaded in the modified device were manually dispersed into an open bench Spraytec laser beam using a 5 mL syringe filled with 3 mL of room air. The exit of the delivery tube was positioned 3 cm from the laser beam and was maintained to ensure consistency across studies. Measurements were repeated for up to three actuations. Table I shows the iterative parameters explored in this study.

Table I.

Device design parameters explored for the modified containment DPI with 3 mL dispersion air.

Experiment	Powder chamber volume (mL)	Device configuration	Number of air inlet holes	Delivery tube internal diameter (mm)	Delivery tube length (mm)
N1	0.68	SS	1	0.89	90
N2	0.21	SS	1	0.89	90
N3	0.21	ST	1	0.89	90
N4	0.21	ST	3	0.89	90
N5	0.21	ST	3	0.60	90
N6	0.21	ST	3	1.17	90
N7	0.21	ST	3	0.89	55
N8	0.21	ST	3	0.89	45

Effect of formulation fill mass and outlet delivery tube length on aerosol performance

To investigate the effect of formulation fill mass on aerosol performance of the spray-dried Survanta-EEG formulation, 3, 5 and 10 mg masses were loaded in the optimized modified device with three different lengths of outlet delivery tube (90, 55 and 45 mm) using a 3 mL dispersion air volume. The emitted mass and aerosol particle size distributions were evaluated using the methods described above. An additional study using the 55 mm delivery tube length and variable fill masses was performed using an increased dispersion air volume of 5 mL. The slightly larger dispersion air volume was investigated as a possible means of delivering a high emitted dose with a single actuation air volume.

Aerosol performance of micronized powder

The applicability of the modified LV-DPI to disperse micronized drug powders was evaluated using micronized albuterol sulfate. Briefly, 10 mg of micronized AS was loaded in the dose containment unit LV-DPI and dispersed using 3 mL of dispersion air. The emitted mass and aerosol characterization was conducted following the procedures mentioned above.

Statistical analysis

JMP® Pro software (version 12.0; SAS Institute Inc., Cary, NC) was used to analyze the data. Statistically significant differences in the data were determined by performing student’s t-test or one-way analysis of variance (ANOVA) followed by Dunnett’s method or Tukey’s HSD (p<0.05). The strength of linear relationships was quantified by determination of the Pearson correlation coefficient.

Results

Survanta-EEG powders

The Survanta-EEG powders were prepared by spray drying. The mean geometric diameter (Dv50) value of the spray-dried powders were 1.3 ± 0.07 µm and 1.0 ± 0.04 µm, respectively at 1.0 and 4.5 bar dispersion pressures and the powders were observed to be monodisperse (Fig. 2). The coefficients of variation in the Dv50 values across the studied batches were <6%, indicating good reproducibility of the powder production process. The span at 4.5 bar dispersion pressure was 1.4 ± 0.1 with particle fractions less than 1 µm of 50.1 ± 2.6% and less than 5 µm of 99.9 ± 0.3%. For the EEG approach, a primary particle size around 1 µm is desirable, with particle fractions less than 1 µm and 5 µm above 30% and 90%, respectively. There was only a small difference (0.3 µm) in the particle size measured at 1 bar, similar to pressures in a DPI and at 4.5 bar, which was used to ensure complete dispersion to primary particles.

Fig. 2.

Particle size distributions (cumulative and frequency distribution) of the spray-dried Survanta-EEG powder batches using the Sympatec at dispersion pressure of 4.5 bar. Markers represent the mean value, error bars represent the standard deviation, n=10.

Effect of iterative design modification on aerosol performance

Reducing the powder chamber volume and removal of both inlet and outlet capillary protrusions with the elimination of the capsule chamber had a significant effect on powder delivery efficiency (p<0.05). The emitted mass for 0.68 mL chamber (N1) was 20.2 ± 3.6% after the first actuation with 3 mL of air, increasing to 40.4 ± 7.5% for the device with 0.21 mL chamber volume (N2) (p=0.0048) (Table II). For the devices with original and reduced chamber volume, the cumulative emitted mass after third actuation was 43.6% and 77.1% of the nominal fill mass, respectively. Improved emitted mass with the reduced powder chamber volume could be due to the removal of the dead space in the original design which enabled the powder chamber to empty more efficiently when the lower 3 mL volume of dispersion air was employed. In addition, removal of the outlet capillary protrusion in the powder chamber may have also contributed to the higher observed emitted mass by eliminating powder loss around the base of the capillary which was observed during the experiment. The mean Dv50 value of the generated aerosols from the SS LV-DPI with reduced chamber volume (N2) was 2.5 μm with particle fractions less than 1 μm of 28.8 ± 3.8% and less than 5 μm of 60.1 ± 4.4%, respectively (Table II).

Table II.

Effect of powder chamber volume, air flow configuration, number of air inlet holes and delivery tube internal diameter on aerosol performance of spray-dried Survanta-EEG powders after dispersion of 3 mg powder with 3 mL dispersion air (values are mean (standard deviation), n≥3).

Experiment	Powder chamber volume (mL)	Device configuration	Air inlet holes	Delivery tube internal diameter (mm)	1st actuation					Cumulative mass emitted after 3 actuations (% nominal)
					Mass emitted (mg)	Mass emitted (% nominal)	Dv50 (µm)	Particle fraction (%)
								<1 µm	<5 µm
N1	0.68	SS	1	0.89	0.62 (0.12)	20.2 (3.6)	–	–	–	43.6 (4.4)
N2	0.21	SS	1	0.89	1.23 (0.22)*	40.4 (7.5)*	2.5 (0.6)	28.8 (3.8)	60.1 (4.4)	77.1 (13.7)
N3	0.21	ST	1	0.89	1.67 (0.16)‡	55.5 (5.7)‡	2.7 (0.6)	23.3 (3.4)	62.5 (10.5)	83.1 (7.9)
N4	0.21	ST	3	0.89	2.12 (0.12)#	70.9 (3.7)#	2.7 (0.2)	26.9 (2.1)	57.5 (2.6)	92.4 (1.9)
N5	0.21	ST	3	0.60	1.05 (0.03)^	33.9 (0.3)^	1.7 (0.1)^	32.6 (0.2)^	73.5 (5.7)^	70.8 (2.6)
N6	0.21	ST	3	1.17	2.26 (0.13)	74.2 (4.7)	6.8 (0.1)^	12.0 (1.1)^	44.7 (0.7)^	98.2 (3.6)

^*Significant difference compared to 0.68 mL powder chamber volume; Student’s t-test, p<0.05

^‡Significant difference compared to SS device configuration; Student’s t-test, p<0.05

^#Significant difference compared to one-0.60 mm air inlet hole; Student’s t-test, p<0.05

^†Statistically significant effect of delivery tube internal diameter; one-way ANOVA, p<0.05

^{^}Significant difference compared to 0.89 mm delivery tube internal diameter; Dunnett’s method, p<0.05

The next iterative design configuration investigated was altering the dispersion airflow pathway from SS to ST design while maintaining the 0.21 mL dose chamber (N3–N6). In the ST design, the dispersion airflow path travels straight through the dose containment unit and the air flow is directed over the top of the powder instead of directing air into the powder bed. After the first actuation, the emitted mass in the ST design (N3) (55.5%) was significantly higher than the SS design (N2) (40.4%) (p=0.0109) (Table II). However, between the two designs, no significant differences were observed in the Dv50 values (2.7 ± 0.6 vs 2.5 ± 0.6 μm for ST and SS design, respectively) and particle fractions less than 1 µm (23.3 ± 3.4% vs 28.8 ± 3.8 % for ST and SS, respectively) and less than 5 µm (62.5 ± 10.5% vs 60.1 ± 4.4% for ST and SS, respectively) (p>0.05) (Table II). These results suggest that the airflow path had a significant influence on powder delivery efficiency but did not affect the powder dispersion. The differences in the emitted mass between the designs could be due to the differences in the turbulent forces within the dose containment unit. Since the ST design directs the inlet airflow path to travel in a single direction over the powder bed within the containment unit, it reduces the turbulent forces which results in more emitted powder (35,36). For the SS design, the inlet airflow path is directed at the powder bed which may create impaction forces that direct the powder to adhere to the walls of the containment unit resulting in a lower emitted mass (35,36). The cumulative mass emitted after three actuations of 3 mL air volume with the ST design (N3) was 83.1% of the nominal mass.

Moving forward with the ST design, increasing the number of inlet holes from one to three (N4–N6) significantly increased the emitted mass (p<0.05). After the first actuation, the emitted mass increased to 70.9% from 55.5% after incorporation of three air inlet holes in the ST design (N4) (Table II). No significant differences were observed in the particle fractions less than 1 and 5 µm between the devices with different number of inlet holes and the Dv50 values remained unaffected. The cumulative mass emitted after the third actuation with 3 mL of air was 92.4% of the nominal fill mass. Iterative changes to the LV-DPI showed a significant effect on emitted mass without affecting the aerosol dispersion properties (Dv50 and particle fractions less than 1 and 5 µm). The next series of iterative changes sought to improve the aerosol dispersion from the LV-DPI.

Effect of delivery tube internal diameter

The delivery tube internal diameter of the best performing ST LV-DPI with three air inlet holes and 0.21 mL powder chamber volume was varied using 0.60 mm and 1.17 mm diameters compared to the original 0.89 mm (N4–N6). Table II shows the summary of the aerosol performance of the different devices. Among the devices, the 0.60 mm ID device (N5) had the lowest emitted mass, indicating relatively poor device emptying compared to the other options (p<0.05). There was no significant difference between the emitted mass of 0.89 mm (N4) and 1.17 mm (N6) ID devices (p>0.05). The emitted masses (% nominal) for 0.89 mm and 1.17 mm ID devices were >70% in contrast to only 33.9% for the 0.60 mm ID device.

The aerosol performance and Dv50 values were significantly different among the devices (p<0.05). Increasing the tube ID also increased the Dv50, which corresponded with a decrease in particle fractions less than 1 and 5 µm (Table II). The Dv50 value for 0.89 mm ID device (N4) (2.7 µm) was significantly larger than 0.60 mm ID device (N5) (1.7 µm) and smaller than 1.17 mm ID device (N6) (6.8 µm). The differences in the aerosol performance among the devices could be due to the differences in the hydrodynamic forces used to disperse the powders (37). Since all the devices had the same total inlet cross-sectional area (1.8 mm), the smallest 0.60 mm ID device had the highest pressure at the outlet, producing improved dispersion (lowest Dv50) but poor emptying (lowest emitted mass). On the other hand, the device with 1.17 mm ID delivery tube had the lowest outlet pressure with the poorest dispersion and good emptying. The particle fractions less than 1 and 5 µm were 26.9% and 57.5%, respectively for the 0.89 mm ID device and were higher than the 1.17 mm ID device (12.0% and 44.7%) and lower than the 0.60 mm ID device (32.6% and 73.5%) (p<0.05). In terms of device emptying (emitted mass) and aerosol dispersion (Dv50 values), the ST device with three air inlet holes and 0.89 mm ID outlet delivery tube was observed to be the best overall performing device and was used in further studies as described below.

Effect of outlet delivery tube length and fill mass

The effect of three different outlet delivery tube lengths (90, 55 and 45 mm) on aerosol performance was studied using three powder fill masses (3, 5 and 10 mg) using the best performing LV-DPI described in the previous section. Table III summarizes the results. For each of the three delivery tube lengths, fill mass had a significant effect on emitted mass. With increasing fill mass of the Survanta-EEG powder, the emitted mass from the device increased on the first actuation, when expressed in terms of mass. However, when expressed as a percentage of the nominal dose, for the 90 mm delivery tube device, the percent emitted after the first actuation was higher for the 3 mg fill mass (70.9%) compared to the 5 and 10 mg fill masses (56.2% and 52.3%, respectively) (p<0.05). After 3 actuations, the cumulative mass emitted for the 3 mg (92.4%) and 5 mg (94.8%) fill masses were significantly higher than the 10 mg fill mass (85.4%) (p<0.05). There were no significant differences in the percent emitted mass of between 5 and 10 mg fill masses on the first actuation and 3 and 5 mg fill masses after 3 actuations (p>0.05). The Dv50 value (2.7 µm) during the first actuation for all three fill masses for the 90 mm delivery tube remained unchanged. However, fill mass had a significant effect on the particle fractions less than 1 µm and less than 5 µm. While, the particle fraction less than 1 µm with the 5 mg fill mass (29.3%) was significantly higher than for the 10 mg fill mass (24.6%), the particle fraction less than 5 µm was significantly lower for the 5 mg fill mass (55.0%) than the 10 mg fill mass (61.8%) (p<0.05). The poor correlation coefficients with p>0.05 observed with the Dv50 and particle fractions less than 1 µm for 90 mm delivery tube (Table IV) also confirmed that the dispersion efficiency was independent of powder fill mass. The particle fraction less than 5 µm had a positive correlation coefficient (0.68) with p-value of 0.042.

Table III.

Effect of delivery tube length and fill mass on aerosol performance of spray-dried Survanta-EEG powders. Values are mean (standard deviation), n=3.

Delivery tube length (mm)	Fill mass (mg)‡	1st actuation					Cumulative mass emitted after 3 actuations (% nominal)
		Mass emitted (mg)	Mass emitted (% nominal)	Dv50 (µm)	Particle fraction (%)
					<1 µm	<5 µm
90	3	2.12 (0.12)	70.9 (3.7)*#	2.7 (0.4)	26.9 (2.1)	57.5 (2.6)	92.4 (1.9)#
90	5	2.83 (0.19)	56.2 (4.4)	2.7 (0.2)	29.3 (1.2)#	55.0 (0.7)#	94.8 (3.1)#
90	10	5.28 (0.43)	52.3 (4.4)	2.7 (0.2)	24.6 (1.2)	61.8 (2.2)	85.4 (3.0)
55	3	2.15 (0.09)	73.6 (3.5)*#	3.0 (0.3)#	23.7 (1.6)#	57.1 (2.2)	94.5 (6.0)
55	5	2.73 (0.23)	54.9 (4.2)	3.2 (0.3)	24.5 (1.8)#	54.7 (1.1)	87.8 (2.7)
55	10	6.42 (0.37)	63.8 (3.6)	4.5 (0.8)	19.6 (1.1)	51.7 (2.6)	90.3 (2.2)
45	3	2.03 (0.15)	66.8 (4.4)#	2.7 (0.6)*#	21.4 (4.3)#	65.3 (13.5)*#	88.3 (3.5)*
45	5	2.88 (0.13)	56.7 (2.5)	7.2 (1.8)	15.2 (1.3)	44.5 (3.7)	76.5 (0.8)
45	10	5.51 (0.61)	54.7 (6.2)	9.1 (0.7)	13.4 (0.2)	39.8 (1.4)	79.8 (6.1)

^‡Statistically significant effect of fill mass; one-way ANOVA, p<0.05

^*Significant difference compared to 5 mg fill mass for the same delivery tube device; Tukey’s HSD, p<0.05

^#Significant difference compared to 10 mg fill mass for the same delivery tube device; Tukey’s HSD, p<0.05

Table IV.

Correlation statistics of powder fill mass with aerosol characteristics for the three delivery tube lengths during the first actuation.

Delivery tube length (mm)	Characteristic	Correlation coefficient (r)	Significance (p-value)	Coefficient of determination (R2)
90	Dv50 values	−0.0286	0.9481	0.0008
	Particle fraction <1 µm	−0.5701	0.1090	0.3250
	Particle fraction <5 µm	0.6843	0.0420*	0.4682
55	Dv50 values	0.8228	0.0065*	0.6770
	Particle fraction <1 µm	−0.7914	0.0111*	0.6263
	Particle fraction <5 µm	−0.7784	0.0135*	0.6059
45	Dv50 values	0.8371	0.0049*	0.7007
	Particle fraction <1 µm	−0.7257	0.0269*	0.5266
	Particle fraction <5 µm	−0.7115	0.0316*	0.5063

^*Significant; Pearson correlation, p<0.05

For the 55 mm delivery tube device, fill mass also had a significant effect on the emitted mass, percent emitted mass, Dv50 and particle fractions less than 1 µm. With the increase in fill masses, the emitted mass increased (Table III). The percent emitted mass after first actuation was significantly higher for the 3 mg fill mass (73.6%) compared to the 5 mg (54.9%) and 10 mg (63.8%) fill masses (p<0.05). The Dv50 value for the 10 mg fill mass (4.5 µm) was significantly larger than the Dv50 values for the 3 and 5 mg fill masses (3.0 and 3.2 µm, respectively) (p<0.05). The particle fraction less than 1 µm was significantly lower at the 10 mg fill mass (19.6%) than the values for the 3 mg (23.7%) and 5 mg (24.5%) with no difference between 3 and 5 mg values. However, among the fill masses, the particle fractions less than 5 µm were not significantly different and the cumulative emitted masses after three actuations were also similar. The positive correlation coefficient with p-value less than 0.05 for Dv50 (Table IV) also supports the observation of decreased powder dispersion (larger Dv50 value) during the first actuation at the higher fill mass with 55 mm delivery tube. The strong negative correlations for the particle fractions less than 1 and 5 µm further confirmed the observed inverse relationship of particle fractions with powder fill mass.

Similar to the 55 mm delivery tube, fill mass for the 45 mm delivery tube had a significant effect on the aerosol performance characteristics. After the first actuation, the percent emitted mass was higher for the 3 mg fill mass than the 10 mg; and after three actuations, the percent emitted mass was higher for the 3 mg than the 5 mg fill mass (p<0.05). The Dv50 value for the 3 mg fill mass was significantly smaller than for the 5 and 10 mg fill masses with the 3 mg fill mass also having the highest particle fractions less than 1 and 5 µm. The correlation coefficients with p-values <0.05 for the Dv50 and particle fractions less than 1 and 5 µm for 45 mm delivery tube (Table IV) also showed a strong correlation of the powder fill mass with Dv50 and particle size fractions less than 1 and 5 µm.

Among the devices with three different lengths of delivery tube, no differences were observed in the emitted mass of powder after the first actuation for both the 3 mg and 5 mg powder fill masses. At the 10 mg fill mass, the 55 mm delivery tube device had the highest emitted mass on the first actuation (63.8% of nominal mass) compared to that for the 90 and 45 mm delivery tube devices (52.3% and 54.7%, respectively). In addition, at the 10 mg fill mass, the cumulative mass emitted after three actuations (90.3% of nominal mass) was also highest for 55 mm delivery tube device. Except for the 3 mg fill mass, for the other two fill masses, the Dv50 increased with decreasing delivery tube length. At both 5 mg and 10 mg fill masses, the 90 mm delivery tube device had the lowest Dv50 (2.7 µm) compared to that for the 55 mm (3.2 and 4.5 µm, respectively) and 45 mm delivery tube devices (7.2 and 9.1 µm, respectively). The particle fractions greater than 1 and 5 µm increased with increasing delivery tube length (Table III).

Effect of dispersion air volume

Figure 3 shows the aerosol performance of spray-dried Survanta-EEG powders dispersed using two different dispersion air volumes at three different fill masses with the modified LV-DPI with 55 mm delivery tube length. Except for the 5 mg fill mass, at each fill mass, no significant differences were observed between the emitted masses for 3 mL and 5 mL dispersion air volumes. At 5 mg fill mass, the percent emitted mass (71.7% of nominal) was higher for 5 mL dispersion air volume than that for the 3 mL volume (54.9% of nominal) (p<0.05). At the lower fill masses of 3 mg and 5 mg, the aerosol Dv50 values were smaller for 3 mL dispersion air volume than the 5 mL dispersion air volume (p<0.05). No differences in the Dv50 values between the two dispersion air volumes were observed at the 10 mg fill mass. Overall, at all fill masses, 5 mL dispersion air volume had better emptying (relatively higher emitted masses) but poorer powder dispersion (higher Dv50 and lower particle fractions less than 1 µm) than the 3 mL dispersion air volume.

Fig. 3.

Emitted mass (primary Y-axis) and particle size (secondary Y-axis) of the spray-dried Survanta-EEG powder dispersed from the modified low-volume dry powder inhaler device using two different dispersion air volumes (3 and 5 mL) at three different fill masses (3, 5 and 10 mg). Markers represent the mean value, error bars represent the standard deviation, n=3.

Aerosol performance with micronized powder

The aerosol performance with the micronized AS powder was performed using the modified LV-DPI with 90 mm and 55 mm delivery tube lengths at 10 mg fill mass. Table V shows the summary of the aerosol performance of the two devices with the micronized AS powder. After first actuation, more than 70% of the nominal fill mass was emitted from both devices. The Dv50 values were 5.8 and 5.3 µm, respectively for 90 mm and 55 mm delivery tube lengths which resulted in the particle fractions less than 1 µm of 13.2% and 9.4%, respectively for these devices. In comparison to the spray-dried Survanta-EEG powders, the Dv50 values were higher and particle fractions less than 1 µm were lower, which could be due to the cohesive properties of the micronized drug despite having a similar primary particle size distribution (mean Dv50: 1.3 µm with span value of 1.9 at 4.5 bar dispersion pressure).

Table V.

Aerosol performance of micronized albuterol sulfate powder from modified low air volume dry powder inhaler at 10 mg fill mass. Values are mean (standard deviation), n≥3.

Delivery tube length (mm)	1st actuation					Cumulative mass emitted after 3 actuations (% nominal)
	Mass emitted (mg)	Mass emitted (% nominal)	Dv50 (µm)	Particle fraction (%)
				<1 µm	<5 µm
90	7.22 (0.36)	72.8 (3.1)	5.8 (1.8)	13.2 (2.0)	48.6 (7.4)	94.4 (0.5)
55	7.35 (0.35)	72.5 (3.0)	5.3 (1.8)	9.4 (3.0)	51.8 (9.3)	90.3 (2.2)

Discussion

Effective powder delivery as an aerosol depends on a combination of both the drug formulation and the delivery device. An inhaler that is able to disperse relatively high powder masses (3–10 mg) with very low volumes of dispersion air is necessary to deliver drugs as aerosols to small animal models and for surfactant aerosol delivery to infants with surfactant deficiency or suffering from respiratory distress syndrome. However, restriction to a low dispersion air volume and preference of delivering the dose in one actuation are challenges to the design optimization. In the current study, we have modified an active DPI which was originally developed for use with 10 mL volumes of dispersion air (34). The original capsule-based chamber was removed and replaced with an integral dose containment chamber with a volume of 0.21 mL. Subsequent iterative design modifications including alteration of the entrainment air flow path through the device, number of air inlet holes, outlet delivery tube lengths and internal diameter, which were then investigated to evaluate their effects on the aerosol performance. The change in the entrainment air flow path from SS to ST and an increase in the number of air inlet holes from one to three increased the percent emitted mass. The increase in the internal diameter of the outlet delivery tube resulted in an increase in both percent emitted mass and aerosol Dv50 values. Evaluation of the effect of outlet delivery tube lengths and fill mass on the aerosol performance revealed an optimal device that was found to have powder dispersion characteristics independent of fill mass with good powder emptying in one actuation with 3 mL dispersion air. At a 10 mg fill mass, this optimal device was able to deliver 5.3 mg of spray-dried Survanta-EEG powder with an aerosol Dv50 of 2.7 µm in a single actuation with 3 mL dispersion air. The observed emitted mass (5.3 mg) is more than double the calculated amount (~2.2 mg) of Survanta-EEG powder needed for a 300 g rodent to produce an effect (38). After three 3 mL actuations, more than 85% of the spray-dried powder was emitted from the device. The low volume of air used to disperse the powder in the optimal dose containment unit DPI is vital to deliver drugs as aerosol to small animal models and preterm infants who have low tidal volumes (5). This device has the potential to be used to deliver spray-dried surfactant powder to small animal models which can also be adapted for use in the treatment of infants receiving non-invasive ventilation. In addition, the aerosol performance of the optimal device with micronized AS powder further showed the potential of the developed device to be used to deliver powders prepared by methods other than the spray drying.

Suitability of the optimal LV-DPI device in combination with spray-dried Survanta-EEG powders was confirmed in a recently published aerosol study (39). The outcome of the in vivo study, as demonstrated by the improved lung compliance and elastance in surfactant depleted rats, confirmed that the optimized DPI enabled efficient delivery of spray-dried surfactant formulations to a small animal model.

Hoppentocht et al. reported variable aerosol performance for powders produced by different methods for the PennCentury™ DP-4M dry powder insufflator which was designed to deliver a small amount (1–2 mg) of dry powder formulation to small animal models using a lower volume of dispersion air (2 mL) (8). For 3 mg of spray-dried colistimethate sodium, the study authors reported an emitted dose of approximately 90% with median particle size ~4 µm with the PennCentury™ device when dispersed with 1 mL of air (1–4 actuations) from the pressurized air pump (8). In contrast, at 3 mg fill mass, our optimal device had a smaller median particle size of 2.7 µm after one actuation with 3 mL air and >92% emitted mass after three 3 mL air actuations. In addition, while the manufacturer recommended dose to be used for DP-4M insufflator is between 1 and 2 mg, our device can be used for emptying up to 10 mg of powder with 3 mL of dispersion air. Okamoto et al. reported a device which consists of a disposable syringe, a disposable needle and two 3-way stopcocks to deliver dry powders to rats (15). Although the device was able to disperse ~5 mg of spray-dried powder in rats with 2 mL of dispersion air, the performance of the device was not confirmed for larger masses (>5 mg) of powder and >2 mL of dispersion air. Chaurasiya et al. developed a simple custom-made device which was capable of delivering dry powder aerosol to mice with 92% delivery efficiency (12). However, the use of this device was limited to 1 mg spray-dried powder with 1 mL of dispersion air.

Walther et al. reported the use of a capsule-based low flow inhaler device to deliver a dry powder synthetic surfactant formulation which dispersed powders at low flow rates of 4–10 L/min, delivering 30 mg of dry powder surfactant from a capsule over 1–3 min (33). Pohlmann et al. used a novel Continuous Powder Aerosolizer (CPA) with a subsequent humidification step to deliver dry powder formulation of recombinant surfactant protein-C (rSP-C) to preterm neonates (31). The delivery efficiency of the CPA was 55% of the loaded mass with the median particle size of 3 to 3.5 µm when dispersed using 14 mL of compressed air to the end of a delivery tube. Ruppert et al. generated aerosols of recombinant surfactant protein-C (rSP-C)-based surfactant formulation from their developed Dry Powder Aerosolizer (32). Although the device was able to aerosolize large amount of surfactant formulation (800 mg/min), the delivery efficiency was dependent on the pressure used to aerosolize the powder. The MMAD of the generated aerosols from this device was 1.6 µm. Recently, Qiu et al. reported a custom-made device for intratracheal administration of dry powder formulation in mice. The device consisted of a 200 μL gel loading pipette tip connected to a 1 mL syringe via a three-way stopcock, which was shown to disperse 1 mg dry powder formulation with 0.6 mL of dispersion air (11). However, the device emptying and the dispersion efficiency of this device was not reported.

Although a considerable number of devices have been reported to deliver dry powder formulation with a lower volume of dispersion air, only a couple of them have used surfactant formulations. While those surfactant formulations were based on synthetic surfactants none contained surfactant proteins SP-B and SP-C together, which are necessary for best regulating the surface tension in the lungs (40). In our current study, we used a naturally derived surfactant in a commercial liquid surfactant replacement product (Survanta®) containing both SP-B and SP-C to prepare our Survanta-EEG powder and evaluated the performance using the optimal LV-DPI device. The developed optimal device is a unique low volume dry powder inhaler that is able to deliver and disperse Survanta-EEG powders of a commercially available natural surfactant formulation, Survanta® intratracheal suspension.

A potential limitation of the optimal LV-DPI device is the requirement of coordination between actuation and inhalation. In addition, the device emptying also partially depends on the operator of the device since the speed of depressing the air syringe varies for different operators. The operator needs practice to ensure that the syringe is emptied in a consistent manner. However, the use of an automated actuation system can address this issue.

Conclusions

An optimized low air volume dry powder inhaler to deliver spray-dried Survanta-EEG powders was successfully developed by a series of design modifications and determining the influence of those modifications and powder fill mass on aerosol performance. The optimized dose containment unit DPI had a 0.21 mL powder chamber, ST airflow path, three-0.60 mm air inlet holes, and 90 mm outlet delivery tube length with 0.89 mm internal diameter. The powder dispersion of the optimal device was independent of fill mass (3–10 mg) with good powder emptying in one actuation with 3 mL of dispersion air. At a powder fill mass of 10 mg, the optimal device was able to deliver 5.3 mg of spray-dried Survanta-EEG powders with an aerosol Dv50 of 2.7 µm in a single actuation with 3 mL air. The cumulative mass emitted after three actuations with 3 mL dispersion air was >85% of the nominal fill mass of 10 mg. The ability of the device to empty high fill masses with good aerosol dispersion with a very low dispersion air volume will be useful to deliver spray-dried Survanta-EEG powders to small rodent animal models as well as low birth weight infants. Further, the aerosol performance of the optimized containment unit DPI with micronized drug (7.2 mg emitted from 10 mg fill mass with Dv50 of 5.8 µm), suggests the applicability of the device to disperse powders prepared by methods other than spray drying.