2022 Volume 70 Issue 4 Pages 245-253
It is mandatory to detect the powder cohesiveness of biopharmaceutical dry powder inhaler (Bio-DPI) formulations and their effect on their performance. Normally, Bio-DPI formulations consist of highly cohesive components with higher drug amounts than small molecules. Herein, a formulation study of a high-drug-ratio Bio-DPI was performed, detecting the risk of powder caking in DPI formulations. The Bio-DPI formulation was manufactured via the spray-dry method followed by mixing with excipients. Powder caking was detected through the void forming index (VFI), which was calculated using pressure drop measured by inverse gas chromatography (iGC). Since VFI can be used to evaluate the structural changes induced by powder caking over time with less than 1 g of sample, VFI is considered suitalbe to apply for DPI formulation screening. The risk of powder caking was detected in spray dryed particles at more than 45% relative humidity (RH) humidity condition, mannitol (as a carrier particle) and magnesium stearate (as a lubricant) were added to the formulations. With formulation screening, addition of more than 40% of mannitol was suggested to reduce the risk of powder caking. Selected DPI formulation remained higher emitted ratio (95.6%), than spray dried particle (52.5%) at 25 °C 65% RH condition for 1-month storage. In conclusion, VFI measurement is useful for selecting the DPI formulation by mitigating powder caking risk with limited samples.
A dry powder inhaler (DPI) formulation is an attractive alternative mode of administration for biopharmaceuticals.1) For biopharmaceutical DPI (Bio-DPI), the dosage of active pharmaceutical ingredients (API) and water adsorption behavior are the two main factors that should be considered in developing formulations. Many biopharmaceutical substances (e.g., peptides, proteins, enzymes, and antibodies) highly adsorb water vapor; hence, increasing the dose of the API may result in powder caking due to the presence of water vapor in the atmosphere. Thus, it is essential to understand the effects of humidity on emission efficiency in developing Bio-DPI formulations.
Emission efficiency is an essential property of DPI formulations to deliver an API to the lungs. As powder caking affects emission efficiency, it is fundamental to investigate the risk of powder caking and evaluate the emission efficiency stability during manufacturing or storage.2,3) Additionally, to optimize the delivery of API, the selection of a device considering emission efficiency is also necessary for the DPI formulation.4) In a DPI formulation study, the stability test is usually adapted to detect the risk of emission efficiency decrease during storage and guarantee the quality of the final product. However, the emission efficiency test is time- and labor-consuming.5) In the early stage formulation development, in which the available amount of API is limited, saving time and API amount are strictly required, especially for expensive biopharmaceutical substances. For this reason, it is necessary to evaluate the powder caking property which affects the emission efficiency with less sample amount in Bio-DPI formulation study.
A number of studies have investigated the powder caking property by using indirect and direct evaluation methods. Indirect methods (e.g., crystallinity change and water adsorption amount) can quanititaviely evaluate the change related to powder caking. Although indirect methods are difficult to apply in the case the changes in the physical properties of the powder are limited. On the other hand, direct methods (e.g., hardness measurement, sieving method, and evaluation of micro adhesion) can be applied regardless of the physicochemical properties of the powder. However, most of direct methods are distructive, it is difficult to evealuate the physical change over time. Previous studies6,7) recommended that the void forming index (VFI) be determined to clarify how powder cohesion occurs, which can also be applied to evaluate the risk of powder caking behavior. VFI was calculated from “Pressure drop” which was measured by inverse gas chromatography (iGC). As “Pressure drop” reflect the gas penetration resistant of the sample powder layer, microstructural change related to powder cohesion can be detected as change of “Pressure drop” from initial state. VFI can be measured without sample distruction, the change can be directly measured over time. With a simple preparation method and sample amount of less than 1 g, VFI can be used to detect powder caking quantitatively under the required humidity conditions. VFI has been used to compare the cohesiveness of lactose and mannitol and to elucidate their cohesion mechanism.6,7) Applying VFI to Bio-DPI formulation development can be expected to clarify the influence of humidity on powder caking and emission efficiency while saving the sample amount and time.
Formulation of a high-dose DPI is challenging for Bio-DPI formulations.8) Most commercial products for DPI formulations (e.g., drugs for chronic obstructive pulmonary disease and asthma) require only several micrograms of API.9) However, as intravenous administration of biopharmaceuticals requires amounts in the order of hundreds of milligrams,10) Bio-DPI formulations are expected to require at least several milligrams of API. Recently, Bio-DPI formulations manufactured via lyophilization,11,12) spray freeze drying,13) or spray-drying using lysozyme as a model drug were reported to have a dose at the order of milligrams. Spray-drying is an attractive technique that has been widely used for commercial production with a scale-up methodology. For this reason, the evaluation and development of a Bio-DPI formulation using the spray-dry method is useful for obtaining knowledge for future production.
Hence, the objectives of this study were: (i) to formulate a high-dose DPI formulation using lysozyme as a model drug, (ii) to investigate the effects of excipients on powder caking in DPI formulations using VFI; and (iii) to evaluate the use of VFI for the development of Bio-DPI formulations.
Lysozyme (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), tyloxapol, histidine, and sucrose (Merck KGaA, Darmstadt, Germany) was used for the spray-dried formulation. For the DPI formulation, Parteck® M-DPI mannitol (M-DPI; Merck KGaA) and magnesium stearate (Mg-St; Mallinckrodt Pharmaceuticals, Dublin, Ireland) were selected as excipients. All other chemicals and solvents were of analytical reagent grade, and deionized water was used for all purposes. Mono Dose Inhaler® (MIAT S.p.a, Milano, Italy) was used as the inhaler device.
MethodsPreparation of Solutions for Spray-DryingThe excipients were dissolved in demineralized water, and lysozyme was added to obtain the desired ratio for spray-drying. The total solid content of the spray solution was 9–12% (w/v), and the pH was adjusted to 5.5. Viscosity was measured using a Lovis 2000 M viscometer (Anton Paar, Graz, Austria).
Spray-DryingThe spray-drying process was performed using a Büchi Mini Spray Dryer B-290 (Büchi Labortechnik, Switzerland). The inlet and outlet temperatures (Tin/Tout) were maintained at 100/60 °C. The volumetric flow rate of the drying air was set to 600 L/min. The resulting aspirator flow rate was 37 m3/h. The spray solution (3 g/mL) was atomized with a two-fluid nozzle using compressed air from an in-house supply. After spray drying, the powders were collected through a high-efficiency cyclone in a glass container and transferred to a glass vial. Before use in experiments, spray-dried particle (SpD) samples were stored with a desiccant (silica gel, 13% relative humidity (RH) for at least one week) and passed through a 42-mesh sieve (screen size of 355 µm).
Preparation of DPI FormulationsThe compositions of the DPI formulations are listed in Table 1. Batches with a total amount of 10 g powder (excipient and drug) were manufactured. The excipient blends were mixed for 16 min at 34 rpm in a Turbula® T2C mixer (Willy A. Bachofen AG, Muttenz, Switzerland). The final powder was manually filled into hypromellose capsules (QualiCaps, Nara, Japan) containing a dosage of 50 mg API per capsule. The powder-filled capsules were stored at 25 °C/60% RH for the stability study or 40 °C/75% RH before the emission efficiency experiments.
| Lot No. | SpD (Lot 40) | SpD (Lot 41) | SpD (Lot 50) | SpD (Lot 51) | |
|---|---|---|---|---|---|
| Components | Lysozyme (mg) | 50 | 50 | 50 | 50 |
| Histidine (mg) | — | — | 1.55 | 1.55 | |
| Sucrose (mg) | — | — | 10 | 10 | |
| Tyloxapol (mg) | — | 0.05 | — | 0.05 | |
| Concentration of spray liquid (% (w/v)) | 9.1 | 9.1 | 11.2 | 11.2 | |
The BET surface area of the powder samples was measured using iGC-SEA (Surface Measurement Systems Ltd., London, U.K.). The samples (100 mg) were packed in silanized glass columns (6 mm outside diameter, 3 mm inside diameter, and 200 mm length; Surface Measurement Systems Ltd.) via vertical tapping. Tapping was continued until no visible cracks, hollows, or channels were observed in the body of the powder. Both ends of the column were loosely stoppered with silanized glass wool. The conditioning of the column packed with the sample powder was carried out at 40 °C/0% RH. The measurements were performed using n-octane (methane was used as an inert reference) under the same conditions to determine the BET surface area of the sample.
Surface EnergyThe surface energy was measured using the same powder sample used in the BET surface area measurement. The conditioning of the column packed with the sample powder was carried out at 30 °C/0% RH, and the measurement was performed under the same conditions to determine the surface energy. Methane was used as an inert reference. N-decane, n-nonane, n-octane, and n-heptane were used to determine the alkane series. Dichrolomethane, ethyl acetate, acetone, and ethanol were used as polar probes. Surface energy profiles were calculated using the iGC standard analysis software.
VFI MeasurementsThe VFI was determined by measuring the surface energy using the same powder sample. Nitrogen gas was supplied at a flow rate of 5.1 mL/min at 30 °C. First, the humidity of the gas was set (30, 45, 60, 75, or 90% RH) and purged for 2 h to encourage powder caking of the sample inside the column. Then, the pressure drop of the iGC was measured after conditioning at 0% RH. The conditioning time was determined as the pressure drop change of all samples reached a plateau in the preliminary study.
VFI was calculated using the following equation:
 |
where Pt is the pressure drop of the sample at 6 h (all samples reached a plateau), and P0 is the pressure drop of the initial sample.
Particle Size MeasurementThe volumetric particle size distribution (PSD) of the powder samples was measured via low-angle laser light diffraction using a Sympatec HELOS/RODOS (Sympatec GmbH, Germany). Each powder sample was dispersed using an applied pressure of 3.0 bar. The cumulative volumetric PSD, d10, d50, and d90 were measured (n = 3). The lenses (R2 lens for the SpD sample and R5 lens for the initial lysozyme) were selected based on preliminary information regarding the PSD of each sample.
Bulk DensityApproximately 1 g of the powder sample was used to determine the bulk density of the samples. Following 240 mechanical taps in a measuring cylinder, the Hausner ratio was calculated using the following equation:
 |
where H is the Hausner ratio, ρtapped is the tapped bulk density of the powder, and ρbulk is the bulk density of the powder.
Scanning Electron Microscopy (SEM)The morphology of the powder particles was determined using a scanning electron microscope (SEM; JSM-IT200, JEOL, Japan). The samples were fixed with self-adhesive carbon tape on aluminum stubs.
Dynamic Vapor Sorption (DVS)Approximately 7 mg of powder was weighed on a dynamic vapor sorption system (Surface Measurement Systems Ltd.) and exposed to a continuous flow (200 mL/min) of carrier gas (nitrogen gas) from 0 to 90% RH at 10% increments at 30 °C.
Next Generation Impactor (NGI)SpD samples were analyzed for their emitted ratio (ER) and fine particle fraction (FPF) using an NGI (Copley Scientific Ltd., U.S.A.). The NGI consisted of an induction port, pre-separator, stages 1 to 7, and a MOC, designed to operate at inlet flow rate between 30 L/min and 100 L/min.14) It was operated at 37 L/min7) with emission of capsules inserted in the Mono Dose Inhaler® over 6.5 s, producing an inhaled volume of 4 L. The cut-off diameters of stages 1 to 7 were 10.5, 5.7, 3.6, 2.1, 1.2, 0.7, and 0.5 µm. The amount of lysozyme at each stage was collected by washing with water and quantified via UV absorbance at 240 nm using Nano-drop One (Thermo Fisher Scientific, U.S.A.).
Emitted Ratio (ER)The emitted ratio was calculated using the following equation:
 |
where E is the emitted ratio, Mbefore is the mass of the sample including the device before vacuum, Mafter is the mass of the sample including the device after vacuum, and mloaded is the mass of the loaded sample.
Fine Particle Fraction (FPF)The fine particle fraction was calculated using the following equation:
 |
where Afine is the total amount of lysozyme collected in the stage smaller than the cut-off diameter of 3.6 µm (Stage 3, 4, 5, 6, 7, and MOC). Atotal is the total amount of lysozyme collected in the NGI, including the induction port, the pre-separator, the MOC, and stages 1 to 7.
To deliver a high dosage drug to the lungs, it is necessary to formulate high drug content particles with sizes of approximately 5 µm.15) The formulations of the prepared SpD samples are listed in Table 1. Sucrose and histidine were added to stabilize the protein from drastic pH changes, condensation, and dehydration during the spray-drying process.10) Tyloxapol, a surfactant commercially used for pulmonary administration, was added to reduce shear stress and interface interaction.
The SEM images and physical characteristics of the prepared SpD samples are shown in Fig. 1 and Table 2. A particle diameter of 5 µm was obtained by adjusting two parameters: the concentration of the spray solution and the spray rate. Addition of excipients significantly affected the characteristics of the powder particles. Dents and hollows were observed in Lot 40, Lot 50, and Lot 51. A slight difference was observed in the specific surface area and surface free energy. Previous studies have shown that particle morphology is affected by the viscosity of the solution16) or by colloidal interactions17) in the spray solution. As there was no significant difference in viscosity between each formulation, the effect of colloidal interaction between precipitated excipients was expected. As the surface charge of the droplet can be decreased with the addition of tyloxapol, it was reasonable that Lot 41 showed a smoother surface than Lot 40. However, Lot 50 and Lot 51, which contained sucrose and histidine, were observed to have dents and hollow structures in both samples. As the isotropic shrinkage of the droplet was promoted by decreasing the charges on the particle surface or increasing ionic strength, it was expected that the addition of tyloxapol might promote the formation of spherical particles; however, the addition of sucrose may suppress the formation of spherical particles.
(a: Lot 40, b: Lot 41, c: Lot 50, d: Lot 51).
| Lot No. | Lysozyme | SpD (Lot 40) | SpD (Lot 41) | SpD (Lot 50) | SpD (Lot 51) | |
|---|---|---|---|---|---|---|
| Viscosity of the spray solution (mPa s) (n = 1) | — | 1.3 | 1.6 | 1.5 | 1.7 | |
| Particle size (µm) (n = 3) | d10 | 10.1 (0.1) | 0.7 (0.0) | 0.7 (0.0) | 1.1 (0.0) | 1.0 (0.0) |
| d50 | 31.6 (0.1) | 3.6 (0.1) | 3.0 (0.1) | 4.5 (0.2) | 4.2 (0.0) | |
| d90 | 63.3 (0.2) | 8.8 (0.4) | 7.2 (0.1) | 11 (0.6) | 10.2 (0.2) | |
| Bulk density (mL/g) (n = 2) | Loose | 2.4 | 3.6 | 3.5 | 4.4 | 4.1 |
| Tapped | 1.7 | 2.5 | 2.2 | 2.9 | 2.9 | |
| Hausner ratio | 0.70 | 0.71 | 0.60 | 0.67 | 0.68 | |
| BET surface area (m2/g) (n = 1) | 0.826 | 2.654 | 2.671 | 3.177 | 2.916 | |
| Surface Energy (mJ/m2) (n = 1) | γd | 31.86 | 59.65 | 50.82 | 55.95 | 42.28 |
| γab | 95.74 | 164.4 | 144.19 | 133.58 | 110.51 | |
| γab/γt | 0.75 | 0.73 | 0.74 | 0.70 | 0.72 | |
Data represented as mean (standard deviation) n = 3.
Biopharmaceuticals such as peptide, proteins and enzymes have high solubility and hygroscopicity. These substances are expected to have high cohesiveness. In this study, DVS and VFI measurements were used to evaluate the cohesiveness of the SpD samples compared with mannitol (M-DPI) (Figs. 2, 3). To clarify the effect of humidity on the cohesiveness of the SpD samples, measurements were conducted at a constant temperature of 30 °C under various humidity conditions. The DVS measurement results revealed high water adsorption of SpD sample (Lot 51), and a significant mass increase was observed at humidity levels above 80% RH. In the case of powder cohesion with crystal change or hydration processes, dehydration can be observed in DVS.18) In comparison, no characteristic moisture absorption or desorption was observed in the SpD sample. Therefore, it was difficult to determine the relationship between powder cohesiveness and humidity.
(a:Lot 51, b: M-DPI).
From the VFI measurement results, a decrease in VFI was detected at a humidity of 45% RH or higher, although there was no significant decrease at 30% RH. Since VFI detects the voids formed in the powder layer due to caking from the pressure loss in the gas passing through the powder layer, a decrease in VFI indicates that the sample tends to form powder caking under the measured conditions. In SpD samples (Lot 51), as VFI decreased sharply from 30% RH to 45% RH, it was predicted that the risk of cohesiveness increased in humidity levels above 45% RH. On contrast, M-DPI which have low risk of powder caking, no change was observed in VFI.
As a result, VFI measurement detected powder caking risk at lower humidity, which was difficult to detect via DVS measurement. Since DVS measures the total amount of water absorption including penetration into the powder, it is difficult to detect slight changes in the powder surface of highly hygroscopic powder.6) It was considered that VFI was able to catch the change because it simply detected the change in the structure. It can be noted that VFI simply evaluates structural changes, the differences can be clearly identified without being affected by the hygroscopicity of the powder itself. Additionaly, from SEM images of SpD sample (Fig. 4) stored in 25 °C 60% RH for 2 h, slight interparticle crosslinking could be detected though (Fig. 4), there were no significant change observed overall. It can be concluded that VFI is a useful assessment for powder caking risk.
As the SpD samples showed a risk of powder caking at a humidity of 45% RH or higher, the stability of SpD samples was evaluated after storage at 25 °C/60% RH and 40 °C/75% RH conditions to confirm the effect of humidity on emission efficiency. The results of the emission efficiency analysis are shown in Table 3 and Fig. 5.
| Time point | Initial | 25 °C/60%RH 1 M | 40 °C/75%RH 1 W |
|---|---|---|---|
| Emitted ratio (%) (n = 3) | 69.7 (5.6) | 52.5 (31.5) | 20.3 (6.2)** |
| Fine powder fraction (%) (n = 3) | 43.8 (1.6) | 29.4 (0.8)** | 14.1 (0.5)** |
Data represented as mean (standard deviation) n = 3. * p < 0.05, ** p < 0.01 difference compared to initial sample by students unpaired t-test.
Error bars denote standard deviation n = 3. * p < 0.05, ** p < 0.01 difference compared to initial sample by students unpaired t-test.
After storage at 25 °C/60% RH for 1 month and 40 °C/75% RH for 1 week, the ER decreased from 69.7 to 52.5% and 20.3%, respectively. After dispersion, a residue was observed in the capsule (Fig. 6), so the dispersion force was considered insufficient to disperse the powder cake formed in the capsule. From these results, it was necessary to add excipients to improve the aerosolization stability of the DPI formulation. As powder caking risk in the SpD sample was suggested via VFI measurement, it was confirmed that VFI could be used as a risk prediction tool for the emission efficiency of DPI formulations.
A list of DPI formulations investigated in this study are presented in Table 4. From the results of “Emission Efficiency of the SpD Samples,” the addition of excipients was necessary to reduce the powder caking risk of the DPI formulations. For the DPI excipients, M-DPI and Mg-St were selected to enhance dispersibility. M-DPI showed low cohesiveness compared to other grades of mannitol, as reported in a previous study.7) Mg-St was expected to decrease powder caking by increasing the surface hydrophobicity of the particles.19) The results of the VFI in each formulation are shown in Table 4 and Fig. 7.
| Lot No. | 51 | 100 | 101 | 102 | 103 | 104 | 105 |
|---|---|---|---|---|---|---|---|
| Spray dried particle (mg) (Lysozyme amount (mg)) | 61.65 (50) | 61.65 (50) | 61.65 (50) | 61.65 (50) | 61.65 (50) | 61.65 (50) | 61.65 (50) |
| M-DPI (mg) | 0 | 10 | 10 | 20 | 30 | 20 | 0 |
| Mg-St (mg) | 0 | 2.5 | 5 | 0 | 0 | 5 | 5 |
| Total (mg) | 61.65 | 74.15 | 76.65 | 81.65 | 91.65 | 86.65 | 66.65 |
| VFI (%) (n = 1) | 14.2 | 20.5 | 22.5 | 63.0 | 49.3 | 47.7 | 18.0 |
Regarding M-DPI, an increase in VFI was observed until the addition of mannitol from 0% (w/w) (VFI of 14.2%) to 40% (w/w) (VFI of 47.7 to 63.0%) (no improvement was observed between 40 and 60% (w/w)). From the VFI results, it can be explained that powder caking could be reduced by adding up to 40%w/w M-DPI to the DPI formulation. From the SEM image (Fig. 8) of Lot 100 and Lot 102, small spray dried particles tend to adhere on coarse mannitol surface. As small particle could be a trigger of powder caking, it was considered that entrapment of the small particle by coarse mannitol reduced the risk of powder caking. However, no improvement was observed in powder caking with the addition of Mg-St. From the SEM image comparing Lot 51 (Fig. 4) and Lot 51 (Fig. 8), no clear difference was observed in particle surface. The surface modification seemed to be insufficient, owing to the weak shear stress in the mixing process. Additionally, the initial ER improved from 69.7% (Lot 51) to 91.8% (Lot 100) and 91.9% (Lot 102) (Table 5) in the presence of coarse mannitol particles. From the above results, the stability test at 25 °C/60% RH was evaluated for the three formulations: Lot 51, Lot 100, and Lot 102.
| Lot No. | 51 | 100 | 102 | |||
|---|---|---|---|---|---|---|
| Time point | Initial | 25 °C/60% RH 1 month | Initial | 25 °C/60% RH 1 month | Initial | 25 °C/60% RH 1 month |
| VFI (%) (n = 1) | 14.2 | 20.5 | 63.0 | |||
| Emitted ratio (%) (n = 3) | 69.7 (5.6) | 52.5 (31.5) | 91.8 (2.9) | 86.1 (5.8) | 91.9 (1.2) | 95.6 (3.7) |
| Fine powder fraction (%) (n = 3) | 43.8 (1.6) | 29.4 (0.8)** | 42.2 (2.1) | 21.9 (0.4)** | 38.5 (1.5) | 21.2 (1.7)* |
Data represented as mean (standard deviation) n = 3. * p < 0.05, ** p < 0.01 difference compared to initial sample by students unpaired t-test.
Furthermore, the sample amount and measurement period required in this screening study should be noted. To evaluate the stability of DPI formulations over time under various conditions for each formulation, several grams of API is necessary to evaluate each formulation. For example, if 50 mg of API was set in each device, three conditions and five time points would require a total of 1.95 g, so API in the order of tens of grams is required in each DPI formulation. However, for VFI measurement, only 50 mg of API (100 mg as DPI formulation) was sufficient in each formulation, regardless of the inhaler device used. Regarding the measurement period, each stability study required approximately one month, whereas, with VFI, it could be evaluated in only 6 h. The required amount of API can be reduced by sorting candidate formulations using VFI measurements before performing the stability test.
Evaluation of Aerosolization Stability of DPI FormulationThe results of emission efficiency after storage at 25 °C/60% RH are shown in Table 5 and Fig. 9. It was confirmed that the DPI formulation, which showed a higher VFI than the SpD sample (Lot 51), can suppress the decrease in the ER during storage at 25 °C/60% RH. No decrease in ER was observed in the formulation containing 40% M-DPI, and a high ER was maintained. This result proved that VFI could predict the stability of DPI formulations and is a useful tool for designing DPI formulations. Additionally, there was a discrepancy between the VFIs, but not in the emission efficiency, of Lot 100 and Lot 102 (Table 5). Lot 102 (Fig. 10a) did not undergo powder caking and maintained a high emission efficiency. Lot 51 (Fig. 10c) formed a powder cake, and the emission efficiency decreased. Lot 100 (Fig. 10b) also formed a powder cake, but the emission efficiency decrease was suppressed because the powder cake was completely dispersed by the airstream generated by the inhaler device. These findings indicate that Lot 100 has the potential risk of powder caking and that this risk cannot be detected from the emission efficiency test on the sample stored at 25 °C/60% RH alone. Interestingly, it was also found that FPF is not always correlated. It can be explained that the desorption of SpD from the carrier particles was due to local changes that are even smaller than the microstructural changes evaluated via VFI. In general, an API-specific approach can be used to improve FPF by modifying the physicochemical properties of API particles. Importantly, the ER should be considered in DPI formulation screening and device selection for early-stage DPI formulation studies.
Error bars denote standard deviation n = 3. * p < 0.05, ** p < 0.01 difference compared to initial sample by students unpaired t-test.
(a: DPI formulation with low powder-caking risk, b: DPI formulation with potential powder- caking risk, c: DPI formulation with high powder-caking risk).
The VFI detects the powder caking property without the effect of device dispersibility, so it can be used for selecting the appropriate device for the formulation to be used. A highly dispersible device is necessary for DPI formulations with a high powder-caking risk (such as those in Figs. 10b and 10c). For a low powder-caking risk DPI formulation (Fig. 10a), devices can be selected with a lower dispersible property. Typically, a complex structure and difficult production are required to design a highly dispersible device. For this reason, the selection of a device based on the powder caking property of a DPI formulation is desirable. Adapting the VFI in DPI formulation development is expected to improve formulation and device screening efficiency by reducing emission efficiency and long-term stability studies.
VFI measurement can be an effective method for selecting candidate formulations to mitigate the risk of powder cohesion with a limited sample (100 mg each formulation) in a short measurement time (6 h). VFI evaluation proved to be a useful tool for designing Bio-DPI formulations, especially in early-stage device development.
Additionally, we proposed a high-dose Bio-DPI formulation that improves emission efficiency by considering the risk of powder caking. VFI measurement in the DPI formulation study clarified that the addition of M-DPI might reduce the risk of powder caking, and a stability study supported this result.
With its high versatility and potential savings in time and labor, we believe that VFI can be adopted in future Bio-DPI formulation designs.
The authors would like to acknowledge Dr. Nantharat Pearnchob-Höhling (Daiichi Sankyo Europe GmbH, Germany) and Dr. Yasuhiro Tsutsumi (Daiichi Sankyo Co., Ltd., Japan) for their valuable advice regarding the preparation of this scientific article. The authors would like to acknowledge Yumiko Fujii and Dr. Norihiro Nishimoto (Daiichi Sankyo Co., Ltd., Japan) for their valuable advice on this scientific article.
Sunao Maruyama: Conceptualization, methodology, data curation, and writing–original draft. Makoto Miyajima: Writing, review, and editing. Etsuo Yonemochi: Conceptualization, methodology, writing–review and editing.
Sunao Maruyama and Makoto Miyajima are employees of Daiichi Sankyo Co., Ltd. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
This article contains supplementary materials.
