Considerations and Approaches for Filling Dry-Powder Inhalers

Published on: 
Pharmaceutical Technology, Pharmaceutical Technology-11-01-2010, Volume 2010 Supplement, Issue 6

The author reviews key considerations for formulating powders for use in inhalers. This article is part of a special Drug Delivery issue.

This article is part of a special issue on Drug Delivery

Dry-powder inhalers (DPIs) are complex multifaceted systems designed to reproducibly deliver an efficacious dose of medicine in an appropriate, aerodynamic range of sizes to enable the treatment of systemic or topical diseases (see Figure 1). Variables that influence the performance and stability of DPIs include the formulation, device, and processes by which the formulation is prepared (e.g. blending, spray-drying) and filled into the device.

Figure 1: Composition of dry-powder inhaler products (FIGURE COURTESY OF THE AUTHOR.)

Factors affecting dry-powder inhalers

Formulation. Most DPI formulations contain a mixture of active pharmaceutical ingredient (API) and an inert carrier material such as lactose. As shown in Figure 2, various attributes of the powder formulation must be considered and managed through the preparation processes and when filling the material into the relevant DPI device or primary device pack (e.g., capsule, preformed blister, blister disc, or geometrical array). In the ideal world, inhalation powder formulations are insensitive to atmospheric and environmental conditions such as light, oxygen, and humidity; are uniform in concentration; flow well; have a low tendency to agglomerate; and do not compact during filling. Under such circumstances, the API might be expected to readily disperse under low energy conditions to produce a uniform and respirable aerosol with an acceptable fine-particle dose. For topical inhalation aerosols, aerodynamic particle-size distributions in the region of 2–5 µm are typically targeted.

Figure 2: Characteristics of dry-powder inhaler formulations. (FIGURE COURTESY OF THE AUTHOR.)

However, powders do not always exhibit these ideal properties. More often, powders are susceptible to moisture which can lead to changes in surface morphology and/or aggregation. In addition, non-uniformity and concentration gradients can result from segregation within the powder bed. These and other factors must be effectively managed during the preparation of formulations and their filling into DPI devices, regardless of device format.

Device. The DPI device into which the formulation is filled can also affect the selection of filling method and equipment. The types of DPIs range from simplistic, low efficiency, passive devices to high(er) efficiency, active delivery devices in which the aerosolization process is driven by an external energy source. In terms of powder filling, the most relevant variables are the size, shape, and array of cavities into which the powder is to be deposited. Reservoir-based devices contain a relatively large powder hopper compared with devices designed to deliver pre-metered powder doses, typically in the 1–30 mg range, usually into either capsules or preformed blister cavities. In this context, the dispensing of powders in the fill weight range of 1–30 mg is defined as microdosing.

Cavity arrangement can affect productivity. For example, the unit operations required to fill and index a circular array of cavities in a disc might require more time than filling a blister strip that indexes linearly. To overcome such device complexities, unique and tailored filling solutions are beneficial.

Although in simple terms, the process of filling DPI products merely requires the reproducible dispensing of powder aliquots into a cavity, the microdosing of ordered mixtures containing irregularly shaped particles of different particle-size distributions and mixed surface morphology represents a significant technical challenge. The need for device-specific equipment for high-speed operations and consistent processing adds to this challenge.

Filling. Figure 3 demonstrates the different factors to be considered (e.g., equipment, speed, scale) when developing either a clinical product and/or a commercial product. Figure 3 also indicates potential batch sizes or the number of filled doses required to support a product evaluation during its development. In formulation feasibility (i.e., pre-investigational new drug/investigational medicinal product dossier), for example, up to and including clinical proof-of-concept (i.e., Phase IIa) batch sizes in the region of 100–5000 doses are adequate not only for enabling clinical evaluation but also for the execution of pharmaceutical development studies. Consider also that fully automated filling solutions are expensive and, as mentioned, tend to be designed to accommodate a specific device format.

Figure 3: Considerations of development stage in filling dry-powder inhalers. IND is investigational new drug. Ph is Phase. NDA is new drug application. MAA is marketing authorization application. (FIGURE COURTESY OF THE AUTHOR.)


The challenge, therefore, has been in the development or identification of fit-for-purpose, manual or semi-automated filling equipment with the capability and functionality to handle inhaled powders in support of early DPI development studies. Although commonly used for early feasibility studies, manual dispensing of single-digit milligram quantities of powder is extremely labor-intensive and requires great manual dexterity. The nature of such manual operations can result in significant fill-weight variability, thereby causing an unnecessary number of rejected doses. Manual operations for inhalation delivery systems also have the potential to cause the powder to compress or to compress in an inconsistent manner. Such factors can contribute to variable delivery from the DPI device causing variability in key product-performance measures such as delivered dose uniformity and fine-particle dose. Such variability can lead to poor decision-making on the part of the inhalation development scientist regarding the suitability of a formulation or product for further development.

Equipment and solutions

To address this capability gap, bench-scale powder handling technologies such as Xcelodose (Capsugel, Cambridge, UK), Powdernium (Symyx Technologies, now Accelrys Inc, San Diego, CA), and the Quantos Perfect Dosing System (Mettler-Toledo, Columbus, OH) have each found niche application in the small-scale microdosing of inhalation powders. These technologies are being used in research and development (R&D) studies as well as in current good manufacturing practice (CGMP) production of Phase I and II clinical-trial materials (1). The data shown in Figure 4a, which are derived from studies conducted in Catalent's Research Triangle Park, North Carolina, facility using the Mettler-Toledo Quantos system demonstrate that powder blends containing micronized API and inhalation-grade lactose can be accurately and reproducibly delivered across, in this instance, a target fill-weight range of 1–25 mg without any powder segregation (2, 3). This assertion is demonstrated by the linear correlation between API assay value per powder aliquot and the total mass of powder dispensed. In this study, percent relative standard deviations (%RSD) of < 1% were achieved for fill weights in the range 5–25 mg and < 2.5% for fill weights between 1 and 2.5 mg (n=10 replicate measurements). These results were obtained without any optimization of the filling process, thereby suggesting that with further optimization, filling precision could be increased.

Figure 4a (left): Effect of powder fill weight on blend concentration-segregation. Figure 4b (right): Effect of filling process on ærosol performance-compaction. NGI is Next Generation Pharmaceutical Impactor (Westech Instruments, Marietta, GA). (FIGURE COURTESY OF THE AUTHOR.)

Additionally, the author received a fine-particle dose of > 50% (i.e., defined as the proportion of the dose delivered from the device with an effective cut-off diameter (< 5 µm), using inertial impaction for the formulation delivered from a simple capsule-based device (i.e., Monohaler, Plastiape, Italy), generated using a Next Generation Pharmaceutical Impactor (NGI, Westech Instruments, Marietta, GA). These data are presented in Figure 4b.

In this latter study, the dose from a single capsule was discharged into the impactor at a flow of 100 L/min. Hence, it can be implied from the comparatively high aerosol efficiency of this simple formulation/device combination that the dry-powder formulation must have been dispensed into the device during the filling process with minimal compaction of the powder blend, thus allowing the API to be readily aerosolized from the carrier material on actuation of the device. Furthermore, it should be noted that through the judicious application of the same processing-equipment configuration, filling procedures, and the use of common personnel, it was possible to achieve effective process transfer and ensure excellent correlation in performance of the clinical product with respect to the early R&D development lots.

Scale up. Although it is possible to conduct effective fit-for-purpose filling trials using small-scale semi-automated equipment, such equipment is not scaleable in a viable manner to allow application for the onward development of a commercial product (1). A commercial product must be produced in large quantities at high speed and with the appropriate degree of consistency required to ensure reproducible product quality. Therefore, one cannot develop a product beyond clinical proof of concept without having in place an effective scale-up path.

As depicted in Figure 3, it is important to bridge the gap from clinical products to commercializable products filled on high-speed, semi-automated filling equipment. It is also crucial to consider the effect of changes to the filling process, equipment, formulation, or to the type of device on the product's pharmaceutical performance (e.g., dose delivery, fine-particle dose), especially when progressing to later stages of clinical development. For example, in these later development stages, higher speed equipment that is tailored to a particular device or product will need to be introduced.

If a simple, low-cost, off-the-shelf, capsule-based inhaler is used for early clinical studies, a more sophisticated inhalation device may be selected for the registration and commercial-product stages. In such a case, it is likely that the new device might generate different aerosol performance and produce differences in dose delivery and fine-particle mass. Technical and clinical studies would be needed to bridge the early clinical and product development data to that of the product under development for commercialization.

In a similar manner, changes made to the mode of device filling or filling equipment can also affect powder flow, segregation within the powder stream or bed, powder compaction, and so forth. Any one of these variables, or many other potential sources of variability, can confer significant changes in the initial performance of the product or in its performance over time (i.e., stability and shelf-life).

Ideally, the inhalation development scientist should consider the selection of a mode or mechanism of powder-formulation filling for early formulation feasibility and then replicate that same unit filling operation throughout the product's development life cycle. Figure 5 provides one example of a scale-up pathway for DPI filling. In this scenario, the unit operation of filling is conducted using a fixed-volume cavity precision bored into a rotating drum. As the formulation is fed to the filling head, an aliquot of powder is drawn into the drum bore through the application of a small, negative air pressure. This negative pressure is maintained as the drum is rotated until the filled cavity is in the six-o'clock position, at which point a small positive pressure is applied to dispense the powder aliquot into the respective dose packaging format (e.g., capsule, preformed blister).

Figure 5: Approaches for filling dry-powder inhalers. IND is investigational new drug. Ph is Phase. NDA is new drug application. MAA is marketing authorization application. Omnidose is a tradename. (FIGURE COURTESY OF THE AUTHOR. FIGURE 5 IS REPRODUCED FROM HARRO HOFLIGER AG/COURTESY AUTHOR.)

Appreciable development work can be required to optimize the cavity geometry and operating pressures for a particular powder (e.g., active concentration, flow, bulk density) and its target fill weight. However, in general terms, the effect of these factors on the filling process are consistent irrespective of process scale. When scaling the filling process, if the unit filling operation is maintained as constant, the major considerations for the process engineer can reduce to those factors affecting throughput product (e.g., uniformity of powder feed, form/seal/cut configuration, number and array of filling heads, in-process verification of accuracy and reproducibility) as opposed to optimization of the unit-filling process. As a result, the likelihood of consistent product quality, consistent product performance and potential to shorten development timelines is greatly enhanced.

In the laboratory scale equipment presented in this example, the process is operator driven (i.e., the powder is dispensed into the small hopper located above the single cavity filling drum). A vacuum line is attached to the filling drum, and using a foot pedal, the operator applies vacuum to the powder to fill the cavity. The operator manually rotates the drum to the dispense position and, again using the foot pedal, applies positive pressure to dispense the powder into the capsule or blister. Operation at this scale requires blisters to be preformed and sealed manually. Intermediate scale semi-automated equipment operates in a similar fashion, however the drum will include multiple filling cavities and contain automation features.

Typically, at this semi-automated scale of operation, the device cavities may be pre-formed, sealed, and cut manually off-line (if blisters) with statistical off-line fill-weight verification. Although significant operator intervention still occurs at this scale, filling throughput in the region of 1000–2000 doses per hour is feasible depending on the device configuration. In the high-speed scenario, all operations occur on-line under high throughput.

Figure 6 contains some examples of alternative proven methods for the filling of powder-based pharmaceutical products along with an opinion on suitability for the processing of inhalation powders. In this context, the dosing principle refers to the specific mode within which the powder is dosed rather than providing any reference to the powder feeding mechanism.

Figure 6: An integrated dry-powder inhaler (DPI) filling process scale-up path. (FIGURE COURTESY OF THE AUTHOR.)


In conclusion, accurate, reproducible, and cost-effective, equipment-based solutions are available to support formulation feasibility and proof-of-concept clinical studies for molecules in DPI devices. However, application of such approaches requires the developer to bridge the technical and clinical programs when working toward product commercialization. Although more expensive, the most robust approach to filling process development, enabling a more rapid path to market, is to adopt a scaleable process from the outset right through the product development life cycle.


This article represents a partial synopsis of a presentation given by Dr. Davies-Cutting at Management Forum, Dry Powder Inhaler Conference, London, July 2010. Special thanks to Harro Höfliger for the use of certain images and data.

Craig J. Davies-Cutting, PhD, is director of inhalation products and technologies, at Catalent Pharma Solutions in Research Triangle Park, North Carolina, tel. 919.465.8430, fax 919.481.4908,


1. Micro-dosing Equipment Fills Niche in R&D, Clinical Trial Materials, Tablets & Capsules, March 2009.

2. L. Mao et al., proceedings of Drug Delivery to the Lungs XIX, Edinburgh, UK, 2008, pp. 195–199.

3. L. Mao et al., "Fast into Man Model for Dry Powder Inhaler (DPI) Development," presented at AAPS, Los Angeles, CA, 2009.