Paul M. Young

Dry powder inhalers (DPIs) have advanced significantly over the past 10–15 years.

  This has been driven by the need to reformulate pressurized metered dose systems to chlorofluorocarbon (CFC)-free systems using hydrofluoroalkanes and the potential to deliver a wider range of therapeutics to the respiratory tract (through improved stability and greater dose choice). 

Apart from medication that specifically targets asthma, other therapies under investigation that use DPI systems include both local delivery (for example, mucolytics, antibiotics, antituberculosis and anticancer drugs) and systemic delivery (for example, insulin, analgesics for cancer pain and drugs for erectile dysfunction).

  In simple terms, the formulation of a DPI generally involves: 					

The production and characterization of a powder suitable for respiratory deposition.

The formulation with or without excipients in a device.

The system is characterized by aerosol performance and stability, and optimized through empirical observation. Although highly simplified here, these key stages of development will be discussed in this article.  

Arguably the nature of the active pharmaceutical ingredient (API) particles is the key to effective DPI delivery. The conventional way to produce a respirable size range API (<5 μm aerodynamic diameter) involves the crystallization from a solution followed by high-energy milling. Although this method is still the most common way to produce micronized APIs, it is not necessarily the most attractive as it requires a multistep process, results in particulates with ill-defined geometries and leads to questions concerning polymorphism and the degree of amorphous content. 

The deliberate control of the API's physical form, geometry and surface composition can be dated back to the 1980s when Gonda and co-workers successfully used elongated particles for their advantageous aerodynamic behaviour,

  and particles coated with hydrophobic additives for the reduction of hygroscopic growth in the airways after inhalation.

  Recently, there has been renewed interest in using engineered particles for DPI formulation to improve the powder dispersibility. The objective is to make 'nonsticky' particles that have a small aerodynamic diameter. The approach includes manipulation of the particle size (including nanoparticles), shape, density, surface morphology and other surface properties.

Single-step production methods have been developed with spray drying being the prominent one for inhaled insulin (approved by FDA) as well as inhaled mannitol, which was approved by the Australian Therapeutic Good Administration (TGA). Furthermore, the addition of binary/ternary components into the initial spray dryer solution may allow increased physical stability in primarily amorphous systems and/or the production of specific morphologies with, for example, improved aerosolization efficiency and reduced density. Other methods of API production include supercritical fluid (SCF) precipitation, high gravity precipitation and production through ultrasonic crystallization.

The formulation of many APIs requires the addition of binary or ternary agents after production to improve powder flow, improve dose reproducibility and act as a diluent for highly active compounds to fulfil the regulatory requirements and produce the anticipated clinical outcome. The traditional approach of powder blends containing the API mixed with lactose as a carrier is still most widely used for DPI products. Furthermore, in recent years, this approach has not only been used for dilution and improving the flowability, but also for enhancing the dispersibility of the micronized API. 

There have been significant advances, especially in understanding the role of carrier properties on the aerosol performance of the API. For example, early work suggested the roughness and the specific size of lactose carriers may be related to the aerosolization performance of an API through variation in contact geometry.

Such observations were substantiated and built upon in recent work, which suggests a complex interplay between performance and the degree of fine lactose,

  and aerosol API performance. The exact mechanism of adhesion and drug liberation during aerosolization is still not clearly understood and is a major focus among researchers within the field. 

Clearly, by controlling such parameters, the aerosolization performance of a particular API may be 'fine tuned' to acquire the chosen respiratory deposition, while maintaining product stability. Subsequently, many relevant carrier engineering technologies have been reported to improve specific performance parameters. 

Some of the key developments include the improvement of API aerosolization performance by controlled crystallization of the carrier: 

 Lactose crystallized from Carbopol gel achieves more regularly shaped lactose crystals with a smoother surface compared with the control lactose, resulting in higher efficiency and reproducibility of drug delivery by DPIs.

 The engineering of lactose carrier surfaces using a proprietary particle smoothing process results in significant differences in surface morphology when compared with the 'as supplied' starting material, and leads to improved aerosolization efficiency from commercially available DPIs.

Controlled etching of the carrier particles to reduce variation in API adhesion.

 Treatment of the lactose carrier by solution-phase variable temperature dissolution with a suitable liquid that can control the particle surface dissolution, combined with the removal of fines produce materials that increase aerosolization of the active from a single dose inhaler.

 Addition of ternary agents, such as magnesium stearate and/or L-Leucine, reduce API carrier adhesion; that is, the use of spray-drying produces a specific physical form of L-Leucine that is suitable as a flow aid during aerosolization.

 Addition of carrier fines can reduce the variation of API-carrier adhesion by filling active sites and/or by forming complex binary/ternary mixtures; that is, investigating the order of mixing of ternary components, examining particle interactions between drug and fine excipient particles (intrinsic or externally added ternary particles); and investigating adhesional properties of the lactose carrier surface using specific techniques such as the atomic force microscopy (AFM) colloid probe technique (CPT).

The inhaler device is as important as the powder formulation for the successful development of DPI products. Powder dispersion in DPI devices is accomplished either passively by the patient's own inspiration, or actively by a suitable component, such as an electronic vibrator or impeller, in the DPI device. Because micronized particles are difficult to deaggregate, the efficacy of passive DPI devices depends critically on a patient's inspiratory flow rate. 

As a result of the recent focus on tightening regulatory requirements, it is desirable for the performance of (breath-actuated) passive inhalers to have minimal dependence on the patient's inspiratory air flow. Furthermore, these DPI devices are now much more sophisticated than they used to be when only single-dose, capsule-loading systems existed. 

From the development perspective, 'passive' multidose DPIs fall into two main categories. They either measure the dose from a powder reservoir or they disperse individual doses that are premetered into blisters by the manufacturer. In general, the premetered drug-in-blister/capsule approach is easier to develop because the reproducibility of the metered dose can be ensured during the drug formulation and packaging, but for the purpose of this article the authors will mainly focus on novel multidose devices. 

One of the most recently marketed multidose DPI products in the US, Asmanex Twisthaler (Schering-Plough, Kenilworth, NJ, USA), for the delivery of mometasone furoate has shown highly reproducible results on emitted dose and fine particle fraction (FPF) using a patented agglomerate formulation.

  In the Twisthaler, the DPI nozzle comprises two parts: a lower swirl chamber and upper chimney in the mouthpiece. A patented fluted chimney design was used to produce a stronger vortex with an increased number of particle collisions with the chimney for deagglomeration.

The Novolizer (Meda Pharma BV, Amstelveen, The Netherlands) is a multidose DPI device already marketed in Europe.

  The Novolizer works on the cyclone principle and air classifier technology. These are robust concepts involving both centrifugal force and drug force acting on the particles swirling in a cylindrical chamber (cyclone), which coupled with collision force achieves drug particle detachment from carrier surfaces.

Another multidose device using a cyclone (in the mouthpiece) for drug deagglomeration is the Next DPI.

  The Otsuka DPI (ODPI [Otsuka Pharmaceutical Co., Ltd, Tokyo, Japan]) disperses a freeze-dry cake containing the active drug by inspiratory air flow of the patient during use.

  This is particularly suitable for delivering biological products, as each dose is individually contained in a vial which eliminates problems caused by contamination and microbial growth. 

Other recent devices include the Acu-Breathe multidose DPI by Respirics Inc. (Raleigh, NC, USA),

  a prototype DPI by Oriel Therapeutics Inc. (Research Triangle Park, NC, USA) using a piezoelectric polymer, which vibrates to deliver the powder,

  and a patented powder dispensing device based on the rotating fluidized bed principle.

  While most DPIs deliver relatively low doses of drug, the Actispire by Britannia Pharmaceuticals  (Redhill, UK) is suitable for high dose delivery (e.g., 20–250 mg of a synthetic lung surfactant).

  This device uses a pressurized gas to aerosolize micronized powder contained in the individual vial. 

There are numerous studies comparing the 

 performance of DPIs, but few provide insights into the mechanisms of powder dispersion in the inhaler devices. Such information is beneficial and would allow rational design of a more efficient inhaler with enhanced performance. The designs of the Twisthaler and the Next DPI were aided by computational fluid dynamics (CFD), but unfortunately, very little details were provided.

Recently, the Aerolizer (Figure 1) was investigated in a series of fundamental studies linking experimental powder dispersion data with CFD simulation of air flow in the inhaler.

  These results provided a deeper understanding of how the inhaler performance is influenced by DPI design features, including the design of a grid, the presence of a capsule and the dimension of the air inlet — all these are common features present in DPIs. 

 The grid was found to have three functions: 					

Helps straighten the air flow in the mouthpiece.

Generates small eddies immediately above the grid in the mouthpiece. 

Allows high-velocity collision between the particles and the grid to occur. A high FPF can be achieved using a grid that will generate more energetic turbulence and higher frequency of particle–grid collision.

 A number of passive DPIs use a capsule to contain a unit dose, followed by emptying the powder into the air stream inside the inhaler after piercing the capsules. It was found that the presence of a spinning capsule in the Aerolizer significantly reduces the turbulence (i.e., the potential for powder deagglomeration). However, the dimension of the pierced hole in the capsule is critical because deagglomeration occurs when the powder is sheared through the holes when it is emptied from the spinning capsule. 

 All DPIs contain some form of air inlet to allow entrainment of air into the device for powder emptying and dispersion as the patient inhales. If the air inlet dimension for the Aerolizer is reduced, it takes longer for the air flow to be fully developed in the inhaler, and the capsule spins faster (as a result of the higher inlet air linear speed) to empty the powder out into the air stream sooner. 

When the powder is emptied into the air stream before the air flow is fully developed, powder deagglomeration will be less effective, making the inhaler performance actually poorer with a smaller air inlet. This was observed for the Aerolizer operating at 60 and 90 L/min with a one-third or two-third inlet size. The results, therefore, highlighted the importance of synchronizing powder release and turbulence build-up in the inhaler. 

The market is rapidly expanding and many novel inhalation devices are in development. Several companies are active in the area, including Aradigm, Batelle Pharma, Chrysalis Technologies, Boehringer-Ingelheim and MAP Pharmaceuticals.

  All are developing 'active' inhalers that offer the advantage of dosing precision and an energy source to enable patient-independent, reproducible aerosol production. 

Recent research into multidose DPIs has been presented in this article. The aerosolization performance of a DPI is a complex system that clearly requires more fundamental studies. It is anticipated that future development of DPI products will rely more on rational design of both the inhaler device and powder formulation properties to attain the optimal products. 

Is a lecturer all at the Faculty of Pharmacy, University of Sydney (Australia). 

Is a former PhD student at the Faculty of Pharmacy, University of Sydney (Australia) and currently senior scientist at Pfizer Ltd (UK). 

Articles

There have been significant advances, especially in understanding the role of carrier properties on the aerosol performance of the API.

Dry powder inhalers: challenges and goals for next generation therapies

Dry powder inhaler (DPI) products typically consist of a device containing a blend of micron-sized drug particles and larger, so-called 'inert' carrier particles. Upon aerosolization, the drug is liberated from the carrier using the inspiration force of the patient to penetrate the lung. The carrier, typically alpha-lactose monohydrate, is used as a diluent and flow aid, improving content uniformity and stability, and to facilitate reproducibility in emitted dose. However, even though there are many marketed DPIs, it is clear that the successful manufacturing of a DPI product still poses many technological, regulatory and scientific challenges.  

The usual approach by pharmaceutical companies is to design — or to obtain the rights to — a novel inhalation device, which can preferably be used for a wide range of different drugs, concentrations and applications. The DPI products are then typically developed using iterative processes, that is, formulations are based on empirical studies (albeit ones informed by experience) until satisfactory DPI performance and blend characteristics are achieved.  

One important aspect of this approach is the tailoring of the carrier material to the required formulation and device characteristics. This can ideally be achieved by using 'as supplied' or 'off-the-shelf' samples of lactose.  

Alternatively, such samples could be processed by the user to produce carriers with the required characteristics. Increasingly, lactose manufacturers are supplying customer grades that are produced to a tailored specification, which might result in supplier/user issues.  

In terms of pharmaceutical drug delivery, DPI carrier technology can be considered a low dose system, which poses challenges both to formulators and, consequently, excipient suppliers. This is exemplified by the fact that a drug dose of 200 μg is considered a relatively 'high' dose in DPIs, and yet this can correspond to <1% drug concentration in the blend; for example, Miflonide (budesonide); Ventolin, Pulvinal (both salbutamol); and Seretide (salmeterol/fluticasone).

Additionally, carrier-based higher potency medications exist that require even lower doses (for example; Foradil, formoterol fumarate, 12 μg, ~0.05% drug load), posing obvious formulation and analytical challenges.  

In terms of DPI formulation development, any carrier for DPI applications should ideally be: 					

 fully characterizable for parameters known to be crucial for performance 

 stable (crystalline and not contain any uncharacterized metastable structures) 

It is difficult for any excipient to meet these criteria and, as with many excipients, there are still questions concerning true physicochemical characteristics, and importantly, their effect on DPI characteristics. Additionally, there are still questions concerning supplier and user characterization methodologies. 

Lactose DPI carriers are commercially available from several sources (Table 1). The volumes of lactose manufactured for this application are very small compared with those used for oral applications.  

 Table 1 Lactose and inhalation grade lactose suppliers.			

However, in such a specific 'technology platform' the excipients provide significant technical and financial challenges to suppliers and users because they will require greater product and regulatory scrutiny for consistency and quality than would be expected for comparable products in general tablet and capsule applications. 

The potential issues concerning any DPI excipient can be generally divided into physicochemical characteristics and residues. 

It is possible to describe lactose monohydrate products in terms such as particle size, shape factors, roughness, moisture sorption, surface area and surface energy, among other parameters. 

However, the exact relationships between such descriptors and functions are, as yet, not fully understood and there is continuing research by the pharmaceutical industry to investigate the relationships between such characteristics and DPI performance. 

 As with all aspects of pharmaceutical development, there are questions concerning excipient (and drug) intra-batch, inter-batch and inter-supplier equivalence. The effect of any such variations in pharmaceutical applications will depend on the robustness of the drug delivery vehicle and formulation specifications. Consequently, any subtle changes in lactose monohydrate powder properties (e.g., particle size distributions and surface roughness) may have dramatic effects on formulation performance, with obvious product and regulatory implications if the product is not sufficiently robust (e.g., one that has not been developed to meet 'Quality by Design' criteria, to accommodate those changes). 

 There are many ways to describe a powder in terms of its particle size distribution (PSD). Particle size descriptors are important at the manufacturing (supplier) and formulation (user) stages of pharmaceutical development. In terms of suppliers, specific PSDs are produced by sieving or classifying. While both methods have different inherent effects on the PSD, for example, in terms of size cut-offs, they can both be used to produce powders with suitable PSDs.

The PSDs, in terms of a product descriptor, can be characterized by suppliers and users by laser light diffraction. It is important that PSD measurement methodologies employed by suppliers and users are consistent and, importantly, validated and understood. This not only includes the actual methodology, but also sample handling procedures and sampling frequencies. 

 Lactose can exist as alpha–monohydrate, beta–anhydrous, stable alpha–anhydrous, unstable alpha–anhydrous and a 'mixed crystal' of anhydrous alpha and beta lactose. Depending on environmental conditions, lactose can also be stable in an amorphous state.  

It can be argued that any excipient used in solid dosage forms, including any ideal carrier for DPIs, should be stable (within known parameters) during the stability claims of the product. Therefore, the presence of any metastable material in the carrier, in particular in the surface of the carrier particles, may be a source of formulation variability, particularly for low dose products. In terms of stability, lactose monohydrate is crystalline and the most stable. All the forms of lactose can be described from a strictly physicochemical viewpoint as metastable; they all revert to lactose monohydrate depending on storage conditions and time. In general, there are analytical techniques available to quantitatively assess the presence of these forms. However, the current analytical limit of about 1–2% may not be sufficient to adequately characterize the bulk and surfaces of the excipient.  

 The subjection of organic materials to mechanical stress can result in the formation of an amorphous material. The amorphous contents of milled and sieved grades of commercially available lactose monohydrates, including DPI grades, have been reported to be 1–4% and <1%, respectively, determined by differential scanning calorimetry or microcalorimetry.

  Obviously, such metastable sites will have different interactions with drug (and excipient) particles than 'stable' crystalline regions.  

Since amorphous lactose can crystallize instantaneously above a water activity of approximately 0.5 at room temperature, there must be questions concerning the physicochemical composition of DPI lactose monohydrates, both when supplied, and during the DPI formulation and product life cycle. 

 Anhydrous lactose can be present at low levels in lactose monohydrate, for example, by crystallization of amorphous lactose — the level of which depends on the crystallization conditions.

  It is metastable and reverts to lactose monohydrate at high water activities over relatively long time scales. Raman microscopy of crystallized amorphous lactose suggested that domains of beta lactose were present in the crystalline surface.

The excipient anhydrous lactose as used in certain inhaler formulations is a good example of the potential difficulties encountered when considering the relationship between purity and pharmacopoieal acceptance — since there is no compendial limit on anhydrous alpha and beta levels posing obvious manufacturing and characterization challenges.  

 The presence of very low levels of residues (including solvents) in excipients is becoming an important issue for users, and the presence of very low levels of Â´non-lactoseÂ´ species in DPI lactose may pose challenges to suppliers and users. Such components are present because of use of the natural product milk, and the manufacturing processes involved in the production of lactose. 

 Lactose, like many excipients, is derived from natural sources, in this case the whey from milk. As with other excipients derived from natural sources (such as microcrystalline cellulose and starch), it would be expected to contain natural 'impurities'. For example, in terms of pharmaceutical applications, the issue of the presence of proteins was recently highlighted by a report that milk proteins have been identified in DPI products.

  Allergic reactions to products containing milk proteins have been reported in patients with a severe milk protein allergy.  

There is no specific assay for proteins/peptides in the monographs for the different polymorphs of lactose. Nevertheless, the monographs describe solution absorbance tests. Of these, perhaps the test for absorbance at 400 nm is a good indicator of protein levels, since absorption at this wavelength is an indication of the degree of protein-based chemical reactions such as the Maillard reaction, an obvious source of potential stability problems.  

Currently available techniques to directly determine the content of proteins in lactose (e.g., via nitrogen determination by Kjeldahl) may, as an analytical technique focusing on the bulk and not on the surface, not be specific enough to sufficiently reflect the influence these impurities may have on the performance of the resulting product. Investigations using X-ray photoelectron spectroscopy (XPS) indicated that up to 1% of atoms in the top 10 nm layer of a specifically produced DPI lactose are nitrogen in the peptide configuration.

  This may indicate a level of up to 10% for peptide impurities in the surface areas.  

The effect of such high levels of these impurities on apparent DPI performance has yet to be determined, but may be important for the performance of any potential protein-based DPIs since these proteinaceous materials may be capable of changing the apparent aerosolization or product stability; for example, in milk powders the Maillard reaction and oxidation are reported to enhance protein interactions.

Lactose as a natural bovine product, derived from cow's milk, which is manufactured by the aid of calf's rennet (or by biotechnologically manufactured Chymosin) might be considered to be a potential risk factor to transmit BSE (bovine spongiform encephalopathy) infectious prions. However, according to the Guideline EMEA/410/01 Rev. 2 "milk...is unlikely to present any risk of TSE (transmissible spongiform encephalopathy) contamination...", irrespective of the geographical origin (and, therefore, to comply with the guideline), provided specific risk reduction factors are being met. Although milk is unlikely to present any risk, the theoretical residual health risk may be further reduced by sourcing the milk from countries thought to be free of BSE and/or by choosing lactose which was manufactured with the aid of biotechnologically engineered Chymosin (to avoid usage of calf's rennet).

The adoption of The International Conference on Harmonization (ICH) Quality 8 (Q8) could represent a milestone in the way that pharmaceutical products are developed and chemistry and manufacturing control (CMC) sections are accepted by worldwide regulatory agencies.

  The text of this guideline states "those aspects of drug substances, 

, and manufacturing processes that are critical and that present a significant risk to product quality, and therefore should be monitored or otherwise controlled, should be identified and discussed".  

ICH Q8 formalizes the ideas of 'Design Space' so that products are better understood and that product failures are less frequent. An important and welcome feature of the development paradigms that this guideline formally introduces is that it is the quality of the data developed that should be important, rather than the quantity of data.  

It can be argued that any of the performance factors noted above could have a significant influence of the performance of a DPI, and thus a sensible design of experiments for a lactose–based DPI product could include, amongst others

 Multiple batches of lactose (perhaps sourced at different times of the year). 

The presence of very low levels of residues (including solvents) in excipients is becoming an important issue for users, and the presence of very low levels of ?non-lactose? species in DPI lactose may pose challenges to suppliers and users.