Ebook Pulmonary drug delivery: Part 2

(BQ) Part 2 book “Pulmonary drug delivery” has contents: Particle engineering for improved pulmonary drug delivery through dry powder inhalers, particle surface roughness – its characterisation and impact on dry powder inhaler performance, drug delivery strategies for pulmonary administration of antibiotics,… and other contents.

8 Particle Engineering for Improved Pulmonary Drug Delivery Through

Dry Powder Inhalers

Waseem Kaialy 1,∗ and Ali Nokhodchi 2,3,†

1School of Pharmacy, Faculty of Science and Engineering, University of Wolverhampton, UK

2School of Life Sciences, University of Sussex, UK

3Drug Applied Research Center and Faculty of Pharmacy, Tabriz University of Medical Sciences,

Iran

Abbreviations

List of Abbreviations

𝜌true or Dtrue True density

Pulmonary Drug Delivery: Advances and Challenges, First Edition Edited by Ali Nokhodchi and Gary P Martin.

© 2015 John Wiley & Sons, Ltd Published 2015 by John Wiley & Sons, Ltd.

FPD Fine particle dose

advan-of localized diseases such as asthma and other pulmonary conditions Delivery via the airways hasalso been used in the treatment or management of systemic diseases (e.g., diabetes) [2, 3] Recombi-nant human deoxyribonuclease (rhDNase, dornase alpha) was the first recombinant protein approvedfor therapeutic use by inhalation delivery [4] It is apparent, therefore, that this route of drug delivery

is important, but performance of inhalation products is still generally poor This is one of the mainreasons for interest in the modification of drug or carrier particles by particle engineering, so as to

enhance the performance of formulations in vivo, and as assessed in vitro.

8.2 Dry Powder Inhalers

Dry powder inhalers (DPIs) comprise a pharmaceutical dosage form of increasing popularity withattractive features such as the provision of a propellant-free means of drug delivery and such formu-lations have rapidly increased in number, worldwide Some of the most commonly used DPI dosageforms are shown in Figure 8.1 More discussion and details on specific inhaler devices are availableelsewhere (see Chapter 3)

8.3 Particle Engineering to Improve the Performance of DPIs

In attempts to enhance the efficiency of delivery from DPIs, several techniques have been utilised toprepare particles of active pharmaceutical ingredients (APIs) and carriers (e.g., lactose or mannitol),

Diskhaler TM

Easyhaler TM Turbulaher TM Aerolizer TM Rotahaler TM Spinhaler TM

Clickhaler TM Handihaler TM Twisthaler TM Novolizer TM

Figure 8.1 Some common DPI devices

under controlled conditions These include simple crystallization techniques, spray-drying, drying, supercritical fluid (SCF) technology and antisolvent technology These are considered in turn

A crystallization process is characterized by the formation of supersaturation, nucleation and tal growth, in addition to secondary phenomena including aggregation, agglomeration, breakage,re-dissolution and aging [5] Crystallization is usually associated with several challenges includ-ing poor mixing and crystal break/agglomerate formation [6] Fundamentally, the difference of thechemical potential between the supersaturated solution and the solid crystal face is the driving forcefor crystallization Typically, supersaturation can be created in the crystallization media by cooling,evaporation of the solvent and/or the addition of an antisolvent

crys-Batch cooling crystallization is a widely used technique for the production of high-value chemicals[7] The slow crystallization rate is the main disadvantage of the cooling crystallization technique [8].This is due to the relatively large width of the metastable zone that requires a high supersaturation

to induce crystallization [9] Nevertheless, slow cooling is advantageous in terms of attaining themaximum yield, minimum agglomeration [10], fewer defects in the crystal lattice [11] and highproduct purity [12]

Antisolvent crystallization is a process where an organic product can be recovered from aqueoussolutions through the addition of non-solvent compounds by which the solute solubility is decreased,without creating a new liquid phase [13] The successful antisolvent must be miscible with the motherliquid but in which the solute is insoluble Under such conditions, solute solubility is reduced butnot completely inhibited Antisolvent crystallization using alcohols does suffer from disadvantagesincluding the requirement for solvent recovery and the risk associated with the use of flammable

solvents at high reaction temperatures [14] Crisp et al [15] reported that the crystal size increased

as the antisolvent proportion decreased The antisolvent crystallization technique has been proven

to be a potential technique for the preparation of particles, such as salbutamol sulphate [16] andbudesonide [17], both generated an improved DPI performance upon aerosolization (Figure 8.2)

It has been documented that the properties of an antisolvent crystallized product are dependent

on several processing parameters such as the type of the antisolvent [18], solution tion [19], agitation intensity [20], antisolvent addition rate [21] and mixing conditions [22] During

concentra-Salbutamol sulphate

Figure 8.2 Engineered drug particles prepared by antisolvent crystallization for DPI systems: (a)

salbuta-mol sulphate (Source: Reproduced from [16], with permission from Elsevier) and (b) budesonide (Source: Reproduced with permission from [17] Copyright © 2008, American Chemical Society)

crystallization, mechanical stirring can introduce random energy fluctuations within the solutionleading to a heterogeneous distribution of local concentrations, resulting in heterogeneous crystalgrowth [23] On the other hand, particles with a narrow size distribution and a regular particleshape can be prepared by suspending the crystals in a gel [24], in which secondary nucleation(heterogeneous nucleation) occurs to a much lesser extent [25]

Spray-drying is a drying technique in which a dry powder is produced by evaporating the liquid fromthe atomized feed when it mixes with the drying hot gas medium In DPIs, the spray-drying techniquehas been employed to produce not only drug particles [26–29] but also carrier particles [26, 30, 31]

A few examples are shown in Figure 8.3

Spray-dried drug particles can produce higher respirable fractions than micronized particles, andthis has been ascribed to their spherical shape resulting in less drug–carrier contact area and in turnless drug–carrier adhesive forces [23] (Figure 8.4) Moreover, spray-dried particles may have morehomogenous particle size distribution (PSD) [33]

One of the most important advantages of spray-drying techniques is the opportunity to generateparticles with pre-determined characteristics, e.g., size, morphology, shape and density, seeking tooptimize the powder properties such as bulk density, flowability and dispersibility [34] Such charac-teristics of the spray-dried particles can be controlled by manipulating several parameters includingthe composition of the solvent [35] and coating of particles by an excipient (e.g., leucine) In addition,other parameters such as solute concentration, solution feed rate, gas feed rate, drying rate, viscosity

of the liquid feed and relative humidity are able to alter the characteristics of the resultant spray-driedparticles [36]

Various methods can be employed to determine aerosol PSDs, which depend on various geometric

used and most relevant parameter to express aerosol particle size [37], a parameter that also

and is defined as the diameter of a sphere having the same volume and a unit density This assumesthat such a ‘hypothetical’ particle impacts on the same stage of the impactor during aerosolization(or has the same impaction characteristics) as the real particles being measured [38] The theoretical

Daeof particles (Dae= Dg× (𝜌true /X)0.5) can be calculated from the particle true density (𝜌true or Dtrue)

Figure 8.3 SEM photographs for different particles used in DPI systems: (a) gentamicin (Source:

Repro-duced from [27], with permission from Elsevier), (b) cromolyn (Source: ReproRepro-duced with permission from [28] Copyright © 2007 Wiley-Liss, Inc.), (c) budesonide (Source: Reproduced from [29] with kind permis- sion from Springer Science and Business Media) and excipient: (d) mannitol (Source: Reproduced from [31], with permission from Elsevier)

Micronized drug

High drug-carrier contact area

Less drug-carrier contact area

Carrier surface

Spray-dried drug

Carrier surface Carrier surface

Figure 8.4 Schematic representation of drug–carrier interactions of a micronized drug and a spray-dried

drug (SEM images taken from Louey et al (Source: Reproduced from [32] with kind permission from Springer Science and Business Media)

Table 8.1 Comparison between drug carrier, Pulmosphere® and large porous

par-ticle DPI formulations

Formulation type Drug-carrier Pulmospheres® Large porous particles Density (g cm –3 ) 1 ± 0.5 ∼0.4 0.1–0.5

Mean geometric

diameter (μm)

the techniques that are dependent on inertial impaction Indeed, aerosol PSD can be expressed in two

An advantage of the spray-drying technique is the capability to produce large porous particles

prepared by a standard, one-step pharmaceutical spray-drying process using ‘generally recognized

as safe’ (GRAS) excipients [42]

The advantages of LPPs are summarized as follows:

a Increased aerosolization efficiency due to lower powder aggregation owing to a lower contact areabetween larger particles resulting, in less total van der Waals forces Several studies have shown

that LPPs increase the amount of respirable particles, both in vitro [43, 44] and in vivo [3, 45].

b Formulations containing LPPs may confer a longer time for drug delivery to occur, because theycan escape the natural clearance mechanisms present in the airways such as the mucociliary esca-lator (Chapter 1) and phagocytosis by alveolar macrophages [41]

c Formulations containing LPPs of APIs with relatively low water solubility (i.e., relativelylipophilic LPPs) have been produced so as to generate sustained release from inhaled products[42]

Despite these advantages, some disadvantages have been stated in connection with LPP tions For example, LPP formulations can impose a limit on the deliverable dose because they cancarry only a small mass of drug due to the low density (by definition) of such particles, so they provideonly a practical means of delivery for potent and low-dose drugs [46]

formula-Generally, there are only two basic strategies by which aerosol particles can be made In strategy 1,

size requiring a geometric size between 1 and 5 μm In strategy 2, the density of particles can range

Biodegradable microspheres have been produced with a sponge-like appearance, having a mean

prepared using a two-step procedure: (1) preparation of a fluorocarbon-in-water emulsion by addingphosphatidylcholine as a surfactant with dispersal using high-pressure homogenization Then, theemulsion is combined with a second solution consisting of the API and other wall-forming materials(e.g., co-surfactants, sugars, salts, etc.), (2) spray-drying of the resulting aqueous dispersion [30]

Advantages of biodegradable microsphere formulations in DPIs include the following [47]:

a Higher respirable fraction due to the good aerodynamic properties ascribed to the hollow porousparticle design

b Biodegradable, non-immunogenic and non-toxic properties

c The possibility of changing the particle characteristics (e.g., morphology, density and size)

Historically, spray-freeze-drying was introduced in 1994 [48] A solution, which includes the API,

is sprayed into a vessel containing a cryogenic liquid such as nitrogen, oxygen or argon This results

in a quick freezing of the generated droplets, which are then lyophilized to produce porous sphericalparticles suitable for inhalation Despite this technique being successfully applied to produce proteinparticles, it still has a number of disadvantages which include the high cost, long processing time,safety concerns and possible denaturing effects of the proteins (due to the stresses associated withfreezing and drying) [49]

A SCF is a material that can be considered to be either a gas or a liquid It has the gaseous erties in terms of penetrability but the liquid-like properties in ability to dissolve materials Thistechnique has been driven by the need to generate particles with controlled physical properties It canprovide an attractive particle engineering option to enhance aerosol performance in inhalation ther-apy [50] For example, engineered drug particles using SCF technology showed reduced surface freeenergy in comparison to micronized drug particles [51] Manipulation of the operating conditionssuch as temperature, pressure, nozzle flow rates and solution concentrations may enable the accuratecontrol of particle size, shape and morphology [52] The main SCF processes can be classified asfollows [53]:

prop-8.3.4.1 Rapid Expansion of Supercritical Fluid Solution (RESS)

In this method, the API is dissolved or solubilized in a SCF, then the rapid expansion of theSCF through a heated orifice leads to a high degree of supersaturation due to the reduction in theSCF density This change results in a reduction of solvation power, producing a precipitation ofthe drug [54]

8.3.4.2 Gas Antisolvent (GAS) Recrystallization

Here the SCF functions as an antisolvent to cause precipitation within the liquid solution This nique has a number of advantages, such as the ability to control particle size of the resulting particlesand producing void-free crystals [55]

tech-8.3.4.3 Solution Enhanced Dispersion by SCF (SEDS)

The SEDS method employs the same principle of the use of antisolvent in solvent-based tion processes [49, 53] The SCF is mixed rapidly with an organic solution containing the API, inwhich the latter is highly soluble This leads to a large volume expansion and reduction in the solventdensity, and results in a high level of supersaturation [36] This technique has demonstrated a highcapability to control the physical properties of particles [56]

5 μm

Figure 8.5 (a) Salmeterol xinafoate (Source: Reproduced with permission from [52] Copyright © 1999,

American Chemical Society), (b) budesonide (Source: Reproduced from [59], with permission from vier) and (c) lactose engineered particles prepared by SCF technology (Source: Reproduced from [60], with permission from Elsevier)

Else-8.3.4.4 Precipitation from Gas Saturated Solutions (PGSS)

This technique is similar to the RESS technique because the SCF functions as a solvent ratherthan antisolvent in both techniques [49] In this technique, the SCF is dissolved in a molten solutebefore it is subjected to the rapid expansion conditions [57] and it has been used to engineer drugssuch as salmeterol xinafoate [52], salbutamol sulphate [58], budesonide [59] and excipients such aslactose [60]

In DPI systems, SCF technology can provide an attractive particle engineering option to enhanceaerosol performance for inhalation therapy For example, both SCF–salbutamol sulphate [58] andSCF–lactose carrier [59] particles generated an improved pulmonary drug delivery of the API from

a DPI A few examples are shown above in Figure 8.5

Since the drug particles in DPI formulations are very fine, the control of their cohesiveness is a keyfactor in determining DPI performance One of the most effective methods in reducing such cohe-siveness between particles in DPI systems is to present the particles as granules, which are easilydisintegrated into aerosol form during inhalation In order to manufacture such ‘soft’ granules, thePSG technique can be applied [61] This procedure consists of two processes that are continuallyalternated, comprising compaction and granulation–fluidization processes Therefore, the PSG pro-cess can be considered as a cyclic fluidization and compaction process that is conducted by alternatingupward and downward gas flow [62]

The resulting formulated particles have excellent characteristics, rendering them attractive for use

in DPI systems These properties include ideal size distributions, the generation of high fine particlefractions (FPFs), the required granule strength for ease of disintegration into aerosol form duringinhalation, excellent granule dispersibility and the provision of good drug re-dispersion (i.e., gener-ates a high percentage of the emitted dose (ED)) [61]

8.4 Engineered Carrier Particles for Improved Pulmonary Drug Delivery from Dry Powder Inhalers

Maximizing the efficiency of drug aerosolization still provides a major challenge in the ment of DPI formulations A slight change in particle physicochemical properties is likely to have a

develop-considerable effect on drug aerosolization behaviour Thus, the use of suitably engineered particlesmight be an essential factor for improving DPI performance Several techniques have been described

to achieve this outcome including antisolvent crystallization [63], batch cooling crystallization [64]and freeze-drying [65] These methods were used to modify two common carriers, lactose and man-nitol particles, to improve the aerosolization of API from DPI formulations

For example, antisolvent crystallization techniques using binary non-solvents includingethanol–butanol, ethanol–acetone and acetone–butanol produced lactose particles which whenincorporated into DPI formulations of albuterol sulphate lead to an enhanced aerosolizationperformance [66–68] In comparison to commercial lactose, engineered lactose particles were lesselongated and more irregular in shape with rougher surfaces In addition the generated lactosealso contained a higher fine particle content and displayed a higher porosity These particlesappeared as ‘secondary’ particles comprised of smaller ‘primary’ subunits having different sizesand morphologies, according to the type of non-solvent used during crystallization Moreover, the

When the volume ratio of non-solvent used during crystallization was changed, the amount of

to optimize the aerosolization performance of lactose in the DPI formulation containing albuterolsulphate [66]

In a further study, it was shown that not only was the efficiency of aerosol delivery dependent uponthe non-solvent type employed but also upon the degree of saturation of lactose solution used duringcrystallization [63] If the crystallization procedure is controlled, lactose particles may be engineeredwith more predictable aerosolization properties

These findings have been extended from lactose to mannitol, when the latter was identified asproviding a possible improved alternative excipient in DPI formulation The morphology of mannitolparticles was dependent on the manufacturing technique employed (Figure 8.6) The efficiency ofdrug dispersion from powders containing crystallized mannitol was influenced by the proportion ofwater present in the acetone or ethanol non-solvent used in the crystallization procedure [69, 70].However, regardless of the ratio of acetone–water or ethanol–water used in the recrystallizationprocedure, all crystallized mannitol carriers showed better performance than commercial mannitolwhen incorporated in DPI formulations [69, 70]

Solid-state analyses demonstrated that all mannitol samples were crystalline with no detected phous content A change in the ratio of acetone–water or ethanol–water led to samples of mannitolwith a different polymorphic content [69, 70]

amor-Formulations containing mannitol crystals grown in solutions having a lower supersaturation (20%,w/v) produced higher FPF of the API in comparison to that from formulations which incorporatedmannitol crystals grown from high supersaturation (50%, w/v) (i.e., 31.6 ± 2.3% vs 14.2 ± 4.4%).This was attributed to the elongated habit, smoother surface and higher ‘intrinsic’ fines content of theformer formulations

Although freeze-dried mannitol did not produce a powder with a smaller geometric size thancommercial mannitol, a smoother surface morphology resulted (Figure 8.6) The use of freeze-dried mannitol generated the weakest salbutamol sulphate–mannitol adhesive forces within thepowder mix, whereas commercial mannitol generated the highest salbutamol sulphate–mannitoladhesive forces It was shown that the smoother the mannitol surface the weaker the salbutamolsulphate–mannitol adhesive forces [64] However, mannitol-containing products with higherpowder porosity and weaker salbutamol sulphate–mannitol adhesive forces generated a higherFPF of salbutamol sulphate It was concluded that the freeze-drying of aqueous mannitol solutionsprovides an attractive approach to preparing an excipient suitable as an excipient for blending into adry powder aerosol formulation Such a strategy might provide an avenue to generate an enhancedpulmonary drug delivery and maximal yield It is a method that is simple, reasonably cost-effectiveand has a low safety risk, since no organic solvents are used

(a) 50 μm (b) 50 μm

Figure 8.6 SEM photographs for (a) mannitol crystallized from acetone, (b) from ethanol, (c) cooling

crystallized mannitol (CCM) (Source: Reproduced with permission from [64] Copyright © 2012, American Chemical Society) and (d) and freeze-dried mannitol (unpublished SEMs)

In addition to ‘engineered lactose’ and ‘engineered mannitol’, an ‘engineered mannitol–lactose’complex was another approach that has been investigated for a better aerosolization performance [71].Antisolvent crystallization has proved to be successful in preparing engineered mannitol, lactose andmannitol–lactose mixtures with improved aerosolization properties (Figure 8.7)

In comparison to commercial carriers, all crystallized mannitol–lactose particles showed a moreregular shape, a higher fines content and a higher specific surface area (Figure 8.7) Carriers crys-tallized using a higher mannitol–lactose ratio produced particles with a higher elongation ratio, amore irregular shape and a smaller true density Mannitol was altered from a spheroidal to a needle

crystal-lization Crystallized mannitol–lactose mixtures did not generate a markedly better aerosolizationperformance than either engineered mannitol and/or engineered lactose alone However, formulatorscan anticipate that an appropriate particle size and a suitable solid-state and morphology of lactosecarrier can be generated and controlled by the judicious addition of mannitol to the crystallizationmedium containing lactose [71]

Among all carriers investigated by Kaialy et al [64, 65], the lowest FPF of salbutamol sulphate was

generated by commercial mannitol (15.4 ± 1.1%) and cooling crystallized mannitol (14.2 ± 4.4%);whereas freeze-dried mannitol produced the highest FPF (46.9 ± 3.6%)

The presence of water in the crystallization of mannitol from either acetone or ethanol using anon-solvent precipitation technique has a considerable effect on the physical properties of the resul-tant engineered mannitol particles The initial degree of supersaturation used during crystallization

of mannitol using a batch-cooling crystallization approach appeared to be a critical factor in the

Carrier product

Elongation ratioRoughness

11.522.53

mannitol:lactose (05:15)

Figure 8.7 SEM images, ( ⧫) elongation ratio (ER), (○) roughness (mean±SE, n ≥3000) for (a) CM

(com-mercial mannitol), (b) CL (com(com-mercial lactose) and different crystallized mannitol–lactose particles: (c) 20:0, (d) 15:05, (e) 10:10, (f) 05:15 and (g) 0:20 (Source: Reproduced from [71] with kind permission from Springer Science and Business Media)

control of the physicochemical and aerosolization properties of crystallized mannitol For example,improved inhalation performance was obtained from the formulations containing mannitol crystal-lized from lower supersaturations, due to a reduction in the SS–mannitol adhesion forces In addition,DPI formulations containing lactose particles crystallized from solutions having a lower degree ofsaturation demonstrated an improved DPI performance in terms of both better drug content homo-geneity and higher amounts of drug likely to be delivered to the lower airway regions It was alsoshown for mannitol–lactose mixtures that the crystallized mannitol–lactose carrier properties weredependent on the mannitol–lactose ratio used during the crystallization

In general, it can be concluded that the carrier particles (lactose or mannitol) with a higher fines tent, higher specific surface area, smaller mean diameter, higher porosity, higher elongation ratio (up

con-to ∼5) and smaller drug–carrier adhesion properties generated a higher FPF of salbutamol sulphateupon aerosolization The use of particle engineering for carrier particles offers great opportunity forimproving DPI performance through careful manipulation of the carrier physicochemical properties

In addition, the DPI performance improvements can be related to changes in key physicochemicalproperties of the carrier, which can be optimized by controlling the crystallization process factors

8.5 Relationships between Physical Properties of Engineered Particles and Dry Powder Inhaler Performance

Pharmacologically potent drugs usually display poor physicochemical properties, thus formulationdevelopment is often considered challenging [64] In fact, the molecular properties that are respon-sible for pharmacological activity can also be responsible for the compound’s pharmaceutical utilitylimitation [72] Particle–particle (drug–drug, drug–excipient and excipient–excipient) interactionsare critical to the performance of DPI formulations These interactions are mainly dependent on thephysicochemical properties of the interacting particles Any small change in the physical properties

of the particles may result in dramatic changes in aerosolization performance [73] Therefore, todevelop a high-quality DPI delivery system, it is critical to have a full understanding of the physico-chemical properties of drug and excipient particles Assessment of the physicochemical properties atthe molecular, particulate and bulk level is necessary for full characterization of the aerosol formu-lations Failure to study one of these areas may lead to a significant lack of understanding in terms

of particle formation processes, performance prediction, batch-to-batch variations, particle tions and the overall DPI performance The formulation design and physicochemical properties ofthe excipient can significantly affect the respiratory deposition pattern of the inhaled drug–carriermixture It is possible that apparently small changes in particle characteristics will result in unac-ceptable variability in aerosol performance Furthermore, there are often multiple factors in play andthus the control of any individual factor would likely be insufficient for optimizing drug deliveryfrom inhaler devices Nevertheless some of the properties that require consideration in this contextare particle size, powder flow properties, particle shape, particle surface texture, fine particle content(and addition) and particle surface area These factors are considered in turn

Particle size is the most important factor in determining the site of deposition in the respiratory ways [74] and moreover, it also affects both the safety and the efficiency of orally and nasally inhaleddrug products [75, 76] In theory, the larger the particle size the stronger the inter-particulate forces.However, the performance of a fine particle powder is determined not only by the inter-particulateforces, but also by the gravitational forces acting upon such particles Whilst van der Waals forcesare directly proportional to the particle size, the gravitational forces are proportional to the cube ofthe particle size, so as a result, fine particles are highly cohesive and have poor flowability [73].Mass median aerodynamic diameter (MMAD) of the aerosol particles is very important, because it

certain diameter, which is approximately 20 μm) However, the total deposited mass of an inhaledaerosol cannot be predicted by the MMAD and geometric standard deviation (GSD) alone [77].Nevertheless, a linear relationship was found between experimental drug MMAD determined by

PSD polydispersity is important in terms of aerosol quality and efficiency An aerosol PSD is ally described by a log-normal distribution and, therefore, the degree of dispersion is best represented

gener-by the GSD Lower GSD values indicate a narrower PSD Aerosol PSD can be designated as beingeither monodisperse or polydisperse If monodisperse, the size of all the aerosol particles is nearly

characterized by non-uniform size and the aerosol particle sizes significantly differ from each other

GSD values of around 2 The value of an aerosol GSD can affect its aerosolization performance Forexample, if two aerosols with the same MMAD of 2 μm are compared, increasing GSD from 1 to3.5 will result in a reduction in the alveolar deposition from 60% to 30% [80] In general, aerosol

particles with a narrow size distribution are preferred in terms of targeting deposition of particleswithin a specific airway region.

refers to the percentage fraction of the drug that is pharmacologically active and in some studies has

been equated to the ‘respirable fraction’ [82] However, an in vitro measured ‘respirable fraction’ might overestimate the actual in vivo ‘respirable fraction’ [83].

During mixing, the larger the carrier particles, the higher the inertial and frictional press-on on) forces, which have the potential to increase the adhesive forces in the mixture, depending on thecarrier payload [84]

(push-It has been shown that the aerosol dispersion is affected by aerosol particle size [32] Generally,

aerosolization [85–87] For example, in the case of patients with severe airflow obstruction, smallparticle aerosols (1.8 μm) were therapeutically more efficacious than large particle aerosols (4.6 and10.3 μm) [88]

Different inhalers might produce different aerosol PSDs with the same drug resulting in

adrenoreceptors exist in high concentrations in the small airways [89] and small aerosol particles

have shown that in patients with severe respiratory tract airway obstruction, the optimal and the mostsuitable particle size of salbutamol sulphate and ipratropium bromide aerosol is approximately about

particle size and the 3 and 6 μm particles are more potent as bronchodilators in comparison to thesame drug with particle size of 1.5 μm [92]

Although aerosol particle size is the most important factor in determining the amount of drugdeposited in the lungs of a healthy adult, in the case of airway obstruction, the effect of aerosol particlesize on therapeutic effect becomes less evident [93] Generally, the total lung deposition increases

in the case of an airflow obstruction [94] In addition, the tendency of the particles to deposit on thecentral airways increases when the lung airways are narrower [95]

Semi-empirical methods have been developed that correlate aerosol particle size with the tion of drug particles in the lungs, and these can be used as a general guide when assessing the effect

deposi-of particle size on the deposition deposi-of particles in the lungs [96] Aerosol particles smaller than 0.5 μmhave two limitations First, they keep moving by Brownian motion and as a result, they settle veryslowly and may not deposit at all because of their high airborne stability [97] Second, 0.5 μm spheri-cal particles can carry into the lungs only 0.1% of the mass that a 5 μm sphere can carry Accordinglyparticles of API as small as 0.5 μm particles may be considered to be an inefficient means of airwaysdelivery

the delivery of APIs to the airways, irrespective of differences in patient lung function However,the optimal particle size range of aerosol particles is dependent both on the drug being used and theprecise site of action of this drug in the lungs, although the latter is still not usually well defined [99]

In fact, the type of the API and inhaler device used in the DPI formulation blends may have

an effect on the preferred carrier size for optimising aerosolization performance For example,carrier fractions containing smaller size particles have been reported to improve aerosolizationperformance only when employing turbulent shear inhalers (e.g., Diskus®, Rotahaler®, Aerolizer®,Handihaler®, Turbuhaler®, Turbulizer®) rather than inhalers generating inertial forces (e.g.,

Nevertheless, carrier powders are usually sieved so as to obtain the 63–90 μm size fractions beforeblending with the drug [101, 102] Theoretically, particle sieving is ideal for a particle size above

75 μm and is less suitable for particles below 38 μm due to the particle cohesiveness Mechanical

sieving is preferable for non-cohesive powders, whereas for cohesive powders the use of an air-jetsieve is preferred [103] The sieving of particles results in a PSD corresponding to the size of thesmallest square apertures that the particles will pass and, therefore, is related to the particle’s widthrather than its length Accordingly, sieving usually results in particles larger than sieve-hole diam-eters, especially for elongated particles, although extensive sieving is expected to favour particlespassing through their shortest diameter [73] In all cases, it is preferable to use carrier particles withnarrow size distributions, as these particles will contain similar amounts of absorbed material per unitmass After the re-dispersion, the large carrier particles will deposit in the mouth and the oropharynxand are cleared, while the small drug particles will partly penetrate the airways according to a leveldependent upon particle size.

Recently, the effect of lactose particle size on the aerosolization of budesonide from a DPI wasreported Generally, the smaller the lactose VMD the higher the RD, the higher the ED, the smaller theMMAD and the higher the FPF of budesonide, which is indicative of an enhanced DPI performance.However, lactose particles with smaller VMD generated higher amounts of budesonide depositing onthe USP throat, which is likely to be disadvantageous in terms of the increased potential for inducing

local side effects, if the deposition patterns were replicated in vivo The smaller the lactose VMD the

poorer the budesonide content homogeneity within the DPI formulation, which is detrimental to theoverall patient safety of the DPI formulation safety [102] (Figure 8.8)

the percentage change in volume of a constant mass of powder due to tapping CI can be calculated

MMDA GSD

Lactose VMD (μm)

(c) 4 3.5 3 2.5 2 1.5

3 2.8 2.6 2.4 2.2 2

0 50 100 150 (f)

350 300 250 200

Lactose VMD (μm) (b)

(e)

y = –12.571n(x) + 63.69

R 2 = 0.9808 40

30 20 10

0 Lactose VMD (μm)

Figure 8.8 Relationships between lactose VMD and (•, % CV)(a), ( ○, RD), (◊, ED)(b), (▴, MMAD),

(Δ, GSD)(c), ( ⧫, IL)(d), (◽, FPF)(e), and (+) constant K(f) of budesonide obtained from formulations

containing different lactose size fraction powders (mean ± SD, n ≥ 3) (Source: Reproduced from [102],

with kind permission from Elsevier)

powder, the poorer is the powder flowability In addition, CI can be expressed in terms of volume of

forces Powder flowability is considered poor when the CI is greater than 25%, and it is consideredgood when the CI is below 15%

CI has the advantage of low cost and being a prescribed USP method, it is easy to apply However,

CI is not an absolute property of the material, being an empirical technique and having no well-builttheoretical basis Thus, CI measurement might vary according to the precise methodology used

is poured onto a flat surface, it will always have the shape of a conical pile The internal dimensional angle measured between the surface of the stable slope cone and the horizontal surface

three-is𝛼, which is related to the several properties of the powder material such as particle shape, density,

surface area and coefficient of friction [109] The higher the angle of repose, the poorer is the powderflowability However, this method is often less accurate for cohesive powders due to orifice blockage,

so it can be applied accurately only to the powders with low to intermediate cohesive force [110] In

flowing, free flowing, fairly free flowing and cohesive, respectively Poor agreement has been reported

assess powder flowability

In DPIs, powder dispersion is related to two processes in sequence: powder fluidization of a powderbed and powder deaggregation [82] Powders with poor flow properties and excessive cohesive forcesresult in poor dispersion properties upon inhalation because of the enhanced particle aggregation anddecreased particle fluidization [111]

Particle shape is one of the most challenging factors in powder technology to control and define Theuse of different preparation methods can result in different particle shapes for the same material, andeven similar crystallization methods for one material may produce samples with different particleshapes Changes in particle shape can affect particle adhesion For example, particles with irregularshape have been reported as having high adhesion properties, although, the reverse trend has alsobeen reported [112] This apparent disparity was ascribed to the dependence of interaction on therelative position of the interacting particles [112]

All the equations that calculate the inter-particulate forces between particles assume an tion between perfectly spherical particles with smooth surfaces Most pharmaceutical solid particlesdeviate from spherical shape [113] and therefore, any attempt to calculate the inter-particle forcesbetween ‘real’ particles with a ‘real’ shape is destined to contain unrealistic assumptions

interac-In DPI systems, particle shape has a significant effect on the theoretical and the experimental sured) MMAD Thus, particle shape can affect particle deposition profile within the airways Theo-

either irregular or aggregated in the dry state [39] The dynamic shape factor is the ratio of the actualdrag force (resistance force) experienced on this particle to the drag force experienced on anotherparticle that has the same volume but spherical shape [114] Increasing the particle dynamic shapefactor decreases the particle MMAD By definition, ER is the particle length, along the longest axis,over the particle width Particles with a higher ER are more elongated and/or more irregular in shape.The use of needle-shaped (or acicular shaped) crystals within a DPI formulation are associated withdrawbacks such as their inherent poor flow properties and the limitations incurred during crystalliza-tion including filter support-overcrowding and solvent inclusion [115] Nevertheless, needle-shapedcrystals have a greater ability to stay airborne in an airflow compared with isometric particles thathave the same geometric mean diameter [116, 117]

Crystal shape can be changed by recrystallization from different solvents as a consequence of theinteractions between a crystal face and the solvent molecule [118] The growth rate of a crystal face

is related to its attachment energy [119] This attachment energy is believed to be independent oftemperature and supersaturation Polar faces of the crystal absorb polar solvents, while non-polarfaces absorb non-polar solvents [120] By knowing the structure of the intended molecule, it can bepredicted in which axis the crystal will grow For example, it is known that hydrophilic groups indrugs can establish hydrogen bonds with polar hydrogen bonding groups (hydrophilic molecules).Changing the solvent polarity can accordingly lead to a change in crystal morphology For instance,

by increasing solvent polarity, crystal growth by establishing hydrogen bonds with polar groups will

be enhanced, resulting in accelerated growth in one direction [118]

A comprehensive study of the influence of the shape of different carrier particles (mainly lactoseand mannitol) on the aerosolization performance showed that carriers with higher ER produced higheramounts of drug delivered to the lower airway regions However, it was proved that the higher thecarrier ER the poorer the flowability, the higher the amounts of drug loss and the higher the amounts ofdrug depositing in the USP throat, all considered disadvantageous in DPI systems [121] (Figure 8.9)

In conclusion, inter-particle forces are inversely related to the distance between the particles, which

is in turn partly dependent on particle shape Generally, any increase in inter-particle distance caused

by changes in particle shape will reduce the inter-particulate forces resulting in a better drug–carrierdetachment and consequently an improved aerosolization performance

100 95 90 85

(d)

50 40 30 20 10 0

ER

Figure 8.9 Relationships between carrier ER and salbutamol sulphate recovered dose (RD), emitted dose

(ED) (a); emission (EM) (b), fine particle dose (FPD) (c) and fine particle fraction (d)(mean±SD, n = 3) (Source: Reproduced from [121], with kind permission from Elsevier)

8.5.4 Particle Surface Texture

The adhesion of particles is a surface phenomenon and, therefore, the drug–carrier adhesion is foundly affected by the surface morphology of carrier particles and the drug particles [122] Particlesurface morphology also affects powder dispersion [123]

pro-Several methods may be applied in order to increase particle surface smoothness, such as treatingthe drug with a series of saturated fatty acids [116], crystallization from carbopol gels [112] and con-trolled temperature etching [124] In addition the modification of drug particle surfaces can be carriedout by coating the particle surfaces with a suitable excipient [125] Such modified particles usuallyexhibit better flowability and dispersibility, which may be ideal for use in carrier-free DPI formula-tions [126] For example, it has been shown that amino acids such as leucine can be included in DPIformulations using powder mixing and/or spray-drying processes to coat the carrier/drug particles[127, 128] When applied to untreated salbutamol sulphate particles, the new modulated surface-treated particles resulted in lower adhesion forces between the particles [128] Moreover, coating ofdrug particles with leucine or phenylalanine reduced the surface energy of the drug particle [129].Wet-smoothing of particles using a high shear mixer with successive steps of lactose surface wet-ting and drying has also been applied to increase carrier surface smoothness This procedure resulted

in particles with a flattened surface and rounded edges, which increased the aerosol flowability andpacking properties [130]

Generally, the use of particles with a smooth surface topography has been reported to decrease thedrug–carrier median separation energy and consequently increase the aerosol drug FPF followingaerosolization For example, increasing the surface smoothness of lactose carrier particles resulted in

an increase in the flowability and the dispersibility of salbutamol sulphate from the Rotahaler® [86].Such resultant effects were attributed to a reduction in the binding sites with multiple contact points

on the particle surface, leading to an easier detachment of the API during inhalation

On the other hand, conflicting results have been reported since a low FPF of the API was mined when carrier particles with a smooth [131] and a rough [132] surface texture respectively This

deter-could be explained by the existence of an optimum particle rugosity, which provides an optimum in

vitro aerosol particle deposition [133] Only small corrugations (asperities of a few nanometers in

size) were sufficient to induce significant reduction in adhesion forces between the particles andconsequently accomplish an increase in the FPF [134] Thus, the carrier surface irregularities willonly reduce the drug–carrier adhesive force where they reduce the total drug–carrier contact area(Figure 8.10) Such optimum rugosity depends upon the scale of roughness in relation to the size ofthe drug/carrier particles [135] This is discussed in more detail in Chapter 9

When comparing different pharmaceutical particles, it has been shown that the preferred particlesare likely to be curved-surface particles containing small asperities [136]

Drug particles

Carrier surface

Figure 8.10 Schematic representation of drug–carrier contact geometry in the case of carrier particles

with (a) smooth surface, (b) optimal rough surface and (c) extensively rough surface

8.5.5 Fine Particle Additives

Similar to surface asperities, the presence of fine particles on the carrier surfaces can decrease thedrug–carrier contact area and increase the drug–carrier separation distance leading to a reduction inthe adhesion forces and as a result improved DPI inhalation performance [112] This technique is one

of the strategies to improve drug aerosolization in DPI formulations without substantially reducingthe carrier mean size [137] For example, the mixing of fine carrier particles with a coarse carrier pow-der improved the aerodynamic properties of salmeterol xinafoate [138], salbutamol sulphate [139]and beclomethasone dipropionate [140] in the corresponding DPI formulation blends The type ofthese fine particles may be a different material from that which comprises the coarse carrier particles.For example, it has been shown that the addition of fine mannitol (4.3 μm), sorbitol (6.3 μm) [141]

or glucose (4.4 μm) [139] to DPI formulations containing dissimilar coarse carrier also facilitatedDPI aerosolization In addition, ternary components have been included in DPI formulation blendssuch as l-leucine and magnesium stearate [142] Similar to fine particles, ternary components mightenhance drug–carrier detachment by decreasing drug–carrier interaction forces

One mechanism that has been proposed for the increased drug deposition induced by the addition offine carrier particles is that the ‘active sites’ or ‘hot spots’ which are high adhesion sites [143] on thecoarse carrier particles are saturated by the fine particle additives or ternary components Such ‘active’sites on the carrier surface are attributed to the existence of different carrier surface morphologies,which results in different physicochemical properties including different adhesive properties to thecarrier surface These ‘active’ sites are proposed to be on the crystal surface where active moleculargroups are presented to the outside because of the displacement of the groups from the crystal lattice

or as a consequence of misrepresentation of the molecular order This results in areas on the crystalsurface with more potential for surface interaction than on the normal crystal surface The presence

of active sites on the surface of carrier particles decreases the apparent ‘respirable’ drug fraction forthe DPI due to the retention of drug particles on these sites within the carrier When fine particleadditives occupy the active sites on the carrier surface, only the passive or low adhesion sites willremain available for drug adhesion [144] leading to a low drug–carrier adhesion and consequentlyincreasing drug–carrier detachment

An alternative theory has also been proposed for the effect of ternary components based on theagglomeration between the ternary component and the similar-sized drug particles This is expected

to decrease the detachment force from the carrier upon inhalation [145], allowing the fine-sized gates to penetrate more deeply into the airways However, both these mechanisms remain hypotheses.Moreover, it should be remembered that the addition of fine particle additives to the DPI formulationhas some limitations For example, a significant portion of the micronized lactose exists in amorphousform and as a consequence the adding fine lactose particles to the DPI formulation will increase theamorphous content within the powder (Chapter 7) This amorphous form is not the preferred powderstructure for inclusion within DPI formulations because it is unstable [146] and is likely to affect thepowder dispersion and flowability as a function of shelf-life [140]

aggre-The engineering of fine particle additives provides a promising approach to enhance the total ability of DPI formulation For example, commercial carrier powders (both mannitol and lactose)might display good flowability but poor aerosolization performance, whereas, in contrast, crystal-lized carrier powders (both mannitol and lactose), despite exhibiting poor flowability, improvedaerosolization performance [147] Interestingly, the use of 5% w/w engineered elongated fine man-nitol as a ternary additive to lactose–salbutamol sulphate DPI formulations produced formulationsthat demonstrated both satisfactory flow properties and improved aerosolization behaviour

Both surface area and morphology measurements are critical elements of any DPI formulation opment The particle surface area can be considered as one component of the measurement of particlesurface geometry [148] Any surface modification that leads to an increase in the surface area of the

devel-aerosol particle requires careful consideration since any increase could closely be associated with theoverall biological response [149].

Generally, the total aerosol surface area of the API contained within a DPI is very large due to theinherent aerosol particle size that is required to be small so as to achieve a therapeutic effect Howeverthe consequences of this large area, is a tendency to a decreased stability of the formulation, since suchsmall particles are able to take up more moisture and gain charge more readily than coarse particles[150] The high moisture absorption can increase the capillary forces between particles, and this inturn might impede flow In addition, an increase in the surface area of the aerosol particles can increasethe inter-particle forces between the aerosol particles leading to an increase in the aerosol particleaggregation [46] Many techniques can be used to measure aerosol particle surface area includinginverse gas chromatography (IGC), atomic force microscope (AFM) [46] and nitrogen adsorption,the latter method being the most commonly used technique [148]

Corrugated particles have a higher surface area than smoother particles with the same volumediameter and a direct linear relationship has been reported between carrier surface area and carriersurface energy [148] Higher aerosol particle surface area may also be attributable to broader PSDs,higher amounts of fines on the carrier surface or higher degrees of particle surface roughness [151]

8.6 Conclusions

In DPIs, it is the physicochemical properties that combine to be a principal factor that determinesoverall DPI aerosolization performance It is apparent that many physicochemical properties withinthe DPI formulation blends can be altered to achieve improved inhalation delivery of the API How-ever, the effect of one physicochemical property is interdependent upon other properties that areknown to affect drug-carrier interaction upon inhalation Studying the effect of one physicochemicalproperty on DPI performance is virtually impossible since a change in one property concomitantlyalters others Therefore, although a number of studies claim to focus on one parameter, the effect onothers that are varied alongside that particular property are often downplayed or neglected.The benefits of using carrier particles with smaller size or more elongated shape in terms of improv-ing the aerosolization efficiency of an API must be balanced against resultant disadvantages, such

as poor flowability and dose uniformity, possible formulation instability and potential for increasedside effects When angular, spherical or elongated mannitol particles are an option, formulators cananticipate better drug delivery to the lung in the case of the elongated form Controlling or adjustingthe porosity of carrier powder may also provide an optimization strategy in promoting the aerosoliza-tion performance of DPI formulations Better drug content homogeneity was obtained in the case ofcarrier powders with better flow properties and narrower size distributions, although poor aerosoliza-tion performance is not an inherent consequence of flow properties Variations in the particle size

Further insight is required into the relationship between DPI particles physicochemical propertiesand their inhalation performance Designing DPI formulation product with more ‘effective’ physic-ochemical properties is a feasible strategy to achieve maximum performance

References

Berlin; 1–50

characterization of supercritically dried insulin-loaded microparticles for pulmonary delivery

European Journal of Pharmaceutics and Biopharmaceutics, 2008; 68(2), 191–200.

3 Wang, J., Ben-Jebria, A., & Edwards, D A (1999) Inhalation of estradiol for sustained

sys-temic delivery Journal of Aerosol Medicine, 12(1), 27–36.

Drug Delivery Reviews, 19(1), 37–46.

Pub-lishing, New York, 1995, pp 1–564

param-eters on yield and purity Journal of Food Science and Technology, 28(3), 167–170.

monohydrate from aqueous solution Journal of Pharmaceutical Sciences, 90(7), 823–832.

Crystal Research and Technology, 42(6), 583–588.

Cam-bridge, UK

nucleation in industrial crystallizers Journal of Crystal Growth, 237, 2130–2135.

crystalline lactose extracted in ternary and quaternary systems Advanced Powder Technology,

excipients from aqueous solutions and the use of preferred orientation in phase identification

by powder X-ray diffraction European Journal of Pharmaceutical Sciences, 42(5), 568–577.

morphology on the aerosolization performance of micronised salbutamol sulphate

Interna-tional Journal of Pharmaceutics, 415(1), 62–72.

budes-onide for dry powder inhalation (DPI) Industrial and Engineering Chemical Research, 47(23),

9623–9627

drowning-out precipitation operation Chemical Engineering Research and Design, 76(7), 809–814.

thiazole-derivative polymorphs – I: Effect of addition rate and initial concentrations Journal

of Crystal Growth, 257(1), 177–184.

habit of paracetamol crystals in anti-solvent crystallization Journal of Crystal Growth, 279(3),

477–488

abecarnil Journal of Crystal Growth, 198, 1307–1314.

and influence of reactor size AIChE Journal, 50(12), 3107–3119.

ana-lyzer Journal of Food Science, 42(4), 1069–1072.

Oxford

25 Hartel, R W., & Shastry, A V (1991) Sugar crystallization in food products Critical Reviews

in Food Science and Nutrition, 30(1), 49–112.

dose inhalers Pharmaceutical Research, 17(2), 168–174.

leucine: Formulation parameters, aerosol performance and in vitro toxicity on CuFi1 cells.

International Journal of Pharmaceutics, 426(1), 100–107.

performance and stability of spray-dried powders for inhalation Journal of Pharmaceutical

Sciences, 97(1), 287–302.

microparti-cles for controlled release inhalation therapy The AAPS Journal, 11(4), 771–778.

Research, 25(5), 999–1022.

par-ticles with tailored surface properties – The influence of carrier surface roughness and shape

European Journal of Pharmaceutics and Biopharmaceutics, 82(1), 194–204.

narrow size distributions produced by jet-milling and spray-drying techniques Pharmaceutical

Research, 21(7), 1200–1206.

particles for pulmonary delivery International Journal of Pharmaceutics, 278(1), 187–195.

the fine particle fraction of inhalation aerosols Pharmaceutica Acta Helvetiae, 71(3), 185–190.

triboelectrification in dry powder inhalers International Journal of Pharmaceutics, 282(1),

107–114

Drug Delivery Reviews, 26(1), 3–15.

Particles Wiley-Interscience, New York, pp 1–504.

aerosols: A critical evaluation of cascade impactor analysis and laser diffraction technique

International Journal of Pharmaceutics, 249(1), 219–231.

glucocor-ticoid particles prepared by a supercritical fluids process European Journal of Pharmaceutical

Sciences, 21(4), 501–509.

inhaled porous drug particles Journal of Pharmaceutical Sciences, 91(7), 1590–1600.

delivery using large, porous inhaled particles Journal of Applied Physiology (Bethesda, MD.:

1985), 85(2), 379–385.

of large porous particles for inhalation Pharmaceutical Research, 16(11), 1735–1742.

pre-sented at the Respiratory Drug Delivery VII, 2(1), 413.

of Biomedical Engineering, 4(1), 93–107.

45 Ben-Jebria, A., Chen, D., Eskew, M L., et al (1999) Large porous particles for sustained tection from carbachol-induced bronchoconstriction in guinea pigs Pharmaceutical Research,

1209–1227

PulmoSphere formulation in healthy volunteers Chest Journal, 124(1), 360–366.

Small Particles of Biologically Active Molecules, US Patent 6569458.

biophar-maceuticals Advanced Drug Delivery Reviews, 58(9), 1009–1029.

survey The Journal of Supercritical Fluids, 20(3), 179–219.

deliv-ery of budesonide from passive DPI devices Journal of Pharmaceutical Sciences, 94(10),

2276–2288

xinafoate Organic Process Research & Development, 3(5), 370–376.

Pharma-ceutical Science and Technology Today, 2(11), 430–440.

engineer-ing with supercritical fluids in pharmaceutics European Journal of Pharmaceutical Sciences,

dense gas techniques for the production of fine particles AAPS PharmSciTech, 5(2), 32–38.

aqueous solutions: The application of SEDS to lactose as a model substance Pharmaceutical

Research, 15(12), 1835–1843.

gas techniques The Journal of Supercritical Fluids, 37(3), 272–278.

with high inhalation performance independent of inhalation patterns Chemical and

Pharma-ceutical Bulletin, 60(3), 334–340.

particle-based lactose blends in passive dry powder inhalers International Journal of Pharmaceutics,

expression of dry powder inhalant by fluorescence and bioluminescence Journal of Controlled

Release, 134(2), 149–154.

pow-ders with coarse lactose for dry powder inhalation Powder Technology, 131(2), 129–138.

on physical property and in vitro aerosol performance of granules prepared by pressure swing

granulation method Aaps Pharmscitech, 4(4), 506–513.

pre-pare engineered lactose with promising aerosolisation properties: Effect of saturation degree

International Journal of Pharmaceutics, 437(1), 57–69.

crystallization on mannitol physical properties and drug dispersion from dry powder inhalers

Crystal Growth & Design, 12(6), 3006–3017.

65 Kaialy, W., & Nokhodchi, A (2013) Freeze-dried mannitol for superior pulmonary drug

deliv-ery via dry powder inhaler Pharmaceutical Research, 30(2), 458–477.

of recrystallised lactose from binary mixtures of ethanol/butanol for improved drug delivery

from dry powder inhalers The AAPS Journal, 13(1), 30–43.

performance of salbutamol sulfate formulated with lactose crystallized from binary mixtures

of ethanol–acetone Journal of Pharmaceutical Sciences, 100(7), 2665–2684.

morphology of crystallised lactose on delivery of salbutamol sulphate from dry powder inhalers

Colloids and Surfaces B: Biointerfaces, 89, 29–39.

of salbutamol from dry powders containing engineered mannitol as excipient International

Journal of Pharmaceutics, 392(1), 178–188.

alterna-tive carrier to enhance deep lung penetration of salbutamol sulphate from dry powder inhaler

Colloids and Surfaces B: Biointerfaces, 79(2), 345–356.

mannitol–lactose mixture on dry powder inhaler performance Pharmaceutical Research,

McGraw-Hill, New York

Press, Boca Raton, FL

Aerosols in Medicine Elsevier, New York, NY, 117–156.

evalua-tion of cascade impactor profiles of pharmaceutical aerosols, part 1: Background for a statistical

method AAPS PharmSciTech, 8(1), E32–E37.

emis-sion, deaggregation and deposition of dry powders for inhalation Journal of Aerosol Science,

lungs: Effects of ventilatory parameters Pharmaceutical Research, 10(6), 871–878.

and laser diffraction techniques for the determination of the particle size of aerosolised powder

formulations International Journal of Pharmaceutics, 358(1), 75–81.

Care, 45(6), 652–666.

the respiratory tract Journal of Pharmacy and Pharmacology, 33(S1), 52P–52P.

the influence of hydrophobic additives Pharmaceutical Research, 10(1), 1–7.

pharmaceu-tical dry powder aerosols Kona, 16, 7–45.

thera-peutical response of salbutamol particles administered from two metered dose aerosols Die

Pharmazie, 46(1), 41–43.

and bulk properties on drug particle detachment from crystalline lactose carrier particles during

inhalation, as function of carrier payload and mixing time European Journal of Pharmaceutics

and Biopharmaceutics, 56(2), 291–302.

carrier particle size and concentration on in vitro deposition International Journal of

Pharma-ceutics, 154(1), 31–37.

Journal of Biopharmaceutical Science, 3(1/2), 101–105.

dispersion of mannitol powders as aerosols Pharmaceutical Research, 16(7), 1098–1103.

bronchodi-latation with nebulised terbutaline in asthmatic subjects Thorax, 41(5), 364–368.

affecting therapeutic effectiveness of aerosolized medications British Journal of Clinical

Phar-macology, 56(6), 588–599.

anticholin-ergic aerosols in patients with severe airflow obstruction Thorax, 51(10), 977–980.

inhaled bronchodilators European Journal of Respiratory Diseases Supplement, 119, 73–78.

particle size in asthmatic patients using monodisperse aerosols Journal of Applied Physiology,

aerosols on lung distribution and pulmonary function in patients with chronic asthma Thorax,

of the regional deposition of inhaled particles in humans The American Industrial Hygiene

Association Journal, 41(6), 399–409.

particles British Journal of Industrial Medicine, 37(4), 337–362.

aerosols from medical inhalers: Their uses and limitations Journal of Aerosol Medicine, 16(4),

341–377

the Lung Butterworths, London, UK, 44–70.

size relationship and shell thickness of inhalable lactose particles during spray drying Journal

of Pharmaceutical Sciences, 92(4), 900–910.

param-eters for the performance of adhesive mixtures in an air classifier during inhalation Drug

Delivery to the Lungs XIII The Aerosol Society, London

100 Malcolmson, R J., & Embleton, J K (1998) Dry powder formulations for pulmonary delivery

Pharmaceutical Science and Technology Today, 1(9), 394–398.

101 Bell, J., Hartley, P., & Cox, J (1971) Dry powder aerosols I: A new powder inhalation device

Journal of Pharmaceutical Sciences, 60(10), 1559–1564.

102 Kaialy, W., Alhalaweh, A., Velaga, S P., & Nokhodchi, A (2012) Influence of lactose carrier

particle size on the aerosol performance of budesonide from a dry powder inhaler Powder

Technology, 227, 74–85.

103 Flament, M., Leterme, P., & Gayot, A (2004) The influence of carrier roughness on adhesion,

content uniformity and the in vitro deposition of terbutaline sulphate from dry powder inhalers.

International Journal of Pharmaceutics, 275(1), 201–209.

104 Nakate, T., Yoshida, H., Ohike, A., et al (2004) Formulation development of inhalation

pow-ders for FK888 with carrier lactose using spinhaler® and its absorption in healthy volunteers

Journal of Controlled Release, 97(1), 19–29.

105 Murakoshi, H., Saotome, T., Fujii, Y., et al (2005) Effect of physical properties of carrier particles on drug emission from a dry powder inhaler device Journal of Drug Delivery Science

and Technology, 15(3), 223–226.

106 Carr, R L (1965) Evaluating flow properties of solids Chemical Engineering, 72(2), 163–168.

107 Carr, R L (1965) Classifying flow properties of solids Chemical Engineering, 72(3), 69–72.

108 Hausner, H (1967) Friction conditions in a mass of metal powder International Journal of

Powder Metallurgy, 3, 7–13.

109 Ridgway, K., & Rupp, R (1969) The effect of particle shape on powder properties Journal of

Pharmacy and Pharmacology, 21(S1), 30S–39S.

110 Thalberg, K., Lindholm, D., & Axelsson, A (2004) Comparison of different flowability tests

for powders for inhalation Powder Technology, 146(3), 206–213.

111 Begat, P., Morton, D A., Staniforth, J N., & Price, R (2004) The cohesive-adhesive ances in dry powder inhaler formulations I: Direct quantification by atomic force microscopy

bal-Pharmaceutical Research, 21(9), 1591–1597.

112 Ming Zeng, X., Martin, G P., Marriott, C., & Pritchard, J (2001) The use of lactose

recrys-tallised from carbopol gels as a carrier for aerosolised salbutamol sulphate European Journal

of Pharmaceutics and Biopharmaceutics, 51(1), 55–62.

113 Hickey, A J., & Concessio, N M (1997) Descriptors of irregular particle morphology and

powder properties Advanced Drug Delivery Reviews, 26(1), 29–40.

114 Baron, P A., & Willeke, K (2001) Gas and particle motion Aerosol Measurement: Principles,

Techniques, and Applications, 2, 61–82.

115 Taulelle, P., Astier, J P., Hoff, C., et al (2006) Pharmaceutical compound crystallization:

Growth mechanism of needle-like crystals Chemical Engineering and Technology, 29(2),

239–246

116 Fults, K A., Miller, I F., & Hickey, A J (1997) Effect of particle morphology on emitted dose

of fatty acid-treated disodium cromoglycate powder aerosols Pharmaceutical Development

and Technology, 2(1), 67–79.

117 Hickey, A., Fults, K., & Pillai, R (1992) Use of particle morphology to influence the delivery

of drugs from dry powder aerosols Journal of Biopharmaceutical Science, 3(1), 2.

118 Di Martino, P., Censi, R., Malaj, L., et al (2007) Influence of solvent and crystallization method

on the crystal habit of metronidazole Crystal Research and Technology, 42(8), 800–806.

119 Meekes, H., Boerrigter, S X., Hollander, F F., & Bennema, P (2003) Needle crystal

morphol-ogy explained Chemical Engineering and Technolmorphol-ogy, 26(3), 256–261.

120 Berkovitch-Yellin, Z (1985) Toward an ab initio derivation of crystal morphology Journal of

the American Chemical Society, 107(26), 8239–8253.

121 Kaialy, W., Alhalaweh, A., Velaga, S P., & Nokhodchi, A (2011) Effect of carrier particle

shape on dry powder inhaler performance International Journal of Pharmaceutics, 421(1),

12–23

122 Ganderton, D., & Kassem, N (1992) Dry powder inhaler Ganderton, D (Ed.) Advances in

Pharmaceutical Sciences Academic Press, London, pp 165–191.

123 Shekunov, B Y., Feeley, J C., Chow, A H., et al (2003) Aerosolisation behaviour of

micro-nised and supercritically processed powders Journal of Aerosol Science, 34(5), 553–568.

124 El-Sabawi, D., Price, R., Edge, S., & Young, P M (2006) Novel temperature controlled

sur-face dissolution of excipient particles for carrier based dry powder inhaler formulations Drug

Development and Industrial Pharmacy, 32(2), 243–251.

125 Bechtold-Peters, K., Nguyen, H., & Rowley, G (2002) Surface modification of powders with

fatty acid or fatty alcohol derivatives or poloxamer for inhalation PCT International

Applica-tion, 1017, 2002.

126 Eerikäinen, H., Watanabe, W., Kauppinen, E I., & Ahonen, P (2003) Aerosol flow reactor

method for synthesis of drug nanoparticles European Journal of Pharmaceutics and

Biophar-maceutics, 55(3), 357–360.

127 Begat, P., Price, R., Harris, H., et al (2005) The influence of force control agents on the

cohesive-adhesive balance in dry powder inhaler formulations Kona, 23, 109–121.

128 Paajanen, M., Katainen, J., Raula, J., et al (2009) Direct evidence on reduced adhesion of

salbutamol sulphate particles due to L-leucine coating Powder Technology, 192(1), 6–11.

129 Chew, N Y., Shekunov, B Y., Tong, H H., et al (2005) Effect of amino acids on the dispersion

of disodium cromoglycate powders Journal of Pharmaceutical Sciences, 94(10), 2289–2300.

130 Ferrari, F., Cocconi, D., Bettini, R., et al (2004) The surface roughness of lactose particles can

be modulated by wet-smoothing using a high-shear mixer AAPS PharmSciTech, 5(4), 69–74.

131 Chew, N Y., & Chan, H (2001) Use of solid corrugated particles to enhance powder aerosol

performance Pharmaceutical Research, 18(11), 1570–1577.

132 Kawashima, Y., Serigano, T., Hino, T., et al (1998) Effect of surface morphology of carrier lactose on dry powder inhalation property of pranlukast hydrate International Journal of Phar-

maceutics, 172(1), 179–188.

133 Vanderbist, F., Baudier, P., & Maes, P (2006) Dry Power Inhaler Excipient, Process for

its Preparation and Pharmaceutical Compositions Containing it PCT Patent Application

WO/1998/050015

134 Chew, N Y., Tang, P., Chan, H., & Raper, J A (2005) How much particle surface

corruga-tion is sufficient to improve aerosol performance of powders? Pharmaceutical Research, 22(1),

148–152

135 Smyth, H D., & Hickey, A J (2005) Carriers in drug powder delivery American Journal of

Drug Delivery, 3(2), 117–132.

136 Li, Q., Rudolph, V., & Peukert, W (2006) London-van der Waals adhesiveness of rough

parti-cles Powder Technology, 161(3), 248–255.

137 Lucas, P., Anderson, K., & Staniforth, J N (1998) Protein deposition from dry powder inhalers:

Fine particle multiplets as performance modifiers Pharmaceutical Research, 15(4), 562–569.

138 Podczeck, F (1999) The influence of particle size distribution and surface roughness of carrier

particles on the in vitro properties of dry powder inhalations Aerosol Science & Technology,

139 Louey, M D., & Stewart, P J (2002) Particle interactions involved in aerosol dispersion of

ternary interactive mixtures Pharmaceutical Research, 19(10), 1524–1531.

140 Zeng, X M., Pandhal, K H., & Martin, G P (2000) The influence of lactose carrier on the tent homogeneity and dispersibility of beclomethasone dipropionate from dry powder aerosols

con-International Journal of Pharmaceutics, 197(1), 41–52.

141 Tee, S., Marriott, C., Zeng, X., & Martin, G (2000) The use of different sugars as fine and

coarse carriers for aerosolised salbutamol sulphate International Journal of Pharmaceutics,

144 Zeng, X M., Martin, G P., Tee, S., et al (1999) Effects of particle size and adding sequence

of fine lactose on the deposition of salbutamol sulphate from a dry powder formulation

Inter-national Journal of Pharmaceutics, 182(2), 133–144.

145 Louey, M D., Razia, S., & Stewart, P J (2003) Influence of physico-chemical carrier

prop-erties on the in vitro aerosol deposition from interactive mixtures International Journal of

Pharmaceutics, 252(1), 87–98.

146 Saleki-Gerhardt, A., Ahlneck, C., & Zografi, G (1994) Assessment of disorder in crystalline

solids International Journal of Pharmaceutics, 101(3), 237–247.

147 Kaialy, W., & Nokhodchi, A (2013) Engineered mannitol ternary additives improve dispersion

of lactose–salbutamol sulphate dry powder inhalations The AAPS Journal, 15(3), 728–743.

148 Sethuraman, V V., & Hickey, A J (2002) Powder properties and their influence on dry powder

inhaler delivery of an antitubercular drug AAPS PharmSciTech, 3(4), 7–16.

149 Maynard, A D., & Maynard, R L (2002) A derived association between ambient aerosol

surface area and excess mortality using historic time series data Atmospheric Environment,

150 Van Campen, L., Amidon, G., & Zografi, G (1983) Moisture sorption kinetics for

water-soluble substances I: Theoretical considerations of heat transport control Journal of

9 Particle Surface Roughness – Its Characterisation and Impact on Dry

Powder Inhaler Performance

Bernice Mei Jin Tan, Celine Valeria Liew, Lai Wah Chan, and Paul Wan Sia Heng* GEA-NUS Pharmaceutical Processing Research Laboratory, Department of Pharmacy,

National University of Singapore, Singapore

Abbreviations

List of Abbreviations

Fsurface Surface factor

* Corresponding author: Email: phapaulh@nus.edu.sg

Pulmonary Drug Delivery: Advances and Challenges, First Edition Edited by Ali Nokhodchi and Gary P Martin.

© 2015 John Wiley & Sons, Ltd Published 2015 by John Wiley & Sons, Ltd.

9.1 Introduction

Dry powder inhalers (DPIs) are used to deliver drugs to the lung for both local (e.g., the treatment

of pulmonary diseases) and systemic applications The dose of drug delivered per actuation is inthe order of tens or hundreds of micrograms [1]; DPIs have the capacity to deliver the highest doses tothe airways in comparison to other inhalation devices [2] Drug particles deposited at target sites in therespiratory system are referred to as fine or respirable particles An upper particle size limit of 4–6 μm

is optimal for central airway deposition while that of the peripheral airway is 2–4 μm [3, 4] Particlesbelow 0.5 μm can enter the deep alveolar lung regions but they remain airborne and are mainly exhaled[5, 6] Particles larger than 10 μm are unable to reach the lower respiratory tract as they are mainlydeposited in the oral, pharyngeal and trachea-bronchial regions [5, 7, 8] A common measure of the

in vitro aerosolisation performance of a DPI is the fine particle fraction (FPF) [9], defined as the mass

fraction of fine particles relative to the nominal dose contained in the device before aerosolisation To

the amount of drug deposited is highly variable [14] This has driven continuing research effort inthe area of formulation and inhaler design in order to improve the efficiency of this delivery system

As a result of small particle size and high surface energy, fine particles tend to be highly erated and poorly dispersible [9] Cohesion between fine particles is predominantly due to the vander Waals, capillary and electrostatic forces [15] These forces have to be overcome for the disper-sion of drug agglomerates into discrete primary particles during inhalation The dispersion problem

agglom-of DPI formulations consisting agglom-of only the micronised drugs, such as the Intal Spincaps™ [16],Bricanyl® Turbohaler and Pulmicort® Turbohaler [17, 18], are addressed by the improved drugdispersion functionality of specially designed inhaler devices In most other products, inert lactosecarrier particles of 40–200 μm are included in the formulation in a typical carrier to drug mass ratio

of 67.5:1 [5, 19–21] In these formulations, drug particles adhere to the surfaces of carrier particlesvia different forms of particle interaction, resulting in a certain degree of ordered mixing [22, 23].This reduces the extent of drug agglomeration, improves powder flow and aids in more precise dosemetering [24]

The importance of various physicochemical characteristics of the carrier, such as the particle size,shape, surface area, electrostatic charge and crystallinity, on the performance of DPIs has beenreviewed extensively in the literature [25] Fundamentally, the study of particle surface properties

is necessary for an understanding of the physical interactions between any two surfaces (drug andcarrier) in contact [26] Surface properties of the carrier determine to a large extent the behaviour ofdrug and carrier particles during aerosolisation [10, 27] However, it still remains a major challenge

to relate the surface profile to the functional characteristics of carrier particles This is partly due tothe difficulties in developing consistent methods for surface characterisation and in defining surfaceroughness

9.2 What is Surface Roughness?

A surface can be visualised as a horizontal plane with surface irregularities (i.e., peaks and valleys)projecting from the plane It can be represented in two dimensions (2D) by a cross-sectional lineprofile or in three dimensions (3D) by the surface topography To facilitate the numerical character-isation of a surface, the acquired surface topography map has to be first filtered into its constituentroughness attributes, waviness and shape profiles based on different spatial frequencies [28, 29].Low- and medium-frequency components constitute the shape and waviness of the surface, whilehigh-frequency components constitute the surface roughness A variety of filtration techniques for

separating roughness from waviness and shape are proposed in the International Organization forStandardization (ISO), American Society of Mechanical Engineers (ASME) and German Institute

of Standardization (DIN) standards [29] Selection of cut-off frequencies for the roughness profile

is also dependent on the lateral resolution of the stylus or optical-based instrument used Due tothis subjectivity in the choice of measurement and image-processing technique, roughness becomes

a scale-dependent attribute of a surface [30] Hence, any study of drug–carrier interactions ing on surface roughness is valid only when it is accompanied with appropriate measurement andimage-processing criteria [31, 32]

focus-The study of surface metrology has historically been essential in many engineering disciplinesespecially where the surface finish of machine parts determines their performance For such engi-neered surfaces, it has been widely accepted that surface features of different spatial frequenciesaffect their functions in different ways [29] For carrier particles in a DPI, the spatial frequencies

of surface features that most critically affect drug adhesion and detachment have not been clearlydefined However, carrier surface models in pictorial form have been proposed to describe the rough-ness of carriers at the different length scales and magnitudes which affect drug dispersion, most often

in relation to the particle size of the drug [6, 33, 34] Figure 9.1 illustrates the concept of surface ness at the different length scales Generally, the large-scale crevices and discontinuities on carriersurfaces that are wider than the diameters of the micronised drug particles and in which the particlescan be trapped are described as the macroroughness of the surface [35] The entrapment of drug par-ticles usually adversely affects drug dispersion as a greater inspiratory effort would be required todislodge the particles trapped in surface crevices On the other hand, nanoroughness constitutes thenanoscale surface asperities which are much smaller in size than drug particles Nanoscale asperitiesestablish a controlled number of contact points between adhering drug particles and carrier surfacesand determine both the true area of contact and their distance of separation [26] The increase innanoscale roughness usually has a positive effect on drug dispersion

rough-Often, surface roughness is quantitatively represented by surface height parameters that are basedonly on the magnitude of asperity heights but provide no additional information on the length scale

of the asperities under investigation In addition, roughness values (usually in the order of 1000 nm

or less) are smaller in magnitude than drug particles (1–5 μm) and the roughness values betweendifferent carrier batches in the same study typically differ by 200 nm or less It is not immediatelyapparent whether the small differences in experimental roughness values do truly signify a transitionfrom a microscale rough surface to a nanoscale smooth surface, as is frequently proposed in differentcarrier surface models No optimal roughness value for carriers has been defined because the differ-ences in other particle properties such as size, size distribution, shape, crystallinity, flow and packingproperties may also alter carrier performance [6, 19, 36] Nevertheless, these surface models stillprovide useful modes for the classification of carrier surfaces to account for the differences in DPIperformance with respect to carrier surface roughness

Carrier surfaceDrug particle

Figure 9.1 Surface roughness at different length scales

9.3 Measurement of Particle Surface Roughness

9.3.1.1 Particle Size

A major challenge in determining the surface roughness of carrier particles is their relatively smallparticle sizes, which typically fall between 40 and 200 μm Due to the random orientation of parti-cles, when presented on the measurement stage and variability in their shapes, it is laborious to locatesufficiently large number of suitable planar surfaces for roughness measurements and to obtain sta-tistically acceptable mean values Tilted or curved surfaces that are very rough often contain surfacefeatures that exceed the range of vertical distance (in the order of a few μm) which most stylus-basedinstruments can measure The size of sampling areas is thus markedly limited and this hampers accu-rate characterisation of surface features at larger length scales With drug particles at the micron-sizerange, other challenges relating to their high surface curvatures and the inadequate resolution of sur-face details are encountered Traditionally, determination of the roughness of irregular particles of

<20 μm was considered neither useful nor reliable [37], as they existed largely as fragments with a

large number of fracture planes and abrupt variations in particle shape Very few studies have gated the surface roughness of micronised particles in DPI research, with the exception of two studiesconcerned with the study of spherical bovine serum albumin particles [38, 39] The use of preciseinstruments such as the atomic force microscope (AFM) and the confocal laser scanning microscope

investi-is necessary for resolving nanoscale surface features [15, 38, 40] Thinvesti-is investi-is because the surface ties that establish a contact with carrier surfaces are of a much smaller scale than the whole particle.The imaging process is much more sensitive to the geometry and wear of the stylus tip due to thegreater need for precise imaging [41]

asperi-9.3.1.2 Measurement Conditions

Conditions for roughness measurements influence both the length scale and the magnitude of ness values Three major considerations in image capture and analysis for obtaining valid roughnessparameters are the scan size [42, 43], image-processing operations [42] and probe tip geometry forstylus-based instruments [27, 41]

rough-Smaller scan areas allow for greater resolution of surface details and less surface tilt The slope on asurface should not exceed 7∘ for high precision roughness measurements [44] Thus, a large number

of images are required to represent the area of interest [45] In 2D line profiling, a characteristiclength, given by the minimum length that allows for representative roughness measurements, existsfor any given surface [42] A minimum sampling length of at least four to five times the characteristiclength is required for representative measurements As it increases further, confidence in a scale-independent roughness measurement also increases For a given surface, the magnitude of roughnessincreases with the sampling length or area because the inclusion of more prominent surface featuresresults in a greater variation of surface heights [46] In one study, the root mean square roughness

to almost 50% of their respective mean values [44] Prior optimisation of the sampling length andmeasurement variables, such as the depth and rate of image scans, was also shown to be importantfor accurate roughness measurement of lactose particles using the scanning probe microscope [33].For the highest resolution of surface details using stylus-based instruments to occur, probe–surfaceinteraction should take place when the surface is perpendicular to the axis of the probe tip [47] Thiscondition is not met when the geometry of the probe tip is not as sharp as the peaks and valleys onhighly irregular surfaces [42] Lateral forces experienced by the probe tip contribute to distortion of

Table 9.1 Summary of the roughness parameters used for inhalable particles

Roughness parameter

(symbol)

Literature references

Applications and comments

Root mean square

roughness (Rq)

15, 38, 57, 46, 58

Square root of the summation of mean height deviation squares Represents the variability of height deviations and is more sensitive to peaks and valleys The ratio

Ra/Rqrepresents the variability of surface heights and a value>1 indicates a high degree of surface irregularity

Maximum

peak-to-valley height (Rt)

33, 46 Sum of height deviations of the highest peak and lowest

valley Ten-point mean height

(Rz)

33, 46 Sum of mean height deviations of the five highest peaks

and five lowest valleys Grey colour variability

(R0)

59, 60 Variability of grey colour along arbitrary lines across the

surface image Flat surface is essential to avoid the tilting angle shadow effect Represents the degree of surface irregularity and correlates with average height deviations measured by atomic force microscopy

Roundness 19, 61 Defined as the ratio Perimeter2∕ (4×𝜋×Area) Is a combined

measure of geometrical shape and surface roughness

Surface factor (Fsurface) 62, 63 Derived from the particle shape factor and its elongation

ratio More accurate roughness measure for cuboidal particles with different shapes and elongation ratios Perfectly smooth cuboidal particles have a value equal to

1 and lower values indicate rougher surfaces Roughness factor 61 Dimensionless index equal to the ratio of the perimeter of

the particle to the perimeter of an ideal shape (sphere or square) with equal area Valid for comparison only when the image magnification is the same

surface features [48] Issues of probe tip wear or breakage are commonly encountered if the probecomes into contact with the surface during scanning [49] Any surface asperity that is sharper thanthe probe tip cannot be accurately reproduced on the surface topography and a mirrored shape of thetip is produced instead [43] In order for accurate representation of the surface, the sample needs to

be relatively smooth on a scale comparable to the radius of the curvature at the apex of the tip [41]

9.3.1.3 Image-processing Operations

The application of algorithms to remove sample tilt and surface curvature is required of particle face images since they are very rarely perfectly flat Subsequently, filtering of the surface profile iscarried out before surface texture parameters can be derived [50] Filtering involves the suppression

sur-of surface components sur-of selected spatial frequencies so that the surface prsur-ofile at the desired lengthscale is obtained The use of standard roughness parameters alone does not always allow meaningfulcomparison of data obtained from different measuring instruments unless surface profiles of identicalspatial frequencies are extracted [29] Until a functional correlation between particular surface com-ponents and carrier performance can be established, neither the shape, nor waviness nor roughness ofparticles should be omitted when undertaking surface analysis [28] The lack of a proper definition

of surface roughness of carrier particles may have contributed to the reported conflicting trends insurface roughness and DPI performance While the standardisation of filtering techniques is neces-sary, there are also concerns that post-processed images may be a misrepresentation of the actualaverage surfaces [42] Hence, a variety of filtration techniques are still being developed for surfaceswith different characteristics and for increasingly complex applications in surface analysis [29].

9.3.1.4 Selection of Roughness Parameters

Surface topography contains a series of data points with each being defined by its x, y and

z-coordinates relative to a mean surface Line profiles are defined by points along the x- (or y) and z-axis Before the advent of 3D surface texture instruments, such as white light interferometers,

surface analysis was largely based on 2D line profilometry Two-dimensional surface parameters aredenoted by different prefixes that reflect the surface components described [50–54] For example,

respectively At present, it is recognised that the spatial distribution of surface features from 3Dmeasurements may allow more comprehensive surface analysis In addition to height parameters,many spatial and hybrid parameters relevant only to 3D surfaces have been derived [55] All 3Dparameters start with the prefix S and are distinguished by the filtration process for the derivation of

Fractal descriptors, which are derived from 2D pixel measurements of particle images, have alsobeen used when surface profilometers are not available However, they can be unreliable for accuratecharacterisation of micro- and nanoscale roughness due to the insufficient image resolution [37].Table 9.1 summarises the common roughness parameters used for inhalable particles

The use of single height parameters to represent surface roughness imparts simplicity for makingstatistical comparisons but they do not contain any spatial information about the surface [30, 42].The ratios of the distances between carrier surface features to drug particle size may determine thelikelihood of drug entrapment [34] Without specifying the length scale of roughness, numericallylarge height parameters may describe surfaces with vastly different qualities For higher roughness atthe greater length scales, surfaces may contain wide and deep surface crevices On the other hand, tallbut closely spaced asperities result in high nanoscale roughness Hence, a combination of height andspatial parameters is necessary to provide complementary information about the specific qualities of

a surface [44, 53]

9.3.2.1 Surface Profiling by Atomic Force Microscopy

The AFM can record surface topography up to sub-nanometre resolution besides measuringadhesion forces between the probe tip and the surface [31, 42, 64] This allows the direct relation

of the strength of drug adhesion to the shape of individual asperities to be determined [43] Surfaceroughness on length scales with many orders of magnitude (ranging from μm to nm range) can

be characterised using the AFM A sharp probe mounted on a cantilever spring moves across thesurface during the scan The force required to lift the probe off the surface can also be measured[43, 65, 66] Deflection of the probe corresponding to the height of each scan point is monitored

by a laser beam reflected onto a position-sensitive photodiode detector The AFM is often used as

a reference method for roughness measurement due to actual contact of the probe with the surfaceand its superior resolution of images [67]

The ability of the AFM to resolve nanoscale surface features is achieved by having long duration

the optical profiler [68] The AFM also exhibits a limited vertical range of the probe deflection and

surfaces with large height deviations often cannot be imaged at all [59, 69] In one study, the surface

characterised using the AFM [70] In contrast, accurate measurements of similarly sized fractions ofrough, fluid-bed granulated lactose particles could not be obtained In another study, the roughness

of several inhalable grades of lactose particles also could not be accurately measured due to therestricted movement of the probe, particularly at very rough sections of the surfaces [58] Hence,the AFM appears to be ideal for smooth to moderately rough surfaces and not ideal for very roughsurfaces especially when the scan area is increased

9.3.2.2 Non-contact Surface Profiling by Optical Methods

order of μm), especially those which the AFM cannot image due to the restricted probe deflection[58, 71] However, most optical profilers exhibit a limited lateral resolution of about 0.3 μm [52]and are more suitable for evaluating the macroroughness of a surface An interferometer is presenttogether with the objective lens in the optical profiler [52] Light from an optical source is split intotwo coherent beams; the reference beam follows a fixed optical path in the interferometer, while thesample beam is directed to and reflected from the sample [71] When the objective lens is positioned

at the plane of precise focus, maximal constructive interference, resulting from the combination ofthe two beams, is detected by the charge-coupled device camera The plane of focus of every point

on the surface is detected through vertical scanning of the objective lens The suitability of using theoptical profiler for determining the roughness of both micron-sized protein particles and coarse lac-tose carrier particles has been evaluated previously [68] Particles were imaged using three differentinstruments: the scanning electron microscope (SEM), optical profiler and the AFM Visual obser-vations of all surface profiles suggested that there was good agreement among the three methods.The root mean square roughness values determined by the optical profiler were slightly higher butcorrelated well with measurements using the AFM However, it has been reported elsewhere that theoptical profiler is unable to image regions of spherical surfaces if the surface slope exceeds 30∘ [52].Sub-micron surface details are likely to be smoothed out due to the limited image resolution [67].Optical artefacts can also contribute to significant deviations in roughness values for surface featureswhich have extreme gradients The extreme sensitivity of the optical profiler to vibration may alsolead to high variability in roughness measurements [68]

Confocal laser scanning microscopy is another optical technique used to produce high resolution3D images of surfaces [71] A spatial pinhole is present, before the photo-detector to eliminate out-of-focus light, so that only light from the focal plane forms the image Images are constructed frompoint-to-point scans across the surface This technique was used to quantify the surface roughness

of spherical spray-dried mannitol particles with mean particle size of 13–14 μm [40] The lateralresolution was reported to be 120 nm with a minimum detectable roughness of 10 nm Fourier trans-formation of the data was required to differentiate the small scale from the large-scale roughness,which was also a component of particle shape As the roughness values obtained corresponded well tovisual observations of the particles under the SEM, this method was proposed as a simple quantitativetechnique for roughness measurement of spherical particles

Light scattering techniques have been used to obtain the surface fractal dimension of particles as aparameter of surface roughness [39] Particles in suspension are subject to an incident light beam that

is scattered in various reflected angles due to the presence of surface irregularities The intensity ofscattered light and scatter angles are measured by detectors positioned at various angles The surfacefractal dimension is obtained from the fractal region of the scattering curve where surface structurescan be resolved In one study, a Mastersizer S laser diffractometer (Malvern, Worcs, UK) was used

to quantify the roughness of 2-μm spray-dried bovine serum albumin particles [39] Due to theirsmall particle size, the roughness values obtained were not considered as physically correct but weremore useful for rank order of the samples The roughness trends obtained were consistent with visualobservations of the samples under the SEM

9.3.2.3 Surface Imaging by Scanning Electron Microscopy

Scanning electron microscopy has been widely used for qualitative roughness analysis due to thepossibility for high magnification and resolution [52] It is routinely used for investigating particlemorphology, ordered structures in powder mixtures [71] and for obtaining fractal descriptors in quan-titative roughness analysis [60] However, only roughness at the macroscale can be evaluated due tolimitations in the image resolution in comparison with the AFM [60] In a study, lactose particlesurfaces were visibly smooth in SEM images but were significantly rougher, with peaks and valleysvarying over 0.5 μm, when measured using the AFM [58] Even though SEM images appear to have a3D structure, the extraction of 3D information is of limited use for accurate roughness quantificationdue to the sophistication and errors in the mathematical methods employed [64] However, 3D recon-struction software has been used on one occasion to create surface line profiles of 80-μm mannitol

different line profiles over the analysed surfaces Similar to the AFM, it is generally time consuming

to obtain a representative number of images for accurate roughness quantification [35]

In addition to qualitative surface analysis, the SEM has been used to investigate agglomerate ture and the organisation of drug and carrier particles in DPI mixtures In one study, images of purebudesonide powder showed that a high proportion of particles with optical diameters of 1–5 μm pre-sented as agglomerates [72] Even after prolonged mixing with carrier lactose in a Turbula® mixer,drug agglomerates still remained but became more predominantly concentrated within the surfacecrevices of carrier particles This suggested that rougher areas of the surfaces were strong bindingsites for the drug and confirmed the highly cohesive nature of budesonide, which was in agreementwith findings from drug–carrier adhesion experiments [15]

9.3.3.1 Determination of Surface Area

The specific surface area of particles has been used as an indirect measure of surface roughness Theparticle size, shape and porosity are additional factors which also influence the surface area [44].Several studies have reported positive correlation between the surface roughness and specific surfacearea of particles Surface area can be measured using nitrogen adsorption with the single-point [70] ormultiple-point Brunauer–Emmett–Teller methods [60] or by air permeametry [6] Nitrogen adsorp-tion method is preferable to air permeametry as it is more accurate, especially when large particlesare involved [60] Among the different nitrogen adsorption techniques, the static volumetric tech-nique was found to be more sensitive to changes in carrier roughness and showed better correlationwith SEM images and AFM measurements The specific surface area of different commercial grades

of recrystallised, fluid-bed granulated and spray-dried lactose batches was positively correlated with

the emitted dose and the FPF in an in vitro study involving the drug, pranlukast hydrate [73] This

was attributed to the greater drug-carrying capacity of rougher lactose particles However, surfacearea measurements alone cannot distinguish between roughness at different length scales and surfaceimaging is still necessary for meaningful roughness characterisation

9.4 Impact of Surface Roughness on Carrier Performance – Theoretical Considerations

Ordered (or interactive) mixing in a DPI formulation is a two-step process involving the breaking up

of fine drug particle agglomerates and the adhesion of these fines to the surfaces of coarse carrier

particles [74] Strong adhesion forces between the drug and carrier particles create ordered structures

in the mixture [75] and result in a considerable degree of blend homogeneity A combination ofinertial, frictional and shear forces is constantly being exerted on adhering drug particles during mix-ing [76] Inertial forces act to induce the detachment of drug particles from carrier surfaces, whilefrictional and shear forces (also known as press-on forces) increase the adhesion of drug to carriersurfaces The true contact area between an adhering drug particle and the carrier surface has been

Surface asperities on rougher particles may restrict the opportunities for a close contact between twocolliding surfaces and reduce the adhesion of drug to carrier surfaces [69, 78, 79] This is desir-able especially during drug detachment from carrier surfaces However, drug particles located in therecesses between the large asperities are also sheltered from inertial and press-on forces, which areessential for the continual breakdown of drug agglomerates and the redistribution of drug particles toother carrier surfaces [76] This may adversely affect blend homogeneity during mixing The inertialforces increase with mixing speed while the press-on forces increase quadratically with mixing dura-tion [80] Hence, the mixing parameters required for blend homogeneity and for the optimal balance

of inter-particulate forces in a DPI formulation may be influenced by the carrier surface properties.The stability of an ordered mixture is determined by its tendency for segregation Gravity promotessegregation, while inter-particulate adhesion promotes the stability of the mixture [81] An increase

in either the carrier surface roughness [61, 74] or the porosity [82] has been shown to reduce regation This has been attributed to the greater density of high-energy binding sites due to moresurface asperities [83] Rougher carrier surfaces also reduced the tendency for removal of drug byabrasive forces and decreased the likelihood of drug particles rolling off the surfaces during capsule

seg-or blister filling [84] While strong adhesive interactions would be ideal fseg-or the fseg-ormation of a stablehomogenous blend [72], they should also be sufficiently weak for drug to be detached from carriersurfaces during aerosolisation [85]

The surface roughness of carrier particles has been found to influence its drug-carrying capacity, oftenmeasured by the total dose emitted from the inhaler The drug-carrying capacity is affected by theavailability of high-energy binding sites and the strength of drug–carrier adhesion A positive rela-tionship between the emitted dose and the specific surface area of particles with different roughnesswas reported [6, 86] While rougher lactose surfaces had higher drug-carrying capacity and tended toexhibit higher emitted doses [33], drug particles were also more firmly adhered and relatively immo-bile [36, 56, 87] Carrier lactose particles that were smoothed with magnesium stearate exhibited

a reduction in the emitted salbutamol dose from 94% to 87% but a corresponding increase in FPFfrom 15% to 26% [86] In another study, a lower density of surface asperities was found to increasethe drug-carrying capacity and emitted dose but resulted in poorer flow properties in pollen-shapedcarrier particles [88] As the de-agglomeration of drug has to occur for the release of respirable par-ticles, higher emitted doses coupled with lower FPF indicate that more drug particles are likely to bedeposited in the oropharyngeal region This potentially results in more local side effects [89]

Adhesion of drug particles to carrier surfaces is affected by the roughness, interfacial surface energy,electrostatic charges and particle deformation behaviour at the contact points [90] The effects ofincreasing roughness on drug adhesion have been explained by different mechanisms, depending

on the length scale of roughness under investigation An increase in the macroroughness signifiesthe presence of more surface crevices and drug-binding sites As a result of strong drug adhesion, a

consisting of terbutaline and lactose was observed [59] In contrast, increasing the nanoscaleroughness increased the heights of small-scale surface asperities The majority of drug particles arethought to settle on surface asperities when their particle size was larger than the arithmetic meanroughness [91] Surfaces with high nanoscale roughness also increased the distance of separationbetween adhering particles and carrier surfaces This reduced the van der Waals forces of attractionbetween them and allowed easier de-aggregation and dispersion during aerosolisation [36, 92].Attempts have been made to model the effects of roughness on particle adhesion [93, 94] It was

Waals force between a large particle (representing the drug particle) and small particle ing a surface asperity) [95] Beyond a roughness value of 1 μm, the separation between particlesreduced the van der Waals forces to almost zero [96] In most models, however, surface asperitieswere assumed to be regularly shaped and spaced across the surface, but in reality, they are randomlyscattered [97] The effects of capillary and electrostatic forces on the total adhesion have often beenignored [98] The distance of separation between the drug and carrier surfaces depends on not onlythe magnitude but also the variability in asperity heights A surface with widely variable asperityheights may contain several high asperities that prevent a contact between the shorter asperities andthe adhering particle, contributing to the reduced true contact area [99] If the material is of low mod-ulus, deformation of the tall asperities allows shorter asperities to be in contact with the adheringsurface, especially with the application of a small force The extent of asperity deformation is diffi-cult to determine, but assumed to be low, in DPI formulations since the drug–carrier adhesion forcestypically lie in the nano-Newton range [43] This raises the importance of surface topography and itsinfluence on drug adhesion

(represent-With the advent of the AFM, the ability to simultaneously map surface topography and measure theadhesion forces between the probe and particle surface allowed the association of individual surfacefeatures to be correlated with the strength of drug adhesion The variability in measurements of forceshas been studied and found to increase with surface roughness [58] Lactose carriers that exhibit largervariability in adhesion force measurements also tend to produce greater standard deviations of FPF

when in vitro studies are conducted [100] The log-normal force distribution curves and the median

force values have been used to describe the strength of drug–carrier adhesion [57, 101] Differentsurfaces have also been ranked according to the degree of adhesion, based on the relative magnitude

of adhesion forces to specific drug substrates [32, 102]

For a binary mixture of drug and carrier, drug dispersion is effected by the transfer of kineticenergy from the inspiratory air stream to displace drug particles from their adhering localities.Three groups of forces, namely the drag and lift, the shear and friction and the inertial forces, aid

in drug detachment (Figure 9.2) [103] The relative contributions of each group of forces depend onthe carrier surface roughness Drag and lift forces are more apparent when the carrier surfaces arerelatively smooth, because drug particles can slide or roll along the surface before detachment [37].When the length scale of roughness is comparable to or larger than the drug particle diameter, thosesituated in the surface crevices are partially sheltered from drag and lift forces [103] Inertial forces,contributed by momentum transfer when a carrier particle collides with the walls of the inhaler orother particles, become the predominant force for detachment [103] In addition, multiple contactpoints in crevices and mechanical interlocking between drug and carrier asperities are possible whenthe surface is very rough [69]

The effect of carrier surface roughness on FPF may be dependent on the predominant mechanism

of drug dispersion, which is specific to each inhaler design Reduction of carrier roughness

is the foremost strategy to reduce drug–carrier adhesion, even though some research workershave suggested otherwise [44, 76, 96] These findings have been reported to be applicable to a

Direction of motion of carrier particle

Drag and lift forces Shear forces

Inertial forces

Direction of airflow

Figure 9.2 The mechanisms of drug detachment from carrier surface

range of commercially available inhalers, including the Diskhaler®, Spinhaler®, Rotahaler® andPulvinal® [69] One exception is the Miat® Monodose Inhaler, which predominantly utilises inertialimpaction forces for drug dispersion when a turbulent airflow is generated in the device The increase

of carrier roughness also did not affect the FPF adversely [76] This was because the rougher carrierparticles increased the ease of the de-agglomeration of drug and carrier particles [104] A similartrend was observed regarding the effect of surface roughness of drug particles on FPF Formulations

different root mean square roughness were tested in a multi-stage liquid impinger [38] The degree

of impaction among the aerosolised particles was modified using throat models with different

significantly by the degree of impaction

of Fine Excipient

The presence of a fine ternary component, most commonly lactose, has been shown to further improvethe FPF compared to binary mixtures of lactose carrier and micronised drugs such as salbutamol [11,105–107], beclomethasone diproprionate [10, 72] and bovine serum albumin [107] The fine excipi-ents typically ranged between 5 and 10 μm and were either incorporated into the formulation or werepresent as intrinsic fines together with the coarse carriers [11, 108] Two theories have been proposed

to explain the effect of surface fines on drug detachment According to the ‘active site’ hypothesis,fine excipients deposit on the surfaces of coarser particles and saturate the ‘high-energy’ binding sites.These active sites are commonly surface crevices and amorphous regions where drug particles wouldnormally be strongly bound [83, 96, 109] Surface layering decreases the macroscale roughness byfilling of fine excipient into the larger crevices, thereby levelling the surface asperities However,the addition of fine excipient increases the nanoscale roughness as the closely spaced fine particlescontribute to small-scale surface asperities When the mixture of the coarse and fine particles is sub-sequently mixed with the drug, the degree of drug entrapment is reduced, leading to a consequentialimprovement in FPF [56, 105] However, the creation of secondary structured, rough carrier sur-faces might also increase the frictional forces during dispersion and impede the detachment of drugparticles sliding and rolling across the surface under the influence of aerosolisation drag forces [37].According to the ‘active site’ hypothesis, mixing of the coarse and fine excipient prior to addition of

micronised drug is necessary for surface layering However, several studies have reported that the in

vitro performance of the DPI was independent of the mixing order of the coarse carrier, fine excipient

and drug [110] Hence, an alternate theory involving the redistribution and displacement of fine ient layers from carrier surfaces by the micronised drug has been proposed Mixed agglomerates ofdrug and fine excipient are formed, in addition to surface layering [1] Rougher carrier surfaces favourthe formation of mixed agglomerates because of reduced contact area and weakened cohesive forcesbetween the coarse carrier and fine excipient [37] In contrast, surface layering predominates forsmoother carrier surfaces and causes drug particles to become embedded in the macrowaviness ofthese layers Other studies have later established that the FPF was independent of the mixing orderonly when the concentration of drug was above a critical range of 0.9–1.5% w/w [111] This sug-gested that the formation of mixed agglomerates had predominated with increasing concentration

excip-of the drug A surface roughness value excip-of 1.3 μm was proposed for the optimal function excip-of lactosecarriers when fine excipient particles are included in the formulation [37]

9.5 Particle Surface Modification

The modification of surfaces to alter their adhesion properties has been widely applied in the electronics [65], pharmaceutical, semiconductor and toner industries [27] Specific to DPI research,roughness modifications have been attempted on both fine drug and coarse carrier particles Surfacemodification may influence the cohesive interactions between drug particles, as well as their aerody-namic properties and may also modulate drug–carrier adhesion [21] However, a clear relationshipbetween roughness and carrier performance is still lacking as both the smoothing and roughening

micro-of carrier particles have led to increased FPF in various in vitro studies [37, 56, 59, 83, 96] Surface

modification can be accomplished either during the process of particle formation or by various ods to induce physical changes to existing surfaces During surface modification, changes in otherparticle properties, such as the size, size distribution, shape, surface area, crystallinity and surfaceenergy, have occurred [21, 112] It is therefore important to consider these effects in tandem with thechanges in surface properties

Spray drying has been most commonly investigated for the surface modification of fine drug particlessince highly spherical particles with homogenous surfaces can be produced Homogenous surfacesoffer similar adhesion conditions to adhering particles, whereas milling, in contrast, causes particles

to have variable surface affinities as mechanical damage creates amorphous regions on crystallinesurfaces [71] An example of the applicability of spray drying to prepare different particles involvedthe parameters being altered to enable the production of either smooth or corrugated bovine serum

atom-isation rate, inlet and outlet temperatures resulted in rougher particles without significant changes totheir median diameter Rough surfaces were formed due to indentations on the surface when the exter-nal crust of protein collapsed into the protein-depleted interior as drying progressed In another study,the roughness of spray-dried bovine serum albumin particles was quantified using the surface fractal

repre-sented a very rough surface It was found that a small increase in surface roughness, from 2.06 to 2.18,was sufficient to improve the FPF significantly when the formulation was tested in a Rotahaler® Noimprovement in FPF was observed when roughness was further increased from 2.18 to 2.41; it wasthought that rougher particles adhered less to the walls of the inhaler and were easier to disperse.However, rougher surfaces could have contributed to higher drag forces on the particles when theywere airborne and this accounted for the constant FPF despite further increase in surface roughness