1
`
`3M COMPANY 2023
`Mylan Pharmaceuticals Inc. v. 3M Company
`IPR2015-02002
`
`

`
`808
`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`816
`
`820
`821
`
`822
`822
`
`824
`
`.......................................................................................... ..
`5.2.2. HFA systems ........................................................... ..
`5.2.3. Vapor pressure effects in terms of mechanisms of atomization ....... ..
`5.3. Drug concentration and drug substance
`5.4. Surfactants .......................................... ..
`5 .4. 1 . Suspension stability.
`5.4.2. Solution stability ...................................................... ..
`5.5. Inverse micelles and disperse systems .................................. ..
`5 .6. Electrostatics ...................................... ..
`
`.......................................................................................... ..
`5.7. Other propellants ............................................................... ..
`.......................................................................................... ..
`5.8. Microbiological growth ...................................................... ..
`6. Influence of device variables ...... ..
`
` .......................................................................................... ..
`
`
`
`6.1. Valve and container design
`6.2. Actuator design...
`7. Conclusions ............. ..
`References .................................................................................... ..
`
`
`
`.......................................................................................... ..
`.......................................................................................... ..
`
`
`
`1. Introduction
`
`Aerosolized beta agonists and anti-allergic com-
`pounds were first
`formulated as pharmaceutical
`aerosols in 1956 using chlorofluorocarbons (CFCs).
`CFC propellants possessed several desirable charac-
`teristics that prompted their use in propellant-driven
`metered dose inhalers (pMDIs) that
`include non-
`toxicity, inertness, and high vapor pressures. Gener-
`ally, metered dose inhaler formulations containing
`CFCS combine a blend of propellants, excipients, and
`drug substance such that formulation factors and
`device characteristics combine to generate an effi-
`cient spray for delivery to the lungs. However, the
`manufacture of CFC propellants was eliminated after
`the signing of the Montreal Protocol [1]. The phase-
`out of CFCs was in response to concern over
`possible detrimental effects on the ozone layer
`originally raised in the 1970s [2]. Since this original
`hypothesis, stratospheric ozone depletion has been
`demonstrated over the Antarctic [3]. In the United
`States, the Food and Drug Administration has recent-
`ly set standards for
`the use of ozone—depleting
`substances and so—called ‘essential—use’ deterrnina—
`
`tions [4]. Specifically, for propellant-driven metered
`dose inhalers determined not
`to have continued
`
`(1)
`the following points were made:
`essentiality,
`products
`that are no longer marketed, and (2)
`products containing ozone—depleting substances mar-
`keted after January 1, 2005, may be proposed to be
`non-essential,
`(3) a moiety can remain on the
`essential—use list until a non—ozone—depleting product
`is marketed (same route of administration, indication,
`
`and convenience), has sufficient supplies to meet
`patient needs, and has sufficient post-marketing data,
`and (4) a CFC-MDI will not be removed until at
`least
`two non—ozone—depleting alternative products
`are marketed under more than one new drug applica-
`tion.
`
`Accordingly, alternative systems for the delivery
`of inhaled medications have been an active research
`
`focus in recent years. Several different approaches
`have been adopted and include altemative propellant
`systems, propellant—free liquid methods, and dry
`powder inhaler-based devices [5—7]. This article will
`outline research efforts that have focused on the
`reformulation of CFC—based metered dose inhalers
`
`using alternative propellants. Predominantly this
`relates to the use of hydrofluorocarbon (HFC) pro-
`pellants and the accompanying formulation strategies
`that have developed alongside this substitution effort.
`These formulation factors, including device charac-
`teristics, are reviewed with respect to the perforrn—
`ance of MDIS.
`
`2. Metered dose inhaler design
`
`types of devices are used to deliver a
`Several
`metered dose of aerosolized medication to the respi-
`ratory tract. However, pMDIs are specifically recog-
`nized as those devices that incorporate a propellant,
`under pressure,
`to generate a metered dose of an
`aerosol through an atomization nozzle. MDIS are the
`most widely used respiratory drug delivery device,
`with an estimated 800 million units produced in 2000
`
`2
`
`

`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`809
`
`‘T Canister
`
`
`
`Spacer Chamber
`
`Fig. 1. Basic components of a pMDI system.
`
`
`
`Actuator
`
`Atomization
`orifice
`
`Drug Substance (suspension
`or solution)
`Liquified propellants /
`excipients
`
`Metering Valve
`
`[8]. MDIs consist of several components (Fig. 1): the
`active substance formulated with propellant and
`excipients, a container, a metering valve crimped
`onto the container, an actuator that connects the
`metering valve to an atomization nozzle, and a
`mouth piece. Additionally, holding chambers or
`spacers may also form part of the delivery system by
`connection to the actuator mouthpiece. A metered
`volume (typically between 20 and 100 pl) of the
`drug/ eXcipient/ propellant blend is expelled from the
`canister via the valve and quickly passes through the
`actuator orifice where primary atomization occurs.
`Characterization of aerosol clouds emitted from
`
`pMDIs is difficult due to the dynamics of the
`atomization process [9]. Individual droplets and the
`surrounding environment of the droplets change
`rapidly in terms of size, velocity, and position after
`being emitted from the actuator orifice. The inter-
`action of this dynamic aerosol plume with the
`geometry of the mouth and airways determines the
`extent of oral and lung deposition. Mechanisms of
`deposition include inertial impaction, sedimentation,
`diffusion, interception, and electrostatic precipitation
`[10]. The relative importance of each deposition
`mechanism on an individual droplet therefore de-
`
`pends on a multitude of factors such as droplet size,
`velocity, evaporation rates, and the anatomical fea-
`tures of the patient’s airways.
`
`3. Measures of the performance of MDIs
`
`3.1. Performance of current pllfll devices
`
`from CFC to HFC
`The transition of pMDIs
`systems has provided an opportunity for the pharma-
`ceutical industry to re-evaluate inhaler performance.
`From one point of view, the task might be seen as
`showing equivalence between the CFC systems and
`their replacements. This approach to the transition is
`less problematic in terms of regulatory approval and
`is therefore more cost effective in the short term.
`
`However, recent phenomenological reports compar-
`ing CFC systems to HFA systems suggest that a
`‘detuning’ of the HFA formulations is necessary to
`attain regulatory equivalence between the systems
`[11]. Thus,
`the opportunity has arisen to improve
`inhaler performance using a systematic approach of
`evaluating the effect of formulation and device
`factors on pMDI performance.
`
`3
`
`

`
`810
`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`Inhaled therapy for treatment of asthma is targeted
`such that
`the active drug substance is delivered
`topically. In terms of bronchodilators this has the
`primary advantage of rapid onset of action, but also
`minimizes the risk of unwanted systemic side—ef—
`fects, which is particularly important
`in corticos-
`teriod therapy. However, simply targeting the lung as
`an organ may not result in optimal therapy, as the
`target receptors for a specific drug may be located in
`particular regions of the lung [12]. Furthermore,
`none of the available anti-asthma drugs are metabo-
`lized in the lungs
`[13]. Consequently,
`the drug
`delivered to the lung will eventually reach the
`systemic circulation and contribute to systemic ac-
`tivity. Thus, the effectiveness of inhaled therapy for
`topical diseases such as asthma depends on the
`device to deliver the correct dose of active drug
`substance to the site of action with minimal deposi-
`tion to other regions that may contribute to unwanted
`side—effects. As a result, measurement and assess-
`ment of inhaled drug delivery is very complex due to
`the multitude of variables that contribute to varia-
`
`tions in delivery such as drug formulation, delivery
`device, administration skill, breathing pattern, and
`lung pathology/ anatomy [14]. Also,
`the optimal
`outcome may not be defined identically between
`scientists, clinicians, patients, regulatory agencies,
`and those paying for the cost of treatment [14].
`
`3.2. In vivo determination of pMDI performance
`
`Measures of respiratory drug delivery performance
`include pharrnacokinetic/pharmacodynamic investi-
`gations,
`in vivo measures,
`in vitro tests, and also
`mathematical models of deposition. All measures are
`important
`ir1 optimizing the development of drug
`formulations and delivery devices for the appropriate
`clinical outcome. Pharmacokinetic/pharrnacodynam-
`ic studies involve measuring plasma drug/physiolog-
`ical marker concentrations and correlating them with
`clinical efficacy and toxicity. This approach allows
`elucidation of the relationship between drug delivery
`and efficacy/toxicity, but may be complicated by
`assay limits,
`local metabolism and gastrointestinal
`absorption. Used in conjunction with in vitro and
`scintigraphic
`studies,
`specific
`characteristics of
`formulation and device development can be related
`to the clinical outcome. Scintigraphic studies allow
`
`visualization of regional lung deposition by incorpo-
`ration of a garnma-radiating nuclide into the formula-
`tion [15]. The site and quantitative amount of
`deposition can be calculated, making this technique
`integral to device and formulation development as
`well as being useful in drug targeting studies [15].
`Used alone, however, this technique does not mea-
`sure clinical efficacy or
`toxicity. A number of
`clinical outcomes are also essential for evaluation of
`
`inhaled therapy. Spirometry, peak expiratory flow
`(PEF), bronchoprovocation testing, measurement of
`inflammation markers, toxicity, quality of life mea-
`sures, and epidemiological studies have all been
`widely used to evaluate the performance of forrnula—
`tions and devices
`in asthma therapy [14]. The
`performance of formulations and devices for bron-
`chodilator delivery to the lung can be assessed using
`changes in lung function (such as spirometry and
`PEF). However, there are no such immediate mea-
`sures to determine the performance of devices con-
`taining anti-inflammatory agents for the prevention
`of asthma. In these cases, the clinical response to
`regional lung delivery may be assessed using tissue
`biopsy and bronchoalveolar lavage samples or in-
`direct measures that
`include urinary and serum
`cortisol levels [16,17].
`
`3.3. Models and mechanisms of deposition within
`the airways
`
`The importance of in vitro measurements in pMDI
`development and optimization can be appreciated
`with an understanding of the mechanisms of particle
`deposition in the lungs. As already mentioned,
`inertial impaction, sedimentation, diffusion, intercep-
`tion, and electrostatic precipitation are the significant
`mechanisms that dictate where a droplet or particle
`will deposit in the airways. Inertial impaction in the
`lung occurs when particles of sufficient momentum
`(a product of mass and velocity) are unable to follow
`the curved streamlines of air within the airways
`during inhalation due to significant centrifiigal forces
`[18]. Sedimentation of particles within the airways is
`related to particle mass and residence times [19].
`Deposition via diffusion is often a small but some-
`times significant mechanism, particularly when par-
`ticle size is sufficiently small [10]. Significant depo-
`sition via interception occurs when the dimensions of
`
`4
`
`

`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`811
`
`the anatomic spaces of the airways become compar-
`able to those of the particle [10]. Finally, electro-
`static precipitation may sometimes occur when
`charged particles, typically charged during atomiza-
`tion, are electrostatically attracted by a charge of
`opposite sign. Such charges may already be present
`on the walls of the device or airways or may be
`induced by the charged particle itself [10]. Data
`indicate that small particles (<1 pm) are most
`influenced, but the overall significance is unknown
`for pMDIs [18]. The extent and location of deposi-
`tion of pMDI aerosols can be modeled empirically or
`by using mechanistically based approaches
`[19].
`Despite being over-simplifications of aerosol deposi-
`tion, these models are very useful in demonstrating
`the effect of certain variables influencing deposition
`in the lung. Although there are a multitude of models
`available in the literat11re [l9,20], each predicts that
`particle diameter and airway diameter have the
`potential
`to influence particle deposition the most
`[19]. Thus, given that particle diameter in under
`more control
`in terms of formulation and device
`
`selection, it is generally accepted that particle size is
`the single most important parameter in pharmaceu-
`tical aerosol delivery.
`
`3.4. In vitro measures ofpMDI performance
`
`Given the complexity of the interactions of aerosol
`clouds emitted by pMDIs and the airways,
`it
`is
`difficult
`to define simple measures of device and
`formulation performance. However, in vitro methods
`have been developed that measure particle size and
`emitted dose characteristics of aerosol drug delivery
`devices. Both theoretical calculations derived from
`models such as those mentioned above and ex-
`
`perimental deposition studies using stable monodis—
`perse aerosols suggest particle size data can be used
`as an estimate of aerosol deposition efficiency [10].
`These predictions indicate that particles greater than
`around 6 pm will be deposited in oropharyngeal
`regions and will not enter the lung [10], while
`particles that traverse the pharynx and upper airways
`are generally less than 6 pm [21]. In addition to
`estimation of deposition characteristics, particle size
`measurements provide simple measures of quality
`control
`for pMDIs and enable a comparison of
`devices and formulations [22—25]. The most com-
`
`mon measure used is the mass median aerodynamic
`diameter (MMAD), which is a statistical measure of
`the aerodynamic size of the aerosol and represents
`the diameter that divides the particle size distribution
`into two halves with respect to mass (i.e. 50% of the
`mass lies in particles above and below the MMAD).
`MMAD is most often measured using inertial impac-
`tion particle size techniques such as cascade impac-
`tion or via multistage liquid impingement analytical
`devices. These methods have been adopted as phar-
`macopoeial standards for various devices, including
`the pMDI [26,27]. Various systems exist for these
`types of particle size techniques, which often yield a
`different quantity and quality of information. Conse-
`quently, data must be carefully analyzed for mean-
`ingful interpretations to be made [28]. In addition to
`MMAD, other parameters obtained from these par-
`ticle sizing techniques that may assist in prediction
`and assessment of
`lung deposition include the
`geometric standard deviation (GSD),
`fine particle
`fraction (FPF), and others
`such as non—balistic
`fraction (NBF) [29].
`Other in vitro methods have been employed to
`evaluate pMDI performance. These include spray
`pattern and geometry measurements
`[30], plume
`velocity calculation [31], spray temperature determi-
`nations [32], and spray force measurements [32].
`
`3.5. Relevance of in vitro and in vivo measures of
`performance
`
`Analyses of lung deposition of particles from a
`wide variety of sources have confirmed that
`the
`extent of particle deposition within the respiratory
`tract
`is related to aerodynamic particle size [33].
`However, there is also evidence to suggest that fine
`particle fractions and MMADs alone do not provide
`sufficient information to predict the respirability of
`aerosols [29,34]. With pMDIs, significant deviation
`of in vitro results from in vivo experiments appears
`to result from the unrealistic inlet tubes or throats
`
`that are used in most cascade impaction and liquid
`impinger systems [34]. It has been suggested that in
`vitro and in vivo correlations can be improved by
`using anatomic throats [35], various breathing pat-
`terns [14], and also by incorporation of other factors
`such as non—ballistic fraction and GSD into the
`
`model rather than MMADs alone [29]. Depending on
`
`5
`
`

`
`812
`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`the goal of the particle size testing these correlations
`may be relevant. Particle size measurements are
`important for quality control, comparisons of forrnu—
`lations and devices, and can also be used to indicate
`relative amounts of lung deposition. They cannot be
`used, however,
`to demonstrate clinical efficacy.
`Ultimately pharmacodynamic and clinical studies
`will be required to demonstrate equivalence or
`efficacy and safety of reformulated products with
`dosage regimen changes.
`
`3.6. Strategies for pMDI optimization and
`improving performance
`
`Unfortunately, a great deal of device design and
`formulation development of pMDIs has been per-
`formed empirically. A satisfactory clinical response
`matched with low apparent systemic adverse effects
`have been interpreted as successful delivery markers
`for anti-asthma medications delivered via the inhala-
`
`tion route. As a consequence, pMDIs are not opti-
`mized systems. Often—cited disadvantages of this
`type of device are the low efficiency of lung
`deposition (often only 20% of the emitted dose
`reaches the lungs) [36] and poor inhaler technique
`[37]. These outcomes are inter-related and must both
`be addressed to improve the performance of pMDI
`device/ formulation combinations. What is required is
`a systematic evaluation of the influence of formula-
`tion factors and device parameters on clinical out-
`comes. However, direct in vivo evaluations of the
`effect of these formulation and device factors would
`
`involve large-scale and costly clinical experiments.
`The multitude of formulation variables and arrays of
`different device parameters would mean that even
`efliciently designed statistical experiments would
`involve massive matrices of controlled experiments.
`Thus more investigations are required that bridge
`different levels of testing such as in vitro measure-
`ments and in vivo outcomes, and in vivo assessments
`and clinical outcomes.
`
`This has been highlighted in recent phenomeno-
`logical studies relating clinical effect of inhaled
`corticosteroids to formulation factors [38]. Reforrnu—
`lation of beclomethasone diproprionate as a solution
`formulation in HFA propellants has resulted in
`greater peripheral lung deposition of the medication
`relative to the CFC equivalents that were formulated
`
`lung
`as suspensions [38]. The greater peripheral
`deposition and reduced oropharyngeal deposition,
`however, does not correspond to an equivalent
`increase in therapeutic effect [39]. The HFC forrnu—
`lation resulted in approximately six times more
`peripheral drug deposition than the CFC product, but
`only twice the therapeutic effect, raising the issue of
`required particle size and target region of the lung
`for optimal inhaled corticosteriod therapy. The par-
`ticle size that maximizes the therapeutic ratio of a
`molecule is likely to be different for a beta-adren-
`ergic agonist than for an inhaled corticosteriod. A
`greater understanding of this relationship will be
`required if we are to achieve improved drug targeting
`with future inhalers [40].
`Although particle diameter and efficiency of dis-
`persion are the primary variables under the control of
`the forrnulator of pMDIs for directing particle depo-
`sition in the airways, inspiratory flow rate is also an
`important factor to consider [41]. Inspiratory flow
`rate can influence the dose emitted firom an inhaler,
`amount
`inhaled, oropharyngeal deposition, and re-
`gional lung deposition of inhaled medications [41].
`Future designs of pMDIs should account for this and
`integrate systems that emit an aerosol at the correct
`time and flow rate. Some systems already incorpo-
`rate breath monitoring and breath-activated aerosols
`[42,43]. In addition, other patient-related factors such
`as inhalation techniques, compliance, and misuse
`need to be addressed. With current pMDI systems,
`effective use is strongly technique dependent [44,45].
`Development of a more ‘forgiving’ formulation and
`device may help to reduce the strong dependence of
`delivery success on patient-related factors
`[38]
`(Table 1).
`
`4. Physicochemical characteristics of HFA
`propellants
`
`Initial screening of alternative propellants iden-
`tified HFA propellants as
`likely candidates
`for
`replacing CFCs. They appeared to have the neces-
`sary physical properties: do not deplete ozone, non-
`flarnmable, sufficient vapor pressures, and,
`impor-
`tantly, they appeared to be as non—toxic as the CFC
`counterparts
`[46]. Although toxicological
`studies
`demonstrated the equivalency of HFAs to CFCs [47],
`
`6
`
`

`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`813
`
`Table 1
`In vivo and m vitro endpoints of pMDI performance
`
`Measure
`type
`
`In vivo
`
`In vitro
`
`Endpoint
`
`Methods
`
`Pharmacokir1etic/
`pharrnacodynarnic
`investigations
`
`Scintigraphic
`studies
`
`Clinical
`measurements
`
`Particle size
`measurement
`
`Spray
`characterization
`
`Formulation
`performance
`
`Plasma drug concentrations
`Physiological markers
`Toxicity measures
`
`Gamma scintigraphy
`Single photon emission computed
`tomography (SPECT)
`Positron emission tomography (PET)
`
`Spirometry and peak expiratory
`flow (PEF)
`Bronchoprovocation testing
`Inflammation marker measurement
`Tissue biopsy
`Bronchoalveolar lavage
`Toxicity
`Quality of life measures
`Epidemiological studies
`
`Inertial methods
`Optical methods
`
`Spray force and velocity
`Plume temperature
`Spray patterns and geometry
`
`Suspension/solution stability
`Emitted dose
`
`Refs.
`
`[17,l29,130]
`
`[131]
`
`[132]
`[133]
`
`[134,135]
`[136,137]
`[138]
`[139]
`[140]
`[141]
`[142]
`[143]
`
`[144,145]
`[31,58]
`
`[32]
`[31]
`[30]
`
`[59,84]
`[86]
`
`it was quickly realized that HFA propellants were
`not ‘drop in’ replacements for CFCS in pMDIs [48].
`While CFCs were good solvents for a number of
`drug candidates [46], HFA propellants are generally
`poor solvents for many anti-asthma drugs and excipi-
`ents currently available for use in pMDIs [5]. These
`
`changes in the physicochemical characteristics of the
`propellant systems impact pMDI design and formula-
`-
`-
`-
`tion. A summary of_ the_phys1cochem1cal nature of
`HFA propellants, primarily 1,1,1,2-tetrafluoroethane
`(HFA 134a), is presented here.
`The structures of common propellants are shown
`in Fig. 2. The physical properties of these propellants
`and other alternatives have been widely reported in
`recent years [5,46,48—5l]. The physical properties of
`each propellant (Table 2) can be related to details of
`the chemical structure and experimental
`investiga-
`tions of propellant physicochemical behavior.
`HFA 134a and 227ea both have high vapor
`pressures and low boiling points. These properties
`
`CI
`
`C'—C—F
`(Ll
`CFC 11
`
`F
`
`F
`\
`CH -CZF
`2
`l
`F
`
`HFA 134a
`
`F
`
`0'
`
`F
`C
`I
`CI
`CFC 12
`
`F
`
`F
`
`F—C—C—F
`('3'
`[I
`CFC 114
`
`F
`
`F
`\C/F
`F\C/
`\
`/ HC/ \
`F
`
`F
`
`F
`HFA 227ea
`
`Fig. 2. Chemical structures of common CFC and HFA pro-
`pellants.
`
`are caused by the strong electronegative repulsive
`interactions between HFA molecules rather
`than
`
`from non-polar characteristics [5]. In fact, HFAS are
`relatively polar. HFA 134a contains two electroposi—
`
`7
`
`

`
`814
`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`Table 2
`Physicochemical properties of pMDI propellants
`
`Property
`
`CFC ll
`
`CFC 12
`
`CFC 114
`
`HFA 134a
`
`HFA 227ea
`
`Thermodynamic
`Boiling point (°C)
`Vapor pressure (kPa)
`Enthalpy vap. (kJ/mol)
`
`Polarity
`Dielectric constant
`Dipole moment
`Induced polarization
`(m3 mol’2X105)
`Solubility parameter
`(Hildabrand units)
`Kauri—Butanol value
`Log P (oct/water)
`Water solubility (ppm)
`
`Liquid phase
`Density (g/cms)
`Viscosity (mPa~s)
`Surface tension (mN/m2)
`
`24
`89
`25.1
`
`2.3
`0.45
`
`2.8
`
`7.5
`60
`2.0
`130
`
`1.49
`0.43
`18
`
`-30
`566
`17.2
`
`2.1
`0.51
`
`2.3
`
`6.1
`18
`2.2
`120
`
`1.33
`0.20
`9
`
`4
`182
`22.1
`
`2.2
`0.58
`
`3.2
`
`6.4
`12
`2.8
`110
`
`1.47
`0.30
`11
`
`-26
`572
`18.6
`
`9.5
`2.1
`
`6.1
`
`6.6
`9
`1.1
`2200
`
`1.23
`0.21
`8
`
`-16
`390
`19.6
`
`4.1
`1.2
`
`6.1
`
`6.2
`13
`2.1
`610
`
`1.42
`0.27
`7
`
`tive protons due to the strongly electron—withdrawing
`effects of the multiple fluorine atoms attached to the
`carbon backbone [5]. Similarly, HFA 227ea has one
`asymmetrical electropositive proton. This polar na-
`ture, relative to CFCs,
`is reflected in the dipole
`moments
`and
`dielectric
`constants.
`The
`
`polarizabilities of HFAs relative to the CFC pro-
`pellants are much lower. This illustrates the strong
`electronegative nature of the fluorine atoms and the
`strength with which associated electrons are held [5].
`An increase in polarizability is associated with
`increased intermolecular attraction. As a conse-
`
`quence, HFAs are relatively polar but have very low
`intermolecular attractive forces when compared with
`CFCs systems.
`impact of these differences
`The most apparent
`between HFAs and CFCs is insufficient capacity to
`solubilize traditional surfactants [48,49]. Thus, it is
`surprising that relatively few published studies have
`investigated the solubilization capacity of HFAs
`[49,50,52,53]. Using co-solvents such as ethanol or
`other solubilizing agents,
`the solubility of surfac-
`tants, drug substances, and excipients in the HFA
`propellants can be increased [5,53,54]. However, if
`surfactants are required in a co—solvent—free system
`(for stability of a suspension formulation), alternative
`surfactants need to be used [5,55].
`
`Many pMDl formulations contain multi-compo-
`nent mixtures of propellant blends and co-solvents.
`Therefore,
`it is important to consider the physico-
`chemical characteristics of these systems in order to
`understand the effects of formulation variables on
`
`product performance. Density, molar volume, and
`vapor pressure are thermodynamic parameters that
`assess intermolecular forces within mixtures and are
`
`readily measurable in pMDls [56]. These can be
`evaluated by comparing theoretical calculations of
`ideal mixtures with the experimental measurements
`of real solutions. Propellant systems that have been
`studied with regard to propellant-driven pMDls
`include HFA l34a/ ethanol, HFA 227ea/ ethanol, and
`HFA 134a/HFA 227ea mixtures [5,53,56]. These
`three miscible components may allow the forrnulator
`to select appropriate densities (for suspension stabili-
`ty), solubility characteristics (for solution and sus-
`pension formulations), and also modify the emitted
`particle size via non-volatile composition effects
`[5,53]. Tzou [56]
`recently demonstrated that
`the
`observed densities of HFA l34a/ ethanol and HFA
`
`134a/HFA 227ea mixtures closely matched ideal
`mixture predictions. Vervaet and Byron [5] reported
`similar result for HFA l34a/ ethanol, HFA 227ea/
`ethanol, and HFA 134a/HFA 227ea mixtures. How-
`ever, when vapor pressure behavior was investigated,
`
`8
`
`

`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`815
`
`positive deviations from Raoult’s law were observed
`with HFA 134a/ethanol and HFA 227ea/ethanol
`
`mixtures [5,53,56]. Blends of HFA 134a and HFA
`227ea did not show any significant deviation from
`theory [56,57]. Vervaet and Byron explain the appar-
`ent atypical observation of deviations from Raoult’s
`law with no accompanied changes in density as a
`surface phenomenon [5]. HFA propellants have a
`higher affinity for
`the gas—liquid interface than
`ethanol, which,
`in turn,
`is surrounded by HFA
`molecules.
`In addition, positive deviations
`from
`Raoult’s law with HFA/ ethanol mixtures indicate
`
`that the intermolecular forces between the compo-
`nents of the mixture (i.e. between ethanol and HFA
`molecules) are less than that between molecules of
`the pure constituents [53,56]. It was suggested that
`this positive deviation in vapor pressure will allow
`forrnulators to use higher concentrations of ethanol
`(for improved solubility) than would be predicted by
`ideality without detrimental effects on droplet size or
`aerosolization [5]. However, recent evidence sug-
`
`Table 3
`Common marketed pMDIs and their general composition
`
`Therapeutic
`group
`
`Bronchodilators
`Maxiair
`Maxair Autohalor
`Proventil
`Proventil HFA
`Tomalate
`
`Ventolin
`Ventolin HFA
`
`Corticosteriods
`Aerobid
`
`Drug
`
`Pirbuterol acetate
`Pirbuterol acetate
`Albuterol sulfate
`Albuterol sulfate
`Bitolterol mesylate
`
`Albuterol sulfate
`Albuterol sulfate
`
`Flunisolide
`
`Azmacort
`
`Triamcinolone acetonide
`
`Beclovent
`Becotide 100
`Flovent 44, 110, 220
`QVAR 50, 100
`QVAR Autohaler 50, 100
`Vanceril
`
`Beclomethasone dipropionate
`Beclomethasone dipropionate
`Fluticasone propionate
`Beclomethasone dipropionate
`Beclomethasone dipropionate
`Beclomethasone dipropionate-
`trichlorofluoromethane clathrate
`
`Other anti-inflammatory
`Intal
`Tilade
`
`Cromolyn sodium
`Nedocromil sodium
`
`gests that increasing ethanol concentrations by 10%
`w/w will have a significant impact on MMAD and
`droplet size, as discussed below [53,58].
`
`5. Influence of formulation variables
`
`The influence of formulation variables will now be
`
`discussed in the context of these physicochemical
`characteristics and the above-mentioned measures of
`
`performance.
`
`5.1. Solution or suspension formulations?
`
`The active drug substance in pMDIs is either
`suspended or dissolved in the propellant or pro-
`pellant mixture (Table 3). Partial solubility of sus-
`pended drug is undesirable as it
`leads to crystal
`growth via a process known as Ostwald ripening
`[59]. As a consequence, particle size changes and
`irregular emitted doses may result. However, if a
`
`Surfactants /
`excipients
`
`Propellant
`system
`
`Formulation
`type
`
`Sorbitan trioleate
`Sorbitan trioleate
`Oleic acid
`Oleic acid
`Ascorbic acid,
`saccharin, menthol
`Oleic acid
`None
`
`Sorbitan trioleate,
`menthol
`
`Oleic acid
`Oleic acid
`
`CFC 11, CFC 12
`CFC 11, CFC 12
`CFC 11, CFC 12
`HFA 134a, ethanol
`38% w/w ethanol,
`CFC 11, CFC 12
`CFC 11, CFC 12
`HFA 134a
`
`CFC11, CFC 12,
`CFC 114
`CFC 12,
`1% W/W ethanol
`CFC 11, CFC 12
`CFC 11, CFC 12
`CFC 11, CFC 12
`HFA 134a, ethanol
`HFA 134a, ethanol
`CFC 11, CFC 12
`
`Suspension
`Suspension
`Suspension
`Suspension
`Solution
`
`Suspension
`Suspension
`
`Suspension
`
`Suspension
`
`Suspension
`Suspension
`Suspension
`Solution
`Solution
`Suspension
`
`Sorbitan trioleate
`Sorbitan trioleate
`
`CFC 11, CFC 12
`CFC 11, CFC 12
`
`Suspension
`Suspension
`
`9
`
`

`
`816
`
`H.D.C. Smyth / Advanced Drug Delivery Reviews 55 (2003) 807-828
`
`solution formulation is chosen the drug must have
`sufficient solubility to allow therapeutic doses to be
`delivered with a few actuations (usually two) of the
`device. Thus, careful selection of drug form and
`propellant system from solubility studies and com-
`patibility investigations is required in pMDI develop-
`ment. In suspension systems, the predominant factor
`limiting the minimum emitted droplet size is the size
`of the suspended particles that will be contained
`within aerosol droplets [60—62]. In solution systems,
`however, the droplet size is primarily governed by
`factors
`such as
`the non-volatile fraction, vapor
`pressure, actuator design, and the physicochemical
`nature of the liquid formulation [62]. With a limited
`range of propellant systems available, the selection
`of a solution or a suspension system may be by
`necessity rather than by choice. However, it should
`be recognized that each system has distinct forrnula—
`tion requirements and may have certain advantages
`for a given drug substance. Suspensions require very
`low solubility of the drug in the formulation. This
`usually results in good chemical stability of the drug
`[63]. However, aggregation and rapid flocculation of
`suspension systems may require addition of stabiliz-
`ing excipients such as surfactants. In solution sys-
`tems, high solubility and good stability of drug in the
`propellant system is required. Most current forrnula—
`tions are suspension systems containing micronized
`drug with particle sizes between 2 and 5 um [64].
`
`5.2. Eflect of vapor pressure
`
`5.2.1. CFC studies
`
`One of the first investigations of the influence of
`formulation factors on particle size was performed
`by Polli and co—workers
`[65]. Using suspension
`systems in CFC propellant blends, the effect of a
`number of formulation factors, including vapor pres-
`sure, was studied. Increased vapor pressure resulted
`in decreased aerosol particle size, ultimately to the
`size of the suspended particle when the vapor
`pressure of the formulation was 77 psig. When
`temperature was used to modify vapor pressure,
`similar trends were als