Modeling and Understanding Combination pMDI Formulations with

Article pubs.acs.org/molecularpharmaceutics

Modeling and Understanding Combination pMDI Formulations with Both Dissolved and Suspended Drugs Stephen W. Stein,† Poonam Sheth,*,‡ Usir S. Younis,§ Erik Mogalian,∥ and Paul B. Myrdal§ †

3M Drug Delivery Systems, 3M CenterBuilding 260-3A-05, St. Paul, Minnesota 55144, United States Cirrus Pharmaceuticals, Inc., 511 Davis Drive, Suite 100, PO Box 14748, Morrisville, North Carolina 27560, United States § College of Pharmacy, The University of Arizona, 1703 East Mabel Street, PO Box 210207, Tucson, Arizona 85721, United States ∥ Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, California 94404, United States Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.molpharmaceut.5b00467

‡

ABSTRACT: A simulation model has been established to predict the residual aerodynamic particle size distribution (APSD) of dual-component pressurized metered dose inhalers (pMDIs). More specifically, this model estimates the APSD of pMDI formulations containing dissolved and suspended compounds for various formulations, and has been verified experimentally. Simulated and experimental data illustrate that APSDs of the dissolved and suspended components of the pMDI are influenced by concentrations of the dissolved and micronized suspended drugs, along with suspended drug size. Atomized droplets from such combination formulations may contain varying number of suspended drug particles and a representative concentration of dissolved drug. These sub-populations of atomized droplets may explain the residual APSDs. The suspended drug follows a monomodal, lognormal distribution and is more greatly impacted by the size and concentration of the suspended drug in comparison to the concentration of dissolved drug. On the other hand, dissolved drug illustrates a bimodal, lognormal residual particle size distribution both theoretically and experimentally. The smaller mode consists of residual particles made of dissolved drug only, while the larger mode consists of residual particles that contain both dissolved and suspended drugs. The model effectively predicted the size distributions of both the dissolved and suspended components of combination formulations (r2 value of 0.914 for the comparison of simulated versus experimental MMAD values for the formulations examined). The results demonstrate that this model is a useful tool that may be able to expedite the development of combination pMDI formulation. KEYWORDS: pressurized metered dose inhalers (pMDIs), suspension formulations, aerodynamic particle size distribution (APSD), simulation model, cascade impactor

1. INTRODUCTION

therapy is suggested with the aforementioned drug classes and an anticholinergic agent. The addition of multiple inhaled drug therapies to a patient’s medication regimen frequently adds confusion and increased chances of poor medication regimen adherence. Thus, being able to combine therapies in one pMDI provides great benefits to patients. Despite these advantages, combining various drugs in one system complicates the formulation process due to differences in drug solubility and effective doses. This article introduces modifications to a previously developed computational model5,6 enabling the assessment of formulation variables that affect solution and suspension combination pMDI formulations. 1.1. Dual-Drug Combination pMDIs. When formulating two or more drugs in one pMDI device, a “coformulation effect” must be considered. A combination inhaler product of

Inhalation drug therapies are widely used for treating lung conditions such as asthma and chronic obstructive pulmonary disease (COPD). These delivery systems allow for treating pulmonary diseases while limiting adverse systemic effects. One of the modalities of delivering drugs to the lung is via the pressurized metered dose inhaler (pMDI). The first pMDI was produced in the mid-1950s by Riker Laboratories (now 3M Pharmaceuticals).1 These devices are now well accepted and highly utilized, with global annual production of over a halfbillion units2 and sales increasing by 10% from 2007 to 2010 (with 277 million in 2010).3 Current therapeutic guidelines for the treatment of asthma and COPD suggest that patients use drugs from different classes on a regular basis to effectively treat their conditions. The National Heart, Lung and Blood Institute’s (NHLBI’s) Guidelines for the Diagnosis and Management of Asthma4 recommend patient use of long-acting β2-agonists along with inhaled corticosteroids to manage their moderately severe chronic condition. If symptoms deteriorate further, a triple © XXXX American Chemical Society

Received: June 16, 2015 Revised: August 6, 2015 Accepted: August 10, 2015

A

DOI: 10.1021/acs.molpharmaceut.5b00467 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Downloaded by SUNY UPSTATE MEDICAL UNIV on August 25, 2015 | http://pubs.acs.org Publication Date (Web): August 24, 2015 | doi: 10.1021/acs.molpharmaceut.5b00467

Molecular Pharmaceutics

Figure 1. Simplistic model of droplet evolution from suspension and solution pMDIs.11

two drugs may have different product performance than if either were formulated separately at doses identical to that present in the combination inhaler product. Furthermore, if the drug concentration of one compound in a dual-drug combination pMDI is altered, it may impact the product performance, especially the aerodynamic particle size distribution (APSD) of the second compound.7 More often than not, product performance for the two drugs in a combination formulation differs. This effect is generally dose-dependent on the individual components of the pMDI formulation.7 1.2. Importance of APSD. Droplets initially atomized from pMDIs are relatively large and may contain propellant, cosolvent, dissolved excipients, and drug(s) in solution and/ or suspension (as presented in Figure 1). These droplets are lognormally distributed6 and undergo a dynamic transformation. Evaporation of the propellant and cosolvent leads to residual particles of nonvolatile matter that can deposit in the human lungs. The ability of a pMDI to deliver drug to the lung is largely dependent on the residual aerodynamic particle size distribution (APSD). APSD is often characterized by mass median aerodynamic diameter (MMAD) and geometric standard distribution (GSD). The aerodynamic diameter of a particle is defined as the diameter of a spherical particle with unit density that has the same settling velocity as the particle of interest. The mass median diameter (MMD) is the physical diameter at which half of the aerosolized mass lies above the

indicated diameter. The GSD represents the spread of the lognormal distribution of particles, where a GSD of 1 indicates a sample of monodispersed particles and a GSD greater than 1 indicates polydispersed particles. For approximately spherical droplets or particles, the MMD can be converted to MMAD by multiplying by the square root of density of the droplet or particle. For solution pMDIs, the MMAD of residual particles increases as a cube-root function of the dissolved concentration of the drug and nonvolatile excipients in the formulation (Figure 2), as predicted by theory.8,9 In contrast, for suspension formulations, the MMAD of residual particles has complex relationships with the suspended drug concentration, the size of the micronized input drug, and other factors6,10 but approximately follows some power function as shown in Figure 2. Particles with aerodynamic diameters of less than approximately 5 μm are more likely to penetrate into the lung.12 Along with clinical factors, formulation parameters (e.g., amount of cosolvent, nonvolatiles, and propellant), device parameters (e.g., valve size and orifice diameter), and variability in micronized drug lot (i.e., MMD and GSD of input suspended drug) can have a significant impact on APSD and thereby the therapeutic effect of an aerosolized product. 1.3. Value of Modeling APSD and Study Objectives. The atomization of nebulized monodisperse suspensions was first investigated by Raabe.13 Subsequently, the delivery from B


Article



perfect a potential pMDI product, it will not eliminate the need for rigorous, confirmatory in vitro testing of pMDI products.

2. MATERIALS AND METHODS 2.1. Experimental Materials and Procedures. Micronized albuterol sulfate of varying particle sizes was provided by 3M Drug Delivery Systems (St. Paul, MN, USA) and Micron Technologies Ltd. (Dartford, Kent, U.K.). Micronized lactose monohydrate of varying particle sizes, oleic acid, valves, and experimental actuators were provided by 3M Drug Delivery Systems. Lactose monohydrate was used as a model suspended drug for this study since it was available in a range of particle sizes. Flunisolide hemihydrate and beclomethasone dipropionate were used as model dissolved drugs in the selected pMDI systems. Pressure resistant glass vials were purchased from Research Products International Corp. (Mt. Prospect, IL, USA). HPLC-grade methanol and phosphoric acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). 200 proof ethanol was purchased from Decon Laboratories (King of Prussia, PA, USA), and 1,1,1,2-tetrafluoroethane (HFA-134a) from Atofina Chemicals Inc. (Philadelphia, PA, UA). 2.1.1. Determining the APSD of Micronized Drug. The particle size distribution of two lots of micronized albuterol sulfate were measured using the model 3321 Aerodynamic Particle Sizer spectrometer (APS) in conjunction with the model 3433 small scale powder disperser (both from TSI Inc., Shoreview, MN, USA). The first drug lot had an MMD of 2.30 μm and a GSD of 1.81. The second drug lot had an MMD of 1.55 μm and a GSD of 1.57. A third lot of albuterol sulfate was obtained by high shear homogenization of the first drug lot in 200 proof ethanol using a technique described by Jinks18 and James et al.19 After high shear homogenization, the particle size of the albuterol sulfate in the slurry was measured using the Malvern Mastersizer 2000 particle size analyzer (Malvern Instruments Ltd., Malvern, Worcestershire, U.K.). The MMD was determined to be 1.06 μm with a GSD of 1.57. In addition, the Malvern Mastersizer 2000 was also used to determine the size of two lots of lactose monohydrate. The first lot has an MMD of 2.12 μm with a GSD of 1.60. The second lot was prepared by high shear homogenization using an Avestin EmulsiFlex-050 (Avestin Europe GmbH, Mannheim, Germany) as described elsewhere.16,20 The MMD of the homogenized lot was 1.06 μm with a GSD of 1.60. 2.1.2. Formulation of pMDIs. 33 pMDIs containing 0 to 1.4% (w/w) of varying sizes of suspended, micronized albuterol sulfate or lactose monohydrate, 0 to 0.3% (w/w) dissolved drug (beclomethasone dipropionate or flunisolide hemihydrate), and 0 to 0.03% oleic acid with approximately 8.2% (w/w) 200 proof ethanol in HFA-134a were prepared in pressure resistant glass vials (see Table 1). Once the glass vials contained the desired amount of ethanol, surfactant, and drug(s), a cold-transfer technique was used to fill the vials with HFA-134a. Each vial was immediately crimped with a 50 μL Spraymiser metering valve using a small-scale bottle crimper. Vials were sonicated for 60 s to disperse the suspension. 2.1.3. Andersen Cascade Impactor (ACI) Testing. Prior to each collection, the stages of the ACI were thoroughly rinsed with 50% (v/v) methanol:water, followed by 100% methanol, and dried. Once dry, the stages and the USP inlet were coated with 50:50 methanol:Pluronic L10 to minimize particle bounce, and the methanol was allowed to evaporate. Actuators with orifice diameters of 0.3 mm were used for all testing. Each pMDI was actuated three times in order to prime the valve; the

Figure 2. Residual MMAD as a function of drug concentration for solution and suspension pMDI formulations. All formulations contain 8% (w/w) ethanol with HFA-134a.6,9

suspension pMDIs has been modeled by Gonda14 and Chan and Gonda,15 who built upon the work of Raabe to model delivery of monodisperse particles contained in polydisperse droplets. The delivery of drug from suspension pMDIs is more complicated than that modeled by Chan and Gonda, since the drug particles are polydisperse and most suspension pMDIs include nonvolatile excipients that modify the APSD. Some of the principles used by Raabe13 and Chan and Gonda14,15 can be applied to modeling polydispersed micronized drug in a suspension pMDI formulation with polydispersed droplets. A current simulation model5,6 effectively simulates the APSD for a single drug suspended in hydrofluoroalkane propellant with ethanol and is summarized in section 2.2. A useful model that predicts APSD from dual-drug pMDIs and does not require technical calculations and experimentation would benefit the formulation and quality control process for pMDI manufacturing. It would facilitate the formulation process by allowing for improved high throughput screening of plausible dual component pMDI products and decreased time and resource investment in current trial-and-error approach to evaluating test pMDIs. Furthermore, it may provide guidance in defining the design space for potentially successful pMDI products with two nonvolatile components, where one is in solution and the other is in suspension. The study described herein seeks to expand on the model developed by Stein et al.6 by including a dissolved drug in the suspension pMDI. This will provide a better understanding of how the APSD of the dissolved or suspended entity is modified by changing the concentration of either drug or changing the size of the suspended drug. Furthermore, the model’s application can be broadened to include predicting APSD for any type of nonvolatile component in solution and suspension in the hydrofluoroalkane−ethanol system. For instance, the model can be extended to evaluate formulations such as (1) suspended excipient with dissolved drug16 or (2) suspended drug with dissolved excipient, such as that seen in Proventil HFA formulation.17 This model would provide better understanding of the sensitivity of various device and formulation components to product performance, which permits a priori evaluation of batch-to-batch variation. While such a model is extremely useful in decreasing the time required to find or C


Article


stem of the valve was subsequently cleaned with the diluent (75:25 methanol:water). The valve stem was then dried, and the pMDI was fitted with a clean actuator. The flow rate through the ACI was adjusted to 28.3 L/min and verified using a TSI series 4000 flow meter (TSI Inc., Shoreview, MN, USA). Each pMDI vial was tested in triplicate. To obtain sufficient drug collection on each stage of the ACI, the number of actuations ranged from 2 to 25 depending on the concentration of each drug in the formulation (Table 1). The valve stem, actuator, USP inlet, jet, stages 0 to 7 of the ACI, and filter were rinsed with known volumes of the diluent, and the amount of drug present for each component was quantified by high performance liquid chromatography (HPLC). 2.1.4. Analytical Assay. The HPLC system consisted of a Waters 2690 separation module coupled with a Waters 996 photodiode array ultraviolet detector (both from Waters Corp., Milford, MA, USA). A gradient separation method was utilized to characterize beclomethasone dipropionate and albuterol sulfate. An Apollo C18 5 μm 150 mm × 4.6 mm column, maintained at 30 ± 2 °C, was used. Injection volume was 50 μL, and the initial mobile phase composition was 65:35% (v/v) methanol:0.1% phosphoric acid in water at a flow rate of 0.85 mL/min. After 3.5 min, the mobile phase linearly changed over 30 s to 80:20% (v/v) methanol:0.1% phosphoric acid with a flow rate of 1.5 mL/min. The data was collected and processed using Waters Empower Pro 2 chromatography data software. The total run time for this method was 12 min. Albuterol sulfate eluted at 1.6 min and was detected at 225 nm; beclomethasone dipropionate eluted at 8.2 min and was detected at 238 nm. Flunisolide hemihydrate was quantified using Supelcosil LC-18 5 μm 150 mm × 4.6 mm column. The mobile phase for this isocratic method was 90:10% (v/v) methanol:0.1% phosphoric acid in water at a flow rate of 1.5 mL/min. Flunisolide hemihydrate was detected at 238 nm. Quantification was conducted based on peak area using a fivepoint standard curve. Lactose monohydrate was not quantified. No leachable and extractable compounds were detected from the pMDIs or bag used to rinse the ACI stages upon analysis of the HPLC data. 2.1.5. Determining APSD of Residual Particles. The HPLC results from the ACI test were used to determine the APSD of the drug delivered in the residual aerosols. Chimera Technologies, Inc., DistFit 2009.01 (Forest Lake, MN, USA) was used to determine the MMAD and GSD of the residual particles. The data was fitted with monomodal and bimodal lognormal distributions based on whether the drug was in suspension or in solution with the ethanol/HFA-134a system. No size information is available for the portion of the drug that deposited on the valve stem, actuator, and USP inlet; thus, these were not included in the APSD calculations. 2.2. Theoretical Framework. A Monte Carlo simulation model was previously established to estimate the APSD of residual particles for pMDIs containing a propellant, ethanol, and a single API in suspension.5,6 This paper introduces modifications to include an additional entity in solution, as depicted in Figure 3. To simplify discussion, the suspended and dissolved entities are labeled as drugs in this study; however, they may also be nonvolatile excipients. This figure consists of two panels: the top panel depicts the physical atomization process while the bottom panel explains how this process is modeled. When a pMDI is actuated, the formulation exits the actuator spray orifice and is broken up into many initial atomized droplets that contain a representative proportion of

Table 1. pMDI Formulations Used for Experimental APSD Testing suspended druga and concn (% w/w) (MMD, GSD)

ethanol concn (% w/w)

no. of actuations per ACI collection

8.1 8.0 8.4 8.0 8.9

15 5 5 5 25

8.7

10

8.6

3

8.1

25

8.1

6

9.0

2

8.2

25

8.2

6

8.7

2

BDP, 0.076

8.2

25

BDP, 0.249

8.4

20

BDP, 0.090

8.2

15

BDP, 0.302

8.0

5

BDP, 0.079

8.3

7

BDP, 0.306

8.4

3

BDP, 0.071

8.0

25

BDP, 0.282

8.3

20

BDP, 0.068

8.1

15

BDP, 0.260

8.0

5

BDP, 0.073

8.0

7

BDP, 0.268

8.0

3

BDP, 0.080

9.9

5

BDP, 0.085

8.0

5

dissolved druga and concn (% w/w)


BDP, 0.084 BDP, 0.167 BDP, 0.253 FH, 0.166 AS, 0.0093 (1.06 μm, AS, 0.215 (1.06 μm, AS, 0.878 (1.06 μm, AS, 0.030 (1.55 μm, AS, 0.410 (1.55 μm, AS 1.028 (1.55 μm, AS, 0.052 (2.30 μm, AS, 0.409 (2.30 μm, AS, 1.096 (2.30 μm, AS, 0.032 (1.55 μm, AS, 0.019 (1.55 μm, AS, 0.532 (1.55 μm, AS, 0.407 (1.55 μm, AS, 0.982 (1.55 μm, AS, 1.039 (1.55 μm, AS, 0.028 (2.30 μm, AS, 0.038 (2.30 μm, AS, 0.672 (2.30 μm, AS, 0.728 (2.30 μm, AS, 1.238 (2.30 μm, AS, 1.356 (2.30 μm, LM, 0.973 (1.06 μm, LM, 1.003 (2.12 μm, AS, 0.402 (1.06 μm, AS, 0.126 (1.06 μm, AS, 0.397 (1.70 μm, AS, 0.121 (1.70 μm, LM, 0.101 (1.06 μm, LM, 0.987 (2.12 μm,

oleic acid concn (% w/w)

0.019

1.57) 1.57) 1.57) 1.57) 1.57) 1.57) 1.81) 1.81) 1.81) 1.57) 1.57) 1.57) 1.57) 1.57) 1.57) 1.81) 1.81) 1.81) 1.81) 1.81) 1.81) 1.60) 1.60) BDP, 0.084

0.018

8.0

5

FH, 0.171

0.017

8.1

5

BDP, 0.083

0.030

7.9

5

FH, 0.169

0.020

8.0

5

FH, 0.169

0.021

8.0

5

FH, 0.164

0.019

7.9

5

1.57) 1.57) 1.60) 1.60) 1.60) 1.60)

a

AS: albuterol sulfate. LM: lactose monohydrate. BDP: beclomethasone dipropionate. FH: flunisolide hemihydrate. D


Article



Figure 3. Depiction of the physical process of atomization (top panel) compared to the proposed simulation algorithm (bottom panel). The brackets toward the top of the “Simulated Process” panel indicate the relationship between each simulation step and its corresponding physical process. Sections highlighted in light red denote the modifications to the model described by Stein et al.5,6

each miscible excipient as that found in the formulation. If the drug is in solution, the atomized droplets will also contain an identical proportion of drug found in the formulation. However, if the drug is suspended in the formulation, varying number of drug particles can be found in each droplet. Most pMDI products on the market deliver tens to hundreds of millions of suspended drug-containing droplets per actuation.21 Over time, the propellant and some of the cosolvent evaporates given their volatile nature, rendering an intermediate droplet that contains some cosolvent, dissolved drug, and/or suspended drug. The cosolvent will completely evaporate over the due course of droplet evolution leaving residual particles that constitute only nonvolatile mass, which can deposit in the human lungs. These residual particles are significantly smaller than their corresponding initial droplets. Since the original model is described in detail elsewhere,5,6 it is only briefly discussed here. The original model enables the prediction of APSD from a suspension formulation, while expanding on previous models by allowing for polydispersed micronized suspended drug and initial droplets. The key inputs are density and concentration of each component of the formulation, the micronized drug size distribution for the suspended component, and the size distribution of the initially atomized droplets. The latter can be estimated using the equations provided by Stein and Myrdal.9 Typically, for HFA134a formulations tested using the USP inlet, the initial droplet MMD is between 8 and 13 μm11 and the GSD is approximately 1.5 to 1.8.9 For the purposes of the experiments described herein, the initial droplet MMD was assumed to be 9.14 μm with a GSD of 1.8, based on calculations using eq 1 and experimental measurements of a pMDI with 8% (w/w) ethanol, 0.083% (w/w) dissolved drug (ρ = 1.25 g/cm3) in HFA-134a with representative device configurations that resulted in a residual particle MMAD of 0.94 μm. Since the nonvolatile component is a small fraction of the formulation, small changes in the nonvolatile concentration are not expected to significantly modify the initial droplet size distribution.

MMDI =

MMADR (ρI C NV )1/3 ρR 1/6

(1)

where MMDI is the initial droplet MMD, MMADR is the residual particle MMAD, CNV (weight fraction) is the concentration of the dissolved nonvolatile component, and ρI and ρR are the densities of the formulation and the residual particles, respectively.9 Once the user inputs the above information, each atomized droplet is simulated individually. The size of a given atomized droplet is determined by a random sampling of droplet size based on the statistics of the initial droplet size distribution entered by the user. Thereafter, the number of drug particles contained within the droplet is determined. The likelihood that a droplet will contain one or more suspended drug particles depends on volume of the droplet and the number of drug particles per unit volume of the formulation (PPUV, see eq 2). It can be described using the Poisson distribution function. The Poisson distribution is used to randomly determine the number of particles in a given droplet based on the PPUV in the formulation and a given droplet’s volume. Large droplets have a higher probability of having one or more drug particles than small droplets. Similarly, droplets from formulations that contain a higher PPUV are more likely to contain drug particles than are droplets of the same size for a formulation with fewer PPUV. Thus, pMDIs with a high drug concentration (or ones with relatively small micronized drug) can have a substantial fraction of drug-laden droplets that contain multiple drug particles (“multiplets”), and a smaller fraction of drugladen droplets that contain a single suspended drug particle (“singlets”). 2

PPUV =

6Cdrug e 4.5ln GSDdrug ρI π (0.0001 MMDdrug )3 ρdrug

(2)

where PPUV (cm−3) is the number of suspended drug particles per unit volume of formulation, Cdrug (weight fraction) is the E


Article



Figure 4. Calculation of aerodynamic diameter for residual particles, where the subscripted variables are defined in the legend. In addition, V (cm3) is the volume, M (g) is the mass, D (cm) is the diameter, and ρ (g/cm3) is the density. Additional variables are defined as follows: χ is the shape factor, and ρ0 (g/cm3) is unit density. See Stein et al.6 for additional details.

concentration of the suspended drug, ρdrug is the suspended drug particle density, ρI is the formulation density, and GSDdrug and MMDdrug (μm) are the geometric standard deviation and mass median diameter, respectively, of the micronized drug used in the formulation. Finally, the size of each suspended drug particle is determined by a random sampling of particle size based on the statistics of the input particle size distribution of the micronized material. This information, in conjunction with the suspended drug density, is used to determine the volume and mass of each residual particle. Once sufficient drug-laden particles are simulated, the APSD for the suspended drug can be determined by fitting a lognormal distribution to the mass versus aerodynamic diameter of residual particles data. For the simulations presented in this paper, at least 30,000 drug-laden particles were simulated.22 2.2.1. Modification of Theoretical Model. As illustrated in Figure 3, atomized droplets from a combination pMDI formulation containing a drug in suspension and a drug in solution may contain only dissolved drug or dissolved and suspended drugs, with ethanol and propellant. These droplets will evaporate over time, rendering residual particles containing only dissolved drug or both dissolved and suspended drugs, with a varying number of suspended drug particles. To account for the additional dissolved entity in the formulation, the user must input the concentration and density of the entity prior to using the model to determine the APSD of the suspended or dissolved component. In order to calculate the amount of dissolved drug that is contained in a dual-component droplet, the simulated total volume of the suspended drug particles is calculated and subtracted from the volume of the initial droplet. It is then assumed that this “void volume” of the droplet is occupied by the propellant, cosolvent, and dissolved drug at proportions

equal to that found in the bulk formulation. Based on the relative composition of the formulation, the volume and the mass of the dissolved drug contained within a given initial droplet can be determined (see eqs 3 and 4). By summing the mass of the suspended and dissolved drugs for each particle, the mass of the residual droplet can be calculated, which then determines the mass-equivalent diameter for a given residual particle (see Figure 4). Thereafter, the aerodynamic diameter of the residual particle is computed, taking into account shape factor. Shape factor is based on the number of suspended drug particles contained within the residual particle.6 Once sufficient droplets are simulated, the mass of suspended or dissolved drug and aerodynamic diameter for each droplet are utilized to determine the APSD for the suspended or dissolved drug. The data is fitted with a monomodal or bimodal lognormal distribution using DistFit. Msoln = ρI Csoln(VI − Vsusp)

(3)

where Msoln (g) is the mass of drug in solution, ρI (g/cm3) is the density of the formulation, Csoln (weight fraction) is the concentration of the dissolved drug in the formulation, and VI and Vsusp (cm3) are the volumes of the initial atomized droplet and suspended drug particles within the droplet, respectively. Vsoln =

Msoln ρsoln

(4)

where Vsoln (cm ) is the volume, Msoln (g) is the mass, and ρsoln (g/cm3) is the density of the dissolved drug. 3

3. RESULTS AND DISCUSSION 3.1. Simulation Model. Figure 5 compares example APSD results from simulated and experimental evaluations of six pMDI product configurations, containing HFA-134a and F


Article



Figure 5. Examples of simulated (Sim) and experimental (ACI) APSDs of the (a) dissolved component and (b) suspended component for a variety of pMDI formulations. All of the formulations contained a nominal 8% (w/w) ethanol with HFA-134a.

ethanol. Some of the configurations presented only contain dissolved drug while other configurations also contain suspended drug with or without oleic acid. The size distribution of the dissolved drug for the formulations evaluated is shown in Figure 5a. For the formulations which also included suspended drug, only the size distribution of the suspended drug component is shown in Figure 5b. Formulations that contain both a suspended and dissolved drug will have two distinct distributions: one corresponding to the dissolved drug and another corresponding to the suspended drug. These distributions may be described by lognormal monomodal or bimodal distributions. The examples in Figures 5a and 5b present simulated (dashed curves) and experimental (solid curves) APSDs of the dissolved nonvolatile component (Figure 5a) and the suspended component (Figure 5b). The value of the aerodynamic diameter corresponding to the 50th percentile of the cumulative mass is the MMAD for the listed entity. The width of the distribution is an indication of the GSD. From this figure, the following can be noted and are later discussed with the support of additional data:

MMAD, which can be seen by comparing distributions in Figures 5a and 5b. (2) The primary particle size of the suspended drug influences the resulting APSD of the suspended component for combination formulations. This is presented by comparing the blue curves (primary particle MMD of 1.06 μm) and green curves (primary particle MMD of 1.70 μm) in Figure 5b. (3) Typically, a formulation that contains only dissolved drug at a low concentration will have a relatively small residual particle MMAD and a narrow distribution, compared to that of the solution component of combination formulations. The inclusion of suspended drug particles has the effect of broadening the size distribution of the dissolved drug component. This is seen when comparing the green to blue and red to purple curves for the formulations presented in Figure 5a. (4) Addition of low concentration of suspended drug in a formulation containing dissolved drug has a lower impact on the resulting APSD for the dissolved drug than would the addition of a high concentration of suspended drug. This is evidenced when the examples in purple, black,

(1) Generally, the suspended drug residual particle MMAD is larger than that of the dissolved drug residual particle G


Article

Molecular Pharmaceutics and orange are compared to those in blue for the formulations presented in Figure 5a. When comparing the dashed to solid lines in Figure 5, it is evident that the model developed is effective at estimating the APSD of a variety of formulations with varying concentrations of dissolved and suspended entities that possess different physicochemical properties. Figure 6 presents a comparison of

Table 2. Suspension Component’s APSD for Representative Experimental Formulations

albuterol sulfate micronized size (MMD) and concn (% w/w) 1.55 μm, 0.03 1.55 μm, 0.4 2.30 μm, 0.03 2.30 μm, 0.4


suspension component

formulation beclomethasone dipropionate concn (% w/w)

MMAD (μm)

GSD

0 0.3 0 0.3 0 0.3 0 0.3

1.68 2.02 2.15 2.33 2.35 2.63 2.81 2.77

1.55 1.53 1.54 1.55 1.74 1.66 1.70 1.64

when dissolved beclomethasone dipropionate concentration increases. The impact that the dissolved drug has on the residual MMAD of the suspended material is most prominent for formulations with low concentrations of suspended drug. To further investigate the trends that dissolved drug can have on the APSD of the suspended component of a formulation, a series of simulations were performed. Figure 7 illustrates the MMAD changes as a function of suspended drug concentration Figure 6. Comparison of experimental residual particle MMAD to simulated MMAD for the formulations detailed in Table 1. All of the formulations contained a nominal 8% (w/w) ethanol with HFA-134a.

MMAD of solution and suspension entities in the combination formulations listed in Table 1. The MMAD of the dissolved entities are presented as “effective MMAD”, which is the MMAD of the distribution assuming a monomodal lognormal distribution. As will be described later, the dissolved component’s APSD from a combination formulation is better described using a bimodal distribution; however, the inclusion of two distributions for dissolved entity of each combination formulation complicates comparisons and is occasionally limited by the sensitivity of the in vitro techniques. As presented in Figure 6, there was good agreement between the MMAD values predicted in the simulations and those measured from the ACI for the formulations evaluated (r2 = 0.914). This demonstrates the utility of the algorithm for predicting residual APSDs for a broad range of pMDI configurations with varying concentrations, primary particle sizes, and drug compounds. This stochastic model can further be utilized to provide insight on factors that affect the APSD of dissolved and suspended nonvolatile entities in such combination pMDI formulations. 3.2. Understanding Suspension pMDI Formulations Containing Dissolved Entity: Effect of Dissolved Material on Suspended Component APSD. Similar to the results and discussion presented in the literature,5,10,16 the suspended component’s APSD in a combination formulation is most affected by the micronized drug or excipient primary particle size and concentration. Examples are presented in Table 2. For a primary particle size of micronized albuterol sulfate at a given beclomethasone dipropionate concentration, increasing the suspended component’s concentration or primary particle size of the micronized drug results in an increase in the residual MMAD of the suspended entity. The coformulated data in Table 2 also indicates that the size distribution for the suspended component of the formulation also tends to increase

Figure 7. Residual APSD (i.e., MMAD) of model suspended component as a function of suspended and dissolved component concentrations. Micronized primary particle MMD of 0.9 μm (top) and 2.2 μm (bottom), both with GSD of 1.8. H


Article



Figure 8. Examples of how dissolved drug concentration can affect the residual aerodynamic diameter of suspension component. The equations are simplifications of the equations presented in Figure 4. Vsoln is the volume of the residual particle occupied by the dissolved nonvolatile drug. Vvoid is the void volume of the residual particle and is assumed to be 25.9% of the volume of the suspended particles in a given residual particle.

versus 1.71 μm) and 1.27 for 2.2 μm MMAD primarily particle size formulations (3.47 μm versus 2.74 μm). The effects seen with the experimental data presented in Table 2 and simulated data presented in Figure 7 may be explained by the varying propensity for atomized droplets to contain multiple suspended particles based on formulation. Recall that APSD is a function of the aerodynamic diameters of the residual particles and the associated masses of the various nonvolatile formulation components. The APSD is lognormally distributed such that a small number of large particles can have a greater impact on the overall MMAD than a large number of small particles. For all of the APSDs presented in this section, two subsets of residual droplets are considered: one population that contains a single suspended drug particle in a residual droplet and another population that contains multiple suspended drug particles in a residual droplet. The latter population typically has residual particle aerodynamic diameters larger than those of the singlet population for a given formulation. The addition of dissolved drug to a suspension formulation has the potential for increasing the residual particles’ aerodynamic diameters by affecting the volume and mass of the residual particles, as exemplified in the three scenarios in Figure 8. For the singlet population, the addition of

and dissolved drug concentration for two different primary particle size distributions for micronized suspended drug. In general, the contour plots between the 0.9 and 2.2 μm raw micronized drug MMD are similar. However, when the dissolved drug component increases from 0 to 0.8%, the relative increase in particle size is slightly greater for 0.9 μm compared to the 2.2 μm input drug size, as evidenced by the contour lines. For instance, the residual MMAD increases from 2.29 to 3.11 μm (36% increase) as the dissolved drug increases from 0 to 0.8% for formulations containing 1% of a 0.9 μm raw micronized MMD drug. In contrast, the same increase in the dissolved drug concentration results only in a 20% increase (from 3.22 to 3.86 μm) for formulations containing a 1% suspension of drug with a 2.2 μm raw micronized MMD particle size. For both plots presented in Figure 7, the influence of dissolved drug concentration has a relatively greater impact on the MMAD as the suspension concentration decreases. To illustrate, compare the aforementioned ratios to the following ratios for the residual MMADs corresponding to that of 0.2% suspension formulations containing 0.8% dissolved drug relative to formulations not containing dissolved drug: 1.52 for 0.9 μm MMD primary particle size formulations (2.58 μm I


Article



Figure 9. Depiction of why the dissolved drug of a combination formulation exhibits a bimodal distribution.

Figure 10. Residual APSDs determined by cascade impaction of a beclomethasone dipropionate (BDP) solution formulation marked with the solid black curve and the solid bars and BDP in a combination formulation containing albuterol sulfate (AS) marked with the dashed curve and the bars with a diagonal fill. The micronized MMD of albuterol sulfate is 1.55 μm. These formulations also contain 8% (w/w) ethanol with HFA-134a.

dissolved drug results in “coating” the suspended drug particles, which increases the residual particle volumes, masses, and aerodynamic diameters (similar to the bottom scenario depicted in Figure 8). For the multiplet population, the impact that the dissolved drug has on the aerodynamic diameter is less straightforward. For these droplets, the addition of dissolved drug increases the mass of the residual particles, but may not impact the volume of the residual particle if the dissolved drug is not sufficient to exceed the void volume between the individual drug particles in the aggregate (compare top and bottom scenarios of Figure 8). The void volume is typically larger in residual particles for large aggregates with many suspended drug particles and/or relatively large input micronized drug size. Thus, the addition of dissolved drug

may not significantly impact the resulting portion of the APSD made up of these particles. 3.3. The Influence of Suspended Drug on Solution Component APSD. Dissolved drug or excipient concentration and initial droplet diameter are the primary determinants for the APSD of solution pMDIs. The residual APSD of solution pMDIs follows a monomodal lognormal distribution. In contrast, the residual APSD of the dissolved component of combination pMDIs follows a bimodal lognormal distribution, as explained in Figure 9. All of the aerosolized droplets from combination solution/suspension pMDIs contain the dissolved nonvolatile drug and any dissolved excipient. For aerosolized droplets that do not contain suspended particles, the droplets contain an amount of dissolved drug that is proportional to the droplet volume and the concentration of the dissolved drug in J


Article


Molecular Pharmaceutics the formulation. For droplets that do contain suspended particles, the remainder of the volume of initial droplets that is not occupied by the suspended drug particle(s) (i.e., the “void volume”) consists of the formulation with some amount of dissolved drug (as determined by the concentration of dissolved drug in the formulation). Both types of aerosolized droplets contribute to the overall APSD of the nonvolatile dissolved component, theoretically resulting in a bimodal lognormal distribution. Many times, due to the formulation variables, it is difficult to resolve the two modes; thus, a monomodal distribution is frequently fitted to these formulations. These monomodal distributions typically have relatively large GSDs, greater than those seen with the suspension counterpart or from solution pMDIs containing only the dissolved drug. For instance, Figure 10 presents the monomodal distribution of a solution formulation containing 0.08% beclomethasone dipropionate and a bimodal distribution of 0.08% beclomethasone dipropionate with 1% suspended albuterol sulfate (micronized drug MMD 1.55 μm). As a solution formulation, with reasonable Chi-square goodness-of-fit, the beclomethasone dipropionate solution formulation has a residual MMAD of 1.03 μm with a GSD of 1.56. The effective APSD for the solution component of the combination formulation has an MMAD of 1.49 μm with a GSD of 2.79. This effective distribution is composed of two lognormal modes for the beclomethasone dipropionate residual APSD: 1.42 μm MMAD (GSD 2.46) and 2.40 μm MMAD (GSD 1.53). The smaller mode is predominantly representative of aerosolized particles that do not contain suspended drug particles. However, this distribution has a larger MMAD and GSD than those seen with the 0.08% beclomethasone dipropionate formulations since it is difficult to experimentally resolve the two modes. The larger mode (2.40 μm MMAD [GSD 1.53]) has a similar distribution to that seen for albuterol sulfate in a 1% albuterol sulfate suspension formulation with a micronized drug size of 1.55 μm containing 0.08% beclomethasone dipropionate (2.45 μm MMAD [GSD 1.60]). Simulations present a further insight into the bimodal distribution phenomenon. Figures 11 and 12 illustrate the simulated ACI profile for 8% ethanol formulations in HFA 134a. Figure 11 presents a formulation that contains 0.08% dissolved beclomethasone dipropionate (black bars) and another formulation that contains 0.08% beclomethasone dipropionate and 1% suspended albuterol sulfate (input particle size of 1.55 μm MMD, red bars). The red bars in Figure 11 are simulated size distribution results for the same formulation for which experimental size distribution results are shown in Figure 10. One of the formulations presented in Figure 12 contains 0.019% oleic acid and 0.166% flunisolide hemihydrate (presented as black bars), and the other formulation contains 0.019% oleic acid, 0.164% flunisolide hemihydrate, and 0.987% of 2.12 μm MMD lactose monohydrate used as a model drug (presented as blue bars). In both examples, the striped bars represent the proportion of dissolved drug mass in residual particles from the combination formulation that do not contain suspended drug particles. The filled blue or red bars represent the proportion of dissolved drug mass in residual particles from the combination formulation that contains suspended particles. This illustrates the cause of the bimodal nature of the dissolved drug from pMDIs with one drug in solution and the other drug suspended in formulation.

Figure 11. Residual APSDs determined by simulation of a 0.08% (w/ w) beclomethasone dipropionate (BDP) solution formulation and 0.08% (w/w) beclomethasone dipropionate in a combination formulation containing 1% (w/w) albuterol sulfate (AS). The micronized MMD of albuterol sulfate is 1.55 μm. These formulations also contain 8% (w/w) ethanol with HFA-134a. This simulated data provides greater insight into the bimodal nature of the dissolved entity in a suspension formulation.

Figure 12. Residual APSD as determined by simulations of dissolved drug in the following two formulations: 0.019% oleic acid, 0.166% flunisolide hemihydrate with 8% ethanol, and 0.019% oleic acid, 0.164% flunisolide hemihydrate, 0.987% 2.12 μm MMD lactose monohydrate with 8% ethanol. The mass predicted on each stage of the ACI is based on the total mass recovered in the ACI through experiments.

In general, the ACI profiles presented in Figures 11 and 12 for the solution pMDI formulation have greater deposition of beclomethasone dipropionate or flunisolide hemihydrate on the lower stages of the ACI (plate 5 and below), which correspond to residual particles with relatively small aerodynamic diameters (≤2.1 μm). For the formulations that contain suspended albuterol sulfate or lactose, more of the total dissolved drug mass (red and blue filled and striped bars) is deposited on the upper stages of the ACI (plate 5 and above). The mass of beclomethasone dipropionate or flunisolide hemihydrate from combination formulations collected on plates 0 to 5 is primarily associated with atomized droplets that contained suspended K


Article


4. CONCLUSIONS A Monte Carlo simulation model to predict the residual aerodynamic particle size distribution (APSD) was refined from a previously existing model.6 Verified experimentally, this model enables the estimation of the APSD of complex pMDI formulations containing dissolved and/or suspended drug(s) and/or excipient(s) for a variety of formulation. This model may be further adapted to determine the APSD of formulations containing multiple dissolved entities23 (using a model without a suspended entity) or cosuspension formulations where two active pharmaceutical ingredients are physically bound to act as a single suspended entity24 (using a model without a dissolved entity). Through the experiments and data presented, it is evident that the concentration of nonvolatile drug and the concentration and size of micronized suspended drug have significant impact on the APSD of both the dissolved and suspended components in combination pMDIs. The size distribution of the suspended drug tends to follow a monomodal lognormal distribution, and it is influenced significantly by the size and concentration of the suspended drug, and to a lesser extent the concentration of the dissolved drug (except in situations of relatively low suspension concentration formulations). The size distribution of the dissolved drug was shown both theoretically and experimentally to consist of a bimodal lognormal distribution. This bimodal distribution consists of a smaller mode that is made up of residual particles that contain only the dissolved drug and a larger mode that is made up of residual particles that contain both the dissolved and suspended drugs.


albuterol sulfate or lactose, in addition to the dissolved drug (red or blue filled bars). In the lower stages, it is clear that the beclomethasone dipropionate and flunisolide hemihydrate mass is associated with droplets that do not contain suspended particles. Figures 11 and 12 also illustrate that the dissolved drug in a combination formulation has a broader size distribution than does a standard solution pMDI formulation. Figure 13 shows experimental measurements of pMDIs containing various concentrations of dissolved becomethasone

Figure 13. MMAD determined by experimental measurements of beclomethasone dipropionate (BDP) in combination pMDI formulations with albuterol sulfate (AS). The data are presented as mean ± standard deviation over three replicates for each data point. The concentrations of beclomethasone dipropionate are 0.08% (unfilled symbols) and 0.3% (filled symbols) with varying albuterol sulfate concentration and micronized sizes (expressed as MMD) with 8% ethanol in HFA-134a.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (919) 321-4045. Fax: (919) 884-2075. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

dipropionate (BDP) and suspended albuterol sulfate (AS) of differing particle sizes. Increasing the concentration of the suspended drug leads to an increase in the residual MMAD of the drug that was dissolved in the formulation. Similar to solution-only pMDI formulations, increasing the concentration of the dissolved drug also increases the MMAD of the dissolved drug. In Figure 13, increasing the concentration of beclomethasone dipropionate from 0.08% (unfilled symbols) to 0.3% (filled symbols) results in an increase in residual MMAD of the beclomethasone in the combination formulation with albuterol sulfate. The influence of the concentration of the dissolved drug on the resulting dissolved drug’s MMAD is diminished for formulations that contain a high concentration of suspended component. For instance, the MMAD difference between the 0.08% and 0.3% beclomethasone dipropionate formulations for the 1% albuterol sulfate (micronized MMD 1.55 μm) is 0.16 μm; for the 1% albuterol sulfate (micronized MMD 2.30 μm), the value is 0.17 μm (see Figure 13). Conversely, the changes in the dissolved component’s concentration has a significant effect on the resulting dissolved drug’s MMAD for formulations that contain a low to moderate concentration of suspended drug. For instance, the MMAD difference between the 0.08% and 0.3% beclomethasone dipropionate formulations for the 0.4% albuterol sulfate (micronized MMD 1.55 μm) is 0.47 μm, and for the 0.4% albuterol sulfate (micronized MMD 2.30 μm), it is 0.28 μm.

ACKNOWLEDGMENTS The authors would like to acknowledge the Pharmaceutical Research and Manufacturers of America (PhRMA) Foundation for a pharmaceutics predoctoral fellowship that supported this work. In addition, they would like to acknowledge Forest L. Danford at The University of Arizona for his assistance in programming the presented model.

■

REFERENCES

(1) Thiel, C. G. From Susie’s Question to CFC Free: An Inventor’s Perspective on Forty Years of MDI Development and Regulation. Respir. Drug Delivery 1996, 1, 115−124. (2) Newman, S. P.; Peart, J. Pressurized Metered Dose Inhalers. In Respiratory Drug Delivery: Essential Theory & Practice; Newman, S. P., Ed.; Sheridan Books, Inc.: 2009; pp 177−216. (3) IMS Health Analysis, IMS MIDAS Quantum, MAT. December 2010. (4) US Department of Health and Human Services, National Institutes of Health, National Heart, Lung, and Blood Institute. National Asthma Education Program. Expert panel on the management of asthma. Expert panel report 3: Guidelines for the diagnosis and management of asthma: Full report. 2007. Available at https:// www.nhlbi.nih.gov/guidelines/asthma. Accessed April 18, 2014. (5) Stein, S. W.; Sheth, P.; Karayiannis, C.; Chiou, H.; Myrdal, P. B. Modeling MDI Delivery: A Priori Predictions, Empirical Models and Experiments. Respir. Drug Delivery 2010, 1, 353−364.

L


Article


Molecular Pharmaceutics (6) Stein, S. W.; Sheth, P.; Myrdal, P. B. A Model for Predicting Size Distributions Delivered from pMDIs with Suspended Drug. Int. J. Pharm. 2012, 422, 101−115. (7) Lechuga-Ballesteros, D.; Noga, B.; Vehring, R.; Cummings, R. H.; Dwivedi, S. K. Novel Cosuspension Metered-Dose Inhalers for the Combination Therapy of Chronic Obstructive Pulmonary Disease and Asthma. Future Med. Chem. 2011, 3, 1703−1718. (8) Meakin, B. J.; Lewis, D. A.; Ganderton, D.; Brambilla, G. Countering Challenges Posed by Mimicry of CFC Performance Using HFA Systems. Respir. Drug Delivery 2000, 1, 99−108. (9) Stein, S. W.; Myrdal, P. B. A Theoretical and Experimental Analysis of Formulation and Device Parameters Affecting Solution MDI Size Distributions. J. Pharm. Sci. 2004, 93, 2158−2175. (10) Sheth, P.; Stein, S. W.; Myrdal, P. B. Factors Influencing Aerodynamic Particle Size Distribution of Suspension Pressurized Metered Dose Inhalers. AAPS PharmSciTech 2015, 16, 192−2015. (11) Sheth, P.; Stein, S. W.; Myrdal, P. B. The Influence of Initial Atomized Droplet Size on Residual Particle Size from Pressurized Metered Dose Inhalers. Int. J. Pharm. 2013, 455, 57−65. (12) Labiris, N. R.; Dolovich, M. B. Pulmonary Drug Delivery. Part II: The Role of Inhalant Delivery Devices and Drug Formulations in Therapeutic Effectiveness of Aerosolized Medications. Br. J. Clin. Pharmacol. 2003, 56, 600−612. (13) Raabe, O. G. The Dilution of Monodisperse Suspensions for Aerosolization. Am. Ind. Hyg. Assoc. J. 1968, 29, 439−443. (14) Gonda, I. Development of a Systematic Theory of Suspension Inhalation Aerosols. I. A Framework to Study the Effects of Aggregation on the Aerodynamic Behaviour of Drug Particles. Int. J. Pharm. 1985, 27, 99−116. (15) Hak-Kim, C.; Gonda, I. Development of a Systematic Theory of Suspension Inhalation Aerosols. II. Aggregates of Monodisperse Particles Nebulized in Polydisperse Droplets. Int. J. Pharm. 1988, 41, 147−157. (16) Stein, S. W. Formulations and Methods for Controlling MDI Particle Size Delivery. US 272,951, 2012. (17) Proventil HFA (albuterol sulfate) Inhalation Aerosol [Package Insert]. Merck & Co. Inc.: Whitehouse Station, NJ, 2012. (18) Jinks, P. A. Preparation and Utility of Sub-Micron Lactose, a Novel Excipient for HFA MDI Suspension Formulations [abstract]. Drug Delivery Lungs 2003, 14, 199−202. (19) James, J.; Crean, B.; Davies, M.; Toon, R.; Jinks, P.; Roberts, C. J. The Surface Characterization and Comparison of Two Potential Sub-Micron, Sugar Bulking Excipients for Use in Low-Dose, Suspension Formulations in Metered Dose Inhalers. Int. J. Pharm. 2008, 361, 209−221. (20) Jinks, P. A.; McKenzie, L.; Lister, J. T. Medicinal Aerosol Formulations. US 81,627, 2004. (21) Stein, S. W. Estimating the Number of Droplets and Drug Particles Emitted from MDIs. AAPS PharmSciTech 2008, 9, 112−115. (22) Stein, S. W. The Influence of Sample Size on the Accuracy of Particle Size Distribution Measurements. Respir. Drug Delivery 2008, 333−336. (23) Adi, H.; Young, P. M.; Traini, D. Co-deposition of a Triple Therapy Drug Formulation for the Treatment of Chronic Obstructive Pulmonary Disease Using Solution-Based Pressurised Metered Dose Inhalers. J. Pharm. Pharmacol. 2012, 64 (9), 1245−1253. (24) Vehring, R.; Lechuga-Ballesteros, D.; Joshi, V.; Noga, B.; Dwivedi, S. K. Cosuspensions of Microcrystals and Engineered Microparticles for Uniform and Efficient Delivery of Respiratory Therapeutics from Pressurized Metered Dose Inhalers. Langmuir 2012, 28 (42), 15015−15023.

M


Modeling and Understanding Combination pMDI Formulations with

Recommend Documents