Rapid and reliable in vitro methods of analysis of aerosol particles are required in product development to ensure product safety and to attain process control quality to establish reproducibility of inhalation aerosols. The particle size distribution for beclomethasone dipropionate delivered from Becotide® and Respocort® inhalers was estimated using an Andersen Cascade Impactor and a scanning electron microscope coupled with an automated image analysis technique. The particle area and the volume-weighted distribution were estimated and compared to the volume-weighted distribution generated using a laser diffraction technique. No significant differences in the measurement of the particle size distributions using either area or the volume-weighted distributions for the small particles were evident. An alternative optical sizing method is described to evaluate particle size distribution and morphology of aerosols administered in metered dose inhalers using a cascade impactor and a scanning electron microscope. This method can directly measure particle size and also permits observation of the particle and surface texture of the particles under investigation.
PDF Abstract XML References Citation
How to cite this article
Drug particle size distribution and shape are important parameters in the evaluation of pharmaceutical formulations. These physical characteristics of drugs affect their dissolution and, hence, the absorption of a drug from different sites in the body which in turn influences its pharmacodynamic response. Determination of particle size and the particle size distribution of pharmaceutical preparations has been investigated using various methods such as inertial impaction[1,2] laser diffraction and microscopy.
Characterization of spherical particles is generally straightforward, however, characterization of irregular shaped particles still remains difficult. Equivalent diameters (Fig. 1) are mainly used to describe the characteristics of irregularly shaped particles. The equivalent diameters of irregular drug particles are solely based on the geometry of the silhouette of the particle.
Microscopy is one of the methods that is used for particle size characterization for inhalation aerosols. It is the only method in which individual particles are viewed and measured. Microscopy provides both qualitative and quantitative information about particle size and shape. The instrumental methods of drug particle analysis are generally very rapid and large numbers of particles can be sized in a short period of time. These instrumental methods, however, do not provide direct observations of the particles as with microscopical methods. In addition, characterization of some aerosolized particles such as those generated from metered-dose inhalers cannot be determined by automated instrumentation techniques. For example laser diffraction is not appropriate for particle sizing from metered-dose inhalers due to low obscuration obtained in the laser diffraction instrument from the spray plume after dilution in the air stream. For these reasons, microscopical analysis is generally considered the reference method in aerosol particle analysis.
Manual techniques that are used for determination of particle size by microscopical analysis are very often tedious. They are also subject to considerable individual variability due to operator fatigue and the small number of particles which can be measured within a realistic time-frame. With the advent of automated image analysis systems, microscopical particle size determination can now be performed much more rapidly and with greater precision on a larger number of particles.
In measuring particle size using an optical technique such as microscopy, it is necessary to assign to each particle a size based on its two-dimensional projected image or silhouette. For spherical particles, it is the diameter of the circular silhouette observed in the microscope.
|Statistical diameters used in particle sizing including Martin's diameter (dM), projected area diameter (dPA) and Feret's diameter (dF )
However, for irregular shaped particles, an equivalent diameter is commonly used (Fig. 1). The equivalent diameters shown are based on the geometry of the silhouette. Feret's diameter which is defined as the length of the projection of a particle along a given reference line or the distance between the extreme left and right tangents that are perpendicular to the reference line. Feret's diameter is a convenient diameter to be employed when the scale along a given axis is available, such as that in a microscope equipped with a filar micrometer. The most commonly used equivalent diameter is the projected area diameter. The projected area diameter is defined as the diameter of the circle that has the same projected area as the particle silhouette. This diameter has the advantage of providing a unique size for a given silhouette, regardless of its orientation. Martin's diameter is the length of the line parallel to a given reference line that divides the projected area (silhouette) of the particle into two equal areas (Fig. 1).
The specific aim of this study was to examine the suitability of particle sizing and size distribution morphology analysis of beclomethasone dipropionate particles delivered from Becotide® and Respocort® metered-dose inhalers using a scanning electron microscope with an automated image analysis technique and through dispersing the aerosol particles from the metered dose inhaler device into an Andersen cascade impactor to simulate the actual dispersion situation of drug inhalation. The projected area, the projected area diameter, Feret's minimum diameter and Feret's maximum diameter were estimated and the volume-weighted parameter are calculated from the area measurement.
MATERIALS AND METHODS
Metered-dose inhalers tested were Respocort® 100 μg inhaler (3M Pharmaceuticals, Australia) and Becotide® 100 μg inhaler (Allen and Hanburys, a Division of Glaxo Wellcome Australia) one device from each inhaler was used for this study.
Alternative method of particle sampling and sizing: Samples of aerosolized particles from the Becotide® and Respocort® metered-dose inhalers were collected using the Andersen Mark II Cascade impactor (Graseby Andersen, Atlanta, CA) as previously described with minor modifications[7-9].
Two carbon tab sample collection plates (12 mm in diameter specially designed to be viewed under the scanning electron microscope) were placed on each stage of the collecting plates of the impactor. The impactor was connected to a vacuum pump calibrated at 60 L min-1. Two actuations from each inhaler were introduced into the impactor through the USP style standard induction port. The vacuum pump was allowed to continue for a further 20 sec after introduction of the dose before the pump was switch off. The impactor was dismantled, the two sample collection plates were removed. The samples were then mounted in the microscope chamber under vacuum prior to image analysis. The scanning electron microscope was a Model 505, Phillips instrument (Holland). The magnification was set to 2500 x, spot size was 50 mm and the electrical power was 15 KV. The scanning electron microscope was calibrated both horizontally and vertically. Ten fields from each sample plate were randomly viewed and imaged. A total of around 150 primary particles were sized on each sample plate at each stage of the impactor. A sum of 2500 primary particles were randomly chosen and sized in the image analysis technique for each inhaler.
The particle size distribution measurements were performed on Kontron Electronics 400 version 3 software (Carl Zeiss Vision Gmbh, Germany) to estimate the projected area, the projected area diameter, Feret's min and Feret's max diameter. Measurement of the various quantiles were performed using (JMP Statistics Software, V3.1 SAS Institute Inc., Cary, NC, USA).
Area measurement: The projected area of the particles can be determined precisely by manually tracing the outline of the particle using the computer cursor, and by counting the number of pixels occupied by the digitalized version of the particle automatically by image analysis software. This type of analysis represents a significant improvement over manual methods that estimate the area of the particle by comparing the projected area of the particle with areas of reference circles. Accordingly, the Feret's minimum and Feret's maximum diameter and the projected area diameter can also be generated automatically.
Volume estimation from the area measurement: The volume of the particle was estimated in this study from the area measured with the aid of the parameter ti, which is related to the thickness of the particle and is defined as the ratio of volume (Vi) to the area (Ai) of ith particle. (Equation1)
The parameter ti represents a third dimension that cannot be determined from two-dimensional microscopical measurements. However, it is assumed that ti will generally fall between two approximations. It is assumed that ti is constant and represents a uniform thickness of all particles. In this case, a volume distribution histogram of particles is independent of the actual value of ti because it is common to both numerator and denominator (Equation 2). The fraction of the particles in the size range β to δ, over n total particles is given as follows:
In the case where ti is constant, the area-and volume-weighted distributions are equal, and the particle size distribution can be represented by the area-weighted distribution.
A second assumption is that the non-measurable, third dimension ti increases in the same proportion as do the x and y dimensions. This condition may closely represent the volume-weighted distribution of the spherical particles. Here, ti is proportional to the square root of Ai Equation 3.
Where k is a constant. Combining equation 3 and 1 and solving for Vi yields Equation 4 which estimates the volume of a particle from a measurement of the area.
The value of k is dependent on the shape of a particle. For example with spherical particles, k is calculated using Equation 5.
The shape of a volume-weighted histogram of particles is independent of the actual value of k. The fraction of particles in the size range β to δ, over n total particles is given as follows:
The constant k which is common to the numerator and denominator, cancels each other so that a volume-weighted distribution histogram is independent of the value of k. With the second approximation, where all dimensions increase proportionally, the volume-weighted distribution is determined by raising each of the measured areas to the power of 3/2.
Characterization of particle size by laser diffraction: Particle size distribution of the original powders of beclomethasone dipropionate from the Becotide® and Respocort® inhalers were determined using laser diffraction. Five actuations from each of the metered-dose inhalers was suspended in 5 ml of distilled water. The suspension was sonicated for 30 sec before being loaded into a stirred sample cell containing water. Measurements were repeated five times, each measurement being one minute apart, to ensure that no dissolution or agglomeration of the particles occurred. The particle size distribution of the powder was determined using a density value of 1.26 g ml-1. The true density of the glucocorticoid powders was determined by the buoyancy method. Powder samples (1 to 2 mg) were placed in a density gradient liquid and centrifuged at 3500 rpm and 5°C for 30 min. The particle density was equal to the liquid density when the particles remained stationary in the liquid after centrifugation.
RESULTS AND DISCUSSION
Alternative method of particle sampling and sizing Particle size and morphology
Particle from stage zero: The scanning electron microscope images of beclomethasone dipropionate particles in stage number zero showed that particles exhibited a wide size distribution which ranged from 1 to 10 μm. The morphology of the particles in this stage varied from large irregular shaped particles with low sphericity to small particles comprising aggregates of around 10 μm in diameter (Fig. 2A and B). Image analysis based on projected area diameter estimated using the automated image analysis technique for beclomethasone dipropionate particles delivered from the Becotide® and Respocort® metered-dose inhalers showed a wide particle size distribution which ranged from size 1.2 to 8.1 μm and from 1.49 to 7.16 μm for the Becotide® and Respocort® inhalers, respectively (Table 1 and 2).
In the dispersion of the inhaler dose into the impactor, particles with aerodynamic diameter ranged from 9.0 to 10.0 μm will deposit on stage number zero. Thus, the size of the deposited particles or particle aggregates shown in the scanning electron microscope images are consistent with the aerodynamic size range in this stage. However, the range of the projected area diameter is inconsistent with the aerodynamic size range of the particles impacted on this stage. This is because the calculated projected area diameter is based on sizing the original particles, and not on the aggregates or conglomerate of particles.
The ratio of Feret's minimum to Feret's maximum diameter provides an estimate of the degree of roundness of the particles. A ratio close to unity means the particle geometrically approaching a spherical shape. The ratio of Feret's minimum maximum diameter is presented in Table 3 and 4. The value of the ratio for stage zero was low compared with the following stages which is entirely consistent with the irregularity of the particles particularly the large drug particles. Thus particles in this stage are a combination of irregular, angular with low sphericity to angular with medium sphericity particles (Fig. 2 ).
Particles from stages one to four: Beclomethasone dipropionate particles deposited on stages one to four consisted mainly of aggregates of primary particles generated from both Becotide® and Respocort® inhaler (Fig. 3). The size of particle aggregates was observed to be smaller at lower stages which is consistent with the expected aerodynamic size range for each stage.
|Particle size distribution of projected area diameter (μm) per quantiles for beclomethasone dipropionate particles delivered from the Becotide® inhaler
|Particle size distribution of projected area diameter (μm) per quantiles for beclomethasone dipropionate particles delivered from the Respocort® inhaler
|Particle size distribution of the ratio (Feret's min / Feret's max) diameter (μm) per quantiles for beclomethasone dipropionate particles delivered from Becotide® inhaler
|Particle size distribution of the ratio (Feret's min / Feret's max) diameter (μm) per quantiles for beclomethasone dipropionate delivered from the Respocort® inhaler
|* Quantiles of the particles in all stages of the impactor
The aerodynamic size range for each stage of the Andersen Cascade Impactor is calibrated over the following ranges from 9 to 10.0, 5.8 to 9.0, 4.7 to 5.8, 3.3 to 4.7, 2.1 to 3.3, 1.1 to 2.1, 0.7 to 1.1, 0.4 to 0.7 μm for stages 0 to 7, respectively.
|Typical scanning electron microscope images for beclomethasone dipropionate particles delivered from the Becotide® inhaler at stage zero of the Andersen Cascade impactor
|Fig. . 3:
|Typical scanning electron microscope images for beclomethasone dipropionate particles delivered from the Becotide® inhaler at stage one of the Andersen Cascade impactor
|Typical scanning electron microscope images for beclomethasone dipropionate particles delivered from the Becotide® inhaler at stage seven of the Andersen Cascade impactor
|Comparison of volume-weighted distribution for beclomethasone dipropionate particles obtained using the microscopical image analysis and Laser diffraction techniques
The large aggregates found in the first stages of the impactor are due to the fact that the aerosol particles are expelled from the inhaler with a high linear velocity (25 to 50 m sec-1) and have a droplet size at the spray jet of 30 to 50 μm. These particles or droplets consist of a powdered drug core coated with surfactant. Depending upon the concentration of the drug in the droplets of propellant and on the degree of propellant evaporation, the powder drug core can consist of either individual drug particles or aggregates. The particle aggregates are observed in the first stages of the impactor, because the propellant droplets do not have enough time to evaporate completely. Thus propellant droplets with a powder core consisting of aggregates will impact on the first stages of the impactor and these aggregates behave as one particle with an aerodynamic diameter comparable with the cut-off diameter of the impactor stage. This observation is supported by the scanning electron microscope images taken of particles from these stages.
The Ferets minimum to maximum diameter ratios for particles size on these stages is higher than that for stage number zero indicating that particles deposited on these stages are more spherical in nature compared with particles collected on stage zero of the impactor. The particle size distribution of the projected area diameter of the beclomethasone dipropionate particles on these stages is not consistent with the expected aerodynamic size range for these stages. The measurement of the projected area diameter is based on sizing of the original particles which are normally small particles forming a big aggregate.
Particles from stages five to seven: The morphology of beclomethasone dipropionate particles deposited on stages five to seven consist mainly of the primary drug particles and a few small aggregates (Fig. 4).
In addition, particles deposited in these stages tend to be more spherical in nature than particles deposited in the earlier stages of the impactor as the Ferets minimum to maximum diameter ratio for these stages is higher than 0.7. Particle size distribution of the projected area diameter of particles in these stages ranged from 0.5 to 2.5 μm and from 0.5 to 2.6 μm for beclomethasone dipropionate delivered from the Becotide® and Respocort® metered-dose inhaler, respectively.
Image analysis and laser diffraction: The volume-weighted distribution obtained from automated microscopical image analysis using the two approximations and that obtained from the laser diffraction technique are indicated in Table 5.
Except quantile under 90% for particles from the Becotide® and Respocort® inhalers which gives higher values, the present data show that the volume-weighted distribution obtained from the microscopical image analysis by the two approximations are in good agreement. The higher values for particles from the Respocort® inhaler in the under 90% quantile could be due to the large particles presented in the stages zero and one. These findings suggests that no significant differences in the results were obtained when the particle size was calculated according to the area or volume-weighted distribution.
Comparing the particle size distribution for beclomethasone dipropionate delivered from the Becotide® and Respocort® inhalers obtained from the automated image analysis technique with that obtained from the laser diffraction technique, it was found that except for the quantile less than 10% which are comparable, the values obtained from the laser diffraction were significantly higher. The cascade impactor deals with particles according to their aerodynamic diamter. If an aerosol droplet consists of an aggregate of small particles, the cascade impactor will deal with the particle as one particle and it will impact according to its aerodynamic diameter, and this is consistent with the cut-off diameter of the stage. However, when we evaluate this droplet using a scanning microscope, each particle in the droplets will be sized and measured, for example if this droplet consists of 5 particles, the 5 particles will be sized separately.
With microscopy there is always a possibility of sampling bias and care must be taken to select particles to size in a way that is not size dependent. One possible mistake is discarding particles that touch the edge of the image rather than selecting particles that have their centroid within a reduced area of interest within the image.
The two methods may compare poorly since the particles collected on the impactor stages contain both primary particles and aggregates of the primary particles and the suspension measured by laser diffraction results is composed of primary particles. However, the primary particles were sized manually on the impactor to be comparable with the laser diffraction technique. A cascade impactor is an official USP technique to simulate the lung and should more accurately reflect the actual situation of inhaler use than particle sizing by a laser diffraction technique.
The image analysis system described in this study provides an alternative method of microscopical particle size determination that can improve the speed and precision compared to manual microscopical determination. In addition to providing validation of instrumental particle size methods, this new method offers an alternative approach for particles size determination for compounds that cannot be analyzed by instrumental methods because of limitations of the instruments or the nature of the particles. Many attempts were made to use a Mastersizer to investigate particle size distribution along with a synchronization unit, and the metered dose inhaler mounting unit. Unfortunately, Malvern Mastersizer does not function with a metered dose inhaler because of the low concentration of the plume (spray).
Investigating particle size distribution of primary particles using microscopical methods of analysis coupled to a cascade may be a useful technique in product development of various xenobiotics including glucocorticoids. The sizing and distribution actually reflect deposition into a simulated lung along with concomitant agglomerization, wheras, laser diffraction only reflects primary particle sizing which does not mimic a clinical situation for particle deposition.