PHARMACEUTICAL TECHNOLOGY EUROPE · N OV E M B E R 1 9 9 9

Position Paper on Particle Sizing:

Sample Preparation, Method Validation
and Data Presentation
1
The particle size of a drug substance has long been known to be a
critical parameter, because of its influence on vital processing factors
such as flowability, mixing, content uniformity, dissolution rate and
bioavailability. Particle sizing methods have suffered from a lack of
understanding and standardization in the pharmaceutical industry.
This article discusses various aspects of developing, implementing
and validating particle sizing methodology.

he Pharmaceutical Analytical Sciences

Richard Bell,
Andrew Dennis,
Barry Hendriksen,*
T Group (PASG) has published proposals
for the qualification of analytical
equipment.1 Regulations and guidelines for
computerized analytical equipment have
Nigel North and been summarized in the Organization for
John Sherwood Economic Co-operation Development’s (OECD)
GLP (good laboratory practice) consensus
paper.2 These requirements need to be built Artwork generated by Peter Fielder-Shaw.
into the process of equipment qualification.
On successful completion of the qualification of an experience in assessing particle size data and could
instrument, method validation can begin. potentially misunderstand poorly presented infor-
Richard Bell, pharmaceutical scientist, Validation of analytical methods in the pharma- mation. An attempt has therefore been made to
Product Development, AstraZeneca ceutical industry is well established and docu- offer guidelines for a simple data presentation
R&D Charnwood, Bakewell Road, mented.3 Method validation guidelines have been template. This document has been developed by,
Loughborough LE11 5RH, UK. published by organizations such as the Interna- and reflects the agreed position of, the members of
Andrew Dennis, manager, tional Conference on Harmonization (ICH), the the PASG particle characterization subgroup.
pharmaceutical sciences, Bristol-Myers United States Pharmacopeia (USP) and the US Food
Squibb Ltd, Pharmaceutical Research and Drug Administration (FDA). According to ICH Sample preparation for particle size
Institute, Reeds Lane, method validation guidelines, laboratories need to analysis
Moreton CH46 1QW, UK. investigate the following parameters: accuracy, Sampling itself will not be considered here. Sample
Barry Hendriksen, research analyst, precision, specificity, detection limit, quantitation preparation will be taken to mean the pretreatment
Pharmaceutical Research, Eli Lilly and limit, linearity and range. and the presentation of the sample to the mea-
Company Ltd, Lilly Research Centre, Erl suring technique in a meaningful manner — that
Wood Manor, Windlesham GU20 6PH, Particle size methodology is, to determine the particle size distribution in the
UK. Validation of particle sizing methodology is not original sample without altering it in the process.
Tel. +44 1276 853 260 included in the ICH guideline, so the applicability If particle reduction processes such as micro-
Fax +44 1276 853 392 of these parameters to this technique has not been nization are used, it is important that they are
E-mail [email protected] considered. PASG therefore set up a subgroup to monitored to retain process control. Different tech-
Nigel North, team leader, investigate the possibility of developing method niques of size measurement impose different con-
Pharmaceutical Technologies, validation guidelines for particle sizing techniques. straints on the preparation process involved.
SmithKline Beecham Pharmaceuticals, ICH terminology and definitions were reviewed and The propensity of particles to form larger
Third Avenue, Harlow CM19 5AW, UK. their applicability to particle sizing methodology agglomerates causes major problems in preparing
John Sherwood, technical expert, assessed. samples for particle size analysis. Not all materials
AstraZeneca Pharmaceuticals Department, Because sample preparation is an important experience this problem, but of those that do, the
Silk Road Business Park, Charter Way, first step in any particle sizing method, guidelines forces between particles may be relatively weak
Macclesfield SK10 4TG, UK. were developed to assist this process and explain its (aggregation according to BS29554), whereby the
importance. It is imperative that presentation of energy needed to separate them is relatively low. In
*To whom all correspondence should data takes account of the customer who will use the this case, the final distribution should resemble the
be addressed. information, because many users have limited initial one, and the method used to effect the
 PA R T I C L E S I Z I N G

separation (the sample preparation) should dispersions, which resulted in small particles should be examined. As guidance, at least
not change the properties of the sample. But dissolving (or precipitating) upon standing. six independent analyses on the same
if the forces between particles are relatively Both temperature and time therefore need to sample are recommended, if sample avail-
strong (agglomeration according to BS29554), be considered. ability permits.
much more energy is needed to separate To overcome the problematic cases cited Intermediate precision. The checking of the
them. Consequently, the separated particles above, sample preparation should be min- precision of a technique within a laboratory
may bear no resemblance to the initial imal and microscopic examination (either should be conducted using, where appro-
agglomerate, and the ‘result’ would be sensi- wet or dry) should be considered as a means priate, different instruments or different per-
tive to the exact method and energy input of visually estimating the particle size data. sonnel. This should be performed on a
2 used. This would be particularly true if the If a numeric quantification is still thought number of samples and repeated as appro-
particles were brittle or fragile, or if they necessary, the particle size distribution priate for each variant.
were easily deformed, in which case the orig- should be measured as a function of the Reproducibility. This parameter indicates pre-
inal particle size is unobtainable in principle. energy input, although no such thing as a cision between laboratories and is appro-
Intuitively, micronized diamond would ‘right’ answer is to be expected with this priate when technical transfer of the method
present no problems because any particle method. Alternatively, simply ensure that the to a different location occurs. This should be
association could be broken up without conditions are rigidly maintained and accept conducted on a number of representative
changing the original hard particles. But soft comparative data. In all cases, the down- samples as appropriate.
plastic particles, once agglomerated, would stream process needs to be considered, Specificity. This is not strictly applicable to
separate into particles bearing no resem- together with the reason for requiring the particle size analysis. It is analogous in some
blance to the original ones. Between these particle size data. For example, if subsequent ways to accuracy. The user should ensure use
two extremes lie the relatively soft organic granulation is envisaged, the sample prepara- of the right technique for the right sample;
molecules, the particle size of which appar- tion method should not use more energy than for example, whether the technique measures
ently depends upon the energy used in the granulation process is likely to impart. agglomerates or primary particles.
sample preparation. Additionally, as emulsion Detection limit. This is neither a simple nor,
droplets represent the softest particles, Guidance for validation of particle usually, a meaningful term in the validation
examining them microscopically under a sizing methods of particle size analysis. A more appropriate
coverslide could flatten the droplets, making This section reviews and discusses potential term is ‘range,’ because it is often difficult to
them look bigger, or breaking them up into criteria for particle sizing method validation. differentiate between quantification and
smaller droplets — depending on the pres- Each criterion has been taken from the ICH detection. The approximate detection limit of
sure applied. When the apparent particle size tripartite guideline, Validation of Analytical the technique should be verified, for
proves to be a function of the energy Methods: Definitions and Terminology.5 The example, by reference to the instrument
expended during the sample preparation, the most relevant and appropriate subject areas specification and confirmed as suitable.
determination of the original particle size for validation of particle sizing methods are Quantification limit. This is not usually con-
therefore becomes inaccessible. summarized in the recommendations. sidered to be a useful term, as it is dependent
If the downstream process involves par- Procedure. Detailed description of the on the sample size employed, the type of
ticle size growth, such as granulation, it proposed technique is required. This should sample and the technique being used.
should be determined which parameter — include a section on why the technique is Linearity. In particle sizing, the absence of a
primary particle size or granule size — is suitable (possibly including a brief micro- linear response is expected — that is, no
required. The relationship between these and scopic evaluation), equivalent to a ‘suitability change in particle size with, for example,
bioefficacy needs to be established. assessment’ or a ‘fitness for use’ statement. concentration or other sample variables.
In one example cited by a company Accuracy. It is questionable whether this is Studies should be undertaken to confirm that
participating in the PASG subgroup, batches applicable for particle sizing, which may not sample concentration, sample mass or similar
of a drug were sized by light scattering, have a right answer, in comparison with a variables do not affect the response.
which was followed by ultrasonication. The chromatography (HPLC) technique, for Range. Ideally, the range of the specific
results appeared to be consistently similar, example, which does — at least theoretically. technique chosen should cover the potential
but microscopic examination of the original For sample characterization, the analysis particle size ranges of the samples to be
samples revealed the presence of larger crys- required is likely to vary in different situa- analysed. This should be described in the
tals not detected by light scattering because tions, for example, agglomerates or primary procedure if appropriate. Choice of the tech-
the sample preparation procedure broke them particles, and number or volume analyses. nique for the potential range of particle size
down into smaller ones. Intra-company and Accuracy may be more appropriately viewed distributions should be considered. Ideally, a
intra-laboratory comparisons have also indi- as calibration or verification of the tech- study with samples representing the extremes
cated that slightly different preparative nique, to a defined and traceable physical of the potential range should be evaluated to
equipment and methods often give widely standard. Calibration, however, is common to confirm suitability. In addition, the method
different results, even when, ostensibly, the all analytical procedures and is not part of should be checked directly using, for
same procedure was adopted. This illustrates method validation. The examination of the example, microscopy.
the importance of considering energy input sample to be tested, by microscopy (quali- Robustness. The lack of sensitivity to minor
during sample preparation. tative not quantitative), should verify changes in the technique should also be
When samples are soluble in the disper- whether the technique is suitable for analysis assessed, such as the amount of ultrasoni-
sing medium, saturated solutions (of analyte of the sample and, in many cases, whether it cation required for sample preparation,
in dispersing medium) may need to be used. is the best assessment of accuracy. different sample weights or vacuum pressure
Significant temperature differences, however, Precision for air jet sieving. This should confirm the
have been experienced between laboratories Repeatability. The repeatability of the whole sample stability and the window of operation
using saturated solutions and those using technique, including the sample preparation, for the procedure. Experimental design may
 PHARMACEUTICAL TECHNOLOGY EUROPE · N OV E M B E R 1 9 9 9

Guidelines for presentation of

particle size data
Drug A Lot Number 1 Particle Size Report Our suggestions are intended as a guideline
Purpose for the presentation of particle size data gen-
This data has been generated as part of the validation of the method for particle size erated by a range of techniques. Figure 1 rep-
determination of Drug A by laser diffraction. resents a specimen template and histogram
Analytical method to illustrate how this might be done. The fol-
Laser diffraction according to Specification/XYZ/1 lowing aspects should be considered individ-
Substance description ually and applied in an appropriate manner.
2 Drug A Lot number 1 is comprised of off white acicular particles (a)
Example photomicrograph
Length of report. If possible keep the report 3
to one page to aid customer appreciation.
Purpose. State the reason the data have
(Note: The laser light diffraction technique been generated. Use appropriate wording to
assumes particles are spherical. The photomicro- minimize risk of subsequent misuse of data.
graph above indicates this is not the case for the Particle size analysis method. State the
majority of particles. Therefore, the suitability of analysis technique and analytical method
the size data has to be judged with this limitation used (with reference to testing protocol).
in mind.) Material description. Provide a brief descrip-
tion of substance using BS2955/USP defini-
tions for particle shape (quoting source of
nomenclature). Include a photomicrograph of
material whenever possible.
Particle size data
Data presentation. Report key parameters
Analysis number: 00000001
with visual illustration such as a histogram.
Equivalent spherical diameter (b) (mean of six measurements and standard deviation) Customer appreciation of data may be
10th percentile (Tm)  71.6 enhanced by explaining reported parameters
50th percentile (Tm)  179.0 as a footnote. Provide location of raw data
90th percentile (Tm)  544.9 storage (such as analysis number) to allow
Example histogram cross-reference to raw data if required.
10
Volume %, 3 results. Stats 1 - S.D.
100 Data precision. The use of electronic calcula-
90 tion gives rise to particle size results being
80 reported that can give the impression of
70 being more accurate than they are. Rounding
60 of figures should normally be employed,
50 using the significant figures approach.
40 Particle size will always be particularly
30 important in pharmaceutical development.
20 With adequate attention to the preparation
10 of the sample, the measured parameters will
0
1.0 10.0 100.0 1000.0 10000.0
0 resemble the actual particle size as reliably as
Particle diameter (m) the sample properties inherently allow.
Adequate validation of the chosen method-
Figure 1: Particle size report. (a) Particle shape defined according to BS29555. (b) The equivalent spherical ology will ensure the quality of the measured
diameter is the diameter of a sphere having the same volume as the actual non-spherical particle measured. parameters. And perhaps above all, the way
the data is presented to the users is vital to
its application and usefulness.
be used to efficiently determine the robust- The validation of a particle sizing tech-
ness of the procedure. nique cannot be completed with the first References
batch of a sample, as data are insufficient to 1. M. Freeman et al., “Position Paper on the Qual-
Recommendations fully validate the procedure at this stage. It ification of Analytical Equipment,” Pharm.
Of the possible validation headings dis- is not always important to know the absolute Technol. Eur. 7(10), 40–46 (1995).
cussed above, the following items are con- particle size of the first batch manufactured, 2. OECD, “The Application of the Principles of GLP
sidered significant and should be included in but it is important to know how the second to Computerized Systems,” OECD Environmental
a particle sizing validation report: and subsequent batches compare with it. The Monograph No. 10 (December 1995).
• the procedure (or reference to it) validation of a particle size method should be 3. M. Swartz and I. Krull, Eds., Analytical Method
• precision in phases, with parts completed when the Development and Validation (Marcel Dekker,
• range (suitability assessment including first sample is analysed and other data com- New York, New York, USA, 1997).
microscopic comparison) pleted or added later, for example, prior to 4. BS2955, Section 3, The Glossary of Terms
• robustness. the regulatory filing. Calibration or verifica- Relating to Particle Technology (British Stan-
Detection limit, quantification limit, accu- tion of an instrument being used is assumed dards Institution, London, UK, 1993).
racy and specificity are not normally consid- to be a normal part of GMP (good manufac- 5. International Conference on Harmonization,
ered appropriate for validation of particle turing practice) and may be used as a suit- Topic Q2A — Validation of Analytical Methods:
sizing methods. ability assessment. Definitions and Terminology (1995).