Abstract
Combined physical and chemical characterization of pharmaceutical particles is important to delivering quality drug products. While a number of particle characterization techniques are available, these techniques are limited in that they typically provide either chemical or physical information, not both. Now, the combination of Raman spectroscopy and image analysis for particle characterization presented in this article provides a unique approach in the characterization of pharmaceutical powders. The size, shape and chemistry of a statistically relevant number of individual particles can be investigated, significantly improving process understanding.Applications examples presented here include: 1) the trace analysis of morphological outliers via coordinated physical and chemical identification, ,and 2) the ability to define physical and chemical attributes of drug product intermediates, granulation, extrudates, and other solid dispersions during process development. In summary, morphologically targeted chemical analysis of particles is shown to deliver data that correlates the physical and chemical characteristics of drug product intermediates with the end performance of a drug product, enhancing process knowledge and the ability to develop for Quality by Design.
Introduction
In the production of solid dosage forms for pharmaceutical products, intermediate particles are often developed with the goal to enhance ultimate performance of the drug product. For example, the active pharmaceutical ingredient (API) may be granulated to improve the flow or compressibility of the powder (1), or, the API may be dispersed within a polymer to form faster dissolving particles for improved bioavailability. Whatever the performance goal, the intermediate particle is almost always a mixture of the API and excipients, and the chemical and physical uniformity of this mixture will affect product quality. Therefore, correlating chemical and physical uniformity of the intermediate with end product performance is a step in developing the process knowledge and deliver quality product (2).
Common techniques for measuring the particle size of drug product intermediate particles include sieve cuts and laser diffraction, whereas chemical uniformity characterization is typically obtained separately by HPLC. As size and chemical information are not obtained simultaneously, and sieve cuts and laser diffraction techniques lack shape information, these techniques are often unable to determine critical attributes of the material(3). Despite this relative lack of information content, sieve cuts are often used as a preparative tool in pharmaceutical development to determine size-dependent characteristics of particles. Assay of the API across sieve cuts is a common analysis that provides size-related information on the chemistry of the material. And provided the sample is of adequate size, any of the bulk analysis techniques for particle characterization, such as ring shear test, tap testing, mercury porosimetry(4) can be utilized.
Chemical uniformity by HPLC assay of the API within a bulk granulate or powder defines only chemical heterogeneity, is subject to sampling variability, (5) and provides no information of the physical heterogeneity of the assayed samples. Image analysis tools allow for the simultaneous interrogation of the size and shape of pharmaceutical particles, and are being used more often due to the improved information content of data. Imaging techniques do not rely on reducing three dimensional particles to a single number that approximates a transformed shape (sphere) of the material as is commonly employed in laser diffraction( 3,5,6). Imaging tools allow for the interrogation of shape parameters based on the 2-dimensional area, the perimeter, the length, and the width of the particles (7). Particularly if automated, image analysis can be used to characterize as many as 20,000 to 1,000,000 particles.
Even with the addition of shape information to particle size provided by image analysis, there are still analytical problems that cannot be solved without the addition of chemical identification. Now, by combining Raman spectroscopy and image analysis into a single instrument platform, a 2-dimensional image of a particle can be associated with its Raman spectrum for coordinated chemical and physical characterization. This further enhancement to particle characterization for pharmaceutical solid dosage development, delivers size, shape AND chemistry of individual particles. Full data sets of these particles enhance the ability to characterize the quality of pharmaceutical particles in development.
Experimental
Instrumentation Overview
Coordinating image analysis and Raman spectroscopy on individual particles starts with image acquisition. A large area must first be scanned to capture a statistically relevant number of particles. During acquisition, size and shape parameters are calculated for all particles. Critical to spectroscopic interrogation, (x,y) spatial coordinates are retained with the size and shape information during image analysis. After acquisition, particles are classified by morphology and size. Individual particles within morphological classifications can then be targeted by their (x,y) coordinate for chemical identification. Chemical identification is performed using Raman spectroscopy. Raman spectra is acquired using a 785nm laser, a holographic transmission-based grating spectrometer, and a 50x dedicated Raman objective to interrogate a 3um sport size on the particle of interest. The individual particle spectra can then be compared to a known library for chemical identification. The resulting data set thus includes the size, shape and chemical identification of individual particles (8,9).
Applications
Example 1: Chemical Identification of Trace Component
In this example, there was a processing issue in the manufacture of a solid dosage form. Specifically, the HPLC assay for API content of the tablets was at times lower than the assay of the precompression blend. In the preparation of the samples for assay a residue was observed in the bottom of the sample flasks. It was hypothesized that this residue was an excipient that was not soluble in the diluent for the assay. To investigate the content of the residue in the assay flasks, a method was developed to examine the morphology and chemistry of individual particles via automated image analysis and coordinated Raman spectroscopy.
The residue was dried and then dispersed on slides for subsequent image analysis. A large sample area was analyzed to obtain reasonable sampling statistics, scanned as described earlier by translating the sample stage through the area of interest, and acquiring visible images at 5X magnification. Resulting individual particle images are processed during acquisition, the result being a data set with a variety of size and shape parameters associated with each particle. Classes based on circular equivalent diameter, convexity, and circularity were determined post data collection. A broad selection of particles in these classes was targeted for chemical identification via Raman spectroscopy. A dedicated 50X objective was used to acquire the Raman spectra. The resulting particle spectra were compared against a library containing pure component Raman spectra of the API and excipients in the formulation. The correlation score was then determined between the particle and each formulation component. In this way, the chemical identity of a broad selection of morphologically classed particles was obtained.
Figure 1- An example an irregular particle that was identified by image analysis and then identified by Raman Spectroscopy as the disintegrant.
Figure 2- The particle image of the first particle identified as API with a correlation score of 0.9526.
Initially all the particles were found to be the disintegrant in the formulation. These disintegrant particles were highly irregular in shape (Figure 1). However, as analysis continued, a single particle was discovered that was significantly correlated with the API (Figure 2), presenting a correlation score of 0.9526, (with a value of 1.0 being a perfect correlation.)
The morphology of this particle was significantly more circular than the average disintegrant. Therefore, other highly circular particles were targeted for chemical identification by Raman spectroscopy. Although not all spherical particle were API, all API were found in a class of particles defined by a larger circular equivalent diameter and a high score for circularity. Limits were set on these attributes to define that class of particles. Although forty percent of the particles with these morphological characteristics were identified as API particles, this morphological class was itself quite rare, comprising only 0.2% of the residue. Therefore less than 0.1% of the residue was API. Without the combined morphological and chemical approach, these particles may never have been identified.
Because the assay loss varied around 6-8%, it was determined that the loss of API was not occurring during HPLC analysis. Rather, the API was being retained in the manufacturing equipment as it was starting up. The process was subsequently optimized to limit this loss.
This experiment informed the analytical and process development for this early stage compound. But, the work also demonstrates the sensitivity made available through morphologically targeted Raman spectroscopy. Using morphology to guide the spectroscopy, a single unique particle can be identified and isolated for spectroscopic analysis. The needle can be identified in the haystack.
Example 2: Monitoring Process Scale Up and Transfer
When processing API into drug product intermediates such as granules, extrudates, or other solid dispersions, the uniformity of these intermediate materials can impact final drug product performance. In this example, the API to excipient ratio of a drug product intermediate was analyzed as the process was transferred from a smaller piece of formulation development equipment to a larger manufacturing unit. The process involved mixing the API and excipients to create a drug product intermediate of a preferred size and morphology, and keeping the component ratios consistent between processes was important for final drug product quality. The drug product intermediate from each piece of process equipment was examined for uniformity across a broad particle size distribution.
Figure 3- The correlation scores of the API in particles across the size distribution are similar between the new and the old processes.
First, the API content was examined as a function of circular equivalent diameter (CE diam). Samples were dispersed on slides and sized by image analysis for CE diam. Correlation scores for the API were then determined on individual particles throughout the particle size distribution. The API correlation scores across the size distribution for the new and older process equipment were compared (Figure 3)
The distribution of API scores across the particle size range were found to be equivalent, confirming that the move to a new, larger size of process equipment did not introduce a bias to the API content within the particle size distribution. This work confirmed an accurate technical transfer between pieces of processing equipment. It is important to note that CE diameter is just one of the attributes that was examined against a chemical correlation score. Other shape factors such as length, width, convexity, circularity, and area were also utilized to cover more attribute space and variability in development.
Example 3: Optimizing a Process During Scale Up for Dissolution Performance
For another scale-up process , the dissolution profile of a final drug product changed as a result of the granulation step moving from lab-scale to engineering scale. The size distribution by sieve cut analysis of the two materials did not adequately differentiate the materials for full process understanding. A higher resolution technique was therefore needed to better examine the quality of the materials. The subsequent utilization of morphologically-targeted Raman spectroscopy was able to identify differences between material that performed well in the lab and material that had reduced performance upon scale-up.
Figure 4- Differences in the lab-scale material (red trace) and the engineering material (green trace) were evident in the particle size distributions, indicated here in the 10-30 μm region of the histogram. These particles were targeted for chemical identification by Raman spectroscopy.
The primary difference observed by image analysis was in the level of fines in the materials, with an increase seen in the engineering batch compared to the laboratory scale (Figure 4). The chemical identify of these fines was then determined by Raman spectroscopy using the combined image analysis/Raman microscopy instrument platform. The results indicated that a majority of the fine material was composed of the ungranulated, micronized, API. Not surprisingly, the ungranulated API was determined to result from under-granulation. Using this coordinated approach to particle characterization, the prevalence of fine particles and their subsequent chemical identification as API was utilized for process development. , The engineering process was optimized to reduce under-granulation and implemented in time for the manufacture of clinical supplies.
Discussion
Today, characterization of drug product intermediates should rely on a multidisciplinary approach to particle characterization, as no single technique can provide data on all possible attributes. Morphologically targeted Raman spectroscopy is a step in the right direction, in that it combines both physical and chemical information, providing highly resolved size analysis, and enabling chemical identification of morphological classes of interest. This multidisciplinary technique provides unique insights into the relationships of physical and chemical attributes within a particle population.
Image analysis has proven to be highly sensitive to morphological outliers, and the addition of Raman spectroscopy allows for the subsequent chemical identification of trace components and unique particles. Additionally, using morphologically targeted Raman spectroscopy, a process can be examined for chemical identity versus morphology at a much higher resolution than other techniques. A chemical assay of sieve cuts will provide the same type of information as the morphologically targeted spectroscopic approach. But, the assay of sieve cuts provides a mean value for 6-12 different data ranges on the particle size axis. The binning for particle attributes is much more highly resolved when using morphologically-targeted spectroscopy. This approach provides a more accurate description of the material and more confidence in developing a process.
The physical and chemical heterogeneity of mixed drug product intermediate particles has been show to be a critical attribute of solid dosages forms. Chemical and physical uniformity can be assessed readily on a statistically relevant number of individual particles using this approach.
These data sets have demonstrated their effectiveness during product development when applied to basic process understanding and process monitoring during scale-up. Coordinated image analysis and Raman spectroscopy thus provides new types of data to be interrogated during pharmaceutical development. Morphologically targeted chemical characterization of particles adds unique value to the characterization of pharmaceutical powders and drug product intermediates.
References
- Laitinen, N. et. Al. New perspectives for visual characterization of pharmaceutical solids. Journal of Pharmaceutical Sciences, 2004. 93(1): 165-176.
- Randall, Cynthia S. Particle Size Distribution. In: Ed. Brittain, Harry G. Physical Characterization of Pharmaceutical Solids. 1995, Marcel Dekker Inc. New York.
- Warman, M. and Hammond, S. Real time particle characterization during production. European Pharmaceutical Review, 2003. 8(4): 49-54.
- Guerin, E. et. Al. Rheological characterization of pharmaceutical powders using tap density, shear cell, and mercury porosimetry. International Journal of Pharmaceutics, 1999. 189: 91-103.
- Muzzio, F.J. et. Al. Sampling and characterization of pharmaceutical and granular blends. International Journal of Pharmaceutics, 2003. 250: 51-64.
- Allen, Terence. Particle Size Measurement, Volume1: Powder Sampling and Particle Size Measurement. 5th Ed. 1997, Chapman and Hall. London.
- Houghton, M.E. and Amidon, G.E. Microscopic characterization of size and shape: An inexpensive and versatile method. Pharmaceutical Research, 1992. 9(7): 856-859.
- Lewis, E.N. Spectrometric Investigation of Heterogeneity. Appl. No. 12/006,677. 4 January 2008.
- Lewis, E.N. Pharmaceutical Mixture Evaluation. Appl. No. 11/265,796. 17 October 2005.
Author Biographies
Justin Pritchard is a scientist at Vertex Pharmaceuticals supporting the implementation of PAT tools during pharmaceutical development. Justin has experience assessing and implementing gap-filling technologies in process analytical technology, physical characterization, and analytical separations. Prior to joining Vertex, he developed analytical methods for Formulation Development department Alkermes and worked in natural product separations at Aphios Corporation. Justin holds a bachelor’s degree in biology from Davidson College.
Martin Warman is a Scientific Fellow at Vertex Pharmaceuticals supporting the use of PAT during QbD. Prior to this he provided PAT consultancy services to the Pharmaceutical industry. He has over 17 years experience working in the field having previously lead the PAT Development Team of Pfizer Global Manufacturing. Before joining Pfizer he worked for Dionex providing application support for liquid and supercritical fluid chromatography, including on-line measurements. He joined Dionex from Shell Research, where he provided analytical support, developing novel environmental analysis methods and technologies in support of Shell’s bio-restoration/bio-remediation projects, and environmental monitoring programs in the North Sea. He has also worked in academia starting his career at the University of Westminster and unusually for an Analytical Chemist, his bachelor’s degree is in Cell Biochemistry and is a qualified microbiologist. This provides extensive, cross industry, experience in application identification, solution specification and delivery, as well as the support of systems during product life-cycle and he has experience in a developing and implementing a wide variety of measurement solutions, from spectroscopic, through chromatographic but including acoustic and particle characterisation.
This article was printed in the May/June 2011 issue of American Pharmaceutical Review - Volume 14, Issue 4. Copyright rests with the publisher. For more information about American Pharmaceutical Review and to read similar articles, visit www.americanpharmaceuticalreview.com and subscribe for free.