BRIMROSE
|
BRIMROSE
|
|
Analysis
Laboratory based methods have served the pharmaceutical industry well. The Role of NIR Spectroscopy in the Measurement of Pharmaceutical Manufacture. By: P.A. Hailey - Pfizer Central Research
Traditional pharmaceutical analysis is focused almost exclusively
on the end products of manufacturing processes. Samples are taken from drug
substance or drug product batches and analyzed in a remote laboratory. The
samples typically go through stages of documentation, sample preparation,
analysis, data analysis and documentation once more, prior to the reporting of
the analytical results. This approach has served the pharmaceutical industry
well but adds significantly to the manufacturing cycle time and does little to
ensure actively the quality of materials as they are manufactured. The
conventional means to demonstrate quality assurance is to increase inspection
levels but this approach could be considered, at best, as a post mortem on the
manufacturing process. Indeed the very nature of the analysis is unlikely to
correlate well with process performance as the sample matrix is destroyed prior
to analysis. The pharmaceutical industry makes measurements on products to
demonstrate that processes are validated and 'in control'. The corollary of
this proposition is that measurements on processes can demonstrate that
products are validated and 'in control'. Process based measurements may offer
the pharmaceutical industry significant quality and cost improvements. What is driving the need for change? That
the pharmaceutical industry is undergoing a period of significant change is
clearly evident. Issues such as the increasing cost and time to bring new
therapies to the market place, high expectation of profitability from the
financial markets and the ever-increasing need to continue and improve the
quality of pharmaceutical products has resulted in a variety of organizational
restructuring and/or mergers. Against this backdrop of a changing business
climate, many organizations are reviewing critically the processes involved in
the research, development and manufacture of pharmaceutical products. The
regulatory environment also presents a further constraint within which changes
must be made. One
major aspect of pharmaceutical manufacture is the requirement to ensure the
quality of materials. The proof of this quality is typically achieved by
testing materials from manufacturing process. Laboratory based methods have
served the pharmaceutical industry well but are often time consuming and add to
the manufacturing cycle time. Changes in analytical philosophy and improvements
in instrumentation could present an opportunity for measurements to be made in
real time to deliver process control. Many other industrial sectors have woken
up to both the economic and quality drivers that necessitate the need to
perform process based measurements. The rapidly changing business and
regulatory climate of the pharmaceutical industry may be about to catalyze a
shift towards process based measurements and NIR spectroscopy could have a
major role in the new testing paradigm. What role does NIR spectroscopy have in the new testing paradigm?
Firstly,
it should be remembered that NIR is but a small region of the electromagnetic
spectrum and offers one possible option for process based analytical
measurements. It is therefore critical that solutions to on-line measurements
should be problem driven rather than technology driven. That said, NIR
spectroscopy does have a significant advantage over some other technologies due
to a vast array of sample presentation options. The absorbances in the NIR
region originate from the fundamental mid-IR absorbances giving rise to
combinations and overtones that are significantly less intense than the
fundamentals. This apparent lack of sensitivity is in fact a distinct advantage
in that it allows analytical measurements to be made without the need to
perform any sample preparation. Additionally, NIR is sensitive to both chemical
and physical effects and as such provides a wealth of information that is
important for measuring process performance. NIR spectroscopy is therefore
particularly well suited to on-line and at-line measurements. Challenges of Introducing NIR
NIR
has a number of challenges, both internally and externally, which must be
overcome to realize the full potential of the technology. Internally there are
a number of philosophical changes required. At its simplest level NIR is a
non-separative technique that typically requires some form of statistical
manipulation of the spectra before useful information can be derived from the
spectral data. The application of these statistical techniques, known as
chemometrics, often presents some difficulty as the overwhelming majority of
the pharmaceutical analytical world is used to dealing with univariate data.
The application of chemometrics to NIR data is almost exclusively multivariate
in nature and consequently the transition from univariate to multivariate data
analysis presents an initial hurdle. Externally, the regulatory authorities are
still developing their understanding and expertise in the field of NIR and
chemometrics and along with industry are struggling with validation issues. The
situation is somewhat hampered by the fact that there are very few academic
experts in the field of the NIR and pharmaceutical analysis. More academic
research is needed to further build upon the confidence in the technique. Most
other spectroscopic techniques are well resourced for fundamental and applied
research, unfortunately NIR is somewhat an orphan technique within academia
circles. This is a situation that industry and instrument vendors should seek
actively to address.
Fiber-optic
probe design is changing and improving but areas such as reaction monitoring
require special attention. The development of probes capable of monitoring
clear liquids through to viscous slurries needs to be demonstrated. Database
transferability is an area that is often quoted as a problem with NIR
methodologies. This however can be a simple exercise depending upon the scope
of the method in question. Simple discrimination between disparate materials
presents a simpler problem than the qualification of closely related materials.
A number of chemometric algorithms are available but largely originating from
the academic world. Developing robust approaches to database transfer founded
on good spectroscopy and sound chemometric principles, is an area that falls
firmly within the remit of instrument vendors.
The
regulatory acceptance of the technique is beginning to grow but this is very
much on a 'case by case' basis and no agreed approach to the fundamental issue
of validation currently exists. Typical NIR validation packages mirror the
widely accepted pharmacopoeial validation guidelines, which are based largely
on separative technologies. Consequently, some of the validation criteria may,
or may not, be applicable to NIR methodologies. Furthermore, NIR methods are
currently required to be supported by reference or primary methods(1).
Ultimately, this may not be necessary or desirable, but in the short term may
deliver the acceptance the industry is seeking. Alongside this, the industry
will need to consider its filing strategy for NIR methods for NCE and
commercial products. In short, the technology has great potential but obstacles
to success still exist.
Application Areas of NIR Spectroscopy in Pharmaceutical manufacture
Pharmaceutical
manufacture is typically a convergent process brining together a number of
components into a final packaged dosage form. Throughout the process,
laboratory based testing is performed to ensure the product is within
specification. Process based measurements are however feasible with NIR
spectroscopy. At-Line Testing of Excipients(1)
At
the commencement of manufacture of a drug product, it is a requirement to
identify the correct material and grade of the pharmaceutical excipients to be
used in the formulation. Testing of the excipients typically involves a
laborious wet chemical identification, which is not really an indicator of the
quality of the material. NIR, being sensitive to physical and chemical
parameters, is an excellent technique for excipient identification.
Typical
NIR spectra for a variety of excipients are shown in Figure 1. A suitable
spectral database can be used to rapidly identify and qualify excipients. NIR
is sensitive to both the physical and chemical characteristics of samples. This
ability to develop a measurement on the 'textual' aspects of the material could
act as an important metric in predicting the process performance of excipients
in manufacturing processes such as blending operations.
Integration
of this 'quality' measurement into 'smart' manufacturing processes could be
used to guarantee successful manufacturing operations by ensuring that the
correct materials of the appropriate quality are used in the manufacture. The
use of the 'textual' information from NIR could prove to be an invaluable
measurement for the point of delivery and point of dispensing testing. In
combination with bar-code readers, weighing stations and electronic batch
documentation a truly smart system can be developed. The concept of a smart
excipient identification system is shown schematically in Figure 2. Blending Operations (2,3)
Often
the next stage in the manufacture of a dosage form is the blending together of
the active component with the excipients to produce a homogeneous blend. This
operation can be performed in a number of ways using a variety of vessels but
for the purposes of discussion, the blending together of non-cohesive powders
will be considered. Typically a vessel is charged with the components of the
formulation and mixed for a given time. At the end of a fixed time period, the
vessel is sampled for analysis, typically HPLC, to determine homogeneity and
potency of the active within a formulation. This temporal approach to blending
does not take into account any quality measurements. Using an on-line approach
to measurement a far greater understanding of the blending process is achieved
and one such approach is shown in Figure 3.
The
NIR spectrophotometer is configured with a fiber-optic probe, which is
interfaced with the blending vessel at the point of rotation. NIR Spectra are
acquired in real-time and using appropriate data pre-processing and chemometric
analysis, blend 'homogeneity' plots are derived.
NIR
spectra collected in real-time from a blending process are shown in Figure 4.
As can be seen, the spectra begin to converge and overlay over each other. The
spectra have been pre-processed and subjected to chemometric analysis to
determine homogeneity of the spectral data.
The
determination of the blending end-pint is shown in Figure 5.
The
approaches to developing control limits are many and varied and can be used to
develop 'smart' manufacturing blenders. A 'smart' blender would be under
software control and would respond to the real-time spectral data. The use of
such a system in the wider context of a clear measurement strategy presents
some interesting opportunities for parametric release.
Quantitative
analyses are also possible, allowing production to proceed directly to the next
stage. One could present an argument that stable blending processes do not
require real-time monitoring but taking this viewpoint would be to miss the
overall objective or real-time measurements; that of parametric release. If,
for example, the blend was monitored in real-time and shown to be within
specification, then the measurement taken during the blending stage could
contribute towards the release of the final drug product on a weight basis
only.
Applications
in drug substance manufacture range from the measurement of physical phenomenon
such as polymorphic conversion, drying and precipitation through to monitoring
covalent bond forming reactions. Monitoring of reaction processes also presents
further advantages in that it could be used to perform 'multiple' measurements
within a single reaction vessel. NIR measurements can also be made on drug
products using reflectance or the transmission spectroscopy although the real
advantages of drug product testing using NIR will probably be best realized as
part of an overall measurement approach.
Finally,
NIR spectroscopy is at an end in itself. The development of NIR methodologies
should not be technology driven but driven by need. To illustrate the point,
the one for one replacement of a mid-IR identity method with a NIR method
offers little advantage to the pharmaceutical analyst. The NIR method may
require no sample preparation and the classification algorithms employed in NIR
are without doubt more rigorous than mid-IF, but it does not product the
quantum gains that the technology really has to offer. Indeed the receiving
location may not be too keen to purchase another expensive instrument when a
perfectly serviceable mid-IR instrument is already available. On the other
hand, if a single measurement could be used to replace many methods, for
example, ID, assay and polymorphic form, then this presents a rather more
attractive option.
NIR
should therefore be seen to offer greater advantages than simply changing from
an old to a new technology within the existing testing paradigm. Smart Manufacture
The
concept of a "smart" manufacturing process is a system or manufacturing
operation responding to analytical data generated in real time. They system
also has an in-built 'artificial intelligence' as decisions are made whether to
continue a manufacturing operation. This concept has been discussed previously.
Dr. Layloff (FDA) predited that: "Processes would be automatic, and
blenders, for instance, controlled chemometrically". The final product
release testing, he suggested, would consist of "basically, just
looking at hardness and apparent particle size, because all the other controls
would assure that every thing was blended properly" (4).
The concept of using parametric release is further discussed in the same
article where Dr. Layloff "envisions a highly automated lab facility
where there would be only a few people employed, 'basically to locate' the
chemometric devices and incoming and in-process materials for analysis."(4).
The
development of parametric release is dependent, clearly defined measurement
strategy. The importance of measurement strategy. The importance of measurement
throughout the process and specifically at critical operations can provide
information on process performance. Measurements from throughout the
manufacturing process begin to build a 'process fingerprint' and the impact of
measurements at the raw material phase may provide predictive information about
how the final drug product may 'process'. These measurements can provide closed
loop feedback of the process and the assurance of the overall product quality
is the summation of all the measurement from throughout the manufacturing
stages. This concept is shown schematically in Figure 6. The goal of achieving
parametric release can only be achieved in the framework of a clear measurement
strategy.
NIR
spectroscopy clearly has a significant role in any measurement strategy. If DR.
Layloff's vision of the future is realized(1), the function of
analytical laboratories will dramatically change over the coming years.
Conclusion and Vision for the Future
NIR
spectroscopy is here and in daily use by a number of pharmaceutical companies
but the true potential of the technology is yet to be fully recognized or
understood. Am opportunity exists to deliver tremendous quality improvements
and cost savings if the technology can be applied appropriately. It requires a
new way of thinking and understanding which will provide both internal and external
challenges. The vision of parametric release may well be assisted by the use of
NIR spectroscopy, indeed the way may already be open to release products based
upon in-process control measurements as describe in the BP, "parametric
release is, in appropriate circumstances, not precluded by the need to comply
with Pharmacopoeia" (5). If NIR technology can be harnessed
within the industry and gain further regulator acceptance, the role of the
pharmaceutical analyst may change significantly over the next decade and NIR
could be the major technique employed throughout the industry. References:
For More Information Please Contact:
process@brimrose.com
|
