Joint Pharmaceutical Analysis Group
What the drug industry thinks about the concept of risk management
The Joint
Pharmaceutical Analysis Group is a focus for the presentation
and discussion of matters of importance to those interested in
pharmaceutical analysis.
The
group’s sponsoring bodies are the Royal Pharmaceutical Society
and the Royal Society of Chemistry. Membership of the group is
open to member of either society.
Pharmacists wishing to join the group should apply in writing,
giving their registration number, to the Secretariat, Joint
Pharmaceutical Analysis
Group, Science Division, Third Floor, Royal Pharmaceutical Society, 1
Lambeth High Street, London SE1 7JN.
A programme of forthcoming scientific
meetings is available from the secretariat. |
The concept of risk management in pharmaceutical manufacturing is not
new. It is the rationale underpinning most current good practice protocols.
However, the disinclination to adopt risk management principles formally
is now widely seen as limiting the development of a structured, cost-effective,
cross-industry, scientific knowledge base.
Many of the tools for risk management
were developed in large engineering projects well before their application
to pharmaceutical processes, said James Kraunsoe of AstraZeneca. For
example, the Ishikawa fish bone analysis was introduced in the 1940s
and applied
to
steel production processes. Fault tree analysis for visually displaying
and evaluating failure paths in a system was used in 1961 in the US Air
Force MinuteMan missile system, and hazard analysis and critical control
point has its roots in food supply chain risk evaluation.
Risk management has been part of process development in AstraZeneca for
many years but was formalised in 2001 by the introduction of the quality
risk assessment concept with its central feature of a risk management
framework. It was initially developed by the PAT (process analytical
technology) Centre
of Excellence in Sweden and is now deployed globally and applied during
the development of about 40 products.
Emphasis is on a structured, harmonised
and documented approach, based on failure mode and effects analysis (FMEA),
a risk assessment technique for systematically identifying potential
failures in a system or a process.
FMEA uses a scoring system for severity (from a slight deviation to a
total product recall), probability (from never to always) and detectability
(from
detection before a unit operation to detection only by the customer).
An important additional feature is comprehensive documentation.
Dr Kraunsoe recognised the potential pitfalls in the risk management
approach. It can be subjective so that on another day or with a different
group of
operatives there may be a different outcome. Nevertheless, experience
suggests that, with the right expertise, the absolute numbers may vary
but the big
picture remains the same.
A proportionate view needs to be maintained
of the regulatory relevance of the risks so that there is a clear understanding
of the difference between what is relevant to the regulatory submission
and what is only a business risk. These may be assessed together, but
must
be prioritised in the submission.
As defined by the International
Conference on Harmonisation, risk management
is a key concept that provides the foundation for science-based decisions
when integrated into quality management systems and other business processes,
said Thorsten Herkert, of AstraZeneca, developing the theme of risk management
in a regulatory submission, including PAT.
The degree of rigour and formality of quality risk management should
be commensurate with the complexity or criticality of the issue being
addressed.
For simple, less critical situations, an informal approach is usually
sufficient; for more complex or critical situations, a more formal approach,
using
recognised tools to conduct and document the quality risk management,
will be beneficial.
The US Food and Drug Administration PAT initiative is designed to do
just that by using an integrated systems approach to regulating pharmaceutical
product quality. The approach is based on science and engineering principles
for assessing and mitigating risks related to poor product and process
quality.
The systems approach is bound up with process understanding. It endeavours
to identify and explain all critical sources of variability, and to verify
process outputs or to predict product parameters. Based on process understanding,
a quality risk analysis (QRA) can be developed to evaluate and mitigate
risks in the process and of final product quality.
A QRA is performed to identify risks and to structure problems (a trouble-shooting
tool). This can also be used to increase process understanding by identifying
knowledge gaps regarding variations in starting materials (physical properties,
which are not picked up by pharmacopoeia tests) and processing conditions.
Real time release — the ability to evaluate and ensure the acceptable
quality of in-process or final product based on process analytical data — is
the logical next step when PAT is rigorously applied. The risk assessment
strategy becomes a core part of the dossier. Tests that are regarded as
essential for good manufacturing processes become important information
on product quality.
The risks need to be classified and controlled in different ways. The
influence of variations should also be controlled with a risk-based concept.
Good
science should always be the focus and there should not be an insistence
on conventional working. Industry needs ongoing debate on these points,
concluded Dr Herkert.
Analytical methodology must be fit for purpose
To understand and control manufacturing processes and products, we need
to ensure that our analytical methodology is fit for purpose, said Phil
Nethercote of Glaxo SmithKline. Methods need to be designed to
quantify target analytes with appropriate accuracy and precision while
being robust,
rugged and efficient to operate.
An approach for applying quality by design
(QbD) principles to the design and evaluation of analytical methods has,
therefore, been developed within GSK, he said.
This approach is based
on concepts similar to those being developed for manufacturing processes;
that is, there is a clear identification of the desired method performance
requirements, risk assessment tools are used to focus on identifying
quality-critical
analytical parameters and statistical approaches, such as design of
experiments (DOE) and measurement system analysis (MSA) are adopted to
provide a
thorough understanding of the impact of variability on the method performance.
The analytical methods quality by design approach represents a paradigm
shift in the way methods are developed, validated, transferred and
controlled. It will result in more robust methods designed with the
end user in mind,
a leaner science and risk-based approach to method validation and transfer,
a method change control process based on structured risk assessments
and reference to existing method understanding, and significantly increased
regulatory flexibility in relation to introducing method improvements.
The quality by design approach is the way forward for method life cycle
management, concluded Dr Nethercote.
The selection of the solid form for development is a milestone in the
conversion of a new chemical entity into a drug product, said Robert
Docherty, of
Pfizer. An understanding of the processes of crystallisation and particle
characterisation of a new active pharmaceutical is crucial at the interface
of drug substance manufacturing and drug product processing.
The development
of maraviroc, a chemokine receptor antagonist for use in HIV infection,
was used as a case study to illustrate how the elements of polymorphic
form, formulation composition and particle size were considered to
show the importance of a fundamental understanding of the material
science
of a new chemical entity in the context of the new emerging quality
by design
regulatory environment.
Maraviroc is highly soluble across the physiologically relevant pH
range, with a classification of III (high solubility/low permeability)
in the
biological classification system. It is chemically and physically stable
with a well characterised and defined polymorph, with excellent stability
with common excipients, giving rapid disintegration in its preferred
excipient combination.
Company knowledge has shown that a microcrystalline
cellulose/dicalcium
phosphate diluent combination is a robust formulation platform for
a dry granulation process, a preferred process within Pfizer, which
has
a wealth
of experience with associated unit operations and understanding of
the performance of the selected excipients in this process. Dr Docherty
concluded
that, because particle size is a non-critical attribute of the drug
substance, crystallisation parameters could also be designated as non-critical.
It is vital to understand the matrix contributions that can compromise
solid dosage form performance, said Roger Hutton,
of GSK, who described the value of terahertz imaging for the non-destructive
testing of solid
dosage forms. The terahertz spectrum lies between the microwave range
and the infrared range. Pulsed imaging is used to detect differences
in refractive
indices and density changes in solids and hence has applications in
three-dimensional analysis of tablets — coatings, water content,
mechanical properties.
Thus this non-destructive technique is useful
in assessing risks associated with tablet manufacture and prediction
of tablet properties. For example,
terahertz spectroscopy can be used to characterise enteric and cosmetic
coatings and predict where failure might occur. Care, however, is
needed in interpretation and results must be compared with other techniques,
such as X-ray microtomography.
Generally, terahertz information is confirmed by X-ray analysis.
However, on occasions there are inconsistencies between the data
received from
each instrument.
Discrepancies may be due to cracks on the tablet
being suffered
in the time between the analyses, but it should also be remembered
that the X-ray method allows a high spatial resolution with a low
sensitivity to density changes, whereas terahertz analysis with
low spatial resolution
has a high sensitivity to density changes.
Data from terahertz analysis helps us understand critical parameters
in risk assessment, concluded Dr Hutton.
Continuous verification approaches are vital
Sampling is pivotal to the integrity and validity of the analytical measurement,
said Dave Rudd, of GlaxoSmithKline, presenting an alternative
view of process analytical technology (PAT). Unawareness of this key consideration
may
seriously compromise any decision-making process.
Using some everyday examples,
Dr Rudd demonstrated that we must learn not to make assumptions about
the distribution of analytical measurements, to use a valid sample size,
to
appreciate the dynamic nature of the population from which our samples
are taken, and to remember the population may not be homogeneous.
For
some, the assessment of risk is all about the balance between the investment
and the return on that investment, he said. For others, it is about
the probability of failure and the consequence of that failure.
Within the context of quality by design, a risk-based approach demands
that emphasis is placed on those activities and aspects of product
and process development that are likely to have greatest impact on
final
product quality — that is, the product performance as experienced
by the patient.
As an example, he showed how a rational scientific approach
to
reduction of moisture (or solvent) content to a pre-determined minimum
could be applied. In such a procedure there is a need to avoid chemical
degradation because this could occur if the drying process is carried
out too aggressively and to avoid adverse changes to physical aspects
of the
particles (such as particle attrition), which could occur if the drying
process is carried out for too long.
It was shown that rather than
measure the properties of the material during a drying stage, it was
more useful
(and, incidentally, cheaper and simpler) to monitor and change the
conditions for the process (air flow rate, drier temperature, duration
of second phase
of drying process). In effect, the focus is away from the product and
towards the process.
It is vital that we develop continuous verification approaches to ensure
(and to demonstrate) that the information we are obtaining, and the
decisions we are making, are founded on a sound scientific basis, Dr
Rudd said.
Chemometrics is the science of relating measurements made on a chemical
system or process to the state of the system via application of mathematical
or statistical methods. Many users of chemometrics are currently not
completely satisfied with, or are even suspicious of, the state of
the art, and not
without reason, said Ralph Marbach, of VTT Optical
Instruments Centre, Oulu, Finland, in a controversial presentation
pointing out that we
must ensure the asset that is chemometrics does not become a liability.
The
conclusions apply to all areas of spectroscopy, but near-infrared spectroscopy
is in particular need of an overhaul of practices. The uncontrolled
application of so-called statistical calibration methods such as principal
component
regression (PCR) and partial least squares regression (PLS) must stop,
he said.
Suppliers of commercial chemometrics software should enable
the user to plot the response spectrum that is implicitly used as the
signal
in their calibrations, and scientific journals should no longer accept
manuscripts that do not show a figure of the response spectrum used.
The mathematical definitions of sensitivity and specificity in the
multivariate case were discussed. Both limits are testable from first
principles — that
is, from measurable pieces of spectroscopic data. Applying the limits post
factum to today’s statistical calibration practice reveals why many
PLS and PCR results are affected by unspecific correlations (which, unlike
spurious correlations, do not disappear when applying PLS or PCR to larger
and larger calibration data sets).
Current practices and standards of testing for specificity were claimed
to be insufficient and misleading in the case of PLS and PCR, and need
to be amended. Taking control over both is the only way to prove specificity
from spectroscopic first principles and to trade off specificity and
sensitivity in an effective and user-controlled way. There is an urgent
need, however,
for the spectrometric community to engage in discussion.
The role of chemometricians will grow in future. The focus will shift
back to spectroscopy and chemistry with a new job description of “responsible
application scientist” being established. With sensible application
of chemometrics the best days of spectrometry are still ahead, said Dr
Marbach. |
The
Pharmaceutical Journal