email for correspondence: peter.scott@astrazeneca.com
Historically,
pharmaceutical production involves the manufacture of the finished
product, followed by laboratory analysis to verify quality. The
disadvantages associated with this approach are continual process
optimization, recurring manufacturing difficulties, and the possibility
of failed batches. The Food and Drug Administration (FDA) is inviting
discussions throughout the pharmaceutical industry concerning
a new mode of operation, which will address these concerns. This
mode of operation is known as Process Analytical Technology (PAT).
This article provides an overview of Process Analytical Technology
and its application to the pharmaceutical industry. Techniques
and terminology common to these methods will be described to provide
an introduction to PAT. The scope of this article is to introduce
the reader to PAT. It, however, is a wide-ranging subject, which
is expanding rapidly.
Process Analytical Technology
Process Analytical Technologies involve the use of raw material
properties, manufacturing parameters, process monitoring, and
chemometric techniques to produce finished products of acceptable
quality. The central point of PAT is to generate product quality
information in real-time. The advantages of PAT are many and varied.
While process monitoring traditionally involved temperature, pressure,
flows, pH and other physical parameters, PAT focusses on the use
of in-line testing using near infrared, Raman, or other physiochemical
techniques as a primary means of process monitoring. The data
retrieved would provide information on the properties of blends,
cores, and other stages in the process. Through the use of probes
in the process, uniformity, drying, and mixing endpoints, and
other targeted stages can be pinpointed to a high degree of certainty.
Sampling error would be minimized with in-line probes placed strategically
through out the production process.
The first step away from off-line testing (laboratory separated
from the production plant), would be at-line testing. This is
the movement of process dedicated testing equipment to the production
line to provide rapid results. One advantage is elimination of
the transfer of samples involving time delays. Along with traditional
tests such as dissolution, assay, friability, hardness, and thickness,
this could also include accelerated dissolution rate analysis,
and NIR tablet analyzers. One approach of process analytical chemistry
is on-line testing, which either draws samples or monitors periodically.
Another mode is known as in-line testing, which places probes
in constant contact with drug product. The advantage of on/in
line testing is better control of the process. Beyond data such
as blending, or drying, the FDA has proposed creating on/at-line
assurance of dissolution rates using analytical data correlations.
Near infrared (NIR) is one of the techniques that has gained recent
recognition as a means to add on or in-line analysis at the production
level. The near-infrared light does not destroy or react with
samples and is able to penetrate into and through solid samples.
While NIR has gotten most of the attention, PAT is not limited
to NIR but can include many other forms of monitoring, such as
Raman, Mid-IR, acoustic emission signals, and other imaging techniques.
Dissolution is the current primary method for evaluating solid
oral dosage form consistency and similarity. Using PAT, processes
would be under such high control that the dissolution results
could be accurately predicted well before the product is analyzed.
Research on the correlation between dissolution results and measured
process parameters would be performed so that the impact of process,
raw materials, and finished product variables would be understood.
The manufacturing process could be continuously monitored and
adjustments made to ensure that the finished product would meet
the desired specifications. Measurements from these techniques
have already been used successfully to give predictive values
for dissolution, content uniformity, assay, moisture, and hardness.
The data produced by these devices are rich with information that
is highly complex. These correlations must be performed with the
use of chemometrics.
Chemometrics
Previously, manufacturing processes have been treated in a univariate
manner, with single parameters tracked by control charts. However,
the reality is that physio-chemical processes are multivariate
with subtle interactions of variables. Chemometrics is the intersection
of chemistry and the mathematics of large matrices of data. Chemometrics
is complex and requires the use of computers and software to perform
the necessary computations. These techniques reduce large amounts
of data into a few recognizable components without any loss of
data. Two chemometric techniques that have been found to be useful
are Principal Component Analysis (PCA) and Partial Least Squares
Regression (PLS). These techniques are recognized for their ability
to eliminate noise, identify latent variables, and extrapolate
missing data.
PCA is a technique of creating data models of previously produced
and tested batches to verify similarity to newly created batches.
One advantage this technique has over the commonly used f2 metric
is that batches are now compared to a substantial compilation
of batches included in a validated model. Trends could potentially
be identified earlier than with an f2 comparison. This could help
improvement of process consistency after scale up and post approval
changes. Another advantage of PCA is that it can handle the large
amount of data produced by dissolution fiber optic (Dis-FO) techniques
without the need to reduce data points.
PLS is used to correlate data, such as finished product dissolution
results, to raw material, process parameters, and in-line readings.
Variables which affect the dissolution rate can be better understood
and monitored. The effect of scale ups and post approval changes
can be quantified. Critical parameters can be controlled, thereby
creating high quality drug product, less level/stage 2 testing,
and minimal product failure. When out of specification results
do occur, drug products can be better investigated through the
use of PLS to determine which underlying variables contributed
to the failing drug product.
Summary
As can be seen in the above discussion, the use of PAT techniques
can be a huge benefit to those who choose to use the technology.
Process Analytical Technology provides better knowledge of raw
materials, manufacturing parameters and their impact on finished
product quality. This will result in a more robust process, better
products, more uniform dissolution results, and a huge cost savings
for the manufacturer. The challenge that dissolution scientists
face is to become familiar with this next generation of pharmaceutical
testing and its potential applications in pharmaceutical testing.
References:
1. Eriksson, L., Johansson, E., Kettaneh-Wold, N. and Wold, S.,
"Multi- and Megavariate Data Analysis, Principles and Applications."
1st Edition, Umetrics Academy, June 07, 2001, ch 3-4.
2. Brown, S., presented at InCINC'94 "Has the 'Chemometrics
Revolution' ended? Some views on the past, present and future
of chemometrics." Department of Chemistry and Biochemistry,
University of Delaware
3. Wold, S., presented at InCINC'94 "Chemometrics; what do
we mean with it, and what do we want from it?" Institute
of Chemistry, Umea University, Umea, Sweden.
4. Zackrisson, G., Ostling, G., Skagerberg, B., Anfält, T.,
"Accelerated Dissolution Rate Analysis (ACDRA) for controlled
release drugs. Application to Roxiam." Journal of Pharmaceutical
& Biomedical Analysis, Vol. 13, No. 4/5, 377-383, 1995.
5. Adams, E., Maesschalck, R., De Spiegeleer, B., Vander Heyden,
Y., Smeyers-Verbeke, J., Massart, D., "Evaluation of dissolution
profiles using principal component analysis." International
Journal of Pharmaceutics, 212, 41-53, 2001.
6. Adams, E., Walczak, B., Vervaet, C., Risha, P., Massart, D.,
"Principal component analysis of dissolution data with missing
elements." International Journal of Pharmaceutics, 234, 169-178,
2002.
7. Kirsch, J., Drennen, J., "Determination of film coated
tablet parameters by near infrared spectroscopy." Journal
of Pharmaceutical and Biomedical Analysis, 13, 1273-1281, 1995.