Professional Documents
Culture Documents
(Download PDF) Pharmaceutical Quality by Design A Practical Approach First Edition Gibson Online Ebook All Chapter PDF
(Download PDF) Pharmaceutical Quality by Design A Practical Approach First Edition Gibson Online Ebook All Chapter PDF
https://textbookfull.com/product/handbook-of-pharmaceutical-wet-
granulation-theory-and-practice-in-a-quality-by-design-
paradigm-1st-edition-ajit-s-narang/
https://textbookfull.com/product/pharmaceutical-drug-product-
development-and-process-optimization-effective-use-of-quality-by-
design-1st-edition-sarwar-beg-editor/
https://textbookfull.com/product/pharmaceutical-preformulation-
and-formulation-a-practical-guide-from-candidate-drug-selection-
to-commercial-dosage-form-mark-gibson/
https://textbookfull.com/product/quality-assurance-and-quality-
control-in-the-analytical-chemical-laboratory-a-practical-
approach-second-edition-konieczka/
Quality by Experimental Design Fourth Edition Thomas B.
Barker
https://textbookfull.com/product/quality-by-experimental-design-
fourth-edition-thomas-b-barker/
https://textbookfull.com/product/reinforced-concrete-design-a-
practical-approach-chapter-12-3rd-edition-svetlana-brzev/
https://textbookfull.com/product/javascript-application-design-a-
build-first-approach-1st-edition-nicolas-bevacqua/
https://textbookfull.com/product/water-distribution-system-
monitoring-a-practical-approach-for-evaluating-drinking-water-
quality-second-edition-abigail-f-cantor/
https://textbookfull.com/product/statistics-for-process-control-
engineers-a-practical-approach-first-edition-myke-king/
Pharmaceutical Quality by Design:
A Practical Approach
ADVANCES IN PHARMACEUTICAL TECHNOLOGY
A Wiley Book Series
Series Editors:
Dennis Douroumis, University of Greenwich, UK
Alfred Fahr, Friedrich–Schiller University of Jena, Germany
Jürgen Siepmann, University of Lille, France
Martin Snowden, University of Greenwich, UK
Vladimir Torchilin, Northeastern University, USA
Forthcoming Titles:
In Vitro Drug Release Testing of Special Dosage Forms
Edited by Nikoletta Fotaki and Sandra Klein
Characterization of Micro‐ and Nanosystems
Edited by Leena Peltonen
Therapeutic Dressings and Wound Healing Applications
Edited by Joshua Boateng
Process Analytics for Pharmaceuticals
Edited by Thomas de Beer, Jukka Rantanen and Clare Strachan
Pharmaceutical Quality
by Design
A Practical Approach
Edited by
WALKIRIA S. SCHLINDWEIN
De Montfort University
Leicester
United Kingdom
MARK GIBSON
A M PharmaServices Ltd
United Kingdom
This edition first published 2018
© 2018 John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by
any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain
permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Walkiria S. Schlindwein and Mark Gibson to be identified as the authors of the editorial material in this work has
been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
111 River Street, Hoboken, NJ 07030, USA
350 Main Street, Malden, MA 02148‐5020, USA
9600 Garsington Road, Oxford, OX4 2DQ, UK
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Boschstr. 12, 69469 Weinheim, Germany
1 Fusionopolis Walk, #07‐01 Solaris South Tower, Singapore 138628
For details of our global editorial offices, customer services, and more information about Wiley products, visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard
print versions of this book may not be available in other formats.
10 9 8 7 6 5 4 3 2 1
Contents
Index 335
List of Figures
Figure 7.22 Prediction profiler for the 10 run fractional factorial. 192
Figure 7.23 Analysis of the central composite design. 194
Figure 7.24 Analysis of the 13 run DSD experiment. 195
Figure 7.25 Comparison of the central composite and definitive screening
design (DSD). 196
Figure 8.1 Notation used in principal components analysis (PCA). 202
Figure 8.2 PCA derives a model that fits the data as well as possible in the least
squares sense. Alternatively, PCA may be understood as maximizing
the variance of the projection coordinates. 203
Figure 8.3 The raw data curves of the 45 batches of the training set. 205
Figure 8.4 Scatter plot of the two principal components. Each point represents
one batch of raw material. The plot is color‐coded according
to supplier. 205
Figure 8.5 Loading line plots of the two principal components. 206
Figure 8.6 Scores and DModX plots for the particle size distribution data set.
Top left: Predicted scores for the test set batches. Top right: Scores
for the training set batches. Bottom left: Predicted DModX for
the test set batches. Bottom right: DModX for the training set batches. 207
Figure 8.7 Line plot of the power spectral density (PSD) curves of the six early
(red color) and three late (black color) batches of supplier L3. 208
Figure 8.8 Spanning batches can be selected using DoE. Such spanning batches
can be subjected to a thorough investigation in order to ensure
production robustness in all part of the PCA score space. 209
Figure 8.9 (Left) Scatter plot of HS_1 against HS_2, (right) PLS t1/t2 score
plot (note the resemblance to Figure 8.8). 210
Figure 8.10 (Left) Summary of fit plot of the PLS model of the SOVRING subset
with complete Y‐data. Five components were significant according to
cross‐validation. (Right) Individual R2Y‐ and Q2Y‐values of the six
responses. The most important responses are PAR, FAR, %Fe_FAR
and %P_FAR. 211
Figure 8.11 (Left) PLS w*c1/w*c2 loading weight plot. Ton_In is the most influential
factor for PAR and FAR. The model indicates that by increasing Ton_In,
the amounts of PAR and FAR increase. There is a significant quadratic
influence of HS_2 on the quality responses %Fe_FAR and %P_FAR.
This nonlinear dependence is better interpreted in a response contour
plot, or a response surface plot. (Right) PLS w*c3/w*c4 weight plot. 212
Figure 8.12 (Left) Response contour plot of PAR, where the influences of
Ton_In and HS_2 are seen. (Right) Response contour plot of FAR. 212
Figure 8.13 (Left) Response contour plot of %P_FAR. (Right) Response contour
plot of %Fe_FAR. 213
Figure 8.14 SweetSpot plot of the SOVRING example, suggesting the SweetSpot
should be located in the upper and right‐hand part. 214
Figure 8.15 A design space plot of the SOVRING example. The green area
corresponds to design space with a low risk, 1%, of failure. 214
Figure 8.16 A schematic representation of how the concepts knowledge space, design
space, and normal operation (control space) region relate to one another. 215
xvi List of Figures
Figure 8.17 The design space hypercube represents the largest regular geometrical
structure that can be inserted into the irregular design space. 216
Figure 8.18 Batch control chart showing predictions for two bad batches.
The BEM fitted using orthogonal PLS readily detects
the deviating batches. 218
Figure 8.19 Batch control chart showing predictions for two bad batches. The BEM
was fitted using classical PLS. 219
Figure 8.20 Contribution plot for batch DoE7 at 1.7 min showing which variables
are contributing to the process deviation. The process upset is caused
by a few of the process parameters. For example, the contribution plot
indicates that the reaction temperature (highlighted by red color)
is much higher than for a normal batch. 219
Figure 8.21 Line plot of the reaction temperature for the non‐NOC batch DoE7.
At the time point of 1.7 min, the reaction temperature is almost 15°
higher than the average temperature across the six NOC batches.
The deviation from normality of the reaction temperature for DoE7
increases further into the lifetime of the batch. 220
Figure 9.1 Elements of a PAT system (each element is further described
in Table 9.1). 230
Figure 9.2 Examples of PAT applications used for an oral solid dosage process
and benefits. 241
Figure 9.3 PAT applications used during processing in lieu of conventional
end‐product QC testing. 244
Figure 9.4 Elements to consider for inclusion in a PAT application, whether it is
for gaining process understanding or for process control. 246
Figure 9.5 Manufacturing process based on hot‐melt extrusion showing the main
unit operations from powders to final product, tablets. 248
Figure 9.6 Preliminary risk assessment (RA) analysis (Ishikawa diagram).
The factors highlighted were considered critical for the
extrusion process. 249
Figure 9.7 Representation of about 5000 UV‐Vis spectra from DoEs 1–3
and verification experiments. 250
Figure 9.8 Raw spectra from a DoE set of runs. 250
Figure 9.9 UV‐Vis spectra from the first screening design (DoE1) showing
a sample of the spectra collected. 251
Figure 10.1 Enhanced QbD lifecycle approach to analytical methods. 258
Figure 10.2 Example of the impact of method variability on overall variability
for drug product assay (LSL=95%LC and USL=105%LC). 261
Figure 10.3 Science‐ and risk‐based approach to analytical method design,
development and lifecycle management. 262
Figure 10.4 Ishikawa (fishbone) diagram for a drug product chromatographic
assay method. 265
Figure 10.5 Gage repeatability and reproducibility study. 266
Figure 10.6 I‐MR chart of drug product assay in manufacturing order –
produced using Minitab® Statistical Software. 269
Figure 10.7 Science and risk assessment process. 270
List of Figures xvii
Figure 10.8 Centred and scaled coefficients for the resolution and separation
factor models. 273
Figure 10.9 Alternative statistical approach – main effect plot amount of TFA
in the mobile phase. 275
Figure 10.10 Centred and scaled coefficients for the signal‐to‐noise model. 276
Figure 10.11 Dissolution data from multivariate experimental study (coloured
by analyst) – individual tablet (open circles) and mean data
(solid circles) – produced using Minitab® Statistical Software. 277
Figure 10.12 Main effects plot from multivariate experimental study – produced
using Minitab® Statistical Software. 278
Figure 11.1 The “V” model concept of validation. 283
Figure 11.2 The ASTM E2500 system lifecycle and validation approach. 285
Figure 11.3 Sequence of activities for formulation, process design, and
optimization incorporating process validation activities according
to the lifecycle approach. 296
Figure 11.4 Stages of process validation showing potential changes. 300
Figure 11.5 The elements of a simple, single‐loop control system for
a liquid process. LT = sensor; LC = controller; LV = actuator. 304
Figure 11.6 Control strategy and design space for an IR tablet. 307
Figure 11.7 Typical batch process flow for a granulated tablet. 309
Figure 11.8 ConsiGma™ continuous tablet production line. 314
Figure 12.1 Summary of the intrinsic value of applying QbD. 322
Figure 12.2 The CTD triangle. 323
Figure 12.3 A typical example of an Ishikawa or fishbone diagram. Key: red =
potential high impact on CQA; Yellow = potential impact on CQA;
green = unlikely to have significant impact on CQA. Example:
roller compaction. 326
Figure 12.4 Types of post‐approval changes available in the EU relative to the
adopted level of risk during the evaluation of the procedure. 330
4 Pharmaceutical Quality by Design
ICH produces guidelines under headings of Safety, Quality and Efficacy. It has eminent
and broad‐ranging groups of experts involved in producing these guidelines, so these
guidances are as near to global as one can obtain, even though they are neither global nor
mandatory, unless – as happens in some cases – regulatory agencies include them in their
national Good Manufacturing Practices (GMPs). Most can be considered, to all intents and
purposes, as internationally applicable.
It was ICH that first brought the term QbD to the pharmaceutical industry when in 2009
it published ICH Q8 (R2) [1], ‘Pharmaceutical Development’.
This was a watershed moment for the industry, as, after this, the importance of taking a
science‐ and risk‐based approach moved to front stage, and even terms like ‘manufacturing
science’ began to be heard.
Science, of course, has always been a fundamental element of the development of pharma-
ceuticals and, historically, innovative application of science has been core to producing the
many life‐saving and life‐enhancing drugs that the industry has produced over time.
So why all this supposedly new approach? What has changed?
Well, the fundamentals driving the need to understand pharmaceutical science remain
the same, but perhaps the following are factors that influenced a change in perspective:
●● The application of science is becoming more complex; for example, biotechnology‐
based drugs are more complicated to understand, make and analyse compared to small
molecules; specialised therapies such as advanced therapeutic medicinal products, gene
therapies, etc. are beginning to emerge.
●● The use and application of more sophisticated tools, for example, process analytical
technology (PAT), [6] has become more commonplace – although this tool is relatively
new for the pharmaceutical industry, it has been in use by other industries for many years.
●● More powerful data processing is now available to enable such tools to be used. An exam-
ple is design of experiments (DoE) (see Chapter 7) and multivariate analysis (MVA) (see
Chapter 8) can now be used for more sophisticated analysis than was possible previously.
●● The industry has become more global, often with many differing countries involved in
the supply chain, which has made it necessary to maintain quality across various interna-
tional boundaries and cultures.
●● The supply chain has become more fragmented and diverse, with many more parties
involved, including contract research organisations (CROs), contract manufacturing
organisations (CMOs) and external suppliers. ‘Virtual’ companies are now emerging, a
role that did not exist a few years ago.
●● Internal organisations are being re‐structured. An ‘over the wall’ attitude for technology
transfer, development and manufacturing is being heavily discouraged. Business benefits
are being seen in having closer working internal partnerships.
Another random document with
no related content on Scribd:
Reference has already been made to Aurispa, who appears to
have been the most important manuscript-dealer of his time, not only
in Venice, but possibly in the world. Aurispa sent various agents to
Greece and to the farther East to collect manuscripts and kept
scribes busied in his work-shop in Venice in preparing authentic
copies of these texts. One of his travellers was Plantinerus, who was
sent to the Peloponnesus in 1415, and who succeeded in securing
there some valuable codices.[343] Plantinerus found, in executing his
commissions, that he had to come into competition with a traveller
sent out by Cosimo de’ Medici on a similar errand.
Venice possessed an advantage over the other Italian cities, not
only in the collection of texts, and in its facilities for manifolding
these, but in its position for securing wide sales for the same in the
cities outside of Italy, with which it was, in connection with its active
commerce, in regular relations. The lines of the Oxford printers,
Theo. Rood and Thomas Hunt, printed in their edition of the Letters
of Phalaris, give an indication of the relations of the English
university in the early part of the fifteenth century with the literary
marts of Southern Europe.
(If you Venetians will send over to us the books which have been
hidden (i. e. difficult or rare books, or possibly books unearthed from
far off Eastern regions) we will find sale for the same.)
There is evidence in fact of a very active book-trade between
Venice and England for many years before the introduction into Italy
of the printing-press. The work of Aldus and of those who were
associated with him in carrying on printing and publishing
undertakings in Venice naturally very largely extended these
relations with the English scholars, but the channels for the same
had already been opened. The manuscript-dealers in Venice fixed
their place of business in the most frequented parts of the city—the
Bridge of the Rialto, and the Plaza of S. Mark.
The trade of the Italian dealers in manuscripts was not brought to
an immediate close by the introduction of printing. The older scholars
still preferred the manuscript form for their books, and found it
difficult to divest themselves of the impression that the less costly
printed volumes were suited only for the requirements of the vulgar
herd. There are even, as Kirchhoff points out,[345] instances of
scribes preparing their manuscripts from printed “copy,” and there
are examples of these manuscript copies of printed books being
made with such literalness as to include the imprint of the printer.
The work of Aldus (continued with scholarly enterprise later by
such men as Froben of Basel and Estienne of Paris) in the printing of
Greek texts, although begun as early as 1495, and although
exercising a very wide influence upon the distribution of Greek
literature, was insufficient to supply the eager demand of the
scholars, while not many other printers were, in the early years of the
exercise of the art, prepared to incur the very considerable risk and
expense required for the production of Greek fonts of type. The risk
was, of course, by no means limited to the cost of the type; the
printers of the earlier Greek books had themselves but slight
familiarity with the literature of Greece, and they were obliged in
many cases to confide the selection and the editing of their texts to
editors to whom this literature was very largely still a novelty. The
printers hardly knew what books to select and they had no adequate
data upon which to base business calculations as to the extent of the
demand that could be looked for for any particular book. The feeling
that they were working in the dark was, therefore, a very natural one.
It was on this ground that, while printing-presses were, during the
century after 1450, multiplying rapidly through Europe, the printing of
Greek books continued to be for a large portion of the period an
exceptional class of undertakings, and work was still found for
scribes who could be trusted to make accurate transcripts of Greek
codices.
Kirchhoff gives the names of the following Italian manuscript-
dealers and scribes whose scholarly activity during the latter half of
the fifteenth century was especially important: Antonius Dazilas,
Cæsar Strategus, Constantius Librarius, Andreas Vergetius, and
Antonius Eparchus. The latter made various journeys to the East in
search of manuscripts. The fact that the dealers in manuscripts very
rarely placed their own names on the copies of the texts sent out
from their work-shops has, in a large number of cases, prevented
these names from being preserved for future record. The names that
have come into record are in the main such as have been referred to
in the correspondence of their scholarly friends and clients. I quote a
few of these references from the lists given by Kirchhoff:
In Bologna the oldest librarius whose work is referred to is Viliaric,
who was called an antiquarius, and whose shop was open in the
beginning of the thirteenth century. In a manuscript, previously
referred to, containing a treatise of Paul Orosius, originally written in
the seventh century, and from which this copy was transcribed early
in the thirteenth, there is at the end an inscription, as follows: