Application of the QbD principles in the development of the casopitant mesylate manufacturing process. Process research studies for the definition of the control strategy of some drug substance-CQAs for stages 2a, 2b, and 2c

Article abstract of DOI:10.1021/op1000622

Casopitant was identified as a potent NK₁ antagonist by GlaxoSmithKline (GSK). It was selected as part of a wide drug discovery programme within GSK for its potential activities on a number of therapeutic targets such as inflammatory bowel dise

Full text of DOI:10.1021/op1000622

Organic Process Research & Development 2010, 14, 805–814
Application of the QbD Principles in the Development of the Casopitant Mesylate
Manufacturing Process. Process Research Studies for the Definition of the Control
Strategy of some Drug Substance-CQAs for Stages 2a, 2b, and 2c
Zadeo Cimarosti,*,† Fernando Bravo,† Damiano Castoldi,† Francesco Tinazzi,† Stefano Provera,‡ Alcide Perboni,†
Damiano Papini,† and Pieter Westerduin†
Chemical DeVelopment, Medicines Research Center, GlaxoSmithKline, Verona, Italy, and Molecular DiscoVery Research,
Medicines Research Center, GlaxoSmithKline, Verona, Italy
Abstract:
assessment tools is therefore key to ensuring that the quality of
the drug substance or drug product is appropriate and consistent.
Regulatory agencies2 fully support this approach and encour-
age its adoption during the development of drug substance and
drug product manufacturing processes. More details on this
approach as applied by GSK have been recently reported in a
previous paper,3 where a detailed description of the elements
of control (the attributes of the input materials, the process
parameters, and the procedure) is also given.
Casopitant was identified as a potent NK1 antagonist by Glaxo-
SmithKline (GSK). It was selected as part of a wide drug
discovery programme within GSK for its potential activities
on a number of therapeutic targets such as inflammatory
bowel disease, overactive bladder, CNS disorders, and
others. The mesylate salt of casopitant was selected for full
development. The manufacturing process to casopitant
mesylate was developed and optimised by following a
Quality by Design approach, whereby a control strategy was
developed, underpinned by process understanding and risk
analysis, for an enhanced level of quality assurance. Quality
process parameters and specifications levels for the Stages
2a, 2b, and 2c are the elements of the control strategy of
the manufacturing process discussed in detail in this paper.
The Design of Experiment approach has been extensively
used to support the definition of the proven acceptable
ranges for the process. The aim is to show the process
development studies carried out to ensure quality control
for the final drug substance.
A QbD approach has recently been applied to the develop-
ment of the manufacturing process for casopitant mesylate 1, a
potent neurokinin receptor (NK1) antagonist.
In this contribution, some of the process understanding
studies carried out on Stages 2a, 2b, and 2c for the definition
of the control strategy are described. The elements of control
discussed are the quality process parameters (the parameters
that have an impact on drug substance-CQAs) and the speci-
fications levels for starting materials and one intermediate of
the manufacturing process. The Design of Experiment (DoE)
approach has been extensively used to support the development
work in all the stages and in particular in the definition of the
proven acceptable ranges (PARs) (the upper and/or lower limits
for process parameter between which the parameter is known
to produce a process output that meets the CQAs) for Stage
2a.
1. Introduction
In the past years, a number of regulatory guidelines (ICH
Q8, ICH Q9, and ICH Q10)1 have been issued, describing a
new approach to process development where the quality is built-
in rather than tested in the product. This approach is called
“Quality by Design” (QbD).
For the reader’s benefit, a Glossary with the definitions of
the terms used within this text is included.
These guidelines are focused on different aspects of QbD.1
For example, ICH Q8 describes an enhanced approach by the
use of process understanding. ICH Q9 describes the risk
management tools that can be used to successfully manage the
risk, and ICH Q10 introduced the concept of a control strategy,
defined as a set of controls, derived from current product and
process understanding that assures process performance and
obtaining drug substance that meets the critical quality attributes
(drug substance-CQAs, the measurable properties that are
critical to ensuring patient safety and efficacy).
2. Synthetic Route
The commercial process to synthesise casopitant me-
sylate 1, is a multistage convergent process, summarised
in Scheme 1.
In Stage 1, the dihydropyridone 10 is converted into the
piperidone 8 by reduction of the double bond and hydrogenoly-
sis of the carbobenzyloxy protecting group. The piperidone 8
is resolved Via dynamic kinetic resolution (DKR) using L-
mandelic acid to yield the (R)-piperidone as the L-mandelate
salt 7.
The development of a robust control strategy supported by
process understanding and by using the appropriate risk
In Stage 2, the (R)-amine 5 is converted into the carbamoyl
chloride 4 by reaction with carbon dioxide and thionyl chloride.
* To
whom
correspondence
should
be
addressed.
E-mail:
(2) See for example: Pharmaceutical cGMPs for the 21st century - A risk
based approach, Final Report; U.S. Department of Health and Human
Services, Food and Drug Administration, Center for Drug Evaluation
and Research (CDER): Silver Spring, MD, U.S.A., 2004.
zadeo.a.cimarosti@gsk.com.
† Chemical Development.
‡ Molecular Discovery Research.
(1) ICH Q8 Pharmaceutical DeVelopment; ICH Q9 Quality Risk Manage-
(3) Cimarosti, Z.; Bravo, F.; Stonestreet, P.; Tinazzi, F.; Vecchi, O.;
Camurri, G. Org. Process Res. DeV. ASAP - DOI: 10.1021/op900242x.
ment; ICHQ 10 Pharmaceutical Quality System.
10.1021/op1000622  2010 American Chemical Society
Published on Web 05/25/2010
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
805

Scheme 1. Commercial route for casopitant mesylatea
a Stage 1a: ethyl acetate, Rh/C, hydrogen. Stage 1b: ethyl acetate, Pd/C, hydrogen. Stage 1c: i) 2-propanol, water, L-mandelic acid; ii) 7 seed; iii) 2-propanol,
cyclohexane. Stage 2e: i) ethyl acetate, carbon dioxide, triethylamine; ii) chloro(trimethyl)silane; iii) pyridine; iv) thionyl chloride; v) malic acid (aq); vi) water; vii)
Na2CO3 (aq) or K2HPO4 (aq); viii) ethyl acetate. Stage 2a: i) ethyl acetate; ii) Na2CO3 (aq); iii) NaCl (aq) or water; iv) ethyl acetate. Stage 2b: i) triethylamine; ii)
1-acetylpiperazine; iii) malic acid (aq); iv) Na2CO3 (aq); v) NaCl (aq) or water; vi) acetonitrile. Stage 2c: i) acetonitrile, 1-acetylpiperazine, NaBH(OAc)3/HCOOH;
ii) ethyl acetate; iii) NH4OH (aq); iv) Na2CO3 (aq); v) NaCl (aq) or water or K2HPO4 (aq); vi) ethyl acetate. Stage 2d: i) ethyl acetate; ii) acetone; iii) methanesulfonic
acid; iv) casopitant mesylate seed; v) isooctane; vi) ethyl acetate.
Scheme 2. Formation of casopitant stereoisomers
Scheme 2) in a ratio of approximately 2:1. Casopitant
mesylate 1 is obtained after a seeded, selective precipita-
tion by addition of methanesulfonic acid in a mixture of
ethyl acetate, acetone, and isooctane.
In this contribution the process research studies carried out
for the definition of the control strategy for Stages 2a, 2b, and
2c of some drug substance-CQAs are reported.
3. Drug Substance-CQAs Discussed for Stages 2a, 2b, and
2c
The drug substance-CQAs discussed in this paper in the
context of Stages 2a, 2b, and 2c are impurities with the potential
to contaminate the drug substance, and they are the stereoiso-
mers of the casopitant and the impurities coming from the
contaminants of the starting material 1-acetylpiperazine. The
structures and the rationale for their formation are reported
below.
3.1. Casopitant Stereoisomers. The synthetic route intro-
duces three stereogenic centres, so that in principle a total of
eight stereoisomers can be formed. In addition to casopitant
mesylate 1 and its enantiomer, there are six other isomers: three
diastereoisomers and each diastereoisomer can exist as an
enantiomeric pair. The formation of each individual stereoiso-
mer is summarised in Scheme 2 and Table 1; all of them might
be present in the drug substance. For this reason all of them
are drug substance-CQAs, and their control has to be ensured
according to the ICH guideline on impurities.
Acylation of the (R)-piperidone 6 (obtained by basification of
its corresponding L-mandelate salt 7) with an excess of the
carbamoyl chloride 4 yields the piperidone-urea 3, followed by
addition of l-acetylpiperazine to react with the excess of
carbamoyl chloride. The piperidone-urea 3 is subjected
to reductive amination with additional 1-acetylpiperazine
in the presence of the reducing system sodium triacetoxy-
borohydride [NaBH(OAc)3]/formic acid to afford a mix-
ture of casopitant 2 and its anti-isomer 11 (showed in
Considering Stages 2a, 2b, and 2c that are discussed in this
contribution, the sources of these stereoisomers are in the
806
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Table 1. Formation of casopitant stereoisomers
It is worth noting that the compounds belonging to the
following pairs 11 and 15, 16 and 18, 17 and 19 and casopitant
and 14 are enantiomers.
3.2. 1-Acetylpiperazine Impurities. Process understanding
studies highlighted also that one of the reagents, 1-acetylpip-
erazine, could contain piperazine, 1-propanoylpiperazine, and
N-(2-aminoethyl)acetamide as impurities. Process understanding
studies showed that these impurities could react in Stage 2c
chemistry by generating three impurities with the potential to
contaminate the drug substance (compounds 20, 21, and 22).
These potential impurities were defined drug substance-CQAs.
Details of their formation are given in Scheme 3 and Table 2.
4. Setting of the Appropriate Specification Limits for the
Key Input Materials of Stages 2a, 2b, and 2c
The experiments that allowed the definition of the specifica-
tion for the (R)-piperidone 6, the (R)-amine 5, and 1-acetylpip-
erazine are reported in this section. These data in combination
with the chemical data reported in the following sections
allowed the definition of the control strategy for Stages 2a, 2b,
and 2c with respect to the drug substance-CQA to be introduced.
4.1. (R)-Piperidone 6 and (R)-Amine 5. These input
materials are considered together as they are both linked with
the drug substance-CQA reported in Scheme 2 (casopitant
stereoisomers 11, 14, 15, 16, 17, 18, and 19). Thus, a spiking
study by using (R)-piperidone mandelate salt 7 [contaminated
with 2% a/a of (S)-piperidone 13], and (R)-amine 5 [contami-
nated with 0.9% a/a of the (S)-amine 12] was carried out. The
final casopitant mesylate was within specification as shown in
Table 3.
a The configuration of the stereocentres according to Scheme 2 is reported.
Scheme 3. Drug substance-CQAs from 1-acetylpiperazine
impurities
Table 3. Spiking experiment for the definition of the
specification limits for 5 and 6
Table 2. Structures of the 1-acetylpiperazine impurities
level of drug
substance-CQA
in the drug
substance
specification
limit of drug
substance-CQA
in the drug substance
(% w/w)
level of
impurity
(% a/a)
impurity
(% w/w)a
(S)-piperidone 13
in (R)-piperidone
salt 7
(S)-amine 12 in
(R)-amine 5
2
16 and 18
0.25
NGTb 0.4
17 and 19 <0.05
NGT 0.1
0.9
11 and 15
14
0.14
<0.05
NGT 0.6
NGT 0.15
a Casopitant stereoisomers 16 and 18, 17 and 19, and 11 and 15 are reported
together as the analytical method was not chiral. 14 is the casopitant enantiomer,
and a dedicated chiral method was used for its detection. b NGT ) Not Greater
Than.
In addition, the process research studies carried out in the
initial phase of the development made it possible to clarify the
following points:
• Epimerisation of the stereogenic centre coming from
the (R)-amine 5 and its derivatives (piperidone-urea 3,
carbamoyl chloride 4 and casopitant 2) do not occur in
the process conditions so, no QCPPs/QPPs are linked
to the level of (S)-amine 12 in (R)-amine 5 and
derivatives.
stereoselectivity of the Stage 2c reaction (the reductive amina-
tion) and in the enantiomer contamination of the (R)-piperidone
mandelate salt 7 and the (R)-amine 5, as summarised in Scheme
2 and Table 1.
• Epimerisation of the stereogenic centre coming from
(R)-piperidone 6 does not occur in:
The same synthetic path is followed also by the enantiomers
of the compounds (R)-amine 5 and (R)-piperidone 6, (S)-amine
12 and (S)-piperidone 13, respectively, leading to the formation
of the other casopitant stereoisomers as detailed in Table 1.
• (R)-piperidone 6 itself in ethyl acetate in presence of
triethylamine in Stage 2b
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
807

Table 4. Spiking of piperazine, 1-propanoylpiperazine and N-(2-aminoethyl)acetamide in 1-acetylpiperazine
level of drug
substance-CQA in
the drug substance
(% w/w)
specification limit of
level of impurity in
1-acetylpiperazine
(% a/a)
drug substance-CQA
in the drug substance
(% w/w)
impurity in
1-acetylpiperazine
drug substance-
CQA affected
piperazine
1-propanoylpiperazine
N-(2-aminoethyl)acetamide
0.29
0.44
0.5
20
21
22
<0.05
0.33
0.34
0.3
0.15
0.15
Scheme 4. (R)-piperidone 6 formation
• piperidone-urea 3 under Stage 2b or 2c conditions
• casopitant 2 under Stage 2c conditions
• Epimerisation has been seen in (R)-piperidone 6, and
it is due to thermal or acidic instability of the molecule
as a consequence of a retro-Michael/Michael mecha-
nism (see Scheme 5). The epimerisation of the (R)-
piperidone 6 is further discussed in section 5.
Scheme 5. Mechanism of racemisation of (R)-piperidone 6
These data from process research studies and the spiking
experiments reported in Table 3 allowed setting the appropriate
specifications for the two input materials considered above. In
particular, limits of 1.5% a/a (S)-piperidone 13 in (R)-piperidone
mandelate salt 7 and 0.7% a/a (S)-amine 12 in (R)-amine 5 have
been set. The limits set on these compounds represent an
element of control over the control strategy.
4.2. 1-Acetylpiperazine. The contaminants of 1-acetylpip-
erazine are described in section 3.2; considering these impurities
and the conditions of Stage 2c, the only approach to ensure
their control was to set appropriate limits of piperazine,
1-propanoylpiperazine, and N-(2-aminoethyl)acetamide in
1-acetylpiperazine.
Spiking experiments have been performed to define the level
of each of these impurities that can be tolerated by the process,
and the results are shown in Table 4. It is worth noting that the
level of these drug substance-CQAs was detected on casopitant
mesylate after Stage 2d. As Stage 2d and its control strategy
will be the subject of a separate discussion, only the general
procedure is reported in this paper (see the Experimental
Section).
7 in ethyl acetate. Sodium carbonate was selected over sodium
bicarbonate for its higher water solubility (to allow more
concentrated solutions) and ethyl acetate was selected as in
common with Stages 2b and 2e.
5.1. Process Understanding Studies. Process studies were
carried out to understand the stability of the (R)-piperidone 6
during the free basing (Scheme 4) as previous knowledge has
highlighted that the stereocentre was not stable chemically and
thermally and racemisation could have occurred.
The mechanism of racemisation is proposed in Scheme 5;
the “opened ring” intermediate was never isolated due to its
transient nature and its reactivity. As stated in the previous
paragraph, this would have had an impact on the further steps
of the synthesis. In particular, the enantiomer of 6 (com-
pound 13) could have reacted in downstream chemistry
to give the diastereoisomers of casopitant reported in
Scheme 2 and Table 1.
An initial risk assessment was carried out, taking into account
the available process knowledge and the reaction mechanism
proposed. Temperature and volumes of ethyl acetate were
identified as potential QCPPs.
On the basis of the data presented in Table 4, and the
specification limits of the corresponding drug substance-CQAs,
the following limits of piperazine, 1-propanoylpiperazine, and
N-(2-aminoethyl)acetamide in 1-acetylpiperazine specification
are proposed:
• 0.3% a/a of piperazine in 1-acetylpiperazine.
• There is approximately a 1:1 ratio between the level of
1-propanoylpiperazine in 1-acetylpiperazine and the
level of 21 in drug substance. Therefore, a limit of
0.10% a/a of 1-propanoylpiperazine in 1-acetylpipera-
zine specification is proposed.
• There is approximately a 1:1 ratio between the level of
N-(2-aminoethyl)acetamide in 1-acetylpiperazine and
the level of 22 in drug substance. Therefore, a limit of
0.15%a/aofN-(2-aminoethyl)acetamideinthe1-acetylpip-
erazine specification is proposed.
These parameters were considered in the context of all the
Stage 2a substeps, in particular:
1. (R)-Piperidone L-mandelate salt 7 was suspended in
ethyl acetate.
2. Aqueous sodium carbonate solution was added, and
the mixture was stirred until dissolution was complete.
3. The two phases were separated.
4. The organic phase was washed with aqueous sodium
chloride solution or water.
5. The two phases were separated.
6. The organic phase was concentrated under vacuum.
5. Process Studies and Control Strategy for Stage 2a
This process step consists of the free basing of the Stage 1
product, this is carried out by adding a solution of sodium
carbonate to a suspension of the (R)-piperidone mandelate salt
Steps 3-5 were not explicitly studied since the chemistry
would not suggest any impact on drug substance-CQAs.
808
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Table 5. Potential QCPPs and PARs for Stage 2a
potential QCPP
PAR
target value
intermediate 7 suspension
temperature
intermediate 6 distillation
temperature
intermediate 6 distillation
endpoint volume
NMTa 25 °C
20 °C
NMTa 25 °C
20 °C
NLTb 2 vol
2.5 vol
a NMT
) Not More Than. No impact on quality from use of lower
temperature. b NLT ) Not Less Than. No impact on quality from use of higher
volume.
Figure 1. Enantiomeric purity of (R)-piperidone L-mandelate
salt 7 in ethyl acetate over time.
5.1.1. Studies on Substep 1. The impact of temperature and
volume of ethyl acetate on the racemisation time course was
investigated using a two-level, full factorial experimental design
with one centre point, comprising five experiments. The
experiments were carried out in the absence of aqueous sodium
carbonate solution. The ranges of these parameters studied were
from 20 to 30 °C for the temperature and 3 to 5 for the volumes
of ethyl acetate.
Data were collected for 48 h. The degree of racemisation
was determined by monitoring the enantiomeric purity of the
(R)-piperidone L-mandelate salt 7 over time. The data are
graphically presented in Figure 1.
Figure 2. Enantiomeric purity of (R)-piperidone 6 in ethyl
acetate.
The rate of racemisation of (R)-piperidone L-mandelate salt
7 in ethyl acetate at 20 °C is very low, as can be seen by the
small extent of racemisation over 48 h, and furthermore is
independent of dilution (volume of ethyl acetate). At 25 °C and
4 volumes, there is an increased rate of racemisation, but (R)-
piperidone L-mandelate salt 7 has acceptable stability for at least
6 h, based on the moderate extent of racemisation at this time
point (the amount of 13 only increased by 0.4% a/a after 5 h at
this temperature). At 30 °C there is a further increase in the
rate of racemisation, resulting in an unacceptable level of (S)-
piperidone 13 at the 6-h time point.
The PAR for temperature of the suspension of (R)-piperidone
L-mandelate salt 7 in ethyl acetate was accordingly set at no
more than (NMT) 25 °C. Note that on a full manufacturing
scale, aqueous sodium carbonate solution is added within a few
hours, well within the time limit for stability. Once the sodium
carbonate solution has been added, (R)-piperidone 6 is stable
in ethyl acetate, giving a very wide margin of control in
manufacture. The PAR for the temperature is summarised in
Table 5.
5.1.2. Studies on Substep 2. Substep 2 was studied by
following the same approach used for Substeps 1 and 6 (see
sections 5.1.1 and 5.1.3, respectively). Initial risk assessment
identified temperature and ethyl acetate volumes as potential
QCPPs; some studies have been carried out (by using the same
ranges used for substep 1), and it was confirmed that the (R)-
piperidone 6 was stable after basic wash. The knowledge from
these experiments, the additional knowledge gained in the
production plant, and performance of representative batches
produced on scale were then used in a failure mode and effects
analysis (FMEA) risk assessment, and the result of the assess-
ment showed that these parameters had no impact on drug
substance-CQA. Thus, these parameters were neither QPPs nor
QCPPs, and they were not considered in the definition of the
control strategy.
5.1.3. Studies on Substep 6. (R)-Piperidone 6 solution is
subjected to a vacuum distillation to azeotropically remove water
prior to the Stage 2b. A two-level, full factorial experimental
design with two centre points, comprising six experiments was
carried out. For the temperature the range 20-30 °C and for
the distillation endpoint, volumes in the range 2-10 were
selected. The study showed that the interaction of high distil-
lation temperature and low distillation endpoint volume in-
creases the racemisation of (R)-piperidone 6. In order to define
the PARs for these potential QCPPs, the stability of the (R)-
piperidone 6 was monitored at 2.5 and 2.0 volumes at 25 °C
for 23 h. These values were investigated to achieve a minimum
final volume, to maximise the reactor capacity for the subse-
quent Stage 2b. Data are graphically presented in Figure 2.
These results show that (R)-piperidone 6 is acceptably stable
to racemisation for up to 23 h at 25 °C at 2.0 volumes. The
PARs for the distillation temperature and distillation end point
volume are summarised in Table 5.
5.2. PARs Determination. Table 5 summarises the PARs
for the potential QCPPs for Stage 2a, along with the proposed
target values. These PARs were further demonstrated in stressed,
verification experiments as explained in section 5.3.
Running Stage 2a within the PARs of the potential QCPPs
listed above is required to ensure that the levels of the drug
substance-CQAs 14, 15, 16, 17, 18, and 19 (see Scheme 2 and
Table 1) will meet the specification criteria for casopitant
mesylate.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
809

Scheme 6. Piperidone-urea 3 formation
5.3. Verification of the PARs. 5.3.1. Details on the Equip-
ment. The PARs listed in section 5.2 were defined on small-
scale laboratory equipment.
In order to confirm that these PARs are applicable to
reactions at scale, and unless otherwise specified, verification
studies were performed using extremes of the Design Space in
2-L-scale equipment configured to mimic full-scale plant
equipment. This was achieved through maintaining geometric
similarity and operating under conditions scaled according to
accepted chemical engineering principles, e.g., using the
constant power-per-unit volume (P/V) principle for scaling
agitation speed.
Scheme 7. Piperidone-urea 3 decomposition
In addition, parameters that were previously investigated
using a univariate approach were introduced in this study to
ensure any interaction with other process parameters was
identified.
5.3.2. Details on the Experiments. The verification experi-
ments were done by carrying out two experiments. The
parameter settings that were selected can be defined as “mild”
or “forcing”, depending on the impact expected on the drug
substance-CQAs.
The “forcing” conditions were identified as those combina-
tions of extremes of the PARs which, on the basis of knowledge
of the process, would have a higher impact on drug substance-
CQAs (high temperature and low volumes), whereas the “mild”
conditions represent the other extreme of the PARs (low
temperature and high volumes). These conditions were used to
ensure that there was no failure of the drug substance quality
at these extremes.
Casopitant mesylate derived from these experiments met the
specifications,confirmingthatthePARsselectedwereappropriate.
5.4. Summary of the PARs and Criticality Assignment
of Potential QCPPs Identified in Stage 2a. The Design Space
knowledge derived from sections 5.1 and 5.3, the intended
operating ranges in the production plant, knowledge gained from
process models established in the laboratory, and performance
of representative batches produced on scale were then used in
a FMEA risk assessment to identify failure modes within the
process that may lead to drug substance-CQAs failure.
The outcome of this risk assessment was that all the
parameters summarised in section 5.2 had very low risks of
failing drug substance-CQAs, and therefore, no QCPPs are
identified within the process; only QPPs remain. Accordingly,
the current PARs for all the QPPs, as well as their target values
and the drug substance-CQAs linked with the parameter, are
summarised in section 5.2.
Table 6. Potential QCPPs and ranges for Stage 2b
parameter
range
concentration of 6 in ethyl acetate
quantity of TEA
quantity of carbamoyl chloride 4
temperature
from 4.5 to 5.5 vol
from 2.3 to 2.7 equiva
from 1.0 to 1.2 equiva
from 75 to 85 °C
(reflux temperature)
a The equivalents are calculated with respect to (R)-piperidone 6.
triethylamine were also selected as a starting point to perform
Stage 2b. An excess of base was required to avoid the
degradation of the piperazine-urea 3 to the unsaturated ketone
23, occurring in strong acidic conditions as shown in Scheme
7. This excess of base also avoids epimerisation of the (R)-
piperidone 6 as demonstrated by stability studies.
An initial risk assessment was carried out, taking into account
the available process knowledge and the reaction mechanism
proposed. The following process parameters were identified as
potential QCPPs, and some DOE studies were carried out to
understand the impact of these potential QCPPs on drug
substance-CQAs. Available knowledge allowed also the iden-
tification of the appropriate ranges for these studies. Table 6
summarises the potential QCPPs identified and the ranges used
in the DOE studies.
The reactions were performed using an automated platform
and gave yields between 82 and 99%. Analysis of the data
coming from the DoE showed that the only factor affecting
the yield was the quantity of carbamoyl chloride 4, whereas
the other factors did not affect this response. As can be seen in
Figure 3, a higher quantity of chloride gave higher yields, and
no curvature was observed; thus, 1.2 equiv of carbamoyl
chloride 4 was selected. Regarding the other parameters, the
centre point of the ranges studied was selected for the
manufacturing process.
The Design Space knowledge derived from these studies,
the additional knowledge gained in the production plant,
knowledge gained from process models established in the
laboratory, and performance of representative batches produced
on scale were then used in a FMEA risk assessment, and the
result of the assessment showed that these parameters had no
impact on drug substance-CQA; thus, these parameters were
6. Process Studies for Stage 2b
In Stage 2b (R)-piperidone 6 was reacted with carbamoyl
chloride 4 to give the piperidone-urea 3 (Scheme 6). The
preparation of the carbamoyl chloride 4 is carried out in Stage
2e, the detailed discussion of which is reported in a previous
paper.4 It is worth noting that no QCPPs/QPPs have been
identified for Stage 2e.
On the basis of the experimental conditions defined for the
preparation of the carbamoyl chloride 4, ethyl acetate and
(4) Guercio, G.; Bacchi, S.; Perboni, A.; Leroi, C.; Bientinesi, I.; Hourdin,
M.; Goodyear, M.; Curti, S.; Provera, S.; Cimarosti, Z. Org. Process
Res. DeV. 2009, 13, 1100–1110.
810
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Figure 3. Result of the DOE study for Stage 2b.
neither QPPs nor QCPPs, and they were not considered in the
definition of the control strategy.
good yield but also to have the highest selectivity, as the
casopitant isomer 11 can also be formed as shown in Scheme
8, where a summary of the main side products is reported.
Many alternative reducing systems were tried in different
solvents, and a summary of them is reported in Table 7.
Strong reducing systems (entries 1, 2, and 3) caused the
formation of the alcohols 24 as the main product; this can be
explained by considering that 1-acetylpiperazine is both poor
nucleophile and weak base. This slows the initial nucleophilic
attack on the carbonyl carbon and leads to slower overall
reaction rates favoring the main side reaction which is the direct
reduction of the carbonyl group to give the mixture of alcohols
24. A solution to this problem was found by using less reactive
hydrides that could minimise the extent of the carbonyl
reduction (entry 4). Unfortunately, the use of NaBH(OAc)3 did
7. Process Studies for Stage 2c
7.1. Initial Process Understanding Studies. Stage 2c is a
reductive amination reaction where the piperidone-urea 3 reacts
with 1-acetylpiperazine in the presence of a reducing system.
The reductive amination reaction promoted by hydride reagents
was selected over alternative approaches. Catalytic hydrogena-
tion (Pt, Pd, Ni), catalytic transfer hydrogenation (with different
hydrogen donor species), and attempts to form intermediates
prone to reduction (formation of enamines and use of titanium
derivatives) failed, giving low yields, high amounts of side
products, or poor selectivity.
Having selected the reduction approach, the major hurdle
to be overcome with this reaction was not only to achieve a
Scheme 8. Stage 2c: product and side products
Table 7. Reducing systems tried for Stage 2c
entry
reducing system
solvent
2 (% a/a)a
2/11 ratio
alcohol 24 (% a/a)b
residual 3 (% a/a)a
1
2
LiBH4
acetonitrile
acetonitrile
acetonitrile
acetonitrile
acetonitrile
acetonitrile
acetonitrile
ethyl acetate
IPA
-
-
-
100
80
50
6.5
18
37
-
-
20
50
0.4
0.3
4
53
15
4
NaBH4
-
3
L-Selectride
NaBH(OAc)3
-
-
4
52
58
46
24
43
63
70
55/45
71/29
77/23
52/48
73/27
68/32
71/29
5
NaBH4/HCOOH (1/3)
NaBH4/CH3COOH (1/1.75)
(CH3)4NBH(OAc)3
NaBH4/HCOOH/Ti(O-iPr)4 (1/3/1)
NaBH4/HCOOH/Ti(O-iPr)4 (1/3/1)
NaBH(OAc)3/HCOOH
6
7
8
25
4
9
10
acetonitrile
1
0.08
a HPLC walk-up method was used; the assumption that 100% area ) area of (2 + 11 + 24 + 3) was made. b Alcohol 24 is a mixture of syn- and anti-isomers.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
811

Table 8. Potential QCPPs after initial risk assessment
understand the impact of these potential QCPPs versus drug
substance-CQAs. Available knowledge allowed the identifica-
tion of the appropriate ranges for these parameters. Table 8
summarises the potential QCPPs identified and the ranges used
in the DOE studies.
A 2 Level Factorial design (1/8 Fractional) with two centre
points and 10 reactions was used to study the reaction; the
responses collected were the yield and the 2/11 ratio.
Analysis of the obtained data showed that the centre point
of the ranges selected (1.61 equiv of NaBH(OAc)3, 5.5 equiv
of formic acid, 2.27 equiv of 1-acetylpiperazine and 15 °C)
gave the best yield and showed an acceptable robustness after
reaction completion, as can be seen in Figure 4. These values
were selected for the manufacturing process.
In this DOE study also the 2/11 ratio (see Scheme 8) was
considered, and the variability seen in these experiments (from
70/30 to 75/25) was not considered significant. Process studies
demonstrated that the performance of Stage 2d did not change
even when a crude coming from the reduction with NaB-
H(OAc)3 with ratio 2/11 ) 55/45 (Table 7, entry 4) was used.
7.3. Final Risk Assessment. The Design Space knowledge
derived from these studies, the additional knowledge gained in
the production plant, knowledge gained from process models
established in the laboratory and performance of representative
batches produced on scale were then used in a FMEA risk
assessment. The result of the assessment showed that these
parameters had no impact on drug substance-CQA; thus, these
parameters were neither QPPs nor QCPPs, and they were not
considered in the definition of the control strategy.
parameter
rangea
amount of NaBH(OAc)3
amount of HCOOH
from 1.55 to 1.72 equiv
from 5 to 6 equiv
from 5 to 25 °C
temperature
stirring time of the mixture
NaBH(OAc)3/formic acid
amount of 1-acetylpiperazine
addition time of the mixture
3/1-acetylpiperazine
from 5 to 55 min
from 2 to 2.54 equiv
from 5 to 55 min
a The range of equivalents reported is related to piperidone-urea 3. The amounts
reported in the experimental procedure are calculated on the (R)-piperidone
mandelate salt 7, assuming a yield of 85% mol for Stage 2b.
not solve the selectivity problem (approx 1/1 diastereoisomeric
ratio between 2 and 11 was observed).
Significant improvements were made when a modification
of the boron substituent of the reducing system was introduced.
During the studies to find the best reducing system, it was noted
that the steric hindrance of the reducing system could have
played a role in the definition of the stereoselectivity of this
reaction. In particular, it was observed that the reduction by
using NaBH4 gave predominantly the syn-alcohol (80 to 20 with
respect to the anti-alcohol) while the use of L-Selectride gave
predominantly anti-alcohol (80 to 20 with respect to the syn-
alcohol). On the basis of these observations, a series of trials
was carried out by using systems with different steric hindrance;
the results are reported in Table 7 (entries 4-10). The best
results in terms of selectivity and yield were obtained with the
systems NaBH4/HCOOH (entry 5), NaBH4/HCOOH/Ti(OiPr)4
(entry 9), and NaBH(OAc)3/HCOOH (entry 10). The latter was
selected as the yield was higher (less alcohol formation), and
the handling in view of plant scale-up was better (less hydrogen
evolution during its preparation).
8. Summary of the Control Strategy for Stages 2a, 2b, and
2c
A summary of the control strategy defined for the drug
substance-CQAs discussed in this paper is reported in Table 9.
7.2. Initial Risk Assessment and DOE Studies. An initial
risk assessment was carried out, keeping into account the
available process knowledge and the reaction mechanism
proposed. The following process parameters were identified as
potential QCPPs, and some DOE studies were carried out to
9. Conclusions
The QbD principles outlined in ICH and other guidances
provide a structured approach to gaining process knowledge
Figure 4. Result of the DOE study for Stage 2c.
812
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

1H NMR (600 MHz, DMSO-d6): δ 7.98 (s, 1H), 7.78 (s,
2H), 7.25 (dd, 1H), 6.98 (dd, 1H), 6.89 (dt, 1H), 5.25 (dd, 1H),
5.16 (q, 1H), 3.63 (m, 1H), 3.56 (m, 1H), 2.75 (dd, 1H), 2.68
(dd, 1H), 2.57 (s, 3H), 2.53 (m, 1H), 2.46 (dt, 1H), 2.27 (s,
3H), 1.58 (d, 3H). MS: m/z 505 [MH]+.
(2R,4S)-4-(4-Acetyl-1-piperazinyl)-N-{(1R)-1-[3,5-bis(tri-
fluoromethyl)phenyl]ethyl}-2-(4-fluoro-2-methylphenyl)-N-
methyl-1-piperidinecarboxamide (Casopitant 2). HCOOH
(0.49 L, 13 mol) was added to a cooled suspension of
NaBH(OAc)3 (0.82 kg, 3.87 mol) in CH3CN (4 L), keeping
the internal temperature between 10-15 °C; then the lines were
washed with more CH3CN (1 L), and the mixture was stirred
for 40 min.
Table 9. Summary of the control strategy
elements of control
starting
materials and
intermediates
quality
process
drug
substance
drug substance
CQAs
specifications parameters specification
casopitant stereoisomers (R)-piperidone
Stage 2a: 7
yes
11a, 14, 15, 16, 17,
18, 19
mandelate salt 7 suspension
temperature
(R)-amine 5
Stage 2a: 6
distillation
temperature
Stage 2a: 6
distillation
endpoint
volume
azine-related impurities 1-acetylpiperazine none
20, 21 and 22
yes
1-Acetylpiperazine (0.7 kg, 5.46 mol) was added neat over
the solution of piperidone-urea 3, and the mixture was diluted
with CH3CN (3 L). The resulting mixture was added over the
previous suspension; fresh CH3CN (4 L) was used to wash the
line. The reaction mixture was stirred at 15 °C for 12 h. The
solvent was evaporated under reduced pressure to 4 L.
The resulting suspension was diluted with fresh EtOAc (4
L), and then washed with ammonia [21% w/w solution (4 L,
∼11.25 M in NH3)], Na2CO3 [15% w/w solution (4 L)]. More
EtOAc (4 L) was added, and the organic layer was washed
with water (4 L). The organic phase was then concentrated to
2.5 L; again fresh EtOAc (4 L) was added, and the solution
was concentrated to 2.5 L to give a solution of casopitant 2.
1H NMR (600 MHz, DMSO-d6): δ 7.99 (s, 1H), 7.68 (s,
2H), 7.18 (dd, 1H), 6.90 (dd, 1H), 6.76 (td, 1H), 5.33 (q, 1H),
4.14 (dd, 1H), 3.38 (m, 5H), 2.71 (s, 3H), 2.72 (m, 1H), 2.54
(m, 1H), 2.47 (m, 2H), 2.41 (m, 2H), 2.34 (s, 3H), 1.95 (s,
3H), 1.85 (m, 1H), 1.77 (m, 1H), 1.62 (dq, 1H), 1.47 (d, 3H),
1.40 (q, 1H).
(2R,4S)-4-(4-Acetyl-1-piperazinyl)-N-{(1R)-1-[3,5-bis(tri-
fluoromethyl)phenyl]ethyl}-2-(4-fluoro-2-methylphenyl)-N-
methyl-1-piperidinecarboxamide Methanesulfonate Salt (Ca-
sopitant Mesylate 1). The solution of casopitant 2 was then
diluted with EtOAc (overall solution of 2 in EtOAc was 4 L)
and acetone (4.5 L) and was heated to the required temperature
(from 39 °C). Then, neat methanesulfonic acid (0.12 L, 1.64
mol) was added, followed by a slurry of 2 (0.005 kg) in EtOAc
(0.05 L) as seed. The obtained suspension was stirred for 1 h.
Then, isooctane (3 L) was added in the required time (1 h),
and the slurry was cooled to 20 °C in 2 h and aged 3 h.
The suspension was filtered, and the solid was washed with
EtOAc (3 × 4 L). The white solid was dried overnight under
vacuum at 40 °C to give the desired casopitant mesylate 1 (0.94
kg, yield 48% mol with respect to 7).
a This stereoisomer is included as monitored with 15 (its enantiomer) in the
achiral HPLC method, 11 is controlled in the final crystallisation step.
and developing robust manufacturing control strategies. This
has been applied successfully to developing a control strategy
for Stages 2a, 2b, and 2c of the casopitant mesylate 1
manufacturing process. On the basis of the process understand-
ing provided, it has been demonstrated that the control of the
drug substance-CQAs discussed can be robustly achieved by
the elements of control defined above for Stages 2a and 2c.
The process understanding generated and the control strategy
proposed could potentially lead to the removal of these drug
substance-CQAs from the drug substance specification, under-
taking a further step towards the full application of the QbD
principles to manufacturing processes.
10. Experimental Section
(2R)-2-(4-Fluoro-2-methylphenyl)-4-piperidinone 6. (R)-
Piridone mandelate salt 7 (1 kg, 2.78 mol) was added to a
mixture of Na2CO3 [15% w/w solution (4 L)] and EtOAc (4
L). The mixture was stirred until dissolution was complete. The
phases were separated, and the organic layer was washed with
NaCl [20% w/w solution (4 L)]. EtOAc (4 L) was added, and
the organic phase was concentrated to 2.5 L to give a solution
of (R)-piperidone 6.
1H NMR (600 MHz, DMSO-d6): δ 7.51 (dd, 1H), 7.01 (m,
2H), 3.98 (m, 1H), 2.85 (m, 1H), 2.74 (br s, 1H), 2.49 (m, 1H),
2.42 (dd, 1H), 2.31 (s, 3H), 2.28 (m, 1H), 2.20 (m, 1H).
(2R)-N-{(1R)-1-[3,5-Bis(trifluoromethyl)phenyl]ethyl}-2-
(4-fluoro-2-methylphenyl)-N-methyl-4-oxo-1-piperidinecar-
boxamide 3. Et3N (0.98 L, 7.03 mol) was added over a solution
of carbamoyl chloride 4 (2.5 L, 3.32 mol), and the resulting
mixture was added to the solution of (R)-piperidone 6; the line
was washed with EtOAc (1 L), and then the mixture was heated
at reflux for 16 h. The reaction mixture was cooled to room
temperature, and 1-acetylpiperazine neat (0.2 kg, 1.56 mol) was
added, followed by a line wash with EtOAc (0.25 L). The
reaction mixture was stirred for 30 min at room temperature;
the organic solution was washed with malic acid 28% w/w
(3 L), Na2CO3 [15% w/w solution (3 L)], and NaCl [20% w/w
solution (4 L)]. CH3CN (4 L) was added, and the solution was
concentrated to 2.5 L; then CH3CN (4 L) was added again,
and the solution was concentrated to 3 L to give a solution of
piperidone-urea 3 in CH3CN.
1H NMR (600 MHz, DMSO-d6): δ 9.57 (br s, 1H), 7.99 (br
s, 1H), 7.68 (br s, 2H), 7.23 (m, 1H), 6.95 (dd, 1H), 6.82 (m,
1H), 5.31 (q, 1H), 4.45 (m, 1H), 4.20 (dd, 1H), 3.99 (m, 1H),
3.56 (m, 1H), 3.47 (m, 3H), 3.37 (m, 1H), 3.15 (m, 1H), 2.96
(m, 1H), 2.87 (m, 1H), 2.80 (t, 1H), 2.74 (s, 3H), 2.36 (s, 3H),
2.30 (s, 3H), 2.13 (m, 1H), 2.08 (m, 1H), 2.10 (s, 3H), 1.87
(m, 1H), 1.73 (m, 1H), 1.46 (d, 3H). MS: m/z 617 [MH]+, as
free base.
Acknowledgment
We thank Annalisa Galgano, Orsola Vecchi, Anna Nicoletti,
Vern De Biasi, Paolo Repeto, Mohammad Yahyah, Robert
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
813

Dennehy, Dario Nicolosi, Luca Martini, Paola Russo, Jill
Trewartha, Maria Concepcion Cerrato-Oliveros, Susanna Gori,
Tiziana Parton, Paul Stonestreet, Neil Hodnett, Tom Thurston,
Matteo Gonzi, Ilaria Bientinesi, Corinne Leroi, Fiona Bird, Jim
Meadows, Tim Walsgrove, Matt Kersey, and Vance Novack
for the helpful discussions.
Risk Assessment) presents a high risk to the process falling
outside the Design Space or proven acceptable ranges.
Quality Process Parameter process parameter that influ-
ences a Critical Quality Attribute but (following a Risk
Assessment) presents a low risk of the process falling outside
the Design Space or proven acceptable ranges.
Control Strategy a (planned) set of controls, derived from
(current) product and process understanding that assures process
performance and product quality. The controls can include
parameters and attributes related to drug substance and drug
product materials and components, facility and equipment
operating conditions, in-process controls, finished product
specifications, and the associated methods and frequency of
monitoring and control. (ICH Q10 definition: words in paren-
theses are felt unnecessary.)
Glossary
Drug Product Critical Quality Attributes or Drug
Substance Critical Quality Attributes measurable properties
of drug product or API that are critical to ensuring patient safety
and efficacy. The property must be within a predetermined range
to ensure product quality. A property which is measured outside
the range indicates a batch failure.
Critical Quality Attributes in the unit operation or stage
inputs, stage outputs, device, etc. measurable properties of inputs
and outputs that (as determined by Risk Assessment) present a
high risk to the process falling outside the design space or
proven acceptable ranges.
Quality Attribute in the unit operation or stage inputs,
stage outputs, device, etc. measurable property of inputs and
outputs that (as determined by Risk Assessment) present a low
risk to the process falling outside the Design Space or proven
acceptable ranges.
Proven Acceptable Range (PAR): upper and/or lower
limits for process parameter or attribute values between which
the parameter or attribute is known to produce a process output
(e.g. intermediate, API or DP) that meets the CQAs. The PAR
may or may not represent the point of failure. The PAR for a
given process parameter or attribute may be dependent upon
the PAR values for one or more other process parameters or
attributes (e.g. multivariate).
Received for review March 4, 2010.
OP1000622
Quality Critical Process Parameter process parameter that
influences a Critical Quality Attribute and (as determined by
814
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development