Study on the HPLC-based separation of some ezetimibe stereoisomers and the underlying stereorecognition process

Article abstract of DOI:10.1002/chir.22829

The enantioseparation of ezetimibe stereoisomers by high-performance liquid chromatography on different chiral stationary phases, ie, 3 polysaccharide-based chiral columns, was studied. It was observed that cellulose-based Chiralpak IC column exhibited the best resolving ability. After the optimization of mobile phase compositions in both normal and reversed phase modes, satisfactory separation could be obtained on Chiralpak IC column, especially in normal phase mode. The use of prohibited solvents as nonstandard mobile phase gave rise to better resolution than that of standard mobile phases (n-hexane/alcohol system). In addition, the presence of ethanol in nonstandard mobile phase has played an important role in enhancing chromatographic efficiency and resolution between ezetimibe stereoisomers. Various attempts were made to comprehensively compare the chiral recognition capabilities of immobilized versus coated polysaccharide-based chiral columns, amylose-based versus cellulose-based chiral stationary phases, reversed versus normal phase modes, and standard versus nonstandard mobile phases. Moreover, possible solute-mobile phase-stationary phase interactions were derived to explain how stationary and mobile phases affected the separation. Then the method validation with respect to selectivity, linearity, precision, accuracy, and robustness was carried out, which was demonstrated to be suitable and accurate for the quantitative determination of (RRS)-ezetimibe impurity in ezetimibe bulk drug.

Full text of DOI:10.1002/chir.22829

Received: 7 November 2017
DOI: 10.1002/chir.22829
Revised: 12 January 2018
Accepted: 15 January 2018
R E G U L A R A R T I C L E
Study on the HPLC‐based separation of some ezetimibe
stereoisomers and the underlying stereorecognition process
Bolin Zhu1 | Yaqi Yao1 | Yu Zhao1 | Tiemin Sun2 | Qing Li1
1 School of Pharmacy, Shenyang
Abstract
Pharmaceutical University, Shenyang,
The enantioseparation of ezetimibe stereoisomers by high‐performance liquid
Liaoning Province, People's Republic of
China
chromatography on different chiral stationary phases, ie, 3 polysaccharide‐based
2 School of Pharmaceutical Engineering,
chiral columns, was studied. It was observed that cellulose‐based Chiralpak IC
Shenyang Pharmaceutical University,
column exhibited the best resolving ability. After the optimization of mobile
Shenyang, Liaoning Province, People's
Republic of China
phase compositions in both normal and reversed phase modes, satisfactory
separation could be obtained on Chiralpak IC column, especially in normal
Correspondence
phase mode. The use of prohibited solvents as nonstandard mobile phase gave
rise to better resolution than that of standard mobile phases (n‐hexane/alcohol
system). In addition, the presence of ethanol in nonstandard mobile phase has
played an important role in enhancing chromatographic efficiency and
resolution between ezetimibe stereoisomers. Various attempts were made to
comprehensively compare the chiral recognition capabilities of immobilized
versus coated polysaccharide‐based chiral columns, amylose‐based versus
cellulose‐based chiral stationary phases, reversed versus normal phase modes,
and standard versus nonstandard mobile phases. Moreover, possible solute‐
mobile phase‐stationary phase interactions were derived to explain how station-
ary and mobile phases affected the separation. Then the method validation with
respect to selectivity, linearity, precision, accuracy, and robustness was carried
out, which was demonstrated to be suitable and accurate for the quantitative
determination of (RRS)‐ezetimibe impurity in ezetimibe bulk drug.
Qing Li, School of Pharmacy, Shenyang
Pharmaceutical University, No. 103,
Wenhua Road Shenhe District 110016,
Shenyang, Liaoning Province, People's
Republic of China.
Email: lqyxm@hotmail.com
Tiemin Sun, School of Pharmaceutical
Engineering, Shenyang Pharmaceutical
University, No. 103, Wenhua Road Shenhe
District 110016, Shenyang, Liaoning
Province, People's Republic of China.
Email: suntiemin@126.com
KEYWORDS
different stationary phases, enantioseparation, ezetimibe stereoisomers, HPLC, separation mechanism
1 | INTRODUCTION
interactions with other medications.1 Ezetimibe used alone
or together with statins proves to be safe and well tolerated.2
Ezetimibe (Ez), 1‐(4‐fluorophenyl)‐3(R)‐[(3S)‐(4‐fluorophenyl)‐
3‐hydroxypropyl]‐4‐(S)‐(4‐hydroxyphenyl)‐2‐azetidinone, as a
selective cholesterol absorption inhibitor, is widely used for
the treatment of hypercholesterolemia. It is effective in lower-
ing serum low‐density lipoprotein cholesterol by inhibition of
the Niemann‐Pick C1‐like1 transporter without affecting the
absorption of fat‐soluble vitamins, triglycerides, or bile acids.
And it has no effect on the activity of major drug metabolizing
enzymes (CYP450), which reduces any potential drug‐drug
In Ez molecule, 3 asymmetric carbons give rise to 8
stereoisomers, among which Ez (namely, (SRS)‐isomer)
has the lowest ED50. Since Ez is the main active ingredi-
ent, any of the other 7 stereoisomers found in the final
product would be taken into account as potential impuri-
ties. However, the synthesis of the final product with the
required stereochemistry is a significant challenge. So, it
is essential to develop an effective method for the
enantioseparation of Ez stereoisomers. Two synthetic
Chirality. 2018;1–10.
wileyonlinelibrary.com/journal/chir
© 2018 Wiley Periodicals, Inc.
1

2
ZHU ET AL.
routes used in present study and the structures of 4
stereoisomers are shown in Figure 1. As depicted, the
yield of chiral impurity is dependent on the synthetic
route adopted. Route 1 allows to obtain Ez and one
impurity (RRS)‐Ez, while 3 undesired stereoisomers of
Ez ((RRS)‐Ez, (SSR)‐Ez, and (RSR)‐Ez) are simulta-
neously present in the final product of route 2. In the
preliminary experiment, the product of route 1 was used
for analysis, and hence, (RRS)‐Ez was the only remaining
impurity. After that, to gain a better understanding of the
stereorecognition process, the enantioseparation of 4
stereoisomers obtained by route 2 was subsequently
examined.
So far to our present knowledge, some techniques have
been reported in the literature for Ez analysis3-5: high‐per-
formance liquid chromatography (HPLC) method for
determination of Ez in pharmaceutical dosage forms,3
liquid chromatography‐mass spectrometric method for
Ez and its glucuronide conjugate (Ez‐glucuronide)
determination in human plasma,4 and spectrophotometric
and TLC‐densitometric methods for the simultaneous
determination of Ez and atorvastatin calcium.5 In contrast
to Ez, research studies focused on the separation of its
stereoisomers have been relatively limited, and only few
workers have attempted chiral resolution of Ez stereoiso-
mers by HPLC.6-8 For example, Filip et al6 achieved the
separation of (RRS)‐Ez from Ez on a Hypersil Gold C18
column by RP‐HPLC analysis. Radhakrishnan and Subba
Rao7 used a Chiralpak IC column for the separation of Ez
and (RRS)‐Ez with mobile phase containing hexane/2‐
proponal/TFA. Similarly, Sun et al8 tested the chiral
recognition of 5 normal phase chiral columns operated in
normal phase mode and reported that Chiralpak IC
column showed the better enantioselectivity. However, to
our knowledge, no paper systematically compared the
enantioseparation ability of immobilized versus coated
polysaccharide‐based chiral columns, amylose‐based
versus cellulose‐based chiral stationary phases (CSPs),
normal versus reversed phase modes, and standard versus
nonstandard mobile phase systems. In addition, few
references focused on the use of nonstandard solvent with
polysaccharide‐based CSPs.
FIGURE 1 Two routes of ezetimibe (Ez) synthesis and chemical structures of Ez, (RRS)‐ Ez, (SSR)‐ Ez, and (RSR)‐Ez

ZHU ET AL.
3
Among HPLC methods for enantioseparation, the use
Shenyang Pharmaceutical University, Shenyang, China.
Two synthetic routes and the corresponding final
products are shown in Figure 1. Methanol (MeOH),
acetonitrile (ACN), CH2Cl2, THF, and n‐hexane of HPLC
grade were purchased from Concord Technology Co. Ltd.
(Tianjin, China). Isopropyl alcohol (IPA) and ethanol
(EtOH) of HPLC grade were purchased from Yuwang
Industrial Co. Ltd. (Shandong, China).
of CSPs has become the most popular tool for determina-
tion of the enantiomeric impurities of wide number of
chiral drugs, as it is rapid, reproducible, and effective.9
The choice of CSP is certainly the dominant factor
contributing to successful separation. Numerous CSPs
are commercially available, and the polysaccharide‐based
CSPs in the Chiralpak and Chiralcel series have been the
most widely utilized materials for enantiomeric resolution
by liquid chromatography.10 Polysaccharide‐based CSPs
are produced either by physical coating or immobilization
of the chiral polymers on silica support. However, the
coated CSPs can only be used with a rather limited eluent,
because some commonly used solvents, such as chloro-
form, acetone, and tetrahydrofuran (THF), can swell or
dissolve the polysaccharide derivatives on silica gel.
However, the use of these prohibited solvents may be
useful for resolving recemates, which could not be
separated by using low‐polarity solvents.11,12 To enhance
the solvent compatibility for these CSPs, the immobilized
ones have been commercialized, which possess
advantages in CSP robustness and extended range of
solvents and applications.13,14 In addition, the coated
CSPs could be used either under normal phase or reversed
phase, while the immobilized CSPs can be used in both
modes.
2.2 | Apparatus
All chromatographic measurements were performed on an
HPLC system consisting of a PU‐1580 pump (Jasco, Tokyo,
Japan) and SPD‐15C UV‐Vis detector. Data acquisition
was performed by using a N2000 Chromatography Data
System obtained from Zhida Information Engineering
Co. Ltd. (Zhejiang, China). The chiral columns, Chiralpak
AD‐H (250 × 4.6 mm), Chiralcel OD‐RH (250 × 4.6 mm)
and Chiralpak IC (250 × 4.6 mm) were obtained from
Daicel Chiral Technologies (China) Co., Ltd, Shanghai,
China. The chiral selectors of these CSPs were tris
(3,5‐dimethylphenylcarbamate) of amylose coated on
5 μm silica gel, tris(3,5‐dimethylphenylcarbamate) of
cellulose coated on 5 μm silica gel, and tris(3,5‐
dichlorophenylcarbamate) of cellulose immobilized on
5 μm silica gel, respectively.
In the present study, attempts have been made to
resolve Ez stereoisomers on both coated and immobilized,
normal and reversed phase CSPs, namely, Chiralpak
AD‐H, Chiralcel OD‐RH, and Chiralpak IC. For the
coated CSP, Chiralpak AD‐H was used in normal phase
mode with mixtures of hexane and alcohol as mobile
phases, and Chiralpak OD‐RH was applied in the
reversed‐phase separation with aqueous mobile phases.
But for Chiralpak IC, as the immobilized CSP, we
uniquely tested unusual mobile phase compositions
containing mixtures of hexane and THF or CH2Cl2 for
separation. This study aimed to explain the separation
results of Ez stereoisomers on various stationary phases
and also the effects of different mobile phases on the
enantioselectivity and resolution in normal and reversed
phase modes. Based on the enantioseparation, the most
suitable CSP and mobile phase were selected for the
quantitative determination of (RRS)‐Ez in Ez bulk drug.
2.3 | Chromatography
Various mobile phase systems were investigated for the
HPLC study. All of them were composed of commonly
used organic HPLC solvents and water. The method
validation was carried out on a Chiralpak IC column
(250 × 4.6 mm id, 5 μm, Daicel Chiral Technologies Co.,
Ltd., Shanghai, China) with a mobile phase consisting of
hexane/IPA (90/10, v/v) at a flow rate of 1.0 mL min−1
.
The injection volume was 20 μL, and the detection
wavelength was set at 232 nm. The temperature was kept
at 25°C except where the effect of temperature was
studied. An amount of 2 mg each Ez and (RRS)‐Ez was
dissolved in 10 mL of mobile phase. The resultant solution
had 0.2 mg mL−1 of individual concentration. The target
analyte concentration of Ez was fixed as 0.2 mg mL−1. A
0.2 mg mL−1 solution of Ez prepared in mobile phase,
which was spiked with 1 μg mL−1 (RRS)‐Ez, was utilized
as standard spiked solution.
All the chromatographic parameters given in this
work were the mean values from 3 consecutive injections.
Retention factor (k′) was calculated from (tR − t0)/t0,
where t0 was the void time estimated from the earliest
baseline perturbation and tR was the retention time of
enantiomeric solute. Separation factor (α) was calculated
from k′2/k′1, where k′1 and k′2 were the retention factors
2 | MATERIALS AND METHODS
2.1 | Chemical and reagents
Ezetimibe, (RRS)‐Ez, (RSR)‐Ez, and (SSR)‐Ez (purity>99.0%) were
purchased from Nanjing Healthnice Pharmaceutical Co.
Ltd. (Nanjing, China). The mixtures of stereoisomers were
synthesized by School of Pharmaceutical Engineering,

4
ZHU ET AL.
of the first and second elutions, respectively. The resolu-
tion (Rs) was directly given by the software and calculated
according to the equation: Rs = 2(t2 − t1)/(W1 + W2),
where W1 and W2 were the peak width of the first and sec-
ond elutions, respectively.
2.4.6 | Accuracy
The accuracy of method was evaluated by spiking Ez
standard with known amount of (RRS)‐Ez at 80, 100,
and 120 of the permitted maximum level (1 μg mL−1) in
triplicate. The recovery was assessed by dividing the
amount of found by the spiked, then multiplied by 100%.
2.4 | Method validation
2.4.7 | Solution stability
2.4.1 | System suitability test
The stability of the standard spiked solution was deter-
mined by keeping at room temperature for 10 hours and
injecting immediately after preparation and every 2 hours
(0, 2, 4, 6, 8, and 10 h).
System suitability testing is an integral part of many
analytical procedures. The tests are based on the concept
that the equipment, electronics, analytical operations,
and samples to be analyzed constitute an integral system.
The system suitability was assessed by 6 replicate analyses
of standard spiked solution.
2.4.8 | Robustness
The evaluation of robustness should be considered during
the development phase and depends on the type of
procedure under study. It should show the reliability of
an analysis with respect to deliberate variations in method
parameters. Based on the proposed method, the effects
2.4.2 | Specificity
The specificity of the analytical method was evaluated by
the analysis of a solution containing Ez and (RRS)‐Ez.
of the proportion of IPA (10
1%), flow rate
(1.0 0.1 mL min−1), wave length (232 5 nm), and col-
umn temperature (20 + 5/10°C) on the enantioseparation
were investigated.
2.4.3 | Linearity
The linearities of Ez and (RRS)‐Ez were determined by
using 6 calibration solutions over a concentration
covering a range of 0.2 μg mL−1 (LOQ) to 1.2 μg mKL−1
(0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 μg mL−1). Regression curve
was obtained by plotting peak area versus concentration,
using the least squares method.
2.5 | Determination of (RRS)‐ezetimibe
impurity in ezetimibe bulk drug
Three batches of Ez bulk drug samples (batch nos.
2013072701, 2013072702, and 2013072703) were analyzed
according to the proposed condition. The mobile phase
was adopted as a diluent to obtain final sample solutions,
which contained ~0.2 mg mL−1 Ez. Then the reference
solutions were prepared by diluting the corresponding
sample solutions to 1.0 μg mL−1 with mobile phase. The
content of (RRS)‐Ez impurity was determined by using
reference solutions run simultaneously.
2.4.4 | LOD and LOQ
LOD and LOQ represent the concentration of analyte that
would yield the S/N of 3 and 10 for peak area, respec-
tively, following the ChP. The LOD and LOQ of Ez and
its isomer were achieved by injecting serial dilution of a
solution of known concentration. The percentage of RSD
of the peak areas of 6 consecutive injections of (RRS)‐Ez
at LOQ concentration should be lower than 10%.
3 | RESULTS AND DISCUSSION
3.1 | Method development and
optimization for chiral separation
2.4.5 | Precision and repeatability
Method precision was determined by measuring the
repeatability (intraday precision) and intermediate preci-
sion (interday precision) of the peak area of (RRS)‐Ez.
The repeatability was achieved by calculating the RSD
in area of (RRS)‐Ez in spiked samples for 6 replicate
injections, and the intermediate precision was expressed
in term of RSD in area obtained for analyses performed
on 3 consecutive days, 6 times each day.
Except Ez as main product, synthetic route 1 produces 1
isomer of Ez, namely, (RRS)‐Ez. To obtain the optimal
condition more effectively, the final product of route
one was analyzed in preliminary experiments. Then, if
better separation can be obtained, the 4 stereoisomers
of route
2
were further selected to study the
stereorecognition process. As described in previous
study, the enantioseparation ability of polysaccharide

ZHU ET AL.
5
derivative depends not only on the substituents but also
on the structure of polysaccharide backbone.15-17
Herein, 3 different polysaccharide‐based CSPs, with
dimethylphenylcarbamates or dichlorophenylcarbamates
of amylose or cellulose as selectors were evaluated in
the following study.
reported the elution order of 2 stereoisomers on different
CSPs. It was found that the elution of (RRS)‐Ez was prior
to Ez on Chiralcel OD‐RH, but this behavior was reversed
when Chiralpak IC was used. Based on the areas of 2
peaks, the concentration of Ez and (RRS)‐Ez existed in
the ratio of 3:1. The large quantity difference between
Ez and (RRS)‐Ez was attributed to the stereoselectivity
of the synthetic route one. It was concluded that the
hydroxy group at the precursor of Ez (benzyl‐Ez) tended
to be converted into the S configuration (Ez) during the
final debenzylation process, so the forming of Ez was
dominant.
Since partial separation was obtained on Chiralcel
OD‐RH, a decrease in ACN proportion was subsequently
performed to improve the resolution. However, until the
retention up to 60 minutes, Ez and (RRS)‐Ez could not
reach the baseline separation. As for Chiralpak IC, the
chiral polymer adhered to the silica‐gel matrix, making
the adsorbent robust and compatible with various solvent
systems.11 Thus, in the following study, we investigated its
enantioselectivity ability for Ez and its stereoisomers in
normal and reversed phase modes, respectively.
3.1.1 | Separation on 3 polysaccharide‐
based chiral columns
The results of separation on 3 CSPs are summarized in
Table 1, with Chiralpak AD‐H operated under normal
phase mode and other 2 columns under reversed phase
mode. It is important to know that except for Chiralpak
IC, other 2 columns, Chiralpak AD‐H and Chiralcel
OD‐RH, are prepared by coating polysaccharides on the
silica support.
It was noticed from the results that no separation was
achieved on Chiralpak AD‐H, with mobile phases
comprising either mixtures of hexane and IPA or hexane
and EtOH. In contrast, partial separation was noticed on
Chiralcel OD‐RH and Chiralpak IC in reversed phase
mode. It is commonly admitted that the mobile phase
composition played an important role in the chiral
recognition ability of these CSPs. Since the mobile phases
used for Chiralpak AD‐H differed to those of the other 2
CSPs, the separation results could not be explained just
in aspect of structural difference of these CSPs.16
Nevertheless, this results supported the conclusion that
only 2 cellulose‐based CSPs (Chiralcel OD‐RH and
Chiralpak IC) showed chiral discrimination capability
toward Ez and (RRS)‐Ez. Chiralpak IC gave in fact better
enantiomeric separation than Chiralcel OD‐RH, as
reflected by the higher resolution (Rs) and shorter
retention (k′). No separation on Chiralpak AD‐H would
be due to no affinity of the solutes to the CSP or because
of the difficulty of the inclusion of solutes into chiral
cavity.18,19
3.1.2 | Separation on Chiralpak IC in
reversed phase mode
In reversed phase mode, the column pressure is high
because of the presence of large amount of water. In this
study, the flow rates (0.4‐0.7 mL min−1) were selected
based on the recommended pressure of this column. A
crucial step to achieve satisfactory resolution is the choice
of proper organic modifiers. Changes in the compositions
of mobile phases could alter the enantiorecognition mech-
anisms, resulting in different separation behaviors.20,21 In
our study, the reversed phase enantioseparation ability of
Chiralpak IC towards Ez and (RRS)‐Ez was evaluated
with various ratios of MeOH/H2O, EtOH/H2O, and
ACN/H2O as mobile phases.
To determine the elution orders of Ez and (RRS)‐Ez on
Chiralcel OD‐RH and Chiralpak IC, peak identification
was established by injecting the reference standard solu-
tions under the same elution conditions. Table 1 also
As indicated by the chromatographic data presented
in Table 2, their retention factors increased as the amount
of MeOH decreased, the same effect but a lesser extent,
was observed for enantioselectivity (α) and resolution
TABLE 1 Results of separation on 3 chiral columns: capacity factor (k′), separation factor (α), and resolution (Rs)
k′
Mobile Phase
(v/v)
Flow Rate
(mL min−1
Column
)
Ez
(RRS)‐Ez
α
Rs
Chiralpak AD‐H
n‐hexane/IPA 85/15
n‐hexane/EtOH 85/15
1.0
1.0
9.97
6.63
9.97
6.63
1.00
1.00
0.00
0.00
Chiralcel OD‐RH
MeOH/H2O 60/40
ACN/H2O 60/40
1.0
1.0
‐
‐
‐
‐
10.41
9.87
1.05
0.76
Chiralpak IC
MeOH/H2O 70/30
0.4
4.87
5.17
1.07
1.12
“‐” means no elution within 60 min.

6
ZHU ET AL.
TABLE 2 Results of separation on Chiralpak IC column in reversed phase mode
k′
Mobile Phase
(v/v)
Flow Rate
(mL min−1
)
Ez
4.83
(RRS)‐Ez
α
Rs
MeOH/H2O 70:30
MeOH/H2O 65:35
MeOH/H2O 62:38
EtOH/H2O 50:50
ACN/H2O 50/50
ACN/H2O 45/55
ACN/H2O 40/60
ACN/H2O 35/65
0.4
0.4
0.4
0.4
0.5
0.5
0.5
0.7
5.17
1.07
1.12
9.24
15.43
9.30
10.07
16.98
10.00
3.40
1.09
1.10
1.07
1.07
1.10
1.11
1.12
1.51
1.55
0.83
1.57
2.13
2.78
2.93
3.16
4.82
5.30
8.57
9.54
17.82
20.03
For abbreviations see heading to Table 1.
(Rs). With the mobile phase consisting of MeOH/H2O
(65/35, v/v) at a flow rate of 0.4 mL min−1, baseline
separation was achieved for the Ez and (RRS)‐Ez along
with the prolonged retention time (near 50 min). A
further increase in water concentration did not contribute
to significant enhancement in resolutions. On the
contrary, the dramatic increase in retention factor was
noticed. In the case of EtOH instead of MeOH in the
mobile phase, the column exhibited poor enantiomer
resolving ability. When replacing EtOH with ACN in the
same volume content, the retention of analytes decreased
significantly and 2 stereoisomers were easily baseline
resolved. As reflected by the higher α of ACN/H2O mobile
phase (Table 2), the separation of Ez stereoisomers with
ACN was significantly superior to those obtained with
other modifiers. The possible reason was that alcohols
might disrupt the hydrogen bond between analyte
molecule and CSP and thus worsen enantioseparation.20
By the optimization of ACN/H2O proportion, the best
separation of Ez and (RRS)‐Ez (Rs = 2.93) was achieved
with ACN/H2O (35/65, v/v) as mobile phase. The corre-
sponding chromatogram is given in Figure 2.
FIGURE 2 Typical high‐performance liquid chromatography
chromatogram of Ez and (RRS)‐Ez. Column: Chiralpak IC; mobile
phase: ACN/H2O (35:65, v/v); flow rate: 0.7 mL min−1
enantioseparation. Results revealed that the π‐π
interaction was the vital driving force that caused chiral
recognition of Ez stereoisomers on Chiralpak IC column.
3.1.3 | Separation on Chiralpak IC in nor-
mal phase mode
In view of the partial resolution and prolonged
retention on Chiralcel OD‐RH, it was remarkable that
the enantiomeric recognition of 2 stereoisomers was more
effective on Chiralpak IC, which might be attributed to
the π‐π interaction of solutes with the CSP.8 The substitu-
tion of electron withdrawing group chlorine at the 3 and 5
position of phenylcarbamate on Chiralpak IC made its
aromatic ring lack of electrons.22,23 Hence, the π‐π
interaction between electron‐rich aromatic ring in solutes
(benzene ring with substituent hydroxyl comprising chiral
carbon atom) and electron‐deficient aromatic ring on CSP
could be accordingly enhanced and improve the separa-
tion ability of Chiralpak IC. By contrast, the substituent
on carbohydrate backbone of Chiralcel OD‐RH was
dimethylphenyl, which played the opposite role in
It is known that the alcohol modifiers used in normal
phase mode have profound influence on chiral selectiv-
ity.24 Since the chiral selector is covalently immobilized
onto the surface of silica, it is possible for Chiralpak IC
to use a wider variety of mobile phases to optimized
enantioseparation. During our experiments, the standard
mobile phases including n‐hexane/IPA and n‐hexane/
EtOH were used. Prohibited solvents such as THF and
CH2Cl2 regarded as nonstandard mobile phases have also
been tried. The separation of Ez and its stereoisomer
impurities with standard and nonstandard mobile phases
were discussed respectively in the following.
The normal phase enantioseparation of Ez and
(RRS)‐Ez was initially tested with mobile phase consisting

ZHU ET AL.
7
of n‐hexane/IPA (90/10, v/v). Satisfactory resolution of
about 4.0 was obtained within 30 minutes, indicating the
good separation performance of Chiralpak IC in normal
phase mode. In view of this result, synthetic route 2 was
adopted to obtain 4 stereoisomers, namely, Ez, (RSR)‐Ez,
(SSR)‐Ez, and (RRS)‐Ez. Herein, these 4 stereoisomers
were selected to gain a better understanding of the
enantioseparation ability of Chiralpak IC.
to retention, resolution, and peak shape was obtained by
using mobile phase consisting of n‐hexane/IPA (90/10,
v/v). The typical chromatogram is shown in Figure 3.
Under the optimum condition, the retention times
were 20.5, 25.8, 31.0, and 36.3 minutes for Ez, (RRS)‐Ez,
(RSR)‐Ez, and (SSR)‐Ez by injection of the pure enantio-
mers under the same experimental conditions. The
contents of 4 isomer were calculated by area normalization
method, which were found to be 32.3%, 17.8%, 17.7%, and
32.2% for Ez, (RRS)‐Ez, (RSR)‐Ez, and (SSR)‐Ez,
respectively. As mentioned in route one, the hydroxy group
at benzyl‐Ez was mainly transformed into the S‐configura-
tion, because of the stereoselectivity of debenzylation
process. In the last process of route 2, 2 benzyl groups were
removed from the 2 enantiomers of benzyl‐Ez, and each
isomer of benzyl‐Ez would produce 2 isomers of Ez
(Figure 1). Similar with route 1, the hydroxy groups at both
benzyl‐Ez in route 2 also tended to translate into the S‐con-
figuration. Thus, (RRS)‐Ez and (RSR)‐Ez were thereof
present in amounts below Ez and (SSR)‐Ez.
The separation results with prohibited solvents such
as THF and CH2Cl2 as nonstandard mobile phases on
Chiralpak IC are presented in Table 3. As seen, the use
of n‐hexane and THF as mobile phase only gave 3 peaks
for 4 stereoisomers, with corresponding contents of
50.5%, 33.0%, and 16.5%, respectively. Meanwhile, the
reconstructed mixture of Ez and (RRS)‐Ez was also
analyzed under the same mobile phase condition, but no
resolution was obtained. The identifications of the second
Two commonly used modifiers, ie, IPA and EtOH,
were used on Chiralpak IC, but the superior of IPA on
enantioselectivity than EtOH was remarkable. With
regard to the structures of IPA and EtOH, one can deduce
that the longer carbon chain of IPA would weaken its
hydrogen bond with CSP, giving rise to the increased
interaction between solute and CSP. Table 3 gives the
separation results of 4 stereoisomers under normal phase
HPLC with various n‐hexane/IPA as mobile phases.
As shown in Table 3, when the concentration of IPA
decreased, the retention was prolonged, along with the
increased resolution (Rs12) and selectivity factor (α12).
This was probably because the IPA in mobile phase
competed with the analytes to form hydrogen bond with
the carbamate moieties of the CSP. Very interestingly,
resolutions (Rs23, Rs34) and selectivity factors (α23, α34)
of the second and third, third, and forth peaks were
differently affected by the variation of IPA proportion.
This may be because of the different conformational
structures of solutes. During optimization of the
chromatographic conditions, the best result with respect
TABLE 3 Results of separation of ezetimibe and its 3 stereoisomers on Chiralpak IC column in standard and nonstandard normal phase
mode
k′
α
Rs
Mobile Phase (v/v)
k1′
k′2
2.47
k′3
3.10
k′4
3.70
α12
α23
α34
Rs12
Rs23
Rs34
n‐hexane/IPA 85/15
2.17
1.14
1.26
1.19
1.74
3.07
2.57
n‐hexane/IPA 88/12
3.50
5.83
9.83
21.13
9.83
6.50
3.60
4.30
2.77
5.13
4.07
7.60
5.17
9.33
6.27
11.10
19.17
45.03
‐‐‐
1.16
1.30
1.33
1.42
1.31
1.36
1.06
1.32
1.25
1.21
1.27
1.23
1.24
1.28
1.09
1.25
1.42
1.27
1.20
1.27
1.21
1.19
1.19
1.17
1.00
1.41
1.30
1.24
1.21
1.24
2.51
4.02
4.71
5.20
3.72
3.06
3.63
4.32
3.81
3.85
3.83
3.28
3.60
4.43
2.01
3.28
3.44
4.22
3.48
5.23
3.15
2.81
2.95
3.05
0.00
5.59
4.97
4.17
3.56
4.78
n‐hexane/IPA 90/10
n‐hexane/IPA 98/8
13.03
29.93
12.90
8.83
16.10
38.46
14.10
11.03
5.43
n‐hexane/IPA 95/5
n‐hexane/THF 90/10
n‐hexane/CH2Cl2 20/80
n‐hexane/CH2Cl2/EtOH 50/50/2
n‐hexane/CH2Cl2/EtOH 50/50/1.5
n‐hexane/CH2Cl2/EtOH 60/40/3
n‐hexane/CH2Cl2/EtOH 60/40/2.5
15.50
7.03
8.93
5.03
9.80
3.83
5.67
7.20
3.47
4.17
6.20
7.90
For abbreviations see heading to Table 1.
Except for the mobile phase consisting of n‐hexane/THF 90/10, k′1, k′2, k′3, and k′4 mean the capacity factors of Ez, (RRS)‐Ez, (RSR)‐Ez, and (SSR)‐Ez. Rs12, Rs23
,
and Rs34 mean the resolutions between Ez and (RRS)‐Ez, (RRS)‐Ez and (RSR)‐Ez, and (RSR)‐Ez and (SSR)‐Ez, respectively. The corresponding selectivity factors
were indicated as α12, α23, and α34. But for n‐hexane/THF 90/10, k′1, k′2, and k′3 mean the capacity factor of Ez/(RRS)‐Ez (eluted as one peak), (SSR)‐Ez, and
(RSR)‐Ez, respectively. Rs12 and Rs23 mean the resolutions between Ez/(RRS)‐Ez and (SSR)‐Ez and (SSR)‐Ez and (RSR)‐Ez, respectively. α12 and α23 mean
the corresponding selectivity factor.

8
ZHU ET AL.
FIGURE 3 Typical high‐performance liquid chromatography
chromatogram of ezetimibe and its 3 stereoisomers. Column:
Chiralpak IC; mobile phase: N‐hexane/IPA (90/10, v/v); flow rate:
1.0 mL min−1
FIGURE 4 Typical high‐performance liquid chromatography
chromatogram of ezetimibe and its 3 stereoisomers. Column:
Chiralpak IC; mobile phase: n‐hexane/CH2Cl2/EtOH (50/50/1.5, v/
v/v); flow rate: 1.0 mL min−1
and third peaks were ascertained with pure enantiomers
of (SSR)‐Ez and (RSR)‐Ez. It was proved that Ez and
(RRS)‐Ez were co‐eluted as the first peak, and the second
and third peaks corresponded to (SSR)‐Ez and (RSR)‐Ez,
respectively. In contrast, the use of THF caused the 2
stereoisomers (SSR)‐Ez/(RSR)‐Ez to be eluted in the
reverse order as compared to those using standard mobile
phase of n‐hexane and IPA. As expected, IPA is a proton
donor that could form hydrogen bond with carbonyl on
the CSP, but THF, as the proton receptor, would interact
with amide on the CSP through hydrogen bond.24 The 2
modifiers of IPA and THF compete with the analytes for
different hydrogen bond binding sites on CSP, thus affect-
ing the enantioseparation in different ways. Then THF
was replaced with CH2Cl2, and 4 stereoisomers could be
resolved to their respective enantiomers. But the broader
peaks and poor symmetry were not negligible under this
condition. However, the addition of EtOH to n‐hexane/
CH2Cl2 mobile phase resulted in improved peak shapes
and better resolution. By adjusting the ratio of ternary
solvent system, 50/50/1.5 of n‐hexane/CH2Cl2/EtOH was
finally adopted to keep sound values of Rs12, Rs23, and
Rs34. Four stereoisomers could be effectively separated
from each other, and the resolutions were 4.32, 4.22, and
4.17 successively. The corresponding chromatogram is
listed as Figure 4. These results indicated that the
presence of alcohol in non‐standard mobile phase system
was beneficial for chiral separation.
on enantioseparation with respect to reversed phase mode
was remarkable, which was accordance with literature. In
the structure of the analyte (Figure 1), the presence of
hydroxyl groups and phenyl rings might be involved in
forming hydrogen bond and π‐π interaction with CSP.
But in case of reversed phase mode, the strong polarity
of mobile phase would compete with solutes for the
hydrogen bond binding sites, resulting in the drastic
decrease in hydrogen bond between solutes and CSP.
Conclusively, the chiral recognition efficiency of
Chiralpak IC was superior to other chiral columns. The
optimum mobile phase compositions and flow rates were
found to be ACN/H2O (35/65, v/v) and 0.7 mL min−1 for
reversed phase mode, n‐hexane/IPA (90/10, v/v) and
1.0 mL min−1 for standard normal phase mode, n‐hexane/
CH2Cl2/EtOH (50/50/1.5, v/v/v) and 1.0 mL min−1 for
nonstandard normal phase mode, and the corresponding
Rs of Ez and (RRS)‐Ez were 2.93, 4.02, and 4.32,
respectively.
3.2 | Validation results of the method
It was remarkable that chiral recognition capacity of
Chiralpak IC was high in comparison to other tested
columns toward Ez and RRS‐Ez, which would produce
satisfactory resolution under both standard and nonstan-
dard normal‐phase mode. Considering the fact that that
the use of CH2Cl2 is not environmental friendly and
supposed to be avoided in all production and quality con-
trol activities, standard mobile phase with n‐hexane/IPA
in the ratio of 90/10 (v/v) was selected and validated for
the separation of Ez and (RRS)‐Ez on Chiralpak IC
column. The method validation was carried out with
respect to suitability, specificity, linearity, precision, accu-
racy, stability, and robustness. The specificity of the ana-
lytical method was evaluated by the analysis of a
solution containing Ez and (RRS)‐Ez.
The chiral discrimination of enantiomers occurs when
they bind with the CSP forming transient diastereomeric
complexes. The phenylcarbamate groups on the CSP can
interact with the solutes through hydrogen bond using
the C═O and NH groups, through dipole‐dipole interac-
tions using C═O moiety, and through π‐π interaction
using phenyl groups.21,25-27 It is reported that these
interactions are relatively weak under reversed phase
mode.21 In this study, the superior of normal phase mode

ZHU ET AL.
9
TABLE 4 System suitability test results
Aa (RSD%)
Rsa (RSD%)
Na (RSD%)
Fa (RSD%)
a
Compound
tR (RSD%)
Ez
21.428 min (0.15)
5 434 987 (1.19)
—
8111 (0.57)
1.122 (0.18)
(RRS)‐Ez
27.908 min (0.15)
25 845 (1.32)
4.286 (0.39)
7631 (1.87)
1.064 (1.78)
aMean value of 6 replicates; tR, A, Rs, N, F mean retention time, peak areas, resolution, theoretical plates, and tailing factor, respectively.
The system suitability was assessed by 6 replicate
analyses of standard spiked solution. The RSD% of peak
areas and retention times for both Ez and (RRS)‐Ez were
within 2%. The efficiency of column as expressed by
number of theoretical plates, resolution, and tailing factor
has also been checked (Table 4). The results were
indicative of suitability for the purpose intended.
Good linearities were observed for Ez and (RRS)‐Ez
over the range of 0.2–1.2 μg mL−1 (0.2, 0.4, 0.6, 0.8, 1.0,
and 1.2 μg mL−1), with the linearity regression equation
y = 56100x − 705 (r = 0.9993) and y = 53600x − 1150
(r = 0.9999) for Ez and (RRS)‐Ez, respectively. The
coefficients of regression of the 2 calibration curves
(r > 0.999) revealed that an excellent correlation existed
between the peak area and concentration. The relative
response factor was found to be 1.05 (in the range of
0.9–1.1), indicating that main component self‐compare
without calibration factor would be used for the
determination of (RRS)‐Ez impurity.
FIGURE 5 A typical chromatogram of bulk drug containing
0.153% of (RRS)‐Ez. The experimental conditions were as indicated
in “2.3”
3.3 | Chiral analysis of bulk drug
The applicability of the proposed method in chiral analy-
sis was evaluated by determination of (RRS)‐Ez in 3
butches of Ez bulk drug. The quantification of (RRS)‐Ez
for 3 butches of bulk drug fell within the limit amount
of 0.5% as required (shown as Figure 5).
Then LOD and LOQ for (RRS)‐Ez were found to be
67 ng mL−1 and 200 ng mL−1, respectively. The RSD% of
6
consecutive injections at LOQ concentration of
(RRS)‐Ez was 9.6%. These results indicated that the
proposed method was sufficiently sensitive for the deter-
mination of (RRS)‐Ez impurity.
4 | CONCLUSION
With regard to precision, the RSD% of the peak area of
(RRS)‐Ez was 0.94% for repeatability and 1.8% for inter-
mediate precision, indicating the good precision of the
method.
In the accuracy study, standard addition method and
recovery experiments were carried out by analyzing
standard spiked solution. The recovery ranged from
96.9% to 103.8%. The overall percent recovery was 100.6%
(RSD 2.3%) for (RRS)‐Ez.
The results of solution stability was that the area
obtained for the 2 isomers after 10 h did not show any
significant change compared to the area of initial analysis.
This indicated that both isomers were stable in the mobile
phase for at least 10 hours when stored at room
temperature.
The resolution between Ez and (RRS)‐Ez was used to
evaluate the method robustness under the above‐men-
tioned varied conditions. The resolution of Ez and
(RRS)‐Ez was always greater than 3.87 under all separa-
tion conditions tested, illustrating sufficient robustness
of the method.
In this paper, a normal phase chiral HPLC method has
been developed and validated for the quantitative deter-
mination of Ez and its stereoisomers on Chiralpak IC
column. During the development and optimization of
the chromatographic conditions, different stationary
phases and mobile phases were used to optimize the
enantioseparation. In contrast, better chiral recognition
and resolution were noticed on Chiralpak IC column. In
normal and reversed phase modes, complete resolution
of Ez and (RRS)‐Ez was achieved on Chiralpak IC.
Besides, the other 2 process‐related impurities of Ez could
also be completely separated in normal phase mode. The
use of CH2Cl2 as nonstandard mobile phases was tested
and satisfactory resolution was obtained. The separation
mechanisms caused by different organic modifiers were
discussed and found that hydrogen bond and π‐π interac-
tion between the CSP and solutes played the crucial role
in the chiral resolution of Ez stereoisomers. The proposed
method was validated, and it was successfully applied for
the quantitative determination of (RRS)‐Ez impurity of Ez
bulk drug.

10
ZHU ET AL.
15. Yashima E. Polysaccharide‐based chiral stationary phases for
high‐performance liquid chromatographic enantioseparation.
J Chromatogr A. 2001;906:105‐125.
ORCID
Qing Li http://orcid.org/0000-0003-3087-4692
16. Wang D‐M, Li F, Jiang Z, Yu L, Guo X‐J. Chiral recognition
mechanisms of four β‐blockers by HPLC with amylose chiral sta-
tionary phase. Iran J Pharm Res. 2014;13:449‐457.
REFERENCES
1. Phan BAP, Dayspring TD, Toth PP. Ezetimibe therapy: mecha-
nism of action and clinical update. Vasc Health Risk Manag.
2012;8:415‐427.
17. Wang T, Chen Y‐W. Application and comparison of derivatized
cellulose and amylose chiral stationary phases for the separation
of enantiomers of pharmaceutical compounds by high‐perfor-
mance liquid chromatography. J Chromatogr A. 1999;855:411‐421.
2. Rosen JB, Jimenez JG, Pirags V, et al. A comparison of efficacy
and safety of an ezetimibe/simvastatin combination compared
with other intensified lipid‐lowering treatment strategies in
diabetic patients with symptomatic cardiovascular disease. Diab
Vasc Dis Res. 2013;10:277‐286.
18. Kaida Y, Okamoto Y. Optical resolution by high‐performance
liquid chromatography on benzycarbamates of cellulose and
amylose. J Chromatogr A. 1993;641:267‐278.
19. Kartozia I, Kanyonyo M, Happaerts T, Lambert DM, Scriba GK,
Chankvetadze B. Comparative HPLC enantioseparation of new
chiral hydantoin derivatives on three different polysaccharide type
chiral stationary phases. J Pharm Biomed Anal. 2002;27:457‐465.
3. Sistla R, Tata VS, Kashyap YV, Chandrasekar D, Diwan PV.
Development and validation of a reversed‐phase HPLC method
for the determination of ezetimibe in pharmaceutical dosage
forms. J Pharm Biomed Anal. 2005;9:517‐522.
20. Geryk R, Kalíková K, Vozka J, Plecitá D, Schmid MG, Tesařová
E. Enantioselective potential of chiral stationary phases based
on immobilized polysaccharides in reversed phase mode.
J Chromatogr A. 2014;1363:155‐161.
4. Li Z‐D, Shi X‐J, Wang B, et al. Determination of ezetimibe and
total ezetimibe in human plasma and its pharmacokinetics. Chin
Pharm J. 2007;42:531‐535.
5. Baghdady YZ, AlGhobashy MA, Abdel‐Aleem AA, Weshahy SA.
Spectrophotometric and TLC‐densitometric methods for the
simultaneous determination of ezetimibe and atorvastatin cal-
cium. J Adv Res. 2013;4:51‐59.
21. Tachibana K, Ohnishi A. Reversed‐phase liquid chromato-
graphic separation of enantiomers on polysaccharide type
chiral stationary phases. J Chromatogr A. 2001;96:127‐154.
22. Chankvetadze B, Yashima E, Okamoto Y. Dimethyl‐, dichloro‐
and chloromethylphenylcarbamates of amylose as chiral sta-
tionary phases for high‐performance liquid chromatography.
J Chromatogr A. 1995;694:101‐109.
6. Filip K, Bańkowski K, Sidoryk K, et al. Physicochemical charac-
terization of ezetimibe and its impurities.
J Mol Struct.
2011;991:162‐170.
23. Kasat RB, Wang NH, Franses EI. Effects of backbone and side
chain on the molecular environments of chiral cavities in
polysaccharide‐based biopolymers. Biomacromolecules. 2007;8:
1676‐1685.
7. Radhakrishnan P, Rao DVS. A validated LC method for the
enantiomeric separation of ezetimibe in bulk drug and pharma-
ceutical dosage forms. Indian Drugs. 2009;46:315‐322.
8. Sun L‐L, Zou Q‐G, Wei P, Ouyang P‐K. Chiral separation and
thermodynamic investigation of ezetimibe optical isomers on a
chiralpak IC column. J Chromatogr Sci. 2016;54:1‐6.
24. Wang T, Jr WR. Effects of alcohol mobile‐phase modifiers
on the structure and chiral selectivity of amylose tris(3,
5‐dimethylphenylcarbamate)
J Chromatogr A. 2003;1015:99‐110.
chiral
stationary
phase.
9. Zhang Z‐F, Yang G‐L, Liang G‐J, Liu H‐Y, Chen Y. Chiral sepa-
ration of tamsulosin isomers by HPLC using cellulose tris(3, 5‐
dimethylphenylcarbamate) as a chiral stationary phase. J Pharm
Biomed Anal. 2004;34:689‐693.
25. Ali I, Aboul‐Enein HY. Impact of immobilized polysaccharide
chiral stationary phases on enantiomeric separations. J Sep Sci.
2006;29:762‐769.
10. Okamoto Y, Kaida Y. Polysaccharide derivatives as chiral sta-
26. O'Brien T, Crocker L, Thompson R, et al. Mechanistic aspects of
chiral discrimination on modified cellulose. Anal Chem.
1997;69:1999‐2007.
tionary phases in HPLC. J Sep Sci. 2005;13:708‐712.
11. Franco P, Senso A, Oliveros L, Minguillón C. Covalently bonded
polysaccharide derivatives as chiral stationary phases in high‐
performance liquid chromatography. J Chromatogr A. 2001;
906:155‐170.
27. Chankvetadze B, Chankvetadze L, Sidamonidze S, Yashima E,
Okamoto Y. High performance liquid chromatography
enantioseparation of chiral pharmaceuticals using tris (chloro‐
methylphenylcarbamate) s of cellulose. J Pharm Biomed Anal.
1996;14:1295‐1303.
12. Zhang T, Nguyen D, Franco P. Reversed‐phase screening strate-
gies for liquid chromatography on polysaccharide‐derived chiral
stationary phases. J Chromatogr A. 2010;1217:1048‐1055.
13. Thunberg L, Hashemi J, Andersson S. Comparative study of
coated and immobilized polysaccharide‐based chiral stationary
phases and their applicability in the resolution of enantiomers.
J Chromatogr B. 2008;875:72‐80.
How to cite this article: Zhu B, Yao Y, Zhao Y,
Sun T, Li Q. Study on the HPLC‐based separation of
some ezetimibe stereoisomers and the underlying
stereorecognition process. Chirality. 2018;1–10.
https://doi.org/10.1002/chir.22829
14. Ikai T, Yamamoto C, Kamigaito M, Okamoto Y. Immobilized‐
type chiral packing materials for HPLC based on polysaccharide
derivatives. J Chromatogr B. 2008;875:2‐11.