Talanta 146 (2016) 83–92

Contents lists available at ScienceDirect

Talanta

http://d0039-91

n CorrE-m

journal homepage: www.elsevier.com/locate/talanta

Novel liquid chromatographic methods for the determination ofvarenicline tartrate

Ramzia I. El-Bagary a,b, Nisreen F. Abo-Talib c, Marwa A. El-Wahab Mohamed c,n

a Pharmaceutical Chemistry Department, Faculty of Pharmacy, Cairo University, Cairo 11562, Egyptb Pharmaceutical Chemistry Deparment, Faculty of Pharmacy, Future University in Egyptc National Organization for Drug Control and Research (NODCAR), P.O. Box 29, Giza 35521, Egypt

a r t i c l e i n f o

Article history:Received 10 January 2015Received in revised form23 July 2015Accepted 28 July 2015Available online 30 July 2015

Keywords:Varenicline tartrateDiode array detectorFluorescence detectorNBD-Cl

x.doi.org/10.1016/j.talanta.2015.07.07740/& 2015 Elsevier B.V. All rights reserved.

esponding author. Fax: þ20 2 33379445.ail address: marsmubarak@yahoo.com (M.A. E

a b s t r a c t

Two simple, sensitive, rapid, and stability-indicating liquid chromatographic (LC) methods have beendeveloped for the determination of varenicline tartrate. They comprised the determination of varenicline(VRC) in the presence of its oxidative degradates and related impurity (N-formyl varenicline) (NFV). Thefirst method was a LC with diode array detection (DAD) at 235 nm using Ristek-Ultras C18 column(100 mm�2.1 mm, 5 mm). Isocratic elution of VRC was employed using a mobile phase consisting ofbuffer mixture (1.2% potassium dihydrogen phosphate and 0.08% octane sulphonic acid): acetonitrile(86:14, v/v), pH (5.0). In the second method; a fluorimetric detection technique was developed, based onprecolumn derivatization of VRC using 7-chloro-4-nitrobenzo-2-oxa-1, 3-diazole (NBD-Cl). The fluores-cence detector (FLD) was operated at 474 nm for excitation and 539 nm for emission. Isocratic elutionwas applied with a mobile phase consisting of methanol-distilled water (70:30, v/v). Separation wasachieved using Symmetrys Waters C18 column (150 mm�4.6 mm, 5 mm). Linearity, accuracy and pre-cision were found to be acceptable over the concentration ranges of 0.5–20.0 mg mL�1 and 0.2–20.0 mg mL�1 with the first and the second method, respectively. The optimized methods were validatedand proved to be specific, simple, and accurate for the quality control of the drug in its pharmaceuticalpreparation.

& 2015 Elsevier B.V. All rights reserved.

1. Introduction

Varenicline tartarate, [7, 8, 9, 10-Tetrahydro-6, 10-methano-6Hpyrazino [2, 3-h] [3] benzazepine (2R, 3R) -2,3dihydroxybutanedioate](Fig. 1a) is a novel drug used in the treatment of smoking cessation[1]. VRC is a partial agonist on α4β2 nicotinic acetylcholine receptor. Itacts by decreasing the degree of cravings and withdrawal symptomsduring the period of smoking cessation [2].

In vivo, studies revealed that in the presence of nicotine, VRCacts as a partial agonist by stimulating the release of dopamineand simultaneously blocking the nicotine receptors [3,4]. VRC isconsidered the most efficient option for smoking cessation whencompared to other pharmacotherapies (e.g., nicotine replacementtherapy and bupropion) [5].

Fewmethods have been described for the determination of VRCin its pharmaceutical preparation using spectrophotometry [6],HPLC [7–9] and LC/MS/MS in plasma [10–13]. Pharmacokineticstudies have been established on VRC [14,15]. Reports on VRC

l-Wahab Mohamed).

related impurities and degradation products have been publishedusing different chromatographic methods [16–21]. In addition,some techniques, such as capillary zone electrophoresis [22] andelectrochemical studies [23,24] have been reported for the de-termination of VRC.

The present work presents two LC methods for the determi-nation of VRC in the presence of either its complete oxidativedegradates or NFV (Fig. 1b–e). Their structures were confirmed byLC/MS/MS. In addition, LC–FLD method was the first fluorimetricmethod used for the determination of VRC based on precolumnderivatization using NBD-Cl. An ideal stability-indicating methodis the one that quantifies the drug as well as efficiently resolves itsdegradation products and related impurities.

Therefore, our aim in this work was to develop simple, accu-rate, specific and reproducible stability-indicating methods for thedetermination of VRC in the presence of its possible degradationproducts and its related impurity. The proposed methods weredeveloped, validated and compared to the manufacturer HPLCmethod using mobile phase consisted of water: 0.3% concentratedphosphoric acid: 0.2% triethylamine (1990: 6: 4), detection at320 nm, flow rate: 1 mL min�1 using C18 column(150 mm�3.9 mm, 5 mm) [25].

HN

N

N

HOOH

O

O

OH

OH

N

N*

*

N

N

N*

O

N

N

N

O

O

N

N

NO

H

d

Fig. 1. Chemical structures of (a) varenicline tartrate, (b) suggested oxidative de-gradate [1], (c) suggested oxidative degradate [2], (d) suggested oxidative de-gradate [3] and (e) N-formyl varenicline.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–9284

2. Experimental

2.1. Instrumentation

� Two HPLC systems were used, For LC–DAD method, Agilent1200 SL series (Agilent, Germany) which comprised binarypump SL model G1312B, autosampler injector ALS SL modelG1329B and DAD SL model G1315C. Separation was performedon a Risteks-Ultra C18 column (100 mm�2.1 mm, 5 mm). ForLC–FLD method, Agilent 1200 series (Agilent, Germany)equipped with quaternary pump model G1311A, autosamplerinjector model G1329A and FLD model G1321A was used. Se-paration was performed on a Symmetrys Waters C18 column(150 mm�4.6 mm, 5 mm).

� Analytical balance (Mettler Toledo, Switzerland) model AB265-S.

� pH meter (Precisa, Switzerland) model pH 900.� Sonicator (Bandelin-Sonorex, Germany) TK 52.� Water bath (Memmert,Germany) model WNB 14.� Variable micropipette (Advantage-Lab, 1000 mL).

2.2. Chemicals and solvents

Acetonitrile and methanol were HPLC grade (Scharlau, Spain),octane sulphonic acid (SDFCL, India), potassium dihydrogenphosphate (Adwic, Egypt), hydrogen peroxide 30% (Panreac, EU),

hydrochloric acid (Riedel–de Haën, Germany), and ammonia(Fisher scientific, UK) were used. Distilled water was produced in-house (Aquatron water still, UK) model A4000D. A solution of 0.2%w/v of NBD-Cl (Sigma, USA) was freshly prepared in methanol.Clark and lubs buffer solution of pH 8.0 was prepared by mixing50 mL of 0.2 mol L�1 aqueous solution of boric acid and potassiumchloride (1 L contains 12.368 g of boric acid and 14.90 g of po-tassium chloride) with 21.3 mL of 0.2 mol L�1 sodium hydroxide(Qualikems, India) in 200 mL standard flask [26], and adjusted bypH meter. All other chemicals and reagents used were of analyticalgrade unless indicated otherwise.

2.3. Samples

Varenicline tartrate was assessed according to the manu-facturer method and its purity was found to be 99.27%71.45.Varenicline related impurity (NFV) CP-697,535 [27] and Champixs

1 mg tablets (batch no: B10572339) labeled to contain 1.71 mgvarenicline tartrate equivalent to 1 mg VRC. They were all suppliedby (Pfizer, Egypt).

2.4. Preparation of VRC solutions

All solutions (VRC, VRC degradates and NFV) were freshlyprepared and protected from light. All calculations were per-formed regarding VRC free base.

A stock standard solution of VRC (100.0 mg mL�1) was preparedin distilled water and used in LC–DAD method. For the LC–FLDmethod, a stock standard solution of VRC (400.0 mg mL�1) wasprepared in methanol. Then further dilutions were made in me-thanol to obtain working standard solutions of VRC; for con-struction of calibration curve (2.0–200.0 mg mL�1) and laboratoryprepared mixtures with either its oxidative degradates or NFV(360.0–40.0 mg mL�1).

2.5. Preparation of forced degradation solutions

For determination of the applicability of the developed analy-tical method as a stability-indicating one, a forced degradationstudy under different conditions was carried out in which VRC wassubjected to the following conditions:

(a) Oxidation with H2O2: VRC was treated with 50 mL of 30% (w/v) hydrogen peroxide and heated in a thermostatic water bathat 90 °C for 4 h to reach complete degradation. Evaporation tilldryness was performed to remove excess hydrogen peroxide,then dissolved in distilled water to obtain stock solution ofVRC oxidative degradates equivalent to 1.0 mg mL�1. Subse-quently further dilution was made in distilled water to obtainsolution of concentration of 100.0 mg mL�1 which was useddirectly for LC–DAD method. While for LC–FLD method, dilu-tions were prepared in methanol to get working oxidativedegraded solutions of (40.0–360.0 mg mL�1) for laboratoryprepared mixtures.

(b) Alkaline conditions: VRC was treated with 10 mL of 5 N-NaOHand heated in a thermostatic water bath at 90 °C for 9 h andthen neutralized by adjusting the pH to 7.0 with 5 N-HCl.

(c) Acidic conditions: VRC was treated with 10 mL of 5 N-HCl andheated in a thermostatic water bath at 90 °C for 9 h and thenneutralized by adjusting the pH to 7.0 with 5 N-NaOH.

The prepared solutions under alkaline and acidic conditionswere completed with distilled water to obtain solutions of con-centration of 100.0 mg mL�1.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–92 85

2.6. Preparation of NFV solutions

For LC–DAD method, a stock solution (100.0 mg mL�1) of NFVwas prepared in distilled water. While for LC–FLD method, a stocksolution (400.0 mg mL�1) of NFV in methanol was prepared andthen further dilutions were prepared in methanol to get workingNFV solutions of (40.0–360.0 mg mL�1) for laboratory preparedmixtures.

2.7. Chromatographic conditions

2.7.1. LC–DAD methodA mobile phase was prepared, consisting of buffer mixture

(1.2% potassium dihydrogen phosphate and 0.08% octane sul-phonic acid) and acetonitrile in the ratio of (86:14, v/v), pH (5.0).The mobile phase was filtered through 0.45 mm membrane filter,degassed and then it was pumped through the column at a flowrate of 0.8 mL min�1. The DAD was operated at 235 nm. Analysiswas performed at ambient temperature and the injection volumewas 5 mL.

2.7.2. LC–FLD methodA mobile phase was prepared, consisting of methanol and

distilled water in the ratio of (70:30, v/v). The mobile phase wasfiltered through 0.45 mm membrane filter, degassed, and pumpedthrough the column at a flow rate of 1 mL min�1. The FLD wasoperated at 474 nm for excitation and 539 nm for emission. Ana-lysis was performed at ambient temperature with an injectionvolume of 20 mL.

2.8. Construction of calibration curves

2.8.1. LC–DAD methodA series of standard solutions were prepared by diluting VRC

stock standard solution (100.0 mg mL�1) with distilled water toreach a concentration range of (0.5–20.0 mg mL�1). Five microliterswere injected for each concentration in triplicate and chromato-graphed under the previously described conditions. The calibra-tion curve was constructed by plotting the peak area against thecorresponding concentration of VRC and the regression equationwas then computed.

2.8.2. LC–FLD methodOne milliliter of each of working standard solutions (2.0–

200.0 mg mL�1) was accurately transferred into a series of 10 mLvolumetric flasks. To each flask 1.6 mL of Clark and Lubs buffer pH8.0 was added followed by the addition of 0.4 mL of (0.2% w/v)NBD-Cl solution. The solutions were heated for 30 min at 70 °C71. The reaction was stopped by cooling under tap water. Onemilliliter of 5 N-HCl was added to each flask and completed tovolume with a mixture of methanol: distilled water (50:50) toobtain solutions over a concentration range of (0.20–20.0 mg mL�1). Twenty microliters were injected for each con-centration in triplicate and chromatographed under the previouslydescribed conditions. The calibration curve was constructed byplotting the peak area against the corresponding concentration ofthe derivatized VRC and the regression equation was thencomputed

2.9. Preparation of laboratory prepared mixtures for stability-in-dicating methods

2.9.1. LC–DAD methodAliquot portions (1.8–0.2 mL) of VRC standard stock solution

(100.0 mg mL�1) were transferred into two series of 10 mL volu-metric flasks, and aliquots (0.2–1.8 mL) from either the oxidative

degraded solution (100.0 mg mL�1) or NFV solution(100.0 mg mL�1) each, were added separately to the same flasksand the volumes were completed to the mark with distilled waterto prepare different mixtures containing 10–90% of either theoxidative degradates or NFV.

2.9.2. LC–FLD methodHalf milliliter of VRC working standard solutions (360.0–

40.0 μg mL�1) were transferred to two series of 10 mL volumetricflasks and half milliliter of either the working oxidative degradatesor NFV solutions (40.0–360.0 μg mL�1) were added separately tothe same flasks to prepare different mixtures containing 10–90% ofeither the oxidative degradates or NFV and then proceeded asdirected under Section 2.8.2.

2.10. Application to pharmaceutical preparation (Champixs 1 mgtablets)

Twenty tablets of 1 mg of Champixs tablets were finely pow-dered, divided into two portions each equivalent to 10 mg of VRCand transferred separately into 100 mL and 25 mL volumetricflasks. One milliliter, 0.25 mL of ammonia solution were added to100 mL and 25 mL volumetric flasks, respectively. Sonication for30 min with distilled water for LC–DAD method and methanol forLC–FLD method. Then the volumes were completed with the samesolvents mentioned and filtered to prepare stock solutions havingconcentrations of 100.0 μg mL�1 and 400.0 μg mL�1 for LC–DADand LC–FLD methods, respectively.

For LC–DAD method, aliquots of 0.2, 0.4 and 0.8 mL weretransferred from the prepared stock solution to 10 mL volumetricflasks and diluted to volume with distilled water. Five microlitersfrom each final solution were injected in triplicate to obtain con-centrations of 2.0, 4.0 and 8.0 μg mL�1. While for LC–FLD method,aliquots of 1.5, 2.5 and 4 mL were transferred from the preparedstock solution to 10 mL volumetric flasks and the volumes werecompleted with methanol, half milliliter of each of these solutionswas transferred to 10 mL volumetric flasks and then proceed asdirected under Section 2.8.2. Twenty microliters from each finalsolution were injected in triplicate to obtain concentrations of 3.0,5.0 and 8.0 μg mL�1

The general procedure described above for each method wasfollowed and the concentration of VRC in its pharmaceutical pre-paration was calculated. The same procedures were repeated ap-plying standard addition technique for both methods.

3. Results and discussion

Development of analytical methods for the determination ofpharmaceuticals in the presence of their degradation productswithout previous chemical separation is always a matter of in-terest. The main task of this work was to establish simple, sensitiveand accurate analytical methods for the determination of VRC inthe presence of both its oxidative induced degradation productsand its related impurity (NFV), in its bulk powder and commercialtablets with satisfactory precision for good analytical practice(GAP).

Under alkaline and acidic stress conditions, only 0.29% and0.31% of VRC was degraded and no degradation was observed incase of oxidation on the basis of the studies of Kadi et al. [7]. Lit-erature survey revealed that 3.54%, 10.54% and 24% of VRC wasdegraded under different oxidative conditions by hydrogen per-oxide [16–18].

In our study, complete degradation of VRC with 30% H2O2 wasachieved by heating in water bath at 90 °C for 4 h (Fig. 2a). Milddegradation of VRC was observed under alkaline and acidic

Fig. 2. Typical chromatograms of LC–DAD method (a) Mixture of 10.0 (mg mL�1) of varenicline, N-formyl varenicline and its oxidative degradates each, (b) 10.0 (mg mL�1) ofvarenicline with 5 N-NaOH showing incomplete degradation, (c) 10.0 (mg mL�1) of varenicline with 5 N-HCl showing incomplete degradation and (d) 10.0 (mg mL�1) ofcomplete oxidative degradates.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–9286

conditions, only 10.13% and 14% of was degraded by 5 N-NaOH and5 N-HCl at 90 °C for 9 h, respectively as shown in (Fig. 2b and c).Thus, LC–DAD method is considered the first stability- indicatingLC method for the determination of VRC in the presence of eitherits oxidative degradates (completely degradated) or NFV.

No reported methods were developed for the determination ofVRC using LC–FLD, thus its derivatization with fluorogenic reagentwas necessary to increase its detectability. NBD-Cl has been usedas a useful derivatizing reagent for primary amines, secondaryamines, thiols, etc. [28–33]. So the development of a fluorimetricstability- indicating method was achieved, through precolumnderivatization of VRC with NBD-Cl, producing a yellow fluorescent

complex (Fig. 3a). While no reaction was formed with either itsoxidative degradates or NFV (Fig. 3b and c).

3.1. Mass spectra of VRC oxidative degradates

Under peroxide stressed conditions, DP-I (4, 6, 7, 8, 9, 10-hex-ahydro-1H-6, 10- methano- pyrazino [2, 3-h] [3] benzazepine- 2,3-dione) was formed up to 24% showing a molecular ion peak atm/z: 244 as suggested by Satheesh et al. [18]. This was confirmedin our work to be one of the oxidative degradates which is iden-tical to the molecular weight of VRC oxidative degradation product[Deg. 3]. LC/MS/MS was used to identify VRC oxidative degradates

Fig. 3. Typical chromatograms of LC–FLD method (a) 10.0 (mg mL�1) of derivatized varenicline, (b) 10.0 (mg mL�1) of derivatized oxidative degradates showing no complexformation and (c) 10.0 (mg mL�1) of derivatized N-formyl varenicline showing no complex formation.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–92 87

(Fig. 4). The mass ion peak of intact VRC was identified at m/z: 212shown in (Fig. 4a). While those of the oxidative degradates were atm/z: 130, 228 and 244 (Fig. 4b–d) which were identical to themolecular weight of the oxidative degradation products (Deg. 1,2 and 3), respectively.

Therefore, one can conclude that carrying out oxidative stresscondition may proceed as shown in (Fig. 5).

3.2. Confirmation of VRC related impurity (NFV)

IR spectrum of intact VRC showed a broad peak of (NH) at3404 cm�1 which disappeared at the NFV spectrum and the ap-pearance of amide carbonyl peak in NFV spectrum at 1660 cm�1 asshown in (Supplementary data).

3.3. Method optimization

3.3.1. LC–DAD methodA simple isocratic HPLC method was developed for the de-

termination of VRC in its pure powder, in its pharmaceutical pre-paration, in the presence of its possible degradation products andNFV using Ristek- Ultras C18 column (100 mm�2.1 mm, 5 mm).The mobile phase used, was a buffer mixture consisting of (1.2%potassium dihydrogen orthophosphate and 0.08% octane sul-phonic acid) and acetonitrile, pH (5.0) in the ratio of (86:14, v/v).The mobile phase was chosen after several trials to reach the op-timum stationary/ mobile phase matching. By increasing acetoni-trile percentage, the degradate peaks enter with each other givingvery poor resolution. By increasing buffer percentage bad resolu-tion was obtained. Reasonable separationwith good resolution in ashort run time was obtained upon using a flow rate of

Fig. 4. Mass spectra of (a) varenicline standard and its three oxidative degradates where (b) Deg. [1], (c) Deg. [2] and (d) Deg. [3].

R.I. El-Bagary et al. / Talanta 146 (2016) 83–9288

0.8 mL min�1. System suitability parameters were tested by cal-culating retention factor, tailing factor, number of theoreticalplates, selectivity factor and resolution. Under the optimumchromatographic conditions, VRC, NFV and its three oxidativedegradates were eluted at 3.098 min, 2.107 min and (1.226, 1.396and 1.807 min), respectively as shown in Fig. 2a. The analyte wasstable in the diluent for at least 24 h. Therefore, the chromato-graphic system described in this work allows complete separationof VRC from its possible degradation products and NFV. Calibrationcurve was obtained by plotting the peak area against the con-centration of VRC. Linearity range was found to be 0.5–20.0 mg mL�1 using the following regression equation:

Y X r31.003 3.0467 0.99992= + =where Y is the peak area, X is the concentration of VRC in mg mL�1

and r2 is the regression coefficient.

3.3.2. LC–FLD methodA simple isocratic HPLC method was developed for the de-

termination of VRC in its pure powder, in its pharmaceutical pre-paration, in the presence of its possible degradation products andNFV using Symmetrys Waters C18 column (150 mm�4.6 mm,5 mm). The mobile phase used, was a mixture of methanol anddistilled water (70:30, v/v). The mobile phase was chosen afterseveral trials to reach the optimum stationary/ mobile phasematching. By increasing the percentage of methanol, poor re-solution was found. While by increasing the percentage of distilledwater, long run time and bad peak resolution was found. Rea-sonable separation with good resolution in a short run time wasobtained upon using a flow rate of 1 mL min�1. System suitability

Possible pathwaysOxidation via 30 %H2O2

HN

N

N

N

N

Chemical Formula: C8H6N2•+

**

[ Deg.1] [ Deg. 2]

[ Deg. 3]

Pathway [2]

Pathway [3]

N

N

N*O

Chemical Formula: C13H14N3O

Pathway[1]

N

N

N

O

O

Chemical Formula: C13H13N3O2

Fig. 5. Suggested schematic diagram of varenicline degradation by 30% H2O2 showing its three degradative products (Pathway: 1, 2 and 3).

R.I. El-Bagary et al. / Talanta 146 (2016) 83–92 89

parameters were tested by calculating retention factor, tailingfactor, number of theoretical plates, selectivity factor and resolu-tion. Under the optimum chromatographic conditions, the deri-vatized VRC complex appeared at 2.782 min while blank showed apeak at 1.309 min (Fig. 3a). No complex was formed with eitherthe oxidative degradates or NFV (Fig. 3b and c). Calibration curvewas obtained by plotting the peak area of the derivatized drugagainst the concentration of VRC. Linearity range was found to be0.2– 20.0 mg mL�1 using the following regression equation:

Y X r17.61 1.6929 0.99982= + =where Y is the peak area of the derivatized drug, X is the

concentration of VRC in mg mL�1 and r2 is the regressioncoefficient.

Different experimental parameters affecting the color devel-opment and its stability were carefully studied and optimized toobtain maximum color intensity.

3.3.2.1. Effect of concentration and volume of NBD-Cl. The con-centration of NBD-Cl solution on the fluorescence intensity (FI)was studied over the range 0.025–0.5% (w/v). It was found that0.4 mL of 0.2% NBD-Cl solution was chosen as to fulfill the opti-mum reaction requirement of VRC.

3.3.2.2. Effect of type, pH and the volume of buffer. Different buffersystems (phosphate, carbonate and Clark and Lubs) were tested.Highest FI was obtained with Clark and Lubs buffer. With otherbuffers either precipitation of white colloid or weak sensitivitieswere observed. This result is in agreement with that of Miyanoet al. [34]. The influence of pH on the reaction was studied bycarrying out the reaction with Clark and Lubs buffer solution overpH values of (7.0–10.0). Best results were obtained with pH 8.0. Athigher pH, sharp decrease in the readings occurred. This was at-tributed probably to the increase for hydroxide ion that holds backthe condensation reaction between VRC and NBD-Cl. The additionof 1.6 mL of Clark and lubs buffer pH 8.0 gave the best readings.

3.3.2.3. Effect of heating temperature and reaction time. The effect oftemperature on FI was studied in the range from 50 to 80 °C fordifferent period. It was found that the most reproducible resultswere attained after heating at 70 °C for 30 min. At a higher tem-perature and a longer heating time, dramatic decrease in the FIvalues were observed.

3.3.2.4. Effect of concentration and volume of HCl. NBD-Cl is hy-drolyzed in alkaline medium to the corresponding hydroxyl

derivative namely, 4-hydroxy-7-nitrobenzo-2-oxa-1, 3-diazole(NBD-OH).Therefore, it was necessary to acidify the reactionmixture to pH 2.0 before carrying out the measurements to de-crease the fluorescence background [35]. The required amount ofHCl for acidification was found to be 1 mL of 5 N-HCl.

3.3.2.5. Effect of diluting solvent. Different diluting solvents weretested to determine the most appropriate one: distilled water,methanol, methanol: distilled water (50:50, v/v), ethanol, acet-onitrile, isopropanol and acetone. A mixture of methanol: distilledwater (50:50, v/v) was found to be the most suitable solventproviding the maximum FI.

3.3.2.6. Stability of the fluorescent derivative. Regarding the stabi-lity of the VRC-NBD complex, FI was stable for at least 3 h.

3.4. Analysis of pharmaceutical preparation

The proposed methods were applied for the determination ofVRC in Champixs 1 mg tablets. Satisfactory results with goodagreement were found. The mean percentage recoveries7S.D. ofthe labeled VRC in Champixs 1 mg tablets were 99.8670.54 and98.9770.33 for LC–DAD and LC–FLD, respectively. While that ofthe added VRC were 100.7370.60 and 99.8371.20 for bothmethods, respectively. The results were summarized as shown inTable 1.

3.5. Method validation

The methods were validated according to the InternationalConference on Harmonization guidelines for validation of analy-tical procedures [36,37].

3.5.1. LinearityThe linearity of both chromatographic methods for the de-

termination of VRC was evaluated by analyzing a series of differentconcentrations of the drug. In LC–DAD method, six concentrationswere chosen ranging from 0.5 to 20.0 mg mL�1. While in LC–FLDmethod, seven concentrations were chosen ranging from 0.2 to20.0 mg mL�1. Linear relationships between the peak areas and theconcentrations of VRC were obtained for both methods. Char-acteristic parameters for regression equations of the adoptedchromatographic methods are given in Table 1 [38].

Table 1Results obtained by LC–DAD and LC–FLD methods for the determination of VRC inbulk and pharmaceutical formulation.

Parameters LC–DAD method LC–FLD method

Calibration range 0.5–20 (mg mL�1) 0.2–20 (mg mL�1)Intercept(a) 3.0467 1.6929Slope(b) 31.003 17.61Limit of detection (LOD) 0.038 (mg mL�1) 0.019 (mg mL�1)Limit of quantitation (LOQ) 0.113 (mg mL�1) 0.059 (mg mL�1)Standard deviation of intercept (Sa) 1.680 1.347Standard deviation of slope (Sb) 0.150 0.120Regression coefficient (r2) 0.9999 0.9998Confidence limit of intercept 3.046774.663 1.692973.464Confidence limit of slope 31.00370.416 17.61 70.309Standard error of estimation 2.573 2.245Drug in dosage form (1 mg tablet) 99.8670.54 98.9770.33

Accuracy� Drug in bulk 99.9771.21 99.3270.92� Drug added 100.7370.60 99.83 71.20

Precisiona

� Intra-day (%R.S.D) 0.27–1.03 0.26–0.38� Inter-day (%R.S.D) 0.15–0.91 0.51–0.61

a The inter-day and the intra-day (n¼3), average of three concentrations (4, 8,12 μg mL�1) for LC- DAD and (2, 10, 18 mg mL�1) for LC-FLD method.

Table 2Determination of VRC in laboratory-prepared mixtures with either its oxidativedegradates or NFV by the two proposed methods.

Method % degorNFV

VRCconc.

Deg orNFVConc.

Recovery % ofVRC with oxida-tive degradates

Recovery % ofVRC with NFV

LC–DAD 10 18 2 99.56 100.4930 14 6 100.36 98.5850 10 10 98.21 100.2970 6 14 100.73 100.1590 2 18 98.70 98.85

Mean7S.D. 99.5171.07 99.6770.89

LC–FLD 10 18 2 99.28 98.8530 14 6 98.59 99.2550 10 10 98.15 99.6870 6 14 100.12 98.9790 2 18 101.75 100.40

Mean7S.D. 99.3871.54 99.4370.63

Fig. 6. Purity spectrum of 10.0 (mg mL�1) of varenicline in LC–DAD method.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–9290

3.5.2. AccuracyAccuracy of the results was calculated by the mean percentage

recovery7S.D. of six different concentrations of VRC. The inter-ference of excipients in the pharmaceutical formulation was alsostudied using the proposed methods. For this reason, standardaddition method was applied to the commercial pharmaceuticalformulation containing VRC. In application of standard additionmethod the mean percentage recoveries and their standard de-viation for the proposed methods were calculated (Table 1). Ac-cording to the obtained results a good precision and accuracy wasobserved for the two methods. Consequently, the excipients inpharmaceutical formulation do not interfere in the analysis of VRCin its pharmaceutical formulation.

3.5.3. PrecisionIn order to judge the quality of the elaborated methods, pre-

cision was determined. For evaluation of the precision estimates,intra-day and inter-day were performed by repeating the assay ofthree different concentrations of VRC; in triplicate, three times inthe same day and assaying the same selected concentrations onthree successive days using the developed chromatographicmethods and calculating the RSD %. Results in Table 1 indicatesatisfactory precision of the proposed methods.

3.5.4. Detection and quantitation limitsLimit of detection (LOD) and limit of quantification (LOQ) were

calculated using the following equations according to the ICHguidelines. LOD¼3.3� s/S and LOQ¼10� s/S, where s is thestandard deviation of the response and S is the slope of the cali-bration curve as shown in Table 1. They confirmed good sensitivityfor the proposed methods and consequently their capability todetermine low amounts of the investigated drug.

3.5.5. SpecificityThe specificity of a method is the extent to which it can be used

for analysis of a particular analyte in a mixture or matrix withoutinterference from other components. The specificity of the pro-posed methods was tested by the analysis of five laboratory pre-pared mixtures containing different percentages of VRC at various

concentrations within the linearity range with its oxidative de-gradates produced in forced degradation studies or with its relatedimpurity (NFV). The laboratory prepared mixtures were analyzedaccording to the previous procedures described under each of theproposed methods. The specificity was demonstrated by thechromatograms recorded for mixtures of VRC with its oxidativedegradates or NFV indicating that the methods enabled highlyspecific analysis of the drug. Well-resolved peaks for VRC, oxida-tive degradates and NFV were observed (Figs. 2a, 3b and 3c). Sa-tisfactory results were obtained (Table 2) indicating the highspecificity of the proposed methods for determination of VRC inpresence of up to 90% of its oxidative degradates or NFV.

In addition, the specificity was enhanced using spectral datafrom DAD and peak purity information, confirming VRC puritysince the peak purity angle is smaller than the peak puritythreshold (Fig. 6).

3.5.6. System suitabilitySystem suitability test is an integral part of liquid chromato-

graphic methods in the course of optimizing the conditions of theproposed methods. The parameters of this test are: column effi-ciency expressed as number of theoretical plates (N), peak re-solution (R), peak tailing (T), selectivity factor (α), and retentionfactor (K) [39]. Good results are listed in Table 3.

3.5.7. Statistical analysisStatistical analysis of the results obtained by both methods and

Table 3System suitability tests for the determination of VRC in the presence of its oxidativedegradates and NFV by the two proposed methods.

Parameters LC–DADmethod

LC–FLDmethod

Reference value

Resolution (R)1st deg. 10.66 – R41.52nd deg. 9.60 –

3rd deg. 6.79 –

NFV 4.90 –

Tailing factor (T)VRC 0.938 0.947 r2 (T¼1 for a typical

symmetric peak)1st deg. 1.007 –

2nd deg. 0.931 –

3rd deg. 0.972 –

NFV 0.975 –

Retention factor (K) 1–10 AcceptableVRC 6.52 1.121st deg. 2.13 –

2nd deg. 2.59 –

3rd deg. 3.48 –

NFV 4.59 –

Selectivity (α)1st deg. 2.53 – 412nd deg. 2.22 –

3rd deg. 1.71 –

NFV 1.47 –

No. of theoreticalplates (N)

VRC 3403 2543 Z2000

Table 4Statistical comparison between the results obtained by the two proposed methodsfor the analysis of VRC and the manufacturer method.

Statistical terms LC–DAD LC–FLD Manufacturer methoda

Mean 99.97 99.32 99.27S.D. 1.21 0.92 1.45S.E. 0.49 0.38 0.65Variance 1.46 0.85 2.10N 6 6 5t-test 0.79 (2.262)b 0.06 (2.262)b

F-ratio 1.44 (5.19)b 2.47 (5.19)b

a HPLC method according to the manufacturer file.b Figures in parenthesis are the theoretical values of t and F at confidence limit

95%.

R.I. El-Bagary et al. / Talanta 146 (2016) 83–92 91

by the manufacturer methods was performed using (t-test) and (F-ratio). No significant difference between them with respect toaccuracy and precision as shown in Table 4.

4. Conclusion

The quality of pharmaceutical products is of vital importancefor patient's safety. The presence of degradation products mayaffect the efficacy and safety of pharmaceuticals. Degradationcould change the chemical, pharmacological and toxicologicalproperties of drugs and have a significant effect on product qualityand safety. Drug stability is regarded as a secure way of ensuringdelivery of therapeutic doses to patients. In this work, simple,sensitive, accurate, precise, reproducible and specific stability-in-dicating LC methods were established for the determination of

VRC in the presence of its degradation products and its relatedimpurity. The behavior of VRC under different stress conditionswas studied. The developed LC–DAD method offers high specificityand good resolution among VRC, its oxidative degradates and itsrelated impurity within suitable analysis time. While, the devel-oped LC–FLD method represents the first fluorimetric method forthe determination of VRC in the presence of either its oxidativedegradates or its related impurity and can thus be used for routineanalysis, quality control and for quality check during stabilitystudies of its pharmaceutical preparations.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.talanta.2015.07.077.

References

[1] ⟨http://www.rxlist.com/chantix-drug.htm⟩.[2] G.M. Keating, K.A. Lyseng- Williamson, Varenicline: a pharmacoeconomic re-

view of its use as an aid to smoking cessation, Pharmacoeconomics 28 (2010)231–254.

[3] H. Rollema, L.K. Chambers, J.W. Coe, J. Glowa, R.S. Hurst, L.A. Lebel, Y. Lu, R.S. Mansbach, R.J. Mather, C.C. Rovetti, Pharmacological profile of the α4β2

nicotinic acetylcholine receptor partial agonist varenicline, an effectivesmoking cessation aid, Neuropharmacology 52 (2007) 985–994.

[4] J.W. Coe, P.R. Brooks, M.G. Vetelino, M.C. Wirtz, E.P. Arnold, J. Huang, S.B. Sands, T.I. Davis, L.A. Lebel, C.B. Fox, Varenicline: an alpha 4 beta 2 nicotinicreceptor partial agonist for smoking cessation, J. Med. Chem. 48 (2005)3474–3477.

[5] G.D. Garrison, S.E. Dugan, Varenicline: a first – line treatment option forsmoking cessation, Clin. Ther. 31 (2009) 463–491.

[6] J.S. Modiya, Development and Validation of New Analytical Methods for Es-timation of Varenicline Tartrate in Bulk and Pharmaceutical Dosage Form,Rajiv Gandhi University of Health Sciences, Karnataka, Bangalore, 2011.

[7] A.A. Kadi, M.S. Mohamed, M.G. Kassem, I.A. Darwish, A validated stability-indicating HPLC method for determination of varenicline in its bulk and ta-blets, J. Chem. Cent. 5 (2011) 30–36.

[8] K.P. Channabasavaraj, J.S. Modiya, H.M. Sharath, Development and validationof RP-HPLC method for estimation of varenicline tartrate in bulk drug andtablet dosage form, Int, J. Pharm. Pharm. Sci. 3 (2011) 59–61.

[9] P. Katakam, R.R. kalakuntla, J.R. Sama, A novel stability indicating RP-HPLCassay method for the determination of varenicline in pharmaceutical for-mulations, Pharmacologia 6 (2012) 169–173.

[10] R.S. Obach, A.E. Reed-Hagen, S.S. Krueger, B.J. Obach, T.N. O’Connell, K.S. Zandi,S. Miller, J.W. Coe, Metabolism and disposition of varenicline, A selective α4β2

acetylcholine receptor partial agonist, in vivo and in vitro, Drug Metab. Dispos.34 (2006) 121–130.

[11] M. Dobrinas, E. Choong, M. Noetzli, J. Cornuz, N. Ansermot, B. Eap, Quantifi-cation of nicotine, cotinine, trans-3′-hydroxycotinine and varenicline in hu-man plasma by a sensitive and specific UPLC–tandem mass-spectrometryprocedure for a clinical study on smoking cessation, J. Chromatogr. B 879(2011) 3574–3582.

[12] M.G. Kassem, M.W. Attwa, A.S. Abdelhameed, G.A.E. Mostafa, An isocratic li-quid chromatography- electrospray ionization tandem mass spectrometricdetermination of varenicline in human plasma and dosage form, Afr. J. Pharm.Pharmaco. 7 (2013) 1343–1350.

[13] A. Al-Haj, M. Alawi, T. Arafat, M.K. Hourani, Method development, validationand bioequivalence of varenicline in human plasma by liquid chromatographytandem mass spectrometry, J. Chromatogr. B 931 (2013) 134–139.

[14] H.M. Faessel, B.J. Smith, M.A. Gibbs, J.S. Gobey, D.J. Clark, A.H. Burstein, Single-dose pharmacokinetics of varenicline, a selective nicotinic receptor partialagonist, in healthy smokers and non-smokers, J. Clin. Pharmacol. 46 (2006)991–998.

[15] H.M. Faessel, M.A. Gibbs, D.J. Clark, K. Rohrbacher, M. Stolar, A.H. Burstein,Multiple dose pharmacokinetics of the selective nicotinic receptor partialagonist, varenicline, in healthy smokers, J. Clin. Pharmacol. 46 (2006)1439–1448.

[16] S.S. Pujeri, A.M.A. Khader, J. Seetharamappa, Stress degradation studies onvarenicline tartrate and development of a validated stability-indicating HPLCmethod, Sci. Pharm. 80 (2012) 115–126.

[17] B. Satheesh, S. Kumar pulluru, V. Raghavan, D. Saravanan, UHPLC separationand quantification of related substances of varenicline tartrate tablet, J. ActaChromatogr. 22 (2010) 207–218.

[18] B. Satheesh, K.K. Sree Ganesh, D. Saravanan, Identification, isolation andcharacterization of an unknown impurity of varenicline, Sci. Pharm. 80 (2012)

R.I. El-Bagary et al. / Talanta 146 (2016) 83–9292

329–336.[19] Varenicline Standards and Impurity Controls, 2007. Available at: ⟨www.free

patentsonline.com/EP2004186.html⟩ (accessed October 2007).[20] N-formyl and N-methyl Degradation Products, 2004. Available at: ⟨www.free

patentsonline.com/y2004/0235850.html⟩ (accessed November 2004).[21] Methods of Reducing Degradant Formation in Pharmaceutical Compositions of

Varenicline, 2008. Available at: ⟨www.freepatentsonline.com/y2008/0026059.html⟩ (accessed January 2008).

[22] H.A. Alhazmi, H.A. Makeen, S.E. Deeb, Capillary electrophoresis of varenicline,Dig. J. Nanomater Bios. 8 (2013) 295–301.

[23] M.M. Aleksic, V. Radulovic, N. Lijeskic, V. Kapetanovic, Electrochemical re-sponse and determination of varenicline at boron doped diamond, glassycarbon and hanging mercury electrodes, Curr. Anal. Chem. 8 (2012) 133–142.

[24] V. Radulović, M.M. Aleksić, V. Kapetanović, An electrochemical study of theadsorptive behaviour of varenicline and its interaction with DNA, J. Serbian.Chem. Soc. 77 (2012) 1409–1422..

[25] HPLC Manufacturer method (Pfizer), 2009.[26] M. Pesez, J. Bartos, Colorimetric and Fluorimetric Analysis of Organic Com-

pounds and Drugs, Marcel Dekker, New York, NY, USA (1974), p. 628–630.[27] M.C. Colomer, Varenicline: First-non nicotinic agent smoking cessation, Of-

farm 26 (2007) 112–114.[28] A.A. El-Emam, S.H. Hansen, M.A. Moustafa, S.M. El-Ashry, D.T. El-Sherbiny,

Determination of lisinopril in dosage forms and spiked human plasmathrough derivatization with 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole (NBD-Cl) followed by spectrophotometry or HPLC with fluorimetric detection, J.Pharm. Biomed. Anal. 34 (2004) 35–44.

[29] N. El- Enany, Spectrofluorimetric and spectrophotometric determination of

gliclazide in pharmaceuticals by derivatization with 4-Chloro-7-nitrobenzo-2-oxa-1, 3-diazole, J. AOAC Int. 86 (2003) 209–214.

[30] N. El-Enany, F. Belal, M. Rizk, Kinetic spectrophotometric determination ofisoxsuprine in dosage forms through derivatisation with 4-Chloro-7-ni-trobenzo-2-oxa-1,3-diazole (NBD-Cl), Sci. Pharm. 74 (2006) 99–119.

[31] N. El-Enany, D. El-Sherbiny, F. Belal, spectrophotometric, spectrofluorimetricand HPLC determination of desloratadine in dosage forms and human plasma,Chem. Pharm. Bull. 55 (2007) 1662–1670.

[32] S. Tatar Ulu, Determination of tianeptine in human plasma using high-per-formance liquid chromatography with fluorescence detection, J. Chromatogr. B834 (2006) 62–67.

[33] S. Erturk, S.M. Cetin, S. Atmaca, L. Ersoy, G. Baktr, A sensitive HPLC method forthe determination of fluoxetine and norfluoxetine in human plasma withfluorescence detection, Ther. Drug Monit. 27 (2005) 38–43.

[34] H. Miyano, T. Toyo’oka, K. Imai, Further studies on the reaction of amines andproteins with 4-fluoro-7-nitrobenzo-2-oxa-1, 3-diazole, Anal. Chim. Acta. 170(1985) 81–87.

[35] H.M. Saleh, S.M. Ghannam, Colorimetric determination of aromatic aminoacids, Alex. J. Pharm. Sci. 14 (2000) 25–29.

[36] ICH, Q2A Text on Validation of Analytical Procedures, International Conferenceon Harmonization, Oct, 1994.

[37] ICH, Q3B Validation of analytical procedures: methodology, InternationalConference on Harmonization, Nov, 1996.

[38] J.C. Miller, J.N. Miller, Statistics and chemometrics for Analytical Chemistry,fifth edition, Prentice Hall, England (2005), p. 256.

[39] A. Weston, P. Brown, HPLC and CE principles and practice, High-PerformanceLiquid Chromatography, Academic press, San Diego, CA (1997), p. 7–15.