A

ioma(0ri(h©

K

1

drtfcogh

0d

Analytica Chimica Acta 599 (2007) 143–150

Determination of nateglinide in human plasma by high-performanceliquid chromatography with pre-column derivatization using

a coumarin-type fluorescent reagent

Danai Malli a, Evagelos Gikas a,b, Andreas Vavagiannis a, Michael Kazanis a,Konstantinos Daniilides a, Dimitra Gennimata c, Irene Panderi a,∗

a University of Athens, School of Pharmacy, Division of Pharmaceutical Chemistry, Panepistimiopolis, Zografou, GR-157 71 Athens, Greeceb GAIA Research Center, Bioanalytical Department, The Goulandris Natural History Museum, 13 Levidou Street, Kifissia GR-145 62, Greece

c Sotiria General Hospital, 152 Mesogion Street, GR-11527 Athens, Greece

Received 28 June 2007; received in revised form 31 July 2007; accepted 1 August 2007Available online 6 August 2007

bstract

A sensitive and selective high-performance liquid chromatographic method has been developed and validated for the determination of nategliniden human plasma. Nateglinide and the internal standard, undecylenic acid, were extracted from plasma by liquid–liquid extraction using a mixturef ethyl acetate–diethyl ether, 50:50 (v/v). Pre-column derivatization reaction was performed using a coumarin-type fluorescent reagent, N-(7-ethoxy-4-methyl-2-oxo-2H-6-chromenyl)-2-bromoacetamide. The derivatization proceeded in acetone in the presence of potassium carbonate

nd catalyzed by 18-crown-6 ether. The fluorescent derivatives were separated under isocratic conditions on a Hypersil BDS-C8 analytical column250.0 mm × 2.1 mm i.d., particle size 5 �m) with a mobile phase that consisted of 65% acetonitrile in water and pumped at a flow rate of.50 mL min−1. The excitation and emission wavelengths were set at 345 and 435 nm, respectively. The assay was linear over a concentration

−1 −1 −1

D38

ange of 0.05–16.00 �g mL for nateglinide with a limit of quantitation of 0.05 �g mL . Quality control samples (0.05, 4.50 and 16.00 �g mL )n five replicates from five different runs of analysis demonstrated intra-assay precision (%coefficient of variation <6.8%), inter-assay precision%coefficient of variation <1.6%) and an overall accuracy (%relative error) less than −3.4%. The method can be used to quantify nateglinide inuman plasma covering a variety of pharmacokinetic or bioequivalence studies.

2007 Elsevier B.V. All rights reserved.

deriv

bladodlC

eywords: Nateglinide; Undecylenic acid; Coumarin; Pre-column fluorescence

. Introduction

Nateglinide, N-(trans-4-isopropylcyclohexyl-carbonyl)--phenylalanine, is an oral insulinotropic agent capable ofestoring the physiological insulin secretion pattern lost inype-2 diabetes mellitus [1]. It increases the insulin releaserom pancreatic ß-cells through inhibition of potassium-ATPhannels [2,3]. The drug has a rapid onset and short durationf insulinotropic action that results in reduction of mealtime

lucose rise and lowers the post-absorptive potential forypoglycaemia.

∗ Corresponding author. Tel.: +30 210 7274820; fax: +30 210 7274747.E-mail address: ipanderi@pharm.uoa.gr (I. Panderi).

atpa

fi

003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.aca.2007.08.004

atization; High-performance liquid chromatography; Human plasma

Only the d-enantiomer of nateglinide has been approved toe used in clinical treatment as it is much more potent than the-enantiomer. After oral administration, nateglinide is rapidlybsorbed with an average elimination half-life of 1.5 h. Therug demonstrates linear pharmacokinetics over a dose rangef 60–240 mg, and the time to peak concentration is not dose-ependent [4]. It undergoes extensive biotransformation in theiver primarily via cytochrome P450 isoenzymes CYP2C9 andYP3A4. The major metabolites of nateglinide are less potentnti-diabetic agents than nateglinide [5]. Single oral administra-ion of 120 mg of the drug, 10 min before a meal, leads to peaklasma concentration levels, Cmax, of 10.08 �g mL−1, occurring

pproximately 0.5–1.0 h after dosing [6].

Up to now, only a few methods have been developed thatocus on the identification and determination of nategliniden biological matrices. In particular, a two column-switching

1 himic

HnCoagtiettTbbnrtmmiAadfl

trlamatcdnti

tobin2flwToir

2

2

P

GdwPlCSAb(

2

(cCodPTl

fc(f

2p

p5Pw0touiee4t

2

1ods

44 D. Malli et al. / Analytica C

PLC-UV method has been developed for the determination ofateglinide and its metabolites in human plasma and urine [7].linical pharmacokinetic studies on nateglinide based on databtained by HPLC-UV and isotope radioactivity methods havelso been published [8,9]. A reversed-phase liquid chromato-raphic method with UV detection has also been reported forhe determination of nateglinide in human plasma [10]. Nateglin-de along with other anti-diabetic drugs has been identified inquine plasma and urine by liquid chromatography coupledo mass spectrometric detection using an ion trap mass spec-rometer equipped with electrospray ionization interface [11].he chiral separation of nateglinide and its l-enantiomer haseen accomplished on a novel molecular imprinted polymer-ased stationary phase [12]. Recently, the bioavailability ofateglinide-hydroxypropyl-�-cyclodextrin complex capsule inabbits has been studied by liquid chromatography coupled toandem mass spectrometric detection [13]. Several chiral nor-

al [14] and reversed-phase [15,16] liquid chromatographicethods have been developed for the determination of nateglin-

de in pharmaceutical preparations and bulk drug substance.lthough the above-described methods include some successful

pproaches, no report has been mentioned in the literature for theetermination of nateglinide in human plasma by a pre-columnuorescence derivatization procedure.

An ever-increasing need exists within bioanalytical labora-ories to improve the sensitivity and selective of assays. Fluo-escence is a low-background technique that shifts the detectionimits of analytical determinations from 100 to 1000 times mores compared to UV detection, exhibiting a dynamic range ofore than three orders of magnitude. As only a small percent-

ge of the known organic substances exhibit native fluorescence,he fluorescence derivatization technique (both pre- and post-olumn) is widely employed in analytical laboratories for theetermination of a great variety of compounds exhibiting low oro fluorescence [17–20]. Fluorescent probes are used either inhe biological field as markers and staining compounds [21], orn chemical analysis as fluorescence inducing reagents [22,23].

In the present work, a micro-bore HPLC method forhe determination of nateglinide in human plasma basedn a pre-column fluorescence derivatization reaction haseen developed and fully validated. To achieve selectiv-ty and high detection sensitivity for the determination ofateglinide, N-(7-methoxy-4-methyl-2-oxo-2H-6-chromenyl)--bromoacetamide (Br-MAMC) was used as the pre-columnuorescent probe [24]. Br-MAMC forms fluorescent derivativesith compounds containing a free carboxylic acid group [25].he proposed method, compared to the already existing meth-ds, provides a useful insight into the quantitation of nategliniden human plasma as it is selective, offers higher sensitivity andequires only 25 �L of human plasma to perform the analysis.

. Experimental

.1. Equipment

The chromatographic equipment used consisted of a Spectra-hysics SP8810 isocratic pump, Spectra-Physics (Darmstadt,

asie

a Acta 599 (2007) 143–150

ermany) and a Rheodyne Model 7725i injector, Rheo-yne (California, CA, USA) with a 20-�L loop. Detectionas performed by a Perkin-Elmer LS-30 spectrofluorometer,erkin-Elmer (Buckinghamshire, UK) using an emission wave-

ength of 435 nm with an excitation wavelength of 345 nm.hromatographic data were processed on a Spectra-PhysicsP4270 integrator, Spectra-Physics (San Jose, CA, USA).

heating block module, Dri-block DB-3, Techne (Cam-ridge, UK) equipped with thermostatically controlled heating25–100 ± 2 ◦C) was used to perform the derivatization reaction.

.2. Materials and reagents

Solvents were of HPLC grade and purchased from MerckDarmstadt, Germany). 1,4,7,10,13,16-Hexaoxacyclooctade-ane,18-crown-6 ether was purchased from Sigma–Aldrichhemie (Steinheim, Germany). Potassium carbonate wasbtained from Merck (Darmstadt, Germany). Water waseionised and further purified by means of a Milli-Q Plus Waterurification System, Millipore, Waters (Milford, MA, USA).he fluorescent reagent, Br-MAMC, was synthesized in our

aboratory according to an already published procedure [24].Nateglinide of pharmaceutical purity grade was obtained

rom Lakshita Pharmachem Ltd. (Mumbai, India), while unde-ylenic acid (internal standard, IS) was purchased from FlukaBuchs, Switzerland). All substances were used without anyurther purification.

.3. Chromatographic conditions and measurementrocedure

Chromatographic separation was achieved using a reversed-hase Hypersil BDS C8 column (250.0 mm × 2.1 mm i.d.,�m particle size), Thermo-Hypersil-Keystone (Bellefonte,A, USA). The mobile phase consisted of acetonitrile mixedith water in a ratio (65:35, v/v). It was filtered through a.45 �m nylon-membrane filter, GelmanSciences (Northamp-on, UK) and degassed under vacuum prior to use. A flow ratef 0.50 mL min−1 with a column inlet pressure of 1350 psi wassed in order to separate the esters of Br-MAMC with nateglin-de and undecylenic acid (IS), respectively. All chromatographicxperiments were conducted at ambient temperature. Initially,xcitation and emission wavelengths were adjusted at 345 and35 nm, respectively. Peak areas were measured for the quanti-ation of the analyte.

.4. Stock and working standard solutions

Stock standard solutions of nateglinide and undecylenic acid,.0 mg mL−1 were prepared by dissolving appropriate amountsf the compounds in acetone. These solutions when stored in theark at −20 ◦C were found to be stable for at least 4 weeks. Stocktandard solutions of the reagent, Br-MAMC, 100.0 �g mL−1,

nd of the 18-crown-6 ether, 1.0 mg mL−1, were prepared by dis-olving appropriate amounts of the corresponding compoundsn acetone. These solutions were stored in the dark under refrig-ration and were found to be stable for a period of 4 weeks.

himic

ps0tip

2c

dafs4

p1is

2

psuuthaTad1utMst14mps

2

cnfptt

liafii

w1noatc

wp1awisumcoscf

3

3

aoaoicv

3aapCtrafet

D. Malli et al. / Analytica C

A series of working standard solutions of nateglinide wererepared by subsequent dilution of the above-mentioned stocktandard solution in acetone to reach concentration ranges of.05–16.00 �g mL−1. A 40.00 �g mL−1 working standard solu-ion of the internal standard, undecylenic acid, was also preparedn acetone. The working standard solutions were freshly pre-ared every week and stored in the dark at −20 ◦C.

.5. Calibration spiked plasma standards and qualityontrol samples

Appropriate aliquots of the above-mentioned working stan-ard solutions of nateglinide were added into eppendorf tubesnd evaporated to dryness prior to the addition of 25 �L of drug-ree human plasma so as to prepare calibration spiked plasmatandards over the concentration range of 0.05, 0.50, 1.50, 3.00,.50, 8.00, 12.00 and 16.00 �g mL−1 for nateglinide.

Quality control (QC) samples were also prepared in humanlasma at four concentration levels (0.05, 0.50, 4.50 and6.00 �g mL−1). Separate stock standard solutions of nateglin-de were used for the preparation of calibration standardolutions and quality control samples.

.6. Sample preparation and derivatization procedure

Extraction, cleanup and derivatization procedures for humanlasma samples were carried out according to the followingteps. Calibration spiked plasma standards, QC samples andnknown samples were spiked with 10 �L of the 40.00 �g mL−1

ndecylenic acid working standard solution prior to the extrac-ion. The samples were acidified by the addition of 40 �L ofydrochloric acid 1.0 M, vortex-mixed for 30 s and 1.0 mL ofmixture of diethyl ether/ethyl acetate (50:50 v/v) was added.he mixture was shaken at 150 × g for 2 min and centrifugedt 3500 × g for 5 min. The aqueous phase was frozen usingry ice for 2 min and the organic layer was transferred into amL eppendorf tube. Contents were evaporated within 3 minnder a gentle stream of nitrogen at 45 ◦C. The residue wasreated with 160 �L of acetone, 80 �L of a 100.0 �g mL−1 Br-

AMC solution, 40 �L of a 1.0 mg mL−1 18-crown-6 etherolution and 2 mg of a fine suspension of K2CO3. The con-ent of the tubes was vigorously mixed on a vortex mixer formin. Then the reaction mixture was allowed to stand for0 min at 45 ± 2 ◦C. Subsequently, 100 �L of each reactionixture were drawn and diluted with 100 �L of the mobile

hase. Aliquots of 20 �L were injected into the chromatographicystem.

.7. Validation procedures

Spiked plasma calibration standards at eight different con-entration levels ranging from 0.05 to 16.00 �g mL−1 forateglinide were prepared and analysed in triplicate in five dif-

erent analytical runs. Calibration curves, based on the ratio ofeak area signal of the ester derivative of the analyte to that ofhe ester derivative of the IS versus the theoretical concentra-ion were prepared for each run. Weighted (1/y2) least-squares

wrFe

a Acta 599 (2007) 143–150 145

inear regressions were used to obtain the equation of the cal-bration curves. QC samples were processed in five replicatest each concentration (0.05, 0.50, 4.50 and 16.00 �g mL−1) forve different analytical runs in order to evaluate the intra- and

nter-assay accuracy and precision.The recovery of the liquid–liquid extraction procedure

as evaluated at three concentration levels (0.50, 4.50 and6.00 �g mL−1) of the analyte and at 8.00 �g mL−1 for the inter-al standard. It was determined by comparing the peak areasbtained from the quality control samples after the extractionnd derivatization procedure, to the peak areas obtained fromhe analysis of standard solutions (in acetone) of equivalentoncentrations after derivatization procedure.

The stability of nateglinide in spiked human plasma samplesas investigated under various storage conditions. In particular,lasma samples spiked at two different concentrations (4.50 and6.00 �g mL−1) of the analyte were stored at ambient temper-ture for 5 h and at −20 ◦C for 2 weeks. Freeze–thaw stabilityas also evaluated by successive cycles of freezing and thaw-

ng; two complete freeze–thaw cycles were performed withamples frozen at −20 ◦C for 7 days per cycle and thawednassisted at room temperature. Absolute peak area measure-ents obtained from the analysis of the stored samples were

ompared to the absolute peak area measurements that werebtained from the analysis of freshly prepared spiked plasmaamples. The analyte was considered stable in the biologi-al matrix when 80–120% of the initial concentration wasound.

. Results and discussion

.1. Optimization of the derivatization procedure

The applicability of Br-MAMC as a fluorescent probe for thenalysis of nateglinide was thoroughly investigated. To achieveptimum conditions, the effects of the reaction time, temper-ture, concentration of Br-MAMC, solvent and of the amountf the derivatization reaction’s catalyst, 18-crown-6 ether, werenvestigated. In every step of the optimization procedure all theontributing factors but one remain constant and the optimizedalue is used for the next experiment.

Fluorescence intensity was assessed initially for 40 min at 30,5, 40, 45, 50, 55 and 60 ◦C. The maximum reaction yield waschieved at 45 ◦C. Lower temperatures were inadequate whereast temperatures higher than 55 ◦C the signal degraded slowly,erhaps through thermal decomposition of the product (Fig. 1A).onsequently the time profile of the reaction yield was moni-

ored for 80 min keeping the optimized temperature value of theeaction at 45 ◦C. The derivatization of the analyte is completednd reaches the maximum reaction yield within 40 min whileurther heating did not increase the reaction yield (Fig. 1B). Theffect of the stoichiometric ratio of the reagent to the analyte onhe signal of the ester derivative of nateglinide with Br-MAMC

as also examined and it was found that the reaction yield

eached a plateau for ratios greater than 1:10 (M/M) (Fig. 1C).inally, the contribution of the reaction’s catalyst, 18-crown-6ther, on the derivatization procedure was evaluated. The role of

146 D. Malli et al. / Analytica Chimica Acta 599 (2007) 143–150

Fig. 1. Optimization of the derivatization reaction: (A) plots of the signal ofthe ester of nateglinide with Br-MAMC vs. temperature; (B) time profile of flu-orescence derivatization reaction of nateglinide with Br-MAMC at 45 ◦C; (C)plots of the signal of the ester of nateglinide with Br-MAMC vs. stoichiomet-r4a

1ttm4cTrt

3

(crt

F4u

Mmarvtsatifcafi1

aiFoa

3

ypstcftwcb

ic ratio of Br-MAMC/nateglinide; reaction time 40 min; reaction temperature5 ◦C; (D) plots of the signal of the ester of nateglinide with Br-MAMC vs. themount in �g of the catalyst (18-crown-6 ether).

8-crown-6 ether is to solvate K+ in its polar cavity formed byhe six oxygen atoms, leaving the anion of the acid countered byhe opposite charge and thus more nucleophilic. The ester for-

ation reached the maximum reaction yield in the presence of0 �g of catalyst whereas further increases in the amount of theatalyst did not cause any increase in the reaction yield (Fig. 1D).he optimum conditions were the following: temperature 45 ◦C,

eaction time: 40 min, ratio of nateglinide/Br-MAMC greaterhan 1:20 (M/M), amount of 18-crown-6 ether 40 �g (Fig. 2).

.2. High-performance liquid chromatography

Chromatography was performed using a Hypersil BDS C8

250.0 mm × 2.1 mm i.d., 5 �m particle size) column whilehromatographic conditions were optimized to achieve adequateetention of the analytes. Initially, the effect of composition ofhe mobile phase on the resolution of the esters was investigated.

wiar

ig. 2. Schematic diagram of the derivatization reaction of N-(7-methoxy--methyl-2-oxo-2H-6-chromenyl)-2-bromoacetamide with nateglinide andndecylenic acid (IS).

ethanol and acetonitrile were tried as organic modifiers in theobile phase in combination with water. In the present study,

cetonitrile was preferred to methanol as it gave better peakesolution. Various combinations of acetonitrile and water witharying content of each component were investigated to iden-ify the optimal mobile phase composition that produced the bestensitivity and peak shape for all the analytes. It was found thatn increase in the content of acetonitrile as organic modifier inhe mobile phase could improve peak shape, whereas an increasen water content broadened the peak. Addition of ammoniumormate to the aqueous component of the mobile phase did nothange the chromatographic behaviour of the analytes. Thus,mobile phase consisting of 65% acetonitrile in water was

nally used. Each chromatographic run was completed within0.0 min.

A representative chromatogram obtained from the analysis ofplasma sample spiked with 8.0 �g mL−1 of the analyte and the

nternal standard and derivatized with Br-MAMC is presented inig. 3. Under the current chromatographic conditions the estersf nateglinide and undecylenic acid with Br-MAMC were elutedt 4.66 and 6.74 min, respectively.

.3. Statistical analysis of data

Spiked plasma calibration standards of nateglinide were anal-sed in triplicate in five analytical runs for the calibrationrocedure. Linear relationships of the ratios of the peak areaignals of the ester of nateglinide with Br-MAMC to that ofhe ester of the IS with Br-MAMC versus the correspondingoncentrations of nateglinide were observed, using a weightingactor of 1/y2. The regression equations of the calibration curves,he standard deviation values of the slopes and intercepts alongith the correlation coefficients are presented in Table 2; the

orrelation coefficient invariably exceeded 0.992. In all cases,ack-calculated concentrations in the calibration curves were

ithin 15% of the nominal values, which is in agreement with

nternational guidelines [26] and indicate that the linear modelcceptably describes the relationship between concentration andesponse.

D. Malli et al. / Analytica Chimic

Fig. 3. A representative chromatogram of a calibration plasma sample spikedwith 8.0 ng mL−1 of nateglinide and 8.0 �g mL−1 of the IS; the retention times ofthe esters of nateglinide and undecylenic acid (IS) with Br-MAMC are 4.66 and6Hr

eevtSupr

tf

aLydpei0rbtlr

tibtp0

tas0tt%n

TA

C

R0

R0

R0

R0

R0

C

.74 min, respectively. Chromatographic conditions: reversed-phase HPLC on aypersil BDS C-8 column; mobile phase acetonitrile and water (65:35, v/v); flow

ate 0.5 mL min−1; fluorescence detection with λex = 345 nm and λem = 435 nm.

A Student’s t-test was performed to determine whether thexperimental intercepts (α) of the above-mentioned regressionquations were significantly different from the theoretical zeroalue. The test is based on the calculation of the quantities= α/S�, where � is the intercept of the regression equations and

� is the standard deviation of α, and their comparison with tab-lated data of the t-distribution. The calculated t-values are alsoresented in Table 1; these values do not exceed the 95% crite-ion of tp = 2.45 for f = 6 degrees of freedom, which denotes that

tce

able 1nalytical parameters of the calibration equations for the determination of nateglinid

onentration range (�g mL−1) Regression equationsa r

un1.05–16.00 RNtg = 0.2404 × CNtg − 0.0020 0

un2.05–16.00 RNtg = 0.2369 × CNtg + 0.0023 0

un3.05–16.00 RNtg = 0.2366 × CNtg + 0.0021 0

un4.05–16.00 RNtg = 0.2382 × CNtg + 0.0026 0

un5.05–16.00 RNtg = 0.2311 × CNtg + 0.0026 0

a Ratio of the peak area amplitude of the analyte ester with Br-MAMC to that of the in

Ntg.b Correlation coefficient.c Standard deviation of slope and intercept.d Standard error of the estimate.e Theoretical value of t at P = 0.05 level of significance, for f = n − 2 = 6 df, 2.45.

a Acta 599 (2007) 143–150 147

he intercept of all regression lines are not significantly differentrom zero.

The LOD was calculated using the equations y − α = 3.3 × S�

nd y − α = b × LOD, while the limit of quantitation,OQ, was attained using the equations y − � = 10 × S� and− α = b × LOQ (where b is the slope and S� is the stan-ard deviation of the intercept of the regression line). Inarticular, LOD and LOQ were calculated taking under consid-ration data obtained from the calibration equations presentedn Table 1. Average values of LOD and LOQ were found to be.0144 ± 0.0021 and 0.0438 ± 0.0063 �g mL−1, respectively. Aepresentative chromatogram obtained from the analysis of alank plasma sample along with a chromatogram obtained fromhe analysis of a sample spiked with 0.05 �g mL−1 of the ana-yte (LOQ level) and the IS are presented in Fig. 4A and B,espectively.

One-way analysis of variance (ANOVA) was used to evaluatehe intra- and inter-assay precision. Results presented in Table 2ndicate that intra-assay coefficients of variation, %CV, wereetween 3.1 and 6.8%, while the inter-assay %CVs were lowerhan 1.6%. The overall accuracy was assessed by the relativeercentage error, absolute %Er, which was ranged from −3.4 to.4% for the analyte.

The selectivity towards endogenous plasma compounds wasested in five different lots of drug-free human plasma bynalysing blanks (non-spiked plasma samples) and plasmaamples spiked with nateglinide at a concentration level of.05 �g mL−1. The concentration of nateglinide obtained afterhe analysis of the five different lots of human plasma was foundo be 0.0477 ± 0.024 �g mL−1 with a relative percentage error,

Er, of −4.5%. The results indicate that the matrix effect doesot appreciably affect the assay.

A thorough investigation was conducted in order to choosehe optimum solvent mixture for the liquid-liquid extraction pro-edure. A number of organic solvents such as hexane, diethylther and ethyl acetate were tested for the extraction procedure

e by HPLC with fluorescence detection

b S.D.c Srd a/Sa

e

Slope Intercept

.998 0.0047 0.0009 0.051 2.25

.9991 0.0041 0.0011 0.044 2.09

.9991 0.0044 0.0009 0.047 2.36

.998 0.0059 0.0011 0.063 2.39

.9990 0.0047 0.0012 0.052 2.17

ternal standard ester with Br-MAMC, RNtg, vs. the corresponding concentration,

148 D. Malli et al. / Analytica Chimica Acta 599 (2007) 143–150

F a repra hic coa ion w

aaet

TA

RM(

RM(

RM(

RM(

RM(

OOII

ig. 4. A representative chromatogram of a blank plasma extract (A), along witht the LOQ level (0.05 �g mL−1) and 8.0 �g mL−1 of the IS (B). Chromatograpcetonitrile and water (65:35, v/v); flow rate 0.5 mL min−1; fluorescence detect

nd lead to poor recoveries (less than 45.0%) for both of thenalyte and IS. It was found that a mixture of diethyl ether andthyl acetate led to almost complete extraction recoveries ofhe compounds. A mixture of diethyl ether and ethyl acetate

(tli

able 2ccuracy and precision evaluation of quality control samples for nateglinide (in five

Concentration (�g mL−1)

0.05 0.50

un 1ean ± S.D. 0.0473 ± 0.0021 0.4

%Er)a −2.7 −0.7

un 2ean ± S.D. 0.0479 ± 0.0027 0.4

%Er)a −2.7 −5.4

un 3ean ± S.D. 0.0492 ± 0.0045 0.4

%Er)a −2.5 −0.2

un 4ean ± S.D. 0.0482 ± 0.0038 0.4

%Er)a −4.9 −1.3

un 5ean ± S.D. 0.0488 ± 0.0026 0.4

%Er)a −2.4 −1.4

verall mean 0.0483 0.4verall accuracy (%Er)a −3.4 −1.8

ntra-assay %CVb 6.8 2.9nter-assay %CVb N.V.c 1.6

a %Er: relative percentage error.b %CV: coefficient of variation; Intra- and inter-assay CV was calculated by ANOVc No significant variation was observed.

esentative chromatogram of a calibration plasma sample spiked with nateglinidenditions: reversed-phase HPLC on a Hypersil BDS C-8 column; mobile phase

ith λex = 345 nm and λem = 435 nm.

50:50, v/v) was finally chosen as the optimum solvent mix-ure for the extraction procedure. The recovery of the proposediquid–liquid extraction procedure was determined by calculat-ng the ratio of the absolute peak areas of extracted spiked plasma

validation days, five replicates per day)

4.50 16.00

97 ± 0.010 4.50 ± 0.23 16.25 ± 0.500.1 1.6

73 ± 0.013 4.405 ± 0.081 16.08 ± 0.63−2.1 0.5

99 ± 0.012 4.44 ± 0.10 15.81 ± 0.72−1.4 −1.2

91 ± 0.011 4.55 ± 0.16 16.09 ± 0.791.1 0.6

93 ± 0.019 4.38 ± 0.14 16.11 ± 0.461.1 0.7

91 4.45 16.07−1.0 0.4

3.4 3.90.5 N.V.c

A.

D. Malli et al. / Analytica Chimica Acta 599 (2007) 143–150 149

Table 3Recovery data for the liquid–liquid extraction procedure for nateglinide and undecylenic acid, IS

Compound

Nateglinide Undecylenic acid

Concentration (�g mL−1) 0.05 8.00 16.00 8.00Mean ± S.D.(n=3) 0.0494 ± 0.0073 8.37 ± 0.35 15.53 ± 0.21 –%Recovery ± S.D. 95.6 ± 1.1 96.9 ± 2.4 96.1 ± 2.0 95.7 ± 3.1

Table 4Stability data for nateglinide in human plasma under various storage conditions (n = 3)

Storage conditions/time Concentration levels (�g mL−1) Calculated concentration (�g mL−1) (mean ± S.D.) %Era

Stability samples Freshly prepared samples

Ambient temperature/5 h 4.50 4.41 ± 0.19 4.515 ± 0.015 −2.316.00 15.77 ± 0.43 15.99 ± 0.14 −0.6

−20 ◦C/14 days 4.50 4.29 ± 0.14 4.65 ± 0.13 −7.716.00 15.61 ± 0.16 15.99 ± 0.41 −2.4

−20 ◦C/1 freeze–thaw cycle 4.50 4.48 ± 0.15 4.56 ± 0.17 −1.816.00 15.99 ± 0.26 16.17 ± 0.13 −1.1

−20 ◦C/2 freeze–thaw cycles 4.50 4.32 ± 0.10 4.56 ± 0.16 −5.31

samplc

scsoae8

nibdp5fshwpridd

3a

pyr1

wPKhm1tP

ohpaaadttpo

4

fph

16.00

a Relative percentage error = [(overall mean assayed concentration of stabilityoncentration of fresh sample) × 100].

amples to the absolute peak areas of aqueous standard solutionsontaining equivalent concentrations of nateglinide (unextractedtandards) that represent 100% recovery. The data, under theptimum extraction conditions, presented in Table 3 indicate anverage recovery of 96.2% for nateglinide, whereas the recov-ry for undecylenic acid (IS) was found to be 95.7 ± 3.1% at.0 �g mL−1.

The stability results presented in Table 4 indicate thatateglinide can be considered stable under the various conditionsnvestigated. In particular, nateglinide concentrations deviatey no more than −7.7% relative to the reference, while noegradation products were observed. Calibration plasma sam-les containing nateglinide may therefore be kept for up toh at ambient temperature, 14 days at −20 ◦C, and after 2

reeze–thaw cycles (7 days per cycle) at −20 ◦C without anyignificant degradation. The stability of the Br-MAMC estersas also been investigated. In particular, a plasma sample spikedith 4.5 �g mL−1 of nateglinide and 8.0 �g mL−1 of the IS wasrocessed according to the sample preparation procedure. Theeaction mixture was kept in an ice-bath and injected period-cal into the HPLC system. After 4 h the signals of the estererivatives of nateglinide and undecylenic acid with Br-MAMCeviated by no more than 4.1 and 3.0%, respectively.

.4. Actual measurement of the analyte in human plasmafter oral administration

The method was applied to the analysis of plasma sam-

les obtained after oral administration of nateglinide. A 57ears old male patient was used in the study. The patient waseceiving 120 mg of nateglinide three times daily (Starlix®

20 mg, Novartis-Pharma) for a long-term treatment along

bawo

4.75 ± 0.28 15.75 ± 0.18 −6.3

e − overall mean assayed concentration of fresh sample)/(overall mean assayed

ith 100 mg of acetylsalicylic acid (Salospir® 100 mg, Uni-harma Hellas) once daily, 20 mg simvastatin (Lepur® 20 mg,leva SA) once daily in the morning, 20 mg omeprazolydrochloride (Penrazol® 20 mg, Elpen SA) once daily in theorning, 150 mg propafenone hydrochloride (Rythmonorm®

50 mg, Vianex SA) three times daily and 16.0 mg candesar-an hydrochloride with 12.5 mg hydrochlorothiazide (Atacandlus®, AstraZeneca) once daily every night.

A blood sample was collected 1 h after dosing with 120 mgf nateglinide and stored in a Venoject® tube containing lithiumeparin as anticoagulant. Immediately after drawing, the sam-le was shaken gently and centrifuged at 4000 rpm for 10 mint 4 ◦C. The plasma sample was analysed 2 days after storaget −20 ◦C. The sample was analysed in triplicate, using plasmaliquots of 25 �L, according to the sample preparation proce-ure (IS was added) in order to calculate the concentration ofhe analyte. The nateglinide plasma concentration was foundo be 7.25 ± 1.27 �g mL−1. The results indicated that the pro-osed method could accurate quantitate nateglinide in presencef other medicines that may be coadministered with the analyte.

. Conclusions

Br-MAMC is a useful fluorogenic labelling reagent suitableor the pre-chromatographic derivatization of nateglinide andermits the determination of nateglinide in a small amount ofuman plasma (25 �L). The method takes full advantage of the

enefits of fluorimetry as it enables the accurate and selectivessay of nateglinide. The LOQ for the analyte was 0.05 �g mL−1

ith a chromatographic run time less than 10.0 min. The methodffers significant improvement in sensitivity and total analysis

1 himic

ttTo

A

p

R

[

[

[

[

[[

[

[

[[[

[

[[

[

50 D. Malli et al. / Analytica C

ime of the assay and proved to be appropriate for the quantifica-ion of nateglinide in the plasma of a patient receiving the drug.hus, it is suitable to support a wide range of pharmacokineticr bioequivalence studies.

cknowledgments

The authors gratefully acknowledge the Sotiria General Hos-ital, Athens, Greece for providing the patient’s plasma samples.

eferences

[1] D.C. Whitelaw, P.M. Clark, J.M. Smith, M. Nattrass, Diabetes Med. 17(2000) 225.

[2] S. Hu, S. Wang, B.E. Dunning, J. Pharmacol Exp. Ther. 291 (1998)1372.

[3] S. Hu, B.R. Boettcher, B.E. Dunning, Diabetologia 46 (S1) (2003) M37.[4] L. Keilson, S. Mather, Y.H. Walter, S. Subramanian, J.F. Mcleod, J. Clin.

Endocrinol. Metab. 85 (2000) 1081.[5] H. Takesada, K. Matsuda, R. Ohtake, R. Mihara, I. Ono, K. Tanaka, Bioorg.

Med. Chem. 4 (1996) 1771.[6] D. Devineni, Y.H. Walter, H.T. Smith, J.S. Lee, J.F. Mcleod, J. Clin. Phar-

macol. 43 (2003) 163.

[7] I. Ono, K. Matsuda, S. Kanno, J. Chromatogr. B 692 (1997) 397.[8] M.L. Weaver, B.A. Orwig, L.C. Rodriguez, E.D. Graham, J.A. Chin, M.J.

Shapiro, J.F. Mcleod, J.B. Mangold, Drug Metabol. Dispos. 29 (2001) 415.[9] H. Sabia, G. Sunkara, M. Ligueros-Saylan, Y. Wang, H. Smith, J. McLeod,

P. Prasad, Eur. J. Clin. Pharmacol. 60 (2004) 407.

[

[

a Acta 599 (2007) 143–150

10] S. Bauer, E. Stormer, J. Kirschheiner, C. Michael, J. Brockmoller, I. Roots,J. Pharm. Biomed. Anal. 31 (2003) 551.

11] E.N.M. Ho, K.C.H. Yiu, T.S.M. Wan, B.D. Stewart, K.L. Watkins, J. Chro-matogr. B 811 (2004) 65.

12] G. Yang, J. Yin, Z. Li, H. Liu, L. Cai, D. Wang, Y. Chen, Chromatographia59 (2004) 705.

13] Z. He, X. Chen, D. Zhong, C. Zhao, X. Liu, R. Zhang, Biomed. Chromatogr.18 (2004) 532.

14] H. Shinkai, M. Nashikawa, Y. Sato, J. Liq. Chromatogr. 12 (1989) 457.15] M. Qi, P. Wang, Y. Sun, Y. Li, J. Liq. Chromatogr. Rel. Technol. 26 (2003)

(1839).16] G. Yang, H. Liu, H. Li, Y. Wang, Z. Li, Y. Chen, Chromatographia 57

(2003) 245.17] S. Mitsuhashi, T. Fukushima, M. Tomiya, T. Santa, K. Imai, T. Toyo’oka,

Anal. Chim. Acta 584 (2007) 315.18] G. Bahrami, B. Mohammadia, J. Chromatogr. B 850 (2007) 417.19] L. Zhang, L. Zhang, W. Zhang, Y. Zhang, Anal. Chim. Acta 543 (2005) 52.20] R. Gatti, M.G. Gioia, A.M. Di Pietra, V. Cavrini, Anal. Chim. Acta 447

(2001) 89.21] T. Baumgart, G. Hunt, E.R. Farkas, W.W. Webb, G.W. Feigenson, Biochem.

Biophys. Acta, 2007, doi:10.1016/j.bbamem.2007.05.012.22] L. Guo, B. Qiu, G. Chen, Anal. Chim. Acta 588 (2007) 123.23] E. Gikas, M. Parissi-Poulou, M. Kazanis, A. Vavagianis, Anal. Chim. Acta

489 (2003) 153.24] E. Gikas, M. Derventi, I. Panderi, A. Vavayannis, M. Kazanis, M. Parissi-

Poulou, J. Liq. Chromatogr. Rel. Technol. 25 (2002) 381.25] S. Rompotis, M. Parissi-Poulou, E. Gikas, M. Kazanis, A. Vavayannis, I.

Panderi, J. Liq. Chromatogr. Rel. Technol. 25 (2002) 2833.26] Guidance for industry. Bioanalytical method validation. FDA May 2001;

www.fda.gov/Cder/guidance/4252fnl.htm.