Mps

RCa

b

c

d

e

a

ARRAA

KTFPUU

HmtQtmSro

h0

Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 140–149

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journa l homepage: www.e lsev ier .com/ locate / jpba

easurement of total and free docetaxel concentration in humanlasma by ultra-performance liquid chromatography–tandem masspectrometry

aül Rigo-Bonnin a,∗, Sara Cobo-Sacristán b, Núria Gonzalo-Diego c, Helena Colom d,armen Munoz-Sánchez c, Ander Urruticoechea e, Catalina Falo e, Pedro Alía a

Laboratori Clínic Department, IDIBELL, Hospital Universitari de Bellvitge, L’Hospitalet de Llobregat, Barcelona, SpainPharmacy Department, IDIBELL, Hospital Universitari de Bellvitge, L’Hospitalet de Llobregat, Barcelona, SpainPharmacy Department, IDIBELL, Catalan Institute of Oncology, Hospital Duran i Reynals, L’Hospitalet de Llobregat, Barcelona, SpainDepartment of Biopharmaceutics and Pharmacokinetics, School of Pharmacy, University of Barcelona, L’Hospitalet de Llobregat, Barcelona, SpainCatalan Institute of Oncology, IDIBELL, Hospital Duran i Reynals, L’Hospitalet de Llobregat, Barcelona, Spain

r t i c l e i n f o

rticle history:eceived 14 April 2015eceived in revised form 18 August 2015ccepted 19 August 2015vailable online 28 August 2015

eywords:otal docetaxelree docetaxelaclitaxelltrafiltrationPLC–MS/MS

a b s t r a c t

Docetaxel is a semi-synthetic taxane with cytotoxic anti-neoplastic activity and, currently used asanticancer agent in several types of cancer. Docetaxel is highly bound to plasma proteins, and this sig-nificantly determines its clearance and activity. Therefore, measurement of free docetaxel in plasmais pharmacologically important when pharmacokinetics is investigated. We developed and validatedchromatographic methods by ultra-performance liquid chromatography-tandem mass spectrometry tomeasure total and free docetaxel concentration in human plasma. The final validated methods involvedliquid–liquid extraction followed by dryness under nitrogen evaporation. To measure free docetaxelconcentration, sample preparation was preceded by ultrafiltration. Chromatographic separation wasachieved using an Acquity® UPLC® BEHTM (2.1 × 100 mm id, 1.7 �m) reverse-phase C18 column at aflow rate of 0.4 mL/min, using isocratic elution mode containing ammonium acetate/formic acid inwater/methanol (30:70 v/v) as mobile phase. Docetaxel and its internal standard (paclitaxel) weredetected by electrospray ionization mass spectrometry in positive ion multiple reaction monitoringmode using mass-to-charge (m/z) transitions of 808.3 → 527.0 (quantifier) and 808.3 → 509.0 (qualifier);and 854.3 → 569.0 (quantifier) and 854,3 → 509,0 (qualifier), respectively. The run time per sample was

3.5 min. The limits of quantification were 1,95 and 0.42 �g/L and linearity was observed between 1.95 and1000 and 0.42–100 �g/L for total and free docetaxel, respectively. Coefficients of variation and absoluterelative biases were less than 13.8% and 10.0%. Recovery values were greater than 79.4%. Evaluation ofthe matrix effect showed ion suppression and no carry-over was observed. The validated methods couldbe useful for both therapeutic drug monitoring and pharmacokinetic studies. They could be applied todaily clinical laboratory practice to measure the concentration of total and free docetaxel in plasma.

Abbreviations: DCX, docetaxel; PK, phamacokinetic; PD, pharmacodynamic;PLC, high-performance liquid chromatography; UV, ultraviolet; MS/MS, tandemass spectrometry; UPLC, ultra-performance liquid chromatography; MTBE, methyl

-butyl ether; LC, liquid chromatography; MS, mass spectrometer; PCX, paclitaxel;C, quality control; LLOQ, lower limit of quantification; ULOQ, upper limit of quan-

ification; ESI, electrospray ionization; MRM, multiple reaction monitoring; m/z,ass-to-charge; EMA, European Medicines Agency; CLSI, Clinical and Laboratory

tandards Institute; IFCC, International Federation of Clinical Chemistry and Labo-atory Medicine; LLOD, lower limit of detection; S/N, signal-to-noise; CV, coefficientf variation; ır, bias; Cmax, maximum concentration; CL, total clearance; AUC, area

ttp://dx.doi.org/10.1016/j.jpba.2015.08.025731-7085/© 2015 Published by Elsevier B.V.

© 2015 Published by Elsevier B.V.

1. Introduction

Docetaxel (Fig. 1) is a semi-synthetic analog of paclitaxel(Taxol®) with cytotoxic anti-neoplastic activity. Docetaxel (DCX)is a promoter of microtubule polymerization leading to cell cycle

under the plasma concentration–time curve; %D, percent deviation; TDM, therapeu-tic drug monitoring.

∗ Corresponding author at: Laboratori Clínic Department, Hospital Universitaride Bellvitge, Feixa Llarga s/n, 08907 L’Hospitalet de Llobregat, Barcelona, Spain.Fax: +34 932607546

E-mail address: raulr@bellvitgehospital.cat (R. Rigo-Bonnin).

R. Rigo-Bonnin et al. / Journal of Pharmaceutical an

ain

svatbbcbfptbwTD

o(ptilGhdtc≥so

using Centrifree® Ultrafiltration Device (with an Ultracel YM-T

Fig.1. Molecular structure of docetaxel and paclitaxel (internal standard).

rrest at G2/M, apoptosis and cytotoxicity [1]. Docetaxel has signif-cant activity in breast, non-small-cell lung, ovarian and head andeck cancers [1].

The pharmacokinetic (PK) profile of DCX is characterized byubstantial interpatient variability. Causes of these inter-individualariations may include age, gender, ethnicity, hepatic impairmentnd genetic polymorphisms [2,3]. Another factor contributing tohis variation is protein binding. In plasma, DCX is extensivelyound to albumin and �1-acid glycoprotein [1,4]. In cancer patients,inding of DCX to �1-acid glycoprotein might contribute to theorrelation between this glycoprotein and clearance of total DCXy lowering concentrations of unbound DCX (free DCX) availableor metabolism [4]. Moreover, in population pharmacokinetic andharmacodynamic (PD) analysis of DCX using total drug concen-rations, high plasma concentrations of �1-acid glycoprotein haveeen significantly associated with low clearance of total DCX asell as mild neutropenia and diminished antitumor efficacy [1,4].

herefore, it might be important to measure concentrations of freeCX when investigating its PK and PD.

Several high-performance liquid chromatography (HPLC) meth-ds for the measurement of DCX in human plasma using ultravioletUV) detection have been previously described [5–9], whichresent low detection capabilities and are not very selective, owingo the presence of endogenous interferences as well as the lim-ted UV absorption characteristics of the taxane moiety and theow wavelengths required to measure the concentration of DCX.reater detection capabilities and more selective HPLC methodsave been developed by the utilization of HPLC coupled with tan-em mass spectrometry (MS/MS) [10–21]. Nevertheless, some ofhese methods present shortcomings. With regard to detection

apability, two of them presented lower limits of quantification5 �g/L [10–11,14] which could be insufficient to perform DCX PK

tudies. Furthermore, the analytical time of each run in these meth-ds was between 4.5–18 min [10–12,14,16–21]. The relatively long

d Biomedical Analysis 117 (2016) 140–149 141

chromatographic run time could not satisfy the requirement of highthroughput measurement in PK studies and therapeutic drug mon-itoring in the daily practice of a clinical laboratory. Also, in somecases they use protein precipitation extraction [14,15], very proneto dirtiness, which often resulted in a significant matrix effect dueto the presence of many residual matrix components. Others usethe cleaner but time-consuming and expensive solid phase extrac-tion [11,19,21]. In addition, to our knowledge, only one of themhas been used to measure DCX concentrations in plasma usingultra-performance liquid chromatography (UPLC)-MS/MS method,in particular for measuring DCX from a lipid microsphere formu-lation in human plasma [13]. UPLC uses high flow and pressureconditions with sub-2 �m particles columns providing, in gen-eral, better performance than HPLC in reduced times even withoutworking under high flow and pressure conditions, due to a reduceddwell and dead volumes (≤150 �L). These characteristics providemore resolution and shorter retention times [22,23].

Other analytical methods have been developed to measure freeDCX concentration in human plasma. Most of them use eitherultrafiltration or equilibrium dialysis to separate free drug fromthe bound fraction [24–28], two use ultrafiltration combined withHPLC-MS/MS for direct measurement of free DCX concentrations[24,25], but only one of them uses UPLC technology [24].

The aim of this work was to develop and validate anUPLC–MS/MS method for the measurement of total and freeDCX concentrations in human plasma within the same run. Fur-thermore, these validated methods were applied to support apopulation pharmacokinetic model allowing dose modification ofdocetaxel in pacients with breast cancer participating in a phase IVclinical trial.

2. Material and methods

2.1. Chemicals and reagents

Certified reference standards with a 93.5% purity of DCX tri-hydrate and a 99.4% purity of paclitaxel (Fig. 1) were purchasedfrom European Pharmacopeia (European Directorate for the Qualityof Medicines-Council of Europe, Strasburg, France). LC-MS-grademethyl t-butyl ether (MTBE), formic acid and ammonium acetatewere obtained from Sigma–Aldrich (Steinheim, Germany). Glacialacetic acid, LC–MS–grade methanol and water and Centrifree®

Ultrafiltration Device (catalog number 4104) were supplied byMerck Millipore Group (Darmstadt, Germany). Drug-free humanplasma was obtained from patients not treated with DCX or pacli-taxel (PCX) in our hospital.

2.2. Plasma sample collection

Approximately, three milliliters of blood were collected in anEDTA-K3 tube (Vacuette, Kremsmünster, Austria) and centrifugedat 1600 × g for 10 min at (4 ± 1) ◦C. Plasma samples obtained werealiquoted and stored at (–75 ± 5) ◦C until analysis.

2.3. Ultrafiltration sample collection

To measure free DCX concentration in plasma a separation of theprotein-bound fraction of DCX had to be achieved. For this purpose,one milliliter of plasma sample was subjected to ultrafiltration

regenerated cellulose membrane 30,000 NMWL units) by centrifu-gation at 1500 × g for 45 min at room temperature (25 ± 3 ◦C) toprovide about 200 �L of the ultrafiltrate. Ultrafiltrate samples wereprocessed on the same day of the analysis.

1 ical an

2

cspfai2f0pflid

ma

2p

u

osp5rfft1tr0ti

ot

2

Ii(gd

2

mtaf2Uu

42 R. Rigo-Bonnin et al. / Journal of Pharmaceut

.4. Preparation of stock and working solutions

Two stock solutions of DCX were prepared in methanol at a con-entration of 5 mg/L for total DCX and 500 �g/L for free DCX. Stockolutions were used for the preparation of working calibrator sam-les and working quality control (QC) samples. Various quantities

rom a set of the DCX stock solutions were diluted with 0.1% (v/v)cetic acid in methanol according to Richheimer et al. [9], resultingn seven working calibrators samples (0, 10, 25, 100, 250, 1000 and500 �g/L for total DCX and 0.0, 1.0, 2.5, 10, 25, 100 and 250 �g/Lor free DCX). The other set of DCX stock solution was diluted with,1% (v/v) acetic acid in methanol to obtain three working QC sam-les (50, 500 and 1500 �g/L for total DCX and 5.0, 50 and 150 �g/L

or free DCX). The working calibrators and QC samples were storedight-protected for up to 6 months at (−75 ± 5) ◦C as 150 �L aliquotsn 1.5 mL-polypropylene microcentrifuge tubes, and thawed on theay of analysis.

Similarly, an IS (PCX) stock solution (10 mg/L) was prepared inethanol and diluted with 0.1% (v/v) acetic acid in methanol to give

n IS working solution (100 �g/L).All solutions were stored at (−75 ± 3) ◦C until use.

.5. Preparation of calibrators and quality control samples inlasma and ultrafiltrate

Drug-free pooled human plasma and ultrafiltrate plasma weresed as a biological matrix for calibrators and QC samples.

According to Mortier et al. [25], for the preparation of plasmar ultrafiltrate calibrators, 250 �L of water, 50 �L of IS workingolution (PCX at 100 �g/L) and 50 �L of working calibrator sam-les were added to 125 �L of drug-free plasma or ultrafiltrate into

mL-glass tubes, resulting in calibrator nominal concentrationsanging between 4.0–1000 �g/L for total DCX and 0.4–100 �g/Lor free DCX. After vortexing for 5 s, 1.5 mL of MTBE was addedor liquid–liquid extraction. Tubes were vortexed for 45 s and cen-rifuged at 1500 × g for 5 min at room temperature. Subsequently,

mL of upper organic layer was transferred to a new 5 mL-glassube and evaporated to dryness under nitrogen at 40 ◦C. Theesidues were redissolved in 200 �L of methanol/water containing.1% acetic acid; 50/50, v/v. After vortexing for 5 s, all volume wasransferred into HPLC vials and placed in the autosampler ready fornjection.

Similary, plasma or ultrafiltrate QC sample were prepared inrder to obtain nominal concentrations of 20, 200 and 600 �g/L forotal DCX and 2.0, 20 and 60 �g/L for free DCX.

.6. Sample preparation

For the preparation of plasma or ultrafiltrated samples, 50 �L ofS working solution (PCX at 100 �g/L), 50 �L of 0.1% (v/v) acetic acidn methanol and 250 �L of water were added to 125 �L of plasmafor total DCX) or 125 �L of ultrafiltrate (for free DCX), into 5 mL-lass tubes. After that, the same liquid–liquid extraction procedureescribed above for plasma or ultrafiltrate calibrators was applied.

.7. Instrumentation

Analyses were conducted using an Acquity® UPLC® integratedeasurement system (Waters, Milford, MA, USA) consisting of a

hermostatic autosampler, a binary solvent delivery manager and

column over a thermostated compartment. Separation was per-ormed on an Acquity® UPLC® BEHTM C18 reverse-phase column,.1 × 100 mm, packed with 1.7 �m particles (Waters, Milford, MA,SA), protected by a pre-column frit (0.2 �m × 2.1 mm). The col-mn chamber was held at a temperature of 35 ◦C.

d Biomedical Analysis 117 (2016) 140–149

The chromatographic separation was performed using twomobile phases isocratically. Mobile phase A consisted of 2 mMammonium acetate with 0.1% formic acid (v/v) in water (pH 4.2),and was also used as a weak wash solvent, while mobile phase Bconsisted of 2 mM ammonium acetate with 0.1% formic acid (v/v)in methanol, and was also used as a strong wash solvent. The iso-cratic program was 30% A: 70% B at a 0.4 mL/min flow rate. Thesample injection volume was 20 �L in a 50-�L loop (partial loopwith needle overfill injection mode) and the autosampler was heldat a temperature of (4 ± 1) ◦C.

Detection was carried out using an Acquity® TQD® tandem-quadrupole MS equipped with a Z-spray electrospray ionization(ESI) source (Waters, Milford, MA, USA) operating in positive mode.Nitrogen was used as the nebulizing and desolvation gas, and argonwas used as collision gas. DCX and PCX were detected by mul-tiple reaction monitoring (MRM) using the following transitionsof mass-to-charge (m/z): DCX, 808.3 → 527.0 (quantifier ion) and808.3 → 509.0 (qualifier ion); PCX, 854.3 → 569.0 (quantifier ion)and 854.3 → 509.0 (qualifier ion). The quantifier to qualifier ratiowas used for peak identification based on criteria set forth by theCLSI C50-A guideline [29]. The scan dwell time was set at 0.2 s forevery channel. The other optimized MS settings employed for DCXand PCX were capillary potential 3.5 kV, extractor voltage 3 V, RFlens voltage 0.1 V, source temperature 120 ◦C, desolvation temper-ature 400 ◦C, desolvation gas flow rate 850 L/h, cone gas flow rate150 L/h, collision gas flow 0.13 mL/min, cone voltage 18 V for DCXand 24 V for PCX and collision energy 10/14 eV for DCX and PCX asquantifier/qualifier ions, respectively.

2.8. Validation procedure

Validation was performed according to the European MedicinesAgency (EMA) and Clinical and Laboratory Standards Institute(CLSI)-International Federation of Clinical Chemistry and Labora-tory Medicine (IFCC) guidelines [29–31].

2.8.1. CalibrationEight-level calibration standards containing DCX were pro-

cessed daily before total and free DCX concentration measurement.Integration of smoothed peak areas and calculation of DCX concen-trations was performed with TargetLynxTM v 4.1 software (Waters,Milford, MA, USA).

According to the EMA guideline [30], calculated concentrationsof the calibration standards were all within ±15% of the nominalvalue, except for the lower limit of quantification (LLOQ) for whichit was within ±20%.

The calibration curves were generated by linear fit of theDCX/PCX standard area response ratio versus DCX concentration(1/X2 weighting; excluding the option to force through the point oforigin).

2.8.2. SelectivityTen different batches of plasma and its plasma ultrafiltrate from

patients not treated with DCX or PCX but receiving, in combinationor alone, other drugs as anticonvulsants, antibiotics, immuno-suppressants, methotrexate or digoxin were used. Two patientsreceived cyclosporine A, mycophenolic acid and everolimus; two,valproic acid and tazobactam; one, ceftriaxone and tacrolimus; one,pheytoin and metotrexate; one, sirolimus and mycophenolic acid;

one, digoxin; one, carbamazepine; and one, cefepime. Drug concen-trations were in the respective therapeutic intervals or maintainedabove the minimal inhibitory concentration of the pathogen.

According to the EMA guideline [30], the absence of interferingcomponents is accepted when the peak area response of interfering

ical an

pL

2

dtcbo

teadtar

2

accfrhlr

2

ie

C

wtTi

cprf

2

cse4pc(tt

ewm(

R. Rigo-Bonnin et al. / Journal of Pharmaceut

eak at the retention time of analyte (DCX) is less than 20% of theLOQ for the analyte (DCX) and 5% for the internal standard (PCX).

.8.3. Limit of detection and limit of quantificationThe CLSI-IFCC C50-A guideline [29] defines the lower limit of

etection (LLOD) as the lowest concentration where the signal-o-noise (S/N) ratio is three or more, and the LLOQ as the lowestoncentration at which the S/N ratio is 10 or more and that coulde estimated with an acceptable inter-day imprecision (coefficientf variation of ≤20%).

To estimate the LLOD and the LLOQ, the lowest working calibra-or samples (10 �g/L for total DCX and 1.0 �g/L for free DCX) wereither not diluted, diluted 2-fold or 5-fold with 0.1% (v/v) aceticcid in methanol. These dilutions were treated with the previouslyescribed procedure for calibrators to obtain nominal concentra-ion in plasma of 4.0, 2.0 and 0.8 �g/L for total DCX and 0.40, 0.20nd 0.08 �g/L for free DCX. Each working sample was processedepeatedly 15 times in one day and in a single series per day.

.8.4. Carry-overIn accordance to the EMA guideline [30], carry-over was

ssessed by injecting a blank sample (0.0 �g/L) after a high-oncentration sample at the upper limit of quantification (nominaloncentration in plasma of 1000 �g/L and 100 �g/L for total andree DCX, respectively). Carry-over is acceptable if the peak areaesponse in the blank sample obtained after measurement of theigh-concentration sample is not greater than 20% of the ana-

yte (DCX) peak area response at the LLOQ, and 5% the peak areaesponse of the internal standard (PCX).

.8.5. Imprecision and biasFour samples, three QC and one LLOQ, were used to estimate

ntra- and inter-day imprecision and bias according to the followingquations:

V(%) =(s

x

).100 ır(%) =

[(x − �

�

)]× 100

here CV, s,x, ır and � are coefficient of variation, standard devia-ion, mean, relative bias and the conventional value, respectively.he conventional value of the QC samples was assigned by weight-

ng procedures.For intra- and inter-day imprecision and bias, 15 aliquots of each

oncentration were tested repeatedly in 1 day and in a single serieser day, for 30 days, respectively. Coefficient of variation and ır

esults were analyzed following the EMA acceptance criteria (15%or QC materials and 20% for LLOQ).

.8.6. Measurement interval (linearity)According to the EMA and the CLSI EP6-A guidelines [30,31], to

alculate the measurement interval, the highest working calibratoramples (2500 �g/L for total DCX and 250 �g/L for free DCX) wereither diluted with 0.1% (v/v) acetic acid in methanol at ratios of:0, 3:1, 2:2, 1:3 and 0:4. These dilutions were treated with thereviously described procedure for calibrators to achieve theoreti-al nominal concentrations in plasma of 1000, 750, 500, 250, 0 �g/Ltotal DCX) and 100, 75, 50, 25 and 0 �g/L (free DCX). According tohe EMA guideline [30], each dilution samples was processed inriplicate, and imprecision and bias should be ±15%.

The measured values were plotted on the y-axis versus thexpected or known values on the x-axis. The measured valuesere considered to be linear following criteria from the CLSI docu-ent EP6-A [31], as assessed in the Analyse-it® statistical software

Analyse-it Software, Ltd., Leeds, UK).

d Biomedical Analysis 117 (2016) 140–149 143

2.8.7. RecoveryA post-extraction spiking experiment was performed to deter-

mine the recovery of the DCX and IS (PCX) in plasma andultrafiltrate following extraction with MTBE. The extraction recov-ery of DCX and free DCX was measured by comparing the mean peakareas obtained from blank plasma or ultrafiltrate samples spikedwith DCX before extraction (A) with those from blank plasma orultrafiltrate to which DCX was added after extraction (B). The ratioA/B was used to evaluate extraction recovery. The recovery wascalculated in six different lots of matrix plasma or ultrafiltrate atthree different concentrations (20, 200 and 600 �g/L for total DCXand 2.0, 20 and 60 �g/L for free DCX). The recovery of PCX wassimilarly studied at the concentration 100 �g/L.

According to the CLSI-IFCC C50-A guideline [29], the variationin recovery between all concentrations should be less than 15%.

2.8.8. Matrix effectAccording to the EMA guideline [30] and Viswanathan et al. [32],

the quantitative measure of the matrix effect can be termed as thematrix factor and defined as the ratio of the peak area response inthe presence of the matrix (measured by analyzing a blank matrixspiked after extraction with analyte) to the peak area response inthe absence of the matrix (pure solution of analyte):

Matrix Factor = Peak Area response in presence of matrix componentsPeak Area response in absence of matrix components

A matrix factor greater than 1 may be due to ion enhancementand that less than 1 may be due to ion suppression. Similarly, theinternal standard can also experience ion enhancement or ion sup-pression.

To take into account the matrix effects of the internal standard(in our case, PCX), a PCX-normalized matrix factor was calculated bydividing the matrix factor of the DCX by the matrix factor of the PCX.The PCX-normalized matrix factor was calculated in six differentlots of matrix plasma or ultrafiltrate at three different concentra-tions (20, 200 and 600 �g/L for total DCX and 2.0, 20 and 60 �g/Lfor free DCX) to determine the variability of the matrix effect insamples from different individuals.

According to the CLSI-IFCC C50-A and the EMA guidelines[29,30], the variability in matrix effect as measured by the coef-ficient of variation should be less than 15% and the variation inmatrix effect between all concentrations should be less than 15%.

2.8.9. Stability studyStability studies included stock solution stabilities of DCX and

PCX, extracted samples in-autosampler stability and short- andlong-term stabilities for DCX and free DCX.

To evaluate the stability of stock solutions, the peak arearesponse of the stock solutions refrigerated at (5 ± 3) ◦C for 1, 3and 7 days and at (−75 ± 5) ◦C for six months were compared withfresh stock at room temperature.

The stability of samples in the autosampler was tested, reinject-ing them after a 6, 12 and 24 h storage at (4 ± 1) ◦C.

Short- and long-term stabilities for DCX and free DCX were con-ducted in various storage conditions at three plasma or ultrafiltrateQC concentration levels (20, 200 and 600 �g/L for total DCX and2.0, 20 and 60 �g/L for free DCX), and 10 replicates of each concen-tration was used. To evaluate short-term stability, the QC sampleswere first stored at (5 ± 3) ◦C for 1, 3 and 7 days and then equili-brated to room temperature and extracted and tested against theirfresh counterparts. For long-term stability evaluation, the QC sam-ples were first frozen at (−75 ± 5) ◦C for 6 months and then thawed

before extraction and tested against fresh calibration and QC sam-ples.

The EMA guideline [30] defines stable samples as having a meanconcentration at each level within ±15% of the nominal concentra-tion.

1 ical an

3

mtraibWf

3

nna4ep

3

cpi4t(

WNctu

4

U1sti

4

trrl

4

r11

4

s

44 R. Rigo-Bonnin et al. / Journal of Pharmaceut

. Application to biological samples

The UPLC–MS/MS methods here described were developed toeasure total and free DCX in human plasma for PK investiga-

ions, from pacients with operable nonmetastatic breast carcinomaeceiving neoadjuvant chemotherapy, and it is currently applied in

phase IV trial research protocol. This study was carried out accord-ng to the principles of the Declaration of Helsinki and was approvedy the local Ethics Committee of Catalan Institute of Oncology.ritten informed consent according to local practice was obtained

or every patient.

.1. Patients and docetaxel administration

Women who had primary operable breast cancer receivingeoadjuvant chemotherapy are being included in the clinical trial,amely patients with tumors >20 mm (T2, T3, T4, N0, N1 and N2)nd the absence of metastatic disease. Chemotherapy consisted of

cycles of intravenously 1h-infusion of 100 mg/m2 dose of DCXvery 21 days, followed by 4 cycles of intravenously doxorubicinlus cyclophosphamide.

.2. Sample collection and pharmacokinetics parameters

Pharmacokinetic sampling occurred for the first and the fourthycles. Approximately, five milliliters of blood were collected inlasma tubes with EDTA-K3 (Vacuette, Kremsmünster, Austria)

mmediately before starting the drug infusion and at 0.5, 1.0, 2.0,.0, 8.0, 12 and 24 h post-dose. Blood samples were quickly cen-rifuged at 1600 × g for 10 min at (4 ± 1) ◦C, aliquoted and stored at–75 ± 5) ◦C until analysis.

Plasma concentration–time data were analyzed using PhoenixinNonlin® software v.7.3 (Certara USA Inc, Princeton, NJ, USA).

on-compartmental model was utilized to calculate the pharma-okinetic parameters. The total clearance (CL) and the area underhe plasma concentration–time curve (AUC0–24) were estimatedsing the linear trapezoidal method from time 0 to 24 h.

. Results

Under the chromatographic conditions described above forPLC–MS/MS methods, the retention times for DCX and PCX were.94 and 1.74 min, respectively. Typical MRM chromatograms arehown in Fig. 2. The UPLC–MS/MS run time was 3.5 min, includinghe time necessary to return to baseline conditions before the nextnjection.

.1. Selectivity

No interfering peaks were present in any plasma or ultrafil-rate sample from patients receiving other drugs. The peak areaesponses observed in all plasma or ultrafiltrate batches at DCXetention time were ≤1.8% of the LLOQ of total and free DCX, beingess than 0.9% at the PCX retention time.

.2. Limit of detection and limit of quantification

The LLOD were 0.96 �g/L (S/N ratio of 7.8) and 0.21 �g/L (S/Natio of 3.2) for DCX and free DCX, respectively. The LLOQ were.95 �g/L (S/N ratio of 17.5, CV of 12.4%) and 0.42 �g/L (S/N ratio of0.2, CV of 19.6%) for DCX and free DCX, respectively.

.3. Carry-over

Peak area responses observed in the blank samples after mea-urement of the high-concentration samples were 5.9% and 7.7%

d Biomedical Analysis 117 (2016) 140–149

of the DCX peak area response at the LLOQ for DCX and free DCX,respectively. On the other hand, peak area responses were 2.6% and3.8% of the peak area response of the PCX.

4.4. Imprecision and bias

Data for intra-day and inter-day imprecision and relative biasare summarized in Table 1. Imprecision and relative bias werelower than the maximum permissible requirements for thesemetrological characteristics (15% for quality control materials and20% for LLOQ).

4.5. Measurement interval (linearity)

The measurement interval was found linear between 1.95 and1000 �g/L and 0.42 and 100 �g/L for DCX and free DCX, respec-tively. The resulting mean measured values of each sample werecompared to predicted concentrations and yielded a linear regres-sion of y = 0.939 × x + 15.4 (r 2 = 0.9909) and y = 0.934 × x + 3.88(r2 = 0.9915) for DCX and free DCX, respectively. Dilution integrityof the samples achieved acceptance criteria for imprecision and bias(15%).

4.6. Recovery

The recoveries of DCX at 20, 200, 600 �g/L and PCX at 100 �g/Lwere (79.4 ± 9.6)%, (83.1 ± 8.2)%, (81.5 ± 6.2)% and (87.1 ± 9.0)%,respectively. In the case of free DCX at 2.0, 20, 60 �g/L andPCX at 100 �g/L the recoveries were (86.2 ± 8.4)%, (89.3 ± 7.2)%,(90.0 ± 6.9)% and (93.4 ± 8.1)%, respectively.

4.7. Matrix effect

Values for matrix factor, variabilities of matrix effect and PCX-normalized matrix factor in different lots of plasma are summarizedin Table 2. Evaluation of the matrix effect showed ion suppressionfor DCX and free DCX. The variation in matrix effect between allconcentrations was less than 15%. As shown in Table 2, the matrixeffects were well compensated by the internal standard, PCX.

4.8. Stability study

Total and free DCX concentrations in plasma were stable duringstorage at (5 ± 3) ◦C for a period of 7 days with an absolute percentdeviation of the nominal concentration (%D) lower than 13.1%. Inplasma sample extracts, total and free DCX concentrations werestable in autosampler at (4 ± 1) ◦C for 24 h (absolute%D ≤ 9.7%).Also, plasma total and free DCX concentrations at (−75 ± 5) ◦C werestable for at least 6 months (absolute%D ≤ 5.9%).

Stock solutions of DCX and PCX stored at (5 ± 3) ◦C were stablefor 7 days (absolute%D of 11.9% and 10.8%, respectively). Also, stocksolution of DCX at (−75 ± 5) ◦C was stable for 6 months (absolute%Dof 6.9%).

Percent deviations were in all cases negative indicating a loss ofDCX or PCX concentration in all matrices with regard to the nominalvalue, i.e., a decomposition or degradation of DCX or PCX occurred.

5. Clinical application

The UPLC–MS/MS method here described was developed tomeasure total and free DCX concentrations in human plasma for

pharmacokinetic investigations and it is currently applied in aresearch protocol.

Total and free DCX concentration values in human plasmaare being used for developing a population PK/PD model inwomen who had primary operable breast cancer receiving

R. Rigo-Bonnin et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 140–149 145

Fig. 2. A representative MRM total ion chromatogram obtained from human (A) plasma patient sample at concentration of 161,7 �g/L and (B) ultrafiltrate plasma patientsample at concentration of 11,0 �g/L.

146 R. Rigo-Bonnin et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 140–149

Table 1Intra-day and inter-day imprecision and bias values obtained in UPLC–MS/MS for docetaxel and free docetaxel concentration in human plasma.

Quantity/theorical concentration in �g/L (Sample type) Intra-day (n = 15) Inter-day (n = 15)

x± s (�g/L) CV (%) ır (%) x ± s (�g/L) CV (%) ır (%)

P—Docetaxel; mass c.2.0 (LLOQ) 1.92 ± 0.19 10.0 −4.0 1.95 ± 0.24 12.4 −2.520.0 (QC1) 19.0 ± 1.31 6.9 −5.0 18.0 ± 1.51 8.4 −10.0200 (QC2) 196 ± 9.8 5.0 −2.0 192 ± 11.7 6.1 −4.0600 (QC3) 592 ± 24.3 4.1 −1.3 589 ± 30.0 5.1 −1.8

P—Docetaxel(free); mass c.0.40 (LLOQ) 0.38 ± 0.07 18.1 −5.0 0.42 ± 0.08 19.6 5.02.0 (QC1) 1.97 ± 0.23 11.6 −1.5 1.94 ± 0.27 13.8 −3.020.0 (QC2) 18.9 ± 1.49 7.9 −5.5 18.6 ± 1.82 9.8 −7.060.0 (QC3) 58.5 ± 3.22 5.5 −2.5 57.9 ± 3.47 6.0 −3.5

n: number of samples processed;x: mean values; s: standard deviation; CV: coefficients of variation; ır: relative bias; LLOQ: low limit of quantification; QC1: internal qualitycontrol 1; QC2: internal quality control 2; QC3: internal quality control 3. According to IUPAC and IFCC: P: plasma; mass c.: mass concentration.

Table 2Matrix factor and PCX-normalized matrix factor for docetaxel in plasma and free docetaxel in ultrafiltrate plasma.

Matrix lot Matrix factor (%) PCX-normalized matrix factor (%)

P—Docetaxel; mass c. 20 mg/L 200 mg/L 600 mg/L 20 mg/L 200 mg/L 600 mg/L

1 84.3 93.4 96.5 94.2 96.7 99.02 77.3 88.6 92.2 96.6 100.5 98.23 83.4 85.9 87.6 94.6 97.7 95.14 88.2 92.7 94.4 94.0 97.2 99.05 69.8 77.6 85.5 90.5 88.2 94.56 74.2 81.1 83.3 87.8 91.7 93.9

x (%) 79.5 86.6 89.9 93.0 95.3 96.6s (%) 6.9 6.3 5.2 3.2 4.5 2.4CV (%) 8.7 7.3 5.8 3.4 4.7 2.4

Matrix lot Matrix factor (%) PCX-normalized matrix factor (%)

P—Docetaxel(free); mass c. 2.0 mg/L 20 mg/L 60 mg/L 2.0 mg/L 20 mg/L 60 mg/L

1 87.7 91.1 95.3 93.8 97.6 96.52 80.9 78.8 92.3 96.1 93.4 100.13 90.1 91.3 90.2 97.9 95.6 100.14 95.4 97.7 94.7 101.3 101.2 96.95 71.1 77.7 79.4 86.6 94.5 95.46 75.4 78.8 83.3 97.0 99.9 98.7

x (%) 83.4 85.9 89.2 95.4 97.0 98.0

x PAC an

nmicmt

bDrbh1Dar1t

pin

s (%) 9.2 8.5

CV (%) 11.1 9.9

¯ : mean values; s: standard deviation; CV: coefficients of variation. According to IU

eoadjuvant chemotherapy. Pharmacokinetic/pharmacodynamicodeling could be guide DCX dosage optimization, thus

ncreasing the clinical effectiveness. Because of its multi-ompartmental behavior and large volume of distribution, aethod with sufficient detection capability is of great impor-

ance.The plasma concentration–time curve of total and free DCX in a

reast cancer patient after i.v. 1h-infusion of 100 mg/m2 dose ofCX for the first and the fourth cycles are shown in Fig. 3. The

esults indicated that the described methods had a sufficient capa-ility of detection to measure total and free DCX concentration inuman plasma for up to 16 h after the administration of a dose of00 mg/m2. The values of CL and AUC0-t in a breast patient for totalCX were 85.67 L/h (44.92 L/h/m2) and 1.86 mg h/L at the first cyclend 92.6 L/h (48.56 L/h/m2) and 1.72 mg h/L at the fourth cycle,espectively. For free DCX, the CL and AUC0–t were 1343 L/h and18.34 �g h/L at the first cycle and 1204 L/h and 132.03 �g h/L athe fourth cycle, respectively.

The pharmacokinetics of DCX has been investigated in severalhase I and phase II trials [1–4,34]. The maximum tolerated dose

s 80–115 mg/m2 and the main dose-limiting toxicity is neutrope-ia. A relationship has been observed between systemic exposure

6.5 5.0 3.1 2.07.2 5.2 3.2 2.0

d IFCC: P: plasma; mass c.: mass concentration.

to DCX, as measured by the AUC, and neutropenia [1]. However,to date, a widely accepted AUC value to establish the best rela-tionship between efficacy and toxicity of DCX, is still lacking.Engels et al. [33] suggested to use a weighted mean for the tar-get AUC (4.90 mg h/L) that was based on several representativeDCX PK studies, including a total of 806 patients treated with DCX100 mg/m2.

The results of our patient were comparable with previousreports in PK studies that used similar doses of DCX. Bruno R.et al. [3] reported a CL mean value of 36.5 L/h (percentile 5–95%:17.5–59.3) and AUC mean value of 4.81 mg h/L (percentile 5–95%:2.93–9.52) with a considerable interpatient PK variability; and Ros-ing H. et al. [34] (who used 1-h intravenous infusion at a dose levelof 100 mg/m2 with PK monitoring) reported a CL mean value of34 L/h/m2 (range: 19.2–53.8) and a AUC mean value of 3.1 mg h/L(range: 1.4–5.2).

Given that DCX metabolization is mainly carried out by CYP450enzymes, there is a potential for altered pharmacokinetics and

therapeutic effect due to high protein-binding in the plasma anddrug–drug interaction. For this representative patient, the percent-age free fraction of drug in each sample in cycles 1 and 4 were(6.01 ± 1.34)% and (7.12 ± 2.24)%, respectively and did not appear

R. Rigo-Bonnin et al. / Journal of Pharmaceutical and Biomedical Analysis 117 (2016) 140–149 147

F ing 1

a

tr

6

mpT(ba[t

6

crpUmopwwpM(pf5(peC

ig.3. Free and total concentration-time curve of docetaxel in a typical patient receivre for the first and the fourth cycles.

o vary with time. These free fraction data agree with all previouseports [1,4,24,25].

. Discussion

UPLC–MS/MS methods were developed and validated for theeasurement of DCX and free DCX concentrations in human

lasma and are being currently applied in a research protocol.hese methods could support the therapeutic drug monitoringTDM) of DCX in different patients, particularly in women withreast cancer receiving neoadjuvant chemotherapy. Intrinsic char-cteristics of the UPLC–MS/MS technique let us achieve similar13,15,24] or shorter [10–12,14,16–21,25–26] retention times thanhose reached using other methods.

.1. Method development

Various combinations of mobile phase and reverse-phase UPLColumns were tested to achieve a good resolution and symmet-ic peaks, a high response, a short retention time and bettereak shape. Different mobile phases were evaluated to improvePLC separation and to enhance MS sensitivity. Several experi-ents were performed testing different mobile phases consisting

f water, on one hand, and acetonitrile and methanol as organichases on the other hand. All these mobile phases were combinedith ammonium acetate (2 mM), with formic acid at 0.1% (v/v) orith both additives. From all the possible combinations, that com-

osed of methanol and water and both additives offered the highestS response. Two kinds of Bridget Ethyl Hybrid UPLC columns

Acquity® UPLC® BEHTM C18 reverse-phase columns) with the samearticle size (1.7 �m) and internal diameter (2.1 mm) but with dif-

erent length (50 mm vs 100 mm) were evaluated jointly. With the

0 mm-lenght BEH column shorter retention times were obtained0.62 min for PCX and 0.73 min for DCX) but it presented widereaks and worst peak shapes, probably, because it was near to thelution front. It was found that the use of an Acquity® UPLC® BEHTM

18 reverse-phase column, 2.1 × 100 mm; 1.7 �m, in combination

h-infusion of 100 mg/m2 docetaxel i.v. every 21 days. Pharmacokinetic data showed

with isocratic elution mode based on 2 mM ammonium acetatewith 0.1% formic acid in methanol–water (70:30, v/v), let us achievethe chromatographic conditions mentioned above. Other parame-ters such as column temperature, flow rate and injection volumewere studied in order to get a fast and reliable separation, and thebest results were obtained when 35 ◦C was used as column temper-ature (versus 25 ◦C, 45 ◦C or 55 ◦C), 0.4 mL/min as flow rate (betterthan 0.3 mL/min or 0.5 mL/min) and 20 �L were injected (versus5 �L or 10 �L). Under all these conditions, retention times of DCXand PCX were constant and reproducible.

All MS parameters were optimized by direct injection of 10 mg/Lof DCX and PCX in a methanol/water solution containing 0.1%acetic acid (50/50 v/v) into the mass spectrometer at a flowrate of 20 �L/min. In our case, the most abundant ions obtainedwere the [M–H]+ adducts. The choice of the monitored ions wasmade after studying the MS/MS fragmentation pattern of DCX andPCX. DCX and PCX were quantified using the MRM mode dueto its high-sensitivity data acquisition when the precursor andthe product ions are monitored. To prevent analytes misidenti-fication, and specifically to confirm the presence of the analytesand the absence of false contributions from similar componentscoeluted in the samples, two MRM transitions were followed forDCX and PCX. One transition was used for quantification (the quan-tifier): 808.3 → 527.0 for DCX and 854.3 → 569.0 for PCX, and theother transition was monitored for identification (the qualifier):808.3 → 509.0 and 854.3 → 509.0 for DCX and PCX, respectively.The quantifier to qualifier ratio was used for peak identificationbased on criteria set forth by the CLSI C50-A guideline [29]. Resultsreport peak ratios of the two peaks which did not deviate from theaverage ratio in the standards by more than the 20%, indicating thatthere is no analyte misidentification.

Although liquid–liquid extraction is not probably the more ade-quate method to prevent the matrix effects, in our evaluation,

a liquid–liquid with MTBE followed by dryness under nitrogenevaporation and further redissolution of the residues in mobilephase simplified the extraction procedures published by others[11,15,19,21], and provided acceptable recoveries and matrix effect

1 ical an

rittstoocpanpreaiiDas

sappbtbscl

6

aadaRnifforafol0ac

7

dtaicd

[

[

[

[

[

[

[

[

48 R. Rigo-Bonnin et al. / Journal of Pharmaceut

esults. The evaluation and the variability of the matrix effectn samples from different individuals are a crucial aspect. Thesewo issues are often not properly studied and could compromisehe analysis performance. An ideal internal standard should be atructural analog or a stable labeled compound, and should trackhe analyte during the extraction and compensate for any analyten the column and any inconsistent response. Due to problemsf availability and the high price of stable labeled compounds, ahemical structural analog of DCX with similar physico-chemicalroperties, PCX, was the first choice for the analysis as otheruthors [10,12,13,16–18]. Although, in our case, DCX and PCX didot elute simultaneously and this could lead to a lack of com-ensation for matrix effect and other characteristics, for example,elated to extraction efficiency, in our evaluation of the matrixffect we observed that the concentration of the three samplesssayed showed a steady value, given that the use of PCX as annternal standard compensates for the ion suppression observedn the DCX. Besides, PCX is unlikely to be co-administered withCX and will therefore be subject to similar matrix effects. Forll these reasons, we considered PCX as an adequate internaltandard.

The major advantage of our method is the possibility of mea-uring total and free DCX concentrations in a same run. Anotherdvantage is a chromatographic run time of only 3.5 min per sam-le, which is considerably shorter than that of all other methodsreviously reported [10–12,14,16–21,25–26]. Although the com-ination of sample preparation and global chromatographic runime can achieve just a moderate throughput, TDM of DCX coulde combined with TDM of other drugs, such as immunosuppres-ants, on the same instrument, which in our case also use the samehromatographic solvents. Hence, delays due to priming and equi-ibration of solvents are minimized.

.2. Validation procedure

In the validation procedure, different analytical characteristicsnd studies were evaluated. No interfering peaks were present inny plasma or ultrafiltrate sample from patients receiving otherrugs indicating that the proposed UPLC–MS/MS methods providecceptable selectivity. Also, no significant carry-over was observed.ecovery values from samples studied were above 79.4% and PCX-ormalized matrix factors ranged between 93.0 to 98.0%. The

mprecision and bias values obtained, for each concentration, wereound to neither exceed the 15% of the CV for QC samples nor the 20%or LLOQ, thus conforming to the EMA criteria [30], and were betterr similar to those of previous publications [12,15–18,20,24]. Theseesults indicate that the proposed UPLC–MS/MS methods providecceptable precision and trueness. Lower limits of quantificationor total and free DCX were better or the same that some meth-ds reported [10,11,13,14,16,25]. UPLC–MS/MS methods presented

inearity within the concentration range of 1.95 and 1000 �g/L and.42 and 100 �g/L for total and free DCX, respectively. Linearitynd the LLOQ obtained were acceptable considering the expectedoncentrations in the PK study.

. Conclusions

In conclusion, the simple UPLC–MS/MS methods that we haveeveloped and validated for measurement of total and free concen-

rations of DCX in human plasma could be useful for both PK studiesnd TDM in the daily practice of the clinical laboratory, consider-ng their time of analysis, practicability and analytical performanceharacteristics as selectivity, precision, trueness, capability ofetection, linearity, recovery and matrix effect. In addition, these

[

d Biomedical Analysis 117 (2016) 140–149

methods allow measurement of both total and free DCX concentra-tions within the same run.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

The authors thank staff members of Pharmacy Department ofCatalan Institute of Oncology, IDIBELL’s Clinical Research Unit (UCI-CEC) and the Biobank HUB-ICO-IDIBELL. This work was supportedby a grant from Ministerio de Sanidad y Consumo (EC11-4054).

References

[1] S.D. Baker, A. Sparreboom, J. Verweij, Clinical pharmacokinetics of docetaxel:recent developments, Clin. Pharmacokinet. 45 (2006) 235–252.

[2] R. Bruno, N. Vivier, J.C. Vergniol, S.L. De Phillips, G. Montay, L.B. Sheiner, Apopulation pharmacokinetic model for docetaxel (Taxotere): model buildingand validation, J. Pharmacokinet. Biopharm. 24 (1996) 153–172.

[3] R. Bruno, N. Vivier, C. Veyrat-Follet, G. Montay, G.R. Rhodes, Populationpharmacokinetics and pharmacokinetic–pharmacodynamic relationships,Invest. New Drugs 19 (2001) 163–169.

[4] H. Minami, K. Kawada, Y. Sasaki, T. Igarashi, T. Saeki, M. Tahara, et al.,Pharmacokinetics and pharmacodynamics of protein-unbound docetaxel incancer patients, Cancer Sci. 97 (2006) 235–241.

[5] L. Zufía López, A. Aldaz Pastor, J.M. Aramendia Beitia, J. Arrobas Velilla, J.Giraldez Deiro, Determination of docetaxel and paclitaxel in human plasmaby high-performance liquid chromatography: validation and application toclinical pharmacokinetic studies, Ther. Drug. Monit. 28 (2006) 199–205.

[6] A. Andersen, D.J. Warren, P.F. Brunsvig, S. Aamdal, G.B. Kristensen, H. Olsen,High sensitivity assays for docetaxel and paclitaxel in plasma usingsolid-phase extraction and high-performance liquid chromatography with UVdetection, BMC Clin. Pharmacol. 6 (2) (2006) 1–10.

[7] J. Ciccolini, J. Catalin, M.F. Blachon, A. Duran, Rapid high-performance liquidchromatographic determination of docetaxel (Taxotere) in plasma usingliquid–liquid extraction, J. Chromatogr. B Biomed. Sci. Appl. 759 (2001)299–306.

[8] M.B. Garg, S.P. Ackland, Simple and sensitive high-performance liquidchromatography method for the determination of docetaxel in human plasmaor urine, J. Chromatogr. B Biomed. Sci. Appl. 748 (2000) 383–388.

[9] S.L. Richheimer, D.M. Tinnermeier, D.W. Timmons, High-performance liquidchromatographic assay of taxol, Anal. Chem. 64 (1992) 2323–2326.

10] S. Gao, J. Zhou, F. Zhang, H. Miao, Y. Yun, J. Feng, et al., Rapid and sensitiveliquid chromatography coupled with electrospray ionization tandem massspectrometry method for the analysis of paclitaxel, docetaxel, vinblastine,and vinorelbine in human plasma, Ther. Drug. Monit. 36 (2014) 394–400.

11] A. Navarrete, M.P. Martínez-Alcázar, I. Durán, E. Calvo, B. Valenzuela, C.Barbas, et al., Simultaneous online SPE-HPLC-MS/MS analysis of docetaxel,temsirolimus and sirolimus in whole blood and human plasma, J. Chromatogr.B Analyt. Technol. Biomed. Life Sci. 921-922 (2013) 35–42.

12] J.J. Hendrikx, A.C. Dubbelman, H. Rosing, A.H. Schinkel, J.H. Schellens, J.H.Beijnen, Quantification of docetaxel and its metabolites in human plasma byliquid chromatography/tandem mass spectrometry, Rapid Commun. MassSpectrom. 27 (2013) 1925–1934.

13] P. Du, N. Li, H. Wang, S. Yang, Y. Song, X. Han, et al., Development andvalidation of a rapid and sensitive UPLC?MS/MS method for determination oftotal docetaxel from a lipid microsphere formulation in human plasma, J.Chromatogr. B Analyt. Technol. Biomed. Life Sci. 926 (2013) 101–107.

14] H. Yamaguchi, A. Fujikawa, H. Ito, N. Tanaka, A. Furugen, K. Miyamori, et al., Arapid and sensitive LC/ESI–MS/MS method for quantitative analysis ofdocetaxel in human plasma and its application to a pharmacokinetic study, J.Chromatogr. B Analyt. Technol. Biomed. Life Sci. 893–894 (2012) 157–161.

15] G. Corona, G.C. Elia, B. Casetta, S. Frustaci, G. Toffoli, High-throughput plasmadocetaxel quantification by liquid chromatography–tandem massspectrometry, Clin. Chim. Acta 412 (2011) 358–365.

16] K.A. Mortier, V. Renard, A.G. Verstraete, A. Van Gussem, S. Van Belle, W.E.Lambert, Development and validation of a liquid chromatography–tandemmass spectrometry assay for the quantification of docetaxel and paclitaxel inhuman plasma and oral fluid, Anal. Chem. 77 (2005) 4677–4683.

17] J. Guitton, S. Cohen, B. Tranchand, B. Vignal, J.P. Droz, M. Guillaumont, et al.,Quantification of docetaxel and its main metabolites in human plasma by

liquid chromatography/tandem mass spectrometry, Rapid Commun. MassSpectrom. 19 (2005) 2419–2426.

18] I.E.L.M. Kuppens, M.J. van Maanen, H. Rosing, J.H.M. Schellens, J.H. Beijnen,Quantitative analysis of docetaxel in human plasma using liquidchromatography coupled with tandem mass spectrometry, Biomed.Chromatogr. 19 (2005) 355–361.

ical an

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

dosing: a randomized pharmacokinetic study, Clin. Cancer Res. 17 (2011)353–362.

[34] N. Rosing, V. Lustig, L.J.C. van Warmerdam, M.T. Huizing, W.W. ten BokkelHuinink, J.H. Schellens, et al., Pharmacokinetics and metabolism of docetaxel

R. Rigo-Bonnin et al. / Journal of Pharmaceut

19] A.G. Grozav, T.E. Hutson, X. Zhoua, R.M. Bukowski, R. Ganapathi, Y. Xua, Rapidanalysis of docetaxel in human plasma by tandem mass spectrometry withon-line sample extraction, J. Pharm. Biomed. Anal. 36 (2004) 125–131.

20] S.D. Baker, M. Zhao, P. He, M.A. Carducci, J. Verweij, A. Sparreboom,Simultaneous analysis of docetaxel and the formulation vehicle polysorbate80 in human plasma by liquid chromatography/tandem mass spectrometry,Anal. Biochem. 324 (2004) 276–284.

21] R.A. Parise, R.K. Ramanathan, W.C. Zamboni, M.J. Egorin, Sensitive liquidchromatography–mass spectrometry assay for quantitation of docetaxel andpaclitaxel in human plasma, J. Chromatogr. B Analyt. Technol. Biomed. LifeSci. 789 (2003) 231–236.

22] M.I. Churchwell, N.C. Twaddle, L.R. Meeker, D.R. Doerge, Improving LC–MSsensitivity through increases in chromatographic performance: comparisonsof UPLC-ES/MS/MS to HPLC-ES/MS/MS, J. Chromatogr. B Analyt. Technol.Biomed. Life Sci. 825 (2005) 134–143.

23] L. Novakova, L. Matysova, P. Solich, Advantages of application of UPLC inpharmaceutical analysis, Talanta 68 (2006) 908–918.

24] P. Du, X. Han, N. Li, H. Wang, S. Yang, Y. Song, Y. Shi, Development andvalidation of an ultrafiltration–UPLC–MS/MS method for rapid quantificationof unbound docetaxel in human plasma, J. Chromatogr. B Analyt. Technol.Biomed. Life Sci. 967 (2014) 28–35.

25] K.A. Mortier, W.E. Lambert, Determination of unbound docetaxel andpaclitaxel in plasma by ultrafiltration and liquid chromatography–andemmass spectrometry, J. Chromatogr. A 28 (2006) 199–205.

26] M.R. Acharya, S.D. Baker, J. Verweij, W.D. Figg, A. Sparreboom, Determinationof fraction unbound docetaxel using microequilibrium dialysis, Anal.Biochem. 331 (2004) 192–194.

27] W.J. Loos, S.D. Baker, J. Verweij, J.G. Boonstra, A. Sparreboom, Clinicalpharmacokinetics of unbound docetaxel: Role of polysorbate 80 and serumproteins, Clin. Pharmacol. Ther. 74 (2003) 364–371.

d Biomedical Analysis 117 (2016) 140–149 149

28] S. Urien, J. Barré, C. Morin, A. Paccaly, G. Montay, J.-P. Tillement, Docetaxelserum protein binding with high affinity to alpha1-acid glycoprotein, Invest.New Drugs 14 (1996) 147–151.

29] Clinical and Laboratory Standards Institute, International Federation ofClinical Chemistry and Laboratory Medicine. Mass Spectrometry in theClinical Laboratory: General Principles and Guidance. C50-A, Wayne, PA: CLSI,2007.

30] European Medicines Agency. Guideline on validation of bioanalyticalmethods; 2011, London. Available at: http://www.ema.europa.eu/docs/enGB/document library/Scientific guideline/2011/08/WC50010968.pdf.(Accessed: 22.02.14.).

31] Clinical and Laboratory Standard Institute, Evaluation of the Linearity ofQuantitative Measurement Procedures: a Statistical Approach; ApprovedGuideline. EP6-A, Wayne, PA: CLSI, 2003.

32] C.T. Viswanathan, S. Bansal, B. Booth, A.J.A.J. De Stefano, M.J. Rose, J. Sailstad,et al., Quantitative bioanalytical methods validation and implementation:best practices for chromatographic and ligand binding assays, Pharm. Res. 24(2007) 1962–1973.

33] F.K. Engels, W.L. Loos, J.M. Van der Bol, P. de Bruijn, R.H. Mathijssen, J. Verweij,et al., Therapeutic drug monitoring for the individualization of docetaxel

administered as a 1-h intravenous infusion, Cancer Chemother. Pharmacol. 45(2000) 213–218.