�sp

Ja

b

c

d

e

a

ARRAA

KACPF9

1

aabdaaa

sltuq

Nf

0d

Journal of Pharmaceutical and Biomedical Analysis 66 (2012) 218– 224

Contents lists available at SciVerse ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

j ourna l ho me p a ge: www.elsev ier .com/ locate / jpba

-Cyclodextrin enhanced on-line organic solvent field-amplified sampletacking in capillary zone electrophoresis for analysis of ambroxol in humanlasma, following liquid–liquid extraction in the 96-well format

i Lia,b, Youwei Bia,b,c, Li Wanga,b,d, Fanlu Suna,b, Zhao Chena,b, Guili Xue,∗, Guorong Fana,b,∗∗

Department of Pharmaceutical Analysis, School of Pharmacy, Second Military Medical University, No. 325 Guohe Road, Shanghai 200433, PR ChinaShanghai Key Laboratory for Pharmaceutical Metabolite Research, No. 325 Guohe Road, Shanghai 200433, PR ChinaSchool of Medicine, Shanghai Jiaotong University, No. 800 Dongchuan Road, Shanghai 200240, PR ChinaDepartment of Pharmaceutical Analysis, China Pharmaceutical University, Nanjing 210009, PR ChinaDepartment of Pharmacy, Kunming General Hospital of Chengdu Military Command, No. 212 Daguan Road, Kunming, Yunnan 650032, PR China

r t i c l e i n f o

rticle history:eceived 5 November 2011eceived in revised form 27 February 2012ccepted 28 February 2012vailable online 8 March 2012

eywords:mbroxol hydrochloride

a b s t r a c t

A field-amplified sample stacking (FASS) and capillary zone electrophoresis (CZE) method is described forthe quantification of ambroxol hydrochloride in human plasma, following liquid–liquid extraction in the96-well format. The separation was carried out at 25 ◦C in a 31.2 cm × 75 �m fused-silica capillary withan applied voltage of 15 kV. The background electrolyte (BGE) was composed of 6.25 mM borate–25 mMphosphate (pH 3.0) and 1 mM �-cyclodextrin. The detection wavelength was 210 nm. Clean-up and pre-concentration of plasma biosamples were developed by 96-well format liquid–liquid extraction (LLE). Inthis study, FASS in combination with �-cyclodextrin enhanced the sensitivity about 60–70 fold in total.

ZElasmaield-amplified sample stacking6-Well format LLE

The method was suitably validated with respect to stability, specificity, linearity, lower limit of quan-titation, accuracy, precision, extraction recovery and robustness. The calibration graph was linear forambroxol hydrochloride from 2 to 500 ng/ml. The lower limit of quantification was 2 ng/ml. The intra-and inter-day precisions of lowest limit of quantification (LLOQ) were 9.61 and 11.80%, respectively.The method developed was successfully applied to the evaluation of clinical pharmacokinetic study of

tablet

ambroxol hydrochloride

. Introduction

Ambroxol (trans-4-(2-amino-3, 5-dibromobenzyl)-minocyclohexanol, Fig. 1), a metabolite of bromhexine [1], isn expectorant and mucolytic agent, which could reduce theronchial hyper-reactivity, stimulate the cellular surfactant pro-uction and increase the amount of antibiotic penetration [2–6]. In

ddition to the mucolytic action, it has also been reported to have

cough-suppressing effect, antioxidant and anti-inflammatoryction [1,7,8]. Ambroxol is administered as hydrochloride in a daily

Abbreviations: CZE, capillary zone electrophoresis; CE, capillary electrophore-is; FASS, field-amplified sample stacking; BGE, background electrolyte; LLE,iquid–liquid extraction; EOF, electroosmotic flow; HD, hydrodynamic; EK, elec-rokinetic; LC–MS/MS, liquid chromatography tandem mass spectrometry; UV,ltraviolet; HPLC, high-performance liquid chromatography; LLOQ, lower limit ofuantification; IS, internal standard; QC, quality control sample.∗ Corresponding author. Tel.: +86 871 5414186; fax: +86 871 5414186.

∗∗ Corresponding author at: School of Pharmacy, Second Military Medical Univ.,o. 325 Guohe Road, Shanghai 200433, PR China. Tel.: +86 21 8187 1260;

ax: +86 21 8187 1260.E-mail addresses: guilixukm@163.com (G. Xu), guorfan@yahoo.com.cn (G. Fan).

731-7085/$ – see front matter © 2012 Published by Elsevier B.V.oi:10.1016/j.jpba.2012.02.026

after oral administration to 12 healthy volunteers.© 2012 Published by Elsevier B.V.

dose of 30–120 mg using oral, rectal, inhalation and intravenousroutes, which produces good results in the treatment of chronicbronchitis [9].

In order to carry out clinical pharmacokinetic studies ofambroxol hydrochloride, a rapid and sensitive analytical methodwas needed, which could allow its determination in plasma. Com-mon separation methods had been reported for the determinationof ambroxol in various matrices based on HPLC determinationswith UV [10], electrochemical [11,12] and mass spectrometricdetection [13–16], and gas chromatography determinations withelectron capture [17,18] and mass spectrometric detection [19].Some of these methods suffered from inadequate sensitivity orlong analysis time. Although LC–MS/MS method can provide excel-lent sensitivity and short run time, the apparatus is expensiveand the matrix effects are difficult to overcome. Capillary elec-trophoresis (CE) is rapidly developing as an alternative analyticaltool with different separation mechanism to HPLC, because of itshigh efficiency, rapid separation, extremely low solvent consump-

tion compared with HPLC, and small sample volume requirement[20]. What is more, the cost of capillaries is far cheaper than thatof HPLC columns. Up to date, only Pérez-Ruiz et al. reported twoCE methods for clinical analysis of ambroxol [9,21]. Disadvantages

J. Li et al. / Journal of Pharmaceutical and Biomedical Analysis 66 (2012) 218– 224 219

a) and

os

ptbeWvtogpFa�irc

o�nwpllo1

2

2

mfDitCC(aCiU

2

itS

Fig. 1. Chemical structures of ambroxol (

f these are tedious derivatization process [9] and low detectionensitivity and long analysis time [21], respectively.

To solve the problem of low sensitivity, field-amplified sam-le stacking (FASS) is usually employed as a simple and efficientechnique in capillary zone electrophoresis (CZE), first introducedy Mikkers et al. [22,23]. It is based on a mismatch between thelectric conductivity of the sample and that of the running buffer.hen compared to injection from a nonstacking sample, con-

entional FASS provided a sensitivity enhancement of about 10-o 20-fold [24]. Shihabi [25] reported that some water-misciblerganic solvents, especially acetonitrile and acetone, bring alongreater degrees (∼30 times) of stacking for the compounds com-ared to that for aqueous buffers or water. However, sometimesASS could result in peak tailing and poor reproducibility, which isttributed to the introduction of matrix in large amount. Therefore,-cyclodextrin which normally acts as carrier electrolyte additive

n capillary electrophoretic (CE) techniques and chiral selector waseported to be used for improving peak shape and separation effi-iency and obtaining good reproducibility of migration times [26].

In this paper, we tried to enhance sensitivity by means ofrganic solvent field-amplified sample stacking combined with-cyclodextrin in capillary zone electrophoresis for the determi-ation of ambroxol hydrochloride in human plasma. The methodas evaluated in terms of selectivity, sensitivity, linearity, accuracy,recision and stability in accordance to the recommendations pub-

ished by the FDA [27], and when combined with 96-well formatiquid–liquid extraction, it was successfully applied to the analysisf ambroxol hydrochloride in clinical pharmacokinetic studies in2 healthy volunteers.

. Experimental

.1. Chemicals and reagents

Ambroxol hydrochloride (99.4% purity) and the tablet for-ulation of ambroxol hydrochloride (30 mg, lot08030151) were

rom Jiangsu Hengrui Pharmaceutical Co. Ltd. (Jiangsu, PR China).iphenhydramine hydrochloride (99.0% purity) was used as the

nternal standard (IS, Fig. 1) obtained from National Institute forhe Control of Pharmaceutical and Biological Product (Beijing, PRhina). Na2B4O7, NaH2PO4 and �-cyclodextrin were acquired fromhina Medicine (Group) Shanghai Chemical Reagent CorporationShanghai, PR China). Human control plasma (sodium heparin asn anticoagulant) was obtained from Kunming General Hospital ofhengdu Military Command (Yunnan, PR China). Water was deion-

zed and purified by using a Milli-Q system (Millipore, Milford, MA,SA) and was used to prepare all aqueous solutions.

.2. Instrumentation

The employed CE system consisted of a Beckman P/ACE MDQnstrument (Beckman Coulter, Brea, CA) equipped with a pho-odiode array detection detector (PDA) and P/ACE System MDQoftware. Detection was performed at 210 nm, where ambroxol had

internal standard diphenhydramine (b).

the maximum absorption. Fused-silica capillaries (31.2 cm × 75 �mi.d., effective length 21 cm) were obtained from Hebei YongnianOptical Fiber Factory (Hebei, China). 96-Well plate refrigerated cen-trifuge (Model SC210A, Thermo Electron, USA) was also used. The2.0 ml Oasis® 96-well plates were purchased from Waters Corpo-ration (Milford, USA).

2.3. Capillary electrophoretic conditions

The background electrolyte (BGE) used in this study was com-posed of 25 mM borate–25 mM phosphate (pH 3.0) and 1 mM�-cyclodextrin. It was prepared by accurately weighing 0.39 gphosphate and 0.24 g borate and making up to 100 ml with deion-ized water. After that, buffer pH was adjusted to 3.0 with 10%phosphoric acid. Then 0.1135 g �-CD was weighed and addedinto the borate–phosphate buffer. The capillary temperature wasmaintained at 25 ◦C and the separation voltage was 15 kV withthe current of about 70 �A. The sample was introduced by usingelectrokinetic (7.5 kV × 15 s) injection modes. BGE were preparedfreshly every day and filtered through a 0.45 �m hydrophilic cellu-lose membrane filter prior to use. A new capillary was conditionedby rinsing with 1 M NaOH, 0.1 M NaOH, H2O and 0.1 M HCl (30 mineach) sequentially. Daily conditioning before start-up was water(2 min), 0.1 M NaOH (10 min), water (2 min), and running buffer(10 min) in regular sequences. Between runs, the capillary wasrinsed with 0.01 M NaOH, H2O and separating buffer (2 min each)sequentially.

2.4. Preparation of stock solutions, calibration samples andquality control samples

Stock solutions of ambroxol hydrochloride and the IS wereprepared in methanol at concentrations of 2 mg/ml and 1 mg/ml,respectively. The ambroxol hydrochloride stock solution wasdiluted with distilled water to working solutions ranging from 20to 5000 ng/ml. A 250 ng/ml IS working solution was obtained bydiluting the stock solution of IS with distilled water. All describedsolutions were protected from light, stored at 4 ◦C.

Calibration samples were obtained by diluting standard work-ing solutions (20 �l) with drug-free human control plasma (180 �l),to span a calibration standard range of 2–500 ng/ml (2, 5, 10, 20,50, 100, 200, and 500 ng/ml). Quality control (QC) samples (5, 20,400 ng/ml) were independently prepared by spiking appropriateamount of the working standard solution in drug-free human con-trol plasma.

2.5. Sample preparation

Samples were prepared using LLE in 96-well format plates. An

eight-channel 300 �l electronic pipetting tool and an eight-channel1200 �l electronic pipetting tool (Eppendorf Xplorer®, EppendorfAG, Hamburg, Germany) were used for liquid transfer steps. Sub-ject plasma samples were thawed at room temperature. 200 �l of

2 nd Bio

swasta15rfowbieA(ct

2

avhi1cTTa

3

3

pIw

3a

(pHrytI6

wcfattBsp

Do

20 J. Li et al. / Journal of Pharmaceutical a

ubject plasma samples, standard curve samples and QC samplesere added individually into a 2 ml deep 96-well plate spiked by

n eight-channel pipetting tool. Aliquots of 30 �l internal standardolution (250 ng/ml) were added, except to the well designated forhe double blank plasma. Then 60 �l 1 M sodium hydroxide weredded to each well of the plate, and vortexed for 30 s. After adding.0 ml of N-hexane, the plate was covered with a mat, vortexed for

min, and centrifuged 10 min at 3500 rpm. The mats were carefullyemoved and 800 �l of the supernatant organic layer was trans-erred from the original sample plates into the respective positionsf new 2.0 ml deep 96-well plates. The height of the transfer stageas carefully chosen so that only the organic layer was aspirated

y an eight-channel pipetting tool. The plates were then placednto a self-constructed 96-well plate evaporator, and the organicxtracts were evaporated to dryness with a nitrogen flow at 30 ◦C.ll dry residues were reconstituted by addition of 50 �l ACN–water

80:20, v/v). Finally, the plates were vortex mixed for 5 min, thenentrifuged for 10 min at 12,000 rpm, and the supernatant wasransferred to the autosampler for injection onto the CE.

.6. Clinical pharmacokinetic study

The method was applied to determine plasma concentrations ofmbroxol hydrochloride from a clinical trial in which 12 healthyolunteers received a single oral dose of 90 mg of ambroxolydrochloride tablet. About 1 ml of blood samples were collected

n heparinized tubes before administration (0 h) and 0.25, 0.50, 1.0,.5, 2.0, 2.5, 3.0, 4.0, 6.0, 9.0, 12.0, 15.0 and 24.0 h after dosing andentrifuged at 3500 rpm for 10 min to separate the plasma fraction.he obtained plasma samples were stored at −20 ◦C until analysis.he study was approved by a local ethics committee and performedfter obtaining written consent from the subject.

. Results and discussion

.1. Method development

In this study, we explored the optimization and validation ofreconcentration process and separation mode. Both ambroxol and

S have strong UV absorption at 210 nm. Therefore the detectionavelength was set at 210 nm.

.1.1. Optimization of borate–phosphate buffer concentrationsnd pH and voltage

Borate–phosphate buffer (pH 3.0) in different concentrations3.125 mM borate–12.5 mM phosphate, 6.25 mM borate–25 mMhosphate, 12.5 mM borate–50 mM phosphate) was studied.igher concentrations of buffer caused the increase in current and

eduction in the EOF, leading to more Joule heating and longer anal-sis time. However, lower concentrations of electrolyte decreasedhe ionic strength of the buffer, which could lower the sensitivity.n order to get the best running time and sensitivity of the analytes,.25 mM borate–25 mM phosphate buffer was selected.

Based on the pKa value (8.2) of ambroxol [28], measurementsere carried out at pH values lower than 7.0. Running buffer

ontaining 6.25 mM borate–25 mM phosphate was studied at dif-erent pH to observe the influence on electrophoretic behavior ofmbroxol and IS in terms of separation efficiency and migrationime. The increase of the pH from 2.5 to 6.0 shortened the migra-ion time of both ambroxol and IS and worsened the resolution.etter resolution and analysis time were obtained when pH < 4,o we chose a running buffer containing 6.25 mM borate–25 mM

hosphate with a pH of 3.0.

Various voltages were tested, including 10, 15, 20, and 25 kV.ue to the big inner diameter (75 �m) and short length (31.2 cm)f capillary, too large separation voltage in our experiment

medical Analysis 66 (2012) 218– 224

induced certain current problems, such as appreciable Joule heat-ing. 15 kV was chosen because it provided optimal separation andefficiency.

3.1.2. Selection of injection modes and sample solventThe sample is usually introduced into the capillary either by

hydrodynamic (HD) or electrokinetic (EK) injection. Under theformer, all components were introduced without discrimination,resulting in more endogenous interference when conducting bio-sample analysis. To avoid those interferences in the CZE mode,the latter was considered a better option for the better sensitiv-ity and selectivity of basic drugs. Furthermore, in our experiment,the peak shapes of both ambroxol and IS in the latter was betterthan those in the former. In addition, EK is reported suitable for on-line FASS [29]. Therefore the electrokinetic injection was chosen asthe optimum.

In order to enhance the response of ambroxol, the composi-tions of sample solvent (10% buffer, distilled water, ACN–distilledwater) were investigated. As shown in Fig. 2, using ACN–distilledwater as sample solvent was good at improving sensitivity. Themechanism of transient pseudo-isotachophoresis has been pro-posed for acetonitrile stacking, in which fast moving anions in thesalts serve as leading ions, whereas water-miscible organic solventsact as the slower moving “terminating ions” [30,31]. For furtherexploration of ACN on ambroxol stacking, different ratios of ACNto distilled water (70%, 80%, 90%) were also investigated. Withthe increase of acetonitrile concentration, ambroxol was stackedmore effectively. However, 90% acetonitrile caused the currentdrop to zero, so we chose 80% ACN–distilled water as samplesolvent.

3.1.3. Concentrations of ˇ-CDAlthough the sensitivity of the analytes was improved over

30-fold, FASS resulted in peak tailing of plasma sample. In ourexperiment, �-CD was added to improve peak shape and repro-ducibility of analytes. Moreover, during this process, the sensitivityof the analytes was improved another 2–3 fold.

In our experiment, using �-CD could get better peak shape andhigher sensitivity. Therefore the effects of �-CD at concentrationsof 0.0, 1.0, 2.0 mM were investigated. As shown in Fig. 3, a runningbuffer containing 1.0 mM �-CD was chosen as the optimum.

3.1.4. Pre-treatment of sampleThe extraction of plasma samples was optimized in our prelim-

inary studies by comparing liquid–liquid extraction, solid-phaseextraction and protein precipitation. Protein precipitation is notsuitable for amboxol extraction from plasma because of lowersensitivity, and more interference of endogenous components inhuman plasma. Solid-phase extraction was time consuming andwasteful. However, the results were satisfactory when N-hexanewas used in liquid–liquid extraction. A 60 �l aliquot of 1 M sodiumhydroxide was added to 200 �l plasma, which enhanced the extrac-tion efficiency of ambroxol and the IS. In order to increase samplethroughput, the LLE 96-well format plates were used, resulting ina shorter sample preparation time.

3.2. Method validation

Analytical method validation was carried out according to therecommendations published by the FDA [27].

3.2.1. Specificity

Specificity of the method was investigated by both peak purity

and spiking experiments with pure standard compounds. Peakpurity was evaluated by means of the P/ACE System MDQ Soft-ware. The total peak purity values of ambroxol hydrochloride and

J. Li et al. / Journal of Pharmaceutical and Biomedical Analysis 66 (2012) 218– 224 221

Fig. 2. The effect of different sample solvents on FASS of ambroxol hydrochloride (400 ng/ml) standard and IS (750 ng/ml). Running buffer, 25 mM phosphate–6.25 mM borate( voltai % ACN

tem

3q

c

FiF

pH 3.0); capillary, 31.2 cm × 75 �m i.d., effective length 21 cm, uncoated; appliednjection, 7.5 kV × 15 s. Electropherograms: (a) 10% buffer; (b) distilled water; (c) 80

he IS were 1.0000. There appeared to be no interference from thendogenous substance, which were analyzed under the same opti-ized condition (Fig. 4).

.2.2. Linearity of calibration curves and lower limits ofuantification (LLOQ)

A weighted (1/x) linear regression was used to perform standardalibration. The mean calibration equation was y = 0.02817 (n = 5,

ig. 3. The effect of different concentrations of �-CD on the peak shape and sensitivity onjection: 7.5 kV × 15 s. Electropherograms: (a) 0 mM �-cyclodextrin; (b) 2 mM �-cyclodig. 2 for other CZE conditions.

ge, 15 kV (+) → (−); column temperature, 25 ◦C; detection 210 nm; electrokinetic–distilled water. Peaks: 1, ambroxol hydrochloride; 2, the IS.

RSD = 0.85%) x + 0.06351 (n = 5, RSD = 4.76%) and r2 = 0.9998 (n = 5,RSD = 0.01%). In this equation, y represents the peak area ratios ofthe analyte to the IS and x represents the plasma concentration ofanalyte in ng/ml. Calibration curves showed excellent linearity in

the range 2–500 ng/ml. The LLOQ, which is determined as the low-est concentration on the standard curve, of ambroxol hydrochloridein human plasma was found to be 2 ng/ml with accuracy of 91.03%and precision of 9.61% (n = 5).

f ambroxol hydrochloride (50 ng/ml) and IS (250 ng/ml) in plasma. Electrokineticextrin; (c) 1 mM �-cyclodextrin. Peaks: 1, ambroxol hydrochloride; 2, the IS. See

222 J. Li et al. / Journal of Pharmaceutical and Biomedical Analysis 66 (2012) 218– 224

Table 1Within- and between-run accuracy and precision in spiked plasma samples.

Sample level Low QC5 ng/ml

Medium QC20 ng/ml

High QC400 ng/ml

Within-run accuracy and precisionValidation run 1Mean ± SD (ng/ml) 5.32 ± 0.43 20.52 ± 0.89 385.87 ± 30.17Accuracy (%) 106.50 ± 8.63 102.58 ± 4.45 96.47 ± 7.54RSD (%) 8.10 4.33 7.81N 5 5 5Validation run 2Mean ± SD (ng/ml) 4.92 ± 0.45 20.30 ± 2.12 392.08 ± 31.49Accuracy (%) 98.41 ± 9.04 101.55 ± 10.58 98.02 ± 7.87RSD (%) 9.19 10.42 8.03N 5 5 5Validation run 3Mean ± SD (ng/ml) 4.97 ± 0.30 19.42 ± 1.13 389.56 ± 9.07Accuracy (%) 97.51 ± 5.97 97.04 ± 5.65 97.37 ± 2.27RSD (%) 6.12 5.82 2.33N 5 5 5Between-run accuracy and precisionMean ± SD (ng/ml) 5.04 ± 0.41 20.07 ± 1.37 389.17 ± 22.43

100.39 ± 6.87 97.29 ± 5.606.84 5.76

15 15

3

itwtw9bto

3

p

Table 2Recoveries of ambroxol hydrochloride and IS in spiked plasma samples (n = 5).

Compound Concentration(ng/ml)

Recovery (%)(mean ± SD)

RSD (%)

Ambroxol 5 78.12 ± 10.79 13.8120 81.46 ± 3.52 4.32

400 85.81 ± 1.76 2.05

Fo

Accuracy (%) 100.80 ± 8.15

RSD (%) 8.08

N 15

.2.3. Accuracy, precision and extraction recoveryFive replicate samples at each QC concentrations were analyzed

n three separate runs. Accuracy was determined by calculatinghe ratios of the predicted concentrations to the spiked values andith the precision expressed as RSD. The results in Table 1 show

hat the within- and between-day variances at three QC levelsere all below 10.42%. It was shown that the accuracy was from

6.47% to 106.50%. The extraction recovery (Table 2) was found toe 78.12 ± 10.79%, 81.46 ± 3.52% and 85.81 ± 1.76% at the concen-ration of 5, 20 and 400 ng/ml, respectively. The extraction recoveryf IS was 82.63 ± 5.22%.

.2.4. StabilityAnalyte stability determinations comprised short-term tem-

erature stability, long-term stability, autosampler stability and

ig. 4. Typical chromatograms of blank plasma (a); blank plasma spiked with 2 ng/ml (LLOf IS at 0.5 h after the dose of 90 mg (c). Peak 1: ambroxol hydrochloride; peak 2: the IS.

IS 250 82.63 ± 5.22 6.32

freeze–thaw cycles stability, which were evaluated by analyz-ing three QC levels in quintuple. The mean values and standard

deviations of the ratios between the concentrations found and ini-tial concentration were used for stability evaluation. The results(Table 3) show that ambroxol hydrochloride had an acceptable

Q) ambroxol hydrochloride and 250 ng/ml IS (b); test plasma spiked with 250 ng/ml

J. Li et al. / Journal of Pharmaceutical and Biomedical Analysis 66 (2012) 218– 224 223

Table 3Stability results of ambroxol hydrochloride in spiked plasma samples (n = 5).

Sample condition Nominal concentration(ng/ml)

Measured concentration(ng/ml) (mean ± SD)

Accuracy (%) RSD (%)

Freeze–thaw stabilitya 5 4.71 ± 0.64 94.30 13.6020 20.80 ± 0.92 104.01 4.44

400 387.44 ± 31.51 96.86 8.13

30-day stabilityb 5 4.53 ± 0.33 90.52 7.3120 19.70 ± 1.21 98.49 6.15

400 392.95 ± 30.60 98.24 7.79

Bench top stabilityc 5 4.96 ± 0.32 99.26 6.4120 18.69 ± 1.03 93.44 5.53

400 407.88 ± 22.26 101.97 5.46

Autosampler stabilityd 5 5.10 ± 0.37 102.08 7.2520 20.43 ± 1.39 102.15 6.82

400 388.34 ± 9.40 97.09 2.42

a After three freeze–thaw cycles.b Stored at −20 ◦C.c Exposed at ambient temperature (25 ◦C) for 8 h.d Kept at ambient temperature (25 ◦C) for 8 h.

Table 4The robustness data of the analytical method upon variation of CZE separation condition (ambroxol hydrochloride (MQC) = 20 ng/ml, n = 5).

Parameter Experimental electrophoretic plate (N)a Resolutiona Found (RSD%) (ng/ml)a

Ambroxol hydrochloride IS

Optimum condition(separation voltage 15 kV;buffer pH 3.0; �-CDconcentration 1 mM)

4063 ± 154 3606 ± 96 4.11 ± 0.98 19.44 ± 1.29 (6.64)

Separation voltage 13.5 kV 3866 ± 182 3584 ± 132 4.77 ± 0.61 18.63 ± 0.85 (4.56)Separation voltage 16.5 kV 3892 ± 146 3585 ± 85 4.50 ± 0.34 19.01 ± 2.09 (10.99)Buffer pH 2.7 3948 ± 125 3627 ± 92 4.26 ± 0.52 19.87 ± 2.23 (11.22)Buffer pH 3.3 4084 ± 103 3559 ± 102 3.81 ± 0.51 20.61 ± 0.90 (4.37)�-CD concentration 0.9 mM 4021 ± 112 3706 ± 127 3.48 ± 0.31 19.26 ± 1.33 (6.91)�-CD concentration 1.1 mM 3932 ± 101 3667 ± 143 5.02 ± 0.45 20.98 ± 1.65 (7.86)

S n the determination of ambroxol hydrochloride (t-test, p > 0.05). Each mean value wasc

ste9r

3

aaodI(dw

3

lhhoatAa

tatistical test results: small changes in experimental conditions have no effect oompared with the mean value obtained by standard conditions.

a Mean ± standard error; RSD, relative standard deviation.

tability at room temperature for 8 h, at −20 ◦C for 1 month, inhe autosampler at room temperature for 8 h after liquid–liquidxtraction and after three freeze–thaw cycles with the values3.44–101.97%, 90.52–98.49%, 97.09–102.15% and 94.30–104.01%,espectively, at the three concentrations studied.

.2.5. RobustnessThe robustness of the method was evaluated by deliberate vari-

tion of the method parameters, such as pH, concentration of �-CDnd voltage. Analyses were carried out in quintuplicate and onlyne parameter was changed in the experiments at a time. Theetermination of 20 ng/ml ambroxol hydrochloride and 250 ng/ml

S spiked plasma samples was performed under various conditionsTable 4). The statistical comparison was done with t-test and noifference was found between results (p > 0.05). Hence, the methodas considered robust and reliable.

.3. Application to clinical pharmacokinetic study

The developed and validated CZE method was used to ana-yze pharmacokinetic profiles of ambroxol hydrochloride in 12ealthy volunteers after a single oral dose of 90 mg of ambroxolydrochloride tablet. Profiles of the mean plasma concentrationf ambroxol versus time are shown in Fig. 5. The value of Tmax

nd Cmax were 1.80 ± 0.80 h and 204.55 ± 35.41 ng/ml, respec-ively. The elimination half-life of ambroxol was 8.14 ± 1.94 h. TheUC0−24 and AUC0−∞ values were 1280.51 ± 450.43 ng h/mlnd 1496.45 ± 601.37 ng h/ml, respectively. Our results of

Fig. 5. Mean plasma concentration–time curve in 12 healthy Chinese subjects whenadministered oral dose of 90 mg of ambroxol hydrochloride.

pharmacokinetic parameters were similar to those resultsobtained by LC/MS/MS [13–16] though the dose of administrationwas different.

4. Conclusion

This work illustrates that preparing the sample in water-miscible organic solvents yields better stacking than that inaqueous buffers, and �-cyclodextrin has an effect on improving

2 nd Bio

ss6bw1hLtopcrwcpta

A

Inat

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[capillary electrophoresis, Electrophoresis 23 (2002) 1612–1617.

[31] M. Wang, Z.W. Cai, L. Xu, Coupling of acetonitrile deproteinization and salting-

24 J. Li et al. / Journal of Pharmaceutical a

ensitivity and reproducibility. Under the proper condition, thistacking CZE method provided a sensitivity enhancement of about0–70 fold in comparison with the same sample dissolved in 10%uffer and separated in buffer without �-cyclodextrin. The methodas validated for the determination of ambroxol hydrochloride in

2 healthy volunteers after a single oral administration of ambroxolydrochloride. The combination of the CZE with 96-well formatLE greatly simplified the preparation process and decreased theime of sample preparation. Furthermore, it saved a great amountf cost of solvent for the determination of large numbers of sam-les. The in vivo analysis process and pharmacokinetic parametersonfirmed that the CZE method was an alternative method foroutine analysis of ambroxol hydrochloride in plasma comparingith the reference method [10,14–16]. The method possessed good

haracteristics of specificity, sensitivity, precision and accuracy androved to be reliable. It is expected that this approach can be appliedo the extraction and analysis of numerous samples in clinical ther-py and pharmacokinetic studies of ambroxol hydrochloride.

cknowledgements

This work was supported by Pharmacokinetic Platform of thennovation Drug Research founded by Ministry of Science and Tech-ology of PR China (2009ZX09301-011-07). The authors gratefullycknowledge laboratory colleagues, Fangyuan Gao and You Li forhe valuable discussion for article composition.

eferences

[1] E. Nemceková, G. Nosálóvá, S. Franová, Ambroxol and protective reflexes of therespiratory tract, Bratisl. Lek. Listy 99 (1998) 111–115.

[2] F. Paganin, O. Bouvet, P. Chanez, D. Fabre, M. Galtier, P. Godard, F.B. Michel, F.Bressolle, Evaluation of the effects of ambroxol on the ofloxacin concentrationsin bronchial tissues in COPD patients with infectious exacerbation, Biopharm.Drug Dispos. 16 (1995) 393–401.

[3] J.C. Wiemeyer, Influence of ambroxol on the broncopulmonary level of antibi-otics, Arzneim.-Forsch. 31 (1981) 974–976.

[4] M.S. Hwang, S. Cho, H. Chung, Y.A. Woo, Nondestructive determination of theambroxol content in tablets by Raman spectroscopy, J. Pharm. Biomed. Anal.38 (2005) 210–215.

[5] J. Spátola, J.J. Poderoso, J.C. Wiemeyer, M. Fernández, R.B. Guerreiro, C. Corazza,Influence of ambroxol on lung tissue penetration of amoxicillin, Arzneim.-Forsch. 37 (1987) 965–966.

[6] K.A. Shaikh, S.D. Patil, A.B. Devkhile, Development and validation of a reversed-phase HPLC method for simultaneous estimation of ambroxol hydrochlorideand azithromycin in tablet dosage form, J. Pharm. Biomed. Anal. 48 (2008)1481–1484.

[7] V. Stetionová, V. Herout, J. Kvetina, In vitro and in vivo antioxidant activity ofambroxol, Clin. Exp. Med. 4 (2004) 152–159.

[8] S. Pfeifer, G. Zissel, K. Kienast, J. Müller-Quernheim, Reduction of cytokinerelease of blood and bronchoalveolar mononuclear cells by ambroxol, Eur. J.Med. Res. 2 (1997) 129–132.

[9] T. Pérez-Ruiz, C. Martínez-Lozano, A. Sanz, E. Bravo, Sensitive method for the

determination of ambroxol in body fluids by capillary electrophoresis and flu-orescence detection, J. Chromatogr. B. 742 (2000) 205–210.

10] M. Nobilis, J. Pastera, D. Svoboda, J. Kvêtina, K. Macek, High-performance liquidchromatographic determination of ambroxol in human plasma, J. Chromatogr.581 (1992) 251–255.

medical Analysis 66 (2012) 218– 224

11] F.J. Flores-Murrieta, C. Hoyo-Vadilo, E. Hong, G. Castaneda-Hernandezet,Assay of ambroxol in human plasma by high-performance liquid chro-matography with amperometric detection, J. Chromatogr. 490 (1989)464–469.

12] H.J. Lee, S.K. Joung, Y.G. Kim, et al., Bioequivalence assessment of ambroxoltablet after a single oral dose administration to healthy male volunteers, Phar-macol. Res. 49 (2004) 93–100.

13] W.Q. Hu, Y. Xu, F. Liu, A.X. Liu, Q.X. Guo, Rapid and sensitive liquid chromatog-raphy tandem mass spectrometry method for the quantification of ambroxolin human plasma, Biomed. Chromatogr. 22 (2008) 1108–1114.

14] F.L. Su, F. Wang, W. Gao, H.D. Li, Determination of ambroxol in human plasma byhigh performance liquid chromatography–electrospray ionization mass spec-trometry (HPLC–MS/ESI), J. Chromatogr. B 853 (2007) 364–368.

15] A.D. Wen, T.J. Hang, S.N. Chen, Z.R. Wang, L.K. Ding, Y. Tian, M. Zhang, X.X. Xu,Simultaneous determination of amoxicillin and ambroxol in human plasmaby LC–MS/MS: validation and application to pharmacokinetic study, J. Pharm.Biomed. Anal. 48 (2008) 829–834.

16] H. Kim, J.Y. Yoo, S.B. Han, H.J. Lee, K.R. Lee, Determination of ambroxol in humanplasma using LC–MS/MS, J. Pharm. Biomed. Anal. 32 (2003) 209–216.

17] J. Schmid, Assay of ambroxol in biological fluids by capillary gas–liquid chro-matography, J. Chromatogr. 414 (1987) 65–75.

18] L. Colombo, F. Marcucci, G.M. Marini, P. Pierfederici, E. Mussini, Determinationof ambroxol in biological material by gas chromatography with electron-capture detection, J. Chromatogr. 530 (1990) 141–147.

19] C.E. Uboh, J.A. Rudy, L.R. Soma, M. Fennell, L. May, R. Sams, F.A. Railing, J.Shellenberger, M. Kahler, Characterization of bromhexine and ambroxol inequine urine: effect of furosemide on identification and confirmation, J. Pharm.Biomed. Anal. 9 (1991) 33–39.

20] K. Ohyamaa, N. Kishikawa, H. Nakagawa, N. Kuroda, M. Nishikido, M. Teshima,H. To, T. Kitahara, H. Sasaki, Simultaneous determination of mycophenolic acidand its acyl and phenol glucuronide metabolites in human serum by capillaryzone electrophoresis, J. Pharm. Biomed. Anal. 47 (2008) 201–206.

21] T. Pérez-Ruiz, C. Martínez-Lozano, A. Sanz, E. Bravo, Determination of bromhex-ine and ambroxol in pharmaceutical dosage forms, urine and blood serum, J.Chromatogr. B 692 (1997) 199–205.

22] F.E.P. Mikkers, F.M. Everaerts, T.P.E.M. Verheggen, High-performance zone elec-trophoresis, J. Chromatogr. 169 (1979) 11–20.

23] C.C. Chen, S.M. Wu, Y.H. Huang, W.K. Ko, H.S. Kou, H.L. Wu, On-linefield-amplified sample stacking in capillary electrophoresis for analysis ofamitriptyline and its metabolite nortriptyline in plasma, Anal. Chim. Acta 517(2004) 103–110.

24] M.C. Breadmore, Recent advances in enhancing the sensitivity of electrophore-sis and electrochromatography in capillaries and microchips, Electrophoresis28 (2007) 254–281.

25] Z.K. Shihabi, Organic solvent high-field amplified stacking for basic compoundsin capillary electrophoresis, J. Chromatogr. A 1066 (2005) 205–210.

26] S. Razee, A. Tamura, T. Masujima, Improvement in the determination of foodadditive dyestuffs by capillary electrophoresis using �-cyclodextrin, J. Chro-matogr. A 715 (1995) 179–188.

27] Guidance for Industry, Bioanalytical Method Validation, US Department ofHealth and Human Services, Food and Drug Adiministration, Center for DrugEvaluation and Research (CDER), Center for Veternary Medicine (CVM), 2001,Available at: http://www.fda.gov/cder/guidance/4252fnl.htm.

28] S. Jacob, A. Shirwaikar, Preparation and evaluation of microencapsulatedfast melt tablets of ambroxol hydrochloride, India J. Pharm. Sci. 71 (2009)276–284.

29] Z. Krivácsy, A. Gelencsér, J. Hlavay, G. Kiss, Z. Sárvári, Electrokinetic injection incapillary electrophoresis and its application to the analysis of inorganic com-pounds, J. Chromatogr. A 834 (1999) 21–44.

30] Z.K. Shihabi, Transient pseudo-isotachophoresis for sample concentration in

out extraction with acetonitrile stacking in chiral capillary electrophoresisfor the determination of warfarin enantiomers, J. Chromatogr. A 1218 (2011)4045–4051.