Sensitive liquid chromatography/mass spectrometry assay for the quantification of azithromycin in human plasma

Copyright © 2007 John Wiley & Sons, Ltd. Biomed. Chromatogr. 21: 1272–1278 (2007)DOI: 10.1002/bmc

1272 F. Liu et al.ORIGINAL RESEARCH ORIGINAL RESEARCH

Copyright © 2007 John Wiley & Sons, Ltd.

BIOMEDICAL CHROMATOGRAPHYBiomed. Chromatogr. 21: 1272–1278 (2007)Published online 29 June 2007 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/bmc.883

Sensitive liquid chromatography/mass spectrometry assayfor the quantification of azithromycin in human plasma

Fei Liu,1 Yu Xu,1* Jinchang Huang,1 Shu Gao2 and Qingxiang Guo1*1Department of Chemistry, University of Science and Technology of China, Hefei 230026, People’s Republic of China2Hefei Con-source Medicine Technology Corporation, Hefei, 230088, People’s Republic of China

Received 17 November 2006; revised 29 April 2007; accepted 11 May 2007

ABSTRACT: A simple, rapid, sensitive and specific liquid chromatography/electrospray ionization mass spectrometry method wasdeveloped and validated to quantify azithromycin in human plasma, using erythromycin as the internal standard (IS). A simplesample preparation method of protein precipitation with methanol was employed. Methanol, acetonitrile and water (12:30:58, v/v/v) were used as the isocratic mobile phase, with 0.1% formic acid and 0.1% ammonium acetate in water. Selected ion monitoringwas specific for azithromycin and erythromycin. The assay was linear over the concentration range 4.69–600 ng/mL. The correla-tion coefficients for the calibration curves ranged from 0.9994 to 0.9998. The intra- and inter-day precisions, calculated from qual-ity control samples, were less than 8.24%. The method was employed in a pharmacokinetic study after oral administration of500 mg azithromycin dispersible tablet to 20 healthy volunteers. Copyright © 2007 John Wiley & Sons, Ltd.

KEYWORDS: azithromycin; HPLC; mass spectrometry; human plasma

INTRODUCTION

Azithromycin [9-de-oxo-9a-aza-9a-methyl-9a-homoerythromycin A dehydrate] is a macrolide anti-biotic. It is chemically related to erythromycin withenhanced spectrum and potency against some bacteria,longer elimination half-life, superior stability in acidicenvironment and higher tissue concentrations (Lalakand Morris, 1993). Low plasma concentration is achi-eved following administration of azithromycin, thusquantification methods for the drug in pharmacokineticstudies and monitoring of its efficacy need to be sensi-tive and specific.

In biological samples several high-performance liquidchromatography (HPLC) methods using electrochemi-cal detection (Shepard et al., 1991; Kees et al., 1998;Taninaka et al., 2000; Gandhi et al., 2000) and fluore-scence detection (Torano and Guchelaar, 1998; Bahramiet al., 2005; Wilms et al., 2005; Bahrami and Mohammadi,2006) have been developed for determination of azi-

thromycin. Fouda and Schneider (1995) described asensitive HPLC method using atmospheric pressurechemical ionization (APCI) mass spectrometric detec-tion for the analysis of azithromycin in serum with alimit of 10 ng/mL. Chen et al. (2006) extracted analytewith methyl tert-butyl ether–hexane (50:50 v/v) andmonitored [M + 2H]2+ by liquid chromatography/massspectrometry (LC/MS). Barrett et al. (2005) used anultrafilter to pretreat sample and Nirogi et al. (2005)used solid-phase extraction to prepare plasma, analysisbeing performed by liquid chromatography/tandemmass spectrometry (LC/MS/MS). All employed compli-cated and time-consuming sample preparation methods.

In the present work, a simple and sensitive LC/MSmethod for the quantification of azithromycin in humanplasma, with erythromycin as the internal standard(IS), is described. Herein a simpler sample preparationmethod of protein precipitation with methanol was em-ployed. This method was applied to a pharmacokineticstudy after administration of 500 mg azithromycin to 20healthy volunteers.

EXPERIMENTAL

Chemicals and reagents. Azithromycin and erythromycin(internal standard, IS) were obtained from the NationalInstitute for the Control of Pharmaceutical and BiologicalProducts (Beijing, China). Methanol and acetonitrile (HPLC-grade) were obtained from Merck (Darmstadt, Germany).

*Correspondence to: Yu Xu, Department of Chemistry, University ofScience and Technology of China, Hefei 230026, People’s Republic ofChina.E-mail: [email protected] Guo, Department of Chemistry, University of Science andTechnology of China, Hefei 230026, People’s Republic of China.E-mail: [email protected]

Abbreviations used: APCI, atmospheric pressure chemical ionization;LLOQ, lower limit of quantitation; SIM, selected ion monitoring.

Contract/grant sponsor: NSFC; Contract/grant number: 20332020.


Quantification of azithromycin in human plasma 1273ORIGINAL RESEARCH

Formic acid and ammonium acetate (HPLC-grade) werepurchased from TEDIA (Fairfield, USA). Distilled water,prepared from demineralized water, was used throughout thestudy. Blank plasma was provided by The First AffiliatedHospital of Anhui Medical University (Hefei, China).

Instruments. An LCQ Advantage MAX ion trap massspectrometer (ThermoFinnigan, USA), fitted with an electro-spray ionization (ESI) source, an Agilent 1100 Quaternarypump and Agilent 1100 autosampler, was used for LC/MSanalysis. Data acquisition was performed using Xcalibur 1.3software (ThermoFinnigan, USA).

LC/MS conditions. Chromatographic analysis was performedusing a Phenomenex Luna CN column (100 × 2.1 mm i.d.,3 µm; Phenomenex, USA) operating at 20°C. A mixture ofwater (containing 0.1% formic acid and 0.1% ammoniumacetate), acetonitrile and methanol (58:30:12, v/v/v) was usedas the isocratic mobile phase at a flow rate of 0.2 mL/min. MSdetection with the ESI source was performed in positive ionmode, using selected ion monitoring (SIM). The protonatedmolecular ion for azithromycin was m/z 749.4, and for ery-thromycin was m/z 734.3. The full-scan mass spectra of proto-nated molecular ions of azithromycin and erythromycin areshown in Fig. 1(A, B). In order to optimize the MS para-meters, a standard solution (1000 ng/mL) of the analyte andIS was infused into the mass spectrometer. For both azithro-mycin and erythromycin, the following optimized parameterswere obtained: temperature of heated capillary 270°C,electrospray voltage 4500 V, capillary voltage 37 V. Nitrogenwas used as sheath gas and auxiliary gas at pressures (arbi-trary units) of 62 and 24, respectively. Scan width for SIMwas m/z 1; and scan time was 0.5 s.

Sample preparation. A 10 µL aliquot of IS solution (1.5 µg/mL erythromycin in methanol) was added to a 100 µL aliquotof plasma sample. The samples were briefly mixed. Then300 µL methanol was added to the mixture, vortex-mixed for3 min and centrifuged for 10 min at 13,400 g. The supernatantwas collected, and a 10 µL aliquot of solution was injectedinto the LC/MS system for analysis.

Calibration standards and quality control samples. A stocksolution of azithromycin was prepared in methanol at theconcentration of 600 µg/mL, and a stock solution of IS wasprepared in methanol at the concentration of 150 µg/mL.Calibration curves for azithromycin were prepared by spikingblank plasma at concentrations of 4.69, 9.38, 18.75, 37.5, 75,150, 300 and 600 ng/mL, and the analyses were performed intriplicate for each concentration. The stock and diluted solu-tions of azithromycin and IS were stored at −20°C.

The quality control (QC) samples were prepared in quintu-plicate using a different stock solution of azithromycin, toobtain plasma concentrations of 9.38, 37.5 and 500 ng/mL,representing low, medium and high concentration QC sam-ples, respectively. The spiked plasma samples (standards andquality controls) were prepared freshly for each analyticalbatch along with the unknown samples.

Method validation. Plasma samples were quantified usingthe ratio of the peak area of azithromycin to that of IS as the

assay parameter. For the calibration standards, peak arearatios were plotted against azithromycin plasma concentra-tions, and linear standard curves were calculated usingweighted (1/x) linear regression.

To evaluate linearity, plasma calibration curves were pre-pared and assayed in triplicate on three separate days. Accu-racy and precision were also assessed by determining QCsamples using quintuplicate (n = 5) preparations of plasmasamples at three concentration levels on three separate days.The accuracy, i.e. percentage concentration deviation, wasexpressed by (mean observed concentration-spiked concen-tration)/(spiked concentration) × 100%, and the precision wasexpressed by relative standard deviation (RSD).

The recoveries of azithromycin at three QC levels wereevaluated by comparing the mean value from analysis of QCsamples with that of samples prepared by spiking extracteddrug-free plasma (spike-after-extraction) samples with thesame amounts of azithromycin at the step immediately priorto chromatography. These spike-after-extraction samplesrepresented 100% recovery. The matrix effect was calculatedby comparing peak areas of azithromycin obtained from thespike-after-extraction samples with those from pure standardsolutions of the homologous concentration.

Azithromycin stability in plasma was assessed by analyzingQC samples at concentrations of 9.38, 37.5 and 500 ng/mL,respectively, in triplicate (n = 3), after exposure to differenttime conditions. The results were compared with those forfreshly prepared QC samples, and the percentage concentra-tion deviation was calculated. The short-term stability wasevaluated after exposure of the plasma samples to room tem-perature (20°C) for 4 and 12 h.

Pharmacokinetic study. The method was used to determinethe concentrations of azithromycin in human plasma fromhealthy volunteers who received oral dosage of azithromycindispersible tablets (containing 250 × 2 mg azithromycin, LivzonPharmaceutical Limited Company, China). Twenty healthymale volunteers were selected for the study. Blood (3 mL)was removed by venepuncture prior to dosage and at 0.5, 1,1.5, 2, 3, 4, 6, 8, 12, 24, 48, 72, 96 and 120 h thereafter. Follow-ing standing for 30 min and centrifugation (4000 g for 10 min),the plasma was removed and stored at −20°C until analysis.

RESULTS AND DISCUSSION

LC/MS conditions

The signal intensities obtained in positive ion modewere much higher than those in negative ion mode. ESIyielded higher signals for m/z 749.4 compared withAPCI. The full-scan mass spectra of protonated mole-cular ions of azithromycin and erythromycin are shownin Fig. 1(A, B).

To achieve maximum peak responses and symmetri-cal chromatographic peaks, mobile phases containingvarying percentages of organic phase and differentcolumns were tested. As a result, methanol, acetonitrileand water (12:30:58, v/v/v) were chosen as the optimizedmobile phase, with 0.1% formic acid and 0.1% ammonium


1274 F. Liu et al.ORIGINAL RESEARCH

Figure 1. Full-scan of [M + H]+ ions of azithromycin (A) and erythromycin (B).



acetate in the water. A Phenomenex Luna CN columnwas chosen.

Sample preparation

Protein precipitation with methanol was chosen for asimple and rapid sample preparation method. Themethod can save much time and simplify the operatingprocess. Different volumes of methanol were evaluatedfor efficiency of protein precipitation; it was found thatthree times the plasma volume can precipitate the plasmaproteins completely, and the chromatographic behaviorof the analytes was not deteriorated by this procedure.

Specificity and matrix effect

The specificity of the method was examined by analyzingblank human plasma [Fig. 2(A)] and plasma spiked withazithromycin (4.69 ng/mL, lower limit of quantitation,LLOQ) and the internal standard [150 ng/mL; Fig. 2(B)].As shown in Fig. 2(A), no significant interferences wereseen from endogenous substances in the SIM channelsfor azithromycin and IS at their corresponding reten-tion times. A fast liquid chromatographic approach wasused to elute azithromycin and erythromycin, givingretention times of 3.90 and 2.80 min. This elution ofazithromycin and erythromycin is well separated fromthe interference of the endogenous materials (retention

Figure 2. SIM chromatograms for azithromycin (I) and erythromycin (IS, II) in humanplasma samples: (A) blank plasma; (B) plasma sample spiked with azithromycin (4.69 ng/mL) and IS (150 ng/mL), LLOQ; (C) a volunteer plasma sample 4 h after an oral dose of500 mg azithromycin.

times of 1.38 and 1.49 min) in human plasma. Fig. 2(C)shows the chromatogram for a volunteer plasma sample4 h after an oral dose of 500 mg azithromycin.

Matrix effects were investigated by analysis of spike-after-extraction samples with pure standard solutions atthe same concentrations. The results revealed no significantdifferences of peak area, and this effect is most likelydue to the sample clean-up with methanol.

Linearity, precision and accuracy

Correlation coefficients >0.999 confirmed that the calibra-tion curves were linear over the concentration range 4.69–600 ng/mL for the analyte. A typical standard curve wasy = 0.011x + 0.002. The LLOQ was confirmed to be4.69 ng/mL, at which the calculated accuracy and preci-sion were below 20%. Both precision and deviationsfrom nominal concentration (accuracy) were less than11.9% for the QC samples. Table 1 shows the intra- andinter-assay precision and accuracy for azithromycin fromQC samples. The intra- and inter-assay relative stand-ard deviations were both measured to be below 8.24%.

Recovery and storage stability

The recoveries of azithromycin were 87.41 ± 2.39, 90.30± 1.10 and 99.97 ± 0.68% at concentrations of 9.38, 37.5and 500 ng/mL, respectively (n = 3). The stability of the



azithromycin prepared sample in the Agilent 1100autosampler (20°C) was determined over 12 h. Theresults are given in Table 2. The sample concentrationsfollowing the stability test period were compared withfreshly prepared QC samples, and reliable stabilitybehavior was observed (within ± 6.8%). The methanolsolution of IS (1500 ng/mL) also proved stable at room

temperature for 4 h and at −20°C for a week. The co-efficient of variation was within 2.0%.

Pharmacokinetic study

The method was applied to the determination the plasmaconcentrations of azithromycin after oral administration

Figure 2. (Continued)



and simple sample preparation procedure, a samplethroughput of 100 per day was routinely achieved. Thissimple and selective method for the determination ofazithromycin in human plasma was readily applicableto the clinical pharmacokinetic study for azithromycin.

CONCLUSION

A simple and sensitive LC/MS method for the quantifi-cation of azithromycin in human plasma was developedand validated. The method satisfied the requirementsof high sensitivity, high selectivity and high throughputfor pharmacokinetic studies.

Acknowledgments

The authors gratefully acknowledge the NSFC (no.20332020) for financial support.

REFERENCES

Bahrami BG, Mirzaeei S and Kiani A. High performance liquid chro-matographic determination of azithromycin in serum using fluore-scence detection and its application in human pharmacokinetic

Table 1. Accuracy and precision for the analysis of azithromycin in human plasma (inprestudy validation, five replicates per day)

Concentration (ng/mL)

Day Spiked Found RSD (%) RE (%)

Day 1 9.38 8.72 ± 0.59 6.72 −7.037.5 35.00 ± 2.88 8.24 −6.7

500 503.70 ± 36.81 7.31 0.7Day 2 9.38 9.38 ± 0.50 5.29 0.0

37.5 38.57 ± 1.23 3.18 2.9500 559.66 ± 10.39 1.86 11.9

Day 3 9.38 9.07 ± 0.31 3.46 −3.337.5 35.95 ± 1.06 2.95 −4.1

500 478.34 ± 24.84 5.19 −4.3Inter-assay 9.38 9.06 ± 0.33 3.66 −3.4

37.5 36.51 ± 1.85 5.07 −2.6500 513.90 ± 41.61 8.10 2.8

RSD, relative standard deviation; RE, relative error.

Table 2. Stability data for assay of azithromycin in human plasma (n ===== 3)

Storage Concentration Concentration foundconditions spiked (ng/mL) (ng/mL) RSD (%) RE (%)

4 h at RT 9.38 9.05 0.16 −3.537.5 35.76 0.37 −4.6

500 502.54 3.49 0.512 h at RT 9.38 9.60 4.01 2.3

37.5 34.94 2.01 −6.8500 486.16 1.15 −2.8

RSD, relative standard deviation; RE, relative error; RT, room temperature.

of 500 mg azithromycin to 20 volunteers. The meanplasma concentration–time curve of azithromycin isshown in Fig. 3. The maximum plasma concentration(Cmax) was 425.71 ± 184.32 ng/mL; the area under thecurve (AUC0−120) was 3880.80 ± 1183.56 ng h/mL; thetime to maximum plasma concentration (Tmax) was 3.13± 2.18 h; and the half-life (t1/2) was 26.32 ± 10.25 h.Because of the relatively short chromatographic run time

Figure 3. Mean plasma concentration–time curve ofazithromycin after an oral dose of 500 mg azithromycin to 20healthy volunteers (n = 20, mean value and SD are plotted).



studies. Journal of Chromatography B Analytical Technologies inthe Biomedical and Life Sciences 2005; 820: 277–281.

Bahrami G and Mohammadi A. A new on-line, in-tube pre-columnderivatization technique for high performance liquid chromato-graphic determination of azithromycin in human serum. Journal ofChromatography B Analytical Technologies in the Biomedical andLife Sciences 2006; 830: 355–358.

Barrett B, Borek-Dohalsky V, Fejt P, Vaingatova S, Huclova J,Nemec B and Jelinek I. Validated HPLC-MS-MS method fordetermination of azithromycin in human plasma. Analytical andBioanalytical Chemistry 2005; 383: 210–217.

Chen BM, Liang YZ, Chen X, Liu SG, Deng FL and Zhou P. Quan-titative determination of azithromycin in human plasma by liquidchromatography–mass spectrometry and its application in a bio-equivalence study. Journal of Pharmaceutical and BiomedicalAnalysis 2006; 42: 480–487.

Fouda HG and Schneider RP. Quantitative-determination of the anti-biotic azithromycin in human serum by high-performance liquid-chromatography (HPLC)–atmospheric pressure chemical-ionizationmass-spectrometry—correlation with a standard HPLC–electro-chemical method. Therapeutic Drug Monitoring 1995; 17: 179–183.

Gandhi R, Kaul CL and Panchagnula R. Validated LC method forin-vitro analysis of azithromycin using electrochemical detection.Journal of Pharmaceutical and Biomedical Analysis 2000; 23: 1073–1078.

Kees F, Spangler S and Wellenhofer M. Determination of macrolidesin biological matrices by high-performance liquid chromatographywith electrochemical detection. Journal of Chromatography A 1998;812: 287–293.

Lalak NJ and Morris DL. Azithromycin clinical pharmacokinetics.Clinical Pharmacokinetics 1993; 25: 370–374.

Nirogi RVS, Kandikere VN, Shukla M, Mudigonda K, MauryaS, Boosi R and Yerramilli A. Sensitive and selective liquidchromatography–tandem mass spectrometry method for the quanti-fication of azithromycin in human plasma. Analytica Chimica Acta2005; 553: 1–8.

Shepard RM, Duthu GS, Ferraina RA and Mullins MA. High-performance liquid-chromatographic assay with electrochemicaldetection for azithromycin in serum and tissues. Journal of Chro-matography Biomedical Applications 1991; 565: 321–337.

Taninaka C, Ohtani H, Hanada E, Kotaki H, Sato H and Iga T.Determination of erythromycin, clarithromycin, roxithromycin,and azithromycin in plasma by high-performance liquid chromato-graphy with amperometric detection. Journal of Chromatography BAnalytical Technologies in the Biomedical and Life Sciences 2000;738: 405–411.

Torano JS and Guchelaar HJ. Quantitative determination of themacrolide antibiotics erythromycin, roxithromycin, azithromycinand clarithromycin in human serum by high-performance liquidchromatography using pre-column derivatization with 9-fluorenylmethyloxycarbonyl chloride and fluorescence detection.Journal of Chromatography B 1998; 720: 89–97.

Wilms E, Trumpie H, Veenendaal W and Touw D. Quantitativedetermination of azithromycin in plasma, blood and isolatedneutrophils by liquid chromatography using pre-column derivatiza-tion with 9-fluorenylmethyloxycarbonyl-chloride and fluorescencedetection. Journal of Chromatography B Analytical Technologies inthe Biomedical and Life Sciences 2005; 814: 37–42.

Documents

Sensitive liquid chromatography/mass spectrometry assay for the quantification of azithromycin in human plasma