1046 JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 DRUGS, COSMETICS, FORENSIC SCIENCES Multiresidue Chromtogrphic Method for the Determintion of Mcrolide Residues in Muscle by High-Performnce Liquid Chromtogrphy with UV Detection JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 MURIELLE JUHEL-GAUGAIN nd BÉATRICE ANGER Community Reference Lbortory for Veterinry Drugs Residues, CNEVA-Fougères, BP203, 35302 Fougères, Frnce MICHEL LAURENTIE Veterinry Phrmcokinetic Drug Unit, CNEVA-Fougères, BP203, 35302 Fougères, Frnce A high-performnce liquid chromtogrphic (HPLC) method for the simultneous determintion of tilmicosin, tylosin, spirmycin, nd its mjor metbolite neospirmycin ws developed tht is suitble for porcine, bovine, nd poultry muscles. Mcrolide residues were extrcted from muscle with cetonitrile, ft ws removed by liquid liquid extrction with isooctne, nd the extrct ws then clened on Bond Elut C 18 crtridges. The HPLC seprtion ws performed on n Inertsil ODS3 C 18 column (150 4 mm) with 0.05% trifluorocetic cid cetonitrile in grdient mode. Two different chromtogrphic grdients were used for tilmicosin tylosin nd spirmycin neospirmycin, nd the detection wvelengths were 287 nd 232 nm, respectively. The method ws vlidted from ½ the mximum residue limit (MRL) to 4 times the MRL with pork muscle smples. Men recoveries were 60, 63.5, 51, nd 42% for tilmicosin, tylosin, spirmycin, nd neospirmycin, respectively. The detection limits re 15 µg/kg for tilmicosin nd tylosin, 30 µg/kg for spirmycin, nd 25 µg/kg for neospirmycin. Linerity, precision, nd ccurcy of the method were lso tested. Mcrolides belong to the fmily of mcrocyclic ntibiotics. They form homogeneous clss of ntibiotics in terms of their chemicl structure (mcrocyclic lctone nucleus) nd ntibcteril spectrum. The 4 mcrolides we hve studied re 16-member-ring mcrolides. These drugs re often used s feed dditives for growth promotion, but they cn lso be used for therpeutic purposes becuse they hve wide rnge ntibcteril spectrum. Their use cn leve residues in edible products tht not only cn hve direct toxic effects but lso my led to llergic rections in consumers nd to the development of resistnt bcteri. Mesures to monitor certin number of residues of phrmcologicl substnces in Received November 4, 1998. Accepted by JM My 10, 1999. frm nimls nd in the fresh met obtined from such nimls re described in Europen Community (EC) Council Directive 96/23/EEC, nd the criteri for routine methods to be used for this purpose re described in EC Commission Decision 93/256/EEC. Mximum residue limits (MRL) in muscle re 50 µg/kg for tilmicosin, 100 µg/kg for tylosin, nd 200 nd 300 µg/kg for the sum of spirmycin nd neospirmycin in bovine nd porcine muscle, respectively. The bibliogrphy in the field of mcrolides residues is very poor. Chromtogrphic methods hve lredy been described for the determintion of mcrolides residues by liquid chromtogrphy with mss spectrometry (LC/MS), gs chromtogrphy/ms (GC/MS), or LC with UV detection (LC UV), but only few were multiresidue methods (1 3). The LC UV method described by Chn (3) included only tylosin nd tilmicosin residues. Some monoresidue methods hve been developed for the determintion of tylosin (4 6) nd spirmycin (7, 8). UV detection ws chosen to develop multiresidue method, but the UV bsorption of spirmycin nd neospirmycin t 232 nm involved n exhustive clenup of the smples to compenste the nonspecificity of this wvelength. Another lterntive ws to develop method with fluorimetric detection by using derivtiztion of the ldehyde function, but tilmicosin would hve been then excluded. Therefore, the purpose of this work ws to develop simple nd relible multiresidue method for the simultneous determintion of tilmicosin, tylosin, spirmycin, nd neospirmycin in muscle by using high-performnce liquid chromtogrphy (HPLC) with UV detection. Experimentl Regents () Solvents. Regent grde methnol, cetonitrile, nd isooctne (Merck, Nogent-Sur-Mrne, Frnce). (b) Wter. Deionized. (c) Ammonium cette nd dipotssium hydrogen phosphte. Regent grde (Merck). (d) 5% Dimethyldichlorosylne in toluene. Supelco (Sint Quentin Fllvier, Frnce). (e) Trifluorocetic cid. UV grde (Merck).
JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 1047 Figure 1. Rection scheme of the formtion of the fluorescent derivtive mong ldehyde group, mmoni, nd cyclohex-1,3-dione. (f) Elution solution. 0.1M methnolic mmonium cette solution. (g) 0.2M buffer solution. Prepred by dissolving 34.84 g dipotssium hydrogen phosphte in 1 L deionized wter. (h) LC mobile phse. 0.05% trifluorocetic cid in wter cetonitrile in grdient mode t flow rte of 0.7 ml/min. Stndrds () Reference stndrds. Tilmicosin nd tylosin were obtined from Eli Lilly nd Co. (Sint Cloud, Frnce), spirmycin from Rhône Merieux (Toulouse, Frnce), nd neospirmycin from Rhône Poulenc Rorer (Vitry sur Seine, Frnce). Figure 2. Typicl liquid chromtogrm exmple of pork muscle fortified with spirmycin, neospirmycin, tilmicosin, tylosin, nd josmycin injected into grdient mode. mau, milli-bsorbnce units.
1048 JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 b b Figure 3. Typicl liquid chromtogrm exmples of muscle smples spiked t the MRL level () nd corresponding stndrds (b) injected with 2 different elution grdients. The top set of chromtogrms re neospirmycin nd spirmycin, nd the bottom set of chromtogrms re tilmicosine nd tylosin.
JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 1049 Tble 1. Recovery results Compound 1/2 MRL b MRL 2 MRL 4 MRL Tilmicosin 61.5 59.9 58.2 44.1 Tylosin 64.6 64.2 63.0 52.2 Neospirmycin 43.0 39.5 41.5 57.2 Spirmycin 51.8 47.4 52.4 61.0 Vlues re given in %. b MRL, mximum residue limit. (b) Individul stock solutions. Four methnolic solutions contining 1 mg/ml of ech stndrd were prepred. For tilmicosin, solution contining 1 mg/ml cis-tilmicosin ws prepred. Solutions were stored below 15 C. (c) Intermedite stndrd solution. A methnolic intermedite solution ws prepred by diluting 0.5 ml tilmicosin, 1 ml tylosin, 2 ml spirmycin, nd 2 ml neospirmycin in 20 ml with methnol. The concentrtions obtined were 25, 50, 100, 100, nd 100 µg/ml, respectively. This solution ws stored below 15 C. (d) Clibrtion stndrds. One milliliter of the intermedite stndrd solution ws trnsferred to flsks of 25, 50, 100, nd 200 ml nd diluted with dipotssium hydrogen phosphte buffer to obtin 4 mixed stndrd solutions of (1) 0.125 µg/ml tilmicosin, 0.25 µg/ml tylosin, nd 0.5 µg/ml spirmycin nd neospirmycin; (2) 0.25 µg/ml tilmicosin, 0.5 µg/ml tylosin, nd 1 µg/ml spirmycin nd neospirmycin; (3) 0.5 µg/ml tilmicosin, 1 µg/ml tylosin, nd 2 µg/ml spirmycin nd neospirmycin; (4) 1 µg/ml tilmicosin, 2 µg/ml tylosin, nd 4 µg/ml spirmycin nd neospirmycin. These solutions were stored in freezer (c 4 C). (e) Spiking solution. Four milliliters of intermedite solution ws diluted in 200 ml deionized wter to obtin solution contining 0.5 µg/mlcis-tilmicosin, 1 µg/ml tylosin, nd 2 µg/ml spirmycin nd neospirmycin. (f) Spiked control smples. Spiked muscle smples were prepred by dding 500 µl spiking solution to 5 g control muscle (free of mcrolides) to obtin spiking levels equl to the MRL for ech mcrolide. Apprtus () Solid-phse extrction (SPE) columns. Bond Elut C 18 (500 mg, 3 or 6 ml; Vrin, Les Ulis, Frnce), 3 ml crtridges were used for porcine nd bovine muscle smples, nd 6 ml crtridges were used for poultry muscle smples. (b) SPE mnifold. Supelco. (c) Filters. Millex HV13 filters (0.45 µm, 13 mm id) (Millipore, Sint Quentin Yvelines, Frnce). (d) Anlyticl column. ODS3 column (150 4mmid, contining C 18 Inertsil, 5 µm prticle size; Interchim, Montluçon, Frnce) nd Lichrospher 100RP18-e gurd column (4 4 mm). (e) Liquid chromtogrph. Series 1050 quternry grdient pump, Series 1050 utosmpler, Series 1050 UV-Vis detector, nd HPLC 2D Chemsttion softwre (Hewlett-Pckrd, Les Ulis, Frnce). Preprtion of Smple Five grms of grinded muscle ws weighed into polypropylene centrifuge tube. Ten milliliters cetonitrile ws dded, the mixture ws homogenized, nd then stirred for 10 min t 100 rpm with the rotry stirrer. Ten milliliters isooctne ws dded nd stirred gin for 5 min t slow speed of c 30 rpm (to void emulsions) with the rotry stirrer. Centrifugtion ws performed for 10 min t 4000 g. The upper lyer (isooctne) ws removed, nd 8 ml of the lower lyer ws pipetted nd diluted with 50 ml deionized wter. SPE Column Clenup A 50 ml reservoir ws connected, nd the Bond Elut crtridge ws ctivted with 1 ml dimethyldichlorosylne followed by 10 ml methnol nd 10 ml deionized wter. The smple extrct ws loded nd pulled through the crtridge t flow rte of no more thn 2 drops/s. The crtridge ws not llowed to dry t this step. The crtridge ws flushed with 10 ml deionized wter nd then dried for 2 min under vcuum. Elution ws performed successively with 1 nd 0.5 ml 0.1M methnolic mmonium solution into previously weighed polypropylene tube. One-hlf milliliter dipotssium hydrogen phosphte buffer ws dded nd the solution ws evported to volume below or equl to 0.5 ml. Phosphte buffer ws dded until the totl content weight is 1 g. The sm- Tble 2. Repetbility nd within-lbortory reproducibility coefficients of vrition Tilmicosin Tylosin Neospirmycin Spirmycin Content Repetbility Withinlbortory reproducibility Repetbility Withinlbortory reproducibility Repetbility Withinlbortory reproducibility Repetbility Withinlbortory reproducibility 1/2 MRL 16.4 16.4 6.3 10.5 9.9 11.9 9.7 14.5 MRL 12.1 15.9 4.8 9.3 9.4 15.7 7.6 15.1 2 MRL 5.9 7.7 3.8 10.3 7.4 10.5 5.5 10.4 4 MRL 8.6 12.9 3.0 10.3 6.6 15.5 6.7 13.9 Vlues re given in %.
1050 JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 Tble 3. Method ccurcy for the determintion of mcrolide residues in muscle Mcrolide Content Accurcy, % Tilmicosin 1/2 MRL 3.822 MRL 1.223 2 MRL 1.6357 4 MRL 3.4097 Tylosin 1/2 MRL 2.185 MRL 1.68269 2 MRL 0.325 4 MRL 3.40288 Neospirmycin 1/2 MRL 2.396 MRL 5.98 2 MRL 1.134 4 MRL 4.98 Spirmycin 1/2 MRL 1.744 MRL 6.897 2 MRL 2.77 4 MRL 2.479 ples were mixed nd filtered through 0.45 µm filters before 100 µl volume ws injected into the LC system. Chromtogrphic Determintion One hundred microliters of ech stndrd, spiked smple extrct, nd test smple extrct were injected. Tilmicosin nd tylosin re detected t 287 nm. Spirmycin nd neospirmycin re detected t 232 nm. () Elution grdient for tilmicosin nd tylosin. (1) 0 min, cetonitrile 0.05% trifluorocetic cid (29 + 71); (2) 11 min, cetonitrile 0.05% trifluorocetic cid (45.5 + 54.5); (3) 11.5 min, cetonitrile 0.05% trifluorocetic cid (50 + 50); (4) 14 min, cetonitrile 0.05% trifluorocetic cid (50 + 50); (5) posttime, 6 min; (6) flow rte, 0.7 ml/min. (b) Elution grdient for spirmycin nd neospirmycin. (1) 0 min, cetonitrile 0.05% trifluorocetic cid (20 + 80); (2) 12 min, cetonitrile 0.05% trifluorocetic cid (27.5 + 72.5); (3) 13 min, cetonitrile 0.05% trifluorocetic cid (50 + 50); (4) 16 min, cetonitrile 0.05% trifluorocetic cid (50 + 50); (5) posttime, 5 min; (6) flow rte, 0.7 ml/min. Results nd Discussion Development During development, we first tried derivtiztion rection with cyclohex-1,3-dione (CHD). CHD is specific fluorogenic regent of the ldehyde function tht ws lredy used for the derivtiztion of josmycin by Leroy et l. (9). In the presence of mmoni, CHD rects with the ldehyde group to give heterocyclic structure with fluorescence properties (Figure 1). Therefore, this regent ws used with spirmycin, neospirmycin, josmycin, nd tylosin but not with tilmicosin, which does not contin ny ldehyde group. The results were stisfctory only with josmycin. The rection with tylosin nd spirmycin gve so mny peks tht it ws lmost impossible to identify ech frgment. It ws then decided to continue the development with UV detection method, which llowed the inclusion of tilmicosin. Becuse the mobile phse is composed of trifluorocetic cid nd cetonitrile, the clibrtion stndrds were first diluted in trifluorocetic cid solution. However, problems of stbility were observed with spirmycin. This compound gve neospirmycin by cidic hydrolysis, nd therefore clibrtion stndrd solution ws then prepred with dipotssium hydrogen phosphte (ph 9.3). A steeply eluting grdient ws first determined to nlyze the 5 compounds, but this grdient involved drifting bseline, prticulrly for the extrcts (Figure 2), nd it ws sometimes difficult to hve precise integrtion of the peks. Two different grdients were then used: one for tilmicosin, tylosin, nd josmycin (UV detection t 287 nd 232 nm) nd nother for neospirmycin nd spirmycin (UV detection t 232 nm). After vlidtion of this method for ll 5 compounds, josmycin Tble 4. Homogeneity study of compounds used in the stbility study Compound Source of vrition Sum of squres df Men squre F Spirmycin Between smples 3055.40 4.00 763.85 2.01 Anlyticl 1903.00 5.00 380.60 Totl 4958.40 9.00 1144.45 Tilmicosin Between smples 22061.40 4.00 5515.35 0.87 Anlyticl 31795.00 5.00 6359.00 Totl 53856.40 9.00 11874.35 Tylosin Between smples 672697.48 4.00 168174.37 1.84 Anlyticl 456278.17 5.00 91255.63 Totl 1128975.65 9.00 259430.00 Criticl vlue of F =5,19(p = 0.05, df1 = 4, df2 = 5); df, degrees of freedom. There re no significnt differences between smples.
JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 1051 ws excluded becuse of poor results in terms of repetbility nd reproducibility. A specific method for josmycin will need to be used. Assy Vlidtion Chromtogrms obtined with this method re presented in Figure 3. The method ws vlidted in the rnge from 25 to 200 µg/kg for tilmicosin, 50 to 400 µg/kg for tylosin, nd 150 to 1200 µg/kg for spirmycin nd neospirmycin. The vlidtion ws performed with pork muscle smples, nd it ws lso verified tht this method cn be used with bovine nd poultry muscle. Becuse the poultry muscle smples sometimes becme plugged during SPE, Bond Elut C 18 (500 mg, 6 ml) cn be used. It ws verified tht the recoveries were not ltered by using 6 ml insted of 3 ml crtridges. The limit of detection (LOD) ws clculted s the pprent content corresponding to the vlue of the men plus 3 times the stndrd devition for the blnk determintions (20 representtive blnk smples). b c Figure 4. Stbility study of spirmycin (), tilmicosin (b), nd tylosin (c) incurred muscle smples stored t 20 C.
1052 JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 Tble 5. ANOVA tble derived from the stbility studies of spirmycin, tilmicosin, nd tylosin Source of vrition Sum of squres df Men squre F Spirmycin Liner regression 16967.0 1.0 16967.0 Residul 84262.1 22.0 3830.1 4.4 Tilmicosin b Liner regression 282139.0 1.0 282139.0 Residul 117338.0 22.0 5333.5 52.9 Tylosin c Liner regression 152696.0 1.0 152696.0 Residul 750300.0 16.0 46893.8 3.3 b c Criticl vlue of F = 4.30 (p = 0.05, df 1=1,df2=22). Criticl vlue of F = 4.30 (p = 0.05, df 1=1,df2=22). Criticl vlue of F = 4.49 (p = 0.05, df 1=1,df2=16). These LODs re 15 µg/kg for tilmicosin nd tylosin, 30 µg/kg for spirmycin, nd 25 µg/kg for neospirmycin. Twelve smples per concentrtion were tested during the vlidtion. Liner clibrtion curves were obtined from ½ MRL to 4 MRL. During extrction, n 8 ml liquot of the 10 ml superntnt ws tken, nd correction fctor of 10/8 ws pplied to the rw dt before the estimted concentrtions were clculted. Recoveries verged 60% for tilmicosin, 63.5% for tylosin, 51% for spirmycin, nd 42.5% for neospirmycin. Tble 1 summrizes the recoveries of ech drug for different levels of spiking. Student s t-test proved tht recoveries re independent of the concentrtions. These recoveries were low compred with monoresidue methods, prticulrly for neospirmycin. However, it is cler tht it ws not possible to obtin high recoveries with method tht must be suitble to severl substnces residues. Furthermore, the 232 nm UV detection required n exhustive clening of the smples, which involved decrese of the recoveries. Precision nd ccurcy of the method were tested nd gve results in ccordnce with EC Commission Decision 93/256/EEC (Tbles 2 nd 3). Suitbility of the Method to Incurred Smples It ws importnt to study incurred smples s well s spiked smples to get informtion bout the metbolism of mcrolides. The incurred smples cn then be used for stbility studies. Four pigs were ordered, nd 3 of them were treted with mcrolides. The fourth one ws slughtered without ny tretment to obtin blnk smples for comprison. Spirmycin nd tylosin were injected by intrmusculr route, nd tilmicosin ws dded to the feed. The pigs were then slughtered t different times, nd the biceps femoris nd glutel muscle were then minced, smpled, nd stored t bout 20 C. The smples were then nlyzed with the previously described method. No prticulr metbolites were detected for spirmycin nd tilmicosin. The quntity of neospirmycin observed ws miniml compred with tht of spirmycin. A smll second pek ws observed with tylosin but ws not present in ll smples. Further studies will be performed by LC/MS to scertin more informtion concerning this compound. A homogeneity study ws then mde before the stbility study ws strted. The homogeneity ws determined for ech mteril ccording to the Interntionl Hrmonized Protocol for Proficiency Testing of (Chemicl) Anlyticl Lbortories (10) by compring the smpling vrince with the nlyticl vrince. The results re given in Tble 4. There ws no significnt differences between smples, nd therefore the mteril ws considered sufficiently homogeneous for the purpose of stbility study. The results of the stbility study re given in Figures 4 c. A sttisticl test ws performed for ech nlyte to verify whether the slope of the liner regression line ws significntly different from 0. Results re given in Tble 5. No significnt slope ws detected with the results concerning tylosin. For spirmycin nd tylosin, slopes of the liner regression lines were significntly different from 0, indicting tht the compounds my be not stble when stored t 20 C. The stbility study will be continued for t lest 1 yer. Conclusion This method llows simultneous determintion of tilmicosin, tylosin, spirmycin, nd neospirmycin between ½ nd 4 MRL. It ws vlidted ccording to EC Commission Decision 93/256/EEC. The suitbility to incurred smples ws verified, nd this method ws determined to be very relible.
JUHEL-GAUGAIN ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 82, NO. 5, 1999 1053 Acknowledgment We thnk Jcques Poirier from Rhône-Poulenc Rorer (Vitry sur Seine, Frnce) for the supply of neospirmycin. References (1) Delépine, B., Hurtud-Pessel, D., & Snders, P. (1996) J. AOAC Int. 79, 397 404 (2) Tktsuki, K., Ushizw, I., & Shoji, T. (1987) J. Chromtogr. 391, 207 217 (3) Chn, W., Gerhrdt, G.C., & Slisbury, C.D.C. (1994) J. AOAC Int. 77, 331 333 (4) Keng, L.J.-Y., & Boison, J.O. (1992) J. Liq. Chromtogr. 15, 2025 2034 (5) Mots, W.A. (1985) J. Assoc. Off. Anl. Chem. 68, 980 984 (6) Delépine, B., Hurtud, D., & Snders, P. (1994) Anlyst 119, 2717 2721 (7) Ngt, T., & Seki, M. (1986) J. Assoc. Off. Anl. Chem. 69, 644 646 (8) Snders, P., & Delépine, B. (1994) Biol. Mss Spectrom. 23, 369 375 (9) Leroy, P., Decolin, D., & Nicols, A. (1994) Anlyst 119, 2743 2747 (10) Thompson, M., & Wood, R. (1993) J. AOAC Int. 76, 926 940