Enhancement of Fluorescence Intensity of Tramadol and Its Main Metabolites in LC Using Pre-Column Derivatization with 9-Fluorenylmethyl Chloroformate

2008, 68, 935940

Gholamreza Bahrami1,2,&, Bahareh Mohammadi1

1 2

Medical Biology Research Center, Kermanshah University of Medical Sciences, Kermanshah, Iran; E-Mail: [email protected] School of Pharmacy, Kermanshah University of Medical Sciences, Kermanshah, Iran

Received: 19 April 2008 / Accepted: 29 July 2008 Online publication: 23 September 2008

Introduction

Abstract
Tramadol was found to exhibit weak uorescence with a maximum emission at 300 nm when excited at 200 nm. Also, uorescence spectra of the drug and its two main metabolites, O-desmethyltramadol and N-desmethyltramadol are not practically identical. Thus low and different sensitivities have been reported for the drug and its metabolites in previously published work. In the present method using 9-uorenylmethyl chloroformate (FMOC-Cl) as labeling agent, equal and magnied uorescence intensity were obtained for the analytes. The drug, its metabolites and an internal standard (oseltamivir phosphate) were extracted from serum by dichloromethane. Pre-column derivatization of the analytes was achieved using FMOC-Cl in the presence of borate buffer (0.1 M, pH 7.5). Liquid chromatography with a mobile phase consisting of a mixture of 0.05 M phosphate buffer containing triethylamine (2 ml L-1; pH = 3.0) and methanol (54:46; v/v) and a Shimpack CLC-ODS column were used for analytical separation of the analytes. The uorescence of the column efuent was monitored at an excitation and emission wavelengths of 265 and 315 nm, respectively. The analytical method was linear over the concentration range of 1.01,280 ng mL-1 of the parent drug and its metabolites and limit of quantication of 1.0 ng mL-1 was obtained for the analytes using 10 lL injection. The method validation was studied and the validated method applied in a bioequivalence study of 2 different tramadol preparations in 24 healthy volunteers.

Keywords
Column liquid chromatography Pre-column derivatization Tramadol 9-Fluorenylmethyl chloroformate

Tramadol ()-trans-2-[(dimethylamino) methyl]-1-(3-methoxyphenyl) cyclo-hexanol (Fig. 1a), a weak l-opioid receptor agonist whose mechanism of action is predominantly based on enhanced serotonergic neurotransmission, is used in the treatment of mild to moderate pain. Tramadol is about 68% bioavailable after a single oral dose and its therapeutic plasma level is within the range of 100300 ng mL-1 [1]. The drug is rapidly and extensively metabolized in the liver to O-desmethyltramadol (ODT; Fig.1b) and N-desmethyl tramadol (NDT; Fig. 1c). ODT is more potent than the parent drug and may account for part of the analgesic eect [1]. Several analytical methods including gas chromatography coupled with nitrogen phosphorus (GCNPD) [2], ame ionization (GCFID) [3], or mass spectrometry (GCMS) [4, 5] detection as well as liquid chromatography (LC) equipped with electrochemical (EC) [6], UV [711], diode array (DAD) [12] or uorescence detections [1316] have been published for the analysis of tramadol in human serum. Limits of quantication (LOQ) of 10 [8], 12.5 [9], 38 [7] and 50 ng mL-1 [12] for the

Original DOI: 10.1365/s10337-008-0806-0 0009-5893/08/12

Chromatographia 2008, 68, December (No. 11/12) 2008 Vieweg+Teubner | GWV Fachverlage GmbH

935

H3C N

CH3

H3C N

CH3

dierent tramadol preparations in 24 healthy volunteers.

H O H3C
Tramadol

H O OH H OH

Experimental
Chemicals
Tramadol and ODT (purity 99.1%) were from Dipharma (Milano, Italy). NDT (purity 99.0%) was from SynFine Research (Richmond Hill, Ontario Canada). Oseltamivir phosphate (I.S.; Fig 1d; purity 99.5%) was from Hetero Labs (Hyderabad, India). All other chemicals were of analytical grade (except methanol which was LC grade) and were purchased from Merck (Darmstadt, Germany). Water was glass-double distilled and further puried for LC with a Maxima purication system (USF ELGA, UK).

O-Desmethyltramadol (ODT)

c
H3C N H

CH3

CH3 O

O O CH3 H N OH O NH2 CH3 Oseltamivir Phosphate (I.S.)

H O H3C

N-Desmethyltramadol (NDT)

Fig. 1. Chemical structures of a tramadol, b O-desmethyl tramadol, c N-desmethyl tramadol, and d oseltamivir phosphate

Instrumentation and Chromatographic Conditions

The LC system used consisted of a two pumps of Shimadzu LC-10A solvent delivery system, a system controller (SCL 10AD), a spectrourometric detector (RF-551) operated at an excitation and emission wavelengths of 265 and 315 nm, respectively, a column oven (CTO-10A), a degasser (DGU-3A) and a data processor (C-R4A) all from Shimadzu, Kyoto, Japan. The analytical column was a Shim-pack CLC-ODS (Shimadzu), 150 9 4.6 mm ID., 5-lm particle size which was protected by a Shim-pack G-ODS guard column (1 cm 9 4.0 mm ID, 5-lm particle size). A mixture of 0.05 M phosphate buer containing triethylamine (2 ml L-1; pH = 3.0) and methanol (54:46; v/v) was used as mobile phase The column oven temperature was set at 58 C and the mobile phase was ltered, degassed and pumped at a ow rate of 2.0 mL min-1 with a back pressure of 150 kg cm-2.

parent drug have been reported using UV detection and injection volumes of 50, 100, 200 and 500 lL, respectively. The drug contains a weakly absorbing chromophore in its molecule, which makes analysis of low tramadol levels by UV detection dicult. Fluorescence detection is more sensitive hence, LOQ of 10 ng mL-1 [14], 3 ng mL-1 [15] and 2.5 ng mL-1 [16] have been reported by uorescence detection using an injection volume of 100 lL. Dierent sensitivities (15 ng mL-1 using LCEC [6], 400 ng mL-1 using LCDAD [12], 1 ng mL-1 [4] and 40 ng mL-1 [5] using GCMS, 2 ng mL-1 using GCNPD [2] and 12.5 ng mL-1 [3] using GCFID) have been reported for analysis of the drug in biological uids in other published methods. However, only in ve of them, ODT metabolite [2, 6, 1416] and in one technique [16] both ODT and NDT have been detected. Long extraction times (20 min [16], 15 min[9], 60 min[5], 15 min [13] and 30 min [14, 15] have been reported in some previously published methods and various

analytical run times ranging from 5 min using a monolithic column [16] to 25 min [9] have been described. Ion pair chromatography [15], back-wash approach [15] or two steps extraction [16] have been used in other published LC methods with uorescence detection. In the present study using 9-uorenylmethyl chloroformate (FMOC-Cl) as pre-column derivatization agent, a new technique is described for simultaneous analysis of tramadol and its two main metabolites in human serum. In our procedure uorescence signal intensity of the analytes is improved and unlike the previously published methods, the same LOQ is obtained for the parent drug and its metabolites. Furthermore, with the exception of the method described by Rouini et al. [16] in which a short run time had been reported using an expensive monolithic column, the analytical run time has been improved. The present technique was applied for determination of the drug and its metabolites in a bioequivalence study following single oral administration of 2

Preparation of Solutions
Stock solutions of tramadol, ODT and NDT (500 lg mL-1) were prepared sep-

936

Chromatographia 2008, 68, December (No. 11/12)

Original

arately by dissolving the analytes in acetonitrile, while stock solution of the I.S. (200 lg mL-1) was prepared in methanol. Dierent working solutions of tramadol and its metabolite ranging from 0.01 to 12.8 lg mL-1were obtained by further dilutions of the stock solutions in acetonitrile. The I.S. stock solution was diluted with methanol to obtain a working standard of 20 lg mL-1. A 400lg mL-1 solution of FMOC-Cl was prepared in acetonitrile. A borate buer (0.05 M) was prepared in water and adjusted to pH 7.5 with 0.05 M potassium hydroxide solution. A stock solution of glycine (20 mg mL-1) was prepared in water. All solutions were stored at 4 C and were stable for at least 3 weeks.

reconstituted in 1 mL of drug-free human serum, mixed for 10 s on a vortex mixer and subjected to extraction, derivatization and analysis as described above. Linearity was checked by preparing standard solutions at nine dierent concentrations ranging from 1.0 to 1,280 ng mL-1. The linearity was also checked for six consecutive days for the solutions of the same concentrations prepared from the stock solutions. Calibration curves (weighted regression line) were obtained by linear least-squares regression analysis of plots of peak-area ratio to I.S. versus drug concentrations.
Accuracy, Precision and Sensitivity

Stability of the drugs in serum samples was evaluated by comparing the peak area ratio values of the analytes following maintenance of the samples (n = 6) up to 30 days at -40 C, 48 h at 28 C and following three thaw-freeze cycles. Stability of solutions of the analytes and the I.S. were studied over a period of four weeks by comparing the peak areas at dierent times.
Application of the Method

Extraction Procedure and Derivatization

Aliquots of blank, calibration standard or unknown human serum samples (1 mL) were pipetted into 100 9 16 mm disposable glass tubes, containing 100 lL of the I.S. To the samples 3 mL dichloromethane were added and after mixing for 30 s on a vortex mixer and centrifugation (5 min at 6,000g), the organic phase was removed and evaporated to dryness at 45 C. The residue was reconstituted in 100 lL of the FMOC-Cl solution and following addition of 25 lL borate buer (0.05 M; pH 7.5) and brief mixing, the samples were kept at 60 C for 10 min. The excess of the labeling agent was reacted with glycine (10 lL; 20 mg mL-1) and after 1 min; a volume of 10 lL of the mixture was injected onto the LC system.

Quality control samples used in method validation were prepared with the drug working solutions to make low (1.0 ng mL-1), medium (50 ng mL-1) and high (500 ng mL-1) concentrations for each compound. Intra-and inter-day variations were calculated for each analyte by repeated analysis (n = 6) of different concentrations in a single analytical run and in ten analytical runs performed on dierent days, respectively. The limits of detection (LOD) and quantication (LOQ) were dened as the concentration of the drug giving a signal-to-noise ratio of 3:1 and 10:1, respectively.
Specicity, Selectivity, Recovery and Stability

Validation of the Method

Calibration Curve and Linearity

Calibration curve samples were prepared within the concentration range of 1.0 1,280 ng mL-1 using pooled human blank serum obtained from normal subjects. In disposable glass tubes (100 9 16 mm), 100 lL from each working solutions of the analytes was evaporated under a gentle stream of nitrogen at 50 C. The residues were Original

The specicity of the method was examined by the presence of disturbing endogenous peaks in 24 human serum samples from dierent volunteers. These samples were pretreated according to the sample preparation procedure except from the addition of the I.S. Selectivity of the assay was examined by analysis of several potentially co-administrated drugs with tramadol. The absolute recoveries of tramadol, ODT and NDT at the above mentioned concentrations as well as the I.S. at applied concentration were calculated in replicates (n = 5) by comparing the respective peak areas obtained by derivatization of the extracted samples from serum, with those obtained after derivatization of the same amounts of un-extracted solutions in acetonitrile.

The present method was applied in a randomized crossover bioequivalence study of two dierent tramadol preparations in 24 male healthy volunteers aged 28.3 3.9 years and weighing 74.5 6.7 kg with normal biochemical parameters. Written informed consent was obtained from the subjects and the study protocol was approved by the Ethics Committee of Kermanshah University of Medical Sciences. After an overnight fasting all the volunteers received a single oral dose of 100 mg tramadol from either Bakhtar Bioshimi (Kermanshah, Iran) or Ortho-McNeil (Ultram, NJ, USA) pharmaceutical companies on two working days separated by a wash-out period of 2 weeks. They were asked to refrain from food or water consumption for 3 h after drug administration. Blood sampling were carried out at suitable intervals up to 24 h and the samples were stored at -40 C until analysis. Pharmacokinetic parameters including maximum concentration (Cmax), area under the concentration time curve from zero to time of last sampling(AUC0 - t) and area under the concentration time curve from zero to innity (AUC0 - ?) were compared using a paired students-test. Bioequivalence between the preparations was determined by calculating 90% condence intervals for the ratio of Cmax, (AUC0 - t) and (AUC0 - ?) values for dierent products, using logarithmic transformed data.

Results and Discussion

Derivatization
The dierent experimental parameters aecting the intensity of the uorescence

Chromatographia 2008, 68, December (No. 11/12)

937

Fig. 2. Typical chromatograms obtained from an extract of a human blank serum spiked with the I.S., b human blank serum spiked with oseltamivir phosphate as the I.S. and 3 ng mL-1 from each O-desmethyl tramadol (ODT), tramadol and N-desmethyl tramadol (NDT), c human blank serum spiked with ODT (120 ng mL-1), tramadol (120 ng mL-1) the I.S. and NDT (75 ng mL-1) d and e serum samples from a volunteer 3 and 12 h after a single oral dose of 100 mg tramadol containing ODT (12.5 and 11 ng mL-1), tramadol (50 and 23 ng mL-1) and NDT (3 and 1.0 ng mL-1)

signal were studied and optimized using the univariate method to obtain maximum sensitivity. Tramadol, its metabolites and the I.S. react with FMOC-Cl in alkaline medium and the derivatization is complete within 10 min at 60 C. The reaction appeared to be highly dependent on pH of buer solution, time as well as temperature of incubation, concentration of the labeling agent and polarity of the medium. The optimal conditions were found to be: FMOC-Cl solution with a concentration of 400 lg mL-1. borate buer solution with a pH of 7.5, reaction medium consisting of wateracetonitrile (1:4 v/v) and a reaction temperature of 60 C for 10 min.

Chromatography
Typical chromatograms of human blank serum spiked with the I.S. (Fig. 2a), human blank serum spiked with tramadol, ODT and NDT (3.0 ng mL-1; Fig.2b) and human blank serum spiked with ODT (120 ng mL-1), tramadol (120 ng mL-1) and NDT (75 ng mL-1) (Fig. 2c) have been shown. ODT, tramadol, the I.S. and NDT were well resolved with retention times of 4.5, 6.1, 8 and 10.1 min. Figures 3d and e show the chromatograms of serum samples obtained at 3 and 12 h after a single oral dose of 100 mg tramadol from a healthy volunteer. Endogenous components and excess of the reagent were

eluted within 2.5 min. The results of the selectivity study showed that there were no interfering peaks from any of the following drugs: acetaminophen, codeine, naproxen, diclofenac, mefenamic acid, ketorolac, ibuprofen, indomethacin, aspirin, sodium salicylate, caeine, phenobarbital, diazepam, carbamazepine, lamotrigine, topiramate, vigabatrin, propranolol, etidronate, and gentamicin.

Sensitivity and Linearity

The LOD and LOQ for the analytes were estimated to be 250 pg mL-1 and 1.0 ng mL-1, respectively, using 10 lL

938

Chromatographia 2008, 68, December (No. 11/12)

Original

injection. The calibration curves for tramadol, ODT and NDT were linear over the concentration ranges of 1.0 1,280 ng mL-1. The correlation coecients (r2) for calibration curves of tramadol, ODT and NDT were equal to or better than 0.9988, 0.9950 and 0.9975, respectively. The equations for means (n = 6) of calibration curves for each analyte were: tramadol; y = 0.2946x 1.919, ODT; y = 0.3002x - 1.6423 and NDT; y = 0.3017x - 1.8524 and CV percent values (slope, intercept) for tramadol, ODT and NDT were (3.8, 6.7), (4.6, 7.6) and (4.2, 6.9), respectively.

Table 1. Intra-and inter-day precision and accuracy for determination of tramadol in human serum by the LC method
Concentration (ng mL-1) (n = 6) Tramadol 1 50 500 ODT 1 50 500 NDT 1 50 500 Intra-day CV% Accuracy% Inter-day CV% Accuracy%a

9.0 4.7 1.6 11.2 5.3 2.05 10.6 3.7 1.8

98.4 97 102.1 97.2 96.3 104.5 93.5 95.5 106

11.9 5.4 1.4 11.7 5.7 2.4 11.7 4.6 1.6

98.3 102.8 97.4 96.6 95.5 103.2 94.4 97.2 103

Stability, Recovery, Accuracy and Precision

Stock solutions of tramadol, its metabolite and the I.S. were stable for 30 days when stored at 4 C and the derivatized solutions were found to be stable (>95%) for 24 h. The concentrations of tramadol in serum stored at -40 C for 30 days or 28 C for 48 h were found to be 101 and 97% from the initial values, respectively. The mean recoveries of tramadol, ODT, NDT and the I.S. from serum were 101 3, 98 4, 98 3 and 89 4%, respectively. The intra-and inter-day accuracy and precision values of the assay method are presented in Table 1. The coecient variation values of both intraand inter-day for all analytes were all less than 12% whereas accuracy never deviated from 100% by more than 6.5%. FMOC-Cl reacts with primary and secondary amines in alkaline conditions as well as with hydroxyl groups. In previously published LC methods, dimethylsulfanilamide [14], sotolol [15] or cis-tramadol [16] had been used as internal standard, however, uorescence spectra of tramadol, their metabolites and applied I.S. in these methods were not practically identical [17]. Thus, different sensitivities for the drug and its metabolites have been reported [1416]. In the procedure described by Nobilis et al. [14] a programmable uorescence detector had been used and excitation and emission wavelengths were changed for detection of the I.S. The present method, however, provides the same sensitivity for the drug and its metaboOriginal

Accuracy has been calculated as a mean percent of the nominal values a Percent of mean deviation from nominal values

320
Tramadol Test

Conc. (ng mL )

240

Tramadol Ref. ODT Test ODT Ref.

-1

160

NDT Test NDT Ref.

80

0 0 5 10 15 20 25

Time (h)
Fig. 3. Mean serum concentrations versus time proles of tramadol and its metabolites for 2 dierent preparations in 24 human volunteers after administration of a single 100 mg oral dose

lites. Tramadol has only a tertiary hydroxyl group, whereas, there are two nucleolic groups in its metabolites (tertiary hydroxyl and secondary amino groups). Although theoretically both of these groups are attackable by FMOCCl not more than one peak was eluted following derivatization of each metabolite. Thus it seems a single uorenyl derivative is formed for each metabolite. A mobile phase with at least 2 mL L-1 of triethylamine was necessary to separate the peaks of the drug from endogenous substances and excess of the reagent which were eluted at rst parts of the chromatogram. Retention behavior of the drug and its metabolites was pH-dependent and their retention times were found to increase proportionally

with the pH of the mobile phase while the retention time of I.S. was positively dependent on the triethylamine proportion of the mobile phase. Thus a mobile phase with a pH of 3.0 and triethylamine of 2 mL L-1 were selected.

Application of the Method

This method has successfully been used for the determination of serum concentrations of the drug and its metabolites in a randomized cross-over bioequivalence study following single oral administration of two dierent tramadol preparations in 24 healthy volunteers. Typical serum concentrationtime proles of the drug and its metabolites have

Chromatographia 2008, 68, December (No. 11/12)

939

Table 2. Mean (SD) pharmacokinetic parameters of tramadol, ODT and NDT for two dierent preparations in 24 human volunteers after administration of a single 100 mg oral dose
Test Tmax (h) Cmax (ng mL-1) AUC0-24 (ngh mL-1) AUC0-? (ngh mL-1) T1/2 (h) Reference Tmax (h) Cmax (ng mL-1) AUC024 (ngh mL-1) AUC0-? (ngh mL-1) T1/2 (h) Tramadol 2.3 (1.0)a 362.3 (95) 3,108 (966) 3,820 (1254) 8.8 (2.5)_ 2.2 (0.9) 353.9 (94) 3,060 (1065) 3,748 (1472) 8.2 (2.4) ODT 2.6 (1.2) 96.3 (38) 986 (335) 1,174 (413) 9.94 (2.9) 2.5 (1.5) 90.6 (32) 962 (402) 1,130 (419) 10.2 (3.6) NDT 3.3 (1.7) 77.4 (35) 885 (315) 1,045 (454) 9.2 (3.8) 3.0 (1.3) 71.5 (37) 844 (352) 996 (536) 9.0 (4.5)

Tmax = Time to maximum concentration, Cmax = Maximum concentration, AUC = Area under the concentration time curve, T1/2 = Elimination half life a Values in the parenthesis are SD of the amounts in 24 healthy volunteers

been shown in Fig. 3 and resulted pharmacokinetic parameters have been summarized in Table 2.

organic solvent (e.g., acetonitrile) and the sensitivity is further improved.

Acknowledgments Conclusions
The present paper is the rst report for simultaneous quantication of tramadol and its metabolites using a one step extraction method and FMOC-Cl as uorescent labeling agent. In our method LOQ of 1 ng mL-1 using 10 lL, injection and 1-mL serum had been obtained for the parent drug and its metabolites. The stability of the derivative also allowed FMOC-Cl derivatives to be extracted using ethyl acetate, after dilution of the reaction mixture with water. In this case the residue can be reconstituted in a smaller volume of an This work was supported by Bakhtar Bioshimi Pharmaceutical Company and partially by Kermanshah University of Medical Sciences.

References
1. Lee CR, McTavis D, Sorkin E (1993) Drugs 46:313340. doi:10.2165/00003495199346020-00008 2. Tao Q, Stone DJ Jr, Borenstein MR, Jean-Bart V, Codd EE, Coogan TP, Desai-Krieger D, Liao S, Raa RB (2001) J Chromatogr B 763:165171. doi:10.1016/ S0378-4347(01)00388-7

3. Ho S-T, Wang J-J, Liaw W-J, Ho Ch-M, Li J-H (1999) J Chromatogr B 736:8996. doi:10.1016/S0378-4347(99)00434-X 4. Sha YF, Shenb S, Duan GL (2005) J Pharm Biomed Anal 37:143147. doi: 10.1016/j.jpba.2004.09.050 5. Gambaro V, Benvenuti C, Ferrari LD, Acqua LD, Fare F (2003) II Farmaco 58:947950. doi:10.1016/S0014-827X(03) 00153-8 6. Valle M, Pavon JM, Calvo R, Campanero MA, Troconiz IF (1999) J Chromatogr B 724:8389. doi:10.1016/S0378-4347 (98)00547-7 7. Kmetec V, Roskar R (2003) J Pharm Biomed Anal 32:10611066. doi:10.1016/ S0731-7085(03)00209-7 8. Gan SH, Ismail R, Wan Adnan WA, Wan Z (2002) J Chromatogr B 772:123 129. doi:10.1016/S1570-0232(02)00065-X 9. Yeh GC, Sheu MT, Yen CL, Wang YW, Liu CH, Ho HO (1999) J Chromatogr B 723:247253 10. Elsing B, Blaschke G (1999) J Chromatogr 612 (2) (1993) 22330 11. Overbeck P, BlaschkeJ G (1999) Chromatogr B Biomed Sci Appl 732(1):185 192 12. Gan SH, Ismail R (2001) J Chromatogr B 759:325335. doi:10.1016/S0378-4347(01) 00237-7 13. Kucuka A, Kadoglub Y, Celebi F (2005) J Chromatogr B 816:203208. doi: 10.1016/j.jchromb.2004.11.031 14. Nobilis M, Kopecky J, Kvetina J, Chladek J, Svoboda Z, Vorsek V, Perlk F, Pour M, Kunes J (2002) J Chromatogr A 949:1122. doi:10.1016/S0021-9673(01) 01567-9 15. Gu Y, Fawcett JP (2005) J Chromatogr B 821:240243. doi:10.1016/j.jchromb.2005. 05.003 16. Rouini MR, Hosseinzadeh Ardakani Y, Soltani F, Aboul-Enein HY, Foroumad A (2006) J Chromatogr B 830:207211. doi:10.1016/j.jchromb.2005.10.039 17. Soetebeer UB, Schierenberg M-O, Moller J-G, Schulz H, Grunefeld G, Andresen P, Blaschke G, Ahr G (2000) J Chromatogr A 895:147155. doi:10.1016/S0021-9673(00) 00704-4

940

Chromatographia 2008, 68, December (No. 11/12)

Original