Chromatographia

, Volume 71, Issue 5, pp 523–527

LC Determination of Tramadol, M1, M2, M4, and M5 in Plasma

Authors

    • Department of Comparative MedicineUniversity of Tennessee
  • Cheryl Greenacre
    • Department of Small Animal Clinical SciencesCollege of Veterinary Medicine
  • Marcy Souza
    • Department of Comparative MedicineUniversity of Tennessee
  • Sherry Cox
    • Department of Comparative MedicineUniversity of Tennessee
Full Short Communication

DOI: 10.1365/s10337-009-1451-y

Cite this article as:
Yarbrough, J., Greenacre, C., Souza, M. et al. Chroma (2010) 71: 523. doi:10.1365/s10337-009-1451-y

Abstract

A new LC procedure for the determination of tramadol, an analgesic, and its metabolites has been developed and validated. Following a liquid–liquid extraction using ethyl acetate and hexane, samples were separated by RP-LC on a Symmetry C18 column and quantified using fluorescence detection at an excitation of 202 nm and an emission of 296 nm. The mobile phase was a mixture of potassium dihydrogen phosphate buffer (0.01 M), 0.1% triethylamine (pH 2.9) and acetonitrile, with a flow-rate of 1.1 mL min−1. The standard curve ranged from 5 to 5,000 ng mL−1. Intra-and inter-assay variability for all the compounds were less that 10% and the average recovery was greater than 90%. This assay is suitable for use in pharmacokinetic studies.

Keywords

Column liquid chromatographyPharmacokineticsTramadolO-Desmethyltramadol

1 Introduction

Recognition of pain and appropriate analgesic therapy is becoming increasingly important in veterinary medicine. Tramadol is a relatively new analgesic in veterinary medicine; however, it has routinely been used for the relief of moderate to severe pain in humans for the last two decades [1, 2]. Tramadol’s analgesic efficacy is a result of complex interactions between opiate, adrenergic, and serotonin receptor systems [3]. It provides analgesia mainly via serotonin and norepinephrine pathways. Tramadol has a wide margin of safety in humans, with minimal respiratory, cardiovascular or gastrointestinal side effects [3, 4]. It does not inhibit the synthesis of prostaglandins and therefore does not produce the serious adverse events generally associated with non-steroidal antiinflammatory drugs (NSAIDs). The minimal side effects of tramadol make it a potentially useful analgesic in veterinary medicine. Few studies have examined the metabolism and analgesic effects of tramadol in animals, but recommended doses based on pharmacokinetic studies are available for cats and dogs [5, 6].

Tramadol is an effective analgesic for canine abdominal and orthopedic surgery [6], and has been used epidurally in horses and found to provide long-term analgesia with no adverse effects [7]. Its overall analgesic efficacy is comparable to that achieved using equianalgesic doses of parenteral morphine or alfentanil in humans [5]. The high safety profile of this drug may make it an effective and safe analgesic in veterinary patients for moderate to severe pain resulting from surgery or osteoarthritis.

Several LC methods have been developed to measure tramadol and M1 in biological fluids [813]. Some of the methods involve the use of UV [810] or electrochemical detection [11]. The use of a fluorescence detector has been used to improve limit of detection [1216]. Mass spectrometer has been used in the study of tramadol and its metabolites [1719]. Other methods use solid phase [8, 12], liquid–liquid extraction techniques [13], or microsome preparation [14]. There are a few methods that measure more than the M1 metabolite [1517, 20, 21] however to our knowledge no method analyzes tramadol and the four metabolites (M1, M2, M4, and M5) in plasma.

We describe a specific and reproducible method to determine plasma concentrations of tramadol and four of its metabolites (M1, M2, M4, and M5). This method has been used successfully to determine pharmacokinetic parameters at this institution in multiple species [2224].

2 Experimental

2.1 Reagents and Standards

Tramadol and its metabolites were provided by Grünenthal (Aachen, Germany). Butorphanol was purchased from US Pharmacopeia (Rockville, MD, USA). Purity levels for the standards were 99.7, 99.8, 100, 94.1, 94.1, and 100% for tramadol, M1, M2, M4, M5, and butorphanol. All other reagent grade chemicals and solvents were purchased from Fisher Scientific (Pittsburg, PA, USA).

Stock solutions of tramadol, M1, M2, M4, M5, and butorphanol (100 μg mL−1) were each prepared in methanol. Dilutions in methanol were prepared to produce 0.1, 1, and 10 μg mL−1 working stocks. Solutions were aliquoted into vials to prevent evaporation and cross contamination, were stored at 4 °C and were stable for 6 months.

2.2 Chromatographic Instrumentation and Conditions

The system consisted of a 2695 separations module and a 2475 fluorescence detector (Waters, Milford, MA, USA.). Separation was attained on a Waters Symmetry C18 (4.6 × 250 mm, 5 μm) column with a 5 μm Symmetry guard column. The mobile phase was a mixture of A: potassium dihydrogen phosphate buffer with 0.1% triethylamine (0.01 M; pH 2.9), and B: acetonitrile. All solutions were prepared fresh daily filtered through a 0.22 μm filter and degassed before their use. The mixture was pumped at a starting condition of 92% A and 8% B (v v−1), changed to 75% A and 25% B over 40 min, then adjusted to 85% A and 15% B over 8 min and then returned to initial conditions for the final 6 min. The flow rate was 1.1 mL min−1. Fluorescence was measured at an excitation of 202 nm and an emission of 296 nm with the gain at 10X. The column was maintained at an ambient temperature of 25 °C.

2.3 Sample Treatment

Tramadol and its metabolites were extracted from plasma samples using a liquid–liquid extraction. Previously frozen plasma samples were thawed and vortexed. Three hundred fifty microliters of plasma was transferred to a screw top test tube, followed by 100 μL of butorphanol (3.5 μg mL−1 internal standard). Seventy microliters of 29.7% ammonium hydroxide were added, vortexed quickly and then 5 mL of ethyl acetate:hexane (40:60) were added. The tubes were vortexed for 1 min, and then centrifuged for 20 min at 1,700×g. The organic layer was removed to a clean glass tube and evaporated to dryness with nitrogen. Samples were reconstituted in 350 μL of mobile phase. The supernatant was transferred to chromatography vials, and 25 μL were injected.

2.4 Validation

For preparation of calibration standards and quality control samples, appropriate aliquots of stock solutions were placed in tubes, evaporated to dryness with nitrogen, and then reconstituted in 350 μL of blank plasma. The final concentration ranges for tramadol, M1, M2, M4, and M5 were 5−5,000 ng mL−1. Quality control standards were 15, 750, and 3,500 ng mL−1. Linearity was assessed by linear regression analysis, and calibration curves needed a correlation coefficient of 0.99 or better to be accepted. The acceptance criterion for each back-calculated standard concentration was a 15% deviation from the nominal value, except the lower limit of quantification (LLOQ) which was set at 20% [25].

3 Results

3.1 Optimization

Several other extraction solutions were tested to determine which would provide the greatest recovery and best chromatogram resolution. Methylene chloride, ethyl acetate: hexane (1:4) and ethyl acetate alone were tested and found to produce either lower recover values or decreased quality of chromotography.

The original procedure was designed to separate and quantitate tramadol, M1, and butorphanol using an isocratic mobile phase composed of 75% A and 25% B. Using this method, all three standards eluted in less than 10 min with good separation. However, when samples from animals dosed with tramadol were analyzed, compounds were present that interfered with the separation of both tramadol and M1. Since there were no interfering peaks in the blank plasma at the zero time point for the animals, interference was suspected to be caused by other metabolites. Additional metabolite standards (M2, M4, and M5) were obtained from Grunenthal. After analyzing these standards we determined that M4 and M5 interfered with the elution of M1, and M2 interfered with the elution of tramadol. The original protocol was modified to achieve separation of these new metabolites. Following protocol modifications, separation of these metabolites was achieved. These metabolites are similar in structure and to achieve separation, a gradient was developed and the run time lengthened. We have analyzed samples from pharmacokinetic studies in multiple species and the majority of the samples have contained the four different metabolites.

3.2 Validation

Endogenous plasma components did not interfere with the elution of tramadol, its metabolites or the internal standard. Blank plasma samples for specificity testing were prepared in the same manner as study samples. Five different blank plasma samples were used in the validation process and a blank sample from each study subject was included in the analysis. Fig. 1 shows chromatograms of a blank canine plasma sample (a) 1,000 ng mL−1 plasma standard (b) and a canine plasma sample after a 4 mg kg−1 IV dose of tramadol (c). Retention times were 17.9, 19.1, 20.1, 38.8, 39.5, and 43.8 min for M1, M4, M5, tramadol, M2, and butorphanol, respectively. Concentrations for M1, M4, M5, tramadol, and M2 were 116, 42, 176, 663, and 542 ng mL−1, respectively.
https://static-content.springer.com/image/art%3A10.1365%2Fs10337-009-1451-y/MediaObjects/10337_2009_1451_Fig1_HTML.gif
Fig. 1

Chromatograms are representative of (a) blank plasma, (b) blank plasma sample spiked with 1,000 ng mL−1 of standards and 350 ng mL−1 of IS, and (c) canine plasma sample after an intravenous dose of 3 mg kg−1 tramadol. Levels for M1, M4, M5, tramadol, and M2 in the canine plasma sample were 116, 42, 176, 663, and 542 ng mL−1 respectively

Due to the broad range of sample concentrations found in the study, a split curve was generated for tramadol and its metabolites. The plasma peak ratio (area of tramadol and its metabolites divided by the internal standard area) versus concentration was plotted and produced two linear curves, one from 5 to 500 ng mL−1 and one from 500 to 5,000 ng mL−1. These two curves produced correlation coefficients ranging from 0.998 to 0.999. The mean slope, Y-intercept, and r2 values are reported for tramadol and each metabolite in Table 1. Intra-assay relative standard deviation (RSD) for plasma spiked with specific concentrations of tramadol and its metabolites ranged from 1.0 to 9.5% (Table 2). The inter-assay RSD ranged from 1.6 to 9.4% (Table 2).
Table 1

Tramadol and metabolites intra-assay linearity (n = 4)

Metabolite

Curve

Mean ± SD

RSD (%)

M1

Low curve

 Slope

0.00185 ± 0.00019

10

 Y-intercept

0.00587 ± 0.00033

5.7

 r2

0.9993 ± 0.00056

0.1

High curve

 Slope

0.00178 ± 0.00056

9.8

 Y-intercept

0.03597 ± 0.00179

5.0

 r2

0.9992 ± 0.00032

0.03

M2

Low curve

 Slope

0.00297 ± 0.00026

8.7

 Y-intercept

0.03926 ± 0.00147

3.8

 r2

0.9991 ± 0.00040

0.04

High curve

 Slope

0.00326 ± 0.00021

6.57

 Y-intercept

−0.27700 ± 0.01766

6.37

 r2

0.9995 ± 0.00017

0.02

M4

Low curve

 Slope

0.00065 ± 0.00006

9.94

 Y-intercept

0.00357 ± 0.00022

6.03

 r2

0.9993 ± 0.00031

0.03

High curve

  

 Slope

0.00067 ± 0.00004

6.31

 Y-intercept

−0.01849 ± 0.00034

1.83

 r2

0.9997 ± 0.00026

0.03

M5

Low curve

 Slope

0.00108 ± 0.00006

5.35

 Y-intercept

0.00309 ± 0.0014

4.50

 r2

0.9996 ± 0.00055

0.06

High curve

 Slope

0.00114 ± 0.00008

6.76

 Y-intercept

−0.06418 ± 0.00383

5.97

 r2

0.9991 ± 0.00085

0.09

Tramadol

Low curve

 Slope

0.00286 ± 0.00008

2.89

 Y-intercept

0.05054 ± 0.00256

5.06

 r2

0.9992 ± 0.00026

0.03

High curve

 Slope

0.00304 ± 0.00006

1.99

 Y-intercept

0.08501 ± 0.00756

8.89

 r2

0.9982 ± 0.00155

0.16

SD standard deviation; n number of curves; RSD relative standard deviation; r2 correlation coefficient

Table 2

Intra-and inter-assay valves

Concentration (ng mL−1)

M1

M2

M4

M5

Tramadol

Concentration measured* (ng mL−1)

RSD (%)

Concentration measured* (ng mL−1)

RSD (%)

Concentration measured* (ng mL−1)

RSD (%)

Concentration measured* (ng mL−1)

RSD (%)

Concentration measured* (ng mL−1)

RSD (%)

Intra-assay accuracy and precision for tramadol and its metabolites (n = 4)

15

15.7 ± 0.55

3.5

15.0 ± 1.42

9.5

14.5 ± 0.82

5.7

15.0 ± 1.16

7.7

15.1 ± 0.59

3.9

750

773 ± 8.1

1.0

750 ± 54.4

7.3

738 ± 27.0

3.7

745 ± 30.1

4.0

751 ± 48.4

6.5

3,500

3,543 ± 106

3.0

3,460 ± 219

6.3

3,497 ± 235

6.7

3,501 ± 211

6.0

3,431 ± 142

4.1

Inter-assay accuracy and precision for tramadol and its metabolites (n = 4)

15

14.7 ± 1.00

6.8

14.7 ± 0.48

3.3

15.4 ± 1.16

7.5

15.1 ± 1.00

6.6

15.6 ± 0.25

1.6

750

767 ± 24.9

3.3

748 ± 19.9

2.7

752 ± 22.6

3.0

760 ± 20.6

2.7

743 ± 40.4

5.4

3,500

3,494 ± 98

2.8

3,412 ± 129

3.8

3,386 ± 175

5.2

3524 ± 332

9.4

3,482 ± 78

2.2

* concentrations are mean ± standard deviation; n number of days; RSD relative standard deviation

3.3 Limit of Quantification, Recovery, and Stability

The LLOQ for tramadol, M1 and M2 was 5 ng mL−1, while M4 and M5 had an LLOQ level of 10 ng mL−1. The recovery of tramadol and its metabolites from spiked plasma was compared with direct injection of analytes at concentrations of 10, 100, and 2,500 ng mL−1. The average recoveries for tramadol, M1, M2, M4, and M5 were 100, 95, 95, 90, and 94%, respectively. The IS had a recovery of 96%. Testing of auto sampler and short term stability of standards for 24 h indicated that tramadol and its metabolites are stable. For the concentrations of 15, 750 and 3,500 ng mL−1, there was a 5% drug loss after 24 h in the auto sampler and a 3% drug loss after 24 h of short term storage in the refrigerator at 4 °C.

4 Discussion

Our method produced a better recovery and LLOQ than those methods using UV [810] and electrochemical detection [11]. Analgesia in humans has been associated with M1 levels below 40 ng mL−1, therefore the LLOQ should be below this value. The Paar et al. [14] method produced a LLOQ five times higher than this value. The methods of Gan et al. [8] and Kücük et al. [12] require the use of solid phase extraction cartridges. Nobilis et al. [13] uses a t-butylmethyl ether extraction technique that requires at least 6 h to perform. Some of the methods [9, 10, 13] quantitate tramadol alone while a few [8, 11, 12] quantitate tramadol and M1. Paar et al. [14] quantified all four metabolites in microsomal fractions but not in plasma. Their method had a higher limit of detection and a runtime of over 60 min. Giorgi et al. [20, 21] and De Leo et al. [17] used methods that require additional back extraction steps which require more time to perform. Also, chromatograms produced by the De Leo et al. [17] extraction were poorly separated. The use of mass spectrometry produces lower LLOQ values [18, 19] however mass spectrometry equipment is expensive and may not be readily accessible to most laboratories.

Other groups [15, 16] have been involved in the analysis of tramadol and various metabolite combinations specifically in humans. Because our laboratory analyzes samples from multiple species, the method was developed to ensure the separation of tramadol, all four metabolites, and the internal standard regardless of species. Ardakani and Rouini [15] do have a slightly lower limit of quantification; however, our levels are more than adequate for pharmacokinetic drug studies. Ardakani and Ruoini [15] used a 100 μL injection volume, while our injection volume is only 25 μL and could be increased if a lower limit of quantification is necessary.

Tramadol and its metabolites were quantified using a liquid–liquid extraction technique. The use of butorphanol as the internal standard allows for the correction of inter- and intra-assay variability in the extraction. The stability studies indicate that samples were stable for 24 h after extraction. Therefore, if an equipment malfunction occurs samples could be reanalyzed. Samples in our studies were thawed one time and analyzed; therefore, freeze–thaw studies were not conducted.

5 Conclusion

This analytical procedure was validated in terms of selectivity, recovery, linearity, LLOQ, precision, and accuracy. The limit of quantification and recovery are more than adequate for use in pharmacokinetic studies. In conclusion our results indicate that this LC procedure represents a highly specific and reproducible method that provides consistent quantification of tramadol and its metabolites in the plasma of multiple species. The method was found to be suitable for the generation of plasma concentration–time curves (Fig. 2).
https://static-content.springer.com/image/art%3A10.1365%2Fs10337-009-1451-y/MediaObjects/10337_2009_1451_Fig2_HTML.gif
Fig. 2

Plasma concentration–time profile for tramadol following intravenous administration of (○) 4 mg kg−1 to an eagle and (□) 2 mg kg−1 to a llama

Copyright information

© Vieweg+Teubner | GWV Fachverlage GmbH 2009