Anal. Chem. 1996, 68, 118-123
A Partially Automated Pretreatment Module for Routine Analyses for Seventeen Non-Steroid Antiinflammatory Drugs in Race Horses Using Gas Chromatography/Mass Spectrometry Soledad Ca´rdenas, Mercedes Gallego, and Miguel Valca´rcel*
Department of Analytical Chemistry, Faculty of Sciences, University of Co´ rdoba, E-14004 Co´ rdoba, Spain Rosa Ventura and Jordi Segura
Department of Pharmacology and Toxicology, IMIM, Barcelona, Spain
A partially automated module for the routine determination of illicit non-steroid antiinflammatory drugs (NSAIDs) in biological fluids from race horses was built, tested, refined, and shown to work. This pretreatment module retains 17 NSAIDs on an Amberlite XAD-2 column before back-elution derivatization with methyl iodide in acetonitrile. Methylated derivatives are manually injected into a gas chromatograph connected to a mass spectrometer. The quantification limits thus achieved are 50-100 ng/ mL in 1 mL of urine or plasma. The proposed method is more expeditious than its manual liquid-liquid and liquid-solid extraction counterparts and is similar in speed to a typical gas chromatographic program. Non-steroid antiinflammatory drugs (NSAIDs) have long been used to treat pain, swelling, and injuries in both human and animal species. In addition to the classical members of this drug family such as aspirin and phenylbutazone, a huge number of new drugs have been introduced by pharmaceutical companies. Although their chemical structure is variable (salicylates, phenylalkyl-, heteroaryl-, and indolealkyl acids, oxicams, pyrazoles, etc.), nearly all of them share an acidic character (pKa e 5.0) and bind tightly to plasma proteins.1 The use of NSAIDs in competing horses (equestrian and racecourse events) has been forbidden by several regulatory authorities to protect the animals from health problems if competing under undue conditions. In some instances, allowances exist for low concentrations of a few selected NSAIDs in biological fluids, although the general regulatory trend is toward zero tolerance. Given the fact that the number of tests for horses is very large worldwide, as is the number of different NSAIDs that could potentially be detected in a single test, the need for rapid, sensitive, and specific analytical methods for analysis is obvious. Urine is the usual fluid of choice for detection of NSAIDs because it can be collected noninvasively and the drugs or their metabolites are present at relatively high concentrations.2 In any case, plasma samples are also readily obtained. (1) Tobin, T. In Drugs and the Performance Horse; Charles C. Thomas Publisher: Springfield, IL, 1981; Chapter 6. (2) Moss, M. S. In Clarke’s Isolation and Identification of Drugs; Moffat, A. C., Ed.; The Pharmaceutical Press: London, 1986; Chapter on Drug Abuse in Sport.
118 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
There are several methods for the identification of NSAIDs in equine biological fluids. The identification of flunixin and its metabolite in urine3 and tolfenamic acid and its metabolite in plasma4 is carried out by using gas chromatography/mass spectrometry (GC/MS); the methods for both substances involve several extractions and a cleanup step by TLC prior to identification. NSAIDs have been identified by GC/MS/MS via liquidsolid extraction and derivatization with CH3I-K2CO3 in acetone.5 Gonza´lez et al.6 recently reported a GC/MS procedure for the detection of NSAIDs in plasma and urine using a manual liquidliquid extraction with ethyl ether and the above derivatization method. Phenylbutazone7 alone and in conjunction with oxyphenbutazone8 was determined in plasma using liquid-solid extraction or direct sample injection in high-performance liquid chromatography (HPLC). NSAIDs are frequently determined in human plasma and urine using GC-FID,9 GC/MS,10 HPLC,10-12 and supercritical fluid chromatography.13 The aim of this work was to develop a simple, rapid, specific, and sensitive method for the determination of illicit NSAIDs in horses by using a mass spectrometer detector. Flow-based methods have high potential since, in contrast to batch methods, containment and washout is inherent, and miniaturization is not hindered by the difficulty of reproducibly dispensing microliter volumes. Various pretreatment flow systems have been coupled (GC-FID and GC-ECD) for automatic conditioning (preconcentration, dilution, derivatization, and solvent changeover) of samples (3) Jaussaud, P.; Courtot, D.; Guyot, J. L.; Paris, J. J. Chromatogr., Biomed. Appl. 1987, 67, 123-130. (4) Jaussaud, P.; Guieu, D.; Courtot, D.; Barbier, B.; Bonnaire, Y. J. Chromatogr., Biomed. Appl. 1992, 111, 136-140. (5) De Jong, E. G.; Kiffers, J.; Maes R. A. A. J. Pharm. Biomed. Anal. 1989, 7, 1617-1622. (6) Gonza´lez, G.; Ventura, R.; Smith, A.; de la Torre, R.; Segura, J. J. Chromatogr., in press. (7) Gupta, R. N. J. Chromatogr., Biomed. Appl. 1990, 530, 160-163. (8) Santasania, C. T. J. Liq. Chromatogr. 1990, 13, 2605-2631. (9) Giachetti, C.; Zanolo, G.; Poletti, P.; Perovanni, F. J. High Resolut. Chromatogr. 1990, 13, 789-792. (10) Singh, A. K.; Jang, Y.; Mishra, U.; Granley, K. J. Chromatogr., Biomed. Appl. 1991, 106, 351-361. (11) Kazemifard, A. G.; Moore, D. E. J. Chromatogr., Biomed. Appl. 1990, 98, 125-132. (12) Caturla, M. C.; Cusido, E. J. Chromatogr., Biomed. Appl. 1992, 119, 101107. (13) Simmons, B. R.; Jagota, N. K.; Stewart, J. T. J. Pharm. Biomed. Anal. 1995, 13, 59-64. 0003-2700/96/0368-0118$12.00/0
© 1995 American Chemical Society
with good results.14,15 Recently, an automated GC/MS method for routine field atmospheric CFC replacement was reported.16 Adaption of a commercially available GC/MS instrument required a microtrap adsorption/desorption system with hardware and software changes in the basic mass spectrometer; the instrument design was highly complex as a result. In this work, a simple flow system was developed for liquid-solid extraction of NSAIDs from horse urine and plasma and methylation of NSAIDs prior to their off-line determination by GC/MS. This off-line method eliminates water, as derivatives from the flow system are collected in vials containing anhydrous sodium sulfate; 1 µL of this dry solution is manually injected into the GC. The module avoids human handling and allows sample processing at rates similar to that of the GC program used. EXPERIMENTAL SECTION Chemicals, Solvents, and Biological Samples. The drug compounds were supplied by the following pharmaceutical manufacturers: flurbiprofen and ibuprofen by Laboratorios Liade S.A. (Madrid, Spain); ibuproxam by Laboratorios Novag S.A. (Barcelona, Spain); indomethacin by Laboratorios Uriach (Barcelona, Spain); phenylbutazone by Laboratorios Miquel (Barcelona, Spain); and others (diclofenac, flufenamic acid, flunixin, ketoprofen, meclofenamic acid, mefenamic acid, naproxen, niflumic acid, oxyphenbutazone, propyphenazone, suxibuzone, tolfenamic acid, and zomepirac) by Sigma (Madrid, Spain). Sodium aluminosilicate pellets were purchased from Sigma, nonionic Amberlite XAD-2 was obtained from Serva Co. (Seville, Spain), and all other reagents (citric acid, methyl iodide, acetic anhydride, K2CO3, sodium sulfate, ammonium hydroxide, ammonium chloride, methanol, n-hexane, acetonitrile, and ethyl ether) were obtained from Merck (Madrid, Spain). All other chemicals and solvents were of analytical grade or better. Horse plasma and urine samples (blanks and positives) were supplied by the Department of Pharmacology and Toxicology of the Instituto Municipal de Investigaciones Me´dicas (IMIM, Barcelona, Spain). Preparation of Standards and Solutions. Stock standard solutions of each NSAID were prepared in methanol (NSAIDs are readily soluble and stable in this solvent) at a concentration of 100 µg/mL and were stored frozen at -20 °C. Stock solutions (10 µg/mL) for optimizing GC conditions were also prepared in methanol. Spiked samples were made by adding appropriate volumes of the stock solutions to urine or plasma blanks to obtain NSAID concentrations of 0.1-2 µg/mL. Standard solutions for construction of the calibration graphs were prepared by adding a few microliters of stock standard solution of each NSAID in methanol to 1 mL of 0.2 mol/L ammonia buffer at pH 9.5. A solution containing 25% v/v methyl iodide and 20% v/v acetic anhydride in acetonitrile was used as eluent/ derivatizing reagent. A 0.2 mol/L NH4OH/NH4Cl buffer solution at pH 9.5 (ammonia buffer) was used to condition the resin and prepare samples. A 2 mol/L citrate buffer at pH 2 was also prepared. (14) Ballesteros, E.; Gallego, M.; Valca´rcel, M. Anal. Chem. 1990, 62, 15871591. (15) Valca´rcel, M.; Ballesteros, E.; Gallego, M. Trends Anal. Chem. 1994, 13, 68-73. (16) Simmonds, P. G.; O’Doherty, S.; Nickless, G.; Sturrock, G. A.; Swaby, R.; Knight, P.; Ricketts, J.; Woffendin, G.; Smith, R. Anal. Chem. 1995, 67, 717-723.
Sample Preparation. No treatment of urine samples of pH 5.5-10 was required, as volumes of 1 mL were directly introduced into the flow system. Whenever sample dilution was needed, 0.2 mol/L ammonia buffer at pH 9.5 was used. Plasma samples were extracted prior to introduction into the flow system. A volume of 1 mL of plasma was pipetted into a glass tube. The plasma was adjusted to pH 2 with 2 mL of 2 mol/L citrate buffer at pH 2 and extracted with 2 mL of ethyl ether. The vial was mechanically shaken for 1 min and centrifuged at 7500g for 10 min. The upper phase was evaporated to dryness under a nitrogen stream, the residue being dissolved in 1 mL of 0.2 mol/L ammonia buffer at pH 9.5 and introduced into the flow system. Apparatus. Methylated NSAID derivatives were determined on a Fisons 8000 GC interfaced to a Fisons MD 800 mass spectrometer, both of which were controlled by a computer equipped with LAB-BASE software (Fisons). Separation of the derivatives was achieved on a 30 m × 0.25 mm i.d., film thickness 0.25 µm, DB-5 fused silica capillary column from J&W Scientific, Cromlab (Barcelona, Spain). Helium was used as the carrier gas at a flow rate of 0.8 mL/min. The GC injection port and GC/MS interface regions were maintained at 280 °C and 300 °C, respectively, and the ion source temperature was 200 °C. The column oven temperature was programmed from 120 to 200 °C (3 min), rising at 15 °C/min, and then to 300 °C (3 min) at 10 °C/min. The mass spectrometer was operated in the scanning mode (range 50-500 m/z) and tuned with the PFTBA to optimize selected ions of m/z 69, 219, 264, and 502. The electron impact ionization potential was 70 eV, and the emission current was 500 µA. Sample injection was done in the split mode (split ratio 1:20). A Gilson Minipuls-2 peristaltic pump fitted with poly(vinyl chloride) and Solvaflex pumping tubes for aqueous and organic solutions, respectively, was employed. Two Rheodyne 5041 injection valves, PTFE tubing of 0.5 mm i.d. and standard connectors for coils and loops, a displacement bottle, and two columns were also used. The resin column was prepared by packing a piece of PTFE tubing (40 mm × 1.5 mm i.d.) with 80 mg of nonionic Amberlite XAD-2; small glass wool plugs were used on the ends to prevent losses of material. The resin was initially conditioned as follows: first, a methanol stream was passed at a flow rate of 0.4 mL/min through the column for 2 min, followed by water for 2 min, and finally, a 0.2 mol/L ammonia buffer solution at pH 9.5 for 3 min. The K2CO3 column was made by packing a glass commercial column of 50 mm × 4 mm i.d. (Omnifit, Cambridge, England) with 150 mg of solid K2CO3 between glass wool plugs. The column packing was critical inasmuch as every two K2CO3 segments (∼0.5 cm long) must be separated by one of sodium aluminosilicate (pellets of 1/16 in., nominal pore diameter 4 Å) to avoid abrupt changes in column compactness (viz., by water traces) that may stop the flow and disconnect the tubing. The useful lifetime of the K2CO3 column was 1 day (∼30 uses). Procedure. The flow system used for the determination of NSAIDs (Figure 1) was operated in two steps for preconcentration and elution-derivatization of the analytes. In the preconcentration step, 1 mL of standard in 0.2 mol/L ammonia buffer at pH 9.5 (0.05-3 µg/mL NSAID) or 1 mL of urine (pH 5.5-10) or pretreated plasma was continuously introduced into the system at 0.4 mL/min and propelled through the Amberlite XAD-2 column located inside the loop of injection valve IV1. A volume of 1 mL of 0.2 mol/L ammonia buffer solution, pH 9.5, was employed as Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
119
Figure 1. Schematic diagram of the experimental setup used for the determination of NSAIDs in horse plasma and urine. P, peristaltic pump; IV, injection valve; W, waste; R, eluent-derivatizing reagent; GC/MS, off-line transfer for GC/MS analysis.
carrier to complete sample introduction. Retention of NSAIDs was instantaneous, and the matrix was sent to waste (W1); simultaneously, the loop of the second injection valve (IV2) was filled with eluent-derivatizing reagent containing 20% acetic anhydride (to condition the eluate) and 25% methyl iodide (derivatizing reagent) in acetonitrile (eluent). In the elutionderivatization step, both injection valves were simultaneous and manually switched, so 200 µL of the eluent-derivatizing reagent was injected into an n-hexane stream and passed through the resin column at 0.4 mL/min to elute retained NSAIDs; then, the eluted fraction was passed through the K2CO3 column, where the flow was halted (1.2 min after injection of the eluent-derivatizing reagent) for 5 min to allow derivatization. Finally, the flow was resumed, and the column effluent was collected for 1 min (∼400 µL) in a glass vial containing 50 mg of anhydrous sodium sulfate. An aliquot (1 µL) of this solution was injected into the gas chromatograph. The resin must then be conditioned with 1.2 mL of 0.2 mol/L ammonia buffer (pH 9.5) for subsequent analyses after each run. RESULTS AND DISCUSSION Existing methods developed for the determination of NSAIDs in biological fluids involve the separation, by solvent extraction or sorption, of the analytes from the urine or plasma sample after derivatization to methyl2,5,17 or trimethylsilyl ethers18 prior to GC analysis. Because the methylated derivatives of NSAIDs are more stable than their trimethylsilylated derivatives, methylation is the more frequent choice. Methyl iodide is widely used for methylation of NSAIDs in the presence of solid K2CO3, which acts as both a catalyst for the derivatization and a provider of the alkaline medium needed for the reaction. Drastic reaction conditions of heating at 100 °C for 30 min are also required.2,6 Conventional methods for the determination of NSAIDs in horse fluids use liquid-liquid or liquid-solid extraction (with Amberlite XAD-2) to separate the analytes from the matrix.2 Initially, the liquidliquid extraction method was selected for automation. Sample pH was adjusted to between 2 and 3 by inserting a stream of 2 mol/L citric acid at pH 2. However, the extraction step posed many problems: experiments with various extractants (n-hexane, ethyl ether, chloroform, and petroleum ether) showed them to be incompatible with methyl iodide and K2CO3. Since the solvents that allowed the methylation (viz., acetone and acetonitrile) were (17) Giachetti, C.; Assandri, A.; Zanolo, G.; Tenconi, A. Chromatographia 1994, 39, 162-169. (18) Tanimura, Y.; Saitoh, Y.; Nakagawa, F.; Suzuki, T. Chem. Pharm. Bull. 1975, 23, 651-658.
120
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
miscible with the aqueous phase and dissolved the solid K2CO3 column, we chose liquid-solid extraction. Selection of Solvents and Reagents. This study was made by using a standard solution of 2 µg/mL phenylbutazone and oxyphenbutazone, its metabolite, in 0.2 mol/L ammonia buffer, pH 9.5. Both NSAIDs exhibited different behavior in the manual procedure using CH3I-K2CO3 as the derivatizing reagent; although phenylbutazone was readily derivatized, oxyphenbutazone required heating for a long time. Two Amberlite XAD-2 and K2CO3 columns were prepared and inserted into a flow manifold such as that depicted in Figure 1. In the manual method,2 NSAIDs were retained on the resin column at pH 9.5, and then the solid phase was conditioned with 10% v/v acetic acid in water for 5 min prior to elution of the acid fraction. To simplify this step and reduce the amount of water present in the system, acetic anhydride was added to the eluent to act as both drying agent and conditioner for the resin at an appropiate pH. Three solvents were found to be compatible with the elution and derivatization. Since derivatization required heating above 50 °C in ethyl acetate and acetone, acetonitrile was selected because it allowed for derivatization at room temperature. Under these conditions, the peak area for oxifenbutazone (m/z 338) was 2 and 10 times greater than that obtained with acetone and ethyl acetate, respectively. Acetyl chloride, dimethyl carbonate, and methyl iodide were assayed as derivatizing reagents. Since phenylbutazone and oxyphenbutazone derivatives were detected only when methyl iodide was used, it was further tested. Since methylation required dry catalytic K2CO3, sodium aluminosilicate pellets were placed at intervals in the column. The eluent-derivatizing reagent was incompatible with pump tubes and displacement bottles because it is miscible with water. An injection valve was therefore included in the flow system to inject the eluent-derivatizing reagent. Study of Variables. This study was carried out by using a manifold similar to that shown in Figure 1. Urine blanks (pH 7.2) were spiked with 2 µg/mL of phenylbutazone and oxyphenbutazone. Portions of 1 mL were used to optimize chemical, physical, and flow variables. The influence of the acetic anhydride concentration was studied over the range 0-40% in acetonitrile (also containing 25% of methyl iodide); chromatographic signals increased with increasing concentration of acetic anhydride up to 10% and 15% for phenylbutazone and oxyphenbutazone, respectively, and remained
Table 1. Figures of Merit of the Calibration Graphs for the Determination of NSAIDs
a
NSAID
m/z
regression equationa
linear range (µg/mL)
quantification limit (ng/mL)
RSDb (%)
ibuprofen flufenamic acid flurbiprofen propyphenazone niflumic acid ibuproxam naproxen flunixin mefenamic acid ketoprofen tolfenamic acid diclofenac meclofenamic acid phenylbutazone oxyphenbutazone indomethacin suxibuzone
161 263 258 215 295 188 185 295 255 268 275 214 242 322 338 371 452
A ) 0.08X + 0.01 A ) 1.80X - 0.10 A ) 0.34X + 0.02 A ) 23.81X - 1.85 A ) 0.89X + 0.04 A ) 1.74X - 0.16 A ) 0.50X + 0.02 A ) 0.56X + 0.03 A ) 1.29X - 0.04 A ) 0.09X + 0.01 A ) 0.78X - 0.01 A ) 1.21X - 0.10 A ) 1.17X + 0.03 A ) 2.03X - 0.30 A ) 0.88X - 0.17 A ) 0.31X + 0.02 A ) 0.80X - 0.10
0.3-3 0.1-3 0.2-3 0.05-3 0.1-3 0.1-3 0.1-3 0.1-3 0.1-3 0.3-3 0.1-3 0.1-3 0.1-3 0.1-3 0.2-3 0.2-3 0.2-3
150 50 150
