Journal of Analytical Toxicology, Vol. 16, May/June 1992
Detection of Diuretics in Horse Urine by GC/MS Heinz-Werner
Hagedorn
and R ( i d i g e r S c h u l z
Institute of Pharmaco/ogy, Toxicology, and Pharmacy, University of Munich, K6niginstraBe 16, D-8000 MOnchen 22, Germany
Abstract I The use of diuretics in horses subject to doping control is prohibited. Thus, a sensitive screening procedure is required to identify the chemically different diuretics. We communicate here a method to detect three commonly employed acidic diuretics: bumetanide, ethacrynic acid, and furosemide. A liquid-liquid extraction on Extrelut 3 ~"~was performed at weak acidic and basic conditions using ethyl acetate as organic solvent. For analysis by GC, the diuretics were methylated oncolumn in the presence of MSTFA/'FMAH, avoiding the commonly employed highly toxic derivatizing agent methyl iodide. For identification of diuretics, we used a mass selective detector operating in the SIM (selected ion monitoring) mode. Confirmation analysis may be obtained with a full scan run. Recoveries for the individual drugs ranged from 31 to 48% at the 100-ng/mL level for 3 mL urine, using calibration curves of drug standards with linearity from 2.5 to 20 ng injected. The limit of detection amounts to 40 ng/mL for the three diuretics. The method permits rapid and sensitive detection of diuretics in horse urine and is recommended for doping control
pK a values, their extraction from urine may cause problems. It is thus a common practice to extract diuretics from urine under acidic and basic conditions (4,12), by adsorption on resin (5), and by Cl8 solid phase material (8). The widely used methyl iodide method (13) of derivatization for GC analysis is time consuming, and because of its high toxicity, should be avoided. These drawbacks of current screening techniques for diuretics encouraged us to develop a more rapid and less toxic method for the detection of commonly used acidic diuretics in horses such as bumetanide, ethacrynic acid, and furosemide. The procedure involves a sequential solvent extraction of urine under acidic and basic conditions on granular Kieselgur, and, subsequently, a onepot derivatization of the diuretics. This method takes advantage of the highly reactive trimethylsilyl groups introduced by MSTFA (14), which are then substituted by methyl groups donated by TMAH. This technique appeared advantageous over current dcrivatization procedures because it is less laborious, very rapid, and sufficiently sensitive to detect low levels of the diuretics.
Experimental Introduction
Diuretics, a class of chemically very heterogeneous compounds, promote the excretion of water and salts. Because of the larger urine volumes produced, the concentration of drugs eliminated by the kidney may be reduced ( 1). In fact, the urine levels of phenylbutazone, a widely used nonsteroidal antiphlogistic dmg in horses, is decreased 40-fold in the presence of furosemide (2), and TLC screening for flunixin has given false-negative results for up to 2 h after application using a diuretic as "masking agent" (3). As expected from these results, diuretics are misused in sports to impede the detection of certain doping agents. For this reason, diuretics have been banned by a number of horse racing authorities. The literature provides several screening procedures for the detection of diuretics, including HPLC (4), TLC (5,6), GC/MS (7,8), and HPLC/MS (9) techniques. With respect to horse urine, sensitive methods have been described for furosemide (10) and bumetanide (11). However, because of the different functional groups (e.g. carboxyl, amino, aminosulfonyl) and their varying
1 94
Reagents and materials. Analytical grade ethyl acetate, methanol, and Extrelut 3 | columns were purchased from E. Merck (Darmstadt, Germany). N-trimethylaniline hydroxide (TMAH, 0.2M in methanol, ready to use) was supplied by Macherey-Nagel (Dtiren, Germany). N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA), ethacrynic acid, bumetanide, and furosemide were purchased from Sigma (Deisenhofen, Germany). Standard drug solutions. Standards o f the diuretics were prepared in methanol. The concentration of each compound was 10 pg/mL. Apparatus. GC was conducted by means of a Hewlett-Packard GC 5890 (GC capillary Hewlett-Packard HP-1, 12.5 m x 0.2 mm i.d., 0.33-Wn film thickness, crosslinked methyl silicone). The detector and injector temperatures were kept at 280~ Automatic injections (2 iaL) were performed with an HP 7673 A autosampler with the oven at 90~ The temperature was stepwise increased to 200~ at 30~ and then to 280~ at 10~ Final time was set to 6 min at 280~ Purge off time was 0.5 min. Helium served as carrier gas. The GC was interfaced to a Hewlett-
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
Journal of Analytical Toxicology, Vol. 16, May/June 1992
Packard MSD 597 mass spectrometer with a Hewlett-Packard 59970C Chem-Station. The MSD was tuned every day with perfluorotributylamine (PFTBA) according to the specifications of the manufacturer. Electron energy was 70 eV. Drug screening was performed in the SIM mode and confirmation was conducted in the full scan mode (50-500 amu). Extraction procedure. Horse urine (3 mL) was adjusted to pH 5, using IM phosphoric acid and submitted to Extrelut 3 | The column was eluted with 15 mL ethyl acetate, and the eluate was passed through a column filled with 1.1 g granular Kieselgur prewetted with 1 mL saturated sodium hydrogen carbonate (pH 8.5). The 1.1 g equals one third of the material present in the original Extrelut 3 | column. This purification step is aimed at retaining stronger acidic matrix materials, permitting the passage of weaker acidic compound such as the diuretics under investigation. After the solvent flow ceased, the column was eluted with additional 2 mL ethyl acetate. The combined eluates were reduced to dryness under vacuum (Speed Vac concentrator, Bachofer, Reutlingen, Germany), and the residues were transferred in ethyl acetate to autosampler vials. After evaporation of the solvent the samples were derivatized. Derivatization. To the residues was added either 40 ILLTMAH, 40 I.tL MSTFA, or MSTFA/TMAH (20 I.tLeach). After the samples were vortexed, a 2-~L aliquot was submitted to GC/MS. Derivatization was accomplished by on-column reaction. Recovery experiments. Horse urine was spiked with the diuretic (stock solution) to a concentration of 100 ng/mL. The extraction procedure and derivatization (20 ILL TMAH plus 20 pL MSTFA) followed the above protocol. Each sample was screened in the MSD running the SIM mode for trimethyl furosemide (m/z 81, m/z 339, m/z 372), ethacrynic acid methyl ester (m/z 243, m/z 261, m/z 316) and trimethyl bumetanide (m/z 318, rn/z 363, m/z 406). Recoveries for the drugs were calculated comparing their individual peak areas with the data (areas) of an external calibration curve by injecting the drugs directly into the GC/MS. Data were processed with the HP 59970C Chem-Station programs. These curves were found to be linear from 1.25-10 ng/~L injected. Recoveries are given in Table I where the spiked amount of the drug refers to 100 percent.
A
,33 : 9t3 :
163
!54
]
~6~ \
eo 4,
]
'~
-%.
4e.
10"
91,
13~
e
tS,2
.1.~,
too
tllO:
tso
,.
.
.~oe
.~se
~,
3N
350
B 261
goi Roi
8e24]
sal
,Q
Detection of Diuretics in Horse Urine by GC/MS Heinz-Werner
Hagedorn
and R ( i d i g e r S c h u l z
Institute of Pharmaco/ogy, Toxicology, and Pharmacy, University of Munich, K6niginstraBe 16, D-8000 MOnchen 22, Germany
Abstract I The use of diuretics in horses subject to doping control is prohibited. Thus, a sensitive screening procedure is required to identify the chemically different diuretics. We communicate here a method to detect three commonly employed acidic diuretics: bumetanide, ethacrynic acid, and furosemide. A liquid-liquid extraction on Extrelut 3 ~"~was performed at weak acidic and basic conditions using ethyl acetate as organic solvent. For analysis by GC, the diuretics were methylated oncolumn in the presence of MSTFA/'FMAH, avoiding the commonly employed highly toxic derivatizing agent methyl iodide. For identification of diuretics, we used a mass selective detector operating in the SIM (selected ion monitoring) mode. Confirmation analysis may be obtained with a full scan run. Recoveries for the individual drugs ranged from 31 to 48% at the 100-ng/mL level for 3 mL urine, using calibration curves of drug standards with linearity from 2.5 to 20 ng injected. The limit of detection amounts to 40 ng/mL for the three diuretics. The method permits rapid and sensitive detection of diuretics in horse urine and is recommended for doping control
pK a values, their extraction from urine may cause problems. It is thus a common practice to extract diuretics from urine under acidic and basic conditions (4,12), by adsorption on resin (5), and by Cl8 solid phase material (8). The widely used methyl iodide method (13) of derivatization for GC analysis is time consuming, and because of its high toxicity, should be avoided. These drawbacks of current screening techniques for diuretics encouraged us to develop a more rapid and less toxic method for the detection of commonly used acidic diuretics in horses such as bumetanide, ethacrynic acid, and furosemide. The procedure involves a sequential solvent extraction of urine under acidic and basic conditions on granular Kieselgur, and, subsequently, a onepot derivatization of the diuretics. This method takes advantage of the highly reactive trimethylsilyl groups introduced by MSTFA (14), which are then substituted by methyl groups donated by TMAH. This technique appeared advantageous over current dcrivatization procedures because it is less laborious, very rapid, and sufficiently sensitive to detect low levels of the diuretics.
Experimental Introduction
Diuretics, a class of chemically very heterogeneous compounds, promote the excretion of water and salts. Because of the larger urine volumes produced, the concentration of drugs eliminated by the kidney may be reduced ( 1). In fact, the urine levels of phenylbutazone, a widely used nonsteroidal antiphlogistic dmg in horses, is decreased 40-fold in the presence of furosemide (2), and TLC screening for flunixin has given false-negative results for up to 2 h after application using a diuretic as "masking agent" (3). As expected from these results, diuretics are misused in sports to impede the detection of certain doping agents. For this reason, diuretics have been banned by a number of horse racing authorities. The literature provides several screening procedures for the detection of diuretics, including HPLC (4), TLC (5,6), GC/MS (7,8), and HPLC/MS (9) techniques. With respect to horse urine, sensitive methods have been described for furosemide (10) and bumetanide (11). However, because of the different functional groups (e.g. carboxyl, amino, aminosulfonyl) and their varying
1 94
Reagents and materials. Analytical grade ethyl acetate, methanol, and Extrelut 3 | columns were purchased from E. Merck (Darmstadt, Germany). N-trimethylaniline hydroxide (TMAH, 0.2M in methanol, ready to use) was supplied by Macherey-Nagel (Dtiren, Germany). N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA), ethacrynic acid, bumetanide, and furosemide were purchased from Sigma (Deisenhofen, Germany). Standard drug solutions. Standards o f the diuretics were prepared in methanol. The concentration of each compound was 10 pg/mL. Apparatus. GC was conducted by means of a Hewlett-Packard GC 5890 (GC capillary Hewlett-Packard HP-1, 12.5 m x 0.2 mm i.d., 0.33-Wn film thickness, crosslinked methyl silicone). The detector and injector temperatures were kept at 280~ Automatic injections (2 iaL) were performed with an HP 7673 A autosampler with the oven at 90~ The temperature was stepwise increased to 200~ at 30~ and then to 280~ at 10~ Final time was set to 6 min at 280~ Purge off time was 0.5 min. Helium served as carrier gas. The GC was interfaced to a Hewlett-
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
Journal of Analytical Toxicology, Vol. 16, May/June 1992
Packard MSD 597 mass spectrometer with a Hewlett-Packard 59970C Chem-Station. The MSD was tuned every day with perfluorotributylamine (PFTBA) according to the specifications of the manufacturer. Electron energy was 70 eV. Drug screening was performed in the SIM mode and confirmation was conducted in the full scan mode (50-500 amu). Extraction procedure. Horse urine (3 mL) was adjusted to pH 5, using IM phosphoric acid and submitted to Extrelut 3 | The column was eluted with 15 mL ethyl acetate, and the eluate was passed through a column filled with 1.1 g granular Kieselgur prewetted with 1 mL saturated sodium hydrogen carbonate (pH 8.5). The 1.1 g equals one third of the material present in the original Extrelut 3 | column. This purification step is aimed at retaining stronger acidic matrix materials, permitting the passage of weaker acidic compound such as the diuretics under investigation. After the solvent flow ceased, the column was eluted with additional 2 mL ethyl acetate. The combined eluates were reduced to dryness under vacuum (Speed Vac concentrator, Bachofer, Reutlingen, Germany), and the residues were transferred in ethyl acetate to autosampler vials. After evaporation of the solvent the samples were derivatized. Derivatization. To the residues was added either 40 ILLTMAH, 40 I.tL MSTFA, or MSTFA/TMAH (20 I.tLeach). After the samples were vortexed, a 2-~L aliquot was submitted to GC/MS. Derivatization was accomplished by on-column reaction. Recovery experiments. Horse urine was spiked with the diuretic (stock solution) to a concentration of 100 ng/mL. The extraction procedure and derivatization (20 ILL TMAH plus 20 pL MSTFA) followed the above protocol. Each sample was screened in the MSD running the SIM mode for trimethyl furosemide (m/z 81, m/z 339, m/z 372), ethacrynic acid methyl ester (m/z 243, m/z 261, m/z 316) and trimethyl bumetanide (m/z 318, rn/z 363, m/z 406). Recoveries for the drugs were calculated comparing their individual peak areas with the data (areas) of an external calibration curve by injecting the drugs directly into the GC/MS. Data were processed with the HP 59970C Chem-Station programs. These curves were found to be linear from 1.25-10 ng/~L injected. Recoveries are given in Table I where the spiked amount of the drug refers to 100 percent.
A
,33 : 9t3 :
163
!54
]
~6~ \
eo 4,
]
'~
-%.
4e.
10"
91,
13~
e
tS,2
.1.~,
too
tllO:
tso
,.
.
.~oe
.~se
~,
3N
350
B 261
goi Roi
8e24]
sal
,Q
