Rapid screening method for diuretics in doping control using automated solid phase extraction and liquid chromatography-electrospray tandem mass spectrometry

Analytica Chimica Acta 502 (2004) 65–74

Rapid screening method for diuretics in doping control using automated solid phase extraction and liquid chromatographyelectrospray tandem mass spectrometry Catrin Goebel, Graham J. Trout, Rymantas Kazlauskas∗ Australian Sports Drug Testing Laboratory, Australian Government Analytical Laboratories, 1 Suakin Street, Pymble, NSW 2073, Australia Received 18 August 2003; received in revised form 25 September 2003; accepted 30 September 2003

Abstract A new method has been developed for the routine detection of diuretics in urine samples collected from athletes. The method developed uses automated solid phase extraction (SPE) with analysis of the extracts by high performance liquid chromatography using electrospray ionisation tandem mass spectrometry. It requires only one injection per sample and is currently capable of detecting 35 diuretics and related compounds at the rate of five samples per hour. All parent compounds can be detected at urinary concentrations significantly below 100 ng/ml with the recovery of most analytes being greater than 80%. Crown Copyright © 2003 Published by Elsevier B.V. All rights reserved. Keywords: Diuretics; Urine; Doping; LC–MS–MS; SPE

1. Introduction Diuretics are pharmaceutical drugs which are used to increase urine flow by promoting the excretion of water by the kidneys. They are used often for the treatment of heart conditions, liver, kidney or lung disease to alleviate salt or water retention. Their potent ability to remove water has caused diuretics to be misused in sport where rapid weight loss is required to meet a weight category. It has also been used to ensure the urine is diluted so that the detection of other banned substances is made more difficult. Thus the use of diuretics in sport is banned at competition. The urine dilution effect of diuretics also allow them to be classified as masking agents and precludes their use both in and out of competition [1]. There are several classes of diuretic drugs—Thiazides (e.g., benzthiazide), loop diuretics (e.g., bumetanide), potassium sparing diuretics (e.g., amiloride), carbonic anhydrase inhibitors (e.g., acetazolamide), osmotic diuretics (e.g., mannitol) and mercurial diuretics (e.g., mersalyl). All these types are banned, however, it is normal to only screen for the first

∗

Corresponding author. Tel.: +61-2-94490111; fax: +61-2-94498080. E-mail address: [email protected] (R. Kazlauskas).

four classes as it is currently impossible to determine from a urine sample whether mannitol has been infused (banned) or ingested (permitted), and mercurial diuretics have been superseded. The drugs are all relatively polar and hence are amenable to analysis by HPLC using C18 type phases with diode array (DAD) or fluorescence detection [2,3]. However, whilst this method may be suitable for screening purposes it suffers from significant interferences due to the background occurring in urine samples. The information provided is insufficient for confirmation and alternative mass spectral confirmation of identity is needed for doping control. Until recently to confirm the identity of diuretic compounds detected by HPLC, GC–MS was used. In order to improve the volatility of the diuretics either methylation or silylation has been required prior to GC–MS. The methylation of these polar drugs has been the most common multiresidue procedure [4] and has allowed GC–MS to be used as a screen to replace the less selective HPLC procedures. The simplest process is an extractive alkylation one, where extraction and derivatisation is carried out in a single step [5]. This method was successfully used in the Sydney 2000 Olympic Games to analyse some 3000 samples in 4 weeks [6]. Recently a rapid method has been published using microwaves to assist the methylation after extraction [7].

0003-2670/$ – see front matter. Crown Copyright © 2003 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2003.09.062

66

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

Two problems with the methylation prior to GC–MS approach are the difficulty in methylating some diuretics, and the toxicity of the methyl iodide used in the derivatisation process. A few drugs such as triamterene and amiloride are analysed as trimethylsilyl (TMS) derivatives as part

O

O

NH2SO2

S

R2

N

X

of other screens. With the advent of more affordable and reliable LC–MS instrumentation, a technique particularly suited to the detection of polar substances, the opportunity has arisen to develop methods for the detection and confirmation of diuretics without the need for derivatisa-

1; X=Cl, R1=CH2SCH2CH=CH2, R2=H Althiazide

N H

, R2=H

1; X=Cl, R1=,

Cyclopenthiazide

R1

1; X=Cl, R1=H, R2=H

1

Hydrochlorothiazide 1; X=CF3, R1=CH2-C6H5, R2=H Bendroflumethiazide

1; X=CF3, R1=H, R2=H 1; X=Cl, R1=

Hydroflumethiazide

, R2=H

Cyclothiazide

1; X=Cl, R1=CH2-S-CH2-CF3, R2=CH3

1; X=Cl, R1=CH-Cl2, R2=H

Polythiazide

Trichlormethiazide H3C O

S

CH3CONH N

SO2NH2 N

NH2SO2

S

CH3CON N

N H

SO2NH2 N

Cl

H3C

OH

H3C

Acetazolamide

O

Methazolamide

Xipamide

O

NH2SO2

S

H N

2; X=Cl, R1=H

2; X=Cl, R1=CH2-S-CH2-C6H5

Chlorothiazide

Benzthiazide X

N

R1

2

Cl

Cl

NH2

F3C

NH2

N NH2SO2

NH2SO2

O

SO2NH2

NH2SO2

Chloroamidophenamide

Chlorexolone

3; R1=NHCH2CH2CH3, R2=OC6H5, R3= SO2NH2, R4=H

R3

R2

3; R1=H, R2=Cl, R3=SO2NH2, R4= NH-2-furanylmethyl

Bumetanide

COOH

R1

SO2NH2

CF3- amidophenamide

Furosemide

3; R1=SO2NH2, R2=0C6H5, R3=1pyrrolidinyl, R4=H

R4

3

Piretanide

O O OCH2COOH CH3CH2

H 3C

NH

O

O

O

CH 3

Cl

N

S

O

Cl

N H

Cl

Ethacrynic acid

N H

C

NH N H

CH 3

N

Torasemide

Fig. 1. Structures of the diuretics and other analytes.

H2N

N

NH2

Amiloride

C

NH2

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

67

Fig. 1. (Continued ).

tion [8,9]. The possibility of automated sample preparation increases the potential for faster and simpler analysis. A robust LC–MS screening procedure for the detection of 32 diuretics and masking agents has been reported [10]. This method was based on liquid liquid extraction, which is difficult to automate, and required two injections per sample. The duplication of analysis was performed because of the need for both positive and negative ionisation in order to detect all the diuretics with sufficient sensitivity. The aim of this work was to develop a simple LC–MS method with automated sample preparation that could replace our existing GC–MS method based on extractive

alkylation. The method was also required to reliably detect those diuretics the existing method could not. Our detection limit was required to be below 100 ng/ml for every diuretic as this is the requirement historically set by the IOC. However the minimum required performance limit (MRPL) may be increased in the proposed World Anti-Doping Agency’s (WADA) international standard for laboratories [11]. Two stimulants pemoline and mesocarb and a masking agent, probenecid, which were detected in our GC–MS diuretic method were also required to be in the menu. The performance of the method has been evaluated for 35 analytes, the structures for which is shown in Fig. 1.

68

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

2. Experimental 2.1. Chemicals and reagents Water was purified using a Milli-Q water purification system (Millipore, Sydney, Australia). All reagents were analytical-reagent or HPLC grade: methanol (Mallinckrodt), acetonitrile (Mallinckrodt), ethyl acetate (Merck), tertiary butyl methyl ether (Mallinckrodt), formic acid 98% (Ajax Chemicals, Australia), acetic acid (Ajax Chemicals, Australia), sodium acetate (Ajax Chemicals, Australia) and ammonium acetate (Ajax Chemicals, Australia). HPLC mobile phases were filtered through either a 0.2 ␮m cellulose nitrate membrane or a 0.2 ␮m PTFE filter. Ultra high purity nitrogen and argon gas were obtained from BOC (Australia). Standards of the following diuretics were obtained: acetazolamide (National Analytical Reference Laboratory (NARL), Australia), althiazide (Searle, Australia), amiloride (Sigma–Aldrich, Australia), bendroflumethiazide (E. R. Squibb & Sons, Australia), benzthiazide (Sigma–Aldrich), bumetanide (NARL), canrenone (Searle), chlorexolone (May & Baker Pharmaceuticals, Australia), chlorothiazide (NARL), chlorthalidone (Ciba–Geigy, Australia), clopamide (NARL), cyclopenthiazide (Ciba–Geigy), cyclothiazide (Sigma–Aldrich), dichlorphenamide (Merck Sharp and Dome, Australia), ethacrynic acid (Merck Sharp and Dome), furosemide (Fisons Pharmaceuticals, Australia), hydrochlorothiazide (Fisons Pharmaceuticals), hydroflumethiazide (Mead Johnson & Company), indapamide (Sigma–Aldrich), mefruside (Bayer, Australia), methazolamide (Lederle, Australia), metolazone (NARL), pemoline (NARL), piretanide (Hoechst), polythiazide (Pfizer Taito), probenecid (Merck Sharp and Dome), quinethazone (NARL), spironolactone (NARL), torasemide (courtesy of Czech Republic IOC Laboratory), triamterene (SmithKline Beecham), trichlormethiazide (Merrell Dow, USA), xipamide (courtesy of Athens IOC Laboratory) and mesocarb (courtesy of the Moscow IOC laboratory). 2.2. Sample preparation To each 2 ml of urine was added 2 ml of pH 5 acetate buffer and 100 ␮l of mefruside internal standard (10 ␮g/ml). The tubes were vortexed and the analytes extracted by passage of the sample through a Varian ABS ELUT Nexus SPE column (60 mg, 3 ml), followed by a 1 ml water wash, a 1 ml wash with 20% methanol in water, and elution with 2 ml methanol. The extractions were carried out using a Gilson ASPEC XL4 programmed for routine sample preparation. The methanolic extract was evaporated to dryness under nitrogen and reconstituted in 200 ␮l of 50% methanol in water. 2.3. Instrumentation A Waters Alliance 2795 separation module equipped with a quaternary pump was used for the LC separation. The mass

spectrometer used was a Micromass Quattro Micro triple stage quadrupole equipped with a Z spray API interface. Full scan and multiple reaction monitoring (MRM) was carried out using both positive and negative modes. 2.4. LC parameters A C18 column (Alltech Prevail, 50 mm ×2.1 mm ×3 ␮m) protected by a C18 guard column (Phenomenex Security Guard 4 mm × 2 mm) was used. The following ternary mobile phase gradient was formed by solvent A (2% aqueous formic acid), solvent B (water), and solvent C (acetonitrile) at a flow rate of 0.2 ml/min—constant 10% A throughout the run, 0% C (0–1 min), 0–80% C (1–6.5 min), 80% C (6.5–7.5 min), 80–0% C (7.5–8 min), and 0% C (8–11 min). The injection volume was 10 ␮l. 2.5. MS parameters The spray conditions of the API interface were: desolvation temperature 180 ◦ C, desolvation gas flow 550 l/h, and cone gas flow 50 l/h. The capillary was set at 3.5 kV and the substance specific cone voltage and collision energy were optimised for each compound. Argon was used as the collision gas at a pressure of 3.3 × 10−3 mbar.

3. Results The method chosen is designed to detect the presence of all the desired analytes at a concentration of less than 100 ng/ml in urine with minimum sample preparation and be suitable for automation using an existing Gilson ASPEC XL4. It was developed from some previous work using LC–MS for the detection of diuretics [12]. As most diuretics are excreted unchanged in the urine it was possible to work with standards for most drugs. Two of the substances are extensively metabolised so it was necessary to use metabolites from administration study urine samples rather than pure substances, these being the hydroxy sulphate metabolite of mesocarb and the oxo-metabolite of mefruside. These and standards of diuretics excreted as the parent compound were analysed to obtain instrumental conditions, retentions times and reference product ion spectra. Both positive and negative modes of ionisation were used for each analyte to obtain mass spectra. The MS–MS data were obtained by choosing a precursor ion and measuring the product ion transitions for each compound for both modes of ionisation. The most appropriate precursor to product ion transition in each ionisation mode was selected for each compound and this was optimised to obtain maximum sensitivity and specificity. An estimate of the limit of detection (LOD) for each compound for both modes was made by extracting blank urine and spiking these extracts with analytes in the concentration range from 5 to 100 ng/ml in the urine. Using an MRPL of

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

69

Table 1 Data obtained for the screening of analytes including mass spectral data and quality data

Substances for which standards are not available to provide accurate recovery data and for which excretion urines were used are shaded.

50 ng/ml (half that of the proposed IOC limit), 24 of the compounds could be detected in positive ion mode and for 19 compounds this mode gave a lower or equal LOD to that found with negative ionisation. However seven compounds were not able to be detected in positive ion mode and thus for any comprehensive screen both modes of ionisation must be used. The mass spectrometer used for this work is capable of switching between positive and negative ionisation modes in a single analysis and so a multifunction MRM method was set up for the detection of all analytes in a screening process. The retention times, precursor and product ions, the cone voltages and collision energies used are shown in Table 1. The data processing method generates a three page graphical and text output. The printout for a 50 ng/ml urine spike is shown in Fig. 2. The method is designed for the rapid qualitative identification of substance(s), that may be present in a urine by monitoring some 36 product ion transitions in a relatively

short LC run and is not expected to give precise quantitation. However for the purpose of validation of the procedure the linearity was evaluated from 10 to 200 ng/ml in urine for all parent drug analytes (these are concentrations at the lower range of those expected after administration of a normal therapeutic dose and in most cases representing levels obtained a considerable time after administration). All gave a linear response with correlation coefficients (R2 ) ranging from 0.975 to 0.994. The recoveries of the diuretics excreted as parent compounds were determined by spiking seven replicates of a blank urine with each analyte at a concentration of 50 ng/ml and comparing these results with an unextracted standard. The values obtained are shown in Table 1 and are all acceptable. As part of the method’s robustness the 20% methanol wash solvent in the sample preparation was changed to 100% water and the recoveries were also determined. It was found that the recoveries were similar although the variabil-

70

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

ity tended to be less, but the background noise for some of the compounds was significantly higher. Whilst the method recovery check is a good indication of the validity of the method it is restricted in that only one urine sample was used for spiking. It was a necessary criterion that the method be capable of detecting all analytes in all urine samples at the MRPL. It is not possible to test all urines and so urine samples from 40 individuals were taken and each was spiked with all the parent compounds at a concentration of 50 ng/ml. The results showed that all analytes could be detected in all 40 urine samples although

for some samples the recoveries of dichlorphenamide and/or quinethazone were lower than expected but the peaks were clearly observable. It was noted that there was a correlation between high urine specific gravity and low recovery of dichlorphenamide and/or quinethazone. After the method had been optimised and in routine use for some months, the method detection limit (MDL) was determined for each parent compound by spiking aliquots of urine with 31 analytes at concentrations of 0.2, 1, 4 and 20 ng/ml. There were seven aliquots of each concentration and the means and standard deviations were calculated. The

Fig. 2. Example of the results printout for the diuretics. The extracted ion chromatograms for each substance at 50 ng/ml (except metabolites and artefacts) show peaks well above the background making detection easy. The tabular output at the end of the printout gives a list of each substance, its mass spectral trace, the retention time of the main peak, the area, the height and the boolean describing the detection of the peak.

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

71

Fig. 2. (Continued )

MDL is the minimum concentration of a substance that can be measured and reported with 99% confidence that the analyte concentration is greater than zero [13]. The results obtained are also shown in Table 1. Whilst the primary objective of the method was to develop an efficient procedure to screen for the presence of diuretics in large numbers of urine samples, data was also collected which would enable the method to be extended to include the confirmation of any suspect samples detected by the screening process. The precursor and product ions and their relative intensities for most of the analytes are set out in Table 2.

4. Discussion The method described is an enhancement of an LC–MS method described previously [12]. It was confirmed that both positive and negative modes of ionisation were required to detect all the desired analytes [10]. The method variations tried went through a series of transitions: (i) The incorporation of both modes of ionisation whilst retaining solvent extraction with ethyl acetate. (ii) The fast methanol water gradient acidified with formic acid used for the HPLC separation in the

72

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

Fig. 2. (Continued ).

original method was slowed in order to increase the separation and allow more analytes to be detected. (iii) The evaluation of solvent systems using water with added ammonium acetate, formic acid, and acetic acid combined with methanol or acetonitrile. The best combination of ionisation efficiency and chromatographic peak shape was found with the formic acid acetonitrile combination. (iv) The evaluation of three C18 columns (50 mm × 2 mm), Luna (Phenomenex), Discovery (Supelco), and Prevail (Alltech). The Prevail column was chosen as the com-

pounds were more evenly distributed across the separation time. When the published method was applied to a variety of urines it was found that the diuretics althiazide, benzthiazide, and polythiazide could not be recovered from some 20% of the urines tested. These compounds all have sulfide side chain substituents and their oxidation has been reported [10]. This was confirmed by the detection of the corresponding sulfoxides and sulfones. The oxidation has been attributed to impurities in the ethyl acetate [10]. We found that the purity of the ethyl acetate does not seem to

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74

73

Table 2 Confirmation data for the diuretics Analyte

RRT

Ionisation mode

m/z 1 (precursor)

Rel int (%)

m/z 2

Rel Int (%)

m/z 3

Rel int (%)

m/z 4

Rel int (%)

Acetazolamide Althiazide Amiloride Bendroflumethiazide Benzthiazide Bumetanide Canrenone Chlorexolone Chlorothiazide Chlorthalidone Clopamide Cyclopenthiazide Cyclothiazide Dichlorphenamide Ethacrynic acid Furosemide Hydrochlorothiazide Hydroflumethiazide Indapamide Mefruside (IS) Mefruside (oxo metab) Mesocarb hydroxysulfate Methazolamide Metolazone Pemoline Piretanide Polythiazide Probenecid Quinethazone Spironolactone Torasemide Triamterene Trichlormethiazide Xipamide

0.72 0.95 0.73 1.01 0.99 1.05 1.14 0.99 0.72 0.86 0.89 1.02 1.00 0.86 1.07 0.95 0.74 0.83 0.98 1.00 0.96 0.85 0.79 0.95 0.79 1.02 1.01 1.06 0.81 1.10 0.93 0.80 0.91 1.04

− − + − − + + + − + + + − − − − − − + + + + − + + + − + − + + + − −

220.9 381.9 170.8 419.9 307.8 364.9 340.9 246.8 213.8 321.5 345.9 362.9 387.6 302.7 242.9 329.0 297.6 329.7 132.0

21 17 49 19 100 26 11 13 81 43 93 100 100 100 74 29 11 15 59

82.7 340.9 142.9 327.8 227.9 283.9 187.1 166.9 179.0 242.9 249.8 252.8 322.1 266.9 207.0 285.0 268.9 238.9 116.9

100 100 76 100 92 33 15 100 100 100 100 16 38 47 23 100 49 77 44

79.6 268.9 115.8 288.9 238.9 240.0 106.9 165.8 114.0 241.0 111.9 236.8 266.8 238.9 191.9 248.9 205.0 160.0 90.8

21 43 84 88 16 100 100 46 24 76 24 15 85 25 27 13 100 38 100

57.7 96.8 59.8 238.9 174.7 183.9 96.8 54.9

43 38 100 51 9 32 28 93

184.9 96.9 82.9 204.4 77.8 68.6 204.8 77.8 77.7

54 43 24 66 38 100 60 51 100

378.6 192.8 234.8 258.9 105.8 362.8 397.8 244.0 287.9 340.8 264.0 253.9 307.8 352.9

20 54 11 19 33 50 37 47 28 56 40 21 65 31

350.6 118.8 77.7 178.9 78.7 281.9 378.0 201.9 252.8 294.9 229.9 237.0 306.0 273.9

11 100 83 100 100 94 19 100 20 10 23 57 77 35

142.9 90.8 57.6 177.7 76.8 237.9 354.8 184.9 245.0 205.0 183.0 194.9 244.0 272.9

100 70 100 29 42 100 15 43 100 14 100 36 68 26

83.8

61

150.9

61

236.0 323.8 71.8 208.9 106.9 168.0 103.9 241.9 77.7

62 100 90 20 100 51 100 100 100

This provides the major product ions and their intensity for the transition from the precursor ion data shown in Table 1.

be the sole cause as most urine samples tested showed no sign of oxidation using the one solvent batch. The fact that some samples repeatedly and reproducibly underwent sulfide oxidation would appear to indicate that some agent in the urine was causing the oxidation in the presence of ethyl acetate. Thus ethyl acetate should not be used for the extraction of diuretics with sulfide side chain substituents. The use of tertiary butyl methyl ether (TBME) was investigated as an alternative extraction solvent. This overcame the oxidation problem but it could not be used generally as it did not extract amiloride and had very low recoveries for acetazolamide, quinethazone and triamterene. The problem with ethyl acetate extraction and the requirement for automation prompted the use of solid phase extraction rather than solvent extraction. Both Nexus ABS ELUT (Varian) and Strata X (Phenomenex) cartridges were evaluated. The Strata X did not extract acetazolamide, hydrochlorothiazide, methazolamide and quinethazone, and had low recoveries for chlorothiazide, chlorthalidone and dichlorphenamide. However, it was found with both SPE cartridges that the extracts were considerably dirtier than

those obtained by ethyl acetate extraction. Thus, for confirmation analyses, it is still preferable to use solvent extraction with ethyl acetate except for the three diuretics with sulfide side chains where TBME should be used. Two stimulants, pemoline and mesocarb, were included in the screen as their detection had been previously undertaken by the extractive alkylation GC–MS screen the LC–MS was to replace. Mesocarb is excreted predominantly as the hydroxy sulfate metabolite [14] and as such is readily detected by LC with UV or DAD detectors. An ethyl acetate extraction from a mesocarb excretion study urine was used to detect and identify the metabolites by comparison with published reference spectra [15]. This extract was then used to select the optimum API and MS conditions (Table 1). Hydrolysis of benzothiadiazines is observed in aqueous solutions [10] and the rate of reaction is dependent on storage conditions and the individual compounds. The products are dependent on the substituents R2 and X (Fig. 3). The majority of the benzothiadiazines hydrolyse to chloroamidophenamide (R2 = H, X = Cl) with other examples include polythiazide (R2 = CH3 , X = Cl)

74

C. Goebel et al. / Analytica Chimica Acta 502 (2004) 65–74 O NH 2SO2

O

O R2

S

NH 2SO2

O S

N

NHR2

Hydrolysis X

N

R1

X

NH 2

H

Fig. 3. Hydrolysis reaction for benzothiadiazines where the product is dependent on the R2 and X substitution.

which hydrolyses to methylchloroamidophenamide and bendroflumethiazide (R2 = H, X = CF3 ) which hydrolyses to triflumethylamidophenamide. Because of the potential for hydrolysis in the urine chloroamidophenamide and triflumethylamidophenamide are monitored in this method. The methylchloroamidophenamide is observed for polythiazide but in our hydrolysis rate experiments the polythiazide proved to be sufficiently stable for screening and not to need the hydrolysis monitored. At present any positives detected are confirmed by our GC–MS methylation method but work is proceeding on developing LC–MS confirmation methods based on the data shown in Table 2. Most compounds have three product ions of sufficient intensity to enable effective confirmation by comparing their ion ratios. Whilst the current screening method only takes 12 min per sample the throughput could be increased in two ways. The first requires no additional instrumentation and merely involves combining two samples prior to extraction. This is possible because of the high sensitivity of the method and since our in house required MRPL is 50 ng/ml whilst the highest MDL is 21 ng/ml. Of course any positive result would require the samples to be analysed individually but as our positive rate is significantly below 1% this would be an infrequent occurrence. The second way of increasing the throughput would be by using multiplexed chromatography where two columns are used to deliver the analytes to the mass spectrometer [16]. This is possible because all the analytes of interest are eluted in approximately 5 min and hence with staggered injections two samples could be analysed in 12 min. A combination of these techniques would enable the analysis to be increased from 5 sample per hour to 20 samples per hour. During the 2000 Sydney Olympic Games our laboratory required four GC–MS instruments to achieve this higher throughput [6].

5. Conclusions The method developed has shown itself to be simple, robust and reliable in the detection of diuretics for sports drug testing. It has been in routine use for more than 12 months involving the analysis of over 6000 urine samples. Only one HPLC column was needed for this period. Use of the

GC–MS diuretics screening proceeding had been discontinued after running both methods in parallel for 1 month. Maintenance of the Quattro Micro has been minimal merely involving occasional washing of the inlet cone to restore sensitivity. Apart from the intended advantages of including more diuretics with faster, simpler and safer sample preparation it has also been found that the interpretation of the data is much easier as interfering peaks are very rarely found in samples despite only one ion transition being monitored for each compound.

References [1] Olympic Movement Anti-Doping Code, Prohibited Classes of Substances and Prohibited Methods, IOC, Lausanne, 2001, Appendix A. [2] S.F. Cooper, R. Masse, R. Dugal, J. Chromatogr. 489 (1989) 65–88. [3] H.-J. Guchelaar, L. Chandi, O. Schouten, W.A. van den Brand, Fresenius J. Anal. Chem. 363 (1999) 700–705. [4] S. Park, J. Park, H. Pyo, Y. Kim, M. Kim, J. Park, J. Anal. Toxicol. 14 (1990) 84–90. [5] A. M Lisi, R. Kazlauskas, G.J. Trout, J. Chromatogr. 581 (1992) 57–63. [6] R. Kazlauskas, Clin. Biochem. Rev. 23 (2002) 35–51. [7] L. Amendola, C. Colamonici, M. Mazzarino, F. Botre, Anal. Chim. Acta 475 (2003) 125–136. [8] V. Sanz-Nebot, I. Toro, R. Berges, R. Ventura, J. Segura, J. Barbosa, J. Mass Spectrom. 36 (2001) 652–657. [9] K. Deventer, F.T. Delbeke, K. Roels, P. Van Eenoo, Biomed. Chromatogr. 16 (2002) 529–535. [10] D. Thieme, J. Grosse, R. Lang, R.K. Mueller, A. Wahl, J. Chromatogr. B 757 (2001) 49–57. [11] Minimum required performance limits for diuretics in the proposed WADA approved international standard for laboratories version 3, http://www.wada-ama.org/docs/web/standards harmonization/ code/min limits 1 2.pdf, accessed 9th July 2002. [12] C. Fouracre, R. Kazlauskas, in: W. Schänzer, H. Geyer, A. Gotzmann, U. Mareck-Engelke (Eds.), Recent Advances in Doping Analysis, vol. 9, Sport und Buch Strauß, Köln, 2001, pp. 321–326. [13] SW-846 Test Methods for Evaluating Solid Waste, vol. 1A, USEPA, 1992. [14] R. Ventura, T. Nadel, M.j. Pretel, A. Solans, J.A. Pascual, J. Segura, in: M. Donike, H. Geyer, A. Gotzmann U. Mareck-Engelke, S. Rauth (Eds.), Proceedings of the 10th Cologne Workshop on Dope Analysis, Sport und Buch Strauss, Köln, 1993, pp. 231–248. [15] S. Yang, S. Zhu, Z. Yang in: W. Schänzer, H. Geyer, A. Gotzmann U. Mareck-Engelke (Eds.), Recent Advances in Doping Analysis, vol. 4, Sport und Buch Strauss, Köln, 1997, pp. 371–376. [16] R. Oertel, K. Richter, J. Fauler, W. Kirch, J. Chromatogr. A 948 (2002) 187–192.