Contents lists available at SciVerse ScienceDirect
Journal of Chromatography B
j ourna l h om epage: www.elsev ier .com/ locate /chromb
fully validated method for the quantification of ethyl glucuronidend ethyl sulphate in urine by UPLC–ESI-MS/MS applied in arospective alcohol self-monitoring study
atalie Kummera,∗, Sarah Willea, Vincent Di Fazioa, Willy Lambertb, Nele Samyna
Federal Public Service Justice, National Institute of Criminalistics and Criminology, Brussels, BelgiumGhent University, Faculty of Pharmaceutical Sciences, Ghent, Belgium
a r t i c l e i n f o
rticle history:eceived 19 November 2012eceived in revised form 8 April 2013ccepted 12 April 2013vailable online 17 April 2013
eywords:thyl glucuronidethyl sulphatealidationPLC–ESI-MS/MSlcohol markerrospective study
a b s t r a c t
A method for the quantification of ethyl glucuronide (EtG) and ethyl sulphate (EtS) in human urineis developed and fully validated according to international guidelines. Protein precipitation is used assample preparation. During the development of the method on an UPLC–ESI-MS/MS system using aCSH C18 column, special attention was paid to reduce matrix effects to improve assay sensitivity and toimprove detection of the second transition for EtS for specificity purposes. The method was linear from0.1 to 10 �g/mL for both analytes. Ion suppression less than 24% (RSD < 15%) was observed for EtG andno significant matrix effect was measured for EtS. The recovery was around 80% (RSD < 14%) for bothcompounds. This method provides good precision (RSDr and RSDt < 10%) and bias (<15%) for internaland external quality control samples. The reproducibility of the method was demonstrated by the suc-cessful participation to proficiency tests (z-score < 0.86). This method was finally used to analyze urinesamples obtained from twenty-seven volunteers whose alcohol consumption during the 5 days beforesampling was monitored. Concentrations between 0.5 and 101.9 �g/mL (mean 10.9, median 1.4) for EtG
and between 0.1 and 37.9 �g/mL (mean 3.6, median 0.3) for EtS were detected in urine samples of volun-teers who declared having consumed alcohol the day before the sampling. EtG and EtS concentrations inurine were highly correlated (r = 0.996, p < 0.001). A moderate correlation between the number of drinksthe day before sampling and the concentration of EtG (r = 0.448, p < 0.02) or EtS (r = 0.406, p < 0.04) wasobserved. Using a cut-off value at 0.1 �g/mL for EtG and EtS, this method is able to detect social alcoholconsumption approximately 24 h after the intake, without showing any false positive result.
. Introduction
Ethyl glucuronide (EtG) and ethyl sulphate (EtS) are twopecific metabolites of ethanol, created by conjugation withDP-glucuronic acid for EtG [1] and with 3′-phosphoadenosine′-phosphosulphate for EtS [2].
Quantification of EtG and EtS in urine is used to detect recentlcohol consumption. These biomarkers extend the detection win-ow relative to blood ethanol measurement and, compared to
ong term biomarkers, allow the detection of drinking of small
uantities. This permits to monitor alcohol consumption duringithdrawal treatment [3,4] or for workplace testing [5,6]. Several
ountries, such as Italy [7] and Germany [8,9], have integrated the
∗ Corresponding author at: National Institute of Criminalistics and Criminology,aboratory of Toxicology, Chaussée de Vilvorde 100, 1120 Brussels, Belgium.el.: +32 02 240 05 53; fax: +32 02 243 46 08.
quantification of EtG and EtS in urine into their licence regrant-ing programme to monitor the abstinence period. In post-mortemcases, the detection of EtG and EtS in urine is usefull to distinguishbetween antemortem alcohol intake and postmortem formation ofethanol [10–12].
EtG and EtS are detectable in urine up to 24 h after intake of0.25 g/kg ethanol and up to 48 h after intake of 0.50 g/kg ethanol[2,13–18]. After alcohol intoxication, they can be detected in urineduring a few days. EtG is eliminated according to a half-timeof 2.5 h [1,13]. After consumption of alcohol and depending onthe amount of consumed alcohol, urinary concentrations for EtGand EtS can vary from some �g/mL [11,18–21] to hundreds of�g/mL [13,15,17,21–23]. Urine samples from alcohol-dependentpatients during detoxification can reach EtG concentrations up to1240 �g/mL [1,2,4,24] and EtS concentrations up to 264 �g/mL
[2].
Due to the possibility of finding EtG and EtS in urine even with-out consumption of alcoholic beverages [18,19,25–28], a cut-offlimit is used to avoid false positive results. However, cut-offs are
ot fixed yet in international guidelines. The ones currently usedary between 0.1 and 1.1 �g/ml [7,9]. A cut-off at 0.1 �g/mL for EtGnd at 0.05 �g/mL for EtS has been proposed to exclude repeatedntake of alcohol [9]. Urine analysis of teetotallers shows no EtG1,29] and no EtS [30] above 0.1 �g/mL.
The most commonly applied technique for quantification oftG and EtS in urine is liquid chromatography coupled to masspectrometry (LC–MS) [4,13] or coupled with tandem mass spec-rometry (LC–MS/MS) [1,7,11,12,17,26,30–33] in combination withimple dilution or protein precipitation as sample preparation. Aew methods have been published using GC–MS [3,29,34,35] or cap-llary zone electrophoresis (CZE) [36–39] for the analysis of EtG andtS in urine or serum.
To decrease matrix effects, especially for EtS, sample prepara-ion should be adapted. Dilution of urine is the easiest ‘samplereparation’ method, however, high matrix effects and higher
nstrument maintenance can be problematic in routine analysis.ven with a 1/20 dilution, relevant matrix effect was observed atow concentrations [30]. Liquid–liquid extraction (LLE) and solidhase extraction (SPE) are conventional sample preparation tech-iques for non-volatile compounds. Due to the high polar andcidic character of EtG and EtS in combination with a differentcidic strength, the development of LLE and SPE is, however, nottraightforward. Protein precipitation can be an alternative clean-p method for this type of analytes [12,17–19], if the matrix effectsre carefully monitored. No matrix effects were reported after pro-ein precipitation [12] using a LC–MS/MS system coupled with ionrap mass spectrometer.
Reversed-phase (RP) chromatography used with negative elec-rospray ionization mode (ESI−) is the most common approach used1,4,11,13,17,23,26,31,33,40]. The retention of very polar acidicompounds, such as EtG (pKa ∼2.84) and EtS (pKa ∼−3.14) [11],s achieved in RP only under highly aqueous conditions. As highlyqueous conditions might not be optimal for ESI ionization, post-olumn addition of an organic modifier is required to enhance theonization of compounds and so to improve sensitivity. A chromato-raphic possibility to improve the ionization is to use a normalhase column [12] or another specific column with particularetention behaviour [11]. Nevertheless, normal phase chromatog-aphy is known to provide variable retention times [41]. The usef no discharge atmospheric pressure chemical ionization (ND-PCI) [7] or atmospheric pressure chemical ionization APCI [30]
s another solution to increase the ionization and so to improve theimit of quantification.
According to international guidelines, forensic analysis byS/MS in MRM mode requires the detection of minimum two
ransitions for each compound; one for identification and one foruantification [33,42]. When LC–MS is used, three characteristic
ons are required. Sometimes it is difficult to find a second transitionor EtS using LC–ESI-MS/MS [12,31], because of the low intensity ofhe second transition and the presence of interfering compoundsn urine.
Low limits of quantification have been reported using aC–MS/MS system coupled with ion trap mass spectrometer7,11,12,43]. Using LC–ESI-MS/MS systems equipped with a tripleuadrupole, only one published method [19] has reported an LOQf 0.1 �g/mL for EtG and EtS. Unfortunately no details of the methodalidation are given in that publication.
As seen before, several LC–MS (/MS) methods have beenescribed for the quantification of EtG and EtS in urine, but mostf them are not fully validated (following all criteria for chromato-raphic assays), especially regarding the measurement of bias and
recision with external and certified quality controls. Moreover, tour knowledge, to date, only two published reports [43,44] havevaluated the reproducibility of the method by participation tonterlaboratory tests.
. B 929 (2013) 149– 154
We describe the development of a simple and robust methodfor the quantification of EtG and EtS in urine using an UPLC–ESI-MS/MS system equipped with a triple quadrupole (QQQ) tandemmass spectrometer, with an LOQ at 0.1 �g/mL both for EtG andEtS. The method was fully validated according to internationalguidelines. A prospective study, based on 27 volunteers, is appliedto evaluate the selected cut-off value (0.1 �g/mL) for EtG andEtS and to estimate the sensitivity and specificity of the method.This method could be used for future surveillance of abstinencein the context of driving licence regranting [7,8] and for thedetection of alcohol intake prior death in selected post-mortemcases [10–12].
2. Experimental
2.1. Chemicals
Ethyl glucuronide (EtG), ethyl sulphate (EtS) and their pen-tadeuterated analogues (EtG-d5 and EtS-d5) were obtained fromSigma–Aldrich (Steinheim, Germany) as a methanolic 1 mg/mLsolution. ULC/MS grade acetonitrile, methanol and 0.1% formicacid in water were purchased from Biosolve (Valkenswaard, TheNetherlands). Blank urine was purchased from Bio-Rad Laborato-ries (Nazareth Eke, Belgium).
2.2. Standard solutions, calibrators and quality control samples(QC)
Two stock solutions, one for calibration (Cal-Stock) and one forinternal quality controls (QC-Stock), with EtG and EtS at a concen-tration of 20 �g/mL were prepared in methanol. The stock solutionwith internal standards (IS) at a concentration of 4 �g/mL was pre-pared in methanol. All solutions were stored at −18 ◦C.
Daily calibration working solutions (Cal-WS) with concentra-tions of 0.1, 0.5 and 10 �g/mL were prepared diluting the Cal-Stocksolution. Calibrators (0.1, 0.25, 0.5, 2.5, 5.0 and 10 �g/mL) were pre-pared by spiking 30 �L of the IS solution to 50 �L of commercialblank urine, an adequate amount of Cal-WS, and methanol until atotal volume of 280 �L was reached.
Daily quality control working solutions (QC-WS) with concen-tration of 0.5 and 5 �g/mL were prepared diluting the QC-Stocksolution. Internal quality controls (0.3, 4 and 7.5 �g/mL) were pre-pared spiking 30 �L of IS to 50 �L of commercial blank urine, anadequate amount of QC-WS and methanol until a total volume of280 �L.
External quality controls Medidrug ETG 1/10-B and MedidrugETG 2/09-B were both purchased from Medichem (Steinenbronn,Germany). Proficiency tests for EtG and EtS in urine organizedby the German Society of Toxicological and Forensic Chemistry(GTFCh) were also performed for quality control purposes.
2.3. Population study
A prospective alcohol self-monitoring study was performed ask-ing 27 volunteers to declare their exact alcohol consumption perday during the 5 days preceding the sampling. Urine samples werecollected in 100 mL urine containers from Sarstedt (Nümbrecht,Germany), transferred into 4 mL Greiner bio-one tubes (Fricken-grasen, Germany) and stored at 2–8 ◦C until analysis. Samples were
analyzed within 5 days after collection.
EtG100 and EtS100 concentration were calculated by normal-izing the measured EtG and EtS to a creatinine concentration of100 mg/dL [2].
atogr
2
ato(or
s
2
XeaCwoAsAfT
p(cvew0pt
2
ii
fwtwat
eupses
TM
N. Kummer et al. / J. Chrom
.4. Sample preparation
Methanol (250 �L) and 30 �L of the IS solution (4 �g/mL) weredded to 50 �L of urine. After precipitation, the sample was cen-rifuged at 14,000 rpm (20,800 × g) during 10 min at 4 ◦C. 250 �Lf the supernatant was transferred to a total recovery glass vialWaters, Zellik, Belgium) and evaporated to dryness under a streamf nitrogen using a heated metal block at 38 ◦C. The residue waseconstituted in 300 �L of 0.1% formic acid in water.
For each authentic sample, an additional 1/1000 dilution wasystematically performed.
.5. Liquid chromatography and mass spectrometry conditions
Analyses were performed on an Aquity UPLC coupled to aevo TQ MS tandem mass spectrometer (Waters, Manchester, UK)quipped with an electrospray ionization source operated in neg-tive mode. Gradient elution was performed on an Acquity UPLCSH C18 (2.1 × 100 mm, 1.8 �m) column (Waters, Milford, MA, USA)ith 0.1% formic acid in water (A) and acetonitrile (B) at a flow rate
f 300 �L/min. The gradient elution started with 99.3% of solution for 2.4 min, decreasing to 40% of solution A at 3.0 min, to 20% ofolution A at 4.4 min. The washing step contains only 2% of solution
and holds from 4.41 to 5.40 min. The initial condition is appliedrom 5.41 min to 7 min. The column temperature was set at 55 ◦C.he injection volume is 5 �L using full-loop mode.
For the MS/MS detection, following parameters were used: tem-erature of source gas (nitrogen) was 150 ◦C, desolvatation gasnitrogen) flow was 900 L/h at 650 ◦C, capillary voltage was 1 KV,one voltage was 26 V with a cone gas flow at 40 L/h, multiplieroltage was 508.42 V, extractor voltage was 3 V, MS mode detectionnergy was 2, MS/MS mode detection energy was 10, ion energy Ias 0.5 V, ion energy II was 0.5 V, entrance and exit potential was
.5 V and collision gas (argon) flow was 0.35 mL/min. Detection waserformed in the multiple reaction monitoring mode (MRM) usinghe appropriate parameters for each compound (Table 1).
.6. Method validation
Selectivity, sensitivity, matrix effects, extraction efficiency, lim-ts, linearity, accuracy and stability were evaluated according tonternational guidelines [45].
To study endogenous interferences, six blank urine samplesrom different individuals were analyzed. To verify that thereere no isotope exchange reactions with non-labelled compounds,
wo zero samples (blank urine spiked with internal standard)ere analyzed. According to the EMEA guideline, interferences are
cceptable in our type of method as long as the signal was lowerhan 20% of the response at the LOQ.
Matrix effects are quantified and evaluated using the post-xtraction addition technique [46]. For the matrix effect, six blankrine samples from different persons were spiked after the sam-
le preparation and compared with compounds spiked at theame theoretical concentration in the mobile phase. Extractionfficiency is evaluated comparing responses of six blank urineamples spiked before sample preparation with responses of six
able 1RM transitions and conditions for EtG, EtS and their deuterated analogues.
Precursor ion (m/z) Product ion (m/z)
EtG (Quantifier) 220.96 74.95
EtG (Qualifier) 220.96 84.96
EtG-d5 225.97 84.90
EtS (Quantifier) 124.90 80.10
EtS (Qualifier) 124.90 96.88
EtS-d5 129.97 97.85
. B 929 (2013) 149– 154 151
blank urine samples spiked after sample preparation. These exper-iments were done at low (0.3 �g/mL), medium (4 �g/mL) and high(7.5 �g/mL) concentration.
The limit of detection (LOD) is determined by analysingdecreasing concentrations of the analytes. The LOD is the lowestconcentration of the analyte for which the signal-to-noise ratio ofboth transitions is at least 3/1.
The limit of quantification (LOQ) is the lowest concentrationof analyte with a signal-to-noise ratio greater than 10/1 for bothtransitions and for which the bias and precision deviation is lessthan 20%. Other identification criteria, such as a stable ion ratio(RSD% < 20) between the quantifier and the qualifier also had to bereached.
The calibration model (n = 6) was tested over the range 0.1, 0.25,0.5, 2.5, 5 and 10 �g/mL. Calibration model and weighting factorwere evaluated for each compound. The goodness of fit was estab-lished as the difference between the calculated calibrator value andits nominal value. The variation coefficient should be lower than15% except for the LOQ (<20%).
Accuracy is measured through the determination of trueness(bias) and precision (repeatability and intermediate precision).Three internal quality controls at low (0.3 �g/mL), medium(4 �g/mL) and high (7.5 �g/mL) concentration and two externalquality controls (ETG 1/10-B and EtG 2/09-B) were analyzed inreplicates on 8 different days. A single factor ANOVA test with sig-nificance level (˛) of 0.05 allows calculating bias, repeatability andintermediate precision with these data. The results are acceptablewhen they are less than 15% (20% for the LOQ). The reproducibil-ity is evaluated by participation in proficiency tests organized byGTFCh.
Freeze/thaw stability, processed sample stability and longterm storage stability are evaluated at low (0.3 �g/mL) and high(7.5 �g/mL) concentration. The mean of the stability should bewithin 90–110% of the mean of the control samples and the 90%confidence interval of the stability sample results should be within±20% of the control samples.
3. Results and discussion
3.1. Method validation
The method was validated for selectivity, sensitivity, matrixeffects, extraction efficiency, limits, linearity, accuracy and stability.
Identification of compounds was based on retention time andon the presence of a stable ratio between the two MRM transitions(<20%). It is well known that EtG and EtS can be present in smallamounts in urine even without voluntary consumption of alcohol[18,19,25–27].
In one blank urine sample (n = 8) EtG has been detected, butthe calculated concentration (0.04 �g/mL) was below the LOQ(0.1 �g/mL). By a combination of retention time and stable ratio
between the qualifier and quantifier, no interfering signal for EtShas been detected in blank urine samples.
Results of matrix effects and extraction efficiency are presentedin Tables 2 and 3.
Table 3Matrix effect and extraction efficiency for EtS.
EtS Low Medium High
Matrix effect calculated as %recovery (RSD %)
106 (9) 95 (8) 88 (3)
Matrix effect (% recovery) 108 (7) 96 (7) 113 (6)
Esa(tE8
0
E
((
ptm
concentration in urine is below 0.1 �g/mL and EtG concentration
F1
compensated by IS (RSD %)Extraction efficiency % (RSD %) 76 (5) 81 (7) 80 (6)
A reproducible ion suppression of less than 24% is observed fortG. The use of EtG-d5 as IS compensated for the matrix effect. Noignificant matrix effects were observed for EtS. Using LC systemsnd simple dilution of urines as sample preparation, matrix effect% recovery) up to 69, 80 and 171% were reported for EtG and upo 94, 110, and 179% for EtS [11,30,43]. The extraction efficiency oftG and EtS is reproducible, concentration independent and about0%.
The LOQ was 0.1 �g/mL for EtG and for EtS (Fig. 1). The LOD was.06 �g/mL for EtG and 0.08 �g/mL for EtS.
For EtG, a weighted (1/x) linear regression line was applied, fortS a 1/x2 weighting was necessary.
The bias of the method is lower than 15%. The repeatabilityRSDr) and intermediate precision (RSDt) are acceptable, with RSD%) lower than 10% (Table 4).
No instability was observed for samples staying in the autosam-
ler during 24 and 72 h. Moreover, EtG and EtS were stable afterhree freeze/thaw cycles, after 2 months at −20 ◦C and after 2
onths at 4 ◦C.
ig. 1. MRM Chromatogram for EtG (m/z 221 → 75 (A), m/z 221 → 85 (B)) and EtG-d5 (30 → 98 (F)) at the LOQ (0.1 �g/mL).
. B 929 (2013) 149– 154
The reproducibility of the method was evaluated via analy-sis of four proficiency tests (EtG 3/11, EtG 1/12, EtG 2/12 andEtG 3/12) organized by the German Society of Toxicological andForensic Chemistry (GTFCh). The z-scores obtained for EtG were−0.22, −0.46, 0.10, 0.23 for EtG 3/11 (1.450 �g/mL), EtG 1/12(0.800 �g/mL), EtG 2/12 (0.556 �g/mL) and EtG 3/12 (0.832 �g/mL),respectively. For EtS, z-scores were 0.57, −0.86, 0.23 and −0.16for EtG 3/11 (0.885 �g/mL), EtG 1/12 (1.100 �g/mL), EtG 2/12(1.070 �g/mL) and EtG 3/12 (0.899 �g/mL), respectively.
Using an Acquity CSH C18 column on an UPLC–ESI-MS/MS sys-tem equipped with a triple quadrupole, allows to fulfil forensicanalysis requirements and to reach an LOQ for EtG and EtS of0.1 �g/mL.
3.2. Population study
Twenty-seven urines samples from volunteers were analyzed.Urine samples from volunteers (n = 14) who did not drink alcoholicbeverages the day before the sampling were all negative for EtGand EtS using a cut-off at 0.1 �g/mL.
In 10 samples from volunteers who declared having con-sumed alcohol the day before the sampling (n = 13) a concentrationbetween 0.5 and 101.9 �g/mL (mean 10.9, median 1.4) for EtG100and between 0.1 and 37.9 �g/mL (mean 3.6, median 0.3) for EtS100was determined (Fig. 2).
In one case (Fig. 2A), no EtS (EtS100 < 0.1 �g/mL andEtG100 = 1.4 �g/mL) was detected after the consumption oftwo alcohol units. In another case (Fig. 2B), EtG was not detected24 h after the ingestion of 1 glass of alcohol (EtG100 < 0.1 �g/mLand EtS100 = 0.1 �g/mL). In one sample (Fig. 2C), neither EtG norEtS (EtG100 and EtS100 < 0.1 �g/mL) were detected 24 h after theconsumption of one glass of alcohol. A recent study [9] has demon-strated that after the consumption of two units of alcohol, EtS
is above 0.1 �g/mL in 6 out of 7 cases. In 2 out of 5 cases, EtGconcentration is above 0.1 �g/mL in urine 24 h after a consumptionof 1 glass of alcohol.
m/z 226 → 85 (C)) and EtS (m/z 125 → 80 (D), m/z 125 → 97 (E)) and EtS-d5 (m/z
N. Kummer et al. / J. Chromatogr. B 929 (2013) 149– 154 153
Table 4Trueness (bias) and precision (repeatability (RSDr) and intermediate precision (RSDt)).
EtG EtS
Nominal value (�g/mL) RSDr (%) RSDt (%) Bias (%) Nominal value (�g/mL) RSDr (%) RSDt (%) Bias (%)
ig. 2. EtG and EtS concentrations normalized to 100 mg/dL creatinine in subjectsho declared having been drinking alcohol 24 h before the sampling.
One volunteer (Fig. 2D), who has declared a consumption ofve alcohol units the day before the sampling, has no EtG and notS in urine (EtG100 and EtS100 < 0.1 �g/mL). The creatinine concen-ration measured in that sample was abnormally low (12 mg/dL).n Germany, a urine creatinine concentration below 20 mg/dL iseclared as “not useable” for analysis [47]. This abnormal dilutionf urine can explain the absence of EtG and EtS in this sample [43].
The three subjects who have declared a consumption of alcohol2, 4 and 6 glasses) 2 days before the sampling and no consumptionhe day before were all negative for EtG and EtS. Kinetic studieshow that EtG and EtS are detectable in urine up to 24 h after intakef about 2 units of alcohol and up to 48 h after intake of about 4 unitsf alcohol [2,13–18].
EtG and EtS were in agreement in 25 out of 27 cases. EtG and EtSoncentration in urine were highly correlated (r = 0.996, p < 0.001).
moderate correlation between the number of drinks the dayefore the sampling and the concentration of EtG (r = 0.448, p < 0.02)nd EtS (r = 0.406, p < 0.04) in urine was observed. This result cane explained by the high inter-individual variation of EtG and EtSoncentration in urine after the consumption of equal amounts ofthanol [17].
Using a cut-off at 0.1 �g/mL, this method is able to detect alcoholonsumption approximately 24 h after the intake, without show-ng any false positive results. A cut-off at 0.1 �g/mL for EtG andt 0.05 �g/mL for EtS has been proposed to exclude the repeatedntake of alcohol [9]. However, knowing that some blank urinesamples contain a small amount of EtG and EtS, a cut-off valuef 0.1 �g/mL for EtG and EtS was chosen to distinguish betweenntentional and unintentional alcohol consumption.
. Conclusion
This report describes a validated method for the quantificationf EtG and EtS in urine by UPLC–ESI-MS/MS using protein precip-tation as clean-up step. The chromatographic run time for one
analysis is 7 min. The recovery was around 80% for both compoundsand the matrix effect calculated as % recovery was on average 80%for EtG and on average 96% for EtS. This method provides good pre-cision (RSDr and RSDt < 10%) and bias (<15%). The validity of themethod was confirmed by successful participation to four profi-ciency tests (z-scores of less than 1). To avoid false positive results,a cut-off value at 0.1 �g/mL should be applied for both analytes.
Acknowledgment
The authors thank all volunteers for their collaboration.
References
[1] N. Stephanson, H. Dahl, A. Helander, O. Beck, Ther. Drug Monit. 24 (2002) 645.[2] A. Helander, O. Beck, J. Anal. Toxicol. 29 (2005) 270.[3] F.M. Wurst, C. Kempter, J. Metzger, S. Seidl, A. Alt, Alcohol 20 (2000) 111.[4] A. Helander, M. Böttcher, C. Fehr, N. Dahmen, O. Beck, Alcohol Alcohol. 44 (2009)
55.[5] G.E. Skipper, W. Weinmann, A. Thierauf, P. Schaefer, G. Wiesbeck, J.P. Allen, M.
[9] M. Albermann, F. Musshoff, E. Doberentz, P. Heese, M. Banger, B. Madea, Int. J.Legal Med. (2012) 1.
10] L. Politi, L. Morini, F. Mari, A. Groppi, E. Bertol, Int. J. Legal Med. 122 (2008) 507.11] W. Bicker, M. Lämmerhofer, T. Keller, R. Schuhmacher, R. Krska, W. Lindner,
Anal. Chem. 78 (2006) 5884.12] Al-Asmari, J. Anal. Toxicol. 34 (2010) 261.13] H. Dahl, N. Stephanson, O. Beck, A. Helander, J. Anal. Toxicol. 26 (2002) 201.14] M.H. Wojcik, J.S. Hawthorne, Alcohol Alcohol. 42 (2007) 317.15] G. Høiseth, J.P. Bernard, N. Stephanson, P.T. Normann, A.S. Christophersen, J.
Mørland, A. Helander, Alcohol Alcohol. 43 (2008) 187.16] G. Høiseth, J.P. Bernard, R. Karinen, L. Johnsen, A. Helander, A.S. Christophersen,
J. Mørland, Forensic Sci. Int. 172 (2007) 119.17] C. Halter, S. Dresen, V. Auwaerter, F. Wurst, W. Weinmann, Int. J. Legal Med.
122 (2008) 123.18] F. Musshoff, E. Albermann, B. Madea, Int. J. Legal Med. 124 (2010) 623.19] A. Thierauf, H. Gnann, A. Wohlfarth, V. Auwärter, M.G.E. Perdekamp, K.-J. But-
tler, F.M. Wurst, W. Weinmann, Forensic Sci. Int. 202 (2010) 82.20] A. Thierauf, C.C. Halter, S. Rana, V. Auwaerter, A. Wohlfarth, F.M. Wurst, W.
Weinmann, Addiction 104 (2009) 2007.21] F.M. Wurst, S. Dresen, J.P. Allen, G. Wiesbeck, M. Graf, W. Weinmann, Addiction
101 (2006) 204.22] G. Høiseth, R. Karinen, A.S. Christophersen, L. Olsen, P.T. Normann, J. Mørland,
Forensic Sci. Int. 165 (2007) 41.23] K. Borucki, R. Schreiner, J. Dierkes, K. Jachau, D. Krause, S. Westphal, F.M. Wurst,
C. Luley, H. Schmidt-Gayk, Alcohol. Clin. Exp. Res. 29 (2005) 781.24] F.M. Wurst, S. Seidl, D. Ladewig, F. Müller-Spahn, A. Alt, Addict. Biol. 7 (2002)
427.25] T.P. Rohrig, J. Anal. Toxicol. 30 (2006) 703.26] G. Høiseth, B. Yttredal, R. Karinen, H. Gjerde, A. Christophersen, J. Anal. Toxicol.
34 (2010) 84.27] A. Costantino, E.J. DiGregorio, W. Korn, S. Spayd, F. Rieders, J. Anal. Toxicol. 30
(2006) 659.28] A. Thierauf, A. Wohlfarth, V. Auwärter, M.G.E. Perdekamp, F.M. Wurst, W. Wein-
mann, Forensic Sci. Int. 202 (2009) e45.29] I. Janda, A. Alt, J. Chromatogr. B: Biomed. Sci. Appl. 758 (2001) 229.30] J. Beyer, T. Vo, D. Gerostamoulos, O. Drummer, Anal. Bioanal. Chem. (2011) 1.31] S. Dresen, W. Weinmann, F.M. Wurst, J. Am. Soc. Mass Spectrom. 15 (2004)
[[[44] R. Agius, T. Nadulski, H.-G. Kahl, B. Dufaux, Forensic Sci. Int. 218 (2012) 10.
54 N. Kummer et al. / J. Chrom
32] F. Wurst, M. Yegles, C. Alling, S. Aradottir, J. Dierkes, G. Wiesbeck, C. Halter, F.Pragst, V. Auwaerter, Int. J. Legal Med. 122 (2008) 235.
33] W. Weinmann, P. Schaefer, A. Thierauf, A. Schreiber, F.M. Wurst, J. Am. Soc.Mass Spectrom. 15 (2004) 188.
34] G. Schmitt, R. Aderjan, T. Keller, M. Wu, J. Anal. Toxicol. 19 (1995) 91.35] I.Á. Freire, A.M.B. Barrera, P.C. Silva, M.J.T. Duque, P.F. Gómez, P.L. Eijo, J. Appl.
Toxicol. 28 (2008) 773.
36] F.A. Esteve-Turrillas, W. Bicker, M. Lämmerhofer, T. Keller, W. Lindner, Elec-
trophoresis 27 (2006) 4763.37] B. Jung, J. Caslavska, W. Thormann, J. Chromatogr. A 1206 (2008) 26.38] L. Krivánková, J. Caslavska, H. Malásková, P. Gebauer, W. Thormann, J. Chro-
matogr. A 1081 (2005) 2.
[[
[
. B 929 (2013) 149– 154
39] M. Nováková, L. Krivánková, Electrophoresis 29 (2008) 1694.40] M. Albermann, F. Musshoff, B. Madea, Anal. Bioanal. Chem. 396 (2010) 2441.41] P. Cabarcos, H.M. Hassan, M.J. Tabernero, K.S. Scott, J. Appl. Toxicol. (2012),
http://dx.doi.org/10.1002/jat.1791.42] L. Rivier, Anal. Chim. Acta 492 (2003) 69.43] M.E. Albermann, F. Musshoff, B. Madea, J. Chromatogr. Sci. 50 (2012) 51.
45] S. Wille, F. Peters, V. Di Fazio, N. Samyn, Accredit. Qual. Assur. 16 (2011) 279.46] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003)
3019.47] T. Arndt, Forensic Sci. Int. 186 (2009) 48.
