General unknown screening in hair by liquid chromatography–hybrid quadrupole time-of-flight mass spectrometry (LC–QTOF-MS)

Forensic Science International 218 (2012) 68–81

Contents lists available at SciVerse ScienceDirect

Forensic Science International journal homepage: www.elsevier.com/locate/forsciint

General unknown screening in hair by liquid chromatography–hybrid quadrupole time-of-flight mass spectrometry (LC–QTOF-MS)§ Sebastian Broecker, Sieglinde Herre, Fritz Pragst * Institute of Legal Medicine, University Hospital Charite´, Turmstraße 21, Building N, 10559 Berlin, Germany

A R T I C L E I N F O

A B S T R A C T

Article history: Received 12 April 2011 Accepted 20 May 2011 Available online 27 October 2011

The retrospective investigation of the exposure to toxic substances by general unknown screening of hair is still a difficult task because of the large number of possible poisons, the low sample amount and the difficult sample matrix. In this study the use of liquid chromatography–hybrid quadrupole time-of-flight mass spectrometry (LC–QTOF-MS) was tested as a promising technique for this purpose. In the optimized procedure, 20 mg hair were decontaminated with water and acetone and two times extracted by 18 h incubation with 0.5 ml of a mixture of methanol/acetonitrile/H2O/ammonium formate at 37 8C. A mixture of deuterated standards from different drug groups was added for quantification and method control. The united extracts were evaporated to a residue of 0.5 ml and 5 ml were injected without clean-up for LC–QTOFMS measurement (instrument Agilent 6530) with positive electrospray ionization and in data dependent acquisition mode. For peak identification the accurate mass data base and spectral library of the authors was used which contains accurate mass CID spectra of more than 2500 and theoretically calculated accurate mass data of more than 7500 toxicologically relevant substances. Validation at the example of 24 illegal drugs, their metabolites and benzodiazepines resulted in limits of detection of 0.003–0.015 ng/mg, and limits of quantification of 0.006–0.021 ng/mg with good accuracy and intra- and interday reproducibility. The matrix effect by ion suppression/enhancement was 72–107% for basic drugs and 42–75% for benzodiazepines. Yields of the hair extraction above 90% were determined for 59 drugs or metabolites. The method was applied to hair samples from 30 drug fatalities and from 60 death cases with known therapeutic drug intake at life time. Altogether 212 substances were identified with a frequency per drug of 1–40 (mean 4.2) and per case of 2–33 (mean 10.2), between them 35 illegal drug related substances and 154 therapeutic drugs. Comparison with the data known from case histories and from the analysis of blood, urine and gastric content showed only a low agreement, with many unexpected drugs detected and many reported drugs not detected in hair. Basic drugs and metabolites such as opioides, cocaine, amphetamines, several groups of antidepressants, neuroleptics, beta-blockers or the metamizole metabolite noramidopyrine were found with high frequency whereas acidic and several neutral drugs such as cannabinoids, salicylic acid, furosemide, barbiturates, phenprocoumone or cardiac glycosides could not be detected with sufficient sensitivity, mainly because of the low ion yield of positive ESI for these compounds. The advantage of a comprehensive acquisition of all substances is paid by a lower sensitivity in comparison to targeted screening LC–MS/MS procedures. In conclusion, the procedure of sample preparation and LC–QTOF-MS analysis proved to be a robust and sensitive routine method in which the qualitative screening for a wide variety of toxic substances in hair is combined with the quantitative determination of selected illegal drugs. ß 2011 Elsevier Ireland Ltd. All rights reserved.

Keywords: CID mass spectra library General unknown screening Hair analysis LC–QTOF-MS Drug abuse

1. Introduction The most important difference of hair as a sample material for toxicological analysis in comparison to other human matrices is the much longer time window which allows the retrospective

§ This paper is part of the special issue entitled: Selected papers from the Chamonix 2011 Society of Hair Testing Meeting, Guest-edited by Pascal Kintz. * Corresponding author. Tel.: +49 30 450 525041; fax: +49 30 450 525904. E-mail address: [email protected] (F. Pragst).

0379-0738/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2011.10.004

detection of chronic exposure to drugs or poisons up to years back [1]. By segment-wise analysis even the approximate time resolved history of the exposure can be demonstrated. Procedures in hair analysis were described for many illegal and therapeutic drugs and frequently reviewed in papers and books [2–5]. They concern mostly single compounds or limited groups of compounds. However, there are cases in which the chronic exposure to any drugs or poisons shall be elucidated without specific information about the substances and whether a substance was involved at all. Examples are chronic criminal administration of drugs or poisons, e.g. for sedation of children or elderly under care, recreational

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

substance abuse with forensic or clinical background, chronic medical treatment in death cases or clinical cases with unknown history. A drug screening in hair can also contribute to the identification of unknown corpses. In such cases, a systematic toxicological analysis (STA), that means the general search for toxic substances in the hair sample, should be performed. From analytical point of view hair differs from other human materials by its lower sample amount which requires more sensitive methods, and by its solid state which requires specific methods of extraction or digestion and makes it difficult to prepare realistic standards and control samples for calibration and validation of quantitative procedures. Apart from these specific features, STA in hair follows the same principles as in blood or urine: preparation of an extract which includes as many as possible toxicologically relevant substances and excludes matrix constituents as far as possible, chromatographic separation of the extract components, characterization of the components by retention time and molecular spectrometry (MS or UV), identification based on library search with these data and (approximate) quantification using the chromatographic peak area. Because of the limitations of hair mentioned above, STA in this material was only seldom described. A procedure by combined use of high performance liquid chromatography with diode array detector (HPLC-DAD) and gas chromatography mass spectrometry (GC–MS) after extraction of 200–250 mg hair in two fractions by 0.1 M HCl (basic drugs) and water (neutral and acidic drugs) and clean-up by solid-phase extraction (SPE) was described by Gaillard and Pepin in 1997 [6]. An overview of screening strategies in hair was given by Sachs [7] in 2006. It was concluded that a real general unknown screening for thousands of substances would not be available with the required sensitivity. Instead, a multianalyte procedure using LC–MS/MS which included 49 antidepressants and neuroleptics, 23 benzodiazepines and 20 opioides was seen as the most promising alternative [7,8]. The availability of time-of-flight mass spectrometers with increased mass resolution and mass accuracy offered new possibilities also for a real STA in hair. Besides the advantage to obtain the molecular formula of an unknown peak from the accurate ion mass and the isotope peak pattern, the principle of TOF-MS enables a comprehensive acquisition of all components of the injected hair extract without any predetermination if they are sufficiently ionized by the ion source of the instrument. Of course, this is limited by the efficiency of the sample preparation to include the broad spectrum of substances deposited in hair. The application of liquid chromatography–time-of-flight mass spectrometry for screening of hair samples was reported by Pelander et al. [9], Liotta et al. [10], and Klose Nielsen et al. [11]. For sample preparation hydrolysis with 1 M NaOH and subsequent SPE [9], incubation of hair with 0.1 M HCl and subsequent liquid/liquid extraction with chloroform/isopropanol [10], or 18 h incubation of the hair pieces with a 25:25:50 (v/v/v) mixture of methanol:acetonitrile:2 mM ammonium formate containing 8% acetonitrile without further clean-up [11] were used. Because of the large number of possible isomers, further evidence in addition to the molecular formula is needed for unambiguous identification of a substance by LC–TOF-MS. For this purpose theoretical and experimental libraries of toxic compounds were used which contained retention times [9,11] or used a metabolomic approach to distinguish between proposed candidates [10]. The screening procedure described by Klose Nielsen et al. was fully validated also for the quantitative determination of 52 pharmaceuticals and drugs of abuse using four deuterated drugs and dibenzepine as internal standards [11]. Time-of-flight mass spectrometry was also used in combination with gas chromatography for hair screening. Gutherya et al. developed a technique using comprehensive twodimensional gas chromatography/time-of-flight mass spectrometry

69

for qualitative analysis of hair extracts after SPE and derivatization [12]. Although this procedure was only applied to three hair samples and 27 drugs or metabolites it should be applicable for real screening purposes. It was shown in previous papers for blood and urine that the accuracy of substance identification in STA by LC–TOF can be much increased by the measurement of accurate mass collision induced dissociation (CID) fragment spectra in a hybrid quadrupole time-of flight mass spectrometer (LC–QTOF-MS) operated in data dependent acquisition mode [13–15]. The principle of this technique was described in detail in [13]. The instrument is operated with steady alternation of MS and MS/MS mode with a cycle time of, for instance, 1.1 s. In MS mode, for 0.33 s the full MS spectrum at the corresponding retention time is recorded, two precursor ions are selected and for each of them the mass dependent collision energy is chosen. After that, the quadrupole selects in succession these two masses and the CID accurate mass spectra are measured during the residual time of the measurement cycle. These two masses are excluded from MS/MS measurement for e.g. 0.15 min in order to enable the acquisition of other co-eluting substances, and the instrument goes on to the next cycle. The identification of a peak is based on MS and MS/MS data and uses the ‘‘Personal Compound Database and Library of Toxic Compounds’’ which contains theoretically calculated accurate mass data and molecular formulas of more than 7500 toxicologically relevant substances and accurate mass collision induced dissociation (CID) spectra of more than 2500 of these substances. A similar approach for STA was also applied to hair samples by Liu et al. [16] who used a hybrid triple-quadrupole linear ion trap mass spectrometer and an in-house spectra library of about 800 toxic substances. The product ions from a specific precursor ion generated by CID were accumulated in the ion trap for adequate intensity before performing the product ion scan. In the present study a method for STA in hair by LC–QTOF-MS was developed and applied to samples from 90 death cases with known exposure to illegal or therapeutic drugs during life time. The results are compared with data from case histories and toxicological investigation of blood, urine and gastric contents. 2. Material and methods 2.1. Chemicals, reagents and reference substances The solvents and chemicals used for the mobile phase were purchased as follows: methanol (LC–MS grade), acetonitrile (LC–MS grade), and ammonium acetate (HPLC grade) from Fisher scientific (Schwerte, Germany), ammonium formate (LC–MS grade) from Agilent Technologies, water (HPLC grade) and formic acid (99+% for analysis) from Acros Organics (Geel, Belgium). All other solvents and reagents used for sample preparation were obtained from Merck (Darmstadt, Germany) in analytical grade purity. Non-deuterated and deuterated standards of illegal and therapeutic drugs, which were used for calibration and validation purposes were purchased from LGC Promochem (Wesel, Germany).

2.2. Instruments and software A 6530 Accurate-Mass LC–QTOF-MS device (Agilent Technologies, Santa Clara, USA) was used for all measurements. The Agilent 1200 SL series HPLC device consisted of a degasser, a thermostated HiP-ALS autosampler, a binary pump Bin Pump SL, and a TCC SL column oven. The QTOF-MS instrument was operated with an electrospray ion source ESI + Agilent Jet Stream Technology in positive ionization mode, a quadruple for isolation of precursor ions with a mass window of 4 m/z in MS/MS mode, a linear hexapole collision cell with nitrogen as collision gas and collision energy 0–40 eV, a TOF-MS with a mass accuracy <3 ppm, mass resolution of 5000–10,000 (100–922 m/z), a measuring frequency of 10,000 transients/s and a detection frequency of 2 GHz (200,000 points/transient). The measurements and post-run analyses were controlled by the software MassHunter Acquisition B.02.01 with Service Pack 3 for the Agilent TOF and QTOF and MassHunter Qualitative Analysis B.03.01 with Service Pack 3. For database and library search, the Personal Compound Database and Library Software B.03.01 was used in combination with the ‘‘Personal Compound Database and Accurate Mass Spectral Library’’ of the authors which was described previously and contains

70

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

accurate mass CID spectra of more than 2500 and theoretically calculated accurate mass data of more than 7500 toxicologically relevant substances [13,15]. For sample preparation a thermomixer TMix 220 (Analytik Jena AG, Germany) and an ultrasonic bath Sonorex RK 103H (Bandelin electronic, Berlin Germany) were applied. 2.3. Hair samples and toxicological data for comparison Drug free hair samples for method optimization and validation were collected from the laboratory stuff and their relatives who assured not to take any drugs with the exception of drinking coffee or tea. Postmortem scalp hair samples were obtained from 30 illegal drug cases and 60 death cases with known therapeutic drug intake at life time which were investigated in the Institute of Legal Medicine of the University Hospital Charite´ Berlin between July 2009 and April 2010. The data about case histories were obtained from the police reports of investigation, including medical records, prescribed or administered drugs, or drugs found at the scene of death. From the illegal drug cases 21 died from drug overdose and 9 from another cause of death with drug abuse known in 5 cases or only detected during toxicological analysis in 4 cases. From the other 60 cases 16 had died in hospital during or after surgery (4), after an accident (5), after an acute illness (1), or after chronic diseases (6). 3 died in a nursing home. 15 were found dead in their apartments with a chronic disease known. 26 cases were suicides with a psychiatric history (19) or after serious diseases (7). From the suicide cases, 12 died by therapeutic drug overdose, 2 by bleeding to death, 3 by hanging, 2 by suffocation under a plastic bag, 1 by electricity, 6 by fall out of a window and 2 were run over by a train. Blood, urine and gastric content of these cases were submitted to toxicological analysis by enzyme immunoassay (EMIT), HPLC-DAD and GC–MS according to validated routine procedures used in the institute. The hair samples were cut directly above the skin at the back of the head and were stored under dry conditions at room temperature. 2.4. Sample preparation In case of longer hair samples, the proximal segment 0–6 cm was analyzed. Shorter hair samples were analyzed in full length. After optimization, the hair samples were prepared according to the following procedure: The hair was decontaminated by gentle shaking for 1 min in water and two times for 1 min in acetone. After drying on a filter paper it was cut to 1–2 mm pieces and about 20 mg were exactly weighed in a 1.5 ml Eppendorf vial. After addition of a mixture of deuterated standards (5 ml of each 1 ng/ml amphetamine-d5, methamphetamined5, MDMA-d5, MDE-d6, MDA-d5, cocaine-d3, benzoylecgonine-d3, methylecgonine-d3, cocaethylene-d3, morphine-d3, 6-acetylmorphine-d3, codeine-d3, dihydrocodeine-d3, methadone-d9, EDDP-d3, alprazolam-d5, alpha-hydroxyalprazolam-d5, diazepam-d5, nordazepam-d5, oxazepam-d5, 7-aminoflunitrazepam-d7, flunitrazepam-d7 and lorazepam-d4) the hair was incubated for 18 h with 0.5 ml of a mixture of methanol/acetonitrile/2 mM ammonium formate (25:25:50, v/v/v) with gentle shaking at 37 8C. Then the mixture was centrifuged for 5 min at 13,200 rpm. The liquid phase was separated and the incubation of the hair pieces was repeated for 18 h with another 0.5 ml of the solvent mixture. Both extracts were united and evaporated in a nitrogen stream to a residue of 0.5 ml in order to remove the most of the organic solvents. 5 ml of the residue was injected for LC–QTOF-MS measurement without further clean-up procedures. 2.5. Measurement by LC–QTOF-MS The chromatographic separation was performed with a column Zorbax Eclipse plus C18, 2.1 mm  100 mm, 3.5 mm (Agilent Technologies) at 30 8C with the eluents A = 10 mM NH4Ac in H2O and B = methanol and the following time program of the gradient: 0 min 10% B, linear to 90% B at 16 min, const. 90% B to 19.8 min, back to 10% B at 20 min and equilibration for 3 min. The flow rate was 0.4 mL/min. The QTOF-MS instrument was operated under the following conditions: Ion source ESI + Agilent Jet Stream Technology in positive ionization mode, quadrupole was used as an ion guide in MS mode and for selection of precursor ions with Dm/ z = 4 in MS/MS mode, collision cell without CID in MS mode and with CID of precursor ions in MS/MS mode at mass dependent ramped CID energy (offset 4 eV, slope 6 eV/100 m/z), TOF-MS with a mass range of 100–1000 m/z in MS mode and 50–600 m/z in MS/MS mode. The scan rate was 3 Hz in MS and MS/MS experiments. The source parameters were: gas temperature 300 8C, gas flow 8 L/min, nebulizer pressure 35 psi, sheath gas temperature 350 8C, sheath gas flow 11 L/min, VCap voltage 3000 V, nozzle voltage 0 V and fragmentor voltage 150 V, reference ions for mass calibration: purine 121.050873 [M+H]+ and 119.036319 [M H] , HP921 = hexakis (1H,1H,3H-tetrafluoropropoxy)phosphazine 922.009798 [M+H]+ and 966.000725 [M+HCO2] . For systematic toxicological analysis the Auto-MS/MS mode (data dependent acquisition) was used with a cycle time of 1.1 s, 0.33 s measurement in MS mode, selection of 2 precursors which were fragmented in the following two MS/MSexperiments and active exclusion after 1 spectrum for 0.15 min.

2.6. Validation For a selection of 24 frequently occurring illegal drugs or metabolites (Table 1), a calibration and validation of the method was performed according to international guidelines [17] using the software Valistat 2.0 of the German Society of Toxicological and Forensic Chemistry (GTFCh) [18]. For this purpose 20 mg of 6 different drug free hair samples were spiked with these drugs at different concentrations and 5 ng of the deuterated standards and analyzed according to the procedure described in Sections 2.4 and 2.5. The method was validated for the following parameters: sensitivity (limits of detection LODs and limits of quantification LOQs), linearity and carryover, intra- and inter-day precision and accuracy, specificity, and matrix effect. The data are shown in Table 1. 2.6.1. Sensitivity 20 mg drug free hair was spiked with all 24 drugs at a concentration between 0.001 and 0.010 ng/mg (in 0.001 ng/mg steps) and the deuterated standards at a concentration of 0.05 ng/mg. The LODs and LOQs were determined according to the DIN 32645. If the calculated LOD was lower than the lowest concentration with a detectable signal that concentration was chosen as the LOD. 2.6.2. Linearity and carryover Seven calibration levels (0.025, 0.05, 0.1, 0.25, 0.5, 1.0 and 2.5 ng/mg) were analyzed 6 times. Two additional levels (5.0 and 10 ng/mg) were measured and were used only in cases were the concentration was higher than 2.5 ng/mg. After the highest level a blank sample was measured to determine the carryover. 2.6.3. Precision and accuracy The precision and the accuracy of the method were evaluated at 0.05 and 1.0 ng/mg which were measured in each six replicates of each level for the within-day (intraday) precision and over 5 days for the between-day (interday) precision. The precision was calculated as the percent relative standard deviation, and the accuracy as the percentage of the measured value related to the reference value. 2.6.4. Specificity Drug free hair samples from six different volunteers were analyzed to examine that there are no interferences between the matrix signals and those from the analytes and the deuterated standards. 2.6.5. Matrix effect Ion suppression or enhancement (%) was determined by measuring the substance mixture at 0.05 and 1.0 ng/mg without and in the presence of the extract from 6 hair samples and calculated as peak area in the presence of matrix/ peak area without matrix  100.

3. Results and discussion 3.1. Optimization of sample preparation The extraction procedure used for STA in hair should include all toxicologically relevant basic, neutral and acidic substances with hydrophilic as well as lipophilic properties with high extraction yield and should avoid hydrolysis or decomposition of the substances. An overview about sample preparation techniques was given in a previous review [2] and includes enzymatic or strongly alkaline digestion of the hair matrix, extraction with aqueous HCl or aqueous buffer, treatment with urea and thioglycolate, extraction with methanol or supercritical fluid extraction. From theses, hair digestion procedures are unsuitable because of analyte degradation. Aqueous extraction at acidic or neutral pH has the disadvantage of excluding lipophilic species. It was shown that swelling of the hair by water or very hydrophilic solvents such as methanol is an essential prerequisite for a high extraction yield [19]. However, methanol as a non-reactive and universal extraction solvent has the disadvantage of a high degree of contamination by matrix constituents and frequently provides only low extraction yields. As an alternative, Kronstrand et al. have developed a simple dissolution extraction method for different drug groups using a mixture of acetonitrile: methanol: 20 mM formate buffer, pH 3.0 (10:10:80, v/v/v) for 18 h at 37 8C [20]. This was optimized by Klose Nielsen et al. to a mixture of methanol, acetonitrile and 2 mM ammonium formate containing 8% acetonitrile (pH 5.3) at a ratio 25:25:50 (v/v/v) for 18 h at 37 8C [11].

Table 1 Validation of hair analysis by LC–QTOF-MS for 24 selected analytes. Analyte

[M+H]+ or [M+Na]+ m/z

RT, min

Internal standard

R2 Type Weight

MS (Auto-MS/MS) LODa, ng/mg

LOQa, ng/mg

6.28

0.001 (0.003)

0.003 (0.009)

7-Aminoflunitrazepam 7-Aminoflunitrazepam-d7

284.1194

7.03

0.005 (0.015)

0.006 (0.018)

Acetylcodeine Cocaethylene-d3

342.1700

9.46

0.001 (0.003)

0.003 (0.009)

Alprazolam Alprazolam-d5

309.0902

11.17

0.005 (0.015)

0.007 (0.021)

Amphetamine Amphetamine-d5

136.1121

3.81

0.002 (0.004)

0.002 (0.004)

Benzoylecgonine Benzoylecgonine-d3

290.1387

4.94

0.001 (0.003)

0.003 (0.009)

Bromazepam a-Hydroxyalprazolam-d5

316.0080

9.57

0.002 (0.006)

0.004 (0.012)

Cocaethylene Cocaethylene-d3

318.1700

9.67

0.001 (0.003)

0.002 (0.009)

Cocaine Cocaine-d3

304.1543

8.62

0.001 (0.003)

0.003 (0.009)

Codeine Codeine-d3

300.1594

6.39

0.001 (0.003)

0.003 (0.009)

Diazepam Diazepam-d5

285.0789

12.61

0.001 (0.003)

0.003 (0.009)

Dihydrocodeine Dihydrocodeine-d6

302.1751

5.04

0.001 (0.003)

0.003 0.00 (9)

EDDP EDDP-d3

278.1903

9.21

0.001 (0.003)

0.003 (0.009)

Flunitrazepam a-Hydroxyalprazolam-d5

314.0935

10.30

0.004 (0.012)

0.006 (0.018)

Lorazepam Lorazepam-d4

343.0012c

11.05

0.004 (0.012)

0.005 (0.015)

MDA MDA-d5

180.1019

4.03

0.002 (0.006)

0.003 (0.009)

MDEA MDEA-d6

208.1332

5.00

0.001 (0.003)

0.003 (0.009)

MDMA MDMA-d5

194.1176

4.30

0.001 (0.003)

0.003 (0.009)

0.999 Linear 1/x 0.998 Linear 1/x 0.998 Linear 1/x 0.999 Linear 1/x 0.999 Linear 1/x 0.999 Linear None 0.998 Linear 1/x 0.996 Linear 1/x 0.999 Linear 1/x 0.997 Linear 1/x 0.999 Linear None 0.997 Linear 1/x 0.999 Linear 1/x 0.997 Linear 1/x 0.996 Linear 1/x 0.999 Linear 1/x 0.999 Linear None 0.999 Linear None

Interday bias, %

Matrix effectb, % Standard deviation, %

0.05 ng/mg

1.0 ng/mg

0.05 ng/mg

1.0 ng/mg

0.05 ng/mg

1.0 ng/mg

0.05 ng/mg

1.0 ng/mg

100

102

2.98

0.54

4.93

5.19

88.54 4.75

78.20 6.76

91

91

3.08

0.90

4.42

1.01

83.20 5.49

79.77 7.14

98

98

2.93

1.17

3.89

1.79

95.48 8.52

75.66 11.48

92

95

4.01

0.71

4.10

2.19

72.57 17.22

67.20 10.59

88

96

1.79

0.19

1.86

2.37

88.22 9.23

76.95 8.50

99

96

1.92

1.47

3.20

2.26

99.29 9.11

84.04 4.80

98

93

1.38

3.41

8.50

4.62

64.66 20.04

62.40 15.12

96

96

1.87

0.88

4.57

1.53

104.25 2.41

86.46 4.41

97

102

1.38

0.37

8.15

3.40

100.09 10.29

86.78 4.10

88

93

2.32

0.96

2.80

6.49

92.43 10.42

76.37 7.82

95

99

1.62

1.47

1.89

1.74

58.87 11.51

55.56 10.08

93

101

2.44

0.63

4.94

5.06

102.10 3.72

87.59 3.49

100

99

1.81

0.64

2.03

1.11

100.99 12.82

90.26 2.79

96

92

4.41

2.67

5.52

4.09

64.33 11.69

66.06 11.43

98

95

6.98

5.69

6.78

6.30

43.53 16.92

44.38 17.44

92

93

2.76

1.23

5.76

2.63

98.91 8.17

74.07 8.36

97

96

3.32

1.60

5.52

1.86

99.84 5.51

87.13 4.48

101

95

3.70

1.30

7.76

1.03

106.88 16.74

88.10 3.23

71

328.1543

Intraday CV, %

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

6-Acetylmorphine 6-Acetylmorphine-d3

Accuracy, %

44.52 19.85 42.36 21.71 3.34 2.65

8.05

48.60 13.09 50.80 15.13 4.00 1.83

2.66

79.64 6.40 87.90 5.51 1.04 0.62

1.42

75.82 13.23 75.05 5.62 3.59 1.11

3.05

86.22 3.75 100.60 6.72 2.09 0.68

11.65 310.2165

Internal standard

1.45

88.34 3.31 94.72 3.27

0.05 ng/mg 1.0 ng/mg

0.97 2.72

0.05 ng/mg

0.51

1.0 ng/mg 0.05 ng/mg 1.0 ng/mg 0.05 ng/mg

101 98 1.86 0.999 Linear 1/x 93 100 1.95 Methamphetamine 150.1278 4.28 0.003 (0.009) 0.003 (0.009) 0.999 Methamphetamine-d5 Linear None 93 97 2.98 Methylecgonine 200.1281 0.80 0.001 (0.003) 0.003 (0.009) 0.999 Methylecgonine-d3 Linear None 98 99 0.33 286.1438 3.63 0.001 (0.003) 0.003 (0.009) 0.998 Morphine Morphine-d3 Linear 1/x 92 96 2.79 Nordiazepam 271.0633 12.12 0.002 (0.006) 0.004 (0.012) 0.999 Nordiazepam-d5 Linear None Oxazepam 309.0401d 94 98 3.14 10.91 0.001 (0.003) 0.003 (0.009) 0.997 Oxazepam-d5 Linear 1/x a LOD and LOQ relate to the MS mode. They are by a factor of 3 higher if the measurement is performed in Auto-MS/MS mode (data in parentheses). b Matrix effect (%) calculated as peak area in the presence of matrix/peak area without matrix  100. c M+2 isotope peak, Na+ adduct. d Na+ adduct. 0.003 (0.009)

LOQa, ng/mg LODa, ng/mg

0.001 (0.003)

Intraday CV, % Accuracy, % R2 Type Weight MS (Auto-MS/MS) RT, min [M+H]+ or [M+Na]+ m/z Analyte

Table 1 (Continued )

Methadone Methadone-d9

Interday bias, %

1.0 ng/mg

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

Matrix effectb, % Standard deviation, %

72

In the present study a comparison of the extraction with methanol and with methanol/acetonitrile/2 mM ammonium formate (25:25:50, v/v/v) was performed for 5 hair samples which contained altogether 17 drugs or metabolites (2 6-acetylmorphine, amitriptyline, benzoylecgonine, cocaine, 2 codeine, 2 EDDP, hydrocortisone, lidocaine, linezolide, 2 methadone, metoclopramide, midazolam, mirtazapine, 2 morphine, nordazepam, quetiapine, and verapamil). 20 mg hair pieces were treated with 1 ml of both solvents under three different conditions: (1) 18 h ultrasonic bath at about 25 8C, (2) 18 h incubation at 37 8C without shaking or stirring and (3) 2 h ultrasonic bath + 16 h incubation + 15 min ultrasonic bath. For correction of instrumental variations a mixture of 10 deuterated standards (7-aminoflunitrazepam-d7, amphetamine-d5, benzoylecgonine-d3, codeine-d3, diazepam-d5, EDDP-d3, MDE-d6, methamphetamine-d5, morphine-d3 and THC-d3 (each 1 ng/mg) was added before extraction. After centrifugation 5 ml were injected for LC–QTOF-MS. The results are shown in Fig. 1. It was generally confirmed that the solvent mixture provided higher extraction yields for most substances. Ultrasonication over the whole time or only for 2 h before and 15 min after 16 h incubation had no essential advantage. All further optimization experiments were performed with the solvent mixture. Shaking during the incubation increased the yield by 5–10%. Increase of the temperature to 60 8C and 80 8C did not essentially improve the extraction yield. No instability of the drugs or the added standards was observed in these experiments. Nevertheless, in order to avoid analyte decomposition, 37 8C was chosen as standard temperature. For examination of the extraction completeness, 17 hair samples with 59 drugs or metabolites were stepwise six times incubated with 1 ml of the solvent mixture for 18 h and the extracts were separately measured. The results are shown in Fig. 2. Several substances such as paracetamol, cocaine, tramadol or zolpidem occurred in more than one hair sample and therefore appear repeatedly in the figure. It is seen that for almost all substances between 60 and 95% of the total amount of all six steps is extracted during the first incubation and that generally the yield of the first two steps exceeds 90% with only a few exceptions. There are some substances which were already completely extracted in the first or second step such as and caffeine, lidocaine, nicotine or paracetamol whereas the extraction of diclofenac seems not to be complete even after 6 incubations. For some drugs occurring in more than one hair sample such as codeine, morphine or zolpidem it is obvious that the extraction yield also strongly differs from hair sample to hair sample. Altogether, it was concluded from these experiments that two successive incubation steps of 18 h are suitable to extract the majority of the relevant substances from hair with sufficient yield for qualitative screening as well as for quantitative purposes. 3.2. Validation for selected substances Although the focus of the systematic toxicological analysis is primarily on substance identification, it is necessary to know the sensitivity with which the identification is possible. A calibration and validation of the method was carried out for 24 drugs or metabolites which were regularly and quantitatively investigated in hair by the laboratory in context of driving ability examination (Table 1). Each 20 mg decontaminated drug free hair pieces were spiked in 6 replicates with 0.025, 0.05, 0.1, 0.25, 0.5, 1.0 and 2.5 ng/ mg of the analytes and analyzed in the presence of the deuterated internal standards (0.25 ng/mg) according the procedure described in Sections 2.4 and 2.5. For quantification by LC–QTOF-MS only the MS data were used. Therefore, these measurements were performed in MS mode of the instrument. Under these conditions the sensitivity is three times higher than in Auto-MS/MS mode since the MS data

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

18 h Ultrasonic bath

73

2 h Ultrason. bath + 16 h incub. + 15 min ultrason. bath

18 h Incubation at 37 °C C

6-Acetylmorphine Amitriptyline Cocaine Codeine 1 Codeine 2 EDDP Lidocaine Linezolide Methadone 1 Methadone 2 Metoclopramide Midazolam Mirtazapine Morphine 1 Morphine 2 Nordazepam Quetiapine Verapamil 0

50

100

150

200

0

50

100

150

200

250

0

50

100

150

200

250

Relative extraction yield (%) Fig. 1. Comparison of extraction yields between methanol (100%) and methanol/acetonitrile/2 mM ammonium formate (25:25:50, v/v/v) with application of ultrasonic bath, incubation at 37 8C and combined ultrasonic bath and incubation.

are accumulated over the whole cycle time of 1.0 s instead of only 0.33 s in Auto-MS/MS mode where the residual cycle time is used for measurement of CID fragment spectra. The corresponding peak area ratio of 3 was also practically confirmed by measuring the same samples in MS- and Auto-MS/MS mode. For evaluation, the selected ion chromatograms of the monoisotopic [M+H]+ (for lorazepam and oxazepam [M+Na]+) ions of the analytes and the internal standards were extracted with a mass window of Dm/z = 30–60 ppm (depending on substance) and the peak area ratios were used for calculation of the calibration curves and further validation parameters. The results are shown in Table 1. The extracted ion chromatogram obtained from a hair sample spiked with 38 drugs at 0.1 ng/mg in Auto-MS/MS mode is shown in Fig. 3. All calibrations curves fit to a linear regression (R2 = 0.996–0.999) and were weighted (1/x) in some cases. The accuracy was between 88 and 102% for all substances at 0.05 and 1.0 ng/mg, better than 30% deviation, as required according to the guidelines [17,18]. The intraday and interday precisions were for both 0.05 ng/mg and 1 ng/mg with values between 0.2 and 8.5% also fully acceptable. The usual estimation of the limits of detection and of quantification LOD and LOQ from the signal to noise ratio was not possible since no chemical noise from matrix was observed in the ion chromatograms extracted with Dm/z = 60 ppm and since the electronic noise was excluded by an instrumental threshold of the sampling rate of 200 counts. Therefore, LOD and LOQ were estimated from specific calibration curves established using samples containing the analyte in the range of between 0.001 and 0.010 ng/mg. As described above, these limits given in Table 1 were determined for the highest sensitive mode of the instrument (MS mode with 1 spectrum/s). For the less sensitive Auto-MS/MS mode these limits are three times higher. Nevertheless, the sensitivity is also in Auto-MS/MS mode clearly below the cut-off values used for driving ability examination (0.05 ng/mg or higher for basic drugs and their metabolites [21]). The ion suppression/enhancement at the concentrations 0.05 ng/mg and 1.0 ng/mg was determined as the quotient of the peak areas obtained by measuring the analytes in the absence of matrix and after spiking of blank matrix extracts from six different hair samples. The data are also given in Table 1. This

matrix effect was acceptable for the basic drugs and increased from 0.05 ng/mg (88–107%) to 1 ng/mg (75–90%). However, in case of the 7 benzodiazepines (with the exception of the basic 7aminoflunitrazepam) a stronger ion suppression to 42–72% was found with no distinct difference between the 0.05 ng/mg and the 1.0 ng/mg samples. Obviously the protonation of these less basic compounds is more affected by the matrix than of the stronger basic drugs. This fact could not be eliminated by including all ionic species of these compounds (adducts with Na+, K+, NH4+). The determination of the hair extraction yield as a part of the validation was described in context of the method optimization (Section 3.1 and Fig. 2). The method was applied to the samples from the proficiency test DHF 1/10 of the German Society of Toxicological and Forensic Chemistry (GTFCh). In Fig. 4 the results are compared with the mean concentrations from all participants. Sample A was a spiked hair sample and sample B was a mixture sample from several drug consumers. A good agreement was seen for all analytes with the exception of benzoylecgonine in sample B. The reason for this deviation could not be cleared. Analytical errors are improbable since all four analytes of this sample were determined from the same sample preparation and in the same measurement and since the peak area of benzoylecgonine-d3 was almost the same in both samples A and B which were measured in the same series. 3.3. STA in hair of postmortem hair samples The method was applied to postmortem hair samples from 30 consumers of illegal drugs and 60 cases with intake of therapeutic drugs in the last time before death which were autopsied in the Institute of Legal Medicine of the University Hospital Charite´ Berlin. Information about case histories and the exposure to drugs were obtained from the police reports. Furthermore, the autopsy reports and the results of the toxicological investigation of blood, urine and gastric content were available for comparison in all cases. 3.3.1. Case examples The application of LC–QTOF-MS to the systematic toxicological analysis of hair shall be demonstrated at two examples.

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

74

6 Acetylmorphin 1 6 Acetylmorphin 2 Ambroxol Amiodarone Amitriptyline 1 Amitriptyline 2 Amitriptylinoxide Atropine Benzoylecgonine 1 Benzoylecgonine 2 Bromazepam Caffeine Carbamazepine 1 Carbamazepine 10,11 10 11 epoxide Carbamazepine 2 Citalopram Clarithromycin Clindamycin Cocaine 1 Cocaine 2 Cocaine 3 Codeine 1 Codeine 2 Codeine 3 Diazepam Diclofenac 1 Diclofenac 2 Dioxoprothazine Diphenhydramine 1 Diphenhydramine 2 Doxepine Fentanyl Fluconazole Levomepromazine Lidocaine 1 Lidocaine 2 Lidocaine 3 Methadone Metoclopramide Metolachlor Metoprolol 1 Metoprolol 2 Metoprolol 3 Mirtazapine Morphine 1 Morphine 2 N Desmethyltramadol

Nicone 1 Nicone 2 Nicone 3 Nicone 4 Nicone 5 Nicone 6 Nicone 7 Norcitalopram Nordiazepam Nortriptyline 1 Nortriptyline 2 Noscapine 1 Noscapine 2 O Desmeth D thyltrama lt d doll Ondansetrone Opipramol Oxycodone Panthenol 1 Panthenol 2 Panthenol 3 Panthenol 4 Papaverine 1 Papaverine 2 Paraceteamol 1 Paraceteamol 2 Paraceteamol 3 Paraceteamol 4 Paraceteamol 5 Phenazone 1 Phenazone 2 Pilocarpine Pipamperone Primidone Pyridoxine Sulfadiazine T Trama dol d l1 Tramadol 2 Tramadol 3 Tramadol 4 Trimethoprim Trimipramine Tripelenamine Tryptamine Zolpidem 1 Zolpidem 2 Zolpidem 3 0%

20%

40%

60%

Extracon yield

80%

100%

1. Extracon 2. Extracon 3. Extracon 4 Extracon 4. 5. Extracon 6. Extracon 0%

20%

40%

60%

80%

100%

Extracon yield

Fig. 2. Extraction yields of 59 drugs or metabolites from 17 postmortem hair samples in six successive 18 h incubation steps with methanol/acetonitrile/2 mM ammonium formate (25:25:50, v/v/v). The total amount of all six steps is assumed to be 100%.

Fig. 3. Extracted ion chromatogram obtained from a hair sample spiked with the standard drug mixture at a concentration of 0.1 ng/mg. 1 = methylecgonine, 2 = morphine, 3 = amphetamine, 4 = MDA, 5 = methamphetamine, 6 = MDMA, 7 = hydromorphone, 8 = clonidine, 9 = dihydrocodeine, 10 = benzoylecgonine, 11 = MDEA, 12 = 6acetylmorphine, 13 = codeine, 14 = 7-aminoflunitrazepam, 15 = tramadol, 16 = cocaine, 17 = nortilidine, 18 = EDDP, 19 = acetylcodeine, 20 = cocaethylene, 21 = bromazepam, 22 = flunitrazepam, 23 = citalopram, 24 = carbamazepine, 25 = a-hydroxyalprazolam, 26 = oxazepam-Na, 27 = lorazepam-Na, 28 = alprazolam, 29 = triazolam, 30 = methadone, 31 = zolpidem, 32 = nordazepam, 33 = haloperidol, 34 = diazepam, 35 = tilidine, 36 = imipramine, 37 = amitriptyline, 38 = clozapine.

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

5 4

LC QTOF MS Mean of parcipants Sample A

Sample B

3 2 1 0

Fig. 4. Application of the LC–QTOF-MS method to the hair samples of the proficiency test DHF 1/10 of the Society of Toxicological and Forensic Chemistry (GTFCh). Sample A = spiked hair, sample B original hair from drug consumers.

3.3.1.1. Case 1: drug fatality. The 30-year-old man was found dead on the floor of his apartment by his brother who was known to the police to be a drug dealer. The diseased had suffered from psychosis and was reported to be under a corresponding medical treatment, but no closer information was available. During autopsy 10 drug packages were found in the stomach. Two damaged packages in brown plastic foil had leaked into the gastric content and contained a mixture of cocaine and lidocaine. The other 8 packages in blue plastic foil were intact and contained a typical illegal heroin preparation with heroin, noscapine, papaverin, caffeine and paracetamol. The toxicological investigation of venous blood, urine, gastric content and hair by immunoassay, HPLC-DAD and GC–MS confirmed an overdose of cocaine and lidocaine as the cause of death with chronic abuse of cocaine and heroin. The hair sample was extracted and analyzed by LC–QTOF-MS in Auto-MS/MS mode as described in the experimental part. The extracted ion chromatogram of the 23 substances which were identified by the general unknown search procedure [13] is shown in Fig. 5. Accurate CID spectra of all substances were present in the library. This is the typical pattern found in hair of cocaine and heroin abusers including the essential drug metabolites (benzoylecgonine, norcocaine, cocaethylene, anhydroecgonine

75

methylester, 6-acetylmorphine, morphine), as well as natural accompanying substances (noscapine, papaverine, acetylcodeine, cinnamoylcocaine) and artificial adulterants (paracetamol, lidocaine, tetramisole, hydroxyzine). For the substance included in the validation (Section 3.2 and Table 1) the concentrations were determined and are also given in Fig. 5. Comparison with the previously determined GC–MS results showed higher concentrations for 6-acetylmorphine, morphine and benzoylecgonine and a lower cocaine concentration (see also Section 3.3.1). As a typical antipsychotic drug risperidone and its main metabolite 9-hydroxyrisperidone were identified. In this way the general information from the case history about the antipsychotic treatment could be confirmed and specified by hair analysis. Although hydroxyzine has an anxiolytic effect, it is mainly used as an antihistaminic and should originate rather from its application as a drug adulterant than from medical treatment. 3.3.1.2. Case 2: death case of a multi-morbid neurologic patient. The 54-year-old man was found dead in his room in a nursing home. He had suffered from a hypoxic brain damage with cerebro-organic psycho syndrome, Korsakow syndrome and epilepsy. He was confined to bed and mainly in a non-orientated state. After an antiinfluenza inoculation he had high temperature and was treated with paracetamol, vancomycin and furosemide. His prescribed chronic medication was amitriptyline (50 mg/day), diazepam (10 mg/day), valproic acid (1500 mg/day), omeprazole (20 mg/ day), quetiapine (400 mg/day), tamsulosine (0.4 mg/day) and lactulose. The toxicological investigation of venous blood, urine and gastric content showed therapeutic concentrations of quetiapine, amitriptyline and diazepam and metabolites of these drugs. No indications for a lethal intoxication were found. The hair sample was extracted and analyzed as described in the experimental part. The extracted ion chromatogram of the 26 drugs or metabolites which were identified by the general unknown search procedure [13] is shown in Fig. 6. With the exception of two of the three hydroxyquetiapines, accurate mass CID fragment spectra of all compounds were present in the spectra library. From the prescribed or administered medication paracetamol, vancomycin, amitriptyline and 6 of its metabolites, omeprazole, quetiapine and three hydroxylated metabolites, tamsulosine and diazepam with its metabolite nordazepam were found. Furthermore, a previous treatment with carbamazepine (antiepileptic), tiapride (neuroleptic), trimethoprim (antibiotic),

Fig. 5. Extracted ion chromatogram of the hair sample extract of case 1. 1 = paracetamol, 2 = anhydroecgonine methylester, 3 = phenethylamine, 4 = cotinine, 5 = morphine (0.16 ng/mg), 6 = nicotine, 7 = benzoylecgonine (22 ng/mg), 8 = tetramisole, 9 = 6-acetylmorphine (2.2 ng/mg), 10 = 3-hydroxylidocaine, 11 = cocaine (27 ng/mg), 12 = norcocaine, 13 = heroin, 14 = acetylcodeine, 15 = cinnamoylcocaine, isomer 1, 16 = cocaethylene, 17 = 9-hydroxyrisperidone, 18 = cinnamoylcocaine, isomer 2, 19 = papaverine, 20 = risperidone, 21 = lidocaine, 22 = noscapine, 23 = hydroxyzine.

76

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

Fig. 6. Extracted ion chromatogram of the hair sample extract of case 2. 1 = panthenol, 2 = paracetamol, 3 = metronidazole, 4 = tiapride, 5 = sulfadicramide, 6 = theophylline, 7 = vancomycin, 8 = caffeine, 9 = trimethoprim, 10 = tamsulosin, 11 = E-10-hydroxynortriptyline, 12 = carbamazepine-10,11-epoxide, 13 = cyclobenzaprine, 14 = E-10hydroxyamitriptyline, 15 = N-desalkylverapamil, 16 = hydroyxyquetiapine (2), 17 = Z-10-hydroxynortriptyline, 18 = Z-10-hydroxyamitriptyline, 19 = 7-hydroxyquetiapine, 20 = omeprazole, 21 = hydroxyquetiapine, 22 = carbamazepine, 23 = nortriptyline, 24 = desmethylnortriptyline, 25 = amitriptyline, 26 = nordazepam, 27 = diazepam, 28 = quetiapine.

Table 2 Substances detected in postmortem hair samples. General overview. Number of hair samples

Detected substances

Substances from smoking, coffee, cacao, spice and hair care Illegal drugs and related substances

Therapeutic drugs and metabolites

Therapeutic drugs reported in case history

Therapeutic drugs detected in blood or urine

Therapeutic drugs in case history and/or blood and urine

Examples of not found therapeutic drugs

a

-

Total: 90 From illegal drug fatalities: 30 From other death cases: 60 Total number: 212 Frequency per druga: 1–40, mean 4.2, median 2 Frequency per casea: 2–33, mean 10.1, median 8 Nicotine 48, cotinine 59, 3-hydroxycotinine 25, caffeine 79, theophylline 44, theobromine 48, piperine 32, panthenol 90 Detected substances: Total 35 15 Illegal drugs and metabolites, 5 accompanying substances, 6 adulterants, 9 other opioids or metabolites, cf. Table 3 Not detected: Cannabinoids: THC, CBD, CBN, THC-COOH Total: 182, cf. Table 4 Parent drugs or key metabolitesb: 154 Metabolites: 28 Frequency per drug: 1–28, mean 3.7, median 2 Number of results: 613 Different drugs: 159 Frequency per reported drug 1–12, mean 2.0 Identified different drugs in hair from case history: 51 (32%) Frequency per identified drug: 1–8, mean 1.96 Total number of reported drugs in case historyc: 318 Detected in hair from total number: 100 (31.4%) Different drugs or key metabolites: 137 Frequency per drug 1–14, mean 2.57 Identified different drugs in hair from blood and urine: 63 (33.3%) Frequency per identified drug: 1–13, mean 2.40 Total number of reported drugs from blood and urinec: 257 Detected in hair from total number: 146 (56.8%) Different drugs or key metabolites: 189 Frequency per drug 1–16, mean 2.52 Identified different drugs: 75 (39.7%) Frequency per identified drug: 1–14, mean 2.61 Total number of drugs from history and or blood and urine: 476 Detected in hair from total number: 175 (36.8%) Acidic drugs: Salicylic acid, furosemide, torasemide, ibuprofen, phenobarbital, pravastatin Neutral drugs: Phenprocoumon, simvastatin, spironolactone Low dose drugs: Digoxin, digitoxin, flupenthixol, glimepiride, L-thyroxin Peptide drugs: Insulin Others: Lithium Instable in blood: Isosorbide dinitrate

Substances from smoking, coffee, cacao, spice and hair care are excluded. Key metabolites were counted if the parent drug is nor stable or is not always found in hair, such as noramidopyrine instead of methamizole or 7-aminoflunitrazepam instead of flunitrazepam. c Total number of reported drugs in the investigated cases, multiples included. b

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

77

Table 3 Frequency of substances identified by LC–QTOF-MS in hair samples of 30 illegal drug fatalities. Heroin related substances Heroin 4 6-Acetylmorphine 19 Morphine 19 Normorphine 3 Codeine 15 Acetylcodeine 13 Papaverine 20 Noscapine 17

Other opioides Methadone 13 EDDP 11 Hydromorphone 3 Hydrocodone 2 Oxycodone 1 Norbuprenorphine 1 Fentanyl 1 Nortilidine 1 Tramadol 2

Amphetamines Amphetamine 7 Methamphetamine 1 MDMA 2

Cocaine related substances Cocaine 20 Benzoylecgonine 18 Norcocaine 5 Cocaethylene 4 Anhydroecgonine methylester 5 Anhydroecgonine 1 Cinnamoylcocaine 4 Adulterants Lidocaine 17 3-Hydroxylidocaine 9 Paracetamol 16 Tetramisole/Levamisole 13 Hydroxyzine 3 Phenacetin 2

Other drugs of abuse Ketamine 1 Norpseudoephedrine 2

that substances resulting from drinking of coffee or consumption of cacao products (caffeine, theophylline, theobromine), from smoking (nicotine, cotinine, 3-hydroxycotinine), from spicy food (piperine) and from hair care products and cre`mes (panthenol) were very frequently seen in the chromatograms. The number of different substances per case varied from 2 to 33. All metabolites detected in this study are present in the CID spectra library [13] and no results from application of the tool ‘‘Find Metabolites’’ were included in these results.

sulfadicramide (anti-infective) and cyclobenzaprine (muscle relaxant) was shown. Valproic acid and furosemide were not seen since the measurement was confined to positive ionization and these anion forming drugs are preferentially analyzed by negative ESI. Caffeine, theophylline and panthenol are frequently occurring substances in many hair samples (see Section 3.3). Nicotine, cotinine and 3-hydroxycotinine as markers for the smoking status were not detected. 3.3.2. Results in 90 death cases An overview of the substances detected in the hair samples from the 90 death cases is given in Table 2. Altogether 212 different substances were identified with a frequency of 1–40. In addition to

3.3.2.1. Illegal drugs and related substances. The illegal drugs and related substances identified in the hair samples of the 30 drug fatalities are given in Table 3. In many cases more than one drug were ng/mg 2.5

ng/mg 7

ng/mg 7

6

6

6-Acetylmorphine

2

Morphine

5

5

4

4

1.5

3

3

1

2

2

1

1

0

0

Codeine

ng/mg 9

10

Cocaine

8

x 0.1

Benzoylecgonine

1.5

5 6

x 0.1

x10

2

2

10 02 24

10 24 41

09 86 67

09 77 75

5 10 019

10 105

09 775

09 935

09 781

09 768

09 736

10 117

09 640

10 079

10 1 105

09 7 775

10 0 019

10 1 117

09 7 781

10 0 079

09 7 736

09 9 935

09 6 640

0 09 915

0 09 6 679

09 7 763

09 7 775

09 7 758

10 2 248

10 2 241

09 9 935

09 7 763

09 7 758

10 2 243

LC QTOF MS

x10

10

0.5

0

GC MS

MDMA

15

1 0

09 75 58

20

1

3

EDDP

30 25

x 0.1

4

4

Amp.

35

6

8

45 40

2

7

10 24 43

09 71 19

09 67 79

09 86 67

10 02 24

10 24 41

09 80 06

10 24 43

09 77 75

10 01 12

ng/mg 50

Case No.

Fig. 7. Comparison of drug and metabolite concentrations in hair samples from drug fatalities measured by GC–MS and later by LC–QTOF-MS.

10 079

Methadone

ng/mg 2.5

09 768

12

09 69 96

09 67 79

09 78 81

09 75 58

09 71 19

09 91 15

09 66 63

10 10 05

10 11 17

0

10 079

ng/mg

10 24 41

10 24 43

09 77 75

09 86 67

10 10 05

09 75 58

10 11 17

09 78 81

09 71 19

79 09 67

0.5

78

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

abused. Heroin abuse was evident in 19 cases from the presence of 6acetylmorphine and morphine. Heroin itself as well as the nonhydrolytic metabolite normorphine were detected only four or three times. The accompanying substances codeine, acetylcodeine, noscapine and papaverine were found in the majority of the cases. The substitution drug methadone (13) was the most abundant non-heroin opioide, but there were also hydromorphone, hydrocodone, oxycodone, buprenorphin, fentanyl, tilidine and tramadol as parent drugs or as metabolites detected in some cases. Cocaine and its main metabolite benzoylecgonine were identified in 17 samples. However, the non-hydrolytic metabolites norcocaine and cocaethylene were seen only in five and four of these cases. Ecgonine methylester was not found. The crack marker anhydroecgonine methylester was present in 5 samples, one of them contained also anhydroecgonine. The accompanying substance cinnamoylcocaine was also detected only five times. Amphetamine was detected in 5 samples, one of them contained also methamphetamine and two others of them contained MDMA. Ketamine and cathine were seen only once. The validated quantitative method described in Section 3.2 was applied to the hair samples from drug fatalities. In Fig. 7 the results are compared with the concentrations of the GC–MS analysis for morphine, 6-acetylmorphine, codeine, methadone, its metabolite EDDP, cocaine, benzoylecgonine, amphetamine and methylenedioxymethamphetamine (MDMA) as far as the previous analysis was carried out. Although for each case the concentrations from

both methods are in the same range, there are considerable differences in the exact values in several cases. This difference is particularly seen for cocaine and benzoylecgonine whereas morphine, methadone, EDDP and amphetamine showed an acceptable agreement between both methods. Besides the usual analytical errors, these differences between the results from both methods can be explained by the fact that the samples were not homogenous and different strands were analyzed, that the hair extraction yield was different, and that ester hydrolysis occurred to a different extent during hair extraction for cocaine, benzoylecgonine and 6-acetylmorphine. D9-Tetrahydrocannabinol (THC), cannabidiol (CBD), cannabinol (CBN) as well as 11-nor-D9-tetrahydrocannabinol-9-carboxylic acid (THC-COOH) could not be detected in hair although the presence of cannabinoids was known from the blood or urine results and also from the GC–MS analysis of the hair samples from 5 of the drug fatalities. One reason for the low sensitivity of the applied method for these drugs is the non-basic structure which decreases the protonation in the ionization by ESI. In addition to that, their accurate ion mass is in the range of many matrix constituents which leads to a high chromatographic background in the extracted chromatograms even at the mass resolution of about 10,000 of the used instrument. The detected medical drugs lidocaine, paracetamol, tetramisole, hydroxyzine and phenacetin are known from the literature as typical adulterants of cocaine and heroin preparations [22–24] and

Table 4 Therapeutic drugs and metabolites identified in 90 hair samples by LC–QTOF-MS. Metabolites are printed italic. Only metabolites present in the CID accurate mass spectra library are included. This table does not contain the substances resulting from abuse of heroin, other opioids, cocaine and amphetamine in the 30 drug fatalities given in Table 2. Drug, frequency

Drug, frequency

Drug, frequency

Noramidopyrine 28 Metoclopramide 25 Lidocaine 23 3-Hydroxylidocaine 5 Tramadol 20 O-Desmethyltramadol 8 Metoprolol 17 Hydroxymetoprolol 6 Paracetamol 16 Diclofenac 13 Diphenhydramine 11 Desmethyldiphenhydramine 6 Climbazole 10 Diazepam 10 Doxepine 10 N-Desmethyldoxepine 9 Nordazepam 10 Trimethoprim 10 Citalopram 9 Norcitalopram 9 Bisnorcitalopram 4 Mirtazapine 9 Nortriptyline 9 E-10-Hydroxynortriptyline 4 Z-10-Hydroxynortriptyline 4 Desmethylnortriptyline 1 Carbamazepine 8 Carbamazepin-epoxide 6 Iminostilbene 2 Phenazone 8 Doxylamine 7 Betazole 6 Metronidazole 6

Ondansetrone 6 Quetiapine 6 7-Hydroxyquetiapine 3 Trimipramine 6 Desmethyltrimipramine 1 Zolpidem 6 Amitriptyline 5 Amitriptylinoxide 2 E-10-Hydroxyamitriptyline 2 Z-10-Hydroxyamitriptyline 2 Ciprofloxacin 5 Morphine 5 Pilocarpine 5 Promethazine 5 Dioxopromethazine 4 Zopiclone 5 Clarithromycin 4 Emepronium 4 Melperone 4 Midazolam 4 a-Hydroxymidazolam 1 Pipamperone 4 Prilocaine 4 Propyphenazone 4 Verapamil 4 Norverapamil 4 N-Desalkylverapamil 4 Amisulpride 3 Clozapine 3 N-Desmethylclozapine 3 Denatonium 3 Fentanyl 3 Fluconazole 3

Haloperidol 3 Hydroxyzine 3 Imipramine 3 Nortilidine 3 Oxycodone 3 Sulfadiazine 3 Torasemide 3 Venlafaxine 3 O-Desmethylvenlafaxine 1 7-Amino-flunitrazepam 2 Atropine 2 Bisoprolol 2 Bromazepam 2 Canrenone 2 Codeine 2 Dihydrocodeine 2 Dyclonine 2 Hyoscyamine 2 Laudanosine 2 Levamisole 2 Levomepromazine 2 Memantine 2 Pethidine 2 Methiomeprazine 2 Octodrine 2 Omeprazol 2 Risperidone 2 9-Hydroxyrisperidone 1 Sulfamethoxazole 2 Tetrazepam 2 Tiapride 2 Tranylcypromine 2 Valpromide 2

Drugs or metabolites detected only once: 7-Aminoclonazepam, 7-aminonitrazepam, acebutolol, acetyldigoxin, allopurinol, ambroxol, aminophenazone, amiodarone, desethylamiodarone, amphothericine, amrinone, anastrozole, articaine, atrazine-desethyl, bamipine, benperidol, bifonazole, chlorhexidine, chlorphenoxamine, clemastine, clindamycin, clomipramine, cyromazine, desacetylmetipranolol, desipramine, desloratadine, dexketoprofen, dextromethorphan, dextrorphan, diltiazem, duloxetin, etomidate, febuprol, gepefrine, histapyrrodine, ibuprofen, isoetarine, lamivudine, levetiracetam, levofloxacin, levomepromazine, linezolid, mazindol, mecetronium, metolachlor, milrinone, myosmine, naloxone, nitrazepam, norcyclzine, ofloxacine, olanzapin, opipramol, de(hydroxyethyl)-opipramol (2), oxcarbazepine, 10hydroxycarbazepine, oxybenzone, pantoprazole, paroxetine, phenytoin, pimobendan, piritramide, pregabalin, propranolol, pseudoephedrine, pyracarbolide, ranitidin, reboxetine, salbutamol, sertraline, norsertraline, simvastatin, sotalol, sulfapyridine, sulpiride, samsulosin, terbinafine, thiabendazol, timolol, tolycaine, triadimefon, tripelenamine, zotepine.

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

79

Table 5 Frequency of the 50 most abundant therapeutic drugs or key metabolites in case histories and/or results of analysis of blood, urine and gastric content of 90 death cases compared to frequency of detection in hair by LC–QTOF-MS. Drug or key metabolite

Case history

Blood and urine

History and/or blood and urine

Detected in hair

Drug or key metabolite

Case history

Blood and urine

History and/or blood and urine

Detected in hair

Metamizol Acetyl salicylic acid Pantoprazol Metoprolol Diazepam Ibuprofen Lidocaine Omeprazole Phenprocoumon Torasemide Amitriptyline Citalopram Diphenhydramine Furosemide Mirtazapine Quetiapine Carbamazepine Diclofenac Metoclopramide Midazolam Paracetamol Promethazine Simvastatin Tramadol Amlodipine

6 12 11 8 4 7 1 7 7 7 2 4 3 6 4 4 3 4 4 1 2 4 6 3 4

15 2 5 5 10 4 9 5 2 1 7 7 7 4 6 4 6 2 6 6 5 3 0 6 4

16 14 14 11 10 9 9 9 8 8 7 7 7 7 7 7 6 6 6 6 6 6 6 6 5

14 0 1 8 7 1 6 3 0 0 4 5 5 0 4 4 5 3 6 3 5 3 0 6 0

Fentanyl Ofloxacin Trimipramine Clozapine Codein Doxepine Doxylamine Insulin Morphine Tilidine Phenobarbital Verapamil Allopurinol Amiodarone Amisulpride Clemastine Enoxaparin Flunitrazepam Lorazepam L-Thyroxin Metronidazol Naloxone Salbutamol Tetrazepam 50 mg Zopiclon

2 2 2 1 2 1 2 4 4 2 1 4 2 0 3 1 3 2 3 3 3 3 3 2 2

4 5 5 4 3 4 3 0a 4 2 3 4 1 3 3 2 0 1 3 0 1 0 0 1 2

5 5 5 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3

1 1 3 2 2 4 3 0 4 2 0 4 0 1 3 1 0 1 0 0 2 0 0 1 3

a

Not determined since other cause of death.

should originate in hair preferentially from this application. However, there are also some other drugs such as benzodiazepines, antipsychotics, antidepressants, antibiotics and non-opioid analgesics which show that some of the illegal drug abuser were under a corresponding medical treatment or abused these drugs too. The occurrence of these therapeutic drugs in hair from drug fatalities is given together with the results from the other 60 cases in Table 4. 3.3.2.2. Therapeutic drugs. In the 90 hair samples 154 parent drugs and 28 metabolites were identified with a frequency of 1–28 (Table 4, altogether 613 results). This included mainly chronically prescribed medicals such as 17 antidepressants, 15 neuroleptics, 11 drugs for treatment of cardiovascular diseases, 10 opioid analgesics, 8 non-opioid analgesics, 7 benzodiazepines, 6 antiepileptics, 6 antihistaminics and 5 hypnotics but also 13 antibiotics and 7 fungicides. These results were compared with the information from the police reports about the case history and with the results of the toxicological analysis of blood, urine and gastric content by immunoassay, HPLC-DAD and GC–MS. This comparison can give only an orientation since the data from case histories were only based on the police reports which included information about the drugs prescribed or administered according to the patient charts or found at the scene of death but were not always complete. It was often not clear whether the drugs were really taken by the diseased and whether it was a chronic or merely an acute treatment. Nevertheless, many of the drugs are typically administered in a long term treatment and, therefore, should be found in hair. An overview of the comparison is given in Table 2. The intake or administration of 159 different medical drugs was documented in the case histories with a frequency of 1–12 per drug (total number 318). From these, only 51 different drugs (32%) were found in hair. Similarly, from the 137 drugs or key metabolites known from analysis of blood, urine or gastric content only 63 (33.3%) were detected in hair. Combination of the information from case history and from toxicological analysis lead to 189 different drugs or key metabolites from which 75 (39.7%) were identified in hair.

In Table 5 the frequency of the 50 most abundant therapeutic drugs or key metabolites in case histories and/or results of analysis of blood, urine and gastric content of the 90 death cases is compared with the frequency of their detection in hair from the same cases by LC–QTOF-MS. A high rate of agreement was found for chronically administered drugs such as medicals for treatment of cardiovascular diseases (metoprolol, verapamil), antidepressants (amitriptyline, citalopram, doxepine, mirtazapine, trimipramine), neuroleptics (amisulpride, clozapine, quetiapine), other drugs prescribed in cases of psychic disorders (diazepam, carbamazepine) and drugs for pain treatment (metamizole, morphine, diclofenac). On the other hand, there were 114 drugs mentioned in the case histories and/or detected in blood or urine which were not found in the hair samples. Reasons for this can be: no or only occasional intake of the drug, low dose and correspondingly low blood concentration, low incorporation rate from blood into hair, insufficient stability in hair, low recovery of sample preparation or insufficient ionization by positive ESI in the mass spectrometric measurement. Some of the not detected drugs are typical for a long-term medical treatment and a regular intake can be assumed. In these cases one or more of the other abovementioned reasons apply. According to the fundamental work of Nakahara [25,26], lipophilic basic drugs have generally a good incorporation rate in hair whereas acidic and hydrophilic drugs are less efficiently incorporated. The occurrence of therapeutic drugs in hair was reviewed by Pragst [27] and Che`ze et al. [28]. Acidic drugs such as salicylic acid (key metabolite of acetylsalicylic acid), valproic acid, furosemide, pravastatin or ibuprofen were not described in hair until now and should have a low incorporation rate in hair. In the present study they could also not be detected since they are not sufficiently ionized by positive ESI. This reason is also true for phenobarbital which was not detected in the present study although it is known to be incorporated in hair with high concentrations [29]. Both low incorporation rate and low ESI yield must be assumed also for neutral substances such as phenprocoumon, simvastatin or spironolactone. No data about

80

S. Broecker et al. / Forensic Science International 218 (2012) 68–81

these three drugs in hair was found in the literature. According to Deveaux et al., digoxin concentrations in hair range from 0.0036 to 0.0114 ng/mg after chronic treatment [30]. In the present study, both digoxin and digitoxin were described in the history of each two cases but were not detected in hair. Further examples of chronically applied drugs with low concentrations in blood which were not found in hair in this study are the oral antidiabetic glimepiride and the anti-hypothyroidism drug Lthyroxin. Oral antidiabetics were frequently abused in criminal cases but only glibenclamide was described in hair with concentrations of 0.65 ng/mg after a 20 mg/day dose and 0.005 ng/mg after a single 5 mg dose [31]. Organic nitrates such as isosorbide dinitrate used as coronary vasodilators are not sufficiently stable in blood for incorporation in hair. The analytical method is not suitable for drugs with peptide structure such as insulin or metal ions such as lithium. Antibiotics such as sulfonamides and penicillin derivatives were not frequently found in hair since they are typical for a short-term application and should have a low incorporation rate in many cases because of their hydrophilic or acidic structure. These examples show that the general unknown screening in hair by LC–QTOF-MS is a large step forward in the elucidation of chronic drug intake. However, the information which can be obtained is by fare not complete and there are still many gaps. The large variability of drugs in structure and in concentration cannot be covered by a single method. In particular, the general unknown screening cannot replace the targeted analysis for specific drugs after a single application, for instance after drug facilitated sexual assault. 4. Conclusion It follows from these investigations that liquid chromatography–hybrid quadrupole time-of-flight mass spectrometry (LC– QTOF-MS) in the data dependent acquisition mode and in combination with a large accurate mass data base and CID spectra library is an efficient way for systematic toxicological analysis in hair. The sample preparation by incubation with a mixture of aqueous formate buffer, methanol and acetonitrile is suitable for a wide variety of toxic substances with different polarities and lipophilicities. The extracts are sufficiently pure for injection into the LC–MS system after a partial evaporation of the solvent. In this way loss of analytes by clean-up steps such as solid phase extraction is avoided. The sensitivity for basic drugs with LOD of 0.003–0.015 ng/mg is sufficient for the detection of chronic intake or administration of many illegal or therapeutic drugs and for measurement of concentrations below the cut-off values used in practice for assessment of drug abuse. The sensitivity can be increased for selected analytes by a factor of 3 or more by measurement in MS or in targeted MS/MS mode of the instrument. The comprehensive acquisition of all substances which were extracted from hair and ionized in the ion source, and the use of deuterated standards enables to combine the general unknown screening with the quantitative determination of selected drugs or metabolites. Furthermore, the measured data file of a sample can be searched for any substance, also if it is not in the database or spectra library, and a confirmation of the structure is to a certain degree possible from the CID spectrum. The results show that despite these large advantages of LC– QTOF-MS for the general unknown screening in hair not all substances to which the person was exposed were found in hair. Beside the general limitations of hair analysis concerning incorporation rate of the substance and its stability in the hair matrix against evaporation, wash-out by shampooing, degradation by oxygen, hair cosmetics or UV light, also methodical reasons decrease the portion of detectable substances. For the method

described in this paper this concerns mainly acidic and neutral drugs and drugs with very low concentrations in hair, for instance after single administration, whereas the extraction procedure seems to be suitable for hydrophilic as well as lipophilic substances. Hydrolytic degradation is improbable by the weakly acidic solvent (pH 5.3). Investigations to further improve the applicability of LC–QTOFMS for systematic screening in hair and in other matrices are in progress with respect to systematic search for metabolites [32], additional use of negative ESI and atmospheric pressure chemical ionization (APCI), extension of the CID spectra library to further toxicological substances including accurate mass APCI spectra, use of targeted MS–MS screening by LC–QTOF-MS and possibilities for a semi-quantitative estimation of concentrations. Acknowledgements The authors thank Agilent Technologies, Inc. (Santa Clara, CA, USA) for generous technical support of these investigations and in particular Bernhard Wuest, Dr. Jerry Zweigenbaum, Dr. Frank Kuhlmann and Dr. Peter Stone for helpful cooperation and discussions. They are also grateful to Dagmar Simmert for thorough and efficient experimental assistance. References [1] F. Pragst, M. Rothe, K. Spiegel, F. Sporkert, Illegal and therapeutic drug concentrations in hair segments—a timetable of drug exposure? Forensic Sci. Rev. 10 (1998) 81–111. [2] F. Pragst, M.A. Balikova, State of the art in hair analysis for detection of drug and alcohol abuse, Clin. Chim. Acta 370 (2006) 17–49. [3] M. Wada, R. Ikeda, N. Kuroda, K. Nakashima, Analytical methods for abused drugs in hair and their applications, Anal. Bioanal. Chem. 397 (2010) 1039–1067. [4] B. Madea, F. Mußhoff (Eds.), Haaranalytik—Technik und Interpretation in Medizin und Recht, Deutscher A¨rzte-Verlag, Ko¨ln, 2004. [5] P. Kintz (Ed.), Analytical and Practical Aspects of Drug Testing in Hair, CRC Taylor & Francis, Boca Raton, FL, 2006. [6] Y. Gaillard, G. Pepin, Screening and identification of drugs in human hair by highperformance liquid chromatography-photodiode-array UV detection and gas chromatography–mass spectrometry after solid-phase extraction. A powerful tool in forensic medicine, J. Chromatogr. A 762 (1997) 251–267. [7] H. Sachs, Screening strategies in hair analysis on drugs, in: P. Kintz (Ed.), Analytical and Practical Aspects of Drug Testing in Hair, CRC Taylor & Francis, Boca Raton, FL, 2006, pp. 187–200. [8] D. Thieme, H. Sachs, Improved screening capabilities in forensic toxicology by application of liquid chromatography–tandem mass spectrometry, Anal. Chim. Acta 492 (2003) 171–186. [9] A. Pelander, J. Ristimaa, I. Rasanen, E. Vuori, I. Ojanpera¨, Screening for basic drugs in hair of drug addicts by liquid chromatography/time-of-flight mass spectrometry, Ther. Drug Monit. 30 (2008) 717–724. [10] E. Liotta, R. Gottardo, A. Bertaso, A. Polettini, Screening for pharmaco-toxicologically relevant compounds in biosamples using high-resolution mass spectrometry: a ‘metabolomic’ approach to the discrimination between isomers, J. Mass Spectrom. 45 (2010) 261–271. [11] M.K. Klose Nielsen, S.S. Johansen, P.W. Dalsgaard, K. Linnet, Simultaneous screening and quantification of 52 common pharmaceuticals and drugs of abuse in hair using UPLC–TOF-MS, Forensic Sci. Int. 196 (2010) 85–92. [12] B. Gutherya, T. Bassindaleb, A. Bassindalec, C.T. Pillinger, G.H. Morgana, Qualitative drug analysis of hair extracts by comprehensive two-dimensional gas chromatography/time-of-flight mass spectrometry, J. Chromatogr. A 1217 (2010) 4402–4410. [13] S. Broecker, S. Herre, B. Wu¨st, J. Zweigenbaum, F. Pragst, Development practical application of a CID accurate mass spectra library of more than 2500 toxic compounds for systematic toxicological analysis by LC–QTOF-MS with data dependent acquisition, Anal. Bioanal. Chem. 400 (2011) 101–117. [14] S. Broecker, F. Pragst, A. Bakdash, S. Herre, M. Tsokos, Combined use of liquid chromatography–hybrid quadrupole time-of-flight mass spectrometry (LC– QTOF-MS) and high performance liquid chromatography with photodiode array detector (HPLC-DAD) in systematic toxicological analysis, Forensic Sci. Int. 212 (2011) 215–226. [15] S. Broecker, S. Herre, F. Pragst, F. Kuhlman, B. Wu¨st, J. Zweigenbaum, Toxicological Screening with the Agilent LC/MS-QTOF and the Personal Compound Database and the Broecker, Herre and Pragst Accurate Mass Spectral Library, Agilent Application Note 5990-6419EN, Agilent Technologies, Inc., Santa Clara, 2010. [16] H.-C. Liu, R.H. Liu, D.-L. Lin, H.-O. Ho, Rapid screening and confirmation of drugs and toxic compounds in biological specimens using liquid chromatography/ion trap tandem mass spectrometry and automated library search, Rapid Commun. Mass Spectrom. 24 (2010) 75–84.

S. Broecker et al. / Forensic Science International 218 (2012) 68–81 [17] F.T. Peters, O.H. Drummer, F. Musshoff, Validation of new methods, Forensic Sci. Int. 165 (2007) 216–224. [18] G. Schmitt, M. Herbold, F. Peters, Methodenvalidierung im forensisch-toxikologischen Labor, Auswertung von Validierungsdaten nach den Richtlinien der GTFCh mit VALISTAT, Arvecon, 2009. [19] M. Rothe, F. Pragst, Solvent optimization for the direct extraction of opiates from hair samples, J. Anal. Toxicol. 19 (1995) 236–240. [20] R. Kronstrand, I. Nystrom, J. Strandberg, H. Druid, Screening for drugs of abuse in hair with ion spray LC–MS–MS, Forensic Sci. Int. 145 (2004) 183–190. [21] W. Schubert, R. Mattern, Beurteilungskriterien–Urteilsbildung in der Medizinisch-Psychologischen Fahreignungsdiagnostik, second ed., Kirschbaum Verlag GmbH, Bonn, 2009. [22] R. Coomber, The adulteration of drugs: what dealers do to illicit drugs, and what they think is done to them, Addict. Res. Theory 5 (1997) 297–306. [23] I. Evrard, S. Legleye, A. Cadet-Taı¨rou, Composition, purity and perceived quality of street cocaine in France, Int. J. Drug Policy 21 (2010) 399–406. [24] A. Chang, J. Osterloh, J. Thomas, Levamisole: a dangerous new cocaine adulterant, Clin. Pharmacol. Ther. 88 (2010) 408–411 (Epub 2010). [25] Y. Nakahara, K. Takahashi, R. Kikura, Hair analysis for drugs of abuse. X. Effect of physicochemical properties of drugs on the incorporation rates into hair, Biol. Pharm. Bull. 18 (1995) 1223–1227.

81

[26] Y. Nakahara, R. Hanajiri, Hair analysis for drugs of abuse XXI. Effect of parasubstituents on benzene ring of methamphetamine on drug incorporation into rat hair, Life Sci. 66 (2000) 563–574. [27] F. Pragst, Medikamente, in: B. Madea, F. Mußhoff (Eds.), Haaranalytik—Technik und Interpretation in Medizin und Recht, Deutscher A¨rzte Verlag, Ko¨ln, 2004, pp. 215–261. [28] M. Che`ze, M. Deveaux, G. Duffort, F. Billaut, G. Pe´pin, Pharmaceuticals in hair, in: P. Kintz (Ed.), Analytical and Practical Aspects of Drug Testing in Hair, CRC Taylor & Francis, Boca Raton, FL, 2006, pp. 163–185. [29] J.P. Goulle´, J. Noyon, A. Layet, N.F. Rapoport, Y. Vaschalde, Y. Pignier, D. Bouige, F. Jouen, Phenobarbital in hair and drug monitoring, Forensic Sci. Int. 70 (1995) 191–202. [30] M. Deveaux, E. Re´vuelta, E. Mornay, V. He´douin, D. Gosset, Immunoassay of digoxin in hair, Forensic Sci. Int. 84 (1997) 219–223. [31] M. Villain, C. Tournoud, F. Flesch, V. Cirimele, P. Kintz, Hair to document exposure to glibenclamide, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 842 (2006) 111–115. [32] S. Broecker, M. Koch, F. Pragst, Detection of metabolites in the general unknown screening of hair by LC–QTOF-MS, Ann. Toxicol. Anal. 23 (2011) S1–S12.