Musshoff SPDE Cannabinoids (PDF)

Forensic Science International 133 (2003) 32–38

Automated headspace solid-phase dynamic extraction for the
determination of cannabinoids in hair samples
Frank Musshoff*, Dirk W. Lachenmeier, Lars Kroener, Burkhard Madea
Institute of Legal Medicine, University of Bonn, Stiftsplatz 12, D-53111 Bonn, Germany
Received 26 August 2002; received in revised form 19 November 2002; accepted 5 January 2003

Abstract
This article describes a fully automated procedure for detecting cannabinoids in human hair samples. The procedure uses
alkaline hydrolysis and headspace solid-phase dynamic extraction (HS-SPDE), followed by on-coating derivatization and gas
chromatography–mass spectrometry (GC–MS). SPDE is a further development of solid-phase microextraction (SPME), based
on an inside needle capillary absorption trap. It uses a hollow needle with an internal coating of polydimethylsiloxane as
extraction and pre-concentration medium.
Ten mg of hair were washed with deionised water, petroleum ether and dichloromethane. After adding deuterated internal
standards, the sample was hydrolyzed with sodium hydroxide and directly submitted to HS-SPDE. After absorption of analytes
for an on-coating derivatization procedure, the SPDE-needle was directly placed into the headspace of a second vial containing
N-methyl-N-trimethylsilyl-trifluoroacetamide before GC–MS analysis. The limit of detection was 0.14 ng/mg for D9-tetrahydrocannabinol, 0.09 ng/mg for cannabidiol, and 0.12 ng/mg for cannabinol. Absolute recoveries were in the range of 0.6 to
8.4%. Linearity was verified over a range from 0.2 to 20 ng/mg, with coefficients of correlation between 0.998 and 0.999. Intraand inter-day precision were determined at two different concentrations and resulted in ranges between 2.3 and 6.0% (intra-day)
and 3.3 and 7.6% (inter-day). Compared with conventional methods of hair analysis, this automated HS-SPDE–GC–MS
procedure is substantially faster. It is easy to perform without using solvents and with minimal sample quantities, and it yields the
same sensitivity and reproducibility. Compared to SPME, we found a higher extraction rate, coupled with a faster automated
operation and greater stability of the device.
# 2003 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Cannabinoids; Hair analysis; SPDE; SPME

1. Introduction
Hair analysis has proved to be a reliable tool for the
retrospective detection of chronic drug abuse in clinical and
forensic toxicology [1,2]. Gas chromatography–mass spectrometry (GC–MS) using selected-ion monitoring (SIM)
seems to be the method of choice for determining the
presence of cannabinoids in hair samples [3–9]. Enhanced
sensitivity was achieved by using the negative ion chemical
ionization (NCI) mode [10–13] or tandem mass spectrometry (MS–MS) [14–18].
*
Corresponding author. Tel.: þ49-228-738316;
fax: þ49-228-738368.
E-mail address: f.musshoff@uni-bonn.de (F. Musshoff).

Besides the parent drug D9-tetrahydrocannabinol (THC),
the cannabis constituents cannabinol (CBN) and cannabidiol
(CBD) are regularly found in hair samples [6]. In our
experience, the main THC metabolite 11-nor-D9-tetrahydrocannabinol-9-carboxylic acid (THC–COOH) was seldom
identified, even in THC positive cases. For a quantitative
determination of THC–COOH in the lower pg/mg range,
MS–MS technique is recommended.
Headspace solid-phase microextraction (HS-SPME) is a
sampling technique that allows an extraction from small
amounts of biological material. HS-SPME is based on the
partitioning of analytes between the sample, the headspace
above the sample, and a coated fused-silica fiber [19,20].
THC, CBD and CBN have been analyzed so far by means of
SPME using the direct extraction technique from an aqueous

0379-0738/03/$ – see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved.
doi:10.1016/S0379-0738(03)00047-1

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

medium (direct immersion, DI-SPME) in water and human
saliva [21] and in hair samples [22]. Sporkert and Pragst [23]
reported a HS-SPME method for determining THC, CBD and
CBN in hair samples. However, the limits of detection were
unsatisfactory. In contrast, we have developed a procedure,
including for the first time a derivatization step, which is
highly recommended for GC–MS analysis of cannabinoids
[24].
However, the main disadvantages of SPME are the fragility
of the fused silica and the unprotected stationary phase coating
on the outer surface of the fiber when extended through the
syringe needle. Its limited flexibility regarding surface area
and film thickness is another weakness of SPME. Murphy [25]
and Gesser and co-workers [26–28] have attempted to overcome these problems while maintaining the advantages of
SPME, by using internally coated needles. An SPME–LC
system known as in-tube SPME that uses an open tubular
fused-silica capillary column was developed by Eisert and
Pawliszyn [29], and was currently reviewed by Kataoka [30].
The solid-phase dynamic extraction (SPDE), recently
developed by Chromtech (Idstein, Germany), is the first
commercially available inside-needle device for headspace
analysis using GC–MS. Stainless steel needles (8 cm) internally coated with a 50 mm film of polydimethylsiloxane
(PDMS) and 10% activated carbon are used. The dynamic
sampling is performed by passing the headspace through the
tube actively by a syringe. The volume of the stationary
phase of the SPDE-needle is approximately 4.40 mm3,
compared with the 0.94 mm3 of a 100 mm PDMS SPME
fiber. A great advantage of the SPDE technique over SPME
is the robustness of the capillary. In contrast to the fragile
SPME fibers, the SPDE device is nearly impossible to
damage mechanically. The SPDE was successfully applied
to the analysis of pesticides in water [31] and amphetamines
and synthetic designer drugs in hair [32].
Headspace microextraction of aqueous samples normally
requires agitation, to facilitate mass transport between the
bulk of the aqueous sample and the fiber [33]. In this study,
we evaluated the extraction process for SPDE using a newly
designed single magnet mixer. The new HS-SPDE method is
compared to our previously developed HS-SPME method
[24] for the analysis of cannabinoids in hair samples.

2. Experimental
2.1. Reagents and materials
The following substances were purchased from Promochem (Wesel, Germany) as methanolic standard solutions:
cannabidiol, cannabinol, D9-tetrahydrocannabinol and D9tetrahydrocannabinol-d3. The solutions were stored at 8 8C
and used after dilution to the required concentrations.
N-methyl-N-trimethylsilyl-trifluoroacetamide (MSTFA) was
obtained from Macherey-Nagel (Du¨ ren, Germany). Chemicals
were purchased from Merck (Darmstadt, Germany).

33

The SPDE equipment (syringes with attached SPDEneedles, SPDE gas station and single magnet mixer) was
kindly donated by Chromtech (Idstein, Germany). The
needles (50 mm  0:8 mm, i.d. 0.4 mm, conical needle tip
with side port) were coated by the manufacturer BGB
Analytik (Anwil, Switzerland) with 50 mm PDMS containing 10% of activated carbon (AC). The needles were
attached to 2.5 ml gas-tight syringes (Chromtech). The
gas station was connected to a high purity nitrogen generator
(Peak Scientific, Inchinnan, Scotland). The gas station is
used to acquire a defined volume of nitrogen before desorption. The syringe adapter heater was set at 90 8C. The
single magnet mixer is a new accessory to the Combi PAL
autosampler; temperature and rotational speed are software
controlled. The standard agitator of the autosampler is used
furthermore for the heating of the derivatization reagent.
2.2. GC–MS method
The GC–MS system used for analysis is an Agilent model
6890 Series Plus gas chromatograph in combination with a
CTC Combi PAL autosampler and an Agilent 5973N mass
selective detector (Chromtech, Idstein, Germany). Data
acquisition and analysis were performed using standard
software supplied by the manufacturer (Agilent Chemstation
C00.01 and CTC Cycle Composer 1.5.2). All steps of the
SPDE methods were fully automated, controlled by the
CTC-Combi-PAL software with custom-made macros.
Substances were separated on a fused silica capillary
column (DB-5MS, 30 m  0:25 mm i.d., film thickness
0.25 mm). Temperature program: 90 8C, hold for 5 min;
30 8C/min up to 190 8C, hold for 10 min; 5 8C/min up to
250 8C, hold for 3 min; 30 8C/min up to 300 8C, hold for
2 min. The temperatures for the injection port, ion source,
quadrupole and interface were set at 260, 230, 150 and
280 8C, respectively. Splitless injection mode was used and
helium with a flow rate of 1.0 ml/min was used as carrier gas.
The inlet nut was modified to accommodate the SPDEneedles with a diameter of 0.8 mm. A 1.5 mm i.d. headspace
insert liner (Supelco, Deisenhofen, Germany) and a conventional septum were used.
To determine the retention times and characteristic mass
fragments, electron impact (EI) mass spectra of the analytes
were recorded by total ion monitoring. For quantitative analysis the chosen diagnostic mass fragments were monitored in
the selected ion monitoring (SIM) mode m/z: 303, 371, 386 for
THC–TMS; 301, 337, 390 for CBD–di-TMS; 367, 368, 382
for CBN–TMS; and 315, 374, 389 for THC–TMS-d3 as
internal standard (target ions are marked). For quantification,
peak area ratios of the analytes to the internal standard were
calculated as a function of the concentration of the substances.
2.3. Preparation of the hair samples
The samples were subsequently washed for 5 min in 5 ml of
deionized water, petroleum ether and, finally, dichloromethane

34

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

using a Vortex Genie 2 mixer (Bender & Hobein AG, Zurich,
Switzerland). After drying the hair samples were cut into
small pieces of about 1 mm. The washing solutions were
analyzed by conventional GC–MS procedures to exclude a
contamination. Ten milligram of hair were submitted to
alkaline hydrolysis into a 10 ml headspace vial in the presence of 1 ml of NaOH (1 M), 0.5 g of sodium carbonate and
80 ml aqueous internal standard solution (250 ng of d3-THC/
ml). After addition of a magnetic stirring bar, the vial was
sealed using a silicone–PTFA septum and a magnetic cap.
2.4. Headspace SPDE procedure
For equilibration the sample solution was mixed for 5 min
at 90 8C in the single magnet mixer (200 rpm, 60 s on time,
2 s off time). With constant stirring of the mixer the SPDEneedle was inserted into the sample vial through the septum
and the plunger was moved up and down at 50 ml/s for 30
times aspirating and dispensing a volume of 1 ml to extract
the analytes dynamically (20 min, Fig. 1). For on-coating
derivatization, the syringe was positioned above a second
vial placed into the agitator oven (90 8C) containing 25 ml of
MSTFA and the plunger was moved up and down for six
times (4 min). After the last filling cycle the syringe was
emptied, moved to the gas station and 1 ml of nitrogen was
aspirated. For desorption of the analytes, the needle was
completely introduced into the hot injection port of the GC
and was hold there for 15 s for thermal equilibration
(260 8C). The plunger was moved slowly down (10 ml/s)

and the analytes were flushed into the GC. Simultaneously
with desorption, the GC run was started.
After removing the SPDE-needle from the injection port,
syringe and needle were cleaned by flushing again 1 ml of
nitrogen into the injection port with now opened split vent,
followed by a blank run to exclude carry-over.
2.5. Method optimization and validation studies
The method conditions that are independent of the extraction mode (SPDE or SPME) like incubation temperature,
alkaline hydrolysis, agitator speed, salt additions, amount of
derivatization reagent were set according to our previously
developed SPME method [24] given in the text previously.
The following SPDE specific parameters were optimized
successively by testing three vials at each temperature and
each point: number of filling cycles for extraction (10–500)
and derivatization (1–10), speed of aspirating the syringe for
extraction (50–250 ml/s) and desorption (10–200 ml/s), flush
gas volume for desorption (250–2500 ml), pre-desorption
time in GC injection port (1–45 s) and desorption temperature (230–280 8C).
For the validation of the method spiked samples containing 2 ng of each analyte/mg hair, respectively, were prepared
and analyzed using the procedures described previously.
Peak purity and selectivity, intra- and inter-day precision
at two different concentrations (2 and 20 ng/mg) were
determined. The linearity of the calibration curve was
evaluated between 0.05 and 20 ng/mg. For the determination
of the limit of detection (LOD) and the limit of quantitation
(LOQ) a separate calibration curve in the range of LOD
(0.01–1 ng/mg) was established [34,35]. For the determination of the absolute recovery, hair samples (10 mg) spiked
with 20 ng of each cannabinoid were analyzed with the
HS-SPDE procedure and results were compared with a
liquid injection of a methanolic solution (20 ng/2 ml).

3. Results and discussion
3.1. Parameter optimization for the SPDE method

Fig. 1. Principle of the SPDE extraction process.

3.1.1. Agitation techniques
The standard agitator of the Combi PAL autosampler
could not be used for shaking during the extraction process,
because the thicker SPDE-needle can not be bent as is
possible with the SPME device. This does not present a
problem for volatile or semi-volatile analytes, because when
the aqueous and gaseous phases are at equilibrium before the
start of sampling, most of the analytes are in the headspace
[33]. The SPDE analysis of amphetamines, for example, was
possible without agitation [32]. In the case of the cannabinoids, the equilibrium is reached slowly without agitation,
leading to long analysis times. With the new single magnet
mixer, the equilibration time can be substantially shortened.
By stirring the analyte, transport in the aqueous phase is

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

35

Fig. 2. Derivatization-time profiles with MSTFA (2 ng/mg of each analyte) (n ¼ 3).

made much faster than in the other two phases. Therefore, it
is no longer the step that limits the diffusion process. In all
further experiments, the samples were stirred magnetically.
3.1.2. Extraction
Extraction time and extraction recovery depend on the
number of filling cycles, the plunger speed, and the volume
aspirated through the syringe. However, even if equilibrium
was not completely reached, 30 cycles were used as a good
compromise between time of analysis and sensitivity. The
optimal extraction flow speed was 50 ml/s. A volume of
1000 ml for aspirating and dispensing produced the best
results.
3.1.3. Derivatization
For THC and CBD, the derivatization reaction started
slowly (1–4 cycles) because of the time needed for MSTFA
to diffuse into the needle coating. The peak areas increased
at five cycles, and the reaction was finished after six cycles.
More derivatization cycles led to a decrease in extraction
recovery, which may have been caused by desorption processes (Fig. 2). The relatively small time window for maximum recovery can be reproducibly adjusted with the
autosampler and had no negative influence on the results.
It is important to note that for each sample, a separate vial
with derivatization reagent had to be used. Otherwise a
carry-over was noticed. The use of 25 ml MSTFA is sufficient for derivatization.
3.1.4. Desorption
The pre-desorption time in the injection port for thermal
equilibration should not be longer than 15 s. With longer
times, a peak tailing was observed, resulting in decreased
sensitivity. In this period of time, the thermal equilibration of
the needle is achieved, so that the analytes are completely
desorbed. The volume and plunger speed have a significant

influence on desorption. At a speed of 50 ml/s, the response
increased with incrementing desorption volume, being highest at a full syringe volume of 2.5 ml. The best response was
reached with the slowest adjustable speed of 10 ml/s and a
nitrogen volume of 1 ml, showing a maximum in the desorption profile curve (Fig. 3). At a plunger speed above 50 ml/s,
the pressure in the injection port was too high, causing the GC
system to show an error message. Additionally, the analytes
had no time to diffuse from the PDMS film into the nitrogen
stream at these faster desorption speeds, causing a decrease in
the chromatographic response and a peak tailing. The optimal
injection port temperature was 260 8C.
Because of the relatively long desorption time, the GC
column was held at 90 8C for 5 min. At higher oven temperatures, peak tailing appeared. Lower temperatures (30, 50
or 70 8C) did not improve the chromatographic separation.

Fig. 3. Effect of the desorption volume and desorption flow speed
on the extraction recovery (CBN, 2 ng/mg hair; n ¼ 3).

36

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

Table 1
Method precision of SPDE in comparison to SPME [24]
Precisiona

CBD
THC
CBN

Intra-day (%) for SPDE

Inter-day (%) for SPDE

Intra-day (%) for SPME

Inter-day (%) for SPME

0.5 ng/mg

20 ng/mg

0.5 ng/mg

20 ng/mg

0.5 ng/mg

20 ng/mg

0.5 ng/mg

20 ng/mg

5.1
5.3
6.0

3.6
3.4
2.3

7.6
5.7
6.1

5.0
3.7
3.3

7.2
5.1
6.8

6.1
1.9
3.3

12.6
5.5
7.2

9.9
3.3
6.7

Samples: 10 mg hair, 1 ml 1 M NaOH, 0.5 g Na2CO3. SPME parameters: incubation (5 min), extraction (25 min) and derivatization (8 min) at
90 8C, desorption (5 min) at 250 8C. SPDE parameters: incubation (5 min), extraction (30 cycles, 200 ml/s) and derivatization (6 cycles,
200 ml/s) at 90 8C, desorption (1 ml nitrogen, 10 ml/s) at 260 8C.
a
Precisions are expressed as R.S.D. (%), intra-day (n ¼ 6), inter-day (n ¼ 18).

Table 2
Extraction yield, limit of detection (LOD) and limit of quantitation (LOQ) of SPDE in comparison to SPME [24]
SPDE

CBD
THC
CBN

SPME

Extraction yielda
(%)

LODb
(ng/mg)

LOQb
(ng/mg)

Extraction yielda
(%)

LODb
(ng/mg)

LOQb
(ng/mg)

1.7
8.4
0.6

0.09
0.14
0.12

0.44
0.45
0.44

1.9
7.5
0.3

0.08
0.05
0.14

0.27
0.27
0.51

Samples and parameters: see Table 1.
a
Extraction yield: The absolute amount of analytes extracted by SPME was calculated by comparison with the corresponding direct
injection of a methanolic sample solution onto the GC column (initial amount: 20 ng, n ¼ 3), yield ¼ peak area of SPME/peak area of liquid
injection100.
b
Limit of detection and quantitation were determined by establishing a specific calibration curve from samples containing the analyte in
the range of LOQ. The limits were calculated from the residual standard deviation of the regression line [34,35].

3.2. Validation
Routine analyses of samples from non-drug users
revealed no interfering peaks from the hair matrix. Peak
purity and selectivity were ensured. Further validation data
were obtained with spiked hair samples and are demonstrated in Tables 1 and 2, in comparison with data previously
obtained trough SPME.
For the semi-volatile analytes, extraction recoveries were
in the range between 0.6 and 8.4%. The determined absolute
recoveries are sufficient because in contrast to a liquid
injection, the total amount absorbed from the SPDE-needle
is transferred onto the GC column. By a liquid injection after
a conventional sample preparation (liquid–liquid extraction
or SPE) only a fraction of the total extract is injected (e.g.
2 ml of 200 ml, which is 1%). Therefore, even for CBN, the
extraction yield of 0.6% is sufficient for a valid analysis.
The detection limits using the SPDE conditions described
previously were 0.09–0.14 ng/mg, which are comparable to
the values obtained with the corresponding SPME method
[24], and the values, obtained with conventional extraction
[6,9] or SPME [22] indicated in literature. If higher sensitivity is required, for example, for THC, which is usually
found at lowest concentrations, the limits can be reduced
by increasing the number of extraction cycles. Using the

described conditions, which were optimized for fast analysis
time, THC could be determined qualitatively in all the cases.
Precision resulted in the ranges between 2.3 and 6.0%
(intra-day) and 3.3 and 7.6% (inter-day). The calibration
curves were constructed from peak areas using the SIM mode
and show a linear relationship for each drug over a range from
0.2 to 20 ng/mg with coefficients of correlation from 0.998 to
0.999. Regarding the validating data, the procedure is sensitive, selective and reproducible. The applicability of the
developed method was demonstrated by analyzing hair samples of individuals with drug abuse in general. A HS-SPME
chromatogram of an authentic hair sample is shown in Fig. 4.
All the three cannabinoids were found in 20 hair samples
from individuals who abuse drugs (Table 3). Conventional
Table 3
THC, CBN and CBD concentrations determined in hair samples of
drug abusers

CBD
THC
CBN
a

Concentration
(ng/mg)

Mean concentration
(ng/mg)

0.81–19.02
0.25–0.73a
0.12a–1.48

4.36
0.36a
0.48

Below limit of quantitation.

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

37

Fig. 4. Reconstructed HS-SPDE SIM chromatogram of an authentic hair sample containing 1.25 ng/mg CBD and 0.44 ng/mg CBN. The
concentration of THC is below the detection limit (approximately 0.25 ng/mg).

procedures involving methanolic extraction resulted in comparable concentrations. The concentrations of THC were
generally in the range between the limits of quantitation
and detection. CBD was the major analyte with the highest
concentration in all samples. In most of the samples (67%),
the concentration of CBN was also superior to THC. This
accords with the findings of Cirimele et al. [6], as well as
those of Strano-Rossi and Chiarotti [22], and our previous
findings using SPME [24], in which CBD and CBN were the
major analytes. In contrast to these studies and our findings,
Baptista et al. [9] found CBD concentrations inferior to those
of THC and CBN.
Both, the new HS-SPDE procedure and the approved
HS-SPME method, seem to be suitable for the determination
of THC, CBN and CBD in hair samples in a convenient onestep method. The number of sources of error is reduced distinctly by the automation of all steps: heating and shaking the
sample, alkaline hydrolysis, absorption, derivatization, and
desorption in the injector of the GC are programmable, which
leads to better reproducibility than with manual operation.
The higher capacity of the SPDE increases the extraction
yield in spite of the shorter extraction time. The main
advantage of SPDE is the stability of the device, which
lasts for more than 350 samplings. With SPME only 90–100
samplings are possible per fiber.

4. Conclusion
The application of fully automated headspace solid-phase
dynamic extraction (HS-SPDE) followed by GC–MS for the
determination of THC, CBD and CBN in hair was tested.
The SPDE as a further development of SPME turned out to
be equally suitable for the requirements of clinical and
forensic toxicology regarding sensitivity and selectivity.

The main advantages are the robustness of the device and
its greater capacity.

Acknowledgements
The authors thank Chromtech (Idstein, Germany) and
BGB Analytik (Anwil, Switzerland) for technical assistance,
valuable discussions and suggestions in the establishment of
the SPDE method. Presented in part at the Workshop 2002 of
the Society of Hair Testing (Berlin, Germany).

References
[1] P. Kintz (Ed.), Drug testing in hair, CRC Press, Boca Raton,
FL, 1996.
[2] H. Sachs, P. Kintz, Testing for drugs in hair: critical review of
chromatographic procedures since 1992, J. Chromatogr. Part
B: Biomed. Sci. Appl. 713 (1998) 147–161.
[3] M.R. Moeller, Drug detection in hair by chromatographic
procedures, J. Chromatogr. 580 (1992) 125–134.
[4] V. Cirimele, P. Kintz, P. Mangin, Testing human hair for
cannabis, Forensic Sci. Int. 70 (1995) 175–182.
[5] C. Jurado, M.P. Gimenez, M. Menendez, M. Repetto,
Simultaneous quantification of opiates, cocaine and cannabinoids in hair, Forensic Sci. Int. 70 (1995) 165–174.
[6] V. Cirimele, H. Sachs, P. Kintz, P. Mangin, Testing human
hair for Cannabis. Part III. Rapid screening procedure for the
simultaneous identification of D9-tetrahydrocannabinol, cannabinol, and cannabidiol, J. Anal. Toxicol. 20 (1996) 13–16.
[7] G. Kauert, J. Ro¨ hrich, Concentrations of D9-tetrahydrocannabinol, cocaine and 6-monoacetylmorphine in hair of drug
abusers, Int. J. Legal Med. 108 (1996) 294–299.
[8] O. Quintela, A.M. Bermejo, M.J. Tabernero, S. Strano-Rossi,
M. Chiarotti, A.C.S. Lucas, Evaluation of cocaine, Forensic
Sci. Int. 107 (2000) 273–279.

38

F. Musshoff et al. / Forensic Science International 133 (2003) 32–38

[9] M.J. Baptista, P.V. Monsanto, E.G. Pinho Marques, A.
Bermejo, S. Avila, A.M. Castanheira, C. Margalho, M.
Barroso, D.N. Vieira, Hair analysis for D9-THC, D9-THC–
COOH, CBN and CBD, by GC–MS-EI: comparison with
GC–MS-NCI for D9-THC–COOH, Forensic Sci. Int. 128
(2002) 66–78.
[10] P. Kintz, V. Cirimele, P. Mangin, Testing human hair for
cannabis. Part II. Identification of THC–COOH by GC–MSNCI as a unique proof, J. Forensic Sci. 40 (1995) 619–622.
[11] D. Wilkins, H. Haughey, E. Cone, M. Huestis, R. Foltz, D.
Rollins, Quantitative analysis of THC, 11-OH–THC, and
THC–COOH in human hair by negative ion chemical ionization
mass spectrometry, J. Anal. Toxicol. 19 (1995) 483–491.
[12] H. Sachs, U. Dressler, Detection of THC–COOH in hair by
MSD-NCI after HPLC clean-up, Forensic Sci. Int. 107 (2000)
239–247.
[13] C. Moore, F. Guzaldo, T. Donahue, The determination
of 11-nor-D9-tetrahydrocannabinol-9-carboxylic acid (THC–
COOH) in hair using negative ion gas chromatography–mass
spectrometry and high volume injection, J. Anal. Toxicol. 25
(2001) 555–558.
[14] C.C. Nelson, M.D. Fraser, J.K. Wilfahrt, R.L. Foltz, Gas
chromatography–tandem mass spectrometry measurement of
D9-tetrahydrocannabinol, naltrexone, and their active metabolites in plasma, Therp. Drug Monit. 15 (1993) 557–562.
[15] T. Mieczkowski, A research note: the outcome of GC–MS–
MS confirmation of hair assays on 93 cannabinoid(þ) cases,
Forensic Sci. Int. 70 (1995) 83–91.
[16] M. Uhl, Determination of drugs in hair using GC–MS–MS,
Forensic Sci. Int. 84 (1997) 281–294.
[17] M. Chiarotti, L. Costamagna, Analysis of 11-nor-9-carboxyD9-tetrahydrocannabinol in biological samples by gas chromatography–tandem mass spectrometry (GC–MS–MS), Forensic Sci. Int. 114 (2000) 1–6.
[18] M. Uhl, Tandem mass spectrometry: a helpful tool in hair
analysis for the forensic expert, Forensic Sci. Int. 107 (2000)
169–179.
[19] J. Pawliszyn, Solid Phase Microextraction—Theory and
Practice, Wiley-VCH, New York, 1997.
[20] J. Pawliszyn, Applications of Solid Phase Microextraction,
RSC, Cambridge, UK, 1999.
[21] B.J. Hall, M. Satterfield-Doerr, A.R. Parikh, J.S. Brodbelt,
Determination of cannabinoids in water and human saliva by
solid-phase microextraction and quadrupole ion trap gas
chromatography–mass spectrometry, Anal. Chem. 70 (1998)
1788–1796.

[22] S. Strano-Rossi, M. Chiarotti, Solid-phase microextraction
for cannabinoids analysis in hair and its possible application
to other drugs, J. Anal. Toxicol. 23 (1999) 7–10.
[23] F. Sporkert, F. Pragst, Use of headspace solid-phase
microextraction (HS-SPME) in hair analysis for organic
compounds, Forensic Sci. Int. 107 (2000) 129–148.
[24] F. Musshoff, H.P. Junker, D.W. Lachenmeier, L. Kroener,
B. Madea, Fully automated determination of cannabinoids in
hair samples using headspace solid-phase microextraction and
gas chromatography–mass spectrometry, J. Anal. Toxicol. 26
(2002) 554–560.
[25] G.E. Murphy, United States Patent 5,565,622 (1996).
[26] M.E. McComb, R.D. Oleschuk, E. Giller, H.D. Gesser,
Microextraction of volatile organic compounds using the
inside needle capillary adsorption trap (INCAT) device,
Talanta 44 (1997) 2137–2143.
[27] S. Shojania, M.E. McComb, R.D. Oleschuk, H. Perreault,
H.D. Gesser, A. Chow, Qualitative analysis of complex
mixtures of VOCs using the inside needle capillary adsorption
trap, Can. J. Chem. 77 (1999) 1716–1727.
[28] S. Shojania, R.D. Oleschuk, M.E. McComb, H.D. Gesser, A.
Chow, The active and passive sampling of benzene, toluene,
ethyl benzene and xylenes compounds using the inside
needle capillary adsorption trap device, Talanta 50 (1999)
193–205.
[29] R. Eisert, J. Pawliszyn, Automated in-tube solid-phase
microextraction coupled to high-performance liquid chromatography, Anal. Chem. 69 (1997) 3140–3147.
[30] H. Kataoka, Automated sample preparation using in-tube
solid-phase microextraction and its application—a review,
Anal. Bioanal. Chem. 373 (2002) 31–45.
[31] J. Lipinski, Automated solid-phase dynamic extraction—
extraction of organics using a wall coated syringe needle,
Fresenius J. Anal. Chem. 369 (2001) 57–62.
[32] F. Musshoff, D.W. Lachenmeier, L. Kroener, B. Madea,
Automated headspace solid-phase dynamic extraction for the
determination of amphetamines and synthetic designer drugs
in hair samples, J. Chromatogr. A 958 (2002) 231–238.
[33] J. Pawliszyn, Solid-phase microextraction, Adv. Exp. Med.
Biol. 488 (2001) 73–87.
[34] DIN 32 645, Chemische Analytik: Nachweis-, Erfassungsund Bestimmungsgrenze, Ermittlung unter Wiederholbedingungen. Begriffe, Verfahren, Auswertung, Beuth Verlag,
Berlin, Germany, 1994.
[35] P.C. Meier, R.E. Zu¨ nd, Statistical Methods in Analytical
Chemistry, Wiley, New York, 2000.