Professional Documents
Culture Documents
SUMMARY. A method for the simultaneous analysis of urine for the major drugs of
abuse is described. The analytical procedure uses solid-phase extraction (SPE), gas
chromatography±mass spectrometry (GC±MS) and a semi-automated identi®cation
process. It allows simultaneous extraction, derivatization and analysis of acidic,
neutral and basic drugs from urine. Urine samples were subjected to enzymatic
hydrolysis followed by SPE using Bakerbond narc-2 columns. The eluant was
selectively derivatized with N-methyl-bis-tri¯uoroacetamide (MBTFA) and N-
methyl-N-trimethylsilyltri¯uoroacetamide+1% trimethylchlorosilane. Analysis was
performed using a GC±MS system operating in full scan mode. A simple macro
programme was written to enhance the mass spectra identi®cation capabilities of the
MS software by producing extracted ion chromatograms (EIC) for the drugs of
interest. Once a suspect compound was indicated by EIC, the mass spectrum of the
compound was searched manually against reference libraries for positive
identi®cation and the retention time checked against that of the standard. This
procedure has increased both the amount and the reliability of information given to
clinicians without increasing the cost per sample. The system has been in routine
operation for 24 months, processing up to 40 urine samples per day, with a usual
turn-around time of 48 h.
690
Analysis of urine for drugs of abuse 691
using GC±MS, a method is needed that allows (CON-DAU) level 3 was obtained from DPC
extraction of the whole range of drugs of abuse (Llanberis, UK). Bakerbond narc-2 jumbo 3-mL
simultaneously. It would then follow that only (125-mg) extraction cartridges were obtained
one injection per sample onto the GC±MS from Mallinckrodt Baker (Milton Keynes, UK).
system would be necessary. This would then The Spe-ed wiz vacuum manifold was obtained
make the use of GC±MS analysis of drug from Applied Separations (Allentown, USA).
treatment samples a realistic proposition and The Techne sample concentrator was obtained
would provide a single assay per sample that from Merck (Poole, UK). MSTFA+1% TMCS
would give an unequivocal evaluation of a was obtained from Pierce & Warriner (Chester,
patient’s drug status. UK).
With the advent of mixed-mode solid-phase
extraction (SPE) columns, it has become Solutions
possible to isolate acidic, neutral and basic Acetate buffer, 1´0 mol/L pH 5´0, was prepared
drugs using a single column.3±5 The method by dissolving 42´9 g of sodium acetate trihydrate
presented here, which was developed in-house, in 400 mL of distilled water; 10´4 mL of glacial
uses a mixed-mode SPE extraction column from acetic acid were added and the solution was
which a whole range of drugs of abuse are diluted to 500 mL with distilled water. The pH
extracted simultaneously. The extracted drugs was adjusted to 5´0 with either 1´0 mol/L
are then derivatized with N-methyl-N-trimethyl- potassium hydroxide or 1´0 mol/L hydrochloric
silyltri¯uoroacetamide+1% trimethylchlorosi- acid. Phosphate buffer, 0´1 mol/L pH 6´0, was
lane (MSTFA+1% TMCS) and N-methyl-bis- prepared by dissolving 13´61 g of anhydrous
tri¯uoroacetamide (MBTFA). The use of a potassium dihydrogen phosphate in 900 mL of
three-stage derivatization procedure utilizing distilled water. The pH was adjusted to 6´0 with
MBTFA and MSTFA+TMCS has been pre- 1´0 mol/L potassium hydroxide and the ®nal
viously described.4 The initial addition of solution was made up to 1´0 L with distilled
MBTFA prevented the loss of volatile primary water. Dilute acetic acid, 10 mmol/L pH 3´3,
and secondary amines by derivatization. It was was prepared by diluting 574 mL of glacial acetic
shown that the addition of the derivatization acid to 1´0 L with distilled water.
reagent to the eluent prior to drying down did
not cause problems in the procedure. Preparation of internal standard solution
A GC±MS system in scan mode is used to MDA-d5 was used as an internal standard; 1 mL
detect and identify the drugs. Using Hewlett- of a 100 mg/mL solution of MDA-d5 in
Packard ChemStation software, at least 61 drugs methanol was diluted to 150 mL with acetate
and metabolites are routinely looked for in all buffer, 1´0 mol/L pH 5´0. MDA-d5 is used as the
samples. These include all the commonly abused internal standard because, like all ampheta-
drugs, plus a whole range of other drugs mines, it is volatile. In addition, it is derivatized
including antidepressants and certain antipsy- by both of the derivatizing agents used. It
chotics that are prescribed to these patients. The therefore monitors all these stages of the
samples are incubated prior to extraction with b- analysis.
glucuronidase to enhance the detection of
benzodiazepines6 and morphine.7 Sample pre-treatment
To 4 mL of urine, 1´5 mL of internal standard
solution and 50 mL (6600 units) of b-glucuroni-
MATERIALS AND METHODS
dase were added. The samples were capped and
Materials incubated overnight at 608C. After incubation
All chemicals were of analyticalreagentgrade and the samples were allowed to cool. The pH of
were obtained from Merck (Poole, UK). Deuter- each sample was then adjusted to 8´0 by the
ated methylenedioxyamphetamine (MDA-d5) addition of 445 mL of 1´0 mol/L potassium
was obtained from Promochem (Welwyn Garden hydroxide solution and adjusted further if
City, UK). All drug standards, b-glucuronidase necessary. The samples were centrifuged at
type H-2 (crude solution from Helix pomatia, 3000 rpm for 10 min.
132 500 units/mL; EC 3.2.1.31), MBTFA and
ammonium hydroxide ACS reagent (28´0±30´0% Extraction procedure
NH3) were obtained from Sigma-Aldrich (Poole, The extraction was performed using Bakerbond
UK). Drugs of abuse urine control module narc-2 columns installed in a Spe-ed wiz vacuum
manifold. The procedure consists of the follow- mode. The mass spectrometer was operated with
ing steps: an electron energy of 70 eV and a mass-to-
charge ratio scan range of 50±550 amu. The ion
1. The columns were conditioned with 2´0 mL source and the quadrupole temperatures were
of methanol followed by 2´0 mL of distilled
maintained at 2308C and 1508C, respectively.
water followed by 2´0 mL of 0´1 mol/L
The GC±MS system was programmed to per-
phosphate buffer pH 6´0. form a 1-mL splitless injection.
2. The supernatant from the pre-treated sam- Data acquisition was performed using an HP
ples was added to the column and pulled Vectra XM series 4 computer, running HP
through at a rate of less than 5 mL/min. G1701AA Enhanced Chemstation software
3. The column was washed with 1 mL of (version G1701AA, revision A.03.00). A macro
distilled water followed by 0´250 mL of was written in-house for the software, which
10 mmol/L acetic acid at pH 3´0. performed target compound analysis on each
4. The samples were eluted from the columns sample for the 61 speci®ed drugs. Once isolated
with 3´0 mL of chloroform±isopropyl alco- from the total ion chromatogram, positive
hol (80:20, v/v), followed by 3´0 mL of identi®cation was provided by three libraries: a
chloroform±isopropyl alcohol±ammonium library developed in-house, containing many
hydroxide ACS reagent (80:20:3, v/v). Both drug tri¯uoroacetamide (TFA) and trimethylsi-
eluates were added to the same tube. lyl (TMS) derivatives (274 compounds); the
P¯eger Maurer Webber library (4367 com-
Derivitization procedure
pounds); and the Wiley 275 library (275 821
1. To each eluent, 25 mL of MBTFA were compounds).
added. After mixing, the solutions were
evaporated under nitrogen at 708C to ap- Control material
proximately 1´0 mL using a Techne sample A control was run with each batch of samples to
concentrator. check retention times and performance of the
2. The remaining solution was transferred to a assay. The control material was diluted with
microvial. The solutions were then evapo- blank urine to give the following concentrations
rated to dryness under nitrogen at 708C. in mmol/L (mg/L): d-amphetamine 9´25 (1´25),
3. The vials were removed from the heating benzoylecgonine 1´3 (0´375), codeine 2´5 (0´75),
block and 50 mL of MSTFA±TMCS were methadone 0´8 (0´25), d-methamphetamine 8´4
added to each. The vials were capped and (1´25), morphine 0´26 (0´075), morphine-3-glu-
incubated at 708C for 60 min. curonide 2´85 (1´325), oxazepam 0´875 (0´25), d-
4. Then 25 mL of MBTFA were added to each propoxyphene 0´55 (0´188), secobarbital 2´1
vial and the vials were incubated for a (0´5), tetrahydrocannabinoicacid 0´183 (0´0625).
further 30 min at 708C.
5. A 1-mL aliquot was injected onto the GC± Recoveries
MS system. Solutions of the drugs of abuse listed in Table 3
were prepared and added to blank urine to give
Instrumental analysis ®nal concentrations in urine of 2´0 mg/L for all
A Hewlett-Packard (UK) model 5973 mass drugs except 9-THC-11-oic acid and MDA-d5.
selective detector with a 6890 series gas chroma- As 9-THC-11-oic acid is usually present in
tograph ®tted with a split/splitless injection port, much lower concentrations, this drug was added
an HP G1513A automatic sampler and the HP to blank urine to give a ®nal concentration of
Enhanced Productivity MS Chemstation soft- 0´10 mg/L. Aliquots (4 mL) of these spiked
ware were used. The analytical column was an urine solutions were extracted in the same
HP-5 MS (30 m60´25 mm i.d., 0´25 mm ®lm manner as the samples. An external standard,
thickness), ®tted with an HP Retention Gap 50 mL of a 1´34 mg/L solution of amitriptyline
(uncoated, deactivated) (1 m60´25 mm i.d., 0 in methanol, was added. The extracts were then
mm ®lm thickness). Temperature conditions were derivatized and run on the GC±MS system as
as follows: initial temperature 808C for 2 min; normal. Solutions of the same drugs were
increased at 168C/min to 2908C; held for prepared in chloroform±isopropanol (80:20,
5´37 min; giving a total run time of 20´50 min. v/v) to give the same ®nal concentrations as in
The ¯ow of the carrier gas (helium) was blank urine. Aliquots (4 mL) of these solutions
maintained at 1´0 mL/min in constant ¯ow were taken and to these were added 50 mL of the
FIGURE 1. Total ion chromatogram from a control sample. 1=amphetamine±TFA; 2=methamphetamine±TFA; 3=methylenedioxyamphetamine-d5±TFA/TMS; 4=
secobarbital±TMS; 5=phencyclidine; 6=methadone; 7=benzoylecgonine±TMS; 8=oxazepam±TMS; 9=codeine±TMS; 10=morphine±TMS; 11= 9-THC-11-oic
FIGURE 2. Total ion chromatogram from a patient’s sample. 1=ecgonine methyl ester±TMS; 2=methylenedioxyamphetamine-d5±TFA/TMS; 3=methadone; 4=2-ethyl-5-
methyl-3,3-diphenyl-1-pyrroline (methadone metabolite); 5=methadone±TMS; 6=cocaine; 7=benzoylecgonine±TMS; 8=desmethyldiazepam±TMS; 9=oxazepam±TMS;
10=codeine±TMS; 11=morphine±TMS; 12=temazepam±TMS. TFA=tri¯uoracetamide; TMS=trimethylsilyl.
Analysis of urine for drugs of abuse 695
same external standard solution. The extracts the extraction column is pre-conditioned to pH
were then derivatized and run on the GC±MS 6´0, it was found that the optimum pH for
system as normal. The ratio of the extracted retention of the whole range of drugs of abuse
drug to external standard was compared to the on the extraction column is 8´0. As endogenous
ratio of non-extracted drug to external standard. compounds found in urine are water-soluble,
By comparing these ratios the percentage they were removed from the column by washing
recovery was calculated. with water. After washing, acetic acid was added
to the column adjusting the pH to 3´0. At this
RESULTS pH, acid and neutral drugs are in the unionized
form and therefore elute in chloroform±isopro-
Figures 1 and 2 show total ion chromatograms
pyl alcohol. The basic drugs were then eluted
from the control (Fig. 1) and from a patient’s
with chloroform±isopropyl alcohol made basic
sample (Fig. 2). It can be seen that the
by the addition of ammonia solution. Water was
chromatograms are relatively clean. A simple
found to provide the optimal type and volume of
macro programme was written in-house to
wash for maximum compound retention by the
produce extracted ion chromatograms (EIC)
extraction column, while removing the interfer-
for the detection of the 61 compounds that each
ing endogenous compounds contained in the
sample is routinely screened for. The mass
urine. The inclusion of methanol in the wash
spectroscopist only views the total ion chroma-
caused a decrease in the recovery of the drugs.
togram to see if there are any unusual peaks
Table 3 shows the results of the recovery
present. Table 1 shows the retention times and
experiments for the major drugs of abuse.
mass spectral data used to perform target
Recoveries for 9-THC-11-oic acid, a strong
compound analysis and produce the EIC. A
acid, and for temazepam and oxazepam, neutral
print-out is issued for each sample showing a
drugs, are relatively low but they are acceptable.
chromatogram of the three selected m/z ions for
The eluent was selectively derivatized with
each compound. If the EIC indicates the
MBTFA, which formed tri¯uoroacetamide
presence of a particular compound by showing
(TFA) derivatives of primary and secondary
that the three ions are present and are present in
amines, and MSTFA+1% TMCS, which
approximately the correct ratio, then the spectra
formed trimethylsilyl (TMS) derivatives of
and retention time are matched manually. The
hydroxyl, acidic and phenolic groups. If the
internal standard is not used for quantitation
evaporation step was performed without the
work; it is added to check that the extraction
addition of MBTFA to the eluent it resulted in
and derivatization procedures have been com-
the loss of low molecular weight stimulants, i.e.
pleted successfully. If the internal standard is
amphetamines. Solans et al.4 observed that, after
shown not to be present in the EIC, then the
the addition of MSTFA+1% TMCS to the
sample is re-run. The reference libraries used
dried residue at 808C for 1 h, the formation of
include the P¯eger Maurer Webber library, the
TMS and TFA derivatives was distorted because
Wiley 275 library and an in-house library. The
the TMS took the place of some TFA in some
latter has been obtained by recording the spectra
amine groups. It was therefore necessary to add
of derivatized standards of the various com-
a further 25 mL of MBTFA to complete the
pounds analysed by GC±MS. Figure 3 shows the
derivatization of the primary and secondary
ion chromatogram for morphine±TMS from a
amines.
patient’s sample compared with that of mor-
Some drugs give more than one derivatization
phine±TMS from the in-house reference library.
product. These include methadone, which has
Looking at the EICs alone will show which
both an underivatized and a TMS derivative
drugs are absent, but to be certain of the
peak, and MDA, which has both a TFA and a
presence of those indicated by the EIC it is
TFA/TMS derivative peak. For each of these
always necessary for a mass spectroscopist to
drugs, the major peak was used to estimate the
match the spectra manually.
limit of detection and calculate percentage
recovery, namely the TMS derivative peak for
DISCUSSION
methadone and the TFA/TMS derivative peak
The Bakerbond narc-2 column contains C18 and for MDA.
cationic exchange adsorbents (aromatic sulpho- The method is used to produce qualitative
nic acid) and the drugs are retained by a mixture results only. To be reported as positive, the drug
of ionic and hydrophobic interactions. Although or metabolite must give a signal-to-noise ratio
FIGURE 3. Comparison of ion chromatograms for morphine±TMS from a patient’s sample (top) and morphine±TMS from the in-house reference library (bottom).
Analysis of urine for drugs of abuse 697
TABLE 1. Retention times and mass spectral data used to perform target compound analysis
TABLE 1. Continued
TABLE 2. Limits of detection and proposed UKNEQAS clinical thresholds for the major drugs of abuse and
metabolites
greater than 10. The limits of detection for the the UKNEQAS cut-off for comparison. The
major drugs of abuse and metabolites are shown method therefore has a higher limit of detection
in Table 2. For comparison, the clinical cut-offs for 9-THC-11-oic acid than that proposed by
proposed by the UK National External Quality UKNEQAS, but it has been agreed that the
Assurance Scheme (UKNEQAS) for drugs of method is sensitive enough for the particular
abuse in urine are also shown. The suggested application it is designed for. The limits of
clinical cut-off for the cannabinoid group is detection for all the other drugs of abuse are
0´05 mg/L, with a cut-off of 0´015 mg/L for 9- acceptable for screening for drug treatment
THC-11-oic acid, the speci®c active metabolite. centres.
As the method described here detects 9-THC- The Toxicology Unit had previously per-
11-oic acid, only 0´015 mg/L has been entered as formed urine drug screening for drug treatment
TABLE 3. Recoveries of the major drugs of abuse and metabolites and concentration of drug solution used
only 4 mL urine and is more sensitive for all the sample of analysis using both methods remains
opioid drugs. Immunoassays are only group equivalent.
speci®c for benzodiazepines, opiates, ampheta-
mines and barbiturates;the new method is speci®c
REFERENCES
for each analyte. There is no need for the two-step
approach of screen followed by con®rmation of 1 Braithwaite RA, Jarvie DR, Minty PSB, Simpson
positives. As previously mentioned, the new D, Widdop B. Screening for drugs of abuse. I:
Opiates, amphetamines and cocaine. Ann Clin
method detects not only drugs of abuse and Biochem 1995; 32: 123±53
metabolites but also a whole range of other drugs 2 Department of Health and Human Services.
including antidepressants, antipsychotics and Mandatory guidelines for Federal workplace drug
various over-the-counter preparations. This testing programs. Fed Regist 1997; 62 (no. 189)
means it has the potential for being used to check 3 Chen XH, Franke JP, Wijsbeek J, de Zeeuw RA.
compliance with therapeutic medication and for Isolation of acidic, neutral and basic drugs from
whole blood using a single mixed-mode solid-phase
corroborating the history given by the patient. extraction column. J Anal Toxicol 1992; 16: 351±5
The method also has the potentialto be used semi- 4 Solans A, Carnicero M, de la Torre R, Segura J.
quantitatively. In addition to screening for the 61 Comprehensive screening procedure for detection of
drugs routinely analysed, other drugs speci®ed by stimulants, narcotics, adrenergic drugs and their
the clinician can be detected, provided standard metabolites in human urine. J Anal Toxicol 1995; 19:
spectra are available. 104±14
5 Lai CK, Lee T, Au KM, Chan AY. Uniform solid-
In conclusion, this method shows that the phase extraction procedure for toxicological drug
powerful technique of GC±MS can be used for screening in serum and urine by HPLC with
the routine analysis of urine samples from photodiode-array detection. Clin Chem 1997; 43:
patients attending drug treatment centres. The 312±25
information produced by this method provides 6 Meatherall R. Optimal enzymatic hydrolysis of
the clinician with a more complete picture of the urinary benzodiazepine congugates. J Anal Toxicol
1994; 18: 382±4
drug status of a patient than is currently 7 Romberg RW, Lee L. Comparison of the hydrolysis
available by conventional means. Not only is rates of morphine-3-glucuronide and morphine-6-
there increased information that is more reliable glucuronide with acid and b-glucuronidase. J Anal
and more speci®c than that produced using the Toxicol 1995; 19: 157±62
conventional approach of a combination of
immunoassay, TLC and GC, but the cost per Accepted for publication 2 March 2000