Journal of Analytical Toxicology, Vol. 23, November/December 1999
Methadone Conversionto EDDP during GC-MS
Analysisof Urine Samples*
E Roark Galloway* and Neal F. Bellet
Microgenics Corporation, 4665 Willow Road, Pleasanton,California 94588
Abstract[
During validation of a gas chromatography-massspectrometry
(GC-MS) method for the methadone metabolite 2-ethylidine-l,5dimethyl-3,3-diphenylpyrrolidine(EDDP), it was noted that
detectable levels of EDDP were found during analysisof extracts
from drug-free urine samplesspiked with methadone. Different
amounts of EDDP were detected by GC-MS during confirmation
analysis;however, levels consistentlyexceeded 50 ng/mL at
methadone concentrations> 10,000 ng/mL. Quantitation of EDDP
was determined by the addition of EDDP-d3 to methadone-spiked
urine samples.Subsequentanalysisof methadone-spikedurine
extracts by high-performanceliquid chromatography(HPLC)
indicated no EDDP as a result of contaminated standard or
conversionduring solid-phaseextraction. Reducingthe GC
injector-port temperature from 260~ to 180~ reduced the
observed EDDP concentration in one sample from 201 ng/mL to
53 ng/mL at the initial methadone concentration of 10,000 ng/mL.
These resultsindicate GC injector-port temperature induces
thermal conversionof methadone to EDDP as an artifact. When
confirmation of methadone and EDDP is critical to determining
individual compliance with maintenance programs, alternative
chromatographic methods (e.g., capillary electrophoresis,HPLC,
or liquid chromatography-massspectrometry) should be
considered.
Introduction
Methadone is primarily administered for chronic maintenance treatment of heroin addiction and has pharmacological
properties similar to morphine (1). Some patients undergoing
methadone treatment for recovery from heroin addiction have
attempted to pass compliance testing by adding a portion of
their supplied methadone to their urine, then diverting the
remainder of drug (2). Detection and accurate quantitation
of both methadone and 2-ethylidine-l,5-dimethyl-3,3diphenylpyrrolidine (EDDP) have been determined to be
important factors in exploring metabolism and efficacyof long* Presentedat the combined Societyof ForensicToxicologistsand The InternationalAssociation
of ForensicToxicologistsmeetingheld in Albuquerque, NM, October5, 1998.
Author to whom correspondenceshould be addressed,
term maintenance therapy (3). Urine EDDP/methadone ratio
can also be used to provide an indication of acute methadone
ingestion versus maintenance therapy (4).
Routine urine monitoring of patients to determine
methadone-treatment compliance includes immunoassay
screening and confirmation analysis for both methadone and
EDDP, the primary metabolite. Recent development of EDDP
specific immunoassays to complement available methadone
screening assays for monitoring patient compliance suggest
confirmation methods include detection and quantitation of
both EDDP and methadone (5,6). Reported analytical techniques for detection and quantitation of methadone and EDDP
have included solid-phase extraction (SPE) (7) followedby detection using gas chromatography (GC) (8), gas chromatography-mass spectrometry (GC-MS) (3,9,10), or capillary
electrophoresis (CE) (11). Enantiomeric determinations of
methadone and EDDPby chiral high-performance liquid chromatography (HPLC) (12-14), liquid chromatography-mass
spectrometry (LC-MS) (15), and CE (16,17) have also been reported. An LC-MS-MS method stating methadone and compounds of similar structure (i.e.,propoxyphene)are thermolabile
and heat decomposition products are commonly encountered
was published; however,no information about specificallyidentified decomposition products was included (18).
The data presented here indicate EDDP is produced as an artifact during routine GC-MS ana]ysis of urine sample extracts
containing high concentrations of methadone. The formation
of EDDP is likely to occur within the high-temperature GC injection port as a heat decomposition product.
Experimental
Materials
Methadone and EDDP perchlorate (1.0 mg/mL methanol solutions) and methadone-d3 and EDDP-d3 perchlorate (100
IJg/mL methanol solutions) were obtained from Radian International (Austin, TX, drug concentration listed as free base).
SKF-525Awas obtained from Research Biochemicals International (Natick, MA) as the hydrochloride salt and prepared
as 1.0-mg/mL methanol stock solution. Trifluoroacetic acid
(sequenal grade) was purchased from Pierce Chemical Co.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
615
Journalof AnalyticalToxicology,Vol. 23, November/December1999
(Rockford, IL). All solvents used were HPLC grade and purchased from Fisher Scientific (Fair Lawn, NJ). Bond Elut Certify (Varian Sample Preparation, Harbor City, CA) 3-mL,
130-rag columns were used for urine sample extraction. Drying
gas used during solid-phase extraction was "bone-dry'-grade
air from Praxair (Oakland, CA).
concentrated NH4OH in a glass reagent bottle with closure.
After the solution was mixed, 78 mL methylene chloride was
added to the mixture. The bottle was again capped and contents
mixed. This elution solvent was prepared fresh daily before use.
All reagents were transferred to 500-mL reservoir bottles used
on the RapidTraceTM solid-phase extraction system.
Reagents
Instrumentation
One liter of potassium phosphate buffer was prepared by dissolving 13.6 g potassium phosphate (monobasic) in deionized
(DI) water. After all solids were dissolved, the pH was adjusted
to 6.0 using 5N KOH stock solution. Acetic acid solution (1M)
was prepared by diluting 28.75 mL glacial acetic acid up to 500
mL with DI water in a volumetric flask. Solid-phase extraction
solvent was prepared by combining 20 mL 2-pmpanol and 2 mL
Automated SPE was completed using a RapidTrace (Zymark,
Hopkinton, MA) automated SPE three-module system. The
system was configured for 3-mL standard SPE columns. Control software version 1.20 was used for system operation. The
sample volume extracted was 2 mL.
GC--MSanalysiswas performed with the lip 5890 series II GC
connected to the model 5971A mass-selectivedetector (Hewlett
Packard, Palo Alto, CA). The column installed was a DB5-MS
(15 m x 0.2-ram i.d., 0.33-1Jm phase, J&W Scientific, Folsom
CA). Helium carrier gas was set to 0.5 mL/min constant flow
using electronic pressure control (EPC). The injector had a 4mm internal diameter silanized glass inlet sleeve with a
silanized glass-woolplug. Injector temperature was initially set
to 260~ and detector temperature was set to 280~ The GC
oven temperature was programmed from 130~ (1-min hold) to
300~ at 25~
then held at 300~ for 3 rain. Samples
were injected in splitless mode, and the split valve was programmed to turn on after 1 min. The MSD data were collected
using MS ChemStation G1034C (version C.03.03) in selected
ion monitoring (SIM)mode using ions with the followingmassto-charge ratios (quantitation ions are underlined): EDDP-d3,
280. 265; EDDP,277. 262, 276; and methadone 294, 223, 295.
HPLC analysis was performed on a Shimadzu model SCL-
A
e
i
o
i
. . . .
......
,
....
,
....
"2
~.
700
Time
7SO
sco
elm
(mln)
, ,,.oo(,,,.,o, ,,,.,o):o,o,oo,.o
B
4OO0O
Table I. Measured EDDP Concentrations Resulting from
Indicated Original Methadone Sample Concentration and
the Effects of Reducing GC Injector Temperature
Methadonespike
(pg/mC)
-
-[
'Z
o
....
, ....
1.0
5.0
10
50
1O0
EDDP
, ....
(rain)
Time
MeasuredEDDP(ng/mt)
GC injector : 180~
GC injector : 260~
0
35
53
134
160
0
126
201
547
665
EDDP Recovery: GC Injector Temperature
i-
~ooo-]
Internal standard
EDOP-ds,
18o0r
,-I--
300
I~U.
180 'C
locmH
WI~ 2 0 0
socot
....
, ....
s.so
, . . . . . . .
e.r
~so
, ....
?.oo
Time
, ....
?.8o
, ....
i,oo
, . . . .
0,80
(mln)
I00
0
0
,0
20
30
4o
so
I
t
I
I
I
t
6o
70
so
9o
:00
~,0
Figure 1. TIC (A) and extracted ion chromatograms (8) indicating an
Methadone (pg/mL)
EDDPpeakdetectedat 6.53 min extractedfrom a 10-p~mL methadone
Figure 2. EDDPdose measuredby GC-MSfrom methadonesampleextracts as a function of GC injector temperature.
sample.
616
Journal of Analytical Toxicology, Vol. 23, November/December 1999
sion was occurring in the GC injection port. Attempts to minimize or eliminate the EDDP signal by lowering the injectionport temperature were investigated. Negative urine aliquots
spiked with methadone were extracted and quantitatively analyzed for EDDP by GC-MS. The same extracts were re-analyzed by GC-MS after the injector-port temperature was
reduced to 180~ from 260~ All other method parameters
were unchanged. Signal response for the monitored EDDP
ions increased relative to the EDDP-d3internal standard as the
amount of methadone spiked into samples increased. As a
10ASsystem controller with dual LC-10 solvent pumps, SIL-10
auto sampler with 500-1aLsample loop, and an SPD-10A(V)detector set at 220 nm (Shimadzu, Columbia, MD). Data collection was completed using VP-Class, version 4.2 software. The
column used was a Zorbax Stable Bond (C18,250 x 4.6 mm, 3.5
IJm, 80 A, MacModAnalytical, Chadds Ford, PA). The mobile
phase used was (A) 0.1% TFA in HPLC-grade water and (B)
0.07% TFA in HPLC-grade acetonitrile. The gradient was programmed to run 40% B (1 min) to 100% B in 20 min, wash at
100% B for 5 rain, then re-equilibrate at 40% B for 7 min.
Methods
GC-MS sample preparation. The SPE
method used was from the Bond Elut Certify
instruction manual (7) and adapted for the
RapidTrace system. Sample volume used was
2 mL, and dried sample extracts were dissolved in 150 IJL ethyl acetate. A 2-1JLsample
was injected and analyzed by GC-MS-SIM.
HPLCsampleanalysis. Samples were prepared using the previously described SPE
method. For HPLC evaluation samples however, no internal standard was added, as the
EDDP-d3 signal would interfere with the
HPLC-UV detection of EDDP. Extracts were
dissolved in 150 tJL ethyl acetate and analyzed
by GC-MS-SIM. These extracts were recovered after GC-MS analysis and evaporated to
dryness. Extracts were then dissolved in 125
1JL initial HPLC mobile phase, prepared by
combining 6.0 mL (A) and 4.0 mL (B). A 100I~L sample was injected into the HPLC for
analysis.
TIC: 1401006.0
A
500000
A
450000.
4 0 0 0 0 0 84
u
i
3OOOOO
~, ~ooo
<
W
w
80000
B.88
1
O
r 9 , i
. . . .
4.50
i . . . .
5.00
i . . . .
5.50
i T 1 ,~
e.OO
9i . . . .
I . .
B.50
7.00
i ....
7,50
i ....
i ....
8.00
8.50
i ....
i ....
9.1111 9.50
Time (mln)
~l~wh
~
41KMT~
~ b
im
ell
m
im
Results
EDDP identification
GC-MS-EI chromatograms from analysis of
the extracts produced a significant signal at
the expected retention time for EDDP.A representative total ion chromatogram (TIC) and
extracted ion chromatograms shown in
Figure 1 illustrate results from a 10 IJg/mL
methadone-spiked sample. Analysisof the ions
monitored for data collection at 6.53 rain
show the same ion ratio pattern and retention
time when compared to analyzed EDDP reference material. A negative urine sample extract and solvent blank injected after the
sample extracts indicated no EDDP carryover
from injection port contamination.
~me(mln~
g
i
"
T
Time
Reduction of GC injection-port
temperature
From the previous investigations, it was
suspected that the observed thermal conver-
~
I
-
i
-
T
-
-
'
I
E
"
-
-
T
~
(rain)
Figure 3. GC-MS chromatogram (A) and HPLC chromatogram (B) for extracted 25-1JglmL
methadone sample. ReferenceHPLC chromatogram (C) of sample containing 1.0 iJglml each
EDDP,methadone,and SKF-525A(IS),which was included for comparison.
617
Journal of Analytical Toxicology, Vol. 23, November/December1999
result, the detected amounts of EDDP increased as a function
of methadone concentration in the original sample, indicated
in Table I. Figure 2 is a graphical presentation of the measured EDDP concentration as the injection-port temperature
was reduced. Measured amounts of EDDP were reduced at the
lower injection-port temperature but still detectable in samples
with initial methadone concentrations greater than 10 pg/mL.
Further reduction of injection-port temperature was not attempted because of concerns about incomplete injected-sample
volatilization, column loading, and injector contamination. It
is possible that cool on-column sample injection could eliminate the observed conversion; however, the instrumentation
used at this facility was not equipped for such an investigation.
Methadone-spiked sample results, HPLC comparison
The extracts analyzed by GC-MS clearly indicate the large
methadone peak detected at 6.91 rain, preceded by a detected
EDDP peak at 6.38 rain. Figure 3 illustrates the resulting
chromatograrns obtained by GC-MS and HPLC for the solidphase extract of a 25-pg/mL methadone sample. The corresponding SIM spectrum observed produced ion ratios
corresponding to EDDP reference material. Integration of the
peaks indicated the EDDP signal generated was 8.49% relative
to methadone at 10 pg/mL, 6.13% at 25 pg/mL, and 6.29% at
50 pg/mL initial methadone concentration. The EDDP concentrations were not determined by GC-MS from this experiment because no EDDP-d3 internal standard or calibration
standards were included as part of the analyses. The GC peak at
6.38 rain and SIM spectra were readily identified as EDDP
when compared to a reference standard.
Figure 3C shows an HPLC reference standard containing
EDDP, methadone, and SKF-525Athat was analyzed to provide
retention information for the expected compounds (20 I~L
injected from a 100-tJg/mLstandard). The corresponding HPLC
chromatograms from the sample extracts did not produce any
corresponding EDDP peak preceding the methadone peak.
Table II lists the chromatographic signal intensities detected
for the samples analyzed by GC-MS and HPLC. If EDDP had
been an actual contaminant from the methadone spike material or been a conversion product resulting from the SPE
preparation procedure, detection of EDDP by the HPLC
method should have correlated with GC data observed and
produced a detectable peak on the HPLC chromatogram. The
HPLC result indicated no integrated peak at the expected retention time of EDDP having 5-10% peak area relative to
methadone, as previously observed from GC-MS results.
Table II. Chromatographic EDDP and Methadone Signal
ResponsesMeasured by GC-MS and HPLC
Responsesignal
Methadone spike
GC-MS
(mg/mL)
EDDP methadone
10
25
50
618
338781
449031
626449
3988225
7320245
9945731
EDDP
HPLC
methadone
No signal 9956832
Nosignal 13604536
No signal 37890788
Discussion
These results indicate that conversion of methadone to
EDDP can occur during routine GC-MS analysis of urine
methadone samples. In reality, samples containing large concentrations of methadone with very minimal amounts of EDDP
would automatically be subject to further investigation. However, laboratories that use both methadone- and EDDPscreening tests could produce conflicting results between
screening and GC-MS confirmation. The significance of these
findings was crucial in evaluating performance of an EDDPspecific immunoassay (5) because several evaluation samples
testing negative by irnmunoassay and HPLC produced measurable EDDP concentrations by GC-MS.
Conclusions
GC-MS analysis can produce an artifact peak detected as
EDDP in samples containing high concentrations of
methadone, which is typical of "spiked" urine samples. This
phenomenon can also produce falsely elevated EDDP levels
detected in true methadone-positive samples. The formation of
EDDP is likely the result of thermal degradation of methadone
in the GC injector port. HPLC, LC-MS, or CE analysis must be
considered as alternative methods used for confirmation analysis of suspected "spiked" urine to avoid inaccurate EDDP
GC-MS results when monitoring methadone compliance by
urine testing.
References
1. R.C. Baselt and R.H. Crave,/. Disposition of Toxic Drugs and
Chemicals in Man, 3rd ed. Year Book Medical Publishers,
Chicago, IL, 1989, pp 512-516.
2. F.R. Goldman and C.I. Thistel. Diversion of methadone: Illicit
methadone use among applicants to two metropolitan drug abuse
programs. Int. J. Addict. 13(6): 855-862 (1978).
3. M.E. Alburges, W. Huang, R.L. Foltz, and D.E. Mood`/. Determination of methadone and its N-demeth,/lation metabolites in biological specimens by GC-PICI-MS. J. Anal. Toxicol. 20:362-368
(1996).
4. A.C. Moffat, Ed. Clarke's Isolation and Identification of Drugs, 2nd
ed. The Pharmaceutical Press, London, U.K., 1986, pp 742-743.
5. CEDIA | EDDP application sheet. Microgenics Corporation,
Pleasanton, CA.
6. Methadone Metabolite Enzyme Immunoassa,/application sheet.
Diagnostic Reagents, Inc., Mountain View, CA.
7. Bond Elut CERTIFY application booklet. Varian Sample Preparation, Harbor City, CA.
8. P. Jacob, III, J.F. Rigod, S.M. Pond, and N.L. Benowitz. Determination of methadone and its primary metabolite in biologic fluids
using gas chromatograph`/with nitrogen-phosphorus determination. J. Anal. Toxicol. 5:292-295 (1981 ).
9. O. Beck, L.O. Bor~us, S. Borg, G. Jacobsson, P. Lafolie, and M.
Stensi~. Monitoring of plasma methadone: Intercorrelation between immunoassa,/and gas chromatograph,/-mass spectrometry. Ther. Drug Monit. 12:473-477 (1990).
Journal of Analytical Toxicology,Vol. 23, November/December1999
10. L.D. Baugh, R.H. Liu, and A.S. Walia. Simultaneous gas chromatography/mass spectrometry assay of methadone and 2-ethyl1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP)in urine. J. Forensic
Sci. 36(2): 548-555 (1991).
11. S. Molteni, J. Caslavska, D. Allemann, and W. Thormann. Determination of methadone and its primary metabolite in human
urine by capillary electrophoretic techniques. J. Chromatogr. B
658(2): 355-367 (1994).
12. O. Beck, L.O. Bor~us, P. Lafolie, and G. Jacobsson. Chiral analysis of methadone in plasma by high performance liquid chromatography. J. Chromatogr. 570:198-202 (1991).
13. R.L.G. Norris, RJ. Ravenscroft, and S.M. Pond. Sensitive high-performance liquid chromatographic assaywith ultraviolet detection
of methadone enantiomers in plasma. J. Chromatogr. B 661:
346-350 (1994).
14. C. Pham-Huy, N. Chikhi-Choffi, H. Galons, N. Sadeg, X. Laqueille,
N. Aymard, E Massicot, J.M. Wamet, and J.R. Claude. Enantioselective high-performance liquid chromatography determination
of methadone enantiomers and its major metabolite in human biological fluids using a new derivatized cyclodextrin-bonded phase.
J. Chromatogr. B 700:155-163 (1997).
15. P. Kintz, H.P. Eser, A. Tracqui, M. Moeller, V. Cirimele, and P.
Mangin. Enantioselective separation of methadone and its main
metabolite in human hair by liquid chromatography/ion spraymass spectrometry. J. Forensic 5ci. 42(2): 291-295 (1997).
16. M. Lanz and W. Thormann. Characterization of the stereoselective metabolism of methadone and its primary metabolite via cyclodextrin capillary electrophoretic determination of their urinary
enantiomers, Electrophoresis 17:1945-I 949 (1996).
17. M. Frost, H. K0hler, and G. Blasche. Enantioselective determination of methadone and its main metabolite 2-ethylidene-1,5dimethyl-3,3-diphenylpyrrolidine (EDDP) in serum, urine and
hair by capillary electrophoresis. Electrophoresis 18:1026-1034
(1997).
18. A.M.A. Verweij, M.L. Hordijk, and P.J.L. Lipman. Quantitative
liquid chromatographic thermospray-tandem mass spectrometric
analysis of some analgesics and tranquilizers of the methadone,
butyrophenone, or diphenylbutylpiperidine groups in whole
blood. J. Anal. Toxicol. 19:65-68 (I 995).
Manuscript received November 17, 1998;
revision received January 21, 1999.
619