The United States Public Health Service Substance
Abuse and Mental Health Services Administration is
alerting medical professionals that a substantial
percentage of cocaine imported into the United
States is adulterated with levamisole, a veterinary
pharmaceutical that can cause blood cell disorders
such as severe neutropenia and agranulocytosis.
Levamisole HCl is the active ingredient in a number
of veterinary drugs approved to treat worm
infestations in animals. Levamisole HCl was also the
active ingredient in a human drug for oral
administration approved on June 18, 1990, as
adjuvant treatment in combination with fluorouracil
after surgical resection in patients with Duke’s stage
C colon cancer.This drug was withdrawn from the
U.S. market around 2000, and it has not been
marketed in the U.S. since then.The objective of this
study was to develop a method to determine the
amount of levamisole in urine samples.The
procedure will be provided to state health
laboratories as needed to be used in the evaluation of
patients that have developed neutropenia or
agranulocytosis in the setting of recent cocaine use. A
gas chromatography–mass spectrometry method was
validated and tested at two different laboratories, and
the method limit of detection for levamisole is 1
ng/mL in urine when using a 5-mL sample.
Confirmation of the stereoisomer of levamisole was
done by high-performance liquid chromatography
using a chiral column.
The United States (U.S.) Substance Abuse and Mental Health
Services Administration issued an alert on September 21, 2009,
warning of numerous cases of agranulocytosis, an acute blood
disorder consisting of a substantial reduction in certain types
of white blood cells. These cases appeared among illicit drug
users in the setting of recent cocaine use adulterated with lev-
amisole (Figure 1), a veterinary pharmaceutical (1). As of July
2009, according to the U.S. Drug Enforcement Administra-
tion 69% of seized cocaine lots entering the U.S. contained lev-
amisole(2).Levamisole is added at approximately6% by weight
to the cocaine hydrochloride bricks (3) with the intent to en-
hance the effects of cocaine (4).
Levamisole had been used historically in humans for such
conditions as cancer and rheumatoid arthritis; hence, its ad-
verse effects on blood cells are known. A study of 60 patients re-
ceiving levamisole for the treatment of rheumatoid arthritis
found that 35% of the patients suffered a persistent decrease in
neutrophil counts (5). Ingestion of cocaine containing lev-
amisole can also cause agranulocytosis (6,7). Snorting,
smoking, or injecting cocaine containing levamisole has been
Determination of Levamisole in Urine by
Gas Chromatography–Mass Spectrometry*
Michael L.Trehy1,†, Daniel J. Brown2, JeffreyT. Woodruff1, Benjamin J. Westenberger1, William G. Nychis3,
Nicholas Reuter4, Joshua G. Schier5, Sara J.Vagi5, and Rong-Jen Hwang6
1U.S. FDA/CDER/DPA, 1114 Market Street, St. Louis, Missouri 63101;2U.S. FDA/ORA, Regional Field Office,
300 River Place, Detroit, Michigan 48207;3U.S. FDA/CDER, Office of Compliance, 10903 New Hampshire Ave, Bldg. 51,
Silver Spring, Maryland 20993;4U.S. PHS/SAMHSA, One Choke Cherry Road, Rockville, Maryland 20857;
5CDC/NCEH/EHHE/HSB MS F-57, 4770 Buford Hwy. NE, Chamblee, Georgia 30341; and6New Mexico DH,
P.O. Box 4700, Albuquerque, New Mexico 87196-4700
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Journal of AnalyticalToxicology,Vol. 35, October 2011
* The findings and conclusions in this article have not been formally disseminated by the Food
and Drug Administration and should not be construed to represent any Agency determination
†Author to whom correspondence should be addressed.
Figure 1. Structure of levamisole, S-(–)-2,3,5,6-tetrahydro-6-phenylimidazo
Journal of AnalyticalToxicology,Vol. 35, October 2011
associated with rapidly developing, life threatening infections
due to the development of neutropenia and agranulocytosis
Pharmacokinetic studies for levamisole have shown that
levamisole is eliminated from plasma with a half-life of ap-
proximately 5.6 h in one study (8) and 241 min in a second
study (9). Only approximately 3.2% of the oral dose was re-
covered as unchanged drug in the urine (8).
An analytical profile for levamisole that summarizes the
physical and spectrometric properties for the determination of
levamisole in illegal drug samples has been prepared (10). Gas
chromatographic (GC) methods have been developed for lev-
amisole in plasma and in animal tissues using a nitrogen-se-
lective thermionic specific detector (8,11) with detection limits
as low as 2 ng/mL in urine.
This gas chromatographic–mass spectrometric (GC–MS)
method was developed and validated to provide a method for
health officials to measure levamisole in urine when unex-
plained occurrences of agranulocytosis occur. Patients will
likely be unaware that they have been exposed to levamisole.
Levamisole hydrochloride, S-(–)-2,3,5,6-tetrahydro-6-
phenylimidazo [2,1-b]thiazole hydrochloride, was purchased
from United States Pharmacopeia lot #F2C122. Tetramisole
hydrochloride, (±)-2,3,5,6-tetrahydro-6-phenylimidazo [2,1-
b]thiazole hydrochloride, was purchased from Acros. Cyhep-
tamide and isoamyl alcohol (3-methyl-1-butanol) Reagent-
Plus®were purchased from Sigma-Aldrich. EMD OminiSolv
grade hexane was purchased from Fisher Scientific.
An Agilent 6890 GC/5975B MS was used in the method val-
idation with Enhanced ChemStation software version MSD
ChemStation D.03.00.552. A Varian 450-GC/Varian 320-MS
TQ MS with Varian MS workstation system control (version
6.9.1) was used in the analysis of urine samples collected in
New Mexico. A 1-µL injection volume was used with splitless
injection for 1 min. An Agilent DB-5MS UI column (30 m ×
0.25 mm, 0.25-µm film thickness) was installed, and a constant
helium carrier gas flow of 1 mL/min was used in the validation
testing with fortified urine samples. In the analysis of urine
samples from New Mexico, two additional columns were em-
ployed. An HP-5MS (15 m × 0.25-mm i.d., 0.25-µm film thick-
ness) column and a DB-17MS (30 m × 0.25-mm i.d., 0.25-µm
film thickness) column were used. The column oven program
initial temperature was 70°C with a 1-min hold, then increased
to 270°C at 20°C/min, and held for 4 min. The transfer line was
set at 280°C, injection port at 250°C, and MS quad at 150°C.
The electron multiplier was set at 200 volts above autotune
value. The source was operated in electron impact mode and
set at 230°C. Selected ion monitoring is used to increase sen-
sitivity. In the method validation using the DB-5MS UI 30-m
column, data acquisition began at 9.5 min monitoring ionsm/z
148 and 204 both with dwell time of 100 ms then switched at
11.6 min to ion m/z 193 with dwell time of 100 ms. Using al-
ternative columns requires determining the actual retention
time of the analytes and adjusting the times required for mon-
itoring the analytes of interest. Levamisole chromatography is
sensitive both to the column type and condition of the column.
Peak tailing significantly increases with column use. After a
year of use in analyzing a variety of samples, removal of the
first meter of the column was required to restore the DB-5MS
UI column to performance similar to that when it was first in-
High-performance liquid chromatography (HPLC)
conditions for chiral separation of tetramisole
Levamisole is one of the two stereo isomers of tetramisole.
One of the samples obtained from the New Mexico Depart-
ment of Health contained 1.2 µg/mL levamisole based on the
GC–MS results. The GC–MS method does not differentiate be-
tween to the two possible stereo isomers. In order to verify that
levamisole was being used and not tetramisole a chiral sepa-
ration was employed on this sample. The separation of the
stereoisomers of tetramisole can be carried by HPLC (12,13).
A Phenomenex Lux 5u Amylose-2 column (250 mm × 4.6 mm)
was used for the separation of the isomers. An Agilent 1290
HPLC system was used for the analysis with a flow rate of
1 mL/min with an eluent composed of 95% acetonitrile/5%
IPA/0.1% DEA and an injection volume of 10 µL. The same
sample extract used for GC–MS analysis was used for the HPLC
Preparation of extraction solution and
internal standard solutions
Five-hundred milliliters of extraction solution was prepared
by transferring 25 mL of isoamyl alcohol to a 500-mL volu-
metric and diluting to volume with hexane. A stock internal
standard solution was prepared by transferring approximately
5 mg of cyheptamide accurately weighed to a 50-mL volu-
metric and diluting to volume with methanol. A working in-
ternal standard solution was prepared by transferring 2 mL of
stock internal standard solution to a 100-mL volumetric and
diluting to volume with methanol containing 1 mL of con-
centrated ammonium hydroxide per liter of methanol in order
to convert the levamisole HCl to its neutral form.
Preparation of standard solutions
Stock levamisole (mw = 204.29) standard solution was pre-
pared by transferring approximately 1.5 mg accurately weighed
levamisole HCl (mw = 240.75) to a 10-mL volumetric and di-
luting to volume with working internal standard solution. A se-
ries of standard solutions was prepared by serial dilution with
working internal standard solution to the analytical range of
interest including the limit of detection (LOD) of 25 ng lev-
amisole per milliliter of solvent. A 25 ng/mL standard is equiv-
alent to a urine sample at 1 ng/mL when extracting a 5-mL
sample with 0.2 mL internal standard solution. The 1 M
sodium hydroxide solution was prepared by transferring 4 g of
sodium hydroxide to a 100-mL volumetric, adding deionized
water to volume, and mixing until dissolved.
Journal of AnalyticalToxicology,Vol. 35, October 2011
Preparation of urine samples for analysis
A 5-mL sample of urine was transferred to a screw-top test
tube when sufficient sample was available and 0.2 mL of
working internal standard solution was added to the test tube.
Two milliliters of 1 M sodium hydroxide and 3 mL of 95:5
hexane/isoamyl alcohol extraction solution were added. The
and then centrifuged using a bench-top centrifuge to break
the emulsion. The upper layer (hexane/isoamyl alcohol) was
transferred to a 5-mL Reacti-vial with a transfer pipette. One
milliliter of hexane was added to the test tube without shaking
and then transferred to the 5-mL Reacti-vial while taking care
to avoid transferring water to the Reacti-vial. The Reacti-vial
was placed in a heating block at 55ºC ± 5ºC and evaporated to
approximately 0.05 mL with a gentle stream of nitrogen. After
30 s. The sample extract was transferred to an autosampler vial
containing 0.25 mL insert for analysis.
Calculation of results
Areas for selected ions were determined at the retention
time for levamisole, fragment ion m/z 148 and molecular ion
m/z 204, and for cyheptamide, fragment ion m/z 193. The
molecular ion 204 is used for quantitation of levamisole. The
ratio of the area for ions m/z 204 to 193 is calculated for stan-
dards and samples. Linear regression analysis is carried out
comparing the ratio of the area for ion 204 to ion 193 versus
the levamisole concentration in the standards. The slope and
intercept obtained from the linear regression analysis are used
to calculate the concentration of levamisole present in the
urine sample extract. The concentration of levamisole in the
urine is determined by multiplying the concentration deter-
mined in the extract by the concentration factor used.
A concentration factor of 25 was used in this work (5 mL/
0.2 mL = 25).
Confirmation of identity
Ion m/z 204 is the levamisole ion used for quantitation, and
ion m/z 148 is a qualifying fragment ion for levamisole. The
area ratio for ions m/z 204 to 148 obtained with unknown
samples should be within ±20% of the ratio calculated for the
levamisole standards for confirmation of identity.
Standard solutions prepared over the concentration range
from 6 to 8300 ng/mL in working internal standard solution
were analyzed to determine linearity, LOD, and limit of quan-
titation (LOQ) for levamisole following ICH guidelines (14).
Urine samples were spiked at four concentrations from 0.8 to
29.7 ng/mL in triplicate. The method robustness was evalu-
ated by testing the method at two laboratories after initial val-
idation was completed. Spiked samples were prepared in trip-
licate and analyzed to determine method recoveries.
Results and Discussion
Method development and optimization
An internal standard technique was developed in order to
avoid problems due to final volume dilution and incomplete re-
covery of sample in the extraction process. Calculations were
minimized by using the same internal standard concentration
in the standards and in the extraction solution. A standard
curve was prepared by plotting the ratio of the area for lev-
amisole (m/z 204) to the area for the internal standard (m/z
193) versus the nanogram of levamisole per milliliter in the
standard solution. The line of best fit was calculated and used
to determine the concentration of levamisole in the urine
sample after an adjustment for the concentration factor of 25.
The standard deviation for five replicate injections of a 223
ng/mL standard was determined to be 3.6%. The linear re-
gression of the response for standards prepared at 6, 20, 66.9,
and 223 ng levamisole/mL was carried out. The correlation co-
efficient squared was 0.992 with a slope of 6.62 × 10–5and in-
as shown in Table I.
Initial analysis was attempted using a thin-film (0.1 µm)
DB-5 column. This column gave tailing factors greater than 2
and consequently made trace analysis difficult. Replacement of
the column with a 0.25-µm DB-5 ultra inert (UI) column re-
sulted in a tailing factor of 1.29 and an improved LOD. The DB-
5 UI column was used for method validation. An HP-5MS (15
m × 0.25 mm, 0.25-µm film thickness) column had a tailing
factor of 2.5 and a DB-17MS (30 m × 0.25 mm, 0.25-µm film
thickness) column had a tailing factor of 1.1 with a 1-µL in-
jection of a 4.8 µg/mL levamisole standard.
After completion of the method validation protocol, the
method was provided to two analysts to evaluate the method
robustness. Analyst 1 used the same GC–MS system used in the
method validation with the substitution of a used DB-5 MS
column with 0.25-µm film thickness at the FDA St. Louis Lab-
oratory. Analyst 2 used an Agilent GC with a mass selective de-
tector (MSD) and a new DB-5 MS column with 0.25-µm film
thickness at the FDA Detroit Laboratory. Analyst 1 obtained
99% recovery of a sample fortified at 8 ng levamisole/mL in
urine, and analyst 2 prepared triplicate samples fortified at
Table I. Spiked Recoveries for Levamisole and Standard
ng levamisole/mL urine
Journal of AnalyticalToxicology,Vol. 35, October 2011
61 ng/mL and obtained blank corrected recoveries of 122.2%,
116.4%, and 113.9% with an average of 117.5% and a percent
relative standard (%RSD) deviation of 3.6%. Precision was
evaluated by determining the %RSD for an approximately 274
ng/mL standard. Analyst 1 had a %RSD of 2.4% with a tailing
factor of 2.9, and analyst 2 obtained a %RSD of 1.6% and a
tailing factor of 1.38.
Stability of the Agilent DB-5MS (30 m
× 0.25 mm, 0.25-µm film thickness) UI
column over a 15-month period was eval-
uated. The column, when initially in-
stalled, had a signal-to-noise peak-to-peak
(S/N) for a 0.223 µg/mL sample of 27 with
a tailing factor of 1.50. After 15 months of
use running a variety of samples, the
column performance had degraded sig-
nificantly and had an S/N of 9 with a
tailing factor of 8.25 with a 0.274 µg/mL
sample. After removal of the first meter of
column the column performance was re-
stored and had an S/N of 36 and tailing
factor of 1.50 with the 0.274 µg/mL
Analysis of specimens
Samples of urine have been analyzed
by health agencies to determine if the
patient’s illness can be linked to lev-
amisole. As part of the method valida-
tion samples previously analyzed by the
New Mexico Health Department were
stored frozen and shipped to the St.
Louis, MO Food and Drug laboratory for
comparative analysis. Results are tabu-
lated in Table II.
The St. Louis, MO Food and Drug Ad-
minstration laboratory analyzed reserve
samples that had been stored at the New
Mexico laboratory. The two methods’ results for samples
2010203487 and 800900598 were similar but differed from
samples 2010203510 and 2010202911. Using the FDA-se-
lected ion monitoring method, all the patients were found to
have at least trace levels of levamisole present. These pa-
tients had tested positive for cocaine or opiates, and lev-
amisole is now commonly found as a cutting agent with co-
Figure 2. Sample extract of urine from patient with agranulocytosis who had taken cocaine within 5 days
of sample collection. Internal standard peak at 12.77 min ion m/z 193 and levamisole peak at 11.04 min
ion m/z 204. Concentration of levamisole in urine was determined to be 0.075 µg/mL.
Table II. Comparison of Results for the Determination of Levamisole in Full Scan GC–MS at the NM SLD with the Selected
Ion Monitoring Method Results Obtained at the FDA Laboratory
Frame* Case Notes
2010203487Negative 1.1 < LOQUnknownTested positive for cocaine but arrived at the hospital unresponsive and later died.
800900598 Positive12129 hCocaine from crack pipe also tested positive (tested by law enforcement out of
state, not SLD).
2010203510 Positive 1.0 < LOQUnknown Tested negative for cocaine.Tested positive for opiates but could not determine
if it was heroin. Unable to contact patient for interview; still trying. Reoccurrence.
2010202911Negative9.3 UnknownThis patient was not interviewed because it was a reoccurrence, and he had
already been interviewed. He tested positive for cocaine, but reported last use
was 2 months prior.
* Approximate time elapsed between last cocaine exposure and specimen collection.
Journal of AnalyticalToxicology,Vol. 35, October 2011
caine (2). A possible explanation for the difference in the re-
sults is that selected ion monitoring is more sensitive than
full scan MS and that levamisole is not stable in urine even
An additional urine sample from a patient with unexplained
agranulocytosis was sent to the FDA for analysis. This patient
had used cocaine within 5 days of the urine analysis but the
exact timing is uncertain. Approximately 2.7 mL of sample
was available for analysis rather than 5 mL. The sample had
been refrigerated after collection and shipped to St. Louis for
analysis. Extraction was done 14 days after original collection.
Sample chromatogram is shown in Figure 2 and was found to
contain 0.075 µg/mL of levamisole. Given the half-life of ap-
proximately 6 h in-vivo for levamisole, it was remarkable to
find levamisole still detectable at easily detectable levels.
Confirmation of stereoisomer
The GC–MS method does not differentiate between the two
stereoisomers of tetramisole. The high concentration of lev-
amisole found in the urine of New Mexico patient 800900598
provided an opportunity to determine if the patient was ex-
posed to a racemic mixture, (±)-2,3,5,6-tetrahydro-6-phenyl-
imidazo[2,1-b]thiazole, or to levamisole alone. This chiral sep-
aration has been reported using chiral
HPLC columns (13,14).
The concentration of levamisole in the
extract of this patient would be approxi-
mately 30 µg/mL, which allowed for de-
tection by HPLC–UV detection. This same
approach should be applicable to HPLC–
MS or HPLC–MS–MS with possibly even
lower LODs. Figure 3 shows the HPLC
chromatograms for the mixed stereo-
isomer tetramisole standard, for the lev-
amisole standard, and for the urine ex-
tract sample using the Phenomenex Lux
5µ Amylose-2 column. Based on the area
percent obtained for the two isomers, the
isomer distribution in the urine extract
was approximately 87% levamisole.
A method for levamisole was developed
and validated to provide health officials
and analysts a means to determine if pa-
tients with signs and symptoms of severe
neutropenia or agranulocytosis have been
exposed to levamisole. The method has
an LOD of 1 ng/mL in a 5-mL sample.
Recoveries were 99% at 8 ng/mL and
117.5% at 61 ng/mL. Method applicability
to patient samples was demonstrated.
The authors would like to acknowledge
the assistance of Professor Geoff Baird at
the University of Washington in evalu-
ating the method and Tammy Hshieh at
the Beth Israel Deaconess Medical Center
for assistance. The authors are particu-
larly grateful to Monica Brackney and
Twila R. Kunde at the New Mexico De-
partment of Health for providing samples
for method validation.
Figure 3. Reconstructed HPLC chromatograms with a Phenomenex Lux 5-µmAmylose-2 250- × 4.6-mm
column: tetramisole (top), levamisole (middle), and urine extract (bottom). Retention time of levamisole
is 6.1 min, and R-isomer has a retention time for 5.3 min.
Journal of AnalyticalToxicology,Vol. 35, October 2011 Download full-text
1. United States Substance Abuse and Mental Health Services Ad-
ministration. Nationwide public health alert issued concerning
life-threatening risk posed by cocaine laced with veterinary anti-
parasite drug. September 21, 2009.
2. Centers for Disease Control. Agranulocytosis associated with co-
caine use—four states, March 2008–November 2009. MMWR
Morb. Mortal. Wkly. Rep. 58(49): 1381–1385 (2009).
3. J.F. Casale, E.M. Corbeil, and P.A. Hays. Identification of lev-
amisole impurities found in illicit cocaine exhibits. Microgram J.
6(1-6): 82–89 (2008).
4. L.P. Raymon and D.S. Isenschmid.The possible role of levamisole
in illicit cocaine preparations. J. Anal. Toxicol. 33: 620–622
5. G.T. Williams, S.A.N. Johnson, P.A. Dieppe, and E.C. Huskisson,
Neutropenia during treatment of rheumatoid arthritis with lev-
amisole. Ann. Rheum. Dis. 37: 366–369 (1978).
6. N.Y. Zhu, D.F. Legatt, and A.R.Turner. Agranulocytosis after con-
sumption of cocaine adulterated with levamisole. Ann. Intern.
Med. 150(4): 287–289 (2009).
7. D.R. Czuchlewski, M. Brackney, C. Ewers, J. Manna,
M.H. Fekrazad, A. Martinez, K.B. Nolte, B. Hjelle, I. Rabinowitz,
B.R. Curtis, J.G. McFarland, J. Baumbach, and K. Foucar. Clini-
copathologic features of agranulocytosis in the setting of lev-
amisole-tainted cocaine. Am.J.Clin.Pathol. 133: 466–472 (2010).
8. E. Kouassi, G. Caillé, L. Léry, L. Larivière, and M.Vézina. Novel
assay and pharmacokinetics of levamisole and p-hydroxyle-
vamisole in human plasma and urine. Biopharm.Drug Dispos. 7:
9. J.M. Reid, J.S. Kovach, M.J. O’Connell, P.G. Bagniewski, and
C.G. Moertel. Clinical and pharmacokinetic studies of high-dose
levamisole in combination with 5-fluoroacil in patients with ad-
vanced cancer Cancer Chemother. Pharmacol. 41: 477–484
10. A.M.M. Valentino and K. Fuentecilla. Levamisole: an analytical
profile Microgram J. 3(3-4): 134–137 (2005).
11. R.Woestenborghs, L. Michielsen, and J. Heykants. Determination
of levamisole in plasma and animal tissues by gas chromatog-
raphy with thermionic specific detection. J. Chromatogr. 224:
12. International Conference on Harmonization (ICH),Validation of
analytical methods: methodology, ICH Q2 B, 1996
13. X.-J. Cai, X.-Z. Xu, and C.-X. Pan. Study of optical isomer sepa-
ration of chiral antifungal drugs tetramisole, miconazole, and
paclobutrazol on two chiral stationary phases. Anal. Lett. 38:
14. L. Peng, S. Jayapalan, B. Chankvetadze, andT. Farkas. Reversed-
phase chiral HPLC and LC/MS analysis with tris(chloromethyl-
phenylcarbamate) derivatives of cellulose and amylose as chiral
stationary phases. J. Chromatogr.A 1217: 6942–6955 (2010).
Manuscript received February 11, 2011;
revision received June 8, 2011.