False-positive results for urine amphetamine and opiate immunoassays in a patient intoxicated with perazine

Full text

o:
9I
11S12 34perazlne
:::B,jfh \
32
(SI 1S2 perazne
::: JL
time (mm) 205 nm
Fig. 1. HPLC-chromatograms of 1:5 (A)and 1:10 (B diluted urine of a perazine intoxicated
patient; 3000 g/L perazine solution dissolved in drug.free urine (C).
Samples contained 1000 Lof diluted urine or perazine solution and 200 L of internal standard solution.
After liquid-chromatographic separation (HPLC).a continuous detection of UVspectra (200-300 nm; 1-nm
steps) was performed. Chromatogramspresentedhere were detected at 205 nm. AU, absorptionunits; IS1,
internal standard 1(Nethylnordiazepam): 1S2, internal standard 2 (chlorpheniramine); 1-4. perazine
metabolites. In fivefold diluted urine, neither Si nor 1S2was found: in 10-fold diluted urine: no 1S2could be
identified.
Perazine,
g/i
Patient’s
urIne 0.1 1
64 neg. 379 pos.
Neg. Neg.
77 neg. 184 pos.
Neg. Neg.
ClinicalChemistry 42, No. 10, 1996 1725
False-Positive Results for Urine
Amphetamine and Opiate
immunoassays in a Patient
Intoxicated with Perazine
To the Editor:
Analytical interference has to be consid-
ered when immunoassays are used in
clinical toxicology and urine testing for
drugs of abuse [1].Group-specific im-
munoassays such as urine tests for am-
phetamines [2, 3] or opiates [4] are par-
ticularly prone to this problem.
We became aware of this problem
when we performed a drug-screening
test on the urine of an unconscious,
intoxicated female who had ingested -
gof perazine and was known to be poly-
toxicomenic. We wanted to know if she
had used any additional drugs of abuse.
The urine drug-screening included rou-
tinely liquid chromatography (HPLC,
Remedi; Bio-Rad, Anaheim, CA) and
four fluorescence polarization immunoas-
says[FPIA (ADx; Abbott, Chicago, IL)] in
sediment-free urine with and without hy-
drolysis [/3-glucuronidase (Sigma Chemical
Co., St. Louis, MO); final activity 1000
kU/L, 30 mm, 56#{176}CatpH 5].
Immunological results were positive
for opiates and amphetamines, including
their related compounds, but negative for
cocaine (benzoylecgonine) and cannabis
metabolites. In the case of HPLC, the
native urine contained very high amounts
of different substances and had to be
diluted five- and 10-fold to achieve an
acceptable separation (Fig. lA,B). None
of these peaks could be identified or
correlated with a substance class because
no similar spectra were in the spectra
library at that time, and peaks cochro-
matographed with internal standard 1
and (or) 2 (Fig. 1A,B). Five peaks repre-
senting the majority of substances (reten-
tion times 3.0, 4.6, 6.4, 14.1, and 15.3
mm) had nearly identical ultraviolet
(UV) spectra.
In contrast to FPIA immunoassays
(ADx), no opiates and amphetamines
were found by HPLC-whether because
of the absence of opiates and amphet-
amines or because dilution took them
below the detection limit. However, this
indicated that high urinary concentra-
tions of these substances were unlikely.
Hydrolysis neither improved separation
nor led to the detection of opiates and
(or) amphetamines. To verify both posi-
tive ADx results, we performed another
immunoassay (Emit-st; Syva Co., San
Jose, CA). Surprisingly, we found no
opiates or amphetamines, the first hint of
cross-reactivity of the FPIA with one or
more substances in the urine sample.
Finally, we could exclude the additional
ingestion of drugs other than perazine by
gas chromatography-mass spectroscopy
(Hewlett-Packard, Palo Alto, CA). Only
the neuroleptic drug perazine and its
metabolites were detected.
The discrepant results between ADx
and the other tests prompted us to exam-
ine perazine as pure compound (dis-
solved in drug-free urine) in vitro by
HPLC and immunological group tests
for opiates and amphetamines. No infor-
mation about cross-reactivity of perazine
with these immunoassays was available
for both systems from the manufacturers
or the literature.
Perazine could be detected at concen-
trations of >1000 jg/L as a single peak
by HPLC (retention time 15.2 mm, Fig.
1C). From its characteristic UV spec-
trum, we could identify perazine in the
patient’s urine at low concentration
(<0.1% peak area) and four of its metab-
olites at high concentrations (>99.9%
peak area of these five peaks; Fig. LA,B).
Additionally, increasing concentra-
tions of perazine as pure compound
(1000-1 000 000 gfL dissolved in drug-
free urine) were determined with Abbott
ADx (cutoff =100 g/L) and Syva
Emit-st (cutoff =300 p.gfL) assays and
resulted in the following:
Amphetamines
ADx, p.gJL 1045 pos.
Emit-st Neg.
Opiates
ADx, g/L 161 pos.
Emit-st Neg.

1726 Letters
For amphetamines, positive results
could be achieved only with the Abbott
ADx assay at a concentration of I g/L
perazine but with a >2.5-fold lower sig-
nal than found in urine of the patient.
We assume that the majority of the ob-
served cross-reactivity might result from
the urinary excreted perazine metabo-
lites, being much more cross-reactive
than that of perazine. Two facts support
this assumption: (a) perazine was almost
exclusively present in its metabolized
form and (b) a complete absorption of 1 g
of perazine by the patient would result in
urinary concentrations of perazine and
its metabolites much below 1 g/L. The
Syva Emit-st assay system showed no
cross-reactivity with perazine or its me-
tabolites.
For opiates, similar high signals were
found in urine samples of the patient and
a 1 g/L perazine solution by the Abbott
ADx. The above-mentioned assumption
of an excretion of perazine metabolites at
urinary concentrations <1 g/L points to
a cross-reactivity of this FPIA with pera-
zine metabolites. In contrast, the Emit-st
assay exhibited no cross-reactivity for
either perazine as pure compound or for
perazine metabolites.
Our findings demonstrate that the
lack of cross-reactivity of pure com-
pounds at high concentrations (in this
case 100 000 g/L) tested with immuno-
assays cannot exclude false-positive re-
sults due to their metabolites. Adults
receive a daily therapeutic dosage of
perazine up to 150 mg per day. Because
the ingestion of 1 g of perazine led to a
>10-fold higher signal with the ADx
amphetamine immunoassay as the cutoff
value, apositive result due to atherapeu-
tic application of perazine seems to be
possible. This might lead to false-positive
results with drugs-of-abuse screening
and misinterpretation of preliminary
positive results. Further studies are
needed to test if such cross-reactivities
occur with other neuroleptic drugs.
References
1. von Meyer L, Biro G, Degel F, Gruhl H, Hailbach
J, van Heijst ANP, KOppel C. Immunchemische
Verfahren. In: Gibitz Hi, Sch#{252}tzH, eds. Einfache
toxikologische Laboratoriumsuntersuchungen
bei akuten Vergiftungen. Weinheim: VCH Ver-
lags.. 1995:232-79.
2. D’Nicuola J, Jones R, Levine B. Smith ML. Eval-
uation of six commercial amphetamine and
methamphetamine immunoassays for cross-re-
activity to phenylpropanolamine and ephedrine
in urine. i Anal Toxicol 1992:16:211-3.
3. Ramos JM, Fitzgerald RL. PokIls A. MDMA and
MDA cross-reactivity observed with Abbott TDx
amphetamine/methamphetamine reagents
[Tech Brief]. Clin Chem 1988:34:991.
4. Cone Ei, Dickerson 5, Paul BD, Mitchell JM.
Forensic drug testing for opiates. IV. Analytical
sensivity, specificity and accuracy of commercial
urine opiate immunoassays. J Anal Toxicol
1992:16:72-8.
Michael Schmolke*
Jurgen Hailbach
Walter G. Guder
Inst. fijp. Kim. Chem.
St#{224}dtischesKrankenhaus Bogenhausen
Englschalkingerstr. 77
81925 Munchen, Germany
*Author for correspondence.
False-Negative Results wIth Emit#{174}
AmphetamIne ConfIrmation Kit
To the Editor:
Testing urine specimens for drugs of
abuse usually involves an immunoassay as
ascreening procedure. False-positive re-
sults due to sympathomimetic amines
occur with some of the amphetamine!
metamphetamine kits, most frequently
with Emit d.a.u.#{174},but also with Emit Il#{174}
immunoassays from Syva Co. (San Jose,
CA) [1]. Rescreening of amphetamine!
metamphetamine-positive samples after
periodate oxidation (Syva confirmation
kit) has been designed to eliminate this
problem. The confirmation kit may,
however, be of limited value. This letter
presents 10 cases that were positive for
amphetamine by gas chromatography-
mass spectrometry (GC-MS), although
they were negative according to the Syva
confirmation analysis.
Our laboratory performs forensic and
drugs-of-abuse analyses in biological
samples. Annually, -20 000 urine speci-
mens are received from prison inmates
and analyzed for drugs of abuse. Amphet-
amine/metamphetammne screening was
performed with polyclonal Emit d.a.u.
amphetamine/metamphetamine with a
300 jg/L cutoff. The analyses were run
on a Monarch 2000 (Instrumentation
Lab., Lexington, MA) at 30 #{176}C.Method
validation experiments have been carried
out to evaluate the normal variability in
the assay. Using a calibrator at the cutoff
value, accuracy was 97%, and precision
in terms of CV was 7.3%. Until recently,
all positive results were subjected to the
Syva confirmation reaction. Briefly,
NaOH and Na104 were added to the
specimens, which were then reanalyzed
for amphetamines. All procedures were
carried out according to Syva’s instruc-
tions. Different batches of confirmation
kits were used. Positive screening results
were analyzed for amphetamine/metam-
phetamine by GC-MS as described pre-
viously [2], with minor modifications.
The cutoff value for GC-MS was 135
g/L (1 imol/L).
In one case, we received two urine
specimens taken 20 mm apart from the
same prisoner. Both samples were
screened for drugs of abuse, and they
screened positive for amphetamines. The
first sample remained positive after the
confirmation reaction, while the second
became negative. Both samples were an-
alyzed by GC-MS, and amphetamine was
detected in both specimens, in concen-
trations of 2300 .tg/L and 640 j.tg/L,
respectively. The notable difference in
the urine amphetamine concentrations
between the two samples could at least
partly be related to a dilution effect, as
the two specimens had somewhat differ-
ent concentrations of creatinine (22.4
and 14.5 mmol/L), and a pH-effect, as
the pH was higher in the second speci-
men (6.0 and 7.0), leading to reduced
urinary excretion of amphetamines.
During the following 3 months, all spec-
imens that screened positive for amphet-
amines were subjected to GC-MS analysis
as well as to the confirmation reaction. Of
-5000 specimens screened, 67 were posi-
tive for amphetamine!metamphetamine.
After the positive samples were assayed by
the confirmation kit, 50 remained positive.
When analyzed on GC-MS, amphetamine
was detected above the cutoff value in 48 of
the 50 specimens confirmed positive, but
also in 9 of the 17 negative specimens
(fable 1). The concentrations of amphet-
amine were between 230 and 490 g/L in
these 9 specimens, and >300 ,.tg/L (the
corresponding Emit cutoft) in 4 specimens.
Altogether, we found false-negative results
in 13% (9 of 67) of the confirmation assays.
Our institute is in charge of a national
control program for laboratories per-
forming urine analyses for drugs of
abuse. Supplemented urine specimens
are sent to -40 other laboratories that
are using different types of immunolog-
ical screening techniques. We added D,L-
amphetamine to two different specimens
at concentrations of 400 and 440 ,.tg/L,