Baden LR, Horowitz G, Jacoby H, Eliopoulos GM. Quinolones and False-Positive Urine Screening for Opiates by Immunoassay Technology. JAMA. 2001;286(24):3115-3119. doi:10.1001/jama.286.24.3115
Author Affiliations: Departments of Medicine (Drs Baden, Jacoby, and Eliopoulos) and Pathology (Dr Horowitz), Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Mass.
Toward Optimal Laboratory Use Section Editor:
David H. Mark, MD, MPH, Contributing Editor.
Context Millions of assays are performed each year to monitor for substance
abuse in various settings. When common medications cross-react with frequently
used testing assays, false-positive results can lead to invalid conclusions.
Objective To evaluate cross-reactivity of quinolone antimicrobials in common opiate
screening assays and to assess the in vivo implications of this phenomenon.
Design, Setting, and Participants The reactivity of 13 quinolones (levofloxacin, ofloxacin, pefloxacin,
enoxacin, moxifloxacin, gatifloxacin, trovafloxacin, sparfloxacin, lomefloxacin,
ciprofloxacin, clinafloxacin, norfloxacin, and nalidixic acid) was tested
in 5 commercial opiate screening assays from September 1998 to March 1999.
In 6 healthy volunteers, we confirmed the cross-reactivity of levofloxacin
or ofloxacin with these opiate screening assays.
Main Outcome Measure Opiate assay activity (threshold for positive result, 300 ng/mL of morphine).
Results Nine of the quinolones caused assay results above the threshold for
a positive result in at least 1 of the assays. Four of the assay systems caused
false-positive results for at least 1 quinolone. Eleven of the 13 compounds
caused some opiate activity by at least 1 assay system. At least 1 compound
caused opiate assay activity in all 5 assay systems.
and pefloxacin were most likely to lead to a false-positive opiate result.
Positive results were obtained in urine from all 6 volunteers.
Conclusion Greater attention to the cross-reactivity of quinolones with immunoassays
for opiates is needed to minimize the potential for invalid test interpretation.
In response to public concerns regarding use of illicit drugs, random
drug testing has become a common practice for employees in the workplace,
for individuals incarcerated or under suspicion by the criminal justice system,
and in other circumstances.1 This practice
has been sanctioned by legislation and affirmed by court decisions, including
the US Supreme Court.2- 6
In general, samples are usually subjected to screening by relatively inexpensive,
rapid, and reliable immunoassays, with samples testing positive requiring
confirmation by an alternative method. When such strict protocols are followed,
false-positive screening test results do not cause problems.
It is now widely appreciated that immunoassays are extremely reliable
and have relatively few false-positive results. As a result, at least for
some applications, it has been advocated that confirmation is not necessary.7,8 Thus, as testing expands beyond the
strictly controlled legal arenas, there is a possibility that positive results
will be acted on in the absence of confirmatory testing. Such is the case
in most hospital laboratories, and thus it becomes important for clinicians
to know that false-positive test results do occur and to request confirmatory
testing by alternative methods (as would be required in legal settings) when
there is a question about the validity of screening results. For this reason,
information concerning therapeutic use of possible cross-reacting prescription
medications should assist authorities in testing and ultimately help the individual
who is being tested.
Meatherall and Dai9 reported in 1997
that ofloxacin could result in a false-positive test result for opiates by
the EMIT [enzyme multiplied immunoassay technique] II assay (Syva, San Jose,
Calif). However, this fact does not appear to be widely known among clinicians.
We encountered a patient in whom a false-positive urine screening test for
opiates during therapy with levofloxacin nearly resulted in his ejection from
a drug-treatment center. As a result, we examined systematically the propensity
of various quinolones to cause false-positive reactions for opiates using
5 major commercially available screening assays.
Thirteen quinolones (levofloxacin, ofloxacin, pefloxacin, enoxacin,
moxifloxacin, gatifloxacin, trovafloxacin mesylate, sparfloxacin, lomefloxacin,
ciprofloxacin hydrochloride, clinafloxacin, norfloxacin, and nalidixic acid)
were either provided by the manufacturer or purchased from a biologic company.
All antibiotics were made soluble as per standard techniques10
to a concentration of 5000 µg/mL to ensure we exceeded possible in vivo
urinary levels and dilutions to concentrations of approximately 1700 µg/mL
and 600 µg/mL were made.
These samples were then analyzed by 5 different commercial immunoassays:
(1) EMIT II reagents, which were run on Hitachi 717 analyzer (Roche Diagnostics,
Indianapolis, Ind), (2) AxSYM fluorescence polarization immunoassay (Abbott
Laboratories, Abbott Park, Ill), (3) CEDIA [cloned enzyme donor immunoassay]
reagents (Microgenics, Concord, Calif), which were run on Hitachi 912 analyzer
(Roche Diagnostics), (4) Roche Abuscreen OnLine reagents (Roche Diagnostics),
which were run on the Dimension XL analyzer (Dade Behring, Newark, Del), and
(5) Beckman opiate reagents, which were run on the Synchron CX analyzer (Beckman
Instruments, Brea, Calif).
For those samples that tested positive in a given assay at the lowest
concentration (600 µg/mL), further dilutions were performed to determine
the approximate lowest concentration that would test positive. Controls with
morphine concentrations of 0, 225, 300, and 375 ng/mL were run simultaneously
on each assay. All assays were run in accordance with the manufacturers' recommendations.
Specifically, the threshold for a positive result was set at 300 ng/mL of
morphine, and samples with error messages not amenable to trouble-shooting
protocols were not run on dilution (as recommended by manufacturers). A more
detailed investigation of the dose-response curves for ofloxacin and levofloxacin
was performed with the EMIT II 717 system, spanning the range of concentrations
from 5000 µg/mL to 0.005 µg/mL.
With approval from the Committee on Human Studies and informed consent
from the participants, in 1999, 6 healthy volunteers were given a single oral
dose of antibiotic (3 received 500 mg of levofloxacin and 3 received 400 mg
of ofloxacin) and urine samples were collected at approximately 6-hour intervals
for the following 48 hours. These samples were analyzed for opiates by the
EMIT II 717 system. Select samples were run on the other 4 assay systems.
The results of the screening assays for the 13 quinolones tested in
vitro are shown in Table 1. A
concentration of 300 ng/mL of morphine was set as the positive threshold,
as previously suggested by the Department of Health and Human Services and
widely used by clinical laboratories.11- 14
By convention, a value of 250 ng/mL would be considered a negative value for
opiates, although substantial opiate activity above baseline (0 ng/mL) is
present. The following quinolones cross-reacted to cause a positive test result
for opiates at concentrations assayed: levofloxacin and ofloxacin (using Abbott
AxSYM, CEDIA, EMIT II, and Roche OnLine assays), pefloxacin (using CEDIA,
EMIT II, and Roche OnLine assays), enoxacin (using CEDIA and EMIT II assays),
gatifloxacin (using EMIT II assay), and lomefloxacin, moxifloxacin, ciprofloxacin,
and norfloxacin (using Roche OnLine assay). Sparfloxacin, clinafloxacin, trovafloxacin,
and nalidixic acid did not cross-react to cause a positive test result with
any of the assays. To properly interpret the clinical relevance of these observations,
the urinary concentrations of these quinolones15- 25
must be considered. Based on these in vitro data and given the anticipated
urinary concentrations, pharmacodynamics, and dosing interval, the quinolones
most likely to cause a false-positive urinary test result for opiates are
levofloxacin and ofloxacin (using CEDIA, EMIT II, and Roche OnLine assays)
and pefloxacin (using CEDIA).
These screening assays for opiates are qualitative (threshold) tests
and should not be used quantitatively. However, it is important to consider
low-concentration opiate cross-reactivity (below the threshold) as this may,
in certain settings, facilitate reaching assay threshold—thus, a false-positive
test result. Detailed analysis of 2-fold serial dilutions of levofloxacin
(Figure 1) and ofloxacin (using
EMIT II assay) demonstrates dose-responsive assay activity between concentrations
of 5 µg/mL to 1250 µg/mL, with the assay threshold being achieved
at approximately 110 µg/mL. The following quinolones have some opiate
activity but below the assay threshold; thus, they have the potential to act
in an additive manner to trigger a positive opiate assay: levofloxacin and
ofloxacin (Synchron assay); pefloxacin (Abbott AxSYM assay); enoxacin (Roche
OnLine and Abbott AxSYM assays); gatifloxacin (Abbott AxSYM, CEDIA, and Roche
OnLine assays); sparfloxacin (CEDIA, EMIT II, and Roche OnLine assays); and
lomefloxacin, moxifloxacin, clinafloxacin, ciprofloxacin, and norfloxacin
(CEDIA and EMIT II assays) (Table 1).
At the concentration levels tested, trovafloxacin and nalidixic acid demonstrated
no detectable opiate cross-reactivity by any of the assays.
A single dose of 500 mg of levofloxacin caused a false-positive test
result using the EMIT II assay within 2 hours for as long as 22 hours in all
3 healthy volunteers (Figure 2).
Ofloxacin (single dose of 400 mg) revealed a similar pattern. Detectable opiate
activity in the urine was seen for more than 30 hours with both antimicrobials.
Selected urine samples from these subjects were run on the other 4 assays
and were found to cause assay positivity above the 375 ng/mL level by the
CEDIA assay and positivity persisted above the 300 ng/mL concentration for
over 24 hours (Table 2). The Roche
OnLine assay found positivity above 375 ng/mL, with persistent (slight) assay
activity at 24 hours. The Synchron assay had activity above the 225 ng/mL
level at 7 hours after dosing, with persistent activity detected at 24 hours.
The Abbott AxSYM assay demonstrated some opiate reactivity; however, this
was below the 225 ng/mL concentration and persisted for over 24 hours.
Few compounds have been identified that cross-react with the common
opiate screening assays26 and include rifampin
(described for the kinetic interaction of microparticles in solution method),
ofloxacin (described for the EMIT method),9
and poppy seeds (described not for cross-reactivity, but detection of minute
amounts of opiates).27- 34
We have shown that several quinolones have the potential to yield false-positive
test results by a number of commonly used opiate screening assay systems currently
used in the United States. It is important to note that several of the quinolones
are metabolized in vivo and the metabolites (eg, norfloxacin as a metabolite
of pefloxacin) may be excreted into the urine.15,35
We did not assess the potential contributions of other metabolites in this
study. Because of the enormous ramifications for an unrecognized false-positive
test result, the results of our study strongly support the use of confirmatory
testing when a person receives a positive opiate test result in the setting
of recent quinolone use.
What all of these assays have in common is that an antibody directed
against opiate epitopes is exposed to labeled drug in the reagent system and
free drug in the sample. With higher concentrations of drug in the sample,
less of the labeled drug in the reagent system binds to the antibody. Depending
on the specific assay, the signal detected (turbidity, enzyme activity, fluorescence
polarization, etc) is changed by the concentration of free drug in the sample
in a predictable, although typically not proportional manner. When the signal
exceeds that of an arbitrary standard (typically, for opiates, 300 ng/mL of
morphine), the assay is considered positive (ie, contains an opiate with a
concentration whose reactivity exceeds that of 300 ng/mL of morphine).
Until recently, the recommended cutoff for opiate assay positivity (by
the Department of Health and Human Services) was 300 ng/mL of morphine.11- 14 To
minimize the unnecessary gas chromotography/mass spectrometry effort and expense
due to poppy-seed food products, the opiate screening threshold was raised
from 300 ng/mL to 2000 ng/mL of morphine in December 1998. This cutoff, however,
has been reported to have only a 70% sensitivity for detecting opiates.11 In part, due to this sensitivity concern and for
purposes of clinical rather than forensic testing, most clinical laboratories
continue to use the 300-ng/mL threshold for assay positivity. Given the clinical
implications of quinolone cross-reactivity with opiate testing, we report
the 300 ng/mL as our cutoff for assay positivity.
Why certain quinolones react with some opiate screening assays is unclear,
as there is no obvious structural similarity between morphine and this class
of drugs nor is there an obvious structural relationship between the quinolones
that cross-react. These data are of particular importance given the widespread
use of these agents, such as for the treatment of community-acquired pneumonia,
nosocomial-acquired pneumonia, sexually transmitted diseases, multidrug-resistant
tuberculosis, and the prophylaxis for possible anthrax exposure.15,36- 45
As many of these infections occur in patients who might be susceptible to
substance abuse, the potential for misinterpretation of testing is self-evident.
In addition, the care of patients in clinical situations may be misguided
by a positive urine test for opiates, such as inappropriately halting the
evaluation of a change in mental status.
It is important to realize that the screening assays are designed to
be positive when the urine concentration of morphine is 300 ng/mL or greater.
These are qualitative not quantitative tests. Different immunoreactive compounds
can have an additive effect on reaching the threshold for a given assay. For
example, if a given quinolone would lead to 225 ng/mL and a poppy seed muffin
to 100 ng/mL of immunoreactivity, consumption of either product alone would
not induce a positive urine opiate test result. However, if both were consumed
simultaneously, the test threshold might be achieved. Thus, a compound that
induces signal activity by a given immunoassay technology, albeit below assay
threshold, could additively contribute to a positive urine screening test
result when other cofactors are present.
When a screening test for drugs frequently abused returns positive,
it is essential to obtain appropriate confirmatory testing, such as gas chromatography,
mass spectrometry, or high-performance liquid chromatography.32,46,47
Quinolones are not misinterpreted as opiates by these methods. Confirmatory
testing should not be done by another immunoassay technique for the reasons
demonstrated in this analysis. Confirmatory testing does not always resolve
the issue of substance abuse as consumption of poppy seeds or medicinally
prescribed opiates have been reported as innocent explanations for positive
opiate screening test results.29,31,46
The major limitations to obtaining confirmatory testing are time and money.
In circumstances in which resources are limited, the significant cost of additional
testing will remain a formidable obstacle to accurate test interpretation.
These data demonstrate the need for vigilance in identifying unintended consequences
of new therapies.