False-Positive Amphetamines in Urine Drug Screens: A 6-Year Review

Abstract

Immunoassays are routinely used to provide rapid urine drug screening results in the clinical setting. These screening tests are prone to false-positive results and ideally require confirmation by mass spectrometry. In this study, we have examined a large number of urine specimens where drugs other than amphetamines may have caused a false-positive amphetamine immunoassay screening result. Urine drug screens (12,250) in a clinical laboratory that used the CEDIA amphetamine/ecstasy method were reviewed for false-positive results over a 6-year period (2015–2020). An additional 3,486 referred samples, for which confirmatory--mass spectrometry was requested, were also reviewed. About 86 in-house samples and 175 referral samples that were CEDIA false-positive screens were further analyzed by an LC–QTOF general unknown screen. Potential cross-reacting drugs were identified, and their molecular similarities to the CEDIA targets were determined. Commercial standards were also analyzed for cross-reactivity in the amphetamine/ecstasy CEDIA screen. Positive amphetamine results in 3.9% of in-house samples and 9.9% of referred tests for confirmatory analysis were false positive for amphetamines. Of these false-positive specimens, on average, 6.8 drugs were detected by the LC–QTOF screen. Several drugs were identified as possible cross-reacting drugs to the CEDIA amphetamine/ecstasy assay. Maximum common substructure scores for 70 potential cross-reacting compounds were calculated. This was not helpful in identifying cross-reacting drugs. False-positive amphetamine screens make up to 3.9–9.9% of positive amphetamine screens in the clinical laboratory. Knowledge of cross-reacting drugs may be helpful when mass spectrometry testing is unavailable.

Introduction

Urine drug screen (UDS) immunoassays are typically performed as the first line of testing for drugs of abuse. They are amenable to automated analyzers and provide rapid results. However, due to the possibility of interference seen in immunoassays, positive results require confirmation by mass spectrometry. In the clinical setting, this sometimes does not occur, but in laboratories that perform confirmatory testing, testing is often batched leading to delays between screening results and confirmatory results. Such delays may lead to inappropriate clinical decisions.

Several case reports on false-positive immunoassays have been published in the literature, with false-positive results being described for the commonly used immunoassay formats, including: enzyme-multiplied immunoassay technique (EMIT^®), cloned-enzyme donor immunoassay (CEDIA^®), kinetic interaction of microparticles in solution (KIMS^®) and glucose-6-phosphate dehydrogenase enzyme immunoassay (DRI^®) (1). In the authors’ laboratory, among the drug screening panels used (benzodiazepines, barbiturates, cannabinoids, amphetamine-type substances (ATSs), opiates and cocaine metabolites), the amphetamine/ecstasy immunoassay screen is the most prone to false-positive results (unpublished observations). Several published case reports have identified differing drugs that are associated with false-positive amphetamine screens, including β-blockers (2–4), antipsychotics (5), antidepressants (6, 7) antiarrhythmics (8, 9) and others; however, the causes of many false-positive immunoassay screens remain unknown.

In this study, we ascertain the prevalence of false-positive CEDIA amphetamine/ecstasy assays seen at a quaternary hospital and attempt to identify drugs and common medicines that contribute to these false-positive readings. A combination of a non-targeted liquid chromatography–quadrupole time-of-flight (LC–QTOF) screen, analyses of spiked drug-free urine (DFU), electronic medical chart review and 2D molecular similarity analysis were employed to identify potential cross-reacting drugs.

Materials and Methods

Sample selection

UDS specimens received from 1 January 2016 to 31 December 2020 at a quaternary hospital in Melbourne, Australia, were included. All urine specimens were initially tested by the CEDIA amphetamine/ecstasy assay (product no. 100,104; Thermo Scientific, Scoresby, Australia) as well as having a urine creatinine level determined using a standard method. A positive cut-off level of 300 µg/L, consistent with the AS/NZS 4308:2008 standard, was used (10). All specimens were subsequently analyzed by either a GC–MS (2015–2016) or an LC–QTOF (2017–2020) method for the presence of ATSs, including methamphetamine, amphetamine, 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine (MDMA), pseudoephedrine, phentermine and several substituted amphetamines, some piperazines and other novel psychoactive substances (11). Samples with a confirmed ATS, including those below the screening cut-off, were excluded.

Ninety-seven in-house UDS specimens with a positive immunoassay result and no ATS detected by confirmatory testing were investigated for drugs that might cause false positives. Eighty-six specimens were analyzed using the LC–QTOF method by a general unknown screen. A further set of 346 specimens, from the same study period, that were presumptively positive for amphetamine screens at the referring institution but tested negative in the confirmation assay for an ATS were also included. Of these, 176 samples were stored and available for LC–QTOF testing.

Ethics approval

All samples had a clinical request for a urine drug screen. Ethical approval was obtained from Alfred Health (HREC 17/35).

LC–QTOF urine drug screen

Chromatographic separation was achieved using an iClass Acquity UHPLC coupled to a Xevo G2-XS mass spectrometer (LC–QTOF; Waters Corporation, Milford, MA, USA). The LC–QTOF analyses were performed by two separate methods, one operated in positive electrospray ionization (ESI) mode and the other operated in negative ESI mode. The detailed LC–QTOF method with validation is described elsewhere (11). Briefly, a 15-min gradient elution (0.4 mL/min) was used for ESI+ mode. Mobile phase A1 consisted of 5 mmol/L ammonium formate at pH 3.0, and Mobile phase B1 consisted of acetonitrile containing 0.1% formic acid. A 7.5-min gradient elution (0.4 mL/min) was used for ESI− mode, Mobile phase A2 consisted of purified water containing 0.001% formic acid, and Mobile phase B2 consisted of acetonitrile containing 0.001% formic acid. Urine specimens were not hydrolyzed and analyzed with a dilution factor of five in 5 mmol/L ammonium formate at pH 3.0. Five microliters was injected onto a Waters^® Acquity HSS C18 column (150 × 2.1 mm, 1.8 µm) operating at 50°C.

Mass data were acquired in MS^E mode, with a mass range of m/z 50–1200, source temperature 120°C, desolvation temperature 500°C, desolvation gas flow at 800 L/h and a cone gas flow of 50 L/h. The QTOF acquired data in two functions: Function 1, a collision energy (CE) of 6 eV, and Function 2, a 10–40 eV CE ramp, yielding high-energy fragments.

A mass correction was performed with leucine enkephalin at 200 ng/L concentration, measured at 20-s intervals throughout each injection at a flow rate of 10 µL/min. The ions monitored were m/z 556.2766 and 554.2620 Daltons in ESI+ and ESI− modes, respectively.

All data were collected and analyzed with Waters UNIFI software, version 1.8. To meet identification criteria, drugs had to meet the following criteria: an accurate mass difference of less than 5 ppm of its theoretical charged mass, elution within a ±0.35-min retention time (RT) window, and have a minimum of one high energy (h/e) fragment, eluting at the same RT. The library contained approximately 2,000 entries including RT and mass spectral data. All chromatograms were reviewed for validity. Drugs identified were considered for spiking experiments and 2D similarity testing.

Preparation of standards to investigate cross-reactivity in the CEDIA assay

Aripiprazole, ranitidine, chlorpromazine, mexiletine, ranitidine, sildenafil and varenicline were purchased from Sigma Aldrich (Sydney, Australia). Atenolol, bisoprolol, lamivudine, meloxicam, metformin, methylphenidate, metoprolol, metaraminol, metformin, modafinil, propranolol, sertraline, tapentadol, tramadol, venlafaxine, verapamil and warfarin were obtained from the Alfred Hospital Pharmacy. Stock concentrations in methanol were prepared for atenolol (25 g/L), bisoprolol (5 g/L), bupropion (25 g/L), chlorpromazine (25 g/L), metformin (100 g/L), metoprolol (20 g/L), mexiletine (25 g/L), mirtazapine (15 g/L), ranitidine (25 g/L), tapentadol (25 g/L), venlafaxine (25 g/L), verapamil (25 g/L) and warfarin (25 g/L). Stock concentrations in water were prepared for lamivudine (30 g/L), meloxicam (7.5 g/L), propranolol (25 g/L), sertraline (3 g/L), sildenafil (3.5 g/L), tramadol (25 g/L) and varenicline (25 g/L). Metaraminol was supplied as a 10 g/L solution. To investigate cross-reactivity, individual stock standards were diluted in DFU to make a range of concentrations and analyzed by the CEDIA amphetamine/ecstasy assay. The percentage of methanol in a diluted standard did not exceed 10%.

Use of the Tanimoto coefficient to investigate compound similarity

Maximum common substructure (MCS) is a metric for investigating chemical similarity and predicting activities (12). MCS similarities with the Tanimoto coefficient were calculated (www.chemminetools.ucr.edu) for the following parent drugs: aripiprazole, atenolol, atomoxetine, bisoprolol, bupropion, bupivacaine, carvedilol, ceftaroline, cetirizine, chloroquine, chlorpheniramine, chlorpromazine, clomipramine, diclofenac, dimethylamylamine, doxylamine, esmolol, fenofibrate, gliclazide, imatinib, labetalol, lamivudine, linezolid, meclofenamic acid, meloxicam, metaraminol, metformin, methylphenidate, metoprolol, mexiletine, modafinil, moxifloxacin, oxymetazoline, ranitidine, sertraline, sildenafil, sitagliptin, sotalol, tapentadol, tetracaine, tramadol, trazodone, varenicline, venlafaxine, verapamil, warfarin and several metabolites against the target compounds amphetamine, methamphetamine and MDMA. Other ATS included in the predication were MDA, phentermine and pseudoephedrine. An MCS score of 1.0 indicates a perfect match, and a score of ≥0.50 was further investigated (13).

Results

CEDIA immunoassay screen results

In the 6-year study period from January 2015 to December 2020, we analyzed 12,250 in-house urine drug screens. Of these specimens, 2,468 (20.1%) tested positive with the CEDIA amphetamine/ecstasy screen at a level of ≥300 µg/L, consistent with the AS/NZS 4308:2008 standard. Of these positive screens, 3.9% were found to be false positive, with no ATSs detected by a confirmatory mass spectrometry method (LC–QTOF in 86 and GC–MS in 11). During the same period, 3,486 referred samples were received for ATS confirmation after initially testing positive for ATSs in a local laboratory. In 346 (9.9%) of those samples, an ATS could not be confirmed by either GC–MS (2015–2016) or LC–QTOF (2017–2020) and were considered false positives (see Figure 1).

The patient demographics and mean amphetamine/ecstasy CEDIA results are shown in Table I. For in-house samples, we had access to the electronic medical record (EMR) to review the therapeutic drugs prescribed. Patients were on average younger in the referred samples (mean 38 years) than in the in-house requests (mean 46 years). The gender distribution was similar (about 60% males). 83.2% of the in-house specimens had positive CEDIA readings between 300 and 500 µg/L, and 16.8% had readings >500 µg/L. In the referral tests, 59.5% (173 cases) had CEDIA results between 300 and 500 µg/L and 31.3% (91 cases) had CEDIA results below the positive cut-off of 300 µg/L. The remaining 9.2% (27 cases) had CEDIA results greater than 500 µg/L. The CEDIA screen result distributions are shown in Figure 2.

Non-targeted LC–QTOF screen

A total of 584 drugs were found in the 86 in-house false-positive urine specimens analyzed by the LC–QTOF screen, with an average finding of 6.8 drugs per sample (SD = 3.6). Drugs had their RTs confirmed locally. The EMR was reviewed for these samples to determine additional prescribed medications that may have contributed to a positive CEDIA amphetamine/ecstasy assay result and were not detected by the LC–QTOF screen. The information was aggregated, and after excluding drugs not reasonably likely to cross-react in the amphetamine screen (i.e., benzodiazepines, opiates, paracetamol, etc.), a list of potential drug candidates was created.

Putative false-positive drugs and the frequency of their detection are listed in Table II. The most frequently encountered drugs associated with false-positive screen tests in the in-house samples were sildenafil (21 cases), ranitidine (12 cases), bisoprolol (7 cases), metoprolol (7 cases), sertraline (5 cases), bupropion (4 cases), mexiletine (4 cases) and tramadol (4 cases). Additionally, single cases of each of the following drugs were identified: atenolol, imatinib, procainamide, tapentadol and a Roundup^® intoxication. No cases with structurally similar drugs metaraminol or methylphenidate (14 detections, no false-positive screens) were identified. In the in-house samples, we could identify possible cross-reacting drugs in 51% of the cases based on findings from the LC–QTOF screen and EMR review.

In the 346 referral samples for ATS confirmation, 291 cases (84.1%) were analyzed in-house by the CEDIA amphetamine/ecstasy screen. This was not a routine component of testing but was regularly performed when an ATS could not be identified. Of the 291 cases tested, 91 cases were negative to the CEDIA screen (31.3%). 176 screen-positive samples (of 200) were available for LC–QTOF analysis to search for drugs that may cause false amphetamine screen results. In these samples, the following drugs were detected: sertraline (20 cases), ranitidine (16 cases), sildenafil (11 cases), metoprolol (9 cases), bupropion (4 cases) and tramadol (4 cases). Additionally, two cases of dimethylamylamine, two cases of tapentadol and one case of atenolol were identified.

Molecular similarity scores

Molecular similarity scores for suspected interfering drugs were compared to the target drugs in the CEDIA amphetamine/ecstasy test. Drugs with significant MCS similarity scores (≥0.50) for amphetamine included bupropion (0.63), metaraminol (0.83), methyldopa (0.67), methylphenidate (0.59), tapentadol (0.53) and varenicline (0.53). Drugs with MCS similarity scores >0.50 to MDMA included atenolol (0.50), bupropion (0.58), metaraminol (0.73), methyldopa (0.76), methylphenidate (0.55), tapentadol (0.50), varenicline (0.50) and venlafaxine (0.55). MCS scores for several drugs of interest and selected metabolites are summarized in Table III.

Cross-reactivity of spiked standards

Aripiprazole, atenolol, bisoprolol, bupropion, chlorpromazine, lamivudine, meloxicam, metaraminol, metformin, methylphenidate, metoprolol, mexiletine, modafinil, ranitidine, sertraline, sildenafil, tapentadol, tramadol, varenicline, venlafaxine, verapamil and warfarin were diluted in DFU and analyzed by the CEDIA amphetamine/ecstasy screening test. Cross-reactivity for each of the parent drugs across a range of concentrations is shown in Figure 3. The following drugs produced a positive result at the 300 µg/L cut-off: atenolol 750,000 µg/L; bisoprolol 400,000 µg/L; bupropion 8,500 µg/L; methylphenidate 125,000 µg/L; metoprolol 300,000 µg/L; metaraminol 500,000 µg/L; mexiletine 25,000 µg/L; ranitidine 225,000 µg/L; sertraline 200,000 µg/L; sildenafil 25,000 µg/L; tapentadol 400,000 µg/L and verapamil 1,250,000 µg/L.

Chlorpromazine, lamivudine, meloxicam, metformin, mirtazapine, propranolol, tramadol, varenicline and venlafaxine tested negative at 2,500,000 µg/L.

Discussion

Immunoassay screens provide rapid results to aid in making clinical decisions. Ideally, positive results require confirmation by mass spectrometry and the identification of the specific drug present. This will ensure the most appropriate clinical decisions and minimize adverse patient care. However, in most clinical scenarios, decisions will be made on immunoassay results as mass spectrometry typically requires transfer to a referral laboratory before testing. Results may take days to weeks to return to the originating laboratory and its clinicians.

Of the urine drug screens used in our laboratory, the CEDIA amphetamine/ecstasy screen (product no. 100,104) is the most susceptible to false-positive results, with 3.9% of all positive screens being a false-positive result. 9.9% of the referred samples for ATS confirmation during the study period were also found to give false-positive screens. This increased incidence is likely related to the differing immunoassay screens used at other institutions and possibly to sample selection bias from the referring institution.

After reviewing the non-targeted LC–QTOF data and available EMRs for medications and excluding drugs unlikely to cross-react, several potential drugs were considered as candidates for causing false-positive screen results.

To investigate the cross-reactivity potential of these drugs, two approaches were taken. The 2D structures of these potential interfering drugs were compared to the immunoassay target molecules amphetamine, methamphetamine and MDMA. Second, drug standards were purchased and diluted in DFU to determine the level, if any, of cross-reactivity in the screening assay.

MCS scores provide a quick estimation of structural similarities between compounds. High similarity scores (MCS ≥0.50) may infer immunoreactivity in the immunoassay screens, an approach Petrie et al. described when investigating bath salt cross-reactivity in amphetamine screens (29). Petrie et al. found that designer drugs with a score of >0.55 similar to amphetamines were likely to cross-react in EMIT, CEDIA and AxSYM immunoassays (29). Compounds with similarity scores below this level had variable cross-reactivity to the different assays studied (29). Several publications have looked at the cross-reactivity of novel psychoactive substances in different ATS immunoassays and found performance differences (30–32).

MCS scores comparing the 2D similarities of the major defining structural groups do not necessarily represent the steric interaction of the target compound with the antibody or components of the assay. This analysis also does not consider any possible similarity with any of the metabolites of these drugs. Reviewing the structure of the sympathomimetic amine, metaraminol, its similarities to methamphetamine are apparent and cross-reactivity in the CEDIA amphetamine/ecstasy is possible (MCS scores 0.73–0.77). However, MCS scores for fenofibrate (MDMA similarity 0.35), ranitidine (amphetamine similarity 0.19) and moxifloxacin (MDMA similarity 0.23) are considerably lower and do not appear to share molecular similarities. However, publications have implicated these drugs in causing false-positive screen results (19, 23–25, 27). Similarly, sildenafil had a low MCS score (0.19–0.21) but was detected frequently in our patient cohort and cross-reacted in the CEDIA method. Conversely, varenicline had a high MCS score but did not cross-react on the CEDIA screen and was also mostly excreted unchanged in the urine. Alternatively, 3D molecular modeling may provide more useful similarity predictions than the 2D approach in this article. This was found acceptable when investigating bath salt cross-reactivity in CEDIA amphetamine/ecstasy assay but less helpful in the EMIT II Plus amphetamines assay or AxSYM II amphetamine/methamphetamine assay (29).

In this study, several drugs were found to be associated with false-positive CEDIA screen results in the in-house samples (Table II), including previously unreported drugs: sildenafil (21 cases), bisoprolol (7 cases), sertraline (5 cases) and tramadol (4 cases). A single case of each of atenolol and tapentadol was also identified. In addition, several cases of bupropion, metoprolol, ranitidine and mexiletine were also associated with false-positive amphetamine screens in this series of specimens.

Sildenafil was found in 21 cases with a false-positive amphetamine screen. Of these, 19 patients had heart failure and were being treated for pulmonary hypertension. They were prescribed either 25 or 50 mg sildenafil three times daily. These are substantial doses as compared to a single dose of 50 mg for erectile dysfunction. Two cases did not have a documented sildenafil prescription. In heart failure patients, urine drug screens were part of their pre-transplantation work-up, and without confirmatory testing, they may not have been listed for transplant.

Several patients were prescribed β-blockers. This study implicates several false-positive amphetamine screens to atenolol, bisoprolol and metoprolol, although this will need to be confirmed in subsequent studies. Along with published case reports for esmolol (4) and labetalol (2), it is reasonable to assume that other β-blockers could cross-react in amphetamine screening assays.

A review of the EMR revealed that one patient had been prescribed metaraminol, a sympathomimetic agent used for the short-term treatment for hypotension. Metaraminol was found to cross-react with the CEDIA screen at 500,000 µg/L. In this patient, metaraminol was not detected by the LC–QTOF screen, and urine levels this high would not be typical. Metaraminol is metabolized to α-methyl-noradrenaline (Corbadrine) and is also a metabolite of methyldopa. Recently, methyldopa was not found to cross-react in the Abbott amphetamines HEIA, although α-methyl-dopamine did cross-react (4). Methylphenidate and verapamil also cross-reacted in the CEDIA assay but at concentrations not routinely encountered. There were no false-positive CEDIA cases in the in-house samples due to these two medications.

Tramadol was another drug highlighted since four in-house samples were identified with large quantities of tramadol. However, tramadol diluted in DFU at 2,500,000 µg/L tested negative in the CEDIA screen, although immunoreactivity was noted below the positive cut-off. Combined with the large amounts of metabolites (e.g., desmethyl) that are also present, we postulate that the metabolite or accumulative effects of parent drugs and metabolites may be enough to trigger a positive result. Tramadol is predominantly excreted in the urine, with 30% of the dose remained unchanged and 60% excreted as metabolites (33).

Measuring the cross-reactivity of the parent drugs with the immunoassay only accounts for the contribution of the parent compound to the screen result. Metabolites, often structurally related to the parent compound, can also contribute to false-positive readings. The trazodone metabolite, meta-chlorophenylpiperazine (mCPP), has been shown to cause false-positive results in the Roche amphetamines II method (6). This is an important consideration as immunoassay manufacturers test their immunoassays against the parent drug and generally do not consider the impact of metabolites. The percentages of a dose excreted unchanged in urine for trazodone, bupropion and mexiletine are low at 0.5%, <1% and <10%, respectively (6, 34, 35).

Mexiletine has been found to cause false-positive amphetamine screen results in several different immunoassays (8, 9). In this cohort, mexiletine was detected in four cases, and mexiletine spiked into DFU at 25,000 µg/L tested positive at the 300 µg/L cut-off. Patients positive for mexiletine were also found to have large peaks, relative to mexiletine, which may correspond to hydroxylated mexiletine metabolites. Although not confirmed with an analytical standard, these structurally related metabolites are likely to cross-react based on their structural similarities.

A limitation of this work is that not all drugs may have been identified. This could be due to incomplete EMRs or drugs not detected by the analytical process. There will be some drugs not included in the LC–QTOF library and drugs that cannot be detected due to poor ionization with ESI, in-source fragmentation or other losses in the analytical process. Additionally, spiking experiments may result in immunoassay cross-reactivity at concentrations that might not be found routinely. For example, methylphenidate is mostly excreted as the metabolite, ritalinic acid, and in a large cohort (n = 9,674), median urine ritalinic acid concentrations were 11,452 µg/L with some larger concentrations found up to 100,000 µg/L (36). Methylphenidate is unlikely to be found at urine concentrations higher than this and is unlikely to trigger a positive CEDIA amphetamine/ecstasy screen. Lawson et al. investigated adherence to antihypertensive medication in a clinical setting with target verapamil concentrations <1,000 µg/L in their assay (37). Additionally, a urine verapamil concentration of 369 µg/L was described in a single fatal overdose (38). Verapamil concentrations in both publications were significantly less than spiking experiments performed in this article. It is also unclear if significant amounts of metaraminol accumulate in the urine (39).

Minimizing false-positive amphetamine screens can be achieved by the use of a higher cut-off concentration. In Australia, 300 µg/L is a common positive cut-off level, which is listed in the AS/NZS4308:2008 standard used for workplace drug testing. If a 500 µg/L positive cut-off were applied to the cases in this study, 83.2% of in-house specimens (79 cases out of 95) and 90.7% of referral specimens (264 cases out of 291) from this study would be reclassified as negative and would not require further testing. On the other hand, a lower cut-off would likely identify a number of specimens where amphetamines would be confirmed by LC–QTOF, although interferences by other substances would likely also increase.

Conclusion

In this series of immunoassay-positive amphetamine/ecstasy screens, we show that 3.9–9.9% of positive results are false positive. We demonstrate that several previously not identified drugs may cause a false-positive CEDIA amphetamine/ecstasy (product no. 100,104) screen and reinforce the importance of confirming presumptive screen results to ensure optimum patient care. Although assessing structural similarities can provide a quick estimate to potential cross-reactivity to the target immunoassay compound, this approach was not helpful in this study. Testing potential cross-reactivities with analytical standards and their metabolites should be undertaken whenever possible.

References