Immunoassays are frequently employed for urine screening

Reduced Interference by Phenothiazines in Amphetamine Drug of Abuse Immunoassays Stacy E. F. elanson, D, PhD; Elizabeth Lee-Lewandrowski, PhD; David A. Griggs, BS, BA; William H. Long, T; James G. Flood, PhD Context. Emergency department physicians frequently request urine drug screens, but many are unaware of their limitations, including the potential for false-positive results. Promethazine, a phenothiazine derivative, is used for the treatment of allergies, agitation, nausea, and vomiting. any patients taking promethazine are subject to urine drug screens and any potential interferences are important to recognize. Design. During an -month period, all patients presenting to the assachusetts General Hospital emergency department who had a finding of promethazine in their serum drug screen, and who also had a urine drug screen performed, were selected for inclusion in the study. The urine drug screen results (n 22 patients/samples) were then studied. Objective. To determine if promethazine use can cause false-positive urine amphetamine results in widely used drug of abuse immunoassays. Results. Thirty-six percent of patients taking promethazine had false-positive test results for urine amphetamines using the EIT II Plus onoclonal Amphetamine/ethamphetamine Immunoassay. Sixty-four percent of patients showed cross-reactivity greater than 20% higher than the blank calibrator rate. In a separate, related study, no promethazine-induced false-positive results were seen with the EIT II Plus, Triage, and TesTcard 9 amphetamine assays, or the Triage methamphetamine assay. Reduced chlorpromazine interference was also seen with these other assays. Conclusions. False-positive urine amphetamine results can be obtained in patients taking promethazine. Promethazine metabolite(s), and not the parent compound, are the likely cause of these urine false-positive results obtained with EIT II Plus onoclonal Amphetamine/ethamphetamine Immunoassay. Immunoassays from different manufacturers can have very different interference profiles, which the pathologist and laboratory scientist must understand and relay to clinicians. (Arch Pathol Lab ed. 2006;130:1834 1838) Immunoassays are frequently employed for urine screening for drugs of abuse because they are simple, easily automated, and available in point-of-care testing format. 1 However, there are some disadvantages to immunoassays. The earlier version of the monoclonal enzyme multiplied immunoassay technique (EIT), EIT dau (Syva Corp, Palo Alto, Calif) amphetamine assay, provided frequent false-positive results caused by phenylpropanolamine, ranitidine, chlorpromazine, and other medications. 2 4 Falsepositive results can be challenging and expensive for a laboratory. In addition, confirmatory analysis of either positive or negative results may be necessary in certain clinical situations. any immunoassay platforms for drug detection are available, and the specificity and sensitivity for detection of use of a particular drug can vary greatly Accepted for publication ay 5, 2006. From the Clinical Laboratory Division, Department of Pathology, Brigham and Women s Hospital and Harvard edical School (Dr elanson); and the Division Of Laboratory edicine, Department of Pathology, assachusetts General Hospital and Harvard edical School (Drs Lee-Lewandrowski, Griggs, Long, and Flood), Boston, ass. The authors have no relevant financial interest in the products or companies described in this article. Reprints: James G. Flood, PhD, Division of Laboratory edicine, assachusetts General Hospital, 55 Fruit St, GRJ 5, Boston, A 024 (e-mail: jflood@partners.org). depending on which manufacturer s assay is used, and which drug within a class is to be detected. A platform should be chosen that most accurately detects the drugs abused by the laboratory s patient population. Clinicians are frequently unaware of the limitations of urine drug screens, and inaccurate interpretation of results can adversely affect patient management. Furthermore, a stigma is attached to patients who have positive test results in the types of tests used to detect drugs of abuse. In 1993 a new assay was developed, the EIT II onoclonal Amphetamine/ethamphetamine (EIT-A) immunoassay (Syva Corp), which reportedly had a higher specificity for amphetamines and methamphetamines. Despite the improved specificity, the EIT-A still provided some false-positive amphetamine results caused by certain classes of medications such as antipsychotics and antihistamines. 3 Studies have demonstrated that chlorpromazine or bupropion use can cause a false-positive result. 4,5 Importantly, chlorpromazine metabolite(s), not the parent drug, were shown to cause the interference. 4 The EIT-A assay was also shown to be more sensitive to bupropion metabolites than the parent compound. 5 Only certain compounds have been tested for cross-reactivity, and the manufacturer provides a list of these in the package insert. any compounds and their metabolites are not screened for potential cross-reactivity, which makes it im- 1834 Arch Pathol Lab ed Vol 130, December 2006 Phenothiazine Interference in Drug of Abuse Immunoassays elanson et al

Table 1. Comparison of 6 Amphetamine/ethamphetamine Urine Drug of Abuse Assays* Assay No. Assay Name Calibrator Compound Calibrator Cutoff, g/l ethod Amphetaine 6 TesTcard 9 ethamphetamine 1 EIT-A d-ethamphetamine 1000 Enzyme multiplied immunoassay technique for automated analyzers 2 EIT-A d-ethamphetamine 1000 See above 3 Triage Amphetamine d-amphetamine 1000 Fluorescence immunoassay for pointof-care testing 4 Triage d-ethamphetamine 1000 See above ethamphetamine 5 TesTcard 9 d-amphetamine 300 Immunochromatography for point-ofcare testing d-ethamphetamine 500 See above * EIT-A indicates EIT II Plus onoclonal Amphetamine/ethamphetamine immunoassay; EIT-A, EIT II Plus amphetamine immunoassay. All assays use antibodies with different specificities to each drug within the amphetamine and/or methamphetamine class. Please refer to the manufacturer s package insert to determine the sensitivity of each assay for specific drugs. These assays are designed to detect drugs in both the amphetamine and methamphetamine class using the d-methamphetamine calibrator. Lower calibrator cutoffs (300 or 500 g/l of d-methamphetamine) are also available from the manufacturer. portant for clinical laboratories to report any new observations. Data from our toxicology laboratory reveal that many patients taking promethazine also have amphetamine detected in their urine, even though the patient denies taking any form of amphetamine. We investigate the validity of these amphetamine results in an attempt to determine any interference of promethazine in the EIT-A assay. Finally, we report results of 6 different test systems on urine samples from patients ingesting promethazine or chlorpromazine. ATERIALS AND ETHODS We identified all patients presenting to the emergency department during an -month period who had promethazine detected in their serum toxicology screen and who also had a urine toxicology screen performed. If promethazine was detected in a patient serum, we then used the laboratory information system to retrospectively identify those patients who also had an EIT- A urine toxicology screen performed at the same time as their serum analysis. All patients with promethazine detected in their serum and who also had an EIT-A urine toxicology screen, regardless of the results, were included in the study (n 22). Results of the EIT-A drug screen were obtained from the laboratory information system and used in the data analysis. Blood and urine samples obtained from patients in the emergency department are sent to the laboratory via a pneumatic tube transportation system and processed immediately on receipt by the laboratory. On completion of analysis (usually 1 3 hours), all urine and serum samples are stored frozen at 20 C. Toxicology analysis for all serum and select urine specimens (as described later in an adjunct study) was performed using liquid chromatography with photodiode array detection (LC- PDA). 6 The detection limits for phenothiazines, amphetamines, and amphetamine-like compounds using LC-PDA are as follows: amphetamine (20 g/l), chlorpromazine (20 g/l), methamphetamine (35 g/l), DA (70 g/l), ephedrine (60 g/l), phentermine (20 g/l), promethazine (20 g/l), and pseudoephedrine (70 g/l). Urine toxicology screens were performed on a Hitachi 9 analyzer (Roche Diagnostics, Indianapolis, Ind) using the EIT- A assay. If the enzyme rate observed using a patient s urine sample exceeds the rate generated using the 1000 g/l d-methamphetamine calibrator, the sample is considered presumptive positive for amphetamines. The assay also has a limit of detection of 1000 g/l for d-amphetamine. In an adjunct study, we selected consecutive patients from the emergency department who were found to have either promethazine (n 3) or chlorpromazine (n 6) in their serum and a positive result from an EIT-A urine amphetamine screen. Their serum and urine samples were stored at 20 C for further toxicology analysis. The LC-PDA analysis of the urine (promethazine) or the LC-PDA analysis of serum (chlorpromazine) was performed to determine whether the patients had taken amphetamines. None of the samples had amphetamines detected. All 9 samples producing false-positive urine amphetamine results were also analyzed using 5 other immunoassay platforms, in addition to the EIT-A assay previously described (Table 1). The EIT II Plus amphetamine assay (EIT-A) (Syva Corp) was performed on a Hitachi 9 analyzer. Similar to the EIT- A, the EIT-A uses a 1000 g/l d-methamphetamine calibrator. Urine sample results with enzyme rates above this cutoff are considered positive. The Triage TOX Drug Screen (Biosite Incorporated, San Diego, Calif) and TesTcard 9 (Varian, Inc, Palo Alto, Calif) point-of-care drugs of abuse immunoassays for amphetamine and methamphetamine were performed according to the manufacturers instructions. The Triage amphetamine and methamphetamine assays use calibrator cutoff concentrations of 1000 g/l. In the TesTcard 9 system, 300 g/l of d-amphetamine and 500 g/l of d-methamphetamine are used as calibrator cutoff concentrations for the amphetamine and methamphetamine assays, respectively. RESULTS We retrospectively reviewed the results of all patients who presented to the assachusetts General Hospital emergency department during an -month period who had promethazine detected in their serum. From these patients, we found 22 who also had a urine toxicology screen performed; these 22 patients were included in this study (Table 2). Using LC-PDA, no amphetamine or amphetamine-related compounds were detected in the serum of any of the patients. Eight (36 %) of the 22 patients taking promethazine had positive test results for urine amphetamine (Table 2 [results in bold]). Results greater than 100% of the cutoff calibrator were considered positive. In addition, 4 (29 %) of the 14 negative results were more than 50% higher than the amphetamine-free blank calibrator used in the assay (Table 2 [results in italic]), suggesting some effect of promethazine on assay results. Sixty-four percent of patients taking promethazine showed cross-reactivity greater than 20% above the blank calibrator rate using the EIT-A assay. In another experiment, we studied paired urine and serum samples from 9 patients in the emergency department using 5 different amphetamine immunoassay plat- Arch Pathol Lab ed Vol 130, December 2006 Phenothiazine Interference in Drug of Abuse Immunoassays elanson et al 1835

Table 2. Urine Enzyme ultiplied Immunoassay Technique II Plus onoclonal Amphetamine/ ethamphetamine Immunoassay (EIT-A) Results in Patients With Promethazine in Their Serum Patient No. 1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 17 18 19 20 21 22 Qualitative Results by EIT-A Cutoff Calibrator, %* 153 122 150 138 29 66 106 130 52 99 6 168 22 15 129 92 * The 1000 g/l methamphetamine cutoff calibrator has an enzyme rate of approximately 70 mau/min more than the blank at 340 nm. Patient results greater than 100% of the activity of the cutoff calibrator are reported as positive (results in bold). Patient results greater than 50% above the amphetamine-free blank calibrator are shown in italic. forms (Table 3). Serum sample results from all these patients were positive for either promethazine or chlorpromazine, and negative for amphetamines using LC-PDA. Each urine sample from these patients tested positive for amphetamines by EIT-A, and LC-PDA did not detect any amphetamines in the urine of the promethazine users. The EIT-A assay uses a 1000 g/l d-methamphetamine cutoff calibrator. All 9 samples were higher than the EIT-A calibrator and the results were reported as positive. The EIT-A assay detects d-amphetamine at about the same concentration as d-methamphetamine, which is 1000 g/l. Eight of the 9 urine samples had EIT-A enzyme rates indistinguishable from the blank calibrator (% of the separation between the blank and cutoff calibrator). The ninth sample had a rate 22% higher than the blank calibrator. Using 300 g/l as the cutoff (an option supported by the manufacturer), this ninth sample had a rate 89% of the cutoff calibrator. In this series of patients, the Triage amphetamine and methamphetamine assays all had negative findings. TesTcard 9 amphetamine tests also all had negative findings. Tes- Tcard 9 methamphetamine test results were negative in 7 samples and positive in 2 samples. COENT Promethazine is used in a variety of clinical situations and its potential cross-reactivity in drugs of abuse assays is clinically relevant. 7 9 Promethazine use appears to be increasing according to findings at this institution. Promethazine was detected in 23 (1.8%) of 1290 serum toxicology samples obtained from patients presenting to the emergency department during a recent 47-day period. A 20- /L limit of detection for promethazine using LC-PDA allows detection of the drug in serum samples from patients taking typical therapeutic doses. As a consequence, our study population resembles a typical one from an emergency department, not a group of patients with known toxic serum promethazine concentrations. ore than one third of our patient population with a positive finding for serum promethazine tested positive for amphetamine in urine using EIT-A. In a previous study, 18 inmates received 50 mg of promethazine every day and their urine was subsequently screened by an earlier-generation EIT-A assay for drugs of abuse. 4 Table 3. Results of Paired Serum/Urine Samples Having Both Serum Toxicology Screens for Promethazine or Chlorpromazine, and a False- Urine Enzyme ultiplied Immunoassay Technique II Plus onoclonal Amphetamine/ethamphetamine Assay (EIT-A)* Patient No. Serum LC-PDA Findings Urine LC-PDA Findings EIT-A 1 Promethazine for amphetamines (146) 2 Promethazine for amphetamines (133) 3 Promethazine for amphetamines (161) 4 Chlorpromazine Not tested (179) 5 Chlorpromazine Not tested (267) 6 Chlorpromazine Not tested (154) 7 Chlorpromazine Not tested (0) 8 Chlorpromazine Not tested (161) 9 Chlorpromazine Not tested (137) EIT-A 1836 Arch Pathol Lab ed Vol 130, December 2006 Phenothiazine Interference in Drug of Abuse Immunoassays elanson et al (22) A TRIAGE TestCard 9 A N N N P N N N P * LC-PDA indicates liquid chromatography with photodiode array detection; EIT-A, EIT II Plus amphetamines assay; C, calibrator compound;, d-methamphetamine calibrator; A, d-amphetamine calibrator; N, negative result using the point-of-care assays; P, positive result using the point-of-care assays. See Table 1 for calibrator compounds and concentrations. The numbers in parentheses are the percent enzyme activity relative to the cutoff calibrator.

Three inmates (17%) had amphetamine detected in their urine. No false-positive results were seen when less than 50 mg/day promethazine was administered. In this study, we demonstrated that false-positive results by EIT- A are caused by promethazine use. The higher falsepositive frequency observed in this study may be explained by the differences in the subject population. All subjects in this study were patients admitted to the emergency department in whom serum and urine toxicology testing was ordered, and no limit was placed on the amount of promethazine the patients had received. A previous study by Smith-Kielland et al 4 indicated that preparations of promethazine up to 3000 mg/l tested negative by the EIT-A assay for amphetamine. In addition, the Syva package insert states that promethazine shows a negative response. This suggests that a promethazine metabolite(s), not promethazine itself, is responsible for the false-positive results described in Table 2. Chlorpromazine metabolites cross-react in the EIT-A assay. 3,4 Promethazine and chlorpromazine possess similar phenothiazine chemical structures, supporting the hypothesis that a promethazine metabolite(s) interferes with the EIT-A assay. The metabolism of promethazine has not been thoroughly investigated in humans, but it is extensively metabolized. The major metabolites are promethazine sulfoxide, norpromethazine, and monodesmethyl promethazine sulfoxide. 10 Primary standards of promethazine metabolites are not currently available so we were unable to definitively determine which of the promethazine metabolites are responsible for the cross-reactivity in the EIT-A assay. anufacturers cannot test for all potential cross-reactants in drugs of abuse assays, and laboratories and clinicians should be encouraged to report false-positive results in these assays. We also suggest that manufacturers expand the information included in package inserts to include whether a compound exhibits significant cross-reactivity compared with a blank solution. We believe this extra information will aid investigators studying suspected false-positive results. Nine samples with false-positive amphetamine results by EIT-A assay were examined by 5 alternate immunoassays (Table 1) to compare cross-reactivity profiles. All 5 methods had reduced interference to both promethazine and chlorpromazine use. Eight of the 9 urine samples tested by EIT-A had enzyme rates indistinguishable from the blank calibrator (ie, %). However, the ninth sample had a rate 22% above the blank. The ninth sample also had the highest degree of interference in the EIT- A assay, and generated a positive result by TesTcard 9 methamphetamine assay. The EIT-A assay can be made more sensitive by using either of the available 300 and 500 g/l cutoff calibrators rather than the 1000 g/l cutoff used in this study. Using a lower cutoff of 300 g/l, the ninth sample had a rate 89% of the cutoff calibrator, extremely close to the rate that would be judged positive, raising the possibility that some cross-reactivity caused by phenothiazine metabolites may still remain in the EIT-A assay. No false-positive results were seen with the Triage amphetamine or methamphetamine assay or the TesTcard 9 amphetamine assay, indicating improved specificity compared with EIT-A. The 2 positive results obtained by TesTcard 9 methamphetamine assay may represent false-positive results. As an alternative explanation, the manufacturer of TesTcard 9 claims a lower detection limit for methamphetamine than the other assays, so it may be detecting amphetamine concentrations that the other immunoassays cannot. However, the negative LC-PDA finding argues against this alternate explanation. Although the sample size is limited, evidence strongly indicates that the EIT-A, Triage, and TesTcard 9 assays provide superior specificity compared with EIT-A in samples from patients receiving promethazine or chlorpromazine. A recent comparison of 6 urine amphetamine immunoassay systems found the EIT-A to be superior in specificity to both EIT-A and the other 4 systems studied. Our data also suggest that EIT-A has superior specificity to EIT-A. However, the sensitivity of EIT-A should be considered. The manufacturer (Syva Corp) states the concentrations of DA ( Ecstasy ) necessary to yield a positive result for the EIT-A and EIT-A assays (using a 1000 mg/l cutoff) as 9140 and 34 274 mg/l, respectively. The concentrations illustrate that the EIT-A assay is less sensitive for detection of DA use than the EIT-A assay. We and others have observed discordant EIT amphetamine-type results, in which the urine EIT-A is positive and the EIT-A is negative. The patients at our hospital with discrepant results had only DA in their serum, consistent with the manufacturer s data that EIT-A is less sensitive to DA detection. In our emergency department population, the frequency of phenothiazine use exceeds that of DA. In addition, our serum toxicology assay has a limit of detection of 70 g/l for DA. Changing from the EIT-A to the EIT-A assay for urine testing at our institution would improve specificity with little cost to sensitivity. In different patient populations, a switch from the EIT-A to the EIT-A assay may have different consequences. The National Academy of Clinical Biochemistry recently recommended that, for patients in emergency departments, highly specific amphetamine immunoassays directed at the illicit amphetamines (such as DA and methamphetamine) are less useful compared with assays sensitive to a broad spectrum of sympathomimetic, stimulant, and illicit amines. 12 We conclude that promethazine use can cause false-positive results in EIT-A urine amphetamine assays. Clinicians should understand this limitation when interpreting urine drug screens from laboratories employing this method. We urge that promethazine be added to the EIT-A package insert listing of compounds that may cause false-positive results. Reduced interference by phenothiazines was observed using these alternative drugs of abuse amphetamine immunoassay platforms; EIT-A, Triage, and TesTcard9. References 1. Colbert DL. Drug abuse screening with immunoassays: unexpected crossreactivities and other pitfalls. Br J Biomed Sci. 1994;51:136 146. 2. Poklis A. Unavailability of drug metabolite reference material to evaluate false positive results for monoclonal EIT dau assay of amphetamine [letter]. Clin Chem. 1992;38:2560. 3. Olsen K, Gulliksen, Christopherson AS. etabolites of chlorpromazine and bromipheniramine may cause false-positive urine amphetamine results with monoclonal EIT dau Immunoassay [letter]. Clin Chem. 1992;38:6 612. 4. Smith-Kielland A, Olsen K, Christopherson AS. False-positive results with EIT II amphetamine/metamphetamine assays in users of common psychotropic drugs [letter]. Clin Chem. 1995;41:951 952. 5. Nixon AL, Long WH, Puopolo PR, Flood JG. Bupropion metabolites produce false-positive urine amphetamine results [letter]. Clin Chem. 1995;41:955 956. 6. Puopolo PR, Volpicelli SA, Johnson D, Flood JG. Emergency toxicology testing (detection, confirmation, and quantification) of basic drugs in serum by Arch Pathol Lab ed Vol 130, December 2006 Phenothiazine Interference in Drug of Abuse Immunoassays elanson et al 1837

liquid chromatography with photodiode array detection. Clin Chem. 1991;37: 2124 2130. 7. TREC Collaborative Group. Rapid tranquillisation for agitated patients in emergency psychiatric rooms: a randomized trial of midazolam versus haloperidol plus promethazine. Br ed J. 2003;327:708 713. 8. Golembiewski JA, O Brien D. A systematic approach to the management of postoperative nausea and vomiting. J Perianesth Nurs. 2002;17:364 376. 9. Cowings PS, Toscano WB, DeRoshia C, iller NE. Promethazine as a motion sickness treatment: impact on human performance and mood states. Aviat Space Environ ed. 2000;71:1013 1022. 10. Baselt RC, ed. Disposition of Toxic Drugs and Chemicals in an, 6th ed. Foster City, NY: Biomedical Publications; 2002:890 891.. Verstraete AG, Vander Heyden F. Comparison of the sensitivity and specificity of six immunoassays for the detection of amphetamines in urine. J Anal Toxicol. 2005;29:359 364. 12. Wu AHB, ckay C, Broussard LA, et al. National academy of clinical biochemistry laboratory medicine practice guidelines: recommendations for the use of laboratory tests to support poisoned patients who present to the emergency department. Clin Chem. 2003;49:357 379. 1838 Arch Pathol Lab ed Vol 130, December 2006 Phenothiazine Interference in Drug of Abuse Immunoassays elanson et al