A Retrospective Analysis of Urine Drugs of Abuse Immunoassay True Positive Rates at a National Reference Laboratory

Transcription

1 Journal of Analytical Toxicology, 2016;40: doi: /jat/bkv133 Advance Access Publication Date: 13 December 2015 Article Article A Retrospective Analysis of Urine Drugs of Abuse Immunoassay True Positive Rates at a National Reference Laboratory Kamisha L. Johnson-Davis 1,2, *, Aaron J. Sadler 2, and Jonathan R. Genzen 1,2 1 University of Utah Health Sciences Center, Salt Lake City, UT 84108, USA, and 2 ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT 84108, USA *Author to whom correspondence should be addressed. ARUP Laboratories, 500 Chipeta Way, MS-115, Salt Lake City, UT 84108, USA. kamisha.johnson-davis@aruplab.com Abstract Urine drug screens are commonly performed to identify drug use or monitor adherence to drug therapy. The purpose of this retrospective study was to evaluate the true positive and false positive rates of one of our in-house urine drug screen panels. The urine drugs of abuse panel studied consists of screening by immunoassay then positive immunoassay results were confirmed by mass spectrometry. Reagents from Syva and Microgenics were used for the immunoassay screen. The screen was performed on a Beckman AU5810 random access automated clinical analyzer. The percent of true positives for each immunoassay was determined. Agreement with previously validated GC MS or LC MS-MS confirmatory methods was also evaluated. There were 8,825 de-identified screening results for each of the drugs in the panel, except for alcohol (N = 2,296). The percent of samples that screened positive were: 10.0% for amphetamine/methamphetamine/3,4-methylenedioxy-methamphetamine (MDMA), 12.8% for benzodiazepines, 43.7% for opiates (including oxycodone) and 20.3% for tetrahydrocannabinol (THC). The false positive rate for amphetamine/methamphetamine was 14%, 34% for opiates (excluding oxycodone), 25% for propoxyphene and 100% for phencyclidine and MDMA immunoassays. Based on the results from this retrospective study, the true positive rate for THC drug use among adults were similar to the rate of illicit drug use in young adults from the 2013 National Survey; however, our positivity rate for cocaine was higher than the National Survey. Introduction Urine drugs screens are often ordered to identify drug use or exposure in the emergency department or other medical settings, to monitor patient adherence to prescribed drug therapy, and to support various forensic applications (1 5). Urine drug screens are frequently performed by immunoassay on routine chemistry analyzers due to ease of use, relatively low cost, rapid result turnaround time and ability to detect several drugs and drug classes. However, immunoassays can be limited in sensitivity and specificity when compared with complex analytical methods such as chromatography coupled to mass spectrometry. Some immunoassay drug screens are prone to interferences due to poor specificity and produce false negative and/or false positive results, due largely to the cross-reactivity profiles of the antibody and/or the cut-off concentration of the assay (6 8). For example, some benzodiazepine immunoassays may or may not detect structurally similar compounds within the same class, such as clonazepam, alprazolam, lorazepam or their metabolites (9 12). Immunoassay specificity varies among commercial vendors, thus cut-offs for specific compounds or cross-reactivity within a drug class are not standardized. In addition, urine adulteration may lead to false negative results (13 16). Consequently, it may be important to follow up a positive or negative drug screen result obtained by immunoassay with definitive testing, based on the clinical scenario (17). The purpose of this retrospective study was to evaluate the overall true positive rate, as well as the performance of our in-house urine The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com 97

2 98 Johnson-Davis et al. immunoassay drug screen to accurately detect positive patient samples for different drugs. In addition, we also wanted to compare the true positive rate of some drugs to the national averages. Our in-house drug screen panel with reflex to confirmation/quantitation can detect several drugs and drug classes, which include: amphetamine/methamphetamine (AMPH), barbiturates (BARB), benzodiazepines (BENZO), cocaine (COC), ecstasy (MDMA), ethanol (ETOH), methadone (MTD, propoxyphene (PPXY), opiates (OPI: morphine, codeine, dihydrocodeine, hydrocodone and hydromorphone), oxycodone and oxymorphone (OXY), phencyclidine (PCP) and marijuana (tetrahydrocannabinol; THC). Methods General Using an IRB approved protocol, a limited dataset of consecutive patient results for the ARUP Drug Panel 9, Urine Screen with Reflex to Confirmation/Quantitation test was obtained from the ARUP data warehouse and used for analysis. All results were derived from routine clinical testing in our laboratory using patient specimens. The data extract included all screening immunoassay and confirmatory test results for each specimen. Screening immunoassays The Drug Screen 9 Panel tests for nine drug classes (AMPH/MDMA, BARB, BENZO, COC, MTD, OPI, PCP, PPXY and THC). Eleven immunoassays are included, as separate immunoassays are used for AMPH and MDMA, as well as for OPI and OXY. The Drug Screen 9A Panel includes the addition of the ETOH enzyme assay, thus a total of 12 analyte assays per specimen. The immunoassay screens were all performed on a Beckman AU5810 chemistry analyzer (Beckman Coulter; Brea, CA) using Syva Emit II Plus reagents (Siemens Healthcare Diagnostics; Newark, DE) for all assays except for OXY (DRI Microgenics; Fremont, CA). All assays (except ETOH) were set up to be used in a qualitative configuration (two point calibration), such that specimens above pre-defined cut-off (Table I) are reported as screen positive and reflexed to definitive testing. Table II describes the compounds used for calibration in each assay and the cross-reactivity of related compounds detected. In this configuration, instrument results are expressed as normalized units ( norm units ), whereby 100 norm units are equivalent to the cut-off concentration for that particular drug. For MDMA, an adjusted clinical cutoff of 108 norm units was used. In-house retrospective analysis was performed to evaluate the range of norm units above 100 for MDMA, which resulted in a false-positive result by the MDMA immunoassay. The number of false positive results was reduced when the norm units for the MDMA assay were greater than 108. Thus, the assay norm unit cut-off was changed from 100 to 108, in an effort to decrease the percent of false positive results for this assay; however, this may have increased the risk of false negatives. For subsequent analysis of MDMA in this report, the Syva Emit II Plus MDMA immunoassay was also configured (for comparative purposes) in a semiquantitative format (five point calibration) where results were reported in units of ng/ml. The semiquantitative method has not been implemented for clinical testing in our laboratory to date. Finally, the ETOH enzyme assay was configured in a semiquantitative format (two point calibration), such that instrument results are expressed as a concentration (mg/dl). Confirmatory assays Specimens with positive screen results were reflexed to definitive testing (LC MS-MS, GC MS, GC FID) before reporting. Characteristics of confirmatory panels are included in Table III; some tests (e.g., ETOH) were single drug confirmations, while other assays consisted of a panel of parent drugs and metabolites (e.g., BENZO). The presence of a single drug above cut-off in such a multianalyte panel was reported as positive for that particular drug. Specimens that were analyzed for BENZO and THC were hydrolyzed to generate free, unconjugated drugs in order to improve analytical sensitivity. The BENZO assay utilizes enzyme hydrolysis with β-glucuronidase and the THC assay utilizes base hydrolysis. Specimens analyzed for other drugs in the panel were not hydrolyzed prior to analysis. Specimen preparation consisted of solid-phase extraction, for the majority of the drugs. However, the BARB assay utilized solid-phase extraction, derivatization, followed by liquid liquid extraction and the COC assay used solid-phase extraction and derivatization prior to analysis. A summary of our method validation protocol consisted of 5 days of imprecision, using three to five concentrations with a replicate of three to five samples per concentration, 2 days of linearity with concentrations that span the analytical measurement range, 2 days of sensitivity to assess the lowest limit of quantification, accuracy study consisting of 20 or more fortified or patient specimens evaluated over a period of 1 3 days and one experiment to individually assess carryover, ion suppression, recovery and interference. Data exclusions and limitations As the data extract did not include patient or specimen identifiers, exclusion of potential multiple drug testing results from the same Table I. Screening Immunoassay Characteristics and Cut-offs Drug Kit Configuration Cut-off (concentration) Cut-off (instrument result) AMPH Syva Emit II Plus Qualitative 300 ng/ml 100 norm units BARB Syva Emit II Plus Qualitative 200 ng/ml 100 norm units BENZO Syva Emit II Plus Qualitative 200 ng/ml 100 norm units COC Syva Emit II Plus Qualitative 150 ng/ml 100 norm units ETOH Syva Emit II Plus Semiquantitative 40 mg/dl 40 mg/dl MDMA Syva Emit II Plus Qualitative 500 ng/ml 100 norm units (Δ to 108) MTD Syva Emit II Plus Qualitative 150 ng/ml 100 norm units OPI Syva Emit II Plus Qualitative 300 ng/ml 100 norm units OXY DRI/Microgenics Qualitative 100 ng/ml 100 norm units PCP Syva Emit II Plus Qualitative 25 ng/ml 100 norm units PPXY Syva Emit II Plus Qualitative 300 ng/ml 100 norm units THC Syva Emit II Plus Qualitative 20 ng/ml 100 norm units

3 Urine Drugs of Abuse Immunoassay Positivity Rates 99 Table II. Calibration Compounds, Cut-offs and Related Compounds that Produce a Result Approximately Equivalent to the Calibrator Cut-off Drug class Compound used for calibration Calibrator cut-off concentration (ng/ml) Related compounds detected at calibrator cut-off (concentration in ng/ml unless otherwise noted) AMPH d-methamphetamine 300 D,L-4-Methylamphetamine (4,400) D-Amphetamine (300) D,L-Methamphetamine (450) D,L-Amphetamine (625) L-Methamphetamine (725) L-Amphetamine (3,450) 1,3-Dimethylpentylamine (3,400) MDA (1,100) MDMA (5,200) MDEA (4,400) BARB Secobarbital 200 Allobarbital (345) Alphenal (284) Amobarbital (348) Aprobarbital (275) Barbital (1,278) 5-Ethyl-5-(4-hydroxyphenyl) barbituric acid (927) Butabarbital (274) Butalbital (304) Butobarbital (349) Cyclopentobarbital (304) Pentobarbital (252) Phenobarbital ( ) Talbutal (194) Thiopental (28,200) BENZO Lormetazepam 200 Alprazolam (65) 7-Aminoclonazepam (2,600) 7-Aminoflunitrazepam (590) Bromazepam (630) Chlordiazepoxide (3,300) Clobazam (260) Clonazepam (580) Clotiazepam (380) Demoxepam (1,600) N-Desalkylflurazepam (130) N-Desmethyldiazepam (110) Diazepam (70) Estazolam (90) Flunitrazepam (140) Flurazepam (190) Halazepam (110) α-hydroxyalprazolam (100) α-hydroxyalprazolam glucuronide (110) 1-N-Hydroxyethylflurazepam (150) α-hydroxytriazolam (130) Ketazolam (100) Lorazepam (600) Medazepam (150) Midazolam (130) Nitrazepam (320) Norchlordiazepoxide (2,600) Oxazepam (250) Prazepam (90) Sertraline (therapeutic doses may produce a positive result) Temazepam (140) Tetrazepam (70) Triazolam (130) COC Benzoylecgonine 150 Cocaine (18 53) Ecgonine (2 6) Table continues

4 100 Johnson-Davis et al. Table II. Continued Drug class Compound used for calibration Calibrator cut-off concentration (ng/ml) Related compounds detected at calibrator cut-off (concentration in ng/ml unless otherwise noted) MDMA MDMA 500 (Concentrations in μg/ml) Methylenedioxyamphetamine (MDA) (0.61) Methylenedioxyethylamphetamine (MDEA) (0.50) N-Methyl-1-(1,3-benzodioxol-5-yl)-2-aminobutane (MBDB) (0.43) 3,4-(Methylenedioxyphenyl)-2-butanamine (BDB) (0.78) Para-methoxyamphetamine (PMA) (22) Para-methoxymethamphetamine (PMMA) (9.0) 4-Hydroxy-3-methoxymethamphetamine (HMMA) (2,100) MTD Methadone 150 None OPI Morphine 300 Codeine ( ) Dihyrocodeine (291) Ethylmorphine (240) Hydrocodone (247) Hydromorphone (498) Levallorphan (>5,000) Levorphanol (1,048) Meperidine (>15,000) 6-Acetylmorphine (435) Morphine-3-glucuronide (626) Nalorphine (5,540) Naloxone (3,60,000) Oxycodone (1,500) Oxymorphone (9,300) OXY Oxycodone 100 Noroxymorphone (at 500,000 has <0.1% cross-reactivity) Noroxycodone (at 50,000 has <0.1% cross-reactivity) Oxycodone (100) Oxymorphone (100) PCP Phencyclidine 25 N, N-Diethyl-1-phenylcyclohexylamine (PCDE) (234) 1-(4-Hydroxypiperidino)phencyclohexane (420) 1-(1-Phenylcyclohexyl)morpholine (PCM) (41) 1-(1-Phenylcyclohexyl)pyrrolidine (PCPy) (54) 4-Phenyl-4-piperidinocyclohexanol (32) 1-[1-(2-Thienyl)-cyclohexyl]morpholine (TCM) (80) 1-[1-(2-Thienyl)-cyclohexyl]piperidine (TCP) (37) 1-[1-(2-Thienyl)-cyclohexyl]pyrrolidine (TCPy) (83) PPXY Propoxyphene 300 Norpropoxyphene (800) THC 11-nor-Δ 9 -THC-9-COOH 20 8-β-11-Dihydroxy-Δ 9 -THC (24) 8-β-Dihydroxy-Δ 9 -THC (26) 11-Hydroxy-Δ 8 -THC (43) 11-Hydroxy-Δ 9 -THC (42) 9-Carboxy-11-nor-Δ 9 -THC-glucuronide (79) References: Beckman Coulter, Inc. (2015, June). Emit II Plus Barbituate Assay. Brea, CA; Beckman Coulter, Inc. (2015, June). Emit II Plus Benzodiazepine Assay. Brea, CA; Beckman Coulter, Inc. (2015, June). Emit II Plus Cannabinoid Assay. Brea, CA; Beckman Coulter, Inc. (2015, June). Emit II Plus Cocaine Metabolite Assay. Brea, CA; Beckman Coulter, Inc. (2015, June). Emit II Plus Opiate Assay. Brea, CA; Beckman Coulter, Inc. (2013, September). Emit II Plus Phencyclidine Assay. Brea, CA; Beckman Coulter, Inc. (2015, June). Emit II Plus Propoxyphene Assay. Brea, CA; Beckman Coutler, Inc. (2013, September). Emit II Plus Amphetamines Assay. Brea, CA; Siemens Healthcare Diagnostics Inc. (2013, April). Emit II Plus Ecstasy Assay. Newark, DE; Thermo Fisher Scientific, Inc. (2014, December). Drugs of Abuse Cross Reactivity Guide. Fremont, CA. patients was not possible. Results were included only in cases in which all screening and confirmatory testing was reported on the same accession number. Results were excluded if the patient gender was not listed. Specimens in which interfering substances were present in the chromatogram of the mass spectrum for any drug component result were reported as unable to quantify due to interfering substance and excluded from analysis. Data analysis Data analysis was performed in Excel 2010 (Microsoft Inc, Redmond, WA) and visualized in SigmaPlot 11 (Systat Software Inc, San Jose, CA). Patient age was rounded to the nearest year to facilitate binning and comparison. Data are presented as mean ± SD unless otherwise indicated. Statistical assessment was conducted using Student s t-test in SigmaPlot. Cross tabulation of specimens with multiple confirm-positive drug classes was conducted in SPSS 18 (PASW; IBM, Armonk, NY). Results Immunoassay screening results A total of 8,825 urine screens were reviewed in this analysis; 45.4% were from males and 54.6% were from females. Overall, the

5 Urine Drugs of Abuse Immunoassay Positivity Rates 101 Table III. Confirmatory Assay Characteristics and Cut-offs Assay Drugs detected Method Cut-off (concentration) AMPH and MDMA Amphetamine LC MS-MS 200 ng/ml Methamphetamine LC MS-MS 200 ng/ml Methylenedioxyamphetamine MDA LC MS-MS 200 ng/ml 3,4-Methylenedioxy-methamphetamine, MDMA LC MS-MS 200 ng/ml Methylenedioxyethylamphetamine, MDEA LC MS-MS 200 ng/ml BARB Amobarbital GC MS 50 ng/ml Butalbital GG MS 50 ng/ml Pentobarbital GC MS 50 ng/ml Phenobarbital GC MS 50 ng/ml Secobarbital GC MS 50 ng/ml BENZO Diazepam LC MS-MS 20 ng/ml Oxazepam LC MS-MS 20 ng/ml Temazepam LC MS-MS 20 ng/ml Nordiazepam LC MS-MS 20 ng/ml Lorazepam LC MS-MS 20 ng/ml Alprazolam LC MS-MS 5 ng/ml α-hydroxyalprazolam LC MS-MS 5 ng/ml Clonazepam LC MS-MS 5 ng/ml 7-Aminoclonazepam LC MS-MS 5 ng/ml Alpha-hydroxytriazolam LC MS-MS 20 ng/ml 2-Hydroxyethylflurazepam LC MS-MS 20 ng/ml Desalkylflurazepam LC MS-MS 20 ng/ml Midazolam LC MS-MS 20 ng/ml Chlordiazepoxide LC MS-MS 20 ng/ml Prazepam LC MS-MS 20 ng/ml Zolpidem LC MS-MS 20 ng/ml Alpha-hydroxymidazolam LC MS-MS 20 ng/ml COC Benzoylecgonine GC MS, LC MS-MS 50 ng/ml ETOH GC/flame ionization 5 mg/dl MTD MTD LC MS-MS 10 ng/ml EDDP LC MS-MS 10 ng/ml OPI/OXY Hydrocodone LC MS-MS 20 ng/ml Hydromorphone LC MS-MS 20 ng/ml Codeine LC MS-MS 20 ng/ml Morphine LC MS-MS 20 ng/ml 6-Acetylmorphine LC MS-MS 20 ng/ml Norhydrocodone LC MS-MS 20 ng/ml Oxycodone LC MS-MS 20 ng/ml Oxymorphone LC MS-MS 20 ng/ml Noroxycodone LC MS-MS 20 ng/ml Noroxymorphone LC MS-MS 20 ng/ml PCP LC MS-MS 10 ng/ml PPXY Propoxyphene and norpropoxyphene LC MS-MS 10 ng/ml THC 11-nor-9-carboxy-delta-9-THC LC MS-MS 5 ng/ml population of females was just slightly younger (41.4 ± 16.3 years; range 0 97) than males (42.9 ± 15.9 years; range 0 91); P < Only a subset of orders (n = 2,296) requested the version of the orderable panel that included ETOH. In total, 99,371 individual screening immunoassays were performed as part of the Drug Screen 9 Panel for this analysis, with an overall screen positive rate of 8.4% (Table IV). The combined opiates (OPI + OXY) screen positive rate was 43.7%. There was a high percent of positive screens observed for OPI (29.9%) and THC (20.3%), followed by OXY (13.8%), BENZO (12.8%) and AMPH (8.0%). There was a low percent ( 3%) of positive screens observed for MTD, COC, MDMA, BARB and ETOH. Positive screens for PCP and PPXY were exceedingly rare (0.05%). Histograms showing all patient instrument results for immunoassay screens are shown in Figure 1 (Parts 1 and 2). As ETOH is configured with a semiquantitative format, its x-axis is expressed as a concentration (mg/dl; Figure 1E). All other assays are expressed in normalized units to the assay concentration cut-off (Table I and Figure 1). Some assays (e.g., MTD, Figure 1G; and OXY, Figure 1I) show an excellent discrimination with bimodal distributions reflecting negative (left peak) and positive (right peak) specimens on opposite sides of the cut-off (dotted line). Other assays (e.g., BENZO, Figure 1C; and OPI, Figure 1H) show more of a continuum of results in-between the bimodal distributions. Two notable distributions, however, were observed in these histograms. For THC, an apparent trimodal distribution of results was observed (Figure 1L). Further investigation revealed that the two smaller peaks in the positive (right-hand) region of the graph corresponded to an old lot (left) and new lot (right) of THC reagents. Screen positivity rates using the old lot (20.7%) and new lot (19.8%) were

6 102 Johnson-Davis et al. Table IV. Summary Results from Immunoassay Screens and Confirmatory Testing Immunoassay screens Confirmation (tested when screen POS) Confirmed POS Specimens Total tested (#) Pos (#) Pos (%) Confirm pos (#) Confirm neg (#) True pos (%) False pos (%) Males (#) Females (#) Males (%) Females (%) AMPH a 8, BARB 8, BENZO 8,825 1, , b 653 b COC 8, ETOH 2, MDMA a 8, MTD 8, OPI c 8,825 2, , OXY c 8,825 1, , c 684 c PCP 8, PPXY 8, THC 8,825 1, , Totals 99,371 8, % 7,142 1, % 14.6% 3,380 3,863 a For AMPH and MDMA, a shared confirmation panel was performed if either screen was positive. This resulted in 0 additional AMPH confirmations and 0 additional MDMA confirmations. b Eight additional positive BENZO confirmations were identified. In these cases, clinicians ordered separate BENZO confirmatory testing, in addition to the negative BENZO screen result from the screen-to-reflex orderable, on the same specimen. c For OPI and OXY, a shared confirmation panel was performed if either screen was positive. This resulted in 93 additional OXY confirmations and 0 additional OPI confirmations. similar, as were the percent of screen positive specimens for THC that ultimately confirmed positive for THC by LC MS (99.2%, old lot; 99.5%, new lot). Finally, for MDMA, there was less of a distinct separation between the peaks of negative and positive results than was observed for other analytes (Figure 1F). Further investigation into MDMA immunoassay performance is presented later in the MDMA Immunoassay Studies section. Confirmatory testing Specimens that screen positive by immunoassay in the Drug Panel 9 test were reflexed to confirmatory testing using methodologies and panels noted in Table III. Some confirmatory tests (e.g., COC) detected one analyte (i.e., benzoylecgonine, method adapted from (18)), while others (e.g., BENZO and OPI/OXY) consisted of a multianalyte panel to quantify parent drugs and/or metabolites (19 21). Methods for AMPH, BARB, ETOH, MTD, PCP, PPXY and THC were adapted from previous published methods (22 27). The identification of one or more analytes in these panels was considered to be a single-positive confirmatory test for the purposes of Table IV summary data. The overall false positive rate (across all immunoassays) based on subsequent confirmatory testing was 14.6% (Table IV). Several immunoassays (e.g., BARB, BENZO, COC, ETOH, MTD, OXY, and THC) were associated with very low rates of false positives (0 3%). AMPH and OPI had a higher percent of false positives (13.8 and 34.0%, respectively). The MDMA false positive rate, however, was 100.0% in this dataset, as none of the 174 screen-positive specimens had positive confirmatory results for MDMA. Of note, the false positive rate for PCP (n = 4) was also 100%. Due to the low true-positive rate for PCP and the withdrawal of PPXY from the US market, these two assays are not beneficial for screening in this population (28 30). The distribution of all drugs and metabolites identified by confirmatory testing panels is presented in Figure 2. In Figure 2A, the most common drugs detected in the amphetamine confirmation drug panel were amphetamine and methamphetamine. Of note, the mass spectrometry method for the amphetamine panel cannot distinguish between D- and L- forms of amphetamines. The confirmed positive specimens could be due to prescription and/or illicit drug use. Moreover, MDMA was not detected at all in this sample population. In Figure 2B and C, the most common drugs detected in the opiate confirmation panel were hydrocodone and OXY, along with their related metabolites, which are highly prescribed opiate drugs. Noteworthy, a small percent (0.9%) of individuals were positive for the heroin metabolite, 6-acetylmorphine. In Figure 2D, some individuals could be prescribed MTD to treat opiate addiction, and MTD and its metabolite, 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP), were detected almost equally among positive MTD patients. Less than 1% (0.89%) of patient samples had positive MTD results, without the presence of EDDP. In addition, less than 0.3% (0.28%) of patient samples was positive for the EDDP without the presence of MTD. In Figure 2E, alprazolam and its metabolite, α-hydroxyalprazolam were the most common drugs detected in the benzodiazepine confirmation panel, as it is currently a highly prescribed benzodiazepine drug. There were also several other common BENZO drug and metabolites detected. Analysis of confirmed positive rates by gender was then conducted. It should be noted that sample sizes in this analysis were slightly larger (see Table IV, legend) for BENZO (includes 8 additional confirmations) and OXY (includes additional 93 confirmations). This was due to clinician-ordered add-on confirmatory testing (for BENZO) and the shared confirmation panel (for OXY) used when either OPI or OXY immunoassay was positive. There was a proportionately higher number of confirmed positive results for COC and THC observed for males, while a proportionally higher number of confirmed positive results for BARB and MTD was observed for females (threshold set as >10% more confirmations than would be predicted based on the frequency of orders for that gender). Age distributions (by gender) of patients with confirmed positive specimens for individual drug class are shown in Figure 3. While age distributions tended to be similar between genders (Figure 3), mild-to-moderate age differences were observed for AMPH (average age: male, 36.2 ± 13.8 years; female, 39.1 ± 12.4 years; P = 0.008),

7 Urine Drugs of Abuse Immunoassay Positivity Rates 103 MTD (average age: male, 48.4 ± 14.2 years; female, 40.8 ± 14.5 years; P < 0.001) and THC (average age: male, 38.7 ± 13.6 years; female, 35.9 ± 13.1 years; P < 0.001). To evaluate patterns of concomitant drug use in this population, a cross tabulation of positivity rates using confirmed positive drug categories was conducted. ETOH, MDMA, PCP and PPXY were excluded from this analysis due to low numbers of confirmed positive results (Table IV). Table V shows the positivity rates (for all drug classes) if a specimen was also positive for the drug listed in the lefthand column. For example, specimens that were true positive for BARB show a high rate of true positivity for BENZO (30.8%), OPI (26.4%) and OXY (19.5%). Specimens that were true positive for BENZO show a high rate of true positivity for OPI (35.0%), OXY (29.7%) and THC (24.8%), but a lower rate of true positivity for BARB (4.3%). Finally, specimens that were true positive for COC showed the highest rate of true positivity for THC (35.0%). Figure 1. Parts 1 and 2. Distribution of all immunoassay screening results. Histograms showing the distribution of instrument results for all immunoassay screens on the Beckman AU5810. X-axes indicate instrument results (in norm units) for Part 1: AMPH (A), BARB (B), BENZO (C), COC (D), MDMA (F) and concentration (mg/dl) for ETOH (E); Part 2 (in norm units): MTD (G), OPI (H), OXY (I), PCP (J), PPXY (K) and THC (L). Y-axes indicate the number of specimens with results at each norm unit or concentration. Note that Y-axes have been cropped to better visualize (and magnify) the distribution of positive specimens for each drug. Vertical dotted lines indicate the cut-off for immunoassay screen for each drug (as indicated in Supplementary Table SI); 100 norm units for AMPH, BARB, BENZO, COC, MTD, OPI, OXY, PCP, PPXY and THC; 108 norm units for MDMA; 40 mg/dl ETOH. MDMA immunoassay studies Given the high rate of false positives for MDMA (Table IV), and the unusual appearance of the MDMA histogram (Figure 1F), further investigation of MDMA immunoassay configuration and screening performance was conducted. As a first step, the assay vendor confirmed appropriate assay configuration on our AU5810 instrument during a site visit. To further evaluate the assay, a series of 40 urine specimens ( previously negative on our amphetamine/mdma confirmation panel) were then fortified with MDMA (n = 10 with concentrations below 500 ng/ml; n = 30 with concentrations above 500 ng/ml) and tested using an LC MS-MS method, as well as with EMIT Ecstasy immunoassays on the AU5810 with both semiquantitative configuration (5 level calibration, using a 500 ng/ml cut-off) and qualitative configuration (2 level calibration, using a 100 norm unit cut-off corresponding to 500 ng/ml). MDMA concentrations as determined by LC MS-MS correlated well with the concentration expected to be found in the fortified specimens (Figure 4A; linear regression, r 2 = 0.997, line not shown), although one minor outlier was observed and was attributed to a pipetting error during sample preparation as the result was confirmed on repeat analysis. Analysis of these fortified samples by MDMA immunoassay using semiquantitative (Figure 4B) and qualitative (Figure 4C) assay configurations, however, demonstrated that there was a good separation of true negative specimens using both configurations (Figure 4B and C; open circles), although there was an overall poor correlation of instrument results (in ng/ml or norm units) versus expected MDMA concentrations above the established 500 ng/ml (or 100 norm unit) cut-offs. Additionally, four clear (and one borderline) false positive results by immunoassay were observed using both calibration configurations (Figure 4B and C, gray triangles; also Table VI). Each of these false positive specimens was then tested for the presence of bupropion and/or trazodone using LC MS-MS methods, as these drugs have previously been associated with false positive MDMA immunoassay results (19, 21, 31). Of these specimens, two contained bupropion only, one contained both bupropion and trazadone, and one contained trazadone only. Possible explanations of these observations, and potential implications for MDMA testing, are included in Discussion. Discussion Urine drug screens are commonly performed to identify drug use or monitor adherence to drug therapy. We performed retrospective

8 104 Johnson-Davis et al. Figure 2. Confirmatory testing panels tabulation of drug positivity rates. Histograms showing the distribution of the number of drugs detected as components of multiple drug panels for AMPH/MDMA (A), OPI (B), OXY(C), MTD (D) and BENZO (E). While OPI/OXY are performed as combined confirmatory panel, they are separated here (B and C) to improve visualization. Abbreviations not otherwise defined in the text: MDA, methylenedioxyamphetamine; MDEA, methylenedioxyethylamphetamine; EDDP, 2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine. Figure 3. Distribution of confirmed true positives by gender and age. Vertical box plots showing the distribution of drug class true positives by age and gender (gray, male; open, female). Box plots show median (middle thick black line), 25th percentile (bottom of box), 75th percentile (top of box), 10th percentile (bottom whisker) and 90th percentile (top whisker). In each box plot, middle thin gray line indicates the mean. Gender differences in the ages of positive specimens for particular drugs are indicated by dotted lines with asterisks (AMPH, P = 0.008; MTD, P < 0.001; THC, P < 0.001). data analysis of our in-house urine drug panel to determine the overall true positive rate and to evaluate the performance of the immunoassays in this panel. Our data showed that there was almost an equal portion of males and females, primarily over the age of 25 years, who were evaluated by this urine drug panel. The overall immunoassay screen positive rate for drugs in this panel was 8.4%. The prescription drugs routinely detected were opiates and benzodiazepines, which are currently used in the treatment for pain management. Of note, out of the total number of positive samples, 29.7% (2,625 out of 8,825 specimens) were positive for AMPH, COC, PPXY and THC. Data analysis also revealed that the majority of individuals who tested positive were also combination drug users. For example, 35% of individuals who were positive for BENZO were also positive for opiates and an additional 25% were also positive for THC. Data in Figure 1 showed that the cut-offs for the majority of the immunoassays were sufficient to distinguish between negative and positive patients. One exception was the immunoassay for MDMA. Based on our investigative experiments for this assay, we discovered that instrument results plateaued once the urine concentrations of MDMA were above 500 ng/ml, regardless of whether the assay was performed in qualitative or semiquantitative configurations. Thus, an ideal clear bimodal separation of patient instrument results was not

9 Urine Drugs of Abuse Immunoassay Positivity Rates 105 Table V. Cross Tabulation of Multiple Drug Class Confirmed Positivity If positive for drug below Then positivity rate for other drugs (%) AMPH BARB BENZO COC MTD OPI OXY THC AMPH BARB BENZO COC MTD OPI OXY THC Figure 4. MDMA investigation. Analysis of fortified urine specimens with MDMA by LC MS-MS (A), immunoassay with semiquantitative configuration (B) and immunoassay with qualitative configuration (C). X-axes (for A C) cropped at 3,000 ng/ml to improve visualization of data points. With cropping, two specimens (X-axis; a: 5,000 ng/ml, b: 10,000 ng/l) are not visible in each graph. These two specimens showed linear detection by LC MS-MS (a: 4,884 ng/ml, b: 9,640 ng/ml), but nonlinearity by semiquantitative immunoassay (a: 951 ng/ml, c: 774 ng/ml) and qualitative immunoassay (a: norm units, b: norm units). All three methods would have interpreted these two specimens as being screen-positive ( 500 ng/ml cut-off). Symbols reflect strict quadrants based on cut-offs ( 500 ng/ ml and/or 100 norm units) if the assay was used as a screening test: open circle, true negative; filled circle, true positive; gray triangles, false positive; gray squares, false negative. as apparent as with other drugs. Investigation into possible reasons for false positives (Figure 4 and Table VI), however, suggested that the presence of other potential interfering drugs (e.g., bupropion and trazadone) may have been a more significant contributor to the MDMA false positive rate observed in our study (as opposed to lack of assay linearity above the positive cut-off) (31). Nonetheless, MDMA EMIT II immunoassay screening (in the absence of definitive testing) would not be advised based on data in the present report. Definitive testing may also be necessary when quantitative results are important for result interpretation or when results are inconsistent with clinical expectations. For example, a clinician may want to investigate an unexpected negative result via immunoassay by testing the same specimen on a confirmatory/definitive assay. The definitive test may be able to detect concentrations lower than the immunoassay positive cut-off concentration, if the immunoassay cut-off concentration is higher than in the definitive test. Definitive testing may also be utilized to confirm adherence by the detection of parent drug and/or metabolites. For example, retrospective data analysis from our inhouse opiate confirmation assay demonstrated that noroxycodone was present in 4.1% (1,811 of 44,397 patient results) of patient samples when oxycodone and oxymorphone were not detected, and noroxymorphone was present in 4.9% (2,201 of 44,397 patient results) of patient samples when oxycodone and oxymorphone were not detected. Therefore, if the immunoassay has poor cross-reactivity to Table VI. Characteristics of Discordant MDMA Results Expected concentration (ng/ml) LC MS-MS result (ng/ml) Immunoassay semiquant configuration (ng/ml) Immunoassay qualitative configuration (norm units) Specimen , Specimen , Specimen Specimen Specimen Correspond to results shown in Figure 4 (gray triangles). these metabolites then the immunoassay result generated would be false negative. Such data have also been reported by Heltsley et al. in 2010 (32). Moreover, false negatives can also be generated in the BENZO assay (11). False negative immunoassay results may occur for drugs that have substandard cross-reactivity (see Table II) such as chlordiazepoxide, norchlordiazepoxide, clonazepam, 7-aminoclonazepam and lorazepam, thus definitive testing may be necessary to confirm adherence to specific BENZO drugs. Some labs have also reported bypassing immunoassay screening for BENZO to perform definitive testing (33).

10 106 Johnson-Davis et al. According to the results from the 2013 National Survey on Drug Use and Health, the rate of illicit drug use among young adults years old was 21.5%, and illicit drug use among adults 26 years and older was 7.3% (34). Furthermore, the survey highlighted that the percent of reported drug use for THC was 19.1% for young adults (18 25) and 5.6% for adults 26 years and older. Cocaine use was reported as 1.1% for young adults and 0.5% for adults 26 years and older. The survey contained data for the following drugs: marijuana, cocaine, heroin, hallucinogens (LSD, PCP, MDMA, peyote, mescaline and psilocybin mushrooms), as well as the nonmedical use of prescription-type pain relievers, tranquilizers, stimulants and sedatives (34). Our test menu at ARUP Laboratories was limited and we could not perform a head-to-head comparison for all of the drugs listed in the 2013 National Survey on Drug Use and Health report. However, we performed a direct comparison for THC and COC positivity rates. The percent of true positives for THC ( 20%) from our drug panel was similar to the national rate of THC use among young adults; however, our COC positivity rate (2.7%) was higher than the national average for adults 18 years and older. The 2013 National Survey on Drug Use and Health also reported the rate of illicit drug use among individuals 12 years or older was higher for males than females (11.5 vs 7.3%, respectively) (34). As noted in Figure 3, in our analysis the overall age for our population of females was just slightly younger (41.4 ± 16.3 years; range 0 97) than males (42.9 ± 15.9 years; range 0 91); P < For confirmed positive specimens, there was no statistically significant difference in age between the overall population of females (44.6 ± 15.0 years; range 0 93) and males (44.7 ± 14.6 years; range 0 87); P = Drug screening (particularly panel-based orderables) are a significant healthcare expenditure. The motivation for our age and gender analysis was that (given the size of our dataset) we might be able to reveal significant trends in drug screen positivity rates that could be used to influence more cost-effective recommendations for testing in specific populations. Ultimately, the differences observed were found to be relatively minor. We believe that publication of these data, however, may be valuable in the literature for those considering similar analyses for comparative purposes. Conclusion Our study highlighted the immunoassay true positive rates from our in-house drug screen panel. There were five assays that had positive rates greater than 5%, which consisted of BENZO, OPY, OXY, THC and AMPH. MDMA, PCP and PPXY had the lowest number of true positives, which suggests that these assays have little clinical utility for screening in this patient population. True positive rates for drug use will vary based on the patient population that is tested (i.e., pain management and drug abstinence). Moreover, the assays for AMPH, MDMA, PCP and PPXY were more prone to false positives than the other analyte assays in the panel. The false positive results were likely due to compounds that have similar structures to the analytes of interest; thus exhibiting cross-reactivity to the antibody in the assay. Numerous publications have listed several drugs as potential agents that may cause false positive results by immunoassay (2, 8, 31, 35 47). The overall false positive rate for this drug panel was about 15%, which supports the need for definitive testing. In contrast, it also highlights the importance for selecting immunoassays with high specificity, which may reduce the number of reflex tests to definitive analysis and the additional testing expenses billed to the client that are caused by false positive immunoassay results. Some of the limitations in our evaluation include a limited test panel, which did not contain all of the drugs and hallucinogens that were listed in the National Survey, thus, we were unable to perform an exact one to one comparison. Another limitation to this retrospective study was that false negative results could not be detected, because negative results by immunoassay were not confirmed. Lastly, prescription information was also not available; therefore we were unable to determine if drug use was illicit or legitimate. Supplementary data Supplementary data are available at Journal of Analytical Toxicology online. Acknowledgments The authors thank David Davis and Lutsia Jensen for assistance in result query and retrieval, as well as Dr Jennifer T. Gosselin for statistical assistance with SPSS. Funding This work was supported by the ARUP Institute for Clinical and Experimental Pathology. References 1. Christo, P.J., Manchikanti, L., Ruan, X., Bottros, M., Hansen, H., Solanki, D.R. et al. (2011) Urine drug testing in chronic pain. Pain Physician, 14, Moeller, K.E., Lee, K.C., Kissack, J.C. (2008) Urine drug screening: practical guide for clinicians. Mayo Clinic Proceedings, 83, Phan, H.M., Yoshizuka, K., Murry, D.J., Perry, P.J. (2012) Drug testing in the workplace. Pharmacotherapy, 32, Hammett-Stabler, C.A., Pesce, A.J., Cannon, D.J. (2002) Urine drug screening in the medical setting. Clinica Chimica Acta, 315, Cone, E.J., Caplan, Y.H. (2009) Urine toxicology testing in chronic pain management. Postgraduate Medicine, 121, Markway, E.C., Baker, S.N. (2011) A review of the methods, interpretation, and limitations of the urine drug screen. Orthopedics, 34, Herring, C., Muzyk, A.J., Johnston, C. (2011) Interferences with urine drug screens. Journal of Pharmacy Practice, 24, Brahm, N.C., Yeager, L.L., Fox, M.D., Farmer, K.C., Palmer, T.A. (2010) Commonly prescribed medications and potential false-positive urine drug screens. American Journal of Health-System Pharmacy, 67, Valentine, J.L., Middleton, R., Sparks, C. (1996) Identification of urinary benzodiazepines and their metabolites: comparison of automated HPLC and GC MS after immunoassay screening of clinical specimens. Journal of Analytical Toxicology, 20, Mikel, C., Pesce, A.J., Rosenthal, M., West, C. (2012) Therapeutic monitoring of benzodiazepines in the management of pain: current limitations of point of care immunoassays suggest testing by mass spectrometry to assure accuracy and improve patient safety. Clinica Chimica Acta, 413, Manchikanti, L., Malla, Y., Wargo, B.W., Fellows, B. (2011) Comparative evaluation of the accuracy of benzodiazepine testing in chronic pain patients utilizing immunoassay with liquid chromatography tandem mass spectrometry (LC/MS/MS) of urine drug testing. Pain Physician, 14, Tenore, P.L. (2010) Advanced urine toxicology testing. Journal of Addictive Diseases, 29, Dasgupta, A. (2007) The effects of adulterants and selected ingested compounds on drugs-of-abuse testing in urine. American Journal of Clinical Pathology, 128,

11 Urine Drugs of Abuse Immunoassay Positivity Rates Jaffee, W.B., Trucco, E., Levy, S., Weiss, R.D. (2007) Is this urine really negative? A systematic review of tampering methods in urine drug screening and testing. Journal of Substance Abuse Treatment, 33, Scholer, A. (2004) The effect of urine manipulation on substance abuse testing. ToxChem und KrimTech, 71, Villena, V.P. (2010) Beating the system: a study of a creatinine assay and its efficacy in authenticating human urine specimens. Journal of Analytical Toxicology, 34, Owen, G.T., Burton, A.W., Schade, C.M., Passik, S. (2012) Urine drug testing: current recommendations and best practices. Pain Physician, 15, ES119 ES Taylor, R.W., Jain, N.C., George, M.P. (1987) Simultaneous identification of cocaine and benzoylecgonine using solid phase extraction and gas chromatography/mass spectrometry. Journal of Analytical Toxicology, 11, Marin, S.J., Roberts, M., Wood, M., McMillin, G.A. (2012) Sensitive UPLC MS MS assay for 21 benzodiazepine drugs and metabolites, zolpidem and zopiclone in serum or plasma. Journal of Analytical Toxicology, 36, Marin, S.J., McMillin, G.A. (2010) LC MS-MS analysis of 13 benzodiazepines and metabolites in urine, serum, plasma, and meconium. Methods in Molecular Biology, 603, Coles, R., Kushnir, M.M., Nelson, G.J., McMillin, G.A., Urry, F.M. (2007) Simultaneous determination of codeine, morphine, hydrocodone, hydromorphone, oxycodone, and 6-acetylmorphine in urine, serum, plasma, whole blood, and meconium by LC MS-MS. Journal of Analytical Toxicology, 31, Middleberg, R.A., Homan, J. (2012) Quantitation of amphetamine-type stimulants by LC MS-MS. Methods in Molecular Biology, 902, Mule, S.J., Casella, G.A. (1989) Confirmation and quantitation of barbiturates in human urine by gas chromatography/mass spectrometry. Journal of Analytical Toxicology, 13, Pontes, H., Guedes de Pinho, P., Casal, S., Carmo, H., Santos, A., Magalhaes, T. et al. (2009) GC determination of acetone, acetaldehyde, ethanol, and methanol in biological matrices and cell culture. Journal of Chromatographic Science, 47, Koster, R.A., Alffenaar, J.W., Greijdanus, B., VanDerNagel, J.E., Uges, D. R. (2014) Application of sweat patch screening for 16 drugs and metabolites using a fast and highly selective LC MS-MS method. Therapeutic Drug Monitoring, 36, Rumpler, M.J. (2014) Quantitative analysis of 11-nor-9-carboxy-tetrahydrocannbinol (THC-COOH) in urine by LC MS-MS following a simple filtration. Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences, 957, Crews, B., Mikel, C., Latyshev, S., West, R., Pesce, A. (2009) Unstable propoxyphene metabolite excreted in human urine is detected by liquid chromatography-tandem mass spectrometry. Journal of Analytical Toxicology, 33, Price, J. (2015) A trend analysis of laboratory positive propoxyphene workplace urine drug screens before and after the product recall. Journal of Addiction Medicine, 9, Puet, B., DePriest, A., Knight, J., Heltsley, R., Black, D.L., Caplan, Y.H. et al. (2013) Urine drug testing of chronic pain patients. V. Prevalence of propoxyphene following its withdrawal from the United States market. Journal of Analytical Toxicology, 37, Rengarajan, A., Mullins, M.E. (2013) How often do false-positive phencyclidine urine screens occur with use of common medications? Clinical Toxicology, 51, Marin, S.J., Doyle, K., Chang, A., Concheiro-Guisan, M., Huestis, M.A., Johnson-Davis, K.L. (2015) One hundred false-positive amphetamine specimens characterized by liquid chromatography time-of-flight mass spectrometry. Journal of Analytical Toxicology, doi: /jat/bkv Heltsley, R., Zichterman, A., Black, D.L., Cawthon, B., Robert, T., Moser, F. et al. (2010) Urine drug testing of chronic pain patients. II. Prevalence patterns of prescription opiates and metabolites. Journal of Analytical Toxicology, 34, Melanson, S.E., Ptolemy, A.S., Wasan, A.D. (2013) Optimizing urine drug testing for monitoring medication compliance in pain management. Pain Medicine, 14, SAMHSA (2013) Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings. default/files/nsduhresultspdfwhtml2013/web/nsduhresults2013. pdf (accessed November 13, 2015). 35. Vincent, E.C., Zebelman, A., Goodwin, C., Stephens, M.M. (2006) Clinical inquiries. What common substances can cause false positives on urine screens for drugs of abuse? The Journal of Family Practice, 55, , Baron, J.M., Griggs, D.A., Nixon, A.L., Long, W.H., Flood, J.G. (2011) The trazodone metabolite meta-chlorophenylpiperazine can cause falsepositive urine amphetamine immunoassay results. Journal of Analytical Toxicology, 35, Jenkins, A.J., Poirier, J.G. 3rd, Juhascik, M.P. (2009) Cross-reactivity of naloxone with oxycodone immunoassays: implications for individuals taking suboxone. Clinical Chemistry, 55, Casey, E.R., Scott, M.G., Tang, S., Mullins, M.E. (2011) Frequency of false positive amphetamine screens due to bupropion using the Syva EMIT II immunoassay. Journal of Medical Toxicology, 7, Vorce, S.P., Holler, J.M., Cawrse, B.M., Magluilo, J. Jr. (2011) Dimethylamylamine: a drug causing positive immunoassay results for amphetamines. Journal of Analytical Toxicology, 35, Ly, B.T., Thornton, S.L., Buono, C., Stone, J.A., Wu, A.H. (2012) Falsepositive urine phencyclidine immunoassay screen result caused by interference by tramadol and its metabolites. Annals of Emergency Medicine, 59, Collins, A.A., Merritt, A.P., Bourland, J.A. (2012) Cross-reactivity of tapentadol specimens with DRI methadone enzyme immunoassay. Journal of Analytical Toxicology, 36, Vidal, C., Skripuletz, T. (2007) Bupropion interference with immunoassays for amphetamines and LSD. Therapeutic Drug Monitoring, 29, Reidy, L., Walls, H.C., Steele, B.W. (2011) Crossreactivity of bupropion metabolite with enzyme-linked immunosorbent assays designed to detect amphetamine in urine. Therapeutic Drug Monitoring, 33, Kaplan, Y.C., Erol, A., Karadas, B. (2012) False-positive amphetamine/ ecstasy (MDMA/3,4-methylenedioxymethamphetamine) (CEDIA) and ecstasy (MDMA/3,4-methylenedioxymethamphetamine) (DRI) test results with fenofibrate. Therapeutic Drug Monitoring, 34, Saitman, A., Park, H.D., Fitzgerald, R.L. (2014) False-positive interferences of common urine drug screen immunoassays: a review. Journal of Analytical Toxicology, 38, Liu, L., Wheeler, S.E., Rymer, J.A., Lower, D., Zona, J., Peck Palmer, O.M. et al. (2015) Ranitidine interference with standard amphetamine immunoassay. Clinica Chimica Acta, 438, Shafiq, Q., Mutgi, A. (2010) Urine opiate screening: false-positive result with levofloxacin. Canadian Medical Association Journal, 182,