Substance abuse is a significant problem in the United

Interpretation and Utility of Drug of Abuse Immunoassays Lessons From Laboratory Drug Testing Surveys Stacy E. F. Melanson, MD, PhD; Leland Baskin, MD; Barbarajean Magnani, PhD, MD; Tai C. Kwong, PhD; Annabel Dizon, MT(ASCP); Alan H. B. Wu, PhD N Context. To assist with patient diagnosis and management, physicians from pain services, drug treatment programs, and the emergency department frequently request that urine be tested for drugs of abuse. However, urine immunoassays for drugs of abuse have limitations. Objectives. To use data from the College of American Pathologists Proficiency Testing Surveys to determine and summarize the characteristics, performance, and limitations of urine immunoassays for drugs of abuse. Design. Six years of urine drug testing proficiency surveys were reviewed. Results. Lysergic acid diethylamide and methaqualone are infrequently prescribed or abused and, therefore, testing may be unnecessary. However, implementation of more specific testing for methylenedioxymethamphetamine and oxycodone may be warranted. Each drug of abuse immunoassay exhibits a different cross-reactivity profile. Depending on the cross-reactivity profile, patients with clinically insignificant concentrations of drugs may have false-positive results, and patients with clinically significant concentrations of drugs may have false-negative results. Conclusions. Laboratory directors should be aware of the characteristics of their laboratories assays and should communicate these characteristics to physicians so that qualitative results can be interpreted more accurately. Furthermore, manufacturer s claims should be interpreted with caution and should be verified in each organization s patient population, if possible. (Arch Pathol Lab Med. 2010;134:735 739) Substance abuse is a significant problem in the United States. 1 More than 1.7 million emergency department visits in 2006 were associated with drug misuse or abuse, of which, 55% involved an illicit drug. In addition, painand addiction-management physicians are treating more patients who require prescriptions for addictive and narcotic drugs. For this reason, drug of abuse urine (DAU) testing is frequently ordered by physicians in a variety of clinical settings to diagnosis overdose, verify compliance, and exclude illicit or unknown prescription drug use. 2 5 The common drugs or classes of drugs that DAU testing uses assays for include amphetamines, barbiturates, benzodiazepines, cannabinoids (tetrahydrocannabinol [THC]), cocaine metabolite/benzoylecgonine, methadone, Accepted for publication July 2, 2009. From the Department of Pathology, Brigham and Women s Hospital, Harvard Medical School, Boston, Massachusetts (Dr Melanson); the Calgary Laboratory Services and the Department of Pathology and Laboratory Medicine, University of Calgary, Calgary, Alberta, Canada (Dr Baskin); the Department of Pathology, Tufts Medical Center, Tufts University School of Medicine, Boston (Dr Magnani); the Department of Pathology and Laboratory Medicine, University of Rochester Medical Center, Rochester, New York (Dr Kwong); the Proficiency Testing Division, College of American Pathology, Northfield, Illinois (Ms Dizon); and the Clinical Laboratory, San Francisco General Hospital, San Francisco, California (Dr Wu). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Stacy E. F. Melanson, MD, PhD, Brigham and Women s Hospital, 75 Francis St, Amory 2, Boston, MA 02115 (e-mail: semelanson@partners.org). opiates, phencyclidine, and tricyclic antidepressants (TCA) and, more recently, methylenedioxymethamphetamine (MDMA) and oxycodone. Immunoassays are used frequently in the clinical laboratories to perform DAU testing because they are easily automated and provide rapid turnaround times. 2,6 Several different immunoassay techniques and platforms are available. 7 Depending on the assay, an antibody is designed to detect a specific class of compounds (ie, barbiturates), a parent drug (ie, methadone), or a metabolite (ie, benzoylecgonine, a metabolite of cocaine). Qualitative results are based on a specific calibrator concentration. Positive results reflect a concentration greater than the calibrator cutoff, whereas negative results reflect concentrations less than the cutoff and do not exclude the presence of the drug or metabolite. Cutoff concentrations are established through a combination of regulatory recommendations (eg, Substance Abuse and Mental Health Services Administration), analytical considerations (ie, maximum discrimination between signal and noise), and historic convention. To effectively manage patients with suspected substance use or abuse, clinicians should appreciate the limitations of DAU testing, including potential interferences. The antibody specificity varies within the drug class, and despite a single cutoff, each individual drug within the class requires a different urine concentration to trigger a positive result. Certain antibodies may also cross-react with medications or their metabolites outside the target drug class leading to false-positive results. These technical aspects can complicate result interpretation. Arch Pathol Lab Med Vol 134, May 2010 Lessons From Laboratory Drug Testing Surveys Melanson et al 735

Table 1. College of American Pathologists Urine Drug Testing Survey Components and 2008 Participant Volumes Mean No. of Participating Laboratories in 2008, n = 3 survey Drug or Class of Drugs mailings Amphetamines 2844 MDMA 371 Barbiturates 2765 Benzodiazepines 2719 BE/Cocaine 2830 THC 2767 Ethanol 523 Methadone 1232 Methaqualone 110 Opiates 2781 Oxycodone 267 Phencyclidine 2325 Propoxyphene 468 Tricyclics 1121 LSD 14 Abbreviations: BE, benzoylecgonine; LSD, lysergic acid diethylamide; MDMA, methylenedioxymethamphetamine; THC, tetrahydrocannabinol. The College of American Pathologists (CAP) Proficiency Testing Surveys are used by thousands of laboratories throughout the world, but primarily in the United States, and provide information on the utility of DAU testing and assay performance. In this study, we compiled data from CAP surveys, investigated the worldwide characteristics and performance of DAU testing, and described the technical aspects that may affect result interpretation. METHODS Laboratories offering DAU testing are required to perform proficiency testing as one method to demonstrate adequate assay performance and their competency to perform testing. The CAP Urine Drug Testing (Screening) (UDS) Proficiency Survey evaluates a laboratory s ability to qualitatively detect the drugs or classes of drugs listed in Table 1. The survey is sent out 3 times per year (ie, A, B, and C survey mailings) and contains 5 urine specimen challenges with various drugs that are unknown to the laboratory. Laboratories provide results for each challenge (ie, drug present or drug absent) for the tests listed in Table 1 that are performed in their laboratory. Each laboratory s performance is compared with its peers that use the same immunoassay technique (Table 2). The CAP provides an acceptable or unacceptable grade for each challenge. An 80% consensus must be reached within each technique for grading to occur. The CAP summarizes the results from all laboratories, which provide robust data for examining the performance and utility of DAU testing worldwide. The summary includes the concentration of each drug in the unknown sample and the number of laboratories with a present or absent response for each immunoassay technique. We reviewed UDS surveys for the past 6 years (2003 2008) for a total of 18 surveys. The number of participants for each drug or class of drugs was recorded. The number of laboratories reporting positive results for each drug and concentration challenged in the survey was examined, and the performance of each immunoassay technique was compared. When the same concentration of a drug was repeatedly challenged, the mean positivity rate was calculated. In addition, cross-reactivity claims provided by the manufacturer were compared with results in the UDS survey. Generic immunoassay techniques, such as lateral flow and immunochromatography, were not included in the analysis. All results are presented as the mean (with parenthetic lowest value, highest value). The approximate number of laboratories participating in each survey discussed in Results is shown in Table 1. The specific surveys from which the data were derived is listed in parentheses in Results. RESULTS The mean participant volume for the UDS survey in 2008 was 3120. Participant volume was the lowest for lysergic acid diethylamide (LSD), with a mean of 14 laboratories (Table 1). Volumes for MDMA, methaqualone, oxycodone, and propoxyphene were also low because fewer than 500 laboratories were enrolled. However, participant volumes for oxycodone, MDMA, and propoxyphene are increasing, whereas participant volumes for LSD and methaqualone have remained relatively low since 2003. Laboratories in the Midwest represent the highest percentage of participants in the LSD survey (ie, 35.7% [5 of 14] of total versus 21.4% [3 of 14] of total in other areas of the country), whereas laboratories in the South represent the highest percentage of participants in the methaqualone survey (ie, 33.9% [37 of 110] from the South, 24.0% [27 of 110] from the Midwest, 25.6% [28 of 110] from the Northeast, and 16.5% [18 of 110] from the West). Oxycodone, MDMA, and propoxyphene testing were all highest in the South. The percentage of laboratories using amphetamine immunoassays at a cutoff of 1000 ng/ml that obtained a positive result when challenged with 2500 ng/ml of MDMA was 42% (1195 of 2844) (range, 2% 98%) (Table 2; UDS-B 2004, UDS-C 2006). The positivity rate decreased to approximately 3.1% (88 of 2844) in the 750 ng/ml challenge (UDS-A 2005). Less than 15% (46 of 304) of laboratories using enzyme immunoassay (EIA; ie, the enzyme multiplied immunoassay technique [EMIT]) and microparticle immunoassay (ie, kinetic interaction of microparticles in solution) had positive results for MDMA. Most opiate immunoassays, which used a 300 ng/ml cutoff, did not detect 7500 ng/ml of oxycodone (UDS-A 2003; Table 2). A mean of 17% (473 of 2781) (range, 0% 54%) of laboratories reported it was present. When challenged with 1500 ng/ml of oxycodone, approximately 2.5% (70 of 2781) of laboratories using an opiate assay with a 300 ng/ml cutoff reported positive results (UDS-A 2004, UDS-B 2005, UDS-B 2006). In addition, although almost all of the participants using immunoassays with a cutoff of 300 ng/ml detected the opiate hydromorphone at 1000 ng/ml, the many laboratories that used fluorescent immunoassay (FIA) and microparticle immunoassay (kinetic interaction of microparticles in solution) did not (UDS-A 2003) (Table 2). Approximately 2700 laboratories using the benzodiazepine immunoassay with a cutoff of 300 ng/ml were challenged with 375 ng/ml of oxazepam (UDS-B 2005). The mean percentage of laboratories reporting its presence was 42% (1142 of 2719) (range, 1% 98%; Figure 1). Although 98% (74 of 75) of laboratories using EIA (EMIT) obtained positive results, only 1% (3 of 318) of laboratories using FIA detected 375 ng/ml of oxazepam. At challeng- 736 Arch Pathol Lab Med Vol 134, May 2010 Lessons From Laboratory Drug Testing Surveys Melanson et al

Immunoassay Technique Table 2. Percentage of Laboratories Producing Positive Results for Specific Challenges Laboratories Reporting Positive Results, % MDMA, 2500 ng/ml a Lorazepam, 750 ng/ml b Oxycodone, 7500 ng/ml c Hydromorphone, 1000 ng/ml d Amitriptyline, 1300 ng/ml e CEDIA n/a 100 29 100 n/a CMI (Triage) 98 100 10 99 99 EIA (DRI) 35 99 n/a n/a 67 EIA (EMIT) 13 76 54 88 100 FIA 48 100 0 13 99 FPIA 57 n/a 2 97 n/a MIA (KIMS) 2 95 7 19 n/a Abbreviations: CEDIA, cloned enzyme donor immunoassay; CMI, colloidal metal immunoassay; DRI, Diagnostic Reagents Inc. (Sunnyvale, California); EIA, enzyme immunoassay; EMIT, enzyme multiplied immunoassay technique; FIA, fluorescence immunoassay; FPIA, fluorescence polarization immunoassay; KIMS, kinetic interaction of microparticles in solution; MDMA, methylenedioxymethamphetamine; MIA, microparticle immunoassay; n/a, information not available. a Immunoassay category: amphetamine; cutoff: 1000 ng/ml; no. of surveys included: 2. b Immunoassay category: benzodiazepine; cutoff: 300 ng/ml; no. of surveys included: 1. c Immunoassay category: opiate; cutoff: 300 ng/ml; no. of surveys included: 1. d Immunoassay category: opiate; cutoff: 300 ng/ml; no. of surveys included: 1. e Immunoassay category: tricyclics; cutoff: 1000 ng/ml; no. of surveys included: 5. es of 750 ng/ml or greater, all laboratories, except those using FIA, detected oxazepam; 750 ng/ml of lorazepam was also challenged (UDS-C 2007). The percentage of laboratories using EIA (EMIT) with a 300 ng/ml cutoff that reported positive results when challenged with 750 ng/ml of lorazepam was less than 80% (57 of 75), whereas all others detected lorazepam (Table 2). Less than 80% (10 of 13) of laboratories using the TCA EIA (Diagnostics Reagents Inc, Sunnyvale, California [DRI]) at a 1000 ng/ml cutoff produced a positive result when challenged with 1300 ng/ml doxepin (UDS-C 2004, UDS-C 2005, UDS-B 2006, UDS-B 2007). During the past 3 years, CAP has provided 3 challenges with 1300 ng/ml of amitriptyline (UDS-C 2005, UDS-C 2006, and UDS-B 2007). Although the positivity rate for most immunoassays using a cutoff of 1000 ng/ml is close to 100% (1121 of 1121), only 67% (9 of 13) of laboratories using EIA (DRI) detected 1300 ng/ml amitriptyline (Table 2). A mean of 88% (2435 of 2767) (range, 71% 100%) of laboratories using THC assays with a 50 ng/ml cutoff reported positive results when challenged with 50 ng/ml of THC (UDS-C 2003, UDS-A 2006; Figure 2). Many assays detected concentrations less than the cutoff. The mean positivity rate was 25% (692 of 2767) (range, 3% 58%) at concentrations of 35 ng/ml (UDS-C 2004, UDS-C 2005). Similarly, all immunoassays using a 300 ng/ml cutoff detected the cocaine metabolite benzoylecgonine at a concentration of 375 ng/ml (UDS-A 2005, UDS-A 2006), and 13% (46 of 355) of FIA users detected a concentration of 188 ng/ml (UDS-A 2008; Figure 3). The manufacturer provides cross-reactivity data for many drugs in the class as well as structurally related compounds. Table 3 lists the cross-reactivity claims for selected immunoassay techniques and illustrates the wide variation between assays. The cloned enzyme donor immunoassay cross-reacts with lorazepam at concentrations less than the cutoff, whereas the cross-reactivity in other techniques is less than 60%. The cross-reactivity claims for the amphetamine and benzodiazepine assays are consistent with the UDS survey results. However, the claims for oxycodone cross-reactivity in the opiate assay are discrepant. Although 6000 ng/ml and 1500 ng/ml of oxycodone should trigger a positive result in the colloidal metal immunoassay (Triage) and EIA (EMIT) assays, respectively, only 10% (40 of 399) of laboratories using colloidal metal immunoassay (Triage) and 54% (230 of 426) using EIA (EMIT) reported present when challenged with 7500 ng/ml of oxycodone. Alternatively, the microparticle immunoassay (kinetic interaction of microparticles in solution) appeared to be more sensitive to oxycodone than the stated claims. The CAP occasionally challenges with red herrings (ie, substances that may cross-react leading to false-positive results). A 9% (209 of 2325) false-positive rate for phencyclidine was seen when laboratories were challenged with 9000 ng/ml of dextromethorphan (UDS-C 2003); 5000 ng/ml of diphenhydramine led to a 2% (22 of 1121) false-positive rate for TCA (UDS-A 2003). Three challenges with 4500 ng/ml of ephedrine produced less than 2% (57 of 2844) false-positive rates in the amphetamine assays and between 1% (4 of 371) and 12% (45 of 371) false-positive rates in the MDMA assays (UDS-B 2004, UDS-A 2006, UDS-B 2007). DISCUSSION The participant volumes suggest that LSD and methaqualone testing may not be necessary, except in specific geographic locations, such as the South and Midwest, where abuse rates may be higher. A laboratory s DAU panel should reflect abuse rates in that area. Testing for drugs that are infrequently abused can be eliminated, which not only avoids problems with poor cross-reactivity and unnecessary follow-up testing but also increases efficiency. The increase in propoxyphene testing may be due to its use in pain management. Assays specific for oxycodone and MDMA were challenged separately starting in 2005. Because these assays are relatively new, it may explain the lower participant volumes. However, the numbers for oxycodone and MDMA, unlike LSD and methaqualone, are trending upward. The National Institute of Drug Abuse, and the Drug Enforcement Agency, publish data on national drug use and substances secured during law enforcement operations. 8,9 In 2007, MDMA seizure and use was highest in the South, consistent with participant volumes for MDMA testing. However, oxycodone seizure and use was higher in the Northeast than would be suggested by participant volumes. Regional differences in UDS survey participation may not reflect the prevalence Arch Pathol Lab Med Vol 134, May 2010 Lessons From Laboratory Drug Testing Surveys Melanson et al 737

Figure 1. Positivity rate of different immunoassay techniques for oxazepam. The percentage of laboratories reporting positive results with a benzodiazepine assay that used a cutoff of 300 ng/ml is plotted at increasing concentrations of oxazepam. The positivity rate for the cloned enzyme donor immunoassay (diamonds, ), colloidal metal immunoassay (squares, &), enzyme immunoassay (enzyme multiplied immunoassay technique; X-shape, 3), fluorescent immunoassay (asterisk, *), fluorescence polarization immunoassay (circle, N ), and microparticle immunoassay (kinetic interaction of microparticles in solution; plus sign, +) is shown. Figure 2. Positivity rate of different immunoassay techniques for tetrahydrocannabinol (THC). The percentage of laboratories reporting positive results with a THC assay that used a cutoff of 50 ng/ml is plotted at increasing concentrations of THC. The positivity rate for the cloned enzyme donor immunoassay (diamonds, ), colloidal metal immunoassay (Triage; squares, &), enzyme immunoassay (Diagnostics Reagents Inc; triangle, m), enzyme immunoassay (enzyme multiplied immunoassay technique; X-shape, 3), fluorescent immunoassay (asterisk, *), fluorescence polarization immunoassay (circle, N ), microparticle immunoassay (kinetic interaction of microparticles in solution; plus sign, +) is shown. Figure 3. Positivity rate of different immunoassay techniques for benzoylecgonine (BE). The percentage of laboratories reporting positive results with a BE assay that used a cutoff of 300 ng/ml is plotted at increasing concentrations of BE. The positivity rate for the cloned enzyme donor immunoassay (diamonds, ), colloidal metal immunoassay (squares, &), enzyme immunoassay (Diagnostics Reagents Inc; triangle, m), enzyme immunoassay (enzyme multiplied immunoassay technique; X-shape, 3), fluorescent immunoassay (asterisk, *), fluorescence polarization immunoassay (circle, N ), microparticle immunoassay (kinetic interaction of microparticles in solution; plus sign, +) is shown. of drug use, but rather a particular regional preference to offer all available immunoassays. Assays specific for MDMA and oxycodone are necessary because of the low positivity rate and cross-reactivity of most amphetamine and opiate assays for these drugs. Laboratories serving emergency departments or pain management and drug treatment programs that have many patients with MDMA or oxycodone abuse or overdose should consider implementing specific assays in their laboratory. If specific assays are not performed or are not warranted, clinicians should know the concentration of MDMA and oxycodone required to trigger a positive screening result in their laboratory. Oxazepam is a metabolite of several different benzodiazepines and may be present in the urine of many patients taking benzodiazepines, primarily as the conjugated metabolite, oxazepam glucuronide. However, most immunoassays are designed to detect the parent compound, not the principal conjugated metabolite found in urine. In addition, the immunoassays do not exhibit 100% crossreactivity for the parent compound oxazepam at the 300 ng/ml cutoff, and urine tests with concentrations of oxazepam greater than 300 ng/ml may be negative. In clinical specimens, glucuronidase can be used to increase positivity in assays designed to detect the parent compound, not the conjugated metabolite, by removing the glucuronide. Laboratory directors may wish to discuss this option with the medical staff. The low cross-reactivity of the FIA assay in CAP proficiency samples is most likely because it is designed to detect the glucuronide metabolite instead of the parent compound, which is present in CAP challenges. Furthermore, similar to oxazepam, the crossreactivity of some assays for lorazepam is low, and urine specimens from patients taking this medication may be negative. Data from UDS surveys illustrate that the urine concentration of a drug, particularly in the opiate, benzodiazepine, or TCA class, may need to be significantly higher than the calibrator cutoff to trigger a positive result. For example, with some opiate immunoassays that use a 300 ng/ml cutoff, a urine concentration of 300 ng/ ml of hydromorphone will not trigger a positive result. In addition, 1000 ng/ml of doxepin will not produce a positive result in most TCA assays that use a 1000 ng/ml cutoff, and 1000 ng/ml of amitriptyline will not consistently produce a positive result in the EIA (DRI) assay. In contrast to opiates, benzodiazepines, and TCA, the positivity rate may be higher for certain drugs, such that drug concentrations lower than the cutoff will trigger a positive result. This can be seen with THC and cocaine metabolite immunoassays. Results from the CAP surveys illustrate that the actual performance of the immunoassay may not reflect the manufacturer s cross-reactivity claims. This may be due to lot-to-lot variability, differences in population, or lack of extensive validation. The cross-reactivity claims stated by the manufacturer are only estimates, and the claims should be verified by the laboratory with their patient population, if possible. 10 Physicians should understand the positivity rate and cross-reactivity of their laboratory s screening immunoassays, realize that the assays may not detect all drugs in a class, and discuss with the laboratory director the most suitable assays to use based on their patient population. If a highly abused drug is not detected well with the current 738 Arch Pathol Lab Med Vol 134, May 2010 Lessons From Laboratory Drug Testing Surveys Melanson et al

Immunoassay Technique Table 3. Manufacturer Cross-Reactivity Claims Cross-Reactivity, % (concentration, ng/ml) Amphetamine Assay for MDMA Benzodiazepine Assay for Lorazepam Opiate Assay for Oxycodone CEDIA 2 (50 000) 122 (246) 3 (10 000) CMI (Triage) 50 (2000) 55 (550) 5 (6000) EIA (EMIT) 5 (20 000) 33 (909) 20 (1500) FPIA 3 (3000) 20 (1500) 3 (10 000) MIA (KIMS),1 (.100 000) 44 (680),1 (.30 000) Abbreviations: CEDIA, cloned enzyme donor immunoassay; CMI, colloidal metal immunoassay; EIA, enzyme immunoassay; EMIT, enzyme multiplied immunoassay technique; FPIA, fluorescence polarization immunoassay; KIMS, kinetic interaction of microparticles in solution; MDMA, methylenedioxymethamphetamine; MIA, microparticle immunoassay. screen, the laboratory may need to change assays. Although the turnaround time is not ideal, physicians can always request that specimens be sent for definitive confirmatory testing, if the result is in question. For THC and benzoylecgonine assays, clinicians should realize that the actual urine drug concentration may be lower than the cutoff concentration. In addition, positive results may not require treatment or cause the presenting signs and symptoms. The number of false-positive results caused by dextromethorphan in the phencyclidine assay, diphenhydramine in the TCA assay, and ephedrine in the amphetamine and MDMA assays was low. 11 13 However, CAP challenged with the parent compound, not the metabolites, and metabolites are most likely for the cause of the cross-reactivity. Clinicians should be aware that falsepositive results can occur, and they should know the likely culprits so they can manage their patients care accordingly. The clinical consequences of misinterpretation of DAU testing can be severe and lead not only to underdiagnoses but also to overdiagnoses. Laboratories should communicate with clinicians and have input into the test menu and test performance. To aid with result interpretation, the laboratory should consider (1) adding a footnote outlining important cross-reactivities and interferents, (2) using a specific toxicology requisition with more detailed description of the assays, or (3) summarizing test performance in a memorandum or table to be circulated or posted on an intranet test menu available for all clinicians. References 1. Department of Health and Human Services. Drug Abuse Warning Network Web site. http://www.dawninfo.samhsa.gov. Accessed May 28, 2009. 2. Hammett-Stabler CA, Pesce AJ, Cannon DJ. Urine drug screening in the medical setting. Clin Chim Acta. 2002;315(1 2):125 135. 3. Wu AHB, McKay 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(3):357 379. 4. Trescot AM, Boswell MV, Atluri SL et al. Opioid guidelines in the management of chronic non-cancer pain. Pain Physician. 2006;9(1):1 39. 5. Heit HA, Gourlay DL. Urine drug testing in pain medicine. J Pain Symptom Manage. 2004;27(3):260 267. 6. Colbert DL. Drug abuse screening with immunoassays: unexpected crossreactivity and other pitfalls. Br J Biomed Sci. 1994;51(2):136 146. 7. Magnani B. Concentrations of compounds that produce positive results. In: Shaw LM, Kwong TC, Rosano TG, Orsulak PJ, Wolf BA, Magnani B, eds. The Clinical Toxicology Laboratory: Contemporary Practice of Poisoning Evaluation. Washington, DC: AACC Press; 2001:481 498. 8. National Institute of Drug Abuse. The Advance Report: Epidemiological Trends in Drug Abuse. http://www.drugabuse.gov/pdf/cewg/advreport108. pdf. Accessed May 28, 2009. 9. Drug Enforcement Administration. National Forensic Laboratory Information System Midyear Report. http://www.deadiversion.usdoj.gov/nflis/2007_ midyear.pdf. Accessed May 28, 2009. 10. Miller JJ, Valdes R. Methods for calculating crossreactivity in immunoassays. J Clin Immunoassay. 1992;15(8):97 107. 11. Marchei E, Pellegrini M, Pichini S, Martin I, Garcia-Algar O, Vall O. Are false-positive phencyclidine immunoassay instant-view multi-test results caused by overdose concentrations of ibuprofen, metamizol, and dextromethorphan? Ther Drug Monit. 2007;29(5):673 671. 12. Asselin WM, Leslie JM. Use of the EMITtox serum tricyclic antidepressant assay for the analysis of urine samples. J Anal Toxicol. 1990;14(3):168 171. 13. Stout PR, Klette KL, Horn CK. Evaluation of ephedrine, pseudoephedrine and phenylpropanolamine concentrations in human urine samples and a comparison of the specificity of DRI amphetamines and Abuscreen online (KIMS) amphetamines screening immunoassays. J Forensic Sci. 2004;49(1):160 164. Arch Pathol Lab Med Vol 134, May 2010 Lessons From Laboratory Drug Testing Surveys Melanson et al 739