Validity Testing of the EZ-SCREEN Cannabinoid Test

Transcription

1 Validity Testing of the EZ-SCREEN Cannabinoid Test Amanda J. Jenkins, Lorrie C. Mills, William D. Darwin, Marilyn A. Huestis, and Edward J. Cone* Addiction Research Center, NIDA, Baltimore, MD John M. Mitchell Navy Drug Screening Laboratory, H2033, Naval Air Station, Jacksonville, FL Abstract Recently, a number of "quick tests" became available for use in on-site drug testing. These tests offer advantages In simplicity, ease of performance, and rapid access to test results. However, there is a paucity of data on the validity of these tests for the detection of drugs of abuse. This report describes a validity study of the EZ-SCREEN cannabinoid test for the detection of cannablnoids in urine. Three healthy, male volunteers with a history of marijuana use participated in the study. Each subject smoked 1, 2, or 4 marijuana cigarettes (2.6% THC) on each test day. Urine samples were collected and incorporated into a specimen set consisting of 178 clinical urine samples, 72 urine samples containing known amounts of drug, and 50 drug-free urine samples. The specimen set was randomized and analyzed under blind conditions by the EZ-SCREEN test and by GC/MS for 11-nor-9- carboxy-/o-tetrahydrocannabinol (THCCOOH). Results were interpreted independently by three readers. Concordance analysis was performed by comparison of results of the EZ- SCREEN test with GC/MS. The EZ-SCREEN test was highly sensitive and produced positive results at a standard THCCOOH concentration of 5 ng/ml. While showing high sensitivity to THCCOOH, the assay demonstrated low crossreactivity with A9-tetrahydrocannabinol (THC) and other cannabinoids. No false-positive results were recorded with 50 drug-free urine samples, but one reader recorded eight undecided results. Overall agreement between the three readers for the EZ-Screen results was approximately 80%. Delayed readings and photocopy readings tended to be less accurate than readings obtained at 3 min. Overall, the EZ- SCREEN test was easy to perform and provided rapid turnaround of results. Because the assay appears to be highly sensitive to THCCOOH in urine, positive results may not be confirmed by GC/MS at a cutoff of 15 ng/ml. Introduction Drug abuse has been cited as a major problem in American industry (1). To combat this problem, many companies have begun *Address correspondence to Dr. Edward J. Cone, ARC, NIDA, P.O. Box 5180, Baltimore, MD testing employees for drugs of abuse (2). An American Management Association (AMA) survey of 1200 United States companies indicated that employee drug testing increased from 51.5% of companies surveyed in January, 1990, to 63% in January, 1991, and to 74.5% in January, 1992 (3). Recent data from the AMA indicates that employee drug testing is performed by 73.2% of manufacturing firms, 63.2% of health care employers, and 20.9% of financial services employers (3). Utility companies (91%) and transportation companies (81%) also have testing programs (4). Testing is conducted by the administrative offices of the United States courts for people who are on probation or parole, by the Federal Bureau of Prisons for prisoners, and by the Federal Aviation Administration for employees undergoing physical examinations (5). Urine test results are used to determine employment eligibility, to provide counseling, and to imprison or otherwise punish offenders (6). Most drug testing programs analyze urine specimens for the "NIDA five" drugs of abuse: cannabinoids, cocaine metabolite, opiates, amphetamines, and phencyclidine. According to the Health and Human Services Mandatory Guidelines for Federal Employees (7), the initial screening tests must be performed by immunoassay. All presumptive positive results must be confirmed by gas chromatography/mass spectrometry (GC/MS). The turnaround time and cost of forensic urine drug testing (FUDT) programs are important factors to consider when implementing a drug testing program. An on-site testing program provides an alternative to the standard drug testing program in which specimens are shipped to commercial laboratories for screening and confirmation. On-site analysis of urine samples has the obvious advantage of rapid access to drug test results; however, concerns have been expressed over the possible misuse of results and the potential inaccuracy of on-site testing methods (8). On-site testing can be implemented by use of immunoassays similar to those employed in commercial urine drug testing laboratories. Alternately, new immunoassay-based "quick tests" are available which offer advantages of simplicity, ease of operation, and self-containment (i.e., instrumentation may not be required). Also, there is no requirement for laboratory facilities with quick tests; and product literature indicates that tests can be performed by individuals with little analytical chemistry experience. Presently, these quick tests are used by the criminal justice system (9), hospital emergency rooms, and drug treatment centers for detection of drugs of abuse in urine. 292 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.

2 The efficacy of quick tests depends upon their "validity" in the detection of drugs of abuse. The term "validity" has been defined as the ability of an assay to detect a drug or its metabolites in biological fluids tbllowing human drug administration (10). This definition encompasses a variety of parameters including the following: sensitivity; specificity; metabolic and pharmacologic variables such as dose, route of administration, concentration, and biological fluid ph; and intersubject variability in absorption, metabolism, and excretion (11). The concept of assay validity also has been expanded to require confirmation of all presumptive positive results by an altemate analytical method such as GC/MS (12). In FUDT analysis, the latter is often a requirement when the analytical data must sustain legal scrutiny. Although there has been increasing use by employers of on-site testing for the detection of drugs of abuse in urine, there is a paucity of information on the validity of quick tests. This report describes a validity study of the EZ-SCREEN cannabinoid test for the detection of cannabinoids in urine. The evaluation was performed with clinical urine specimens collected under controlled conditions from subjects who smoked marijuana. Standards containing known concentrations of cannabinoid metabolites and drug-free urine samples were also included in the analysis. Concordance analysis was performed by comparison of the results of the EZ-SCREEN test to GC/MS assay. Experimental Drugs and chemlcals Marijuana cigarettes (2.6% A9-tetrahydrocannabinol [THC]), placebo marijuana cigarettes (0% THC), THC, 11-nor-9-carboxy-A9-tetrahydrocannabinol (THCCOOH), 1 l-hydroxy-thc (11-OH-THC), cannabidiol, cannabinol, and cannabigerol standards were obtained from the National Institute on Drug Abuse (Rockville, MD). All organic solvents were HPLC grade and chemicals were reagent grade. Subjects, dosing, and specimen collection Three healthy male volunteers between the ages of 21 and 55 years with a history of marijuana use participated in the study. The subjects were marijuana-free for a minimum of three days as evidenced by urinalysis (EMIT d.a.u. TM assay, 100 ng/ml cutoff, Syva Co.) prior to the study. On the test day, each subject smoked two marijuana cigarettes between 0900 and 0930 hours and two cigarettes between 1300 and 1330 hours. The subjects had the option of omitting the fourth marijuana cigarette. Active (MJ) and placebo (P) marijuana cigarettes were smoked during each test session to provide the following combinations of doses: E E R P; MJ, P, P, P; MJ, P, MJ, P; MJ, MJ; and MJ, MJ, MJ, MJ. All urine samples were collected throughout the study. After collection, samples were combined into individual 12-h pools. Samples were frozen immediately after processing. THCCOOH standards were prepared in drug-free urine at the following concentrations: 5, 10, 15, 20, 50, 100, 250, 500, and 2000 ng/ml. The remaining cannabinoid standards were prepared in a similar manner in concentrations of 25, 100, and 500 ng/ml. All standards were assayed by EZ-SCREEN and GC/MS in triplicate. A specimen set consisted of 178 clinical urine samples, 72 urine drug standards, and 50 drug-free urine samples. The drugfree urine samples were collected from 16 donors (eight males and eight females). All drug-free donor specimens were validated as negative for cannabinoids by the TDx cannabinoid urine assay (Abbott Laboratories Inc.) at a detection level of 25 ng/ml. All samples were coded and randomized before analysis. EZ-SCREEN test The test was performed on freshly thawed samples which were allowed to reach room temperature prior to analysis. The specimens were mixed, and particulate matter was allowed to settle prior to testing. The EZ-SCREEN test was performed according to the manufacturer's procedure (13). Briefly, the procedure consisted of the addition of sample to the sample port of the test card. Positive and negative controls were added to their respective ports on the same test card. Enzyme conjugate, a wash reagent, and substrate were added to each port, and results were read after 3 rain. The negative control exhibited a grey-blue color. The positive control exhibited a lighter color than the negative control. The sample was considered positive () for cannabinoid metabolites when the color of the sample was equal to or lighter than the adjacent positive control (13). The sample was considered negative (-) for cannabinoid metabolites when the color of the sample was darker than the adjacent positive control. An undecided reading () was recorded when color interpretation was considered borderline between positive and negative. An invalid result was recorded if the negative and positive controls developed the same degree of color or if both controls failed to develop any color. Three readers determined the resuits independently. The cards were read at 3 min (according to the test procedure) and again at 60 min. Immediately after the visual reading at 3 min, the cards were photocopied. The photocopied records were read under blind conditions, independent of influence from original results. GC/MS assay The measurement of THCCOOH in urine by GC/MS was performed as previously described (14). The limit of quantitation and detection for this assay was 0.8 ng/ml. The upper limit of linearity was 1000 ng/ml. Results Cllnlcal assay results and concordance analysls Three hundred urine samples, consisting of clinical specimens from marijuana users, control urine from drug-free subjects, and urine drug standards, were analyzed by the EZ-SCREEN test and by GC/MS assay in randomized order under blind conditions. The results for three marijuana smokers are tabulated in chronological order in Table I. On different days, Subjects A and C smoked either placebo, one, two, or four marijuana cigarettes; Subject B smoked one marijuana cigarette one day and two marijuana cigarettes on two others, but did not participate in higher dose experiments. Prior to their participation in the study, the three subjects were required to test negative for cannabinoids by Emit d.a.u. (100 ng/ml cutoff) for three consecutive days. Specimens collected just prior to the subjects' participation indicated the presence of 2.2 ng/ml of THCCOOH for Subject A, whereas specimens collected from Subjects B and C were negative by GC/MS. Prior to the next drug administration, subjects were again required to test negative for three consecutive days. Preadministration urine samples contained THCCOOH in concentrations ranging from 0.5 ng/ml to 6.4 ng/ml. Both negative and 293

3 Table I. QualitativeDetection of Cannabinoidswith the EZ-SCREENTest and GC/MS In Urine Specimens from Marijuana Smokers Subject: Dose* End of collection GC/MS, EZ-SCREENt interval THCCOOH Reader (h) (ng/ml) End of collection GC/MS, EZ-SCREENt Subject: interval THCCOOH Reader Dose* (h) (ng/ml) 1 2 Subject A: 1 MJ Cigarette Subject A: 2 MJ Cigarettes I _ 3.6 _ 4.3 _ _ * MJ = marijuana (2.6% THC); P = placebo marijuana cigarette (0% THC). Individual doses are listed in chronological order. t = Positive reading; - = negative reading; _ = undecided reading. Subject A: _ Placebo r Subject A: MJ Cigarettes _ Subject B: MJ Cigarette _ Subject B: MJ Cigarettes (M J, R M J, P) _

4 positive results were obtained by EZ-SCREEN. The negative results generally were associated with samples that tested in the 0-2 ng/ml range for THCCOOH by GC/MS, whereas undecided and positive results were associated with samples that contained greater than 2 ng/ml. Following marijuana smoking, both EZ-SCREEN and GC/MS results were highly positive. Samples continued to test positive by EZ-SCREEN until THCCOOH concentrations declined below 10 ng/ml. Below this level, EZ-SCREEN results consisted of positive, negative, and undecided readings. Negative EZ- SCREEN results were generally associated with THCCOOH concentrations less than 5.0 ng/ml. Only one specimen collected after marijuana smoking (Table I, Subject B, 1 MJ cigarette, 144 hrs, GC/MS value of 1.1 ng/ml) tested negative by EZ-SCREEN by all three readers. Concordance analysis of EZ-SCREEN results versus GC/MS analysis for THCCOOH with the 178 clinical specimens is shown in Table II. There was a mean of 64 positive readings by EZ-SCREEN, with specimens containing THCCOOH at a concentration equal to or greater than 15 ng/ml. In addition, a mean of 41 readings were positive with a concentration in the ng/ml range, and a mean of 12 readings were positive with concentrations less than 5 ng/ml. All negative EZ-SCREEN results were associated with specimens containing THCCOOH in concentrations less than 15 ng/ml. A mean of 49 readings were undecided. The majority of undecided interpretations occurred Table II. Concordance Analysis of Results by the EZ-SCREEN Test and GC/MS Assay of Clinical Specimens from Marijuana Smokers GC/MS, THCCOOH (ng/ml) < >15 Reader Reader Reader EZ-SCREEN(N=178) Positive Negative Undecided Table I (continued). Qualitative Detection of Cannabinoids with the EZ-SCREEN Test and GC/MS In Urine Specimens from Marijuana Smokers End of collection GC/MS, Subject: interval THCCOOH Dose* (h) (ng/ml) EZ-SCREENt Reader End of collection GC/MS, EZ-SCREENt Subject: interval THCCOOH Reader Dose* (h) (ng/ml) Subject 8: 2 MJ Cigarettes (M J, M J) Subject C: 4 MJ Cigarettes Subject C: 1 MJ Cigarette "4" Subject C: MJ Cigarette (continued) Subject C: MJ Cigarettes Subject C: Placebo * MJ = marijuana (2.6% THC); P = placebo marijuana cigarette (0% THC). Individual doses are listed in chronological order. t = Positive reading; - = negative reading; = undecided reading. 295

5 Table III. Analysis of Cannabinoid Standards and Drug-Free Urine Samples by EZ-SCREEN Test and GC/MS GC/MS, EZ-SCREEN* Concentration THCCOOH Reader Standard (ng/ml) (ng/ml) Standard GC/MS, EZ-SCREEN~ Concentration THCCOOH Reader (ng/ml) (ng/ml) THCCOOH 11-OH-THC THC Cannabidiol Jr _ 25 0 Jr 0 -- Jr _ Jr Jr _ _ Jr Cannabigerol _ ~ ~ ControISubjectA Control Subject B ControISubject C Control Subject D Control Subject E Control Subject F ~ Jr -- Control Subject G Control Subject H Jr - Control Subject I Control Subject J Control Subject K Control Subject L Control Subject M Cannabinol _ _ _ - Control Subject N Control Subject Control Subject P Jr -- 9 = Positive reading; - = negative reading; _ = undecided reading

6 with specimens at concentrations less than 5 ng/ml. Only 0-11% of the undecided results were associated with specimens that contained THCCOOH in concentrations in excess of 15 ng/ml. No invalid results were recorded at 3 min. Sensitivity and specificity Standards and drug-free urine samples were assayed in triplicate at varying concentrations to assess the sensitivity and specificity of the EZ-SCREEN test (Table III). All three readers detected THCCOOH at concentrations as low as 5 ng/ml. Some undecided results were recorded for THCCOOH at concentrations as high as 500 ng/ml. Positive results were noted with l l-oh-thc and THC at concentrations greater than or equal to 100 ng/ml. Only negative and undecided interpretations were recorded for cannabinol, cannabidiol, and cannabigerol, with the exception of a single positive reading for cannabidiol at 500 ng/ml and two positive readings for cannabinol at 500 ng/ml. These three readings were obtained from Reader 3. Approximate concentrations of cannabinoids which produced a positive result in the EZ-SCREEN test are summarized in Table IV. Drug-free control urine samples were negative for THCCOOH by GC/MS (Table III). EZ-SCREEN results of these samples were consistently negative by Readers 1 and 3; Reader 2's results consisted of both negative and undecided readings. Reader concordance and methodological comparisons Concordance in EZ-SCREEN results was 79% between Readers 1 and 2 and 78% between Readers 1 and 3. Reader 2 agreed with Reader 3 on 84% of occasions (Table V). Overall agreement between the three readers was approximately 80%. Concordance between readings (Reader 1) taken at 3 min (as specified by the manufacturer) and at 60 min after the EZ- SCREEN test are shown in Table VI. Three invalid results were Table IV. Approximate Concentrations of Cannabinoids in Urine that Produce a Positive Result with the EZ-SCREEN Test Cannabin0ids Concentration (ng/ml) THCCOOH 5 11-OH-THe 100 THe 100 Cannabinol 500 Cannabidiol 500 Cannabigerol >500 Table Vo Concordance Between Readers of the EZ- SCREEN Test (N = 178) with Clinical Specimens from Marijuana Smokers Positive result Negative result Undecided Test Reader Reader Reader result Reader Positive obtained at 60 min, whereas none were obtained at the 3-min reading. The results recorded at 60 min were in agreement with the 3-min results 71% of the time. The discordant results were primarily associated with positive to undecided changes and undecided to negative changes, when comparing the 3-min readings to the 60-min readings. Table VI also lists results from readings of photocopied results of the EZ-SCREEN test. The results were photocopied immediately after the 3-min reading and were read by the readers at a later time, independent of their original results. The results from the photocopied readings were in agreement with the 3-rain results 75% of the time. The discordant results were primarily associated with positive to undecided changes and undecided to negative changes, when comparing the 3-min readings to the photocopied readings. Discussion The performance of the EZ-SCREEN cannabinoid test was evaluated with clinical specimens obtained from marijuana smokers, drug-free specimens, and urine drug standards. Results were compared to those obtained by GC/MS assay for THCCOOH. Generally, the EZ-SCREEN test was simple to perform and provided rapid results. The assay was highly sensitive and produced positive results at a standard THCCOOH concentration of 5 ng/ml, the lowest concentration tested. Additionally, 11 clinical specimens collected from human subjects after smoking marijuana tested positive by EZ-SCREEN with concentrations less than 5 ng/ml as determined by GC/MS. These findings indicated that the detection limit of the EZ-SCREEN test is substantially lower than the detection average of 41 ng/ml specified by the manufacturer (13). Indeed, the detection limit is likely to be in the range of 2-5 ng/ml THCCOOH. In addition to its sensitivity, the EZ-SCREEN test was highly selective and accurate. While showing high sensitivity to THC- COOH, the assay demonstrated relatively low cross-reactivity with THC and other cannabinoids. When the assay was challenged with 50 drug-free urine samples collected from eight male and eight female subjects, no false-positives results were recorded; one reader, however, recorded eight undecided results. The absence of false-positive results in these samples, that is, positive EZ-SCREEN test results for specimens that contained THC- COOH at concentrations less than 5.0 ng/ml, added additional confidence to the validity of the EZ-SCREEN test. The advantages and disadvantages of on-site testing have been reviewed (8,15). Cone (8) indicated that when balancing the cost Table VI. Concordance of EZ-SCREEN Test Results (N = 178) for Clinical Marijuana Specimens at 3-Min Readings Versus 60-Min Readings and Versus Photocopied Results (for Reader 1) 6O-min results* Photocopied results* 3-min results Pos Neg Undecided Invalid Pos Neg Undecided Positive Negative Negative Undecided Undecided * Pos = positive; neg = negative. 297

7 and expediency of an on-site drug testing program with testing at a commercial forensic urine drug testing (FUDT) laboratory, it is important to consider the quality of on-site testing results. On-site tests are not drag-specific and all presumptive positives must be confirmed, preferably by GC/MS. On-site tests may be more or less sensitive than the tests used in FUDT laboratories. If the assay is more sensitive, as appears to be the case for the EZ- SCREEN test for cannabinoids, some presumptive positive results would not be confirmed by GC/MS. Consequently, presumptive positive results from quick tests should not be used, prior to confirmation by GC/MS, in decision-making processes regarding an individual's employment or liberty. In summary, few evaluations have been conducted on quicktest assays targeted for the on-site drug testing market. Because several companies are now manufacturing and marketing on-site drug testing kits, it is critical that validity evaluations be performed as these new products appear and are used for FUDT purposes. References 1. R. Lewy. Preemployment qualitative urine toxicology screening. J. Occup. Med. 25 (8)" (1983). 2. H. Arkin, R.L. DuPont, and L.B. Tepper. Preemployment drug screens. J. Am. Med. Assoc. 255: (1986) AMA Survey on workplace drug testing and drug abuse policies. Summary of key findings. American Medical Association Research Report, New York, NY, R. Eisner. Drug-of-Abuse. Diagnostics and Clinical Testing 27:20-23 (1989). 5. R.E. Willette. Drug testing programs. In Urine Testing fordrugs of Abuse, NIDA Research Monograph 73. R.L. Hawks and C.N. Chiang, Eds., U.S. Government Printing Office, Washington, DC, 1986, pp A.J. McBay. Problems in testing for abused drugs. J. Am. Med. Assoc. 255:39-40 (1986). 7. Department of Health and Human Services. Mandatory guidelines for federal workplace drug testing programs; Final guidelines; Notice. Fed. Regist. 53 (69): (April 11, 1988). 8. E.J. Cone. On-site drug testing. Expediency versus accuracy? Employment Testing, BWR 4(12): (1990). 9. On-site drug testing kits start moving into corporate market. Drug Detection Report 1(5): 1-6 (1991 ). 10. C.W. Gorodetzky. Validity of urine tests in monitoring drug abuse. In Report of the Thirty-Fourth Annual Scientific Meeting Committee on Problems of Drug Dependence, Ann Arbor, MI, 1972, pp C.W. Gorodetzky. Detection of drugs of abuse in biological fluids. In Handbook of Experimental Pharmacology. G.V.R. Born, O. Eichler, A. Farah, H. Herken, and A.D. Welch, Eds., Springer- Verlag, Berlin, Germany, 1977, pp E.J. Cone, S.L. Menchen, and J. Mitchell. Validity testing of the TDx cocaine metabolite assay with human specimens obtained after intravenous cocaine administration. Forens. Sci. Int. 37: (1988). 13. EZ-SCREEN CANNABINOID/COCAINE package insert., Rev. 12/88, Environmental Diagnostics, Inc., Burlington, NC. 14. B.D. Paul, L.D. Mell, Jr., J.M. Mitchell, R.M. McKinley, and J. Irving. Detection and quantitation of urinary 11-nor-Aa-tetrahydrocannabinol-9-carboxylic acid, a metabolite of tetrahydrocannabinol, by capillary gas chromatography and electron impact mass fragmentography. J. Anal ToxicoL 11:1-5 (1987). 15. M.R. Harkey and G.L. Henderson. Quick tests for drugs of abuse: A critical review of the technology. Report commissioned by the California Department of Alcohol and Drug Programs under contract no. D , Sacramento, CA, Manuscript received August 14, 1992; revision received January 20,