Technical Note Evaluation of the Integrated E-Z Split Key Cup II for Rapid Detection of Twelve Drug Classes in Urine Dina N. Greene, Christopher M. Lehman, and Gwendolyn A. McMillin* University of Utah Health Sciences Center, Department of Pathology, Salt Lake City, Utah Abstract The availability of point-of-care (POC) medical devices for drug testing has surged. Reduction in turnaround time, and hence, rapid results are attractive, particularly to acute care facilities, rehabilitation facilities and specialized clinics such as pain management clinics. Here we describe our validation results for the Integrated E-Z Split Key Cup II, a low-cost, rapid urine test that utilizes competitive immunoassay technology to detect 12 drugs or drug classes of commonly abused drugs. Positivity is based on the absence of a colored band at a labeled portion of the detection strip; a negative result produces a distinct, colored band. Using reagent-grade standards, the apparent cut-off for each of the drugs was challenged. The stability of the results was monitored over time. Five urine samples known to be negative for all drug categories and 24 patient samples confirmed positive for a total of 95 drugs and/or drug metabolites claimed to be detected by the device were tested. Adulterants and potential cross-reacting compounds were also evaluated. One false-positive result for benzodizepines was observed. One false-negative result for barbiturates was observed, but was resolved. Overall, the cups demonstrated excellent sensitivity, specificity, and diagnostic efficiency for all drugs represented. Introduction Point-of-care (POC) testing provides a means for health practitioners to receive accurate and timely laboratory test results at the bedside. By definition, a POC test is performed near the patient, outside of a clinical laboratory, and therefore has the potential to make results available within minutes of specimen collection. Many POC testing methods are subject to minimal regulatory requirements, and can be performed with specimens such as urine, capillary blood, or oral fluid, that do not require phlebotomy. POC testing may improve patient counseling * Author to whom correspondence should be addressed: Department of Pathology, University of Utah School of Medicine, ARUP Laboratories, 500 Chipeta Way, Salt Lake City, UT 84108. Email: gwen.mcmillin@aruplab.com. and management because results are available during a patient visit. Successful examples of POC testing used routinely include glucose and prothrombin time monitoring to guide glycemic and anti-coagulation management. Advances in technology have increased the availability of POC devices, and the potential clinical benefits are attractive; however, to be effective, POC must be easy to use, cost effective, and meet accepted quality standards. Recent work has shown that many POC devices fail to meet quality standards, emphasizing the need for stringent validation criteria (1). Inappropriate drug use and drug abuse remains a common problem clinically. It is well recognized that self-reporting of drug use is of limited accuracy and utility. For example, when screening for prenatal exposure to methamphetamine, a notable percent of women that denied abuse gave birth to infants who tested positive (2). Further, a large review of national surveys showed that although some people will self-report drug history accurately, external factors such as pregnancy and need for approval increase the rate of inaccurate reporting (3). Because patient management decisions regarding drug use and abuse depend on accurate drug histories, detection of drugs using POC testing may benefit outpatient drug rehabilitation facilities, whether used to monitor abstinence, verify compliance with prescribed therapy, or identify inappropriate drug use. Specifically, a counselor may better manage the patient session having real-time drug screen results available rather than waiting for results from a test that was sent to a central laboratory and may require several days to generate results. Here, we evaluate the performance of the Integrated E-Z Split Key Cup II, a user-friendly and cost-effective POC device that detects 12 commonly abused drugs (amphetamine, barbiturates, benzodiazepines, buprenorphine, cocaine, marijuana, methadone, methamphetamine, MDMA, opiates, oxycodone, and propoxyphene). Using a validation procedure consistent with CLIA guidelines, we evaluated the ability of the drug screen cup to detect purified reagent standards for each drug class at concentrations above and below the cutoff; the stability of the results over time; the ability of the cup to accurately identify multiple drugs in patient urine; and the effect of potential cross-reacting, interfering compounds, and common adulter- 46 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher s permission.
ants on the results. To the best of our knowledge, this is the only published validation for a POC device that detects 12 drug classes. Materials and Methods Table I. Limit of Detection Measured for Each Drug Class Published Confirmed Drug Reference Cutoff Concentration LOD Class Standard (ng/ml) (ng/ml) Amphetamine d,l-amphetamine 3000 1250 Amphetamine d-amphetamine 1000 420* Barbiturates Secobarbital 300 375 Benzodiazepines Oxazepam 300 450 Buprenorphine Buprenorphine 10 10 Cocaine Benzoylecognine 150 150 Marijuana 11-nor- 9 -THC-9-COOH 50 62.5 Methadone Methadone 300 450 Methamphetamine d,l-methamphetamine 4500 1250 Methamphetamine d-methamphetamine 1000 300* MDMA MDMA 500 375 Opiate Morphine 300 375 Oxycodone Oxycodone 100 100 Propoxyphene Propoxyphene 300 300 * These values were calculated using measured cut offs for the racemic mixture and the published cut offs for d,l and d isoforms (amphetamines) or d and l isoforms (methamphetamine). 3,4-Methylenedioxymethamphetamine. POC device The Integrated E-Z Split Key Cup II (Lot DOA9100739) was purchased from Rapid Detect (Poteau, OK). The manufacturer s published cutoffs for the 12 drug classes detected by the device are listed in Table I. The drugs are detected based on competitive immunoassays. There are six test strips/cup; each strip detects specific drugs. Each strip also has its own colored control band. Urine is distributed into the cup. The key is removed from the cap and pushed into the siding of the apparatus. This unlocks the urine and allows it to wick the membrane, traveling to the top of the membrane via capillary action. Embedded in the membrane are drug-protein conjugates; each test strip has corresponding anti-drug mouse monoclonal antibodies that will travel with the urine. The anti-drug antibodies are labeled with a colored conjugate. Therefore, if there is drug in the urine, the drug will sequester the antibody, and no band will appear at the specific area of the membrane containing the drug-protein conjugate. In contrast, if the patient is negative for a drug, the anti-drug antibody will migrate to and bind the embedded drug, which is visualized by a colored band. More specifically, if urine is positive for a drug, the band is absent; if urine is negative for a drug, the band is present (Figure 1). The control band on each strip is used to detect proper volume of specimen, membrane wicking, and correct procedural technique. The control band does not specifically detect the presence of urine. Embedded in the membrane at the control position is rabbit IgG. Disbursed with the anti-drug mouse monoclonal antibodies are color conjugated goat polyclonal anti-rabbit IgG. Hence, any liquid that properly moistens the membrane will result in the appearance of a strong band in the control region of the strip. Limit of detection and time course study The cutoff for each drug was challenged using Cerilliant purified reference standards (Round Rock, TX). The standard specific to the drug class was the calibrator published in the product package insert. The standards were run using 0%, 25%, 50%, 75%, 100%, 125%, and, if needed, 150% of the apparent drug-specific cutoff. We used five separate cups for each percent tested and defined our limit of detection to be the concentration that gave a positive drug result in all five cups after 5 min, consistent with manufacturer s instructions. To evaluate the stability of test results, the cups were re-read after 30, 60, 120, and 1440 min. Figure 1. Images generated by photocopying the Integrated E-Z Split Key Cup II after screening urine positive for methamphetamine, amphetamine, and propoxyphene. Positive control is indicated by the C ; drug tested indicated either by number or T. Patient urine samples Patient urine previously confirmed positive or negative by validated gas chromatography mass spectrometry (GC MS) or liquid chromatography (LC) tandem MS assays for drugs of abuse were pooled to create 24 samples containing various amounts of the drugs of interest (2 5 drugs/urine pool, see Table II for a detailed description of sample composition) and 5 negative samples. If necessary, the pools were diluted with blank urine. The 24 drug-positive pools and 5 drug-negative urines were blinded before analysis. Thirty-five milliliters of the 47
Table II. Detailed Description of the 24 Patient Samples Screened Using the EZ Cup Drug(s) Concentration Positive Using Sample Present (ng/ml) E-Z Cup 1 THC 160 THC Morphine 420 Opioids Oxycodone 170 Oxycodone Amphetamine 170 Amphetamine Methamphetamine 4700 Methamphetamine Methadone 460 Methadone 2 THC 95 THC Benzoylecognine 1220 Cocaine Methamphetamine 2000 Methamphetamine Amphetamine 360 Amphetamine 3 THC 65 THC Lorazepam 150 Benzodiazepines Alprozolam 30 α-hydroxyalprazolam 130 7-Aminoclonazepam 150 Amphetamine 3680 Amphetamine Methamphetamine 2600 Methamphetamine 4 THC 210 THC Nordiazepam 830 Benzodiazepines Oxazepam 1660 Temazepam 1920 Lorazepam 470 Morphine 350 Opioids Oxycodone 550 Oxycodone Norbuprenorphine 13 Buprenorphine 5 THC 60 THC MDMA 410 MDMA Oxycodone 135 Oxycodone Benzoylecognine 280 Cocaine 6 THC 55 THC MDMA 600 MDMA Nordiazepam 450 Benzodiazepines Oxazepam 640 Temazapam 340 Alprazolam 50 Methadone 1600 Methadone Propoxyphene 370 Propoxyphene 7 α-hydroxyalprazolam 180 Benzodiazepines 7-Aminoclonazepam 40 Alprazolam 50 Nordiazepam 280 Oxazepam 1100 Temazepam 880 Amphetamine 4690 Amphetamine Oxycodone 2800 Oxycodone Morphine 3000 Opioids Propoxyphene 310 Propoxyphene 8 Alprazolam 370 Benzodiazepines Amphetamine 2000 Amphetamine urine in question was allowed to react in the drug screen cup for 5 min before results were read. If there was a false-positive or false-negative result, the urine was poured into a fresh cup to assess reproducibility. If a specimen was positive for a drug that was unexpected, the drug identity was confirmed and the concentration was quantified using an appropriate, previously validated GC MS or LC MS MS assay. The Institutional Review Board of the University of Utah in Salt Lake City, Utah approved all studies using human samples. Adulterants and additional drug specificity The following adulterants were individually dissolved in 30 ml drug positive urine: 5 ml dry eye drops, 5 g Comet Cleaner with Bleach, 3 ml Backdown antibacterial hand soap, 5 ml 5.0% gluteraldehyde, 5 ml 1.0 M sodium nitrite, 6.5 g pyridinium chlorochromate (PCC), and 5 ml 0.1 M acetic acid. The urine used had previously screened positive for multiple drugs, including THC. Further adulteration studies were accomplished by testing 30 ml of non-urine liquid: Monster energy drink, Kroger sterile eye drops, and water. For all screens, results were read after incubating for 5 min. Cross-reactivity was assessed using either purified reference standards (Cerilliant) or patient samples confirmed positive for the drug. The standards or the patient samples were diluted in blank urine, and results were read after 5 min. Results Limit of detection and time course The limit of detection was defined as the drug concentration that gave a positive result for all 5 cups after 5 min. This value was identical to the manufacturer s published cutoff concentration for the following drug classes: buprenorphine, cocaine, opiates, and propoxyphene (Table I). Three of the drug classes showed a limit of detection at 125% of the cutoff concentration: opiates, barbiturates, and marijuana. Two drug classes, methadone and benzodiazepines, required 150%; 3,4-methylenedioxymethamphetamine (MDMA) was the only 48
Table II (continued). Detailed Description of the 24 Patient Samples Screened Using the EZ Cup Drug(s) Concentration Positive Using Sample Present (ng.ml) E-Z Cup 8 Buprenorphine 920 Buprenorphine Methamphetamine 1230 Methamphetamine 9 Oxycodone 750 Oxycodone Benzoylecognine 640 Cocaine Norbuprenorphine 25 Buprenorphine 10 Methadone 1110 Methadone Morphine 840 Opioids Buprenorphine 150 Buprenorphine Benzodiazepines* 11 Hydromorphone 40 Opioids Morphine 1510 Oxycodone 400 Oxycodone Oxymorphone 20 12 Lorazepam 350 Benzodiazepines Butalbitol 4830 Barbiturates Codeine 340 Opioids Morphine 11 Methadone 780 Methadone Oxycodone 1860 Oxycodone Propoxyphene 210 Propoxyphene Benzoylecognine 60 Cocaine Norbuprenorphine 7 Buprenorphine Buprenorphine-glucuronide 29 13 MDMA 2000 MDMA Methamphetamine Propoxyphene 26 Propoxyphene Butalbitol* 3400 14 Oxazepam 2930 Benzodiazepines Lorazepam 380 Methamphetamine 800 Methamphetamine Amphetamine 1550 Amphetamine MDMA 400 MDMA Propxyphene 1370 Propoxyphene 15 Methadone 475 Methadone Morphine 830 Opioids Codeine 90 6-Acetylmorphine 170 Methamphetamine 2030 Methamphetamine MDMA 480 MDMA THC 4 THC 16 Propoxyphene 90 Propoxyphene Amphetamine 3200 Amphetamine Codeine 510 Opioids Buprenorphine 9 Buprenorphine Norbuprenorphine 29 Norbuprenorphine-glucuronide 1100 THC 18 THC * False-positive or false-negative result. drug to give consistent positive results at 75% of the expected value. The observed cutoff levels for methamphetamine and amphetamine were measured using a racemic mixture standard. The antibodies against these stimulants are designed to more specifically detect the illegal d-form of the drugs, and hence the published cutoff for the racemic mixture is greater. The observed cutoffs for the racemic mixtures were much lower (< 45%) than the published cutoffs: 1250 ng/ml for both amphetamine and methamphetamine compared to the published 3000 and 4500 ng/ml, respectively. Using these values and assuming that there would be a proportional increase in sensitivity with a pure d-isomer, the cutoff concentrations for d-amphetamine and d-methamphetamine were calculated. For example, the measured LOD was 58% less than the published LOD for the racemic mixture of amphetamines. Thus, we assumed that the d-isomer LOD would behave similarly, and the LOD was calculated to be 58% less than the manufacturer states. A comparison of the published cutoff concentrations and the validated concentrations are listed in Table I. One advantage of most POC devices, including the Integrated E-Z Split Key Cup II, is the rapid production of results. According to the manufacurer s instructions, results should be visualized after 5 min. In order to determine how incubation time influences test results, cups used for the LOD study were resulted after 5, 30, 60, 120, and 1440 min. The results were similar for every concentration tested. The results for the cups incubated with 125% or similar concentration of the published cutoff are shown in Table III. In general, results were stable up to 60 min, but began to give false-negative results after 2 h. Results interpreted after 1440 min were markedly affected, producing a significant amount of false-negative results. Of note, the two drugs that were determined to require 150% of standard to produce consistently positive results at 5 min (methadone and benzodiazepines) had improved sensitivity after 30 min. These two drugs also showed consistently positive results after 120 min using concentrations at 100% of 49
the published cutoffs; concentrations at 75% of the published cutoff did not give positive results, even after overnight incubation. Patient urine samples Twenty-four urine samples confirmed positive for a total of 95 drugs of abuse and 5 urine samples negative for all relevant drugs were screened (29 samples total; Tables II, IV, and V). One false-positive and one-false negative result were observed. The false-positive result was for benzodiazepines. Given that 13 true-positive urines screened positive and 26 true-negative Table II (continued). Detailed Description of the 24 Patient Samples Screened Using the EZ Cup Drug(s) Concentration Positive Using Sample Present (ng/ml) E-Z Cup 50 17 Benzoylecognine 170 Cocaine Phenobarbitol 2370 Barbiturates Temazepam 270 Benzodiazepines Oxazepam 325 Nordiazepam 65 18 Methamphetamine 6100 Methamphetamine THC 6 THC Phenobarbitol 1030 Barbiturates Amphetamine 1600 Amphetamine Lorazepam 60 Benzodiazepines 7-Aminoclonazepam 30 Buprenorphine-glucuronide 5 Buprenorphine Norbuprenorphine 4 Norbuprenorphine-glucuronide 13 19 THC 2 THC Phenobarbitol 590 Barbiturates α-hydroxyalprazolam 900 Benzodiazepines Alprazolam 320 Benzoylecognine 700 Cocaine 20 Methamphetamine 5300 Methamphetamine Amphetamine 830 Amphetamine Propoxyphene 440 Propoxyphene 21 Tramadol 16030 22 Hydrocodone 8300 Opioids Oxazepam 130 Benzodiazepines 23 7-Aminoclonazepam 750 Benzodiazepines Hydrocodone 1050 Opioids THC 16 THC 24 Fentanyl 5000 THC 25 THC Nordiazepam 440 Benzodiazepines Oxazepam 1800 Temazapam 1400 7-Aminoclonazepam 580 Phenobarbitol 404 Barbiturates Oxycodone 400 Oxycodone samples screened negative, the specificity and sensitivity were calculated to be 96.3% and 100.0%, respectively. The diagnostic efficiency was 96.6%. The root cause of the false-positive urine result was evaluated by testing for the presence of a reacting compound (22 benzodiazepines/metabolites were examined); however, none was detected. The sample also tested negative for sertraline, a commonly prescribed antidepressant that the manufacturer suggests can cause positive benzodiazepine results, leaving the false positive unresolved. The false-negative result was for barbiturates. This sample was confirmed to be positive for butalbitol at concentrations well above the cutoff. When the specimen was decanted into a new screening cup, the barbiturates gave a positive result, suggesting that there may be variation between cups. Thus, with 5 true-positive screening results and 23 true-negative screening results, the sensitivity, specificity, and diagnostic efficiency are 83.3%, 100.0%, and 96.6%, respectively. All other drug classes had a specificity, sensitivity, and diagnostic efficiency of 100% for the samples tested. It is noteworthy that many of the drugs gave positive results at concentrations well below the published cutoff and the limit of detection quantified using purified standards. This result is not unexpected, considering the large amounts of metabolites structurally similar to the parent drug of interest that are expected to be present in the urine of drug users. In accordance, the patient results for methadone were positive at levels well below the reagent standard cutoff because the assay recognizes both methadone and 2-ethylidene- 1,5-dimethyl-3,3-diphenylpyrrolidine (EDDP) equally. Similarly, high levels of MDMA (> 2000 ng/ml) can lead to falsepositive results for methamphetamine. Although MDMA is not a metabolite of methamphetamine, the structure of MDMA and the structures of MDMA metabolites are very similar to methamphetamine. This cross-reactivity was reported by the manufacturer and supported by our results (Table IV). Adulterants The use of household or proprietary chemicals to alter drug screen results is recognized to occur. Thus, knowing the effect that frequently used adulterants might have on the Integrated E-Z Split Key Cup II is important. Samples that had previously screened positive for multiple drug classes were manipulated with liquid soap, Kroger sterile eye drops, pow-
dered bleach, acetic acid, gluteraldehyde, or sodium nitrite. There were no instances of these substances giving a falsenegative result; gluteraldehyde gave a false-positive result for methamphetamine. Both the liquid soap and the gluteraldehyde weakened the intensity of the control band. False positives Table III. Time Course of Results Using Purified Reagent Standards at 125% the Published Cutoff* 5 30 60 120 1440 Drug (min) (min) (min) (min) (min) d,l-methamphetamine 5 5 5 4 0 d,l-mdma 5 5 5 4 0 11-nor- 9 -THC-9 COOH 5 5 5 5 5 Oxazepam (BZO) 3 5 5 5 4 Methadone 2 5 5 5 4 Secobarbitol (BAR) 5 5 5 5 5 d,l-amphetamine 5 5 5 5 4 Morphine 5 5 5 5 5 Oxycodone 5 5 5 5 5 Propoxyphene 5 5 5 5 2 Benzoylecgonine (COC) 5 5 5 5 5 Buprenorphine 5 5 5 5 2 * The results represent the number of cups positive from a total of 5. Bold results indicate that 100% of cups were positive for a drug. or weakened intensity of negative bands is expected after the addition of these substances, as all are known to interfere with protein interactions, and hence, antibody antigen interactions. Most interesting was the sample adulterated with PCC. The addition of PCC inhibited the binding of all antibodies to the membrane giving a false-positive result for all drugs being tested, while also blocking the formation of the control band. A more troubling result was obtained after the addition of non-urine liquid to the screening cup. Because the control band is only used to detect proper volume of specimen, membrane wicking, and correct procedural technique, it will produce a colored line regardless of the matrix added to the cup, not specifically in the presence of urine. The addition of water, Monster energy drink, or Kroger sterile eye drops all gave results that mimicked blank urine. Cross-reactivity of additional drugs The manufacturer provided an extensive list of compounds that do or do not cross-react with the various drug classes screened by the Integrated E-Z Split Key Cup II. However, there were still commonly prescribed or abused drugs that were not present on their list. For these compounds, patient urine or blank urine plus purified reference standard were screened using the cup at concentrations similar to what would be observed in urine. Patient urine was used to test cross-reactivity of fentanyl (5000 ng/ml) and tramadol (16,030 ng/ml); purified reagent standards were used to test cross-reactivity of Table IV. Overview of Patient Urine Screening Results: Compounds that Screened Positive for the Various Drug Classes, the Range of Concentrations Present in the Specimens, the Number of Samples (n) for the Specific Compound, and the Mean Concentration of All Positive Specimens Concentration % Drug Range Mean Cross- Class Compound n (ng/ml) (ng/ml) Reactivity* Amphetamines Amphetamine 8 170 4700 1920 100% Cocaine Benzoylecognine 6 60 1220 510 100% Methamphetamine Methamphetamine 8 1230 17340 5410 100% MDMA 1 2000 2000 50% MDMA MDMA 5 400 2000 780 100% Barbiturates Butalbitol 2 3400 4830 4120 12% Phenobarbitol 4 400 2370 1100 300% Methadone Methadone 5 475 1600 890 100% Opiates Morphine 6 420 3010 1160 100% Codeine 2 345 830 590 100% Hydrocodone 2 1050 8300 4680 0.60% Oxycodone Oxycodone 9 170 4100 1270 100% Propoxyphene Propoxyphene 7 15 1370 400 100% Buprenorphine Buprenorphine 2 150 920 540 100% Norbuprenorphine 4 4 29 20 70% Marijuana 11-nor- 9 -THC-9 COOH 12 2 210 60 100% * Estimated from the detection concentration listed in the package insert by dividing the LOD for the drug class standard by the LOD of the specific drug present. 51
tapentadol (2500 ng/ml), phentermine (17,000 ng/ml), ritalinic acid (17,000 ng/ml), selegiline (10 ng/ml), desmethylselegiline (1000 ng/ml), zolpiclone (5000 ng/ml), and zolpidem (5000 ng/ml). The results revealed no apparent cross-reactivity with any of the tested compounds. markedly decreased compared to surrounding bands, the specimen was resulted as positive. On confirmation, the specimen was found to contain quantities of benzodiazepines well below the cutoff. Thus, interpretation of results may be subjective. To Discussion Rapid detection of drugs in patient specimens may be beneficial to acute care facilities, rehabilitation facilities, and specialty clinics such as those used for pain management, and emergency departments. In this study we evaluated the performance of the Integrated E-Z Split Key Cup II, a chromatographic immunoassay that qualitatively detects multiple drugs and drug metabolites in urine. Overall, the device met quality standards, displaying high diagnostic sensitivity, specificity, and diagnostic efficiency. The device is accurate, easy to use, and cost effective, making it an ideal screen for institutions where timely results can have a profound effect on patient care. The tight seal of the cup allows the results to easily be photocopied (Figure 1), facilitating patient record maintenance and standardization of reporting procedures. Although CLIA has identified the cup to be a moderately complex test, the drug screening cup is remarkably easy to use. Many POC devices are waived, meaning that the basic regulatory guideline is to follow the manufacturer s instructions. However, because the Integrated E-Z Split Key Cup II is a moderately complex test, the institution employing the screen must follow more stringent guidelines regarding the training of testing personnel (including regular competency evaluations). Most importantly, when implementing a moderately complex test, the institution must demonstrate proficiency through participation in an accredited program such as that offered through the College of American Pathologists. The results of the cup are generally very clear, especially if the urine is negative for a drug/multiple drugs. However, if the cup is closely scrutinized the reader may be able to convince himself or herself that a drug is or is not present. For example, the cup used to analyze benzodiazepine sample 9 (Table V) showed a very faint band. Because the intensity was Table V. Concentrations of Benzodiazepines and Metabolites Present in Patient Urine Samples that Tested Positive for Benzodiazepines Using the E-Z Cup Concentration % Cross-Reactivity Sample Benzodiazepine (ng/ml) Relative to Standard* 1 Lorazepam 150 20% Alprazolam 30 150% α-hydroxyalprazolam 130 25% 7-Aminoclonazepam 150 40% 2 Nordiazepam 830 80% Oxazepam 1660 100% Temazepam 1920 300% Lorazepam 470 20% 3 Nordiazepam 450 80% Oxazepam 640 100% Temazepam 340 300% Alprazolam 50 150% 4 α-hydroxyalprazolam 180 25% 7-Aminoclonazepam 40 40% Alprazolam 50 150% Nordiazepam 280 80% Oxazepam 1100 100% Temazepam 880 300% 5 Alprazolam 370 150% 6 Lorazepam 350 20% 7 Oxazepam 2930 100% Lorazepam 380 20% 8 Temazepam 270 300% Oxazepam 325 100% 9 Lorazepam 60 20% Nordiazepam 70 80% 7-Aminoclonazepam 30 40% 10 α-hydroxyalprazolam 900 25% Alprazolam 320 150% 11 7-Aminoclonazepam 750 40% 12 Nordiazepam 440 80% Oxazepam 1800 100% Temazepam 1400 300% 7-Aminoclonazepam 580 40% 13 Oxazepam 130 100% * Estimated from the detection concentration listed in the package insert by dividing the LOD for the drug class standard by the LOD of the specific drug present. 52
minimize bias of results, it is helpful to employ the following guidelines: 1. read the results at least an arm s length distance; 2. blind the reader to the patient name; and 3. run a negative control or blank specimen in parallel with the patient sample. Archiving a photocopy of the cup may assist with resolving any discrepancies or inconsistencies over time and among analysts. Confirmation of any result inconsistent with expectations will also improve interpretation of results. However, sensitivity for detection of drugs, and hence, interpretation of results, is also a function of available confirmation testing. Thus, positive results in an immunoassay such as the EZ-Cup may reflect drugs or metabolites that are not detected by the confirmation testing available. It is likely that this latter scenario explains the false-positive benzodiazepine result observed in this study. These experiments clearly support the diagnostic capabilities of the drug testing device. However, it is important to note that the study is limited by the use of samples with a high prevalence population (24 positive/29 negative patient samples). The study was designed in this manner mainly for overall efficiency, but also because we expected the device to be utilized in facilities with high prevalence of drug abuse (such as rehabilitation and pain management centers) and not for the general population. A more broadly applicable study design would have been to test individuals at random and thus measure the predictive values of the device in a lower prevalence population. Government agencies require random and regular drug abuse testing. The initial screening cutoffs for the mandatory drug classes are published by the Substance Abuse and Mental Health Services Administration (SAMSHA) and were recently modified in May 2010 to be 50 ng/ml marijuana; 150 ng/ml cocaine; 2000 ng/ml opiates; 25 ng/ml phencyclidine; 500 ng/ml amphetamines; 500 ng/ml MDMA; and 10 ng/ml 6- acetylmorphine. With these cutoffs, this device would be appropriate for only MDMA, cocaine, and marijuana. Although this cup is not ideal for a SAMSHA-certified agency, the excellent performance of the screening cup encourages the validation and use of similar devices in institutions where rapid initial testing may be practical and advantageous. Acknowledgments The authors would like to thank Chantry Clark and Heidi Carlisle for their helpful laboratory assistance. We are grateful for the funding to support this study provided by ARUP Laboratories and the ARUP Institute for Clinical and Experimental Pathology. References 1. E. Lenters-Westra and R.J. Slingerland. Six of eight hemoglobin A1c point-of-care instruments do not meet the general accepted analytical performance criteria. Clin. Chem. 56(1): 44 52 (2010). 2. T.R. Gray, L.L. LaGasse, L.M. Smith, C. Derauf, P. Grant, R. Shah, A.M. Arria, S.A. Della Grotta, A. Strauss, W.F. Haning, B.M. Lester, and M.A. Huestis. Identification of prenatal amphetamines exposure by maternal interview and meconium toxicology in the Infant Development, Environment and Lifestyle (IDEAL) study. Ther. Drug Monit. 31(6): 769 75 (2009). 3. E.R. Harrison, J. Haaga, and T. Richards. Self-reported drug use data: what do they reveal? Am. J. Drug Alcohol Abuse 19(4): 423 441 (1993). Manuscript received June 15, 2010; revision received July 21, 2010. 53