Abstract

On-site drugs of abuse testing devices have undergone continuous improvement. We evaluated three devices with different designs: an automated reader, the Multi-Drug Screen Test Device with DxLINK (DxLINK; Innovacon, Alere, San Diego, USA) and two colorimetric immunoassays, the One Step Multi-Line Screen Panel with Integrated E-Z Split Key Cup II (E-Z Cup; Innovacon, Alere) and the One Step Multi-Drug Screen Panel card (Multi4 card; Alere, Abon Biopharm, Hangzhou, China). Eleven drugs [amphetamine, secobarbital, oxazepam, buprenorphine, benzoylecgonine, methylenedioxymethamphetamine (MDMA), 11-nor-9-carboxy-Δ9-tetrahydrocannabinol (THC), methamphetamine, methadone, morphine and nortriptyline] were tested using the DxLINK and E-Z Cup. Four drugs (benzoylecgonine, THC, methamphetamine and morphine) were tested using the Multi4 card using control materials (Detectabuse Stat-Skreen; Biochemical Diagnostics, Edgewood, NY, USA). The concentrations (−50%, −25%, +25%, +50% and 3× cut-off values) of the control materials were confirmed by mass spectrometry. Concordance rates were calculated around cut-offs. All devices showed high overall agreement rates of >90% with a few exceptions: the DxLINK exhibited lower sensitivity for benzoylecgonine, methadone and nortriptyline (60% and 30%, 92% and 40%, and 96% and 60% sensitivity at +50% and +25% cut-off levels, respectively). The E-Z Cup exhibited lower sensitivity for oxazepam and nortriptyline (97% and 50%, and 97% and 40% sensitivity at +50% and +25% cut-off levels, respectively). We additionally evaluated test-band color by visual inspection using a standard color-scale card. When detailed color criteria for determination of positivity were applied for the E-Z Cup, using slightly less stringent criteria, oxazepam, buprenorphine, MDMA and nortriptyline showed increases in sensitivity from 70–80% to 90–100%, all with a specificity above 98%. Overall, all devices exhibited satisfactory performance at ±50% cut-off levels for commonly used drugs, with the exception of lower sensitivity for cocaine testing for DxLINK. Careful evaluation of devices and elaborate calibration of visual interpretation for determining positivity may help improve the performance of these devices.

Introduction

In 2012, the United Nations Office on Drugs and Crime (UNODC) reported that ~243 million people worldwide use drugs; this number has increased annually since 2006 (1). According to a report by the South Korean National Police Agency, the number of drug offenders has gradually increased since 2012 (2). During the first half of 2015, 3,370 narcotic offenders were taken into custody, which was an increase of 22.5% compared to the same period in the previous year (2). To control this nationwide trend of increasing drug abuse, robust identification of and prompt legal actions against drug abusers are needed, and rehabilitation treatment should be initiated immediately to promote recovery. To achieve these goals, rapid-result drug tests are needed.

In the absence of on-site drug-testing devices, clinical specimens from individuals suspected of drug abuse are shipped to a laboratory capable of conducting drug testing using immunoassays and other methods. Consequently, there is a significant delay before test results are obtained (3). On-site drug testing can reduce the costs of industrial drug testing by reducing the number of samples that are sent to reference laboratories. On-site drug-testing protects the chain of custody, reducing the risk of contamination or damage during transportation. It can also protect the confidentiality of people being tested and thereby reduces anxiety surrounding test results.

Many different on-site drug-testing devices have been developed and are now widely used, and on-site devices capable of testing for multiple drug classes are being advanced and evaluated (311). However, performance evaluations of these newly developed devices have been limited, and confirmatory evaluations of screening-stage performance are lacking.

Even when applied to the same drug and employing the same immunoassay technique, on-site devices for drug-testing fall into a wide variety of categories depending on the methods used for result interpretation and mode of operation. The devices evaluated in this study are all based on competitive binding immunoassays and include the Multi-Drug Screen Test Device with DxLINK Technology (DxLINK; Innovacon, Inc., Alere, San Diego, USA), the One Step Multi-Line Screen Test Panel with Integrated E-Z Split Key Cup II (E-Z Cup; Innovacon, Inc., Alere) and the One Step Multi-Drug Screen Test Panel (Multi4 card; Abon Biopharm, Alere, Hangzhou, China). These devices all enable qualitative screening for dozens of commonly abused drugs or drug classes, including amphetamine, barbiturates, benzodiazepines, buprenorphine, cocaine, cannabinoids, methadone, methamphetamine, methylenedioxymethamphetamine (MDMA), opiates and tricyclic antidepressants (TCA), and are based on US Substance Abuse and Mental Health Services Administration (SAMHSA)-recommended cut-off concentrations for these drugs. In all of these assays, if a given drug is present in a urine sample, it competitively binds to a specifically colored antibody, which prevents the antibody from binding to antibody-drug conjugates that are present in the detection region of the test. Consequently, the test line disappears, indicating a positive result for that drug (3). The Multi4 card represents the simplest technique for determining whether a result is positive or negative; its results are based on any change in the color of the test line 3–5 minutes after the immersion of the strip in a urine sample. The E-Z cup includes both a cup and an attached key for increased convenience; a certain amount of specimen is applied to the test strip at a predetermined pressure using the apparatus. This device allows testers to read changes in the color of the control and test lines and to interpret the results in the same manner as the Multi4 card. The DxLINK uses an automated reader: an image of a screening device is electronically captured by a scanner, and the test system software analyzes the test and control lines to report the polarity of the results. Devices with automated readers are designed to reduce clerical errors by eliminating the potential subjectivity of the visual inspection of test results and enable rapid reading of results for multiple drugs. In addition, automatically generated test results provide traceability of the individual who operated the reader, convenient built-in quality control checks, automatic recording of test results, and automatic image capture and interfacing of the reader and a laboratory's information-storage system.

To evaluate the drug-screening performance of these three devices, control materials at various concentrations above and below SAMHSA-recommended cut-off concentrations were employed. The actual concentrations of the control materials were determined using a reference method. The results of these tests were used to assess test performance around the cut-off levels of the devices.

Materials and Methods

Devices

The DxLINK and E-Z Cup can detect 11 classes of drugs: amphetamine, barbiturates, benzodiazepines, buprenorphine, cocaine, MDMA, cannabinoids, methamphetamine, methadone, morphine and TCA. The manufacturers’ suggested cut-off values are shown in Table I, and drug testing using these devices was conducted in accordance with the manufacturers’ instructions. Among the One Step Multi-Drug Screen Test Panel (Abon Biopharm, Alere), the Multi4 test card can detect four classes of drugs (cocaine, cannabinoids, methamphetamine and high-level opiates). When reading the test results, the automated reader of the DxLINK scans an image, and the results are automatically reported as either presumptive positive or negative based on the cut-offs from its embedded software. These results were accepted without any modification in this study. The E-Z Cup and Multi4 card test strips were also visually inspected by two or more test readers, and test positivity was determined based on the intensity of the test line. According to the manufacturer's instructions, if a red or pink test line appears, the test should be considered negative, even for a very faint pink line. Because very faint test lines may cause some confusion in the interpretation of results, we additionally recorded the color intensity of the test lines using values ranging from Grade 1 (G1) through G10 as compared with a standard color-scoring card (Gold Color Card; Abon Biopharm) provided by the manufacturer, which enabled more objective evaluation of the intensity and positivity of the results. This color-scale card was developed for the purpose of improving the calibration of immunologic assays when developing and maintaining devices. On a standard color card, 11 color scales that include a range of light to dark colors are designated into 11 different grades (G1, G2, G3, G3.5, G4, G5, G6, G7, G8, G9 and G10). When using a double-antibody sandwich assay, a test line with G1–G2 color intensity, which is very faint and hardly visible, is generally considered to be negative, whereas a test line with G3 intensity is sometimes considered positive and sometimes negative. In devices using competitive binding immunoassays like the E-Z Cup and Multi4 card, lower grades indicate higher drug concentrations because the test line becomes fainter with increasing drug concentration. The manufacturer of the Multi4 card provided additional performance evaluation data for using this standard color-scoring card that had been previously submitted to the Korea Food and Drug Administration (KFDA) during the device's approval process. Using these data, results with G3 or below were recorded as positive, which was interpreted as the presence of drugs in a sample at a concentration above the established cut-off concentration. Conversely, results with G3.5 or above were recorded as negative.

Table I.

Cut-off Concentrations for Drugs and Drug Metabolites

Drug class Common calibratorsa Cut-off value (ng/mL) Devices 
Amphetamine (AMP) d-Amphetamine 1,000 DxLINK, EZ Cup 
Barbiturates (BAR) Secobarbital 300 DxLINK, EZ Cup 
Benzodiazepines (BZO) Oxazepam 300 DxLINK, EZ Cup 
Buprenorphine (BUP) Buprenorphine 10 DxLINK, EZ Cup 
Cocaine (COC) Benzoylecgonine 300 DxLINK, EZ Cup, Multi4 card 
MDMA d,l-MDMA 500 DxLINK, EZ Cup 
Cannabinoids (THC) (−)-(THC) 50 DxLINK, EZ Cup, Multi4 card 
Methamphetamine (MET) d-Methamphetamine 1,000 DxLINK, EZ Cup, Multi4 card 
Methadone (MTD) Methadone 300 DxLINK, EZ Cup 
Morphine (MOP 300) Morphine 300 DxLINK, EZ Cup 
Opiate (OPI 2000) Morphine 2,000 Multi4 card 
TCA Nortriptyline 1,000 DxLINK, EZ Cup 
Drug class Common calibratorsa Cut-off value (ng/mL) Devices 
Amphetamine (AMP) d-Amphetamine 1,000 DxLINK, EZ Cup 
Barbiturates (BAR) Secobarbital 300 DxLINK, EZ Cup 
Benzodiazepines (BZO) Oxazepam 300 DxLINK, EZ Cup 
Buprenorphine (BUP) Buprenorphine 10 DxLINK, EZ Cup 
Cocaine (COC) Benzoylecgonine 300 DxLINK, EZ Cup, Multi4 card 
MDMA d,l-MDMA 500 DxLINK, EZ Cup 
Cannabinoids (THC) (−)-(THC) 50 DxLINK, EZ Cup, Multi4 card 
Methamphetamine (MET) d-Methamphetamine 1,000 DxLINK, EZ Cup, Multi4 card 
Methadone (MTD) Methadone 300 DxLINK, EZ Cup 
Morphine (MOP 300) Morphine 300 DxLINK, EZ Cup 
Opiate (OPI 2000) Morphine 2,000 Multi4 card 
TCA Nortriptyline 1,000 DxLINK, EZ Cup 

aThe same calibrators were used for all three devices.

Test materials

The commercially available control samples known as Detectabuse Stat-Skreen Liquid controls (Biochemical Diagnostics, Edgewood, NY, USA) were used for evaluations. All three devices used the same calibrators (Table I), and we used matched control materials during the calibration process. SAMHSA-recommended cut-off concentrations were used as reference concentrations, and the Stat-Skreen Low-Opiate and High-Opiate controls included materials such as d-amphetamine, secobarbital, oxazepam, benzoylecgonine, (−)-11-nor-9-carboxy-Δ9-tetrahydrocannabinol (THC), d-methamphetamine, methadone, nortriptyline and morphine. Morphine was categorized into two groups, low- and high-opiate controls, which corresponded to concentrations of 300 and 2,000 ng/mL, respectively. Stat-Skreen with a low-opiate concentration was used to evaluate the DxLINK and E-Z Cup devices, whereas Stat-Skreen with a high-opiate concentration was used to evaluate the Multi4 card. In all cases, evaluations were based on the cut-off value (Table I) for each device. The Stat-Skreen low-opiate and high-opiate controls consisted of a negative control and target materials at concentrations of −50%, −25%, +25%, +50% and 3× the cut-off, in addition to materials at the reference cut-off; all of these concentrations were evaluated in this study. Stat-Skreen with MDMA and buprenorphine was distributed as a separate product at a single 2× cut-off concentration. Series of concentrations of buprenorphine and MDMA reference materials were made at our laboratory. Material at a 2× cut-off concentration was mixed with the negative control material at the desired concentration (−50%, −25%, +25%, +50% and 1× cut-off concentration). All samples were aliquoted according to the volume needed and stored at <−20°C until use.

Mass spectrometry

The concentrations of amphetamine, secobarbital, oxazepam, benzoylecgonine, MDMA, methamphetamine, methadone, morphine and THC in each Stat-Skreen control were certified by test laboratories that verify and monitor the usefulness of devices for measuring drugs of abuse in accordance with the standards set by the US SAMHSA, College of American Pathologists (CAP) and American Association of Clinical Chemistry (AACC). The concentrations of eight drugs, excluding THC, were determined by gas chromatography-mass spectrometry at the Center for Advanced Laboratory Medicine (CALM), UC San Diego Department of Pathology (San Diego, USA), whereas the concentration of THC was measured by liquid chromatography-mass spectrometry in the test laboratories located at Seoul National University Bundang Hospital, Department of Laboratory Medicine (Bundang, Republic of Korea). For secobarbital, buprenorphine and nortriptyline, mass spectrometry results obtained from the SAMHSA or a CAP-certified test laboratories were regarded as the actual concentrations, in line with manufacturer recommendations. The mass spectrometry results are presented in Table II. For most drugs, the results were within ±15% of the target concentrations, but the concentration of THC was significantly lower (range from −18.5 to −49.3%) than the target concentration (Table II). For the THC control, the actual concentration of the −25% cut-off level-labeled material was −48% of the cut-off concentration, which was closer to the −50% cut-off level, and the actual concentration of the cut-off-labeled material was −29% of the cut-off, which was closer to the −25% cut-off concentration. Materials that fell into the range between the +25% and +50% cut-off concentrations actually ranged from −17 to +10% of the cut-off concentration, which was far less than the expected values.

Table II.

Mass Spectrometry Results Compared to Each Target Concentration for the Control Material

Control material Drug concentration, ng/mL (Percentage of cut-off level, %)a 
−50% Cut-off −25% Cut-off Cut-off +25% Cut-off +50% Cut-off 2× Cut-off 3× Cut-off 
Amphetamine 503 (50) 711 (71) 909 (91) 1,145 (115) 1,419 (142) NA 2,844 (284) 
Secobarbitalb 147 (49) 228 (76) 294 (98) 363 (121) 441 (147) NA 885 (295) 
Oxazepam 159 (53) 244 (81) 339 (113) 407 (136) 500 (167) NA 946 (315) 
Buprenorphine 5 (54) 8 (81) 11 (108) 14 (135) 16 (162) 22 (216) NA 
Benzoylecgonine 144 (48) 219 (73) 280 (93) 344 (115) 474 (158) NA 861 (287) 
MDMA 253 (51) 380 (76) 506 (101) 633 (127) 759 (152) 1,012 (202) NA 
THC 10 (22) 28 (52) 35 (71) 42 (83) 55 (110) NA 125 (251) 
Methamphetamine 487 (49) 711 (71) 935 (94) 1,155 (116) 1,427 (143) NA 2,755 (276) 
Methadone 142 (47) 216 (72) 283 (94) 336 (112) 434 (145) NA 835 (278) 
Morphine, Low 137 (46) 217 (72) 288 (95) 359 (120) 421 (140) NA 820 (273) 
Morphine, High 907 (45) 1,418 (71) 1,644 (82) 2,345 (117) 2,779 (139) NA 5,604 (280) 
Nortriptylineb 530 (53) 790 (79) 1,070 (107) 1,330 (133) 1,600 (160) NA 3,020 (302) 
Control material Drug concentration, ng/mL (Percentage of cut-off level, %)a 
−50% Cut-off −25% Cut-off Cut-off +25% Cut-off +50% Cut-off 2× Cut-off 3× Cut-off 
Amphetamine 503 (50) 711 (71) 909 (91) 1,145 (115) 1,419 (142) NA 2,844 (284) 
Secobarbitalb 147 (49) 228 (76) 294 (98) 363 (121) 441 (147) NA 885 (295) 
Oxazepam 159 (53) 244 (81) 339 (113) 407 (136) 500 (167) NA 946 (315) 
Buprenorphine 5 (54) 8 (81) 11 (108) 14 (135) 16 (162) 22 (216) NA 
Benzoylecgonine 144 (48) 219 (73) 280 (93) 344 (115) 474 (158) NA 861 (287) 
MDMA 253 (51) 380 (76) 506 (101) 633 (127) 759 (152) 1,012 (202) NA 
THC 10 (22) 28 (52) 35 (71) 42 (83) 55 (110) NA 125 (251) 
Methamphetamine 487 (49) 711 (71) 935 (94) 1,155 (116) 1,427 (143) NA 2,755 (276) 
Methadone 142 (47) 216 (72) 283 (94) 336 (112) 434 (145) NA 835 (278) 
Morphine, Low 137 (46) 217 (72) 288 (95) 359 (120) 421 (140) NA 820 (273) 
Morphine, High 907 (45) 1,418 (71) 1,644 (82) 2,345 (117) 2,779 (139) NA 5,604 (280) 
Nortriptylineb 530 (53) 790 (79) 1,070 (107) 1,330 (133) 1,600 (160) NA 3,020 (302) 

Abbreviations: NA, not assessed.

aPercentage of target cut-off value as indicated in Table I.

bVerified concentrations provided by manufacturers.

Performance evaluation

Repeated measurements were performed with the three devices using control materials at various concentrations to assess imprecision around the cut-off. After testing twice a day for 10 days, a total of 20 repeated-measurement results were obtained for the E-Z Cup and Multi4 card. For the DxLINK, a total of 10 repeated-measurement results were obtained after testing once a day for 10 days. For the E-Z Cup, two different assay lots were used: one spanned from Day 1 to the morning of Day 7 and the other from the afternoon of Day 7 to Day 10. For the materials with −50%, +50% and 3× (2× for buprenorphine and MDMA) cut-off concentrations, 40 additional aliquots (15, 15 and 10, respectively) were prepared to perform a random blinded test with each device. The test results were then recorded. To determine the cut-off concentrations at which positive or negative results were expected, probit analysis was performed using the results of 20 repeated measurements of the control materials at the cut-off concentration and at the −50%, −25%, +25%, +50% and 3× cut-off levels.

Statistical analysis

True positive and negative results were interpreted based on the drug concentrations of the control materials that were verified by mass spectrometry. Then, the results generated by the devices were compared with the recorded results to calculate positive and negative agreement fractions. Receiver operating characteristic (ROC) curves were used to compute the color grade and the area under the curve (AUC) to assess the sensitivity and specificity of the test strips for the drug cut-off concentrations. SPSS (version 17.0, SPSS Inc., Chicago, IL, USA) was used for these analyses.

Results

Performance evaluation around cut-off levels

At ±50% of each cut-off level, the three on-site drug-testing devices exhibited high positive and negative agreement (greater than 90%) with the control materials, as confirmed by mass spectrometry, for most drug classes (Table III). All three devices exhibited negative results that were 100% consistent with the control material at a −50% cut-off level across all drug groups. For the DxLINK, there was 100% positive agreement with the +50% cut-off level control materials for six drug classes and greater than 90% positive agreement for buprenorphine, methadone and nortriptyline. For benzoylecgonine, the device exhibited 60% positive agreement, which is somewhat low. The E-Z Cup showed 100% positive agreement for eight drug classes and greater than 95% positive agreement for oxazepam and nortriptyline. The Multi4 card exhibited 100% positive agreement for all three drugs tested, excluding THC, for which the measurement at the +50% cut-off was not available.

Table III.

Performance Evaluation Using Control Specimens

Drug group DxLINK E-Z Cup Multi4 card 
Positive agreement (%) Negative agreement (%) Overall agreement (%) Positive agreement (%) Negative agreement (%) Overall agreement (%) Positive agreement (%) Negative agreement (%) Overall agreement (%) 
±50% Cut-off 
 d-Amphetamine 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Secobarbital 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Oxazepam 25/25 (100) 25/25 (100) 50/50 (100) 34/35 (97) 35/35 (100) 69/70 (99) NA NA NA 
 Buprenorphine 23/25 (92) 25/25 (100) 48/50 (96) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Benzoylecgonine 15/25 (60) 25/25 (100) 40/50 (80) 35/35 (100) 35/35 (100) 70/70 (100) 35/35 (100) 35/35 (100) 70/70 (100) 
 MDMA 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 THC NA 10/10 (100) NA NA 20/20 (100) NA NA 20/20 (100) NA 
 d-Methamphetamine 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) 35/35 (100) 35/35 (100) 70/70 (100) 
 Methadone 23/25 (92) 25/25 (100) 48/50 (96) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Morphine, Low 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Morphine, High NA NA NA NA NA NA 35/35 (100) 35/35 (100) 70/70 (100) 
 Nortriptyline 24/25 (96) 25/25 (100) 49/50 (98) 34/35 (97) 35/35 (100) 69/70 (99) NA NA NA 
 Overall 235/250 (94) 260/260 (100) 485/500 (97) 348/350 (99) 370/370 (100) 698/700 (99.7) 105/105 (100) 125/125 (100) 210/210 (100) 
±25% Cut-off 
 d-Amphetamine 9/10 (90) 8/10 (80) 17/20 (85) 19/20 (95) 19/20 (95) 38/40 (95) NA NA NA 
 Secobarbital 10/10 (100) 9/10 (90) 19/20 (95) 20/20 (100) 19/20 (95) 39/40 (98) NA NA NA 
 Oxazepam 9/10 (90) 8/10 (80) 17/20 (85) 10/20 (50) 20/20 (100) 30/40 (75) NA NA NA 
 Buprenorphine 8/10 (80) 8/10 (80) 16/20 (80) 18/20 (90) 20/20 (100) 38/40 (95) NA NA NA 
 Benzoylecgonine 3/10 (30) 10/10 (100) 13/20 (65) 20/20 (100) 19/20 (95) 39/40 (98) 18/20 (90) 19/20 (95) 37/40 (93) 
 MDMA 10/10 (100) 1/10 (10) 11/20 (55) 19/20 (95) 20/20 (100) 39/40 (98) NA NA NA 
 THC NA 8/10 (80) NA NA 14/20 (70) NA NA 20/20 (100) NA 
 d-Methamphetamine 9/10 (90) 8/10 (80) 17/20 (85) 20/20 (100) 19/20 (95) 39/40 (98) 20/20 (100) 18/20 (90) 38/40 (95) 
 Methadone 4/10 (40) 9/10 (90) 13/20 (65) 18/20 (90) 20/20 (100) 38/40 (95) NA NA NA 
 Morphine, Low 10/10 (100) 5/10 (50) 15/20 (75) 20/20 (100) 19/20 (95) 39/40 (98) NA NA NA 
 Morphine, High NA NA NA NA NA NA 20/20 (100) 20/20 (100) 40/40 (100) 
 Nortriptyline 6/10 (60) 9/10 (90) 15/20 (75) 8/20 (40) 20/20 (100) 28/40 (70) NA NA NA 
 Overall 78/100 (78) 83/110 (75) 153/200 (77) 172/200 (86) 209/220 (95) 367/400 (92) 58/60 (97) 77/80 (96) 115/120 (96) 
Drug group DxLINK E-Z Cup Multi4 card 
Positive agreement (%) Negative agreement (%) Overall agreement (%) Positive agreement (%) Negative agreement (%) Overall agreement (%) Positive agreement (%) Negative agreement (%) Overall agreement (%) 
±50% Cut-off 
 d-Amphetamine 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Secobarbital 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Oxazepam 25/25 (100) 25/25 (100) 50/50 (100) 34/35 (97) 35/35 (100) 69/70 (99) NA NA NA 
 Buprenorphine 23/25 (92) 25/25 (100) 48/50 (96) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Benzoylecgonine 15/25 (60) 25/25 (100) 40/50 (80) 35/35 (100) 35/35 (100) 70/70 (100) 35/35 (100) 35/35 (100) 70/70 (100) 
 MDMA 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 THC NA 10/10 (100) NA NA 20/20 (100) NA NA 20/20 (100) NA 
 d-Methamphetamine 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) 35/35 (100) 35/35 (100) 70/70 (100) 
 Methadone 23/25 (92) 25/25 (100) 48/50 (96) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Morphine, Low 25/25 (100) 25/25 (100) 50/50 (100) 35/35 (100) 35/35 (100) 70/70 (100) NA NA NA 
 Morphine, High NA NA NA NA NA NA 35/35 (100) 35/35 (100) 70/70 (100) 
 Nortriptyline 24/25 (96) 25/25 (100) 49/50 (98) 34/35 (97) 35/35 (100) 69/70 (99) NA NA NA 
 Overall 235/250 (94) 260/260 (100) 485/500 (97) 348/350 (99) 370/370 (100) 698/700 (99.7) 105/105 (100) 125/125 (100) 210/210 (100) 
±25% Cut-off 
 d-Amphetamine 9/10 (90) 8/10 (80) 17/20 (85) 19/20 (95) 19/20 (95) 38/40 (95) NA NA NA 
 Secobarbital 10/10 (100) 9/10 (90) 19/20 (95) 20/20 (100) 19/20 (95) 39/40 (98) NA NA NA 
 Oxazepam 9/10 (90) 8/10 (80) 17/20 (85) 10/20 (50) 20/20 (100) 30/40 (75) NA NA NA 
 Buprenorphine 8/10 (80) 8/10 (80) 16/20 (80) 18/20 (90) 20/20 (100) 38/40 (95) NA NA NA 
 Benzoylecgonine 3/10 (30) 10/10 (100) 13/20 (65) 20/20 (100) 19/20 (95) 39/40 (98) 18/20 (90) 19/20 (95) 37/40 (93) 
 MDMA 10/10 (100) 1/10 (10) 11/20 (55) 19/20 (95) 20/20 (100) 39/40 (98) NA NA NA 
 THC NA 8/10 (80) NA NA 14/20 (70) NA NA 20/20 (100) NA 
 d-Methamphetamine 9/10 (90) 8/10 (80) 17/20 (85) 20/20 (100) 19/20 (95) 39/40 (98) 20/20 (100) 18/20 (90) 38/40 (95) 
 Methadone 4/10 (40) 9/10 (90) 13/20 (65) 18/20 (90) 20/20 (100) 38/40 (95) NA NA NA 
 Morphine, Low 10/10 (100) 5/10 (50) 15/20 (75) 20/20 (100) 19/20 (95) 39/40 (98) NA NA NA 
 Morphine, High NA NA NA NA NA NA 20/20 (100) 20/20 (100) 40/40 (100) 
 Nortriptyline 6/10 (60) 9/10 (90) 15/20 (75) 8/20 (40) 20/20 (100) 28/40 (70) NA NA NA 
 Overall 78/100 (78) 83/110 (75) 153/200 (77) 172/200 (86) 209/220 (95) 367/400 (92) 58/60 (97) 77/80 (96) 115/120 (96) 

Data are presented as number of samples with positive or negative results per total number of true positive or negative samples, and calculated percentages are presented within parenthesis. Abbreviations: NA, not assessed.

The results produced by the devices at the ±25% level of each cut-off confirmed that both false-positive respectively false-negative results were increased at this level than the ±50% level. The DxLINK had low positive agreements of 30% and 40% for benzoylecgonine and methadone, respectively, whereas the positive agreement for nortriptyline was 60%. Greater than 80% positive agreement was observed for the other drugs. Greater than 80% negative agreement was observed for most drugs, but a low negative agreement of 10% was observed for MDMA, and 50% negative agreement was observed for morphine. For the E-Z Cup, 95% negative agreement was observed for all drugs except THC, whereas 50% and 40% positive agreement were observed for oxazepam and nortriptyline, respectively. Greater than 90% positive agreement was observed for all other drugs. The Multi4 card exhibited greater than 90% positive and negative agreements for the three drugs tested with this method.

Comparing the above performance data to those reported by the manufacturer, at the ±50% cut-off levels, the manufacturers of all three devices reported results with 100% positive and negative agreement for all drug classes; therefore, the 60% positive agreement for our results for benzoylecgonine from the DxLINK was significantly lower than that claimed by the manufacturer. At the ±25% cut-off levels, the manufacturer of the DxLINK device claimed 77–80% positive agreement for amphetamine, secobarbital, morphine and MDMA; however, our data presented higher positive agreement (90–100%). In addition, the manufacturer of the DxLINK device claimed 77% negative agreement for secobarbital, which was lower than our result of 90%. Furthermore, this manufacturer also reported 83% positive agreement for benzoylecgonine and 80% negative agreement for MDMA, whereas our results for these drugs were 30% and 10%, respectively. Nortriptyline presented rather low positive agreement (77% and 60% for the manufacturer's and our results, respectively) and morphine presented rather low negative agreement (67% and 50% for the manufacturer's and our results, respectively) in both our and the manufacturer's data. At ±25% cut-off levels, the manufacturer of the E-Z Cup device presented 67% positive agreement for MDMA and morphine, 77% positive agreement for secobarbital and 67% negative agreement for MDMA and methadone, which were lower than our results (all ranged from 95 to 100%). For nortriptyline, the positive agreement reported by the E-Z Cup manufacturer is 83%, which is higher than our result of 40%.

Test accuracy for randomly selected samples

A total of 40 samples, including 15, 15 and 10 samples of the materials with −50%, +50% and 3× (2× for buprenorphine and MDMA) cut-off concentrations, respectively, were randomly tested and separately analyzed for accuracy assessment. All devices exhibited 100% negative agreement for the negative specimen at the −50% cut-off concentration, with the exception of 87% (13/15) agreement for methadone and buprenorphine and 47% (7/15) agreement for benzoylecgonine for the DxLINK. All three devices exhibited 100% positive agreement for specimens at the +50% cut-off concentration. For the materials at 3× or 2× the cut-off concentration, 100% positive agreement was reported, except for 90% (9/10) agreement values for methadone and buprenorphine by the DxLINK and E-Z Cup, respectively. Due to the instability of the control materials, assessments of high-concentration THC could not be made.

Cut-off level evaluation

The control materials were measured repeatedly at different concentrations to determine the cut-off concentrations at which 50% positive or 50% negative results were expected. The measurements confirmed that most drugs had a deviation within ±15% of the established cut-off level (Table IV;

). The DxLINK produced somewhat lower cut-off levels for MDMA and morphine (68% and 75% of each manufacturer's cut-off, respectively), whereas methadone and benzoylecgonine had somewhat higher cut-off levels (125% and 145% of each manufacturer's cut-off, respectively), consistent with the results from evaluating positive and negative agreement. For the E-Z Cup, oxazepam and nortriptyline exhibited somewhat higher cut-off levels (124% and 128% of each manufacturer's cut-off, respectively).
Table IV.

Evaluation of Cut-off Levels for DxLINK, E-Z Cup and Multi4 Card

Drugs (Drug group) Percentage of SAMHSA cut-off level, % (95% confidence interval)a 
DxLINK E-Z Cup Multi4 card 
Amphetamine 88.2 (77.2–98.7) 97.8 (90.5–104.2) NA 
Secobarbital 91.7 (80.0–101.3) 105.1 (99.4–111.5) NA 
Oxazepam 95.0 (83.0–105.8) 123.7 (116.5–129.6) NA 
Buprenorphine 102.0 (88.6–114.1) 105.1 (98.8–111.5) NA 
Benzoylecgonine 144.8 (128.4–167.8) 102.8 (96.7–109.0) 101.1 (93.7–107.8) 
MDMA 68.3 (60.1–76.8) 107.3 (102.0–114.3) NA 
THC 85.2 (74.1–91.6) 78.4 (74.3–83.1) NA 
Methamphetamine 102.2 (89.7–112.8) 100.5 (93.9–106.5) 89.2 (82.3–95.6) 
Methadone 125.1 (109.2–142.3) 100.3 (93.3–106.6) NA 
Morphine, Low 74.9 (65.9–84.2) 84.9 (79.1–93.2) NA 
Morphine, High NA NA 101.9 (95.5–107.9) 
Nortriptyline 112.8 (99.4–123.8) 127.5 (122.0–132.9) NA 
Drugs (Drug group) Percentage of SAMHSA cut-off level, % (95% confidence interval)a 
DxLINK E-Z Cup Multi4 card 
Amphetamine 88.2 (77.2–98.7) 97.8 (90.5–104.2) NA 
Secobarbital 91.7 (80.0–101.3) 105.1 (99.4–111.5) NA 
Oxazepam 95.0 (83.0–105.8) 123.7 (116.5–129.6) NA 
Buprenorphine 102.0 (88.6–114.1) 105.1 (98.8–111.5) NA 
Benzoylecgonine 144.8 (128.4–167.8) 102.8 (96.7–109.0) 101.1 (93.7–107.8) 
MDMA 68.3 (60.1–76.8) 107.3 (102.0–114.3) NA 
THC 85.2 (74.1–91.6) 78.4 (74.3–83.1) NA 
Methamphetamine 102.2 (89.7–112.8) 100.5 (93.9–106.5) 89.2 (82.3–95.6) 
Methadone 125.1 (109.2–142.3) 100.3 (93.3–106.6) NA 
Morphine, Low 74.9 (65.9–84.2) 84.9 (79.1–93.2) NA 
Morphine, High NA NA 101.9 (95.5–107.9) 
Nortriptyline 112.8 (99.4–123.8) 127.5 (122.0–132.9) NA 

Data are presented as cut-offs. Abbreviations: NA, not assessed.

aPercentage of target cut-off value as indicated in Table I.

Evaluation of color interpretation criteria

We evaluated the performance of color interpretation by visual inspection by comparing against a standard color-scale card provided by the manufacturer (Alere, Abon Biopharm). Based on the data provided by the manufacturer, concentrations at G3 or below were recorded as positive and interpreted as indicative of the presence of drugs within a sample at a concentration above the cut-off level, whereas concentrations above G3.5 were recorded as negative. Changes in the sensitivity and specificity of the drug-test results were assessed based on various criteria associated with positive results for different color grades, such as those below G2, below G3.5 or below G4. The results are summarized in Table V. Moreover, the color grades expected to have the highest sensitivity and specificity for each drug in these tests were also examined via ROC analysis. The E-Z Cup had the highest sensitivity and specificity for amphetamine, secobarbital, benzoylecgonine and methamphetamine when concentrations below G3 were interpreted as positive based on the reference criteria. The device exhibited the highest sensitivity and specificity for THC, methadone and morphine when concentrations below G2 were interpreted as positive, for buprenorphine and MDMA when concentrations below G3.5 were interpreted as positive, and for oxazepam and nortriptyline when concentrations below G4 were interpreted as positive (Table V). For the Multi4 card, results were most consistent with true values when concentrations below G3 were read as positive for benzoylecgonine, methadone and high-level morphine, whereas higher sensitivity for THC was observed when reading concentrations below G4 as positive. With the positive and negative results reclassified based on these new criteria, very high sensitivity and specificity of greater than 90% and 80% (range, 90.6–100% and 87.4–98.7%), respectively, were observed for all drugs. When a less stringent cut-off of G3.5 was applied for oxazepam, buprenorphine, MDMA and nortriptyline, for which >98% specificity can be achieved, the sensitivity for these drugs increased from 70–80% to 90–100% (71.4–89.5%, 85.7–100%, 83.8–97.1% and 68.6–88.6% sensitivity from G3 to G3.5, respectively). A high AUC of greater than 0.98 was also observed for most drugs. There were no statistically significant differences in color grade (P = 0.513) or percentages of positive results (P = 0.848) between the two different assay lots evaluated with the E-Z Cup.

Table V.

Sensitivity and Specificity According to Different Color Grade Criteria

Tests AUC G2a G3 G3.5 G4 
Sen (%) Spe (%) Sen (%) Spe (%) Sen (%) Spe (%) Sen (%) Spe (%) 
E-Z Cup 
 Amphetamine 0.982 87.1b 96.8b 98.8c 87.4c 100 69.5 100 60.0 
 Secobarbital 0.995 81.2 100 100c 94.8c 100 76.8 100 63.2 
 Oxazepam 0.993 50.4 100 71.4 100 89.5b 98.7b 100c 89.3c 
 Buprenorphine 0.999 61.9 100 85.7 100 100c 98.7c 100 76.0 
 Benzoylecgonine 0.994 83.5 100 100c 92.6c 100 69.5 100 64.2 
 MDMA 0.996 58.1 100 83.8 100 97.1c 98.7c 100 82.7 
 THC 0.988 98.5c 93.9c 100 77.4 100 59.1 100 53.0 
 Methamphetamine 0.992 83.5b 100b 100c 90.5c 100 70.5 100 64.2 
 Methadone 0.988 95.3c 95.8c 97.7 88.4 100 66.3 100 60.0 
 Morphine 0.945 90.6c 87.4c 100 77.9 100 68.4 100 58.9 
 Nortriptyline 0.992 53.3 100 68.6 100 88.6b 98.7b 98.1c 92c 
Multi4 card 
 Benzoylecgonine 0.986 85.9b 97.9b 97.7c 89.5c 100 78.9 100 58.9 
 THC 0.935 6.0 99.2 36.0 97.7 50.0b 96.2b 92c 87.7c 
 Methamphetamine 0.984 90.6b 95.8b 100c 81.1c 100 63.2 100 42.1 
 Morphine, High 0.992 87.1 99.0 100c 91.6c 100 76.8 100 56.8 
Tests AUC G2a G3 G3.5 G4 
Sen (%) Spe (%) Sen (%) Spe (%) Sen (%) Spe (%) Sen (%) Spe (%) 
E-Z Cup 
 Amphetamine 0.982 87.1b 96.8b 98.8c 87.4c 100 69.5 100 60.0 
 Secobarbital 0.995 81.2 100 100c 94.8c 100 76.8 100 63.2 
 Oxazepam 0.993 50.4 100 71.4 100 89.5b 98.7b 100c 89.3c 
 Buprenorphine 0.999 61.9 100 85.7 100 100c 98.7c 100 76.0 
 Benzoylecgonine 0.994 83.5 100 100c 92.6c 100 69.5 100 64.2 
 MDMA 0.996 58.1 100 83.8 100 97.1c 98.7c 100 82.7 
 THC 0.988 98.5c 93.9c 100 77.4 100 59.1 100 53.0 
 Methamphetamine 0.992 83.5b 100b 100c 90.5c 100 70.5 100 64.2 
 Methadone 0.988 95.3c 95.8c 97.7 88.4 100 66.3 100 60.0 
 Morphine 0.945 90.6c 87.4c 100 77.9 100 68.4 100 58.9 
 Nortriptyline 0.992 53.3 100 68.6 100 88.6b 98.7b 98.1c 92c 
Multi4 card 
 Benzoylecgonine 0.986 85.9b 97.9b 97.7c 89.5c 100 78.9 100 58.9 
 THC 0.935 6.0 99.2 36.0 97.7 50.0b 96.2b 92c 87.7c 
 Methamphetamine 0.984 90.6b 95.8b 100c 81.1c 100 63.2 100 42.1 
 Morphine, High 0.992 87.1 99.0 100c 91.6c 100 76.8 100 56.8 

Abbreviations: Sen, sensitivity; Spe, specificity.

aValues equal or below a specific grade (G2, grade 2; G3, grade 3; G3.5, grade 3.5; and G4 grade 4) were recorded as positive.

bCut-offs for >95% specificity should be obtained separately for optimal sensitivity.

cSensitivity and specificity of the test according to optimal color grade cut-offs for determining positivity.

Discussion

This is the first study to compare the performance of a newly developed on-site drug-testing devices with automated reader as compared with color interpretation-based immunoassays which are capable of simultaneously detecting 11 drug classes. Mass spectrometry-based measurements of control material concentrations showed that 10 of the 11 drugs examined in this study were measured within 15% of certified concentrations. However, the concentrations of THC were significantly lower than the expected concentrations described by the manufacturers. Previous studies have reported that the concentration of THC can substantially decrease when it is stored in a plastic vial (12, 13). The THC used in this study was shipped from the manufacturer to our laboratory and to a reference laboratory in plastic vials, which inevitably led to several freeze–thaw cycles during transportation and over the course of the experiments. Therefore, breakdown of THC during storage may have decreased its concentration and caused the difference between the two results. As such, THC was excluded from the agreement analysis because reliable test results are expected to be difficult to obtain due to the instability of the material.

In this study, all three on-site drug-testing devices exhibited mostly acceptable results at drug concentrations around the cut-off level. Greater than 90% positive agreement was observed for most classes of drugs at the +50% concentration of each cut-off level. However, the DxLINK showed only 60% positive agreement for benzoylecgonine. This result was lower than the data claimed by the manufacturer, where the positive agreement was 100%. Evaluations at the cut-off level showed that the cut-off level of benzoylecgonine was higher than the SAMHSA cut-off level for cocaine. Based on these results, it is suggested that greater attention is needed when interpreting drug-test results for cocaine generated by the DxLINK. When suspected drug abusers present negative results, retesting using a more sensitive assay for cocaine may be needed. Greater than 80% positive agreement was observed at +25% of each cut-off level for most drug classes, but the DxLINK exhibited >50% false negatives for benzoylecgonine and methadone, whereas the E-Z Cup exhibited >50% false negatives for nortriptyline. However, in most real testing situations, fewer specimens with drug concentrations near cut-off levels will be encountered (14). Therefore, all three devices are expected to exhibit better performance in the field.

For the negative specimens, at −50% of each cut-off level, all three devices yielded 100% agreement. The devices also showed greater than 80% agreement for most drug classes at −25% of each cut-off level. In particular, the E-Z Cup exhibited greater than 95% negative agreement for all drug classes at −25%. For the DxLINK, MDMA and morphine resulted in a false-positive rate of 50%. In specimens from real drug abusers, such false-positive rates may be attributed to the cross-reactivity of narcotic-class drugs (3, 1517). However, the results of this study suggest that the false-positive rates observed for some drug classes may be attributable to difficulties in interpreting test lines rather than to cross-reactivity because the control materials used in this study contain no known cross-reactivity. The presence of a faint test line G3.5 or above in the E-Z Cup and Multi4 card devices was interpreted as a negative result as described by the manufacturers. Due to the nature of visual reading, test lines may appear very faint or almost invisible when drugs are present at concentrations near the cut-off level, which may cause some confusion in the interpretation of test results and ultimately increase the false-positive rate. However, our results indicate that for most drugs it is possible to obtain a relatively accurate reading near the cut-off. Compared to the test results from visual reading of the E-Z Cup and Multi4 card devices, the DxLINK, which employs automated scanning and reading of test lines, did not exhibit superior agreement. Rather, it yielded a greater number of false positives for morphine and MDMA within −25% of the cut-off level. These results suggest that, for the DxLINK, morphine and MDMA cases where a very faint test line appears at a concentration slightly lower than the cut-off should be considered positive.

With regard to the accuracy of the randomly tested samples, both the E-Z Cup and Multi4 card showed greater than 90% agreement. Overall, the accuracy of the DxLINK exhibited greater than 90% agreement; however, greater than 50% false-negative results were obtained for benzoylecgonine at +15% of the cut-off and greater than 10% false-negative results were obtained for methadone and buprenorphine, which indicates lower agreement.

Re-evaluation of cut-off levels using measurements from control materials confirmed a ±15% deviation relative to the established cut-off level for most drugs. However, for the DxLINK, the cut-off levels for MDMA and morphine decreased, whereas the cut-off levels for methadone and cocaine increased. The E-Z Cup exhibited increased cut-off levels for oxazepam and nortriptyline; compared to a previous evaluation (18), these results are consistent with reports that an increase in the limit of detection (LOD) of up to 150% of the SHAMSA cut-off for methadone and oxazepam can be found with this device. In contrast, while previous results showed increases in the LOD for morphine, secobarbital and THC of 125%, no clear trend was observed in this study. The results for benzodiazepine are not comparable because this drug was not previously evaluated. Additionally, the results for amphetamine and methamphetamine obtained in previous works and the current study are difficult to compare because in the study by Greene et al. the compounds were evaluated as a d,l-amphetamine and d,l-methamphetamine racemic mixture, whereas our study used the pure d-form of the compounds, which are the illegal isoforms of the drugs. d,l-amphetamine and l-methamphetamine have higher cut-off values than the d-forms of these compounds (18).

Another previous study comparing four on-site drug-testing devices using urine specimens (19) from the Triage Panel for Drug Abuse Plus TCA (Biosite Diagnostics, San Diego, CA) exhibited the best results for seven classes of drugs, producing 100%, 70% and 80% positive agreement for five drug classes, secobarbital and oxazepam, respectively, in control materials within +25% of the cut-off level. Additionally, 95–100% sensitivity was observed for the drug-supplemented urine specimens. The Rapid Drug screen had low sensitivity, showing a 20–85% sensitivity distribution, for the drug-supplemented urine specimens, indicating a difference in some measures (19). When we compared control materials with +25% cut-off levels with the Triage results, the DxLINK presented equivalent sensitivity for d-amphetamine and morphine, slightly higher sensitivity for secobarbital and oxazepam, and lower sensitivity for benzoylecgonine. The E-Z Cup presented slightly higher sensitivity for secobarbital and lower sensitivity for oxazepam. Both the DxLINK and E-Z Cup can be considered advantageous over the Triage on-site drug-testing kit, which permits simultaneous testing of up to 7 drug classes, as the former permit simultaneous drug testing of up to 11 drug classes.

The devices evaluated in this study all utilize multiplexed assays that simultaneously test multiple drugs. This type of assay also enables testers to perform multiple tests simultaneously. However, intrinsic limitations of multiplexed assays include the difficulty of calibrating drug classes and potential interactions. In addition, the interpretation of very faint or almost invisible lines near a cut-off level can be difficult. Considering these characteristics, in this study, we interpreted visible inspection results using a standardized color grade for various drug classes. Evaluations of the E-Z Cup and Multi4 card devices involved visual inspection of color grades, and the highest sensitivity and specificity were obtained with the control material concentrations. Many drugs produced identical sensitivity and specificity using cut-off of ≤G3, as suggested based on the evaluation data provided by the manufacturer, whereas increased sensitivity and specificity were obtained using varied criteria, such as ≤G2 or ≤G4, for some drug classes. Although the manufacturer's instructions indicate that positive results are obtained when no line is visible, it may be confusing for test readers when a line is very faint. Therefore, using finely calibrated standards for visual interpretation and thorough training of test readers may help improve the performance of devices that allow test results to be read on-site.

In summary, the DxLINK and E-Z Cup devices exhibited differences in overall performance for the detection of varied drug classes, with somewhat high false-positive and false-negative rates. This suggests a need for careful interpretation when using these two devices. Because cocaine testing using the DxLINK presented rather low sensitivity, when suspected drug abusers present negative results with this device, retesting using a more sensitive assay may be required. Moreover, confirmation of a specimen that initially tests positive by a drug-testing device through a more specific test using mass spectrometry must be performed because these devices lack 100% specificity (15). In Korea, the prevalence of drug abusers is <1%, and the positive predictive value of such devices is lowered in low-prevalence settings, such as in random screening. Additionally, there might be differences in sensitivity and specificity among different drug classes; thus, for drug-testing devices that require visual inspection such as the E-Z Cup, an increase in sensitivity and specificity can be expected when applying varied color grade criteria for different drug classes during on-site reading. For the DxLINK, automated reading prevents users from obtaining additional information but can help reduce reader subjectivity during visual inspection. Additionally, results are read quickly and clerical errors are reduced.

The present study investigated whether on-site drug-screening kits are capable of detecting multiple drugs simultaneously. Despite several previous studies examining on-site drug testing, on-site drug tests capable of simultaneously identifying >10 classes of drugs have rarely been investigated. Furthermore, mass spectrometry was applied to determine the true concentrations of control materials to increase test reliability. Moreover, the results of the devices were compared to directly measured values.

Conclusion

The performance of three on-site drug-testing devices, the DxLINK, E-Z Cup, and Multi4 card, was evaluated in this study. All three devices exhibited satisfactory performance with 100% specificity at −50% cut-off levels and >90% sensitivity for most of the drug classes studied, with the exception of the 60% sensitivity found for benzoylecgonine for the DxLINK at +50% cut-off levels. The DxLINK presented less sensitive screening results for cocaine and methadone than the other devices; however, it also was more sensitive for oxazepam and nortriptyline than the E-Z Cup. Overall, these devices can simultaneously test many classes of drugs with performance comparable to that of previously evaluated devices. In addition, the devices are appropriate for on-site drug screening because they are easy to use and results can be obtained within 5–10 minutes.

Supplementary data

Authors’ Contributions

G.C.K. designed the study, performed data management and statistical analysis, and helped write the manuscript. Y.P. and J.K. performed the experiments and shared responsibility for data management and statistical analysis. S.Y.K. and H.K. shared responsibility for data management, statistical analysis and writing the manuscript. J.K. and S.H.K. shared responsibility for study design, data interpretation and manuscript revision for important intellectual content. All authors read and approved the final manuscript.

References

1
Fedotov
,
Y.
United Nation Office on Drug and Crime, World Drug Report 2014. http://www.unodc.org (accessed June 15, 2014).
2
Korean Association Against Drug Abuse
. (
2015
) Statistics on Drug Abuse reported by the South Korean National Police Agency 2015. http://www.drugfree.or.kr/information/index.html?contentsNum=4&headNum=2 (accessed June 30, 2015).
3
Taylor
,
E.H.
,
Pizzo
,
P.
(
2004
)
Evaluation of the DrugCheck 9 on-site immunoassay test cup according to a standard method validation protocol
.
Journal of Analytical Toxicololy
 ,
28
,
190
197
.
4
Buechler
,
K.F.
,
Moi
,
S.
,
Noar
,
B.
,
McGrath
,
D.
,
Villela
,
J.
,
Clancy
,
M.
, et al.
(
1992
)
Simultaneous detection of seven drugs of abuse by the Triage panel for drugs of abuse
.
Clinical Chemistry
 ,
38
,
1678
1684
.
5
Koch
,
T.R.
,
Raglin
,
R.L.
,
Kirk
,
S.
,
Bruni
,
J.F.
(
1994
)
Improved screening for benzodiazepine metabolites in urine using the Triage panel for drugs of abuse
.
Journal of Analytical Toxicology
 ,
18
,
168
172
.
6
Rohrich
,
J.
,
Schmidt
,
K.
,
Bratzke
,
H.
(
1994
)
Application of the novel immunoassay TRIAGE to a rapid detection of antemortem drug abuse
.
Journal of Analytical Toxicology
 ,
18
,
407
414
.
7
Jenkins
,
A.J.
,
Darwin
,
W.D.
,
Huestis
,
M.A.
,
Cone
,
E.J.
,
Mitchell
,
J.M.
(
1995
)
Validity testing of the accuPINCH THC test
.
Journal of Analytical Toxicology
 ,
19
,
5
12
.
8
Towt
,
J.
,
Tsai
,
S.C.
,
Hernandez
,
M.R.
,
Klimov
,
A.D.
,
Kravec
,
C.V.
,
Rouse
,
S.L.
, et al.
(
1995
)
ONTRAK TESTCUP: a novel, on-site, multi-analyte screen for the detection of abused drugs
.
Journal of Analytical Toxicology
 ,
19
,
504
510
.
9
George
,
S.
,
Braithwaite
,
R.A.
(
1995
)
A preliminary evaluation of five rapid detection kits for on site drugs of abuse screening
.
Addiction (Abingdon, England)
 ,
90
,
227
232
.
10
Crouch
,
D.J.
,
Hersch
,
R.K.
,
Cook
,
R.F.
,
Frank
,
J.F.
,
Walsh
,
J.M.
(
2002
)
A field evaluation of five on-site drug-testing devices
.
Journal of Analytical Toxicology
 ,
26
,
493
499
.
11
Gronholm
,
M.
,
Lillsunde
,
P.
(
2001
)
A comparison between on-site immunoassay drug-testing devices and laboratory results
.
Forensic Science International
 ,
121
,
37
46
.
12
Buchan
,
B.J.
,
Walsh
,
J.M.
,
Leaverton
,
P.E.
(
1998
)
Evaluation of the accuracy of on-site multi-analyte drug testing devices in the determination of the prevalence of illicit drugs in drivers
.
Journal of Forensic Sciences
 ,
43
,
395
399
.
13
Christophersen
,
A.S.
(
1986
)
Tetrahydrocannabinol stability in whole blood: plastic versus glass containers
.
Journal of Analytical Toxicology
 ,
10
,
129
131
.
14
Kadehjian
,
L.J.
(
2001
)
Performance of five non-instrumented urine drug-testing devices with challenging near-cutoff specimens
.
Journal of Analytical Toxicology
 ,
25
,
670
679
.
15
Moeller
,
K.E.
,
Lee
,
K.C.
,
Kissack
,
J.C.
(
2008
)
Urine drug screening: practical guide for clinicians
.
Mayo Clinic Proceedings
 ,
83
,
66
76
.
16
Merigian
,
K.S.
,
Browning
,
R.
,
Kellerman
,
A.
(
1993
)
Doxepin causing false-positive urine test for amphetamine
.
Annals of Emergency Medicine
 ,
22
,
1370
.
17
Merigian
,
K.S.
,
Browning
,
R.G.
(
1993
)
Desipramine and amantadine causing false-positive urine test for amphetamine
.
Annals of Emergency Medicine
 ,
22
,
1927
1928
.
18
Greene
,
D.N.
,
Lehman
,
C.M.
,
McMillin
,
G.A.
(
2011
)
Evaluation of the integrated E-Z split key((R)) cup II for rapid detection of twelve drug classes in urine
.
Journal of Analytical Toxicology
 ,
35
,
46
53
.
19
Peace
,
M.R.
,
Tarnai
,
L.D.
,
Poklis
,
A.
(
2000
)
Performance evaluation of four on-site drug-testing devices for detection of drugs of abuse in urine
.
Journal of Analytical Toxicology
 ,
24
,
589
594
.

Author notes

Seon Young Kim and Hyunjin Kim contributed equally as first authors.

Supplementary data