Anna Pehrsson, Tom Blencowe*, Kari Vimpari, Kaarina Langel, Charlotta Engblom, and Pirjo Lillsunde

Transcription

1 An Evaluation of On-Site Oral Fluid Drug Screening Devices DrugWipe 5 + and Rapid STAT Using Oral Fluid for Confirmation Analysis Anna Pehrsson, Tom Blencowe*, Kari Vimpari, Kaarina Langel, Charlotta Engblom, and Pirjo Lillsunde National Institute for Health and Welfare, Drug Analytics Unit, P.O. Box 30, FI Helsinki, Finland Abstract In this study, the performance of two on-site oral fluid drug-testing devices, DrugWipe 5 + (Securetec) and Rapid STAT (Mavand), was assessed. The results obtained by the devices were compared with gas chromatography mass spectrometry confirmation analysis results in oral fluid. Sensitivity, specificity, and accuracy of the tests, as well as positive and negative predictive values, were calculated based on the classified results of the comparison. Both of the devices were evaluated for their ability to meet toxicological cutoffs as set in the DRUID project ( as well as those reported by the manufacturers. The evaluation was performed for relevant drug groups of both devices: amphetamines, cannabis, cocaine, and opiates. Additionally, Rapid STAT has a test for benzodiazepines included in the same device. Both tests seemed to perform quite well for amphetamines although they also gave negative results for cases with high concentrations. Also, the benzodiazepine test of Rapid STAT was at a relatively good level although only half of the positive test results were true positives using the test cutoffs. The same phenomenon was detected for the cannabis tests of both devices. The proper evaluation of cocaine and opiates tests was not applicable because of the very low number of positive cases. Introduction Driving under the influence of drugs (DUID) is known to be a serious threat to public health and road traffic safety. In the past, selective police enforcement on suspected drugged drivers has been performed in a limited number of EU member states, mostly using urine-based screening devices. The experiences of the police have shown that performing on-site urine screening is very complicated at the roadside, often potentially intrusive or vulnerable to adulteration. In addition, a positive urine test * Author to whom correspondence should be addressed. National Institute for Health and Welfare, Drug Analytics Unit, P.O. Box 30, FI Helsinki, Finland. may indicate exposure to a drug several days previously since the analytes accumulate in urine. Detection of drugs in oral fluid (OF) indicates more recent drug use and this matrix is suitable for inexpensive, noninvasive, and easy-to-use diagnostic aids for the detection of the illicit use of drugs (1,2). Onsite screening for drugs of abuse in OF provides a relatively quick and increasingly effective means of detecting if a motorist has consumed drugs or medicines that are of concern to traffic safety. Nevertheless, a positive OF on-site screening result can only be interpreted as a possible indication of drug use rather than definitive proof. Correlations between presence of drugs of abuse in OF and pharmacological and physiological effects have been reported (3,4). The state of impairment of a drug user is related to the physiologically active fraction of the substance in question. Only the free, physiologically active, component of a drug is transferred from blood to OF. In whole blood many drugs are highly bound to blood proteins and this fraction of the drug is physiologically inactive (5,6). Weakly basic drugs, on the other hand, are detected in OF in higher concentrations than in blood. This results from the relative acidity of OF compared to plasma and consequent ion trapping (7,8). For many drugs, transfer of the free component from blood to OF occurs via simple passive diffusion (1), although transfer is affected by various factors such as salivary ph and the characteristics of the drug (e.g., acidic or basic, lipid solubility). Thus the presence of a substance in OF may correspond to its presence in blood (9) and OF is a potential alternative matrix as a confirmation sample for toxicological analysis. However, knowledge concerning the OF/blood ratios for drugs remains incomplete and results of previous comparisons of blood and OF analyses vary quite widely (10). Nonetheless, in the Rosita-2 project, which evaluated the usability and analytical reliability of on-site OF drug testing devices in , a comparison of results from laboratory analyses in OF and blood concluded that OF is a good screening fluid for the presence or absence of amphetamines, cannabis, cocaine, and opiates in blood (11). In this publication, the reliability of two on-site OF screening devices, DrugWipe 5 + (Securetec Detektions-Systeme AG, Reproduction (photocopying) of editorial content of this journal is prohibited without publisher s permission. 211

2 Munich, Germany) and Rapid STAT (Mavand Solutions GmbH, Mössingen, Germany), is evaluated by comparing the laboratory OF results with those of the device. In addition to the Rosita-2 project (11) and the preceding Rosita project (12), the performance of DrugWipe devices has been evaluated in several studies (9,13 21). These evaluations were based on widely varying study protocols, (e.g., there were some differences in result interpretation and a large variation in the number and/or type of cases/samples tested). There are hardly any evaluations (22) for Rapid STAT since it is a relatively new OF screening device. This study was produced under the on-going EU project Driving under the Influence of Drugs, Alcohol and Medicines (DRUID). Methods Sample collection Altogether 221 subjects participated in the study on voluntary basis. An ethical approval for the study was obtained from the coordinative ethical committee of the Hospital District of Helsinki and Uusimaa, Finland. A written consent was obtained from all of the participants. A total of 136 subjects were tested with the DrugWipe 5 + and 132 with the Rapid STAT. Subjects were recruited from real police DUID cases (n = 58), a roadside study (n = 83), a rehabilitation clinic (n = 21) and the personnel of the Drug Analytics Unit (n = 59). An OF sample was collected from all the subjects for confirmation analysis. Participants suspected of DUID were taken to the Police Surgeon Station for blood sampling and clinical evaluation by a physician as part of normal police procedure. The on-site tests and collection of the OF samples were performed by the police officer or by the physician. For the nonsuspected drivers or the laboratory personnel, the screening tests and OF sample collection were conducted by researchers for the DRUID project. Clinic nurses tested the subjects and collected their OF samples at the rehabilitation clinic. All OF samples were frozen ( 20 C) until analysis. For OF collection, the Saliva Sampler (StatSure Diagnostic Systems, Framingham, MA) device was used. This device was chosen for sample collection based on a previous study (23). The device consists of an absorptive cellulose pad with a volume adequacy indicator and a plastic tube containing buffer solution. The window of the stem turns blue when 1 ml of OF is collected. When collecting the sample, OF is first allowed to gather in the mouth. The collection pad is placed under the tongue and removed when the indicator window has turned completely blue. The pad is then placed into the collection tube. In the laboratory, the collection pad is disconnected from the stem and dropped to the bottom of the tube, and a filter is inserted into the tube to recover the OF-buffer solution. On-site screening devices Both devices are lateral flow immunoassays in which each analyte is represented by a separate line in the test window. The DrugWipe 5 + device comprises an OF collector, a detection element, and an integrated liquid ampoule. First, the OF collector is detached from the test body. The tongue or the cheek of the tested person is wiped, and the collector is then reattached to the test body. The ampoule is pressed to let the buffer solution flow to the test strips. The results can be read when red control lines indicating a successful test appear on both strips. Positive screening results are indicated by red lines in the designated positions. The total time required for testing is 3 10 min. The Rapid STAT device comprises a collection stick, an aroma field to increase salivation, a buffer solution and a test strip. OF is collected by rotary movements of the microfiber collector stick inside the cheeks and gums. The collection stick is then put into the buffer bottle and agitated before removal. Seven drops of the buffer fluid mixture are pipetted to each well of the test device. The lid is closed to the first position and left for 4 min. The test is started by pressing down the lid completely to let the buffer flow to the test strips. Red control lines indicate a successful test. A positive screening result is indicated by the absence of a red line in the designated positions. The results should be read within 8 min of the buffer flowing; the total time needed for testing is 7 12 min. Table I. List of Analyzed Substances and Their Cutoffs in Oral Fluid Used for the Confirmation Analysis Substance Cutoff (µg/l) 6-Acetylmorphine (6-AM) 5 Alprazolam 1 Amphetamine 25 Benzoylecgonine 10 Bromazepam 5 Chlordiazepoxide 10 Clonazepam 1 Cocaine 10 Codeine 20 Diazepam 5 Flunitrazepam 1 Lorazepam 1 MDA 25 MDEA 25 MDMA 25 Methamphetamine 25 Midazolam 2 Morphine 20 Nitrazepam 2 Nordiazepam 1 Oxazepam 5 Temazepam THC 1 212

3 Substances analyzed and cutoffs All OF confirmation samples collected for the study were analyzed for a range of substances. The substances analyzed and the corresponding cutoffs are presented in Table I. These cutoffs were set for a separate epidemiological study within the DRUID project. They are typically lower than the screening device cutoffs (e.g., 9 -THC) or at an equivalent level (e.g., amphetamines). These cutoffs were set to allow optimal detection of positive DUID cases for all countries involved in the project. The drug screening performance of the devices were therefore evaluated against these DRUID cutoffs as well as the manufacturers stated cutoffs. Analytical methods The StatSure Saliva Sampler collection device is designed to collect 1 ml of OF, and the device also includes 1 ml of a buffer solution. Therefore, the concentrations of substances found in the OF sample were determined using a calibration curve, which was based on the assumption that the OF-tobuffer ratio is 1:1. The standards used for the calibration curve were adjusted accordingly. In reality, the amount of collected OF in the device will vary to some extent. This can be because of intrinsic variation in the collection volume when the indicator turns blue or, for example, if the subject is unable to provide a full 1 ml of OF due to a dry mouth. Thus, it was necessary to adjust the concentration of analyte determined in analysis according to the actual volume of collected OF. This was calculated using the following formula (Eq. 1): C corrected = C uncorrected (1 + w w ) Eq. 1 2 (w w ) wherein w = average weight of empty StatSure device; w = weight of sample and StatSure device; C uncorrected = uncorrected concentration of analyte; and C corrected = concentration of analyte corrected for volume of oral fluid collected. The density of the buffer OF mixture was assumed to be 1 g/ml for this adjustment. Samples were prepared by taking 1 ml of thawed sample (OF + buffer) for analysis. Standards for analysis were made by spiking 0.5 ml of blank OF with stock solution containing all analytes. For the standards preparation, blank OF was collected from laboratory personnel, frozen and thawed. Deuterated analogues of the analytes were used as internal standards. Standards were prepared by adding 0.5 ml of StatSure buffer to 0.5 ml standards in OF. Analysis of illicit and medicinal drugs was carried out using gas chromatography electron impact ionization-mass spectrometry (GC EI-MS) or GC negative chemical ionization (NCI)-MS. A full description of sample preparation and analysis and validation results were published previously (24). The laboratory participated in an OF external quality control program organized by RTI International (Research Triangle Park, NC) for all the DRUID project toxicology laboratories. Based on the results the performance of all the OF analytical methods was at an excellent level (25). Interpretation of screening test results Based on the GC MS confirmation results of the OF samples and the DRUID cutoffs, the results were classified as true positive (TP), true negative (TN), false positive (FP), and false negative (FN). Using these classifications, several parameters for the evaluation can be calculated, namely sensitivity, specificity, accuracy and prevalence. Also, positive predictive value (PPV) and negative predictive value (NPV) were calculated according to the Bayesian method. The equations used in these calculations are presented in Table II. For PPV and NPV calculations, the prevalence of the substances in question among the DUID suspected population in Finland in the year 2008 were used. These were taken from the database of the Drug Analytics Unit. Altogether 4419 cases were investigated in that year. It should be noted that these prevalence values were calculated from blood sample results and are not based on OF. The most common findings were sleep-inducing and sedative substances 59% and amphetamines 56%. Other findings were cannabis 20%, methamphetamine 8.0%, morphine 0.9%, and cocaine 0.7%. Evaluation of results The devices included in this study were evaluated according to two sets of criteria. The first criteria were the DRUID cutoffs for the relevant substances (Table I). For this evaluation, each test result was interpreted as TP, TN, FP, or FN according to the concentration of the individual substances detected in the confirmation sample. So, for example, with an amphetamines test detecting amphetamine, methamphetamine, MDA, and MDMA, the investigated case was interpreted as TP if any of the detected substances was found at a concentration above the relevant DRUID cutoff value (25 µg/l in OF for each substance). However, a positive test case for which concentrations of, for example, both amphetamine and methamphetamine were found at a level only slightly lower than the DRUID cutoff (i.e., both only just below 25 µg/l) would be interpreted as FP. Thus, no cross-reactivity was taken into account. The second set of criteria was the cutoffs and cross-reactivities of the devices claimed by the manufacturers. For both devices, the manufacturers provided information regarding the cutoff concentrations for substances at which the device could be expected to give a positive result (Table III). Table II. Equations Used for the Device Performance Evaluation Parameter PPV NPV Calculation* sens prev sens prev + (1 spec) (1 prev) spec (1 prev) spec (1 prev) + prev (1 spec) * Abbreviations: sens, sensitivity; spec, specificity; and prev, prevalence. 213

4 In reality, the presence of two or more substances, both relevant to a specific test (e.g., amphetamines) can give rise to cross-reactivity, wherein the combined concentration of the two substances is sufficient that it can be expected to give a positive result for the test. However the individual concentrations at which each substance alone is expected to give a positive result can differ within each drug group; therefore the combined concentration is not a straightforward sum of each relevant substance concentration, rather it is necessary to take the individual cutoffs for each analyte into account when calculating the combined concentration. This can be done by means of a cross-reactivity equation. For example, Table IV shows the cutoffs and resultant cross-reactivities for the Drug- Wipe amphetamines test. The combined concentration of analytes for this example (cf. Table IV) may hence be calculated by the following equation: c(amphetamines) = 100% c(amp) + 200% c(mamp) + 200% c(mdma) + 200% c(mda) + 77% c(mdea) This combined concentration of the amphetamines group analytes is directly comparable to the target compound cutoff of 50 µg/l for amphetamine. The terms c(amp), c(mamp), c(mdma), c(mda), and c(mdea) refer to the confirmed concentrations of amphetamine, methamphetamine, MDMA, MDA, and MDEA, respectively. Using the manufacturer cutoffs and derived cross-reactivity equations the on-site test results can be directly compared to the concentrations of substances found in the confirmation sample to ascertain whether the case is classified as TP, TN, FP, or FN. Results DrugWipe 5 + The results of the test evaluation are shown in Table V. Amphetamine-type stimulant drugs. Altogether 38 cases (28%) had amphetamine-type stimulants (ATS) detected in their OF. Amphetamine was found in all the cases that were positive for ATSs. The concentration range for amphetamine was µg/l. Methamphetamine was found in 11 cases (8.1%) for which the range was ,700 µg/l. In one case, MDA (281 µg/l) and MDMA (4850 µg/l) were found in addition to amphetamine. MDEA was not found in any of the samples. The combined concentration of the relevant analytes for the DrugWipe 5 + amphetamines test, directly comparable to the manufacturers stated cutoff for the target compound amphetamine, was calculated using the equation presented in the section Evaluation of results. There were five FN cases, all of which contained only amphetamine from the ATS group. The concentration range for these five cases was µg/l. This concentration range is clearly a lot higher than both the DRUID and manufacturer cutoffs, but still relatively low in comparison to the overall range of concentrations found for amphetamine. The ATS concentration, taking into account cross-reactivity, for more than 80% of the TP cases was higher than 3620 µg/l (the maximum amphetamine concentration found in the FN cases). There were also 5 FP cases observed, each of which contained no ATS drugs. Cannabis. There were 21 (16%) cases which were 9 -THC positive in the OF sample, the concentration range was µg/l. Only 9 -THC was measured in the study hence no cross-reactivity was taken into account. There were 12 FN cases according to the DRUID cutoff. The concentration range found in these cases was µg/l. For comparison, the corresponding range for 9 TP cases was µg/l. It appeared that there was no clear distinction in concentration between the cases that tested 9 -THC negative and those that tested positive. The manufacturer launched a new version of the DrugWipe 5 + test with an enhanced test strip for cannabis during the testing period. In this study, 12 tests with the improved cannabis test strip were made. All of the cases tested negative for Table III. Detected Drug Groups and their Cutoff Values for the DrugWipe 5+ and Rapid STAT Devices Securetec DrugWipe 5+ Mavand Rapid STAT Detected drug group Cutoff Detected drug group Cutoff (target compound) (µg/l) (target compound) (µg/l) Amphetamines 50 Amphetamines 25 (d-amphetamine) (d-amphetamine) Methamphetamine 25 Methamphetamine 25 (d-methamphetamine) MDMA 25 MDMA 50 Cocaine 30 Cocaine 12 (benzoylecgonine) (benzoylecgonine) Opiates (codeine) 10 Opiates (morphine) 25 Cannabis ( 9 -THC) 30 Cannabis ( 9 -THC) 15 Benzodiazepines 25 (oxazepam) Table IV. Substance Cutoffs and Cross-Reactivities for the DrugWipe 5+ Amphetamines Test Cutoff in OF Cross-Reactivity Substance (µg/l) (%) Amphetamine (AMP)* Methamphetamine (MAMP) MDMA MDA MDEA * Target compound of test. 214

5 cannabis. However, one case contained 6.38 µg/l of 9 -THC. Opiates. Two opiate positive cases were detected. One contained 15.9 µg/l of morphine and 2.47 µg/l of the heroin metabolite 6-AM. Another OF sample contained 105 µg/l of codeine. The combined concentration of the relevant analytes for the DrugWipe 5 + opiates test, directly comparable to the manufacturers stated cutoff for the target compound morphine, was calculated using: Table V. Results of the Test Evaluations According to the DRUID and DrugWipe 5+ Cutoffs* According to DRUID Cutoff c(opiates) = 100% c(morphine) + 100% c(codeine) + 100% c(ethylmorphine) When interpreting the device result according to the test cutoff, both of the cases were FN. When utilizing the DRUID cutoff (20 µg/l, slightly higher than the test cutoff), the morphine-positive case was interpreted as TN and only the codeine case was FN. Cocaine. None of the cases were found to contain cocaine or the metabolite benzoylecgonine. All of the According to Test Cutoff AMP CAN OPI COC AMP CAN OPI COC TP FP TN FN N of successful tests Failed Missing analysis Total N of cases Sensitivity 87% 43% n.a. n.a. 87% 63% n.a. n.a. Specificity 95% 96% 100% 100% 95% 94% 100% 100% Accuracy 93% 88% 99% 100% 93% 92% 99% 100% Prevalence 28% 16% 0.7% 0% 28% 6.0% 1.5% 0% PPV 96% 75% n.a. n.a. 96% 71% n.a. n.a. NPV 85% 87% n.a. n.a. 85% 91% n.a. n.a. * Prevalence shown is calculated for study population; PPV and NPV calculated using DUI suspect prevalences from 2008 (in blood samples). Abbreviations: amp = amphetamines (incl. methamphetamine), can = cannabis, opi = opiates, and coc = cocaine. n.a. - calculation not applicable. cases were screened as negative with the DrugWipe 5 + test. For a small number of cases (n = 9) the analysis for benzoylecgonine was not successfully completed. Cocaine was analyzed separately from benzoylecgonine, and no trace was found in these samples. Nonetheless, they have not been included in the specificity, accuracy, and prevalence calculations. Rapid STAT The results of the test evaluation are shown in Table VI. Amphetamine-type stimulant drugs. Altogether 27 cases (22%) were positive for ATSs. All of these positive cases contained amphetamine. The concentration range for amphetamine was ,700 µg/l. Methamphetamine was found in four cases (3.1%), with the concentration range of ,700 µg/l. MDA, MDMA, or MDEA were not found in any of the samples. The test cutoff for both amphetamine and methamphetamine tests is Table VI. Results of the Test Evaluations According to the DRUID and Rapid STAT Cutoffs* According to DRUID Cutoff According to Test Cutoff AMP MAMP CAN OPI COC BZO AMP MAMP CAN OPI COC BZO TP FP TN FN N of successful tests Failed Missing analysis Total N of cases Sensitivity 72% n.a. 68% n.a. n.a. 68% 72% n.a. 85% n.a. n.a. 86% Specificity 100% 97% 89% 100% 100% 97% 100% 97% 88% 100% 100% 88% Accuracy 94% 97% 86% 100% 100% 89% 94% 97% 88% 100% 100% 88% Prevalence 20% 3.1% 15% 2.4% 0% 27% 20% 3.1% 10% 2.4% 0% 11% PPV 100% n.a. 61% n.a. n.a. 97% 100% n.a. 64% n.a. n.a. 91% NPV 74% n.a. 92% n.a. n.a. 68% 74% n.a. 96% n.a. n.a. 81% * Prevalence shown is calculated for study population; PPV and NPV calculated using DUI suspect prevalences from 2008 (in blood samples). Abbreviations: amp = amphetamine, mamp = methamphetamine, can = cannabis, opi = opiates, coc = cocaine, and bzo = benzodiazepines. n.a. = calculation not applicable. 215

6 25 µg/l, which is the same as the DRUID cutoff. Together with the high concentrations encountered in the samples, this means only one interpretation of the test results needs to be made. Amphetamine test. The combined concentration of the relevant analytes for the Rapid STAT amphetamines test, directly comparable to the manufacturers stated cutoff for the target compound amphetamine, was calculated using: c(amphetamines) = 100% c(amp) + 50% c(mda) + 0.5% c(mdea) + 1% c(mdma) There were 7 FN cases in the concentration range of µg/l. The situation was similar to that seen for the Drug- Wipe 5 + test: cases with concentrations amounting to several hundreds or even thousands of µg/l of ATSs still gave a negative screening result with the device. Still, the concentration range of the FN cases was relatively low in comparison to overall range found for amphetamine. Methamphetamine test. The combined concentration of the relevant analytes for the Rapid STAT methamphetamine test, directly comparable to the manufacturers stated cutoff for the target compound methamphetamine, was calculated using: c(methamphetamines) = 100% c(mamp) + 50% c(mdma) + 5% c(mdea) % c(mda) There were only four positive cases (3.1%) for methamphetamine. Four negative cases also gave a positive result for methamphetamine with the test. Three of these FP cases contained very high concentrations of amphetamine (44,700, 21,000, and 15,300 µg/l). The manufacturer of the device has reported 1000 µg/l of amphetamine to be the highest concentration that has not showed cross-reactivity. Hence, it can be speculated that these three false positives are a result of high amphetamine concentrations, although the manufacturer does not report amphetamine as a cross-reacting compound for the Rapid STAT methamphetamine test. Cannabis. Altogether 19 cases (14%) were positive for 9 - THC. The concentration range for these positive cases was µg/l. There were 6 FN cases according to the DRUID cutoff. The concentration range of these cases was µg/l. For Table VII. Benzodiazepine Findings in OF Samples for the Rapid STAT Evaluation Range Average Median Analyte N (µg/l) (µg/l) (µg/l) Nordiazepam Clonazepam Alprazolam Diazepam , Oxazepam Temazepam comparison, the corresponding range for the 13 TP cases was µg/l. Again, similarly to the results observed for the DrugWipe 5 + test, the range of 9 -THC concentrations which can give a positive test result is not distinct from that which gives a negative screening result. The largest FN concentration (2020 µg/l), although not typical, was alarmingly high. Opiates. Three positive opiate cases were identified. In two of these cases, codeine was the only opiate found (107 µg/l and 105 µg/l). In the third case, the findings were morphine (30.4 µg/l), codeine (1490 µg/l), and buprenorphine (8.8 µg/l). All three cases were tested positive with the Rapid STAT device. The combined concentration of the relevant analytes for the Rapid STAT opiates test, directly comparable to the manufacturers stated cutoff for the target compound morphine, was calculated using: c(opiates) = 100% c(morphine) + 50% c(6-am) % c(codeine) Cocaine. None of the cases included in the study contained cocaine or benzoylecgonine. All the cases tested negative with the Rapid STAT device. Similarly, as for the Drugwipe 5 + results the analysis for benzoylecgonine was not successfully completed in a small number of cases (n = 12). No trace of cocaine was found in these samples. They have not been included to the specificity, accuracy and prevalence calculations. Benzodiazepines. Altogether there were 37 cases (28%) which had benzodiazepines in their OF samples (according to the DRUID cutoffs). The most common benzodiazepine findings are shown in Table VII. The device s benzodiazepine test failed for three of these positive cases. Both of the substances lorazepam and midazolam were found twice, from four different samples. Concentrations for these cases were 1.6 and 44.7 µg/l lorazepam and 12.8 and 47.8 µg/l midazolam. The combined concentration of the relevant analytes for the Rapid STAT benzodiazepine test, directly comparable to the manufacturers stated cutoff for the target compound oxazepam, was calculated using: c(benzodiazepines) = 400% c(diazepam) + 100% c(alprazolam) + 200% c(flunitrazepam) + 4% 100% c(lorazepam) % c(clonazepam) + 100% c(nordiazepam) + 100% c(oxazepam) + 200% c(nitrazepam) + 200% c(temazepam) + 20% c(triazolam) There were several cases in which benzodiazepines were found below the DRUID cutoff values. Very low concentrations of oxazepam and diazepam were detected in 12 and 11 cases, respectively. Temazepam was found at very low levels in four cases. Lorazepam and chlordiazepoxide were detected in one case each. However, in all but one case other benzodiazepines were detected at levels above the DRUID cutoffs. 216

7 There were 11 FN cases according to the DRUID cutoffs. The concentrations confirmed in these cases were not substantially lower than in TP cases. Consequently, no real separation of low and high concentrations was found between the cases that tested negative or positive. Only two cases were interpreted as FN when the test cutoff and cross-reactivity calculations were used. Discussion Compared with the results obtained for the DrugWipe 5 in the Rosita-2 project (11), the sensitivity and specificity of the DrugWipe 5 + amphetamines test were approximately at the same level according to this study. In the context of cannabis related 9 -THC screening, the results of DrugWipe 5 + obtained in this study have somewhat improved from those presented for the DrugWipe 5 in the Rosita-2 project (11). Sensitivity increased from about 34% to 43%, whereas specificity remained at approximately the same level (92% in Rosita-2, and 96% in this study). The sensitivity of the DrugWipe devices for both the amphetamines and cannabis tests did not reach the values obtained in our previous study based on OF confirmation samples (21) and in a study based on blood confirmation samples by Wille et al. (22). In contrast, the specificity of both tests seemed to be at the better level. Unfortunately, it was not possible to properly evaluate the cocaine and opiates tests due to the low prevalence of these substances in Finnish traffic. There is a need for more evaluations of the performance of the Rapid STAT device to allow such a comparison as presented above. In this study, neither of the devices met the very strict criteria set in the Rosita 1 project (12): sensitivity and specificity should exceed 90% while accuracy should exceed 95%. There were 12 FP cases for 9 -THC screening with the Rapid STAT device at the DRUID cutoff (1 ng/ml), from which only one contained any findings at all for the drug in the OF sample. The possibility exists that the microfiber collection stick employed by this device was so efficient that no drug was left for the OF confirmation sample collection that normally would have occurred after the screening. However, the authors contend that this is probably not the case because the Rapid STAT instructions of use are to collect the sample by brushing around the gums and cheek area of the mouth, whereas the cellulose pad of the Saliva Sampler is placed under the tongue. Furthermore, in eight of the relevant cases a DrugWipe 5 + test was also performed prior to the Rapid STAT test and in only one instance was this also positive for 9 -THC, although the lower sensitivity of this device, as found in this study, should not be forgotten. Without further knowledge of the immunological assay used in the device for 9 -THC, the authors are unable to conjecture as to the reasons for such a relatively high number of FP cases, except to mention the generally problematic nature associated with screening for this drug. The DrugWipe evaluation results of this study and the Rosita-2 project are readily comparable for amphetamines and cannabis tests: cutoffs employed in this study (DRUID) and the Rosita-2 project for amphetamines are similar while the cutoff for 9 -THC has decreased from 2 to 1 µg/l between the successive EU projects. Corresponding Rosita-2 cutoffs were also employed in our previous study (21), whereas comparison to the results presented by Wille et al. (22) is merely suggestive due to the different matrix and cutoffs used for confirmation analysis. In addition, the number of cases evaluated for each substance group was rather limited (usually less than 50, and varies depending on the device and even for the individual tests comprised within a device) in this latter study (22). From the Finnish point of view, it would be desirable to find an on-site screening device that could reliably detect the use of amphetamines and benzodiazepines in particular. The use of either or both of these substance groups is confirmed in the vast majority of suspect DUID cases in Finland. Benzodiazepines and amphetamines have been the most prevalent substances (76% and 46%, respectively) in suspect DUID cases in Finland in the period Based on this register-based study, it seems likely that amphetamines will become even more prevalent than benzodiazepines (26). Also, an appropriate test for 9 -THC must not be forgotten since 20% of the suspected DUID cases in Finland in 2008 show signs of the previous use of cannabis. Perhaps unsurprisingly, the aforementioned substance groups are also common findings in DUID cases in the other Nordic countries (27). It must be noted that the Drug- Wipe devices have already been in regular use by the Finnish police for several years, and the police officers have been mainly satisfied with the operability of the device. Conclusions Regarding the results for sensitivity, specificity and accuracy both tests seem to perform quite well for the ATS drugs, although the sensitivity of the Rapid STAT test is only moderate. Nonetheless, it is evident that both screening devices can give negative results for cases that contain concentrations of hundreds or even thousands of µg/l of ATS drugs. In addition, the DrugWipe 5 + device also produced a positive test result for some cases that contained no ATSs at all, which is not acceptable. The performance of the benzodiazepine test of Rapid STAT seems to be at a relatively good level already if sensitivity, specificity, accuracy, PPV, and NPV are criteria for test performance. However it should be noted that out of all the positive test results only half were TP when test cutoffs were used as a reference point. Unfortunately, the same phenomenon can be seen for the Rapid STAT cannabis and methamphetamine tests as well as for the cannabis test in the DrugWipe 5 + device. This is definitely not acceptable for an on-site device. There would seem to be a lot of development work still ahead on detection of these drug classes for the Rapid STAT test. As previously mentioned the manufacturer of the DrugWipe launched a newer version of DrugWipe 5 + with an enhanced test strip for cannabis during the study. The results from Wille et al. (22), in which the newer device was evaluated (personal communication with Securetec Detektions-Systeme AG, manufacturer of DrugWipe 5 + ), already appear to be better than in this evaluation; however, as was already discussed, the results are not di- 217

8 rectly comparable becasue their evaluation was made using blood as a confirmation sample. Unfortunately, the number of positive cases for cocaine and opiates was very low, which made sensitivity calculations not applicable. However, both devices achieved very good results for specificity for these substance classes. Acknowledgments This report has been produced under the project Driving Under Influence of Drugs, Alcohol, and Medicines (DRUID) financed by the European Community within the framework of the EU 6th Framework Program. This report reflects only the authors view. The European community is not liable for any use that may be made of the information contained herein. References 1. J.K.M. Aps and L.C. Martens. Review: the physiology of saliva and transfer of drugs into saliva. Forensic Sci. Int. 150: (2005). 2. P. Lillsunde. Analytical techniques for drug detection in oral fluid. Ther. Drug Monit. 30: (2008). 3. D.B. Menkes, R.C. Howard, G.F.S. Spears, and E.R. Cairns. Salivary THC following cannabis smoking correlates with subjective intoxication and heart-rate. Psychopharmacology 103: (1991). 4. E.J. Cone, J.M. Oyler, and W.D. Darwin. Cocaine disposition in saliva following intravenous, intranasal, and smoked administration. J. Anal. Toxicol. 21: (1997). 5. N. Samyn, A. Verstraete, C. van Haeren, and P. Kintz. Analysis of drugs of abuse in saliva. Forensic Sci. Rev. 11: 1 19 (1999). 6. W. Schramm, R.H. Smith, and P.A. Craig. Drugs of abuse in saliva: a review. J. Anal. Toxicol. 16: 1 9 (1992). 7. W.M. Bosker and M.A. Huestis. Oral fluid testing for drugs of abuse. Clin. Chem. 55: (2009). 8. E.J. Cone and M.A. Huestis. Interpretation of oral fluid tests for drugs of abuse. Ann. N. Y. Acad. Sci. 1098: (2007). 9. D.J. Crouch, J.M. Walsh, L. Cangianelli, and O. Quintela. Laboratory evaluation and field application of roadside oral fluid collectors and drug testing devices. Ther. Drug Monit. 30: (2008). 10. S.M.R. Wille, E. Raes, P. Lillsunde, T. Gunnar, M. Laloup, N. Samyn, A.S. Christophersen, M.R. Moeller, K.P. Hammer, and A.G. Verstraete. Relationship between oral fluid and blood concentrations of drugs of abuse in drivers suspected of driving under the influence of drugs. Ther. Drug Monit. 31: (2009). 11. A.G. Verstraete and E. Raes, Eds. Rosita-2 Project: Final Report, Academia Press, Ghent, Belgium, 2006, pp A.G. Verstraete, Ed. ROSITA, Roadside Testing Assessment: Final Report, Rosita consortium, Ghent, Belgium, 2001, pp P. Kintz, V. Cirimele, and B. Ludes. Codeine testing in sweat and saliva with the Drugwipe. Int. J. Legal Med. 111: (1998). 14. N. Samyn and C. van Haeren. On-site testing of saliva and sweat with Drugwipe and determination of concentrations of drugs of abuse in saliva, plasma and urine of suspected users. Int. J. Legal Med. 113: (2000). 15. M. Grönholm and P. Lillsunde. A comparison between on-site immunoassay drug-testing devices and laboratory results. Forensic Sci. Int. 121: (2001). 16. S. Pichini, M. Navarro, M. Farre, J. Ortuno, P.N. Roset, R. Pacifici, P. Zuccaro, J. Segura, and R. de la Torre. On-site testing of 3,4- methylenedioxymethamphetamine (ecstasy) in saliva with Drugwipe and Drugread: a controlled study in recreational users. Clin. Chem. 48: (2002). 17. J.M. Walsh, R. Flegel, D.J. Crouch, L. Cangianelli, and J. Baudys. An evaluation of rapid point-of-collection oral fluid drug-testing devices. J. Anal. Toxicol. 27: (2003). 18. D.J. Crouch, J.M. Walsh, R. Flegel, L. Cangianelli, J. Baudys, and R. Atkins. An evaluation of selected oral fluid point-ofcollection drug-testing devices. J. Anal. Toxicol. 29: (2005). 19. P. Kintz, W. Bernhard, M. Villain, M. Gasser, B. Aebi, and V. Cirimele. Detection of cannabis use in drivers with the Drugwipe device and by GC MS after Intercept device collection. J. Anal. Toxicol. 29: (2005). 20. O.H. Drummer, D. Gerostamoulos, M. Chu, P. Swann, M. Boorman, and I. Cairns. Drugs in oral fluid in randomly selected drivers. Forensic Sci. Int. 170: (2007). 21. A. Pehrsson, T. Gunnar, C. Engblom, H. Seppä, A. Jama, and P. Lillsunde. Roadside oral fluid testing: comparison of the results of Drugwipe 5 and Drugwipe Benzodiazepines on-site tests with laboratory confirmation results of oral fluid and whole blood. Forensic Sci. Int. 175: (2008). 22. S.M.R. Wille, N. Samyn, M. del Mar Ramírez-Fernández, and G. De Boeck. Evaluation of on-site oral fluid screening using Drugwipe-5 +, RapidSTAT and Drug Test 5000 for the detection of drugs of abuse in drivers. Forensic Sci. Int. 198: 2 6 (2010). 23. K. Langel, C. Engblom, A. Pehrsson, T. Gunnar, K. Ariniemi, and P. Lillsunde. Drug testing in oral fluid evaluation of sample collection devices. J. Anal. Toxicol. 32: (2008). 24. T. Blencowe, A. Pehrsson, and P. Lillsunde, Eds. Analytical evaluation of oral fluid screening devices and preceding selection procedures. DRUID report; K. Pil, F.M. Esposito, and A. Verstraete. External quality assessment of multi-analyte chromatographic methods in oral fluid. Clin. Chim. Acta 411: (2010). 26. K. Ojaniemi, T. Lintonen, A. Impinen, P. Lillsunde, and A. Ostamo. Trends in driving under the influence of drugs: a register-based study of DUID suspects during Accid. Anal. Prev. 41: (2009). 27. A.S. Christophersen, G. Ceder, J. Kristinsson, P. Lillsunde, and A. Steentoft. Drugged driving in the Nordic countries a comparative study between five countries. Forensic Sci. Int. 106: (1999). Manuscript received July 1, 2010; revision received September 7,