Analytical approaches of European Union laboratories to drugs of abuse analysis

^a Author for correspondence. Fax 34-3-2213237; e-mail rtorre{at}imim.es.

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 
We report a survey on urine drug testing within a total of 269 laboratories of the European Union. Clinical laboratories predominated over forensic laboratories (59.5% vs 28.5%). Screening without identification/quantification was the common approach used by clinical laboratories, whereas screening with identification/quantification was the approach used by almost all forensic laboratories. Screening was primarily performed by immunoassay in both types of laboratories. Gas chromatography coupled to mass spectrometry was the main analytical method used for specific identification/quantification of drugs, but other methods (including immunoassays) were also used. Cutoff values applied varied by laboratory type, country, and method used. A high percentage of laboratories did not use or report cutoff values. Overall, countries of the European Union vary significantly in regards to drugs tested, analytical approach, and screening and identification cutoff values. It is recommended to clearly state the analytical method and the cutoff values used when reporting results for drugs of abuse testing.

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 
Countries in the European Union (EU)¹ differ in their approaches to the drugs of abuse problem; consequently, there are differences concerning drugs of abuse testing. The development of an internal single market in the EU and the associated free exchange of laboratory services bring new relevance to intercomparison of data. Since 1993, free circulation of workers in the EU has added a further need for harmonized criteria in workplace drug testing.

The analytical strategy used for drug testing may depend to a great extent to potential consequences of results. In general terms, analysis of body fluids for drugs of abuse takes place in two different environments: the clinical setting (therapeutic care of drug addicts) and the penalty setting (e.g., forensic medicine, prisons, workplace, insurance, driving, sports). In clinical practice, the analytical result is only one of a series of factors that affect the decision-making process and must be assessed as a complement to the patient–physician relationship. By contrast, in the penalty model, sanctions against the individual providing the specimen are mostly based on analytical results; the reliability of these results, therefore, is essential.

Important aspects of urine drug testing include the analytical approach, the methods used, and the cutoff concentrations applied. Usually a two-step procedure is followed in drugs of abuse testing, i.e., a preliminary screening of groups of substances (e.g., opiates), and the identification of specific substances (e.g., morphine, codeine, 6-acetylmorphine), sometimes accompanied by their quantification. However, given the lack of specific guidelines, different analytical strategies may be considered acceptable by different EU analysts.

In this context, in 1993 and 1994 a survey was undertaken in the European Community to examine the reliability of urine drug testing. The survey was supported both by the DG V/F/1 (Directorate General V, Employment, Industrial Relations and Social Affairs, Public Health and Safety at Work) of the European Commission and by the Institut Municipal d'Investigació Mèdica, Barcelona, Spain. The purpose of the study was not only to assess the quality of analysis performed (1), but also to gain insight into the analytical strategies applied in the different European countries and laboratories involved in drugs of abuse testing.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 
In the 1993 and 1994 survey among laboratories of the EU member states performing drugs of abuse analyses, six test samples of sterile urine containing several drugs and (or) their metabolites were provided for analysis under routine conditions. The drug menu by class of substances included amphetamines, opiates (morphine-related compounds), methadone, cocaine, and cannabinoids, whereas the drug menu for identification/quantification consisted of amphetamine, methamphetamine, morphine, codeine, 6-acetylmorphine, methadone, 1,5-dimethyl-3,3-diphenyl-2-ethylidene-pyrrolidine (EDDP), benzoylecgonine, ecgonine methyl ester, and 11-nor-9-carboxy--tetrahydrocannabinol. The content of samples was validated by a group of seven reference laboratories.

The results collection form required participating laboratories to provide information regarding type of laboratory (forensic, clinical, research, other) and type of institution (commercial or noncommercial). Information regarding the analytical methods used at each step (drug screening, identification, and quantification) was requested. A coded list of analytical methods most often used in drug testing was provided. Moreover, the form for results reporting was designed in such a way as to gain insight into analytical cutoff concentrations used both at screening and identification procedures.

More than 300 centers from a list supplied by the European Commission were invited to participate in the study. The list was based on a questionnaire on performances and capabilities of European Analytical Toxicology Labs. (2).

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 
participating laboratories
The total number of participating laboratories was 269 (including the seven reference laboratories), of which 32 (12.5%) were commercial laboratories and 225 (87.5%) were noncommercial, depending on whether organizations declared that they were profit- or non-profit-oriented. Data from five laboratories were not available. Laboratories from all countries belonged mainly to noncommercial institutions. With regard to the type of centers, 59.5% were clinical laboratories (range 31–87%), 28.5% were forensic laboratories (range 5–61%), and 12% were "other" (range 6–40%), including research laboratories (5%) (Table 1 ).

analytical approaches
A comparison of the rate of each analytical approach performed by laboratories, depending on whether they were clinical or forensic, is shown in Fig. 1 . In general, a high percentage of clinical laboratories performed only screening tests without identification or confirmation of positive urine drug test screen results; however, the analytical approach of clinical laboratories was not homogeneous in all countries (Table 2 ). For example, the percentages of Spanish and Portuguese clinical laboratories participating exclusively in the screening phase of the survey were clearly above the mean (71% and 100%, respectively). In contrast, all Danish clinical laboratories performed at least one identification of specific substances. However, none of the clinical laboratories from Denmark or Portugal performed quantification of substances, whereas >30% of laboratories from The Netherlands and Italy performed at least one quantitative analysis. With regard to the forensic laboratories (Table 2 ), almost all of them performed identification of positive results, regardless of country. Quantification of specific substances was carried out by 83% of German laboratories, whereas ~50% of those in Italy did not report quantification results. Analytical approaches of research-oriented laboratories were as follows: 45% of laboratories performed only screening vs 55% of laboratories performed at least one identification. A total of 27% of research laboratories performed quantification.

View larger version (42K): [in this window] [in a new window]   Figure 1. Percentages of testing being performed for each analytical approach (screening, identification, quantification) by clinical and forensic laboratories.

All participating laboratories screened for opiates. Amphetamines, cocaine, and cannabinoids were analyzed by almost 96–97% of laboratories, and methadone received the lowest rate of analysis (87%). When considering the rate of screening analysis performed in each country, the lowest percentage was recorded for Portugal; half of its participating laboratories did not screen for amphetamines and/or methadone. In Spain, the rate for methadone screening was 70%, also below the mean.

There were large differences among EU member states in the rate of identification of specific substances (Table 3 ). Although few Portuguese laboratories performed identification, they covered nearly the full menu of drugs. In contrast, all Danish laboratories performed identification but the rate of analyses for each specific substance was low. Nearly the full menu of drugs was also analyzed by French and German laboratories performing identification, while in The Netherlands the rates of identification of some substances were low. The largest differences in identification rates by country corresponded (in decreasing order) to 6-acetylmorphine, EDDP, benzoylecgonine, ecgonine methyl ester, and 11-nor-9-carboxy--tetrahydrocannabinol. As expected, more substances were identified by forensic laboratories than by clinical laboratories. The greatest difference was found for 11-nor-9-carboxy--tetrahydrocannabinol. Important differences were also noticed in the identification rates of 6-acetylmorphine, benzoylecgonine, and ecgonine methyl ester.

analytical methods
The use of either immunological or chromatographic techniques was highly dependent on the purpose of the analytical step. The percentages of analytical methods used for each analytical step and for each country are shown in Table 4 (only those countries with a significant number of participating laboratories are included).

Screening for groups of substances was mainly performed by immunological techniques (91%), particularly enzyme immunoassay (EIA; 54%) and fluorescence polarization immunoassay (FPIA; 29%). In the group of chromatographic techniques, thin-layer chromatography (TLC) was the most common (5%). A large proportion of laboratories from Denmark and the UK used EIA, whereas FPIA was more frequently used in The Netherlands. The use of TLC as screening method was concentrated mainly in Denmark and the UK. When comparing clinical and forensic laboratories, no differences were observed in the percentages of use of immunoassay vs chromatographic methods for screening purposes. Rates of use of EIA and FPIA were also similar for both types of laboratories. Nevertheless, there were differences with regard to the chromatographic method used. Of screening analyses, 6% were performed by TLC by clinical laboratories while only 1% of such analyses by forensic laboratories were by TLC.

Identification/quantification of specific substances was mainly performed by chromatographic methods (98% for identification, 90% for quantification). Gas chromatography coupled to mass spectrometry (GC/MS) was the main analytical technique used for identification and quantification of specific substances (48% and 55%, respectively). A few laboratories used immunoassay for identification, apparently ignoring the fact that nonspecificity of immunoassays permits only presumptive identification of drug classes so that specific identification of drugs must await confirmation testing. A greater percentage of quantitative analyses was performed by immunological methods (10%), especially FPIA. Because immunoassays generally involve cross-reactivity with metabolites, the quantified value is the sum of analyte and metabolite reactivities and results are actually in "calibrator-equivalent" units. GC/MS was the identification method most often used in all countries but Denmark and the UK, where TLC predominated. Regarding quantification, differences in GC/MS rate of use among countries were also observed, Portugal and Spain being the countries where more quantifications were performed by GC/MS (>80%). In contrast, in Denmark GC/MS was not used by any laboratory. The Netherlands and Italy were the countries with higher rates of quantitative analyses performed by immunoassay. Analytical methods used in clinical and forensic laboratories for identification and quantification of specific substances are shown in Fig. 2 . TLC was the analytical method most often used by clinical laboratories for identification analyses, GC/MS the more frequent among forensic laboratories. The rate of use of GC/MS for quantitative analyses was also higher among forensic laboratories. The use of immunoassays for quantification was mainly concentrated in clinical laboratories (28%); few quantitative analysis were performed by forensic laboratories using immunoassays.

View larger version (24K): [in this window] [in a new window]   Figure 2. Distribution of the percentages of testing being performed with each analytical method by clinical and forensic laboratories for identification and quantification of specific substances.

cutoff concentrations
Given that specific regulations on cutoff (or threshold) concentrations (3) for positivity have not yet been developed, laboratories were requested to report the cutoff values they used for drug screening and drug identification purposes (Table 5 ).

In the screening step, although the cutoff ranges were wide, there was good agreement in cutoff concentrations reported by laboratories. There were some target values that most of them tended to use. On the other hand, as many as 22% of screening analyses were performed without the cutoff values being reported. For the amphetamines group, European countries used more frequently the cutoff of 300 µg/L rather than the 1000 µg/L applied by the Substance Abuse and Mental Health Services Administration (SAMHSA) accredited laboratories (4). In screening for cannabinoids, three cutoff values (100, 50, or 20–25 µg/L) were applied, this being the group of substances with the largest differences. For opiates, methadone, and cocaine groups, >85% of laboratories used a cutoff value between 200 and 300 µg/L. Laboratories using TLC showed higher cutoff values than laboratories using Emit (Behring). In contrast, lower cutoff concentrations for some analytes were observed among laboratories that used FPIA for screening purposes as compared with those that used Emit. Because screening techniques were mainly commercial immunoassays, one could expect that the cutoff recommended by a manufacturer would be applied, and this was done in most cases (e.g., 250 µg/L for methadone screening applied by laboratories using FPIA; 20 µg/L cutoff by Emit and 25 µg/L by FPIA for cannabinoids screening), although some laboratories reported different cutoff concentrations than those expected, such as in 7.3% of analysis performed with Emit.

When comparing clinical and forensic laboratories in the screening step, the mean cutoff values were lower in forensic laboratories for all groups of drugs (Fig. 3 ). The percentage of laboratories not reporting a cutoff value, and probably not applying a cutoff concentration in the screening analysis, was greater for forensic than for clinical laboratories.

View larger version (34K): [in this window] [in a new window]   Figure 3. Mean cutoff concentration applied by clinical and forensic laboratories for each analyte (left bars) and percentage of laboratories not reporting a cutoff concentration (right bars) for screening purposes (top panel) or identification (bottom panel).

Mean cutoff concentrations applied by countries for each analyte are shown in Table 6 . To obtain overall information on differences observed among European countries, we established an arbitrary scoring system as follows: For each analyte, the mean cutoff concentrations applied by countries were sorted from the lowest to the highest, after which a rank number from 1 to 9 (because 9 countries were evaluated) was assigned to each and the numbers assigned to each analyte (5 in screening) were summed for each country. The evaluation range thus went from 5 to 45 arbitrary units (AU) for the screening, with the lowest value corresponding to the country in which the lowest cutoff concentrations were applied. Results obtained are shown in Fig. 4 . The analytical method and proportion between clinical and forensic laboratories were the main factors that influenced the cutoff values used.

View larger version (38K): [in this window] [in a new window]   Figure 4. Display of countries for application of cutoff concentration, from the lowest to the highest values: arbitrary system based on the mean cutoff used by each country for each analyte (see text). The scale ranges from 5 to 45 and from 9 to 81 arbitrary units for the screening and identification analyses, respectively.

For the identification of specific substances, there was a lack of agreement in the cutoff concentrations, and the cutoff ranges were even wider than in the screening step. Identification and screening cutoff criteria for major urinary metabolites coincided in almost 30% of laboratories. Theoretically, for the chromatographic identification of a main metabolite, one would expect lower cutoff concentrations than those applied for immunoassay screening of the group of drugs. This was observed for 45.9% of analyses performed but, interestingly, 7.4% of the laboratories used higher cutoff values in drug identification than in drug screening. The remaining laboratories used the same cutoff concentrations in both analytical steps. Agreement was even lower for identification of the minor metabolites, where, in general, the cutoff values used were lower than for the major metabolites.

When comparing clinical and forensic laboratories in the drug identification step, the mean cutoff values used by the forensic laboratories were lower for all substances except ecgonine methyl ester (Fig. 3 ). Discrepancies in the mode were found for the following substances: amphetamine, 300 µg/L for clinical laboratories vs 200 µg/L for forensic laboratories; 6-acetylmorphine, 300 vs 10–50 µg/L; EDDP, 300 vs 200 µg/L; ecgonine methyl ester, 20 vs 100 µg/L. As found in the screening step, a sizable number of laboratories did not report cutoff concentrations (33% of identification analysis were performed without cutoff specification), but unlike the case in screening, forensic laboratories seemed to have standardized cutoff concentrations for the identification of specific substances more often than did the clinical laboratories.

Mean cutoff concentrations applied by countries for each specific analyte are shown in Table 7 . An arbitrary scoring system following the same procedure as in the screening evaluation was established but the evaluation range varied between 9 and 81 AU because 9 variables were evaluated (ecgonine methyl ester was excluded because data from only a few countries were available); again, the lowest value corresponding to the country in which the lowest cutoff concentrations were applied. The results are shown in Fig. 4 . In those countries in which the use of TLC predominated (UK, Denmark), cutoff values were higher than those reported in countries using GC/MS (Portugal, France, Germany). Additionally, those countries with a greater participation of clinical laboratories reported higher cutoff concentrations. [Note that rates of clinical and forensic laboratories differed between screening and identification because many clinical laboratories did not perform identification (see Table 2 ).]

Differences among countries in the use of cutoff values not only reflect a lack of consensus in analytical/toxicological terms but also may be interpreted as eventual differences in the degree of social tolerance for the consumption of a given drug.

analytical results for survey samples
For identification of specific substances, an evaluation of results reported by laboratories for each specific drug present in the test samples was performed according the cutoff concentrations applied by each laboratory. Results were classified into three groups: (a) results reported by laboratories using cutoff values clearly below the drug concentration present in the test sample—for these cases, positive results were expected, in which case negative results were considered falsely negative; (b) results reported by laboratories using cutoff values in the range of drug concentration present in the test sample (concentration mean ±30%)—taking into account that quantitative analyses have a margin of error, a critical evaluation of these cases was not possible and this group was considered separately; (c) results reported by laboratories using cutoff values clearly above the drug concentrations present in the sample, in which case positive results were considered falsely positive—on the assumption that the analytical methods used had no sensitivity to detect the substances and (or) that the toxicological criteria of these laboratories should have indicated not to report results less than this concentration. Results are shown in Fig. 5 . False-positive cases accounted for almost 50% of the results; 4% were false-negatives.

View larger version (24K): [in this window] [in a new window]   Figure 5. Evaluation of results reported by participating laboratories for identification of specific substances. The cutoff concentration reported had been used for classifying each result for each substance and for each laboratory as a real positive (positive analytical result: cutoff < concentration; left) or a real negative (negative analytical result: cutoff > concentration; right) sample. In cases where the cutoff applied was similar (±30%) to the concentration in the sample, a separate group was considered (middle). Results were considered falsely negative (FN) when a laboratory reported a result as negative but the analyte was present in a concentration higher than the cutoff used by the laboratory. Results were considered falsely positive (FP) when a laboratory reported a result as positive but the analyte was present in a concentration less than the laboratory's cutoff value.

	Conclusions

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 
The total number of laboratories participating in this study was large enough to draw several conclusions about the situation of drugs of abuse testing in the EU. Because participation was voluntary, some bias in the sample of participating laboratories in relation to the whole population of European laboratories cannot be excluded. In five countries (Czech Republic, Finland, Greece, Ireland, Luxembourg), the number of participating laboratories was too small and were not included in some evaluations by individual countries.

A predominant percentage of the EU laboratories performing drugs of abuse analysis were noncommercial. Moreover, a higher proportion of clinical than forensic laboratories performed drugs of abuse analysis, although large variations in this ratio among countries were observed.

Differences in the analytical approach were clearly associated with the type of laboratory (5)(6)(7)(8). As could be expected, forensic laboratories performed more identification and quantification analyses than clinical laboratories did. This finding is probably related to the consequences that may be derived from the result of the analysis. In addition, large differences in analytical approaches favored in each country were seen that were independent of the type of laboratory involved in drug testing.

Regarding the analytical methodology, a large proportion of laboratories used commercial immunoassays for screening purposes. Differences observed between countries in which immunoassay was more often used were related to the position of each manufacturer in the market of the countries. This observation has an impact not only on the predominance of each immunoassay but also on the cutoff concentrations applied in each country for each analyte. There was a clear agreement between laboratories that used immunoassays for screening purposes. Inappropriate methods such as immunoassays were used either in the identification and quantification approach by some laboratories. Although a sizable number of clinical laboratories used TLC for the specific identification of substances, forensic laboratories more often used instrumented chromatographic methods, particularly GC/MS. In reference to the technology for quantification of specific compounds, in some countries some of the clinical laboratories used immunoassays rather than chromatographic techniques for this.

Cutoff values routinely applied were more influenced by the analytical method used than by the clinical or toxicological significance of the test [(9)(10)(11), and IC Dijkhuis, unpublished]. In the screening step, agreement in cutoff concentrations applied was greater because screening analyses were mainly performed with commercial immunoassays. The cutoff concentrations reported by the laboratories were usually coincident with the cutoff standardized for the commercial immunoassay used; therefore, immunoassay kit manufacturers have played a very important role in the standardization of cutoff concentrations applied by European laboratories. For the identification of specific substances, however, there was great variability in reporting cutoff concentrations. No specific values, really, could be considered as cutoff modes. The cutoff applied was more related to the sensitivity of the analytical method than to the real toxicological criteria for evaluating the results. Studies addressing the assessment of more adequate cutoff concentrations have usually been promoted by manufacturers of immunoassays and have focused on screening for groups of substances by these techniques. The studies have seldom been performed to evaluate the recommended cutoff concentrations for identification of specific substances. This is one reason for the lower agreement in the cutoff values used for drug identification purposes.

Forensic laboratories reported cutoff values in the screening analytical step less frequently than clinical laboratories but did specify cutoff concentrations for the identification of specific substances more often. On the other hand, the cutoff concentrations used by the forensic laboratories were lower than those of the clinical ones. As a result of large differences in the analytical methods used, there was great variability in the analytical cutoff concentrations applied for reporting positive results.

There appears to be some misunderstanding about the meaning of a cutoff value. An important number of analyses performed in samples containing concentrations less than a reported cutoff value were reported as positive, which clearly indicates that, in practice, the cutoff value reported was not applied by some laboratories.

In the US, the development of SAMHSA (formerly National Institute on Drug Abuse; NIDA) guidelines has provided some uniformity in drug-testing procedures, particularly at the workplace. In some countries, the lack of specific rules has led some European laboratories to adopt SAMHSA regulations. The substantial variability observed among different countries of the EU in the individual approaches to urine drug testing with respect to the analytical strategy applied is of great concern in situations where the same individual can be tested in different EU countries. The diversity of cutoffs in use most probably reflects the many different reasons for and purposes of the drug of abuse testing programs. Given the current situation, however, different results can be expected from a single sample, depending on the analytical criteria applied by the laboratory and the country in which the sample is analyzed. This situation generates confusion and may consequently affect the credibility of the laboratories. Thus, in reporting analytical results, the cutoff criteria and analytical methods used should be formally stated. Although reaching a general consensus on criteria to analyze drugs of abuse is quite difficult, some specific areas such as workplace drug testing, where up-to-date clinical and forensic laboratories are involved, will help bring about this consensus. In fact, recommendations for the reliable detection of drugs of abuse in urine with particular reference to the workplace have recently been published (12).

	Acknowledgments

 
We are indebted to M.T. Van der Venne and J.C. Berger (Commission of the European Union), Ch. Gilliard and F. Parmentier (Belgium), K. Worm (Denmark), P. Lafargue and P. Mangin (France), M.R. Möller and R.K. Müller (Germany), H. Tsoukalis (Greece), A. Pierce (Ireland), U. Avico, E. Sternieri, S.D. Ferrara (Italy), R. Wenning (Luxembourg), R.A. De Zeeuw and I.C. Dijkhuis (The Netherlands), D.A. Carrondo Tome dos Reis (Portugal), and M.D. Osselton and J. Williams (UK), members of the European Toxicological Working Group, for their valuable assistance in the design of the study. The following laboratories are also acknowledged for their support as reference laboratories: Institute of Toxicology (Oslo, Norway), Institute of Legal Medicine (Padova, Italy), National Poisons Unit (London, UK), Psychiatric Diagnostic Labs. America (South Plainfield, NJ), Smithkline Beecham Lab. (Atlanta, GA), Foothill Hospital (Calgary, Canada), Institute für Rechtsmedizin (Homburg/Saar, Germany), and National Lab. of Forensic Chemistry (Linkoping, Sweden). We are grateful to the Dirección General de Farmacia y Productos Sanitarios (Spain) and to the NIDA Drug Supply System for their supply of reference substances, to the Department of Psychiatry of Hospital del Mar for their supply of clinical urines, to the Clinical Trials Unit of the Department of Pharmacology and Toxicology at IMIM for their supply of excretion studies, and to Anna Artola, Alicia Redón, Marta Carnicero, and Marisa González for excellent technical assistance, Javier Morano for computer work, Rosa Herrera for secretarial help, and Marta Pulido for editing the manuscript and editorial assistance.

The Commission of the European Union has provided financial support for this study (re 92 CVVE 1-264-0; 93202764 05E01; 96CVVF2-201-0). Neither the European Commission nor any person acting in the name of the Commission is to be held responsible for the use made of the information contained in this report.

	Footnotes

 
Drug Abuse Research Unit, Institut Municipal d'Investigació Mèdica (IMIM), Autonomous University of Barcelona, Doctor Aiguader 80, E-08003 Barcelona, Spain.

¹ Nonstandard abbreviations: EU, European Union; EDDP, 1,5-dimethyl-3,3-diphenyl-2-ethylidiene-pyrrolidine; EIA, enzyme immunoassay; FPIA, fluorescence polarization immunoassay; TLC, thin-layer chromatography; GC/MS, gas chromatography–mass spectrometry; and SAMHSA, Substance Abuse and Mental Health Services Administration.

	References

Top Abstract Introduction Materials and Methods Results and Discussion Conclusions References

 

1. Badia R, Segura J, Artola A, De la Torre R. Survey on drugs of abuse testing in the European Union. J Anal Toxicol 1998;22(in press).
2. Compendium of facilities for drugs of abuse testing and general analytical toxicology in the European Community. In: de Zeeuw RA, Franke JP, van der Venne MT, eds. 1995 Report EUR 15980 EN. ISSN 1018–5593 Directorate General Employment, Industrial relations and Social Affairs. Health and Safety Directorate. Luxembourg: European Commission..
3. De Cresce R, Mazura A, Lifshitz M, Tilson J. Testing for drugs. Ch. 3. In: Drug testing in the workplace. Chicago: ASCO Press, 1989:81..
4. . DHHS/SAMHSA. Mandatory guidelines for federal workplace drug testing programs; notice. Fed Regist 1994;59:29908-29931.
5. Visher C. A comparison of urinalysis technologies for drug testing in criminal justice. Res. Report NCJ 132397 1991:1-44 Natl Inst of Justice Washington, DC. .
6. Catlin D, Cowan D, Donike M, Fraisse D, Oftebro H, Rendic S. Testing urine for drugs. J Autom Chem 1992;14:85-92.
7. Spiehler VR, O'Donnell CM, Gokhale DV. Confirmation and certainty in toxicology screening. Clin Chem 1988;34:1535-1539. [Abstract/Free Full Text]
8. Goldberger BA, Cone EJ. Confirmatory tests for drugs in the workplace by gas chromatography–mass spectrometry. J Chromatogr A 1994;674:73-86. [Web of Science][Medline] [Order article via Infotrieve]
9. Liu RH, Edwards C, Baugh LD, Weng JL, Fyfe MJ, Walia A. Selection of an appropriate initial test cut-off concentration for workplace drug urinalysis —cannabis example. J Anal Toxicol 1994;18:65-70. [Web of Science][Medline] [Order article via Infotrieve]
10. Kelly TH, Foltin RW, Emurian CS, Fischman MW. Performance-based testing for drugs of abuse: dose and time profiles of marijuana, amphetamine, alcohol and diazepam. J Anal Toxicol 1993;17:264-272. [Web of Science][Medline] [Order article via Infotrieve]
11. Cone EJ, Dickerson S, Paul BD, Mitchell JM. Forensic drug testing for opiates. V. Urine testing for heroin, morphine, and codeine with commercial opiate immunoassays. J Anal Toxicol 1993;17:156-164. [Web of Science][Medline] [Order article via Infotrieve]
12. de la Torre R, Segura J, Williams J, de Zeeuw RA. European Union Toxicology Experts Working Group. Recommendations for the reliable detection of illicit drugs in urine in the European Union, with special attention to the workplace. Ann Clin Biochem 1997;34(4):339-344.