August cbr 110806 final.indd

Drug Testing in Oral Fluid
Olaf H Drummer
Victorian Institute of Forensic Medicine and Department of Forensic Medicine, Monash University, 57-83 Kavanagh St, Southbank, Vic. 3006, Australia For correspondence: Prof Olaf Drummer e-mail: [email protected] Over the last decade there have been considerable developments in the use of oral fl uid (saliva) for drug testing. Oral fl uid can provide a quick and non-invasive specimen for drug testing. However, its collection may be thwarted by lack of available fl uid due to a range of physiological factors, including drug use itself. Food and techniques designed to stimulate production of oral fl uid can also affect the concentration of drugs. Current applications are mainly focused on drugs of abuse testing in employees at workplaces where drug use has safety implications, in drivers of vehicles at the roadside and in other situations where drug impairment is suspected. Testing has included alcohol (ethanol) and a range of clinical tests eg antibodies to HIV, therapeutic drugs and steroids. Its main application has been for testing for drugs of abuse such as the amphetamines, cocaine and metabolites, opioids such as morphine, methadone and heroin, and for cannabis. Oral fl uid concentrations of basic drugs such as the amphetamines, cocaine and some opioids are similar or higher than those in plasma. Tetrahydrocannabinol (THC), the major species present from cannabis use, displays similar concentrations in oral fl uid compared to blood in the elimination phase. However, there is signifi cant local absorption of the drug in the oral cavity which increases the concentrations for a period after use of drug. Depot effects occur for other drugs introduced into the body that allow local absorption, such as smoking of tobacco (nicotine), cocaine, amphetamines, or use of sub-lingual buprenorphine. Screening techniques are usually an adaptation of those used in other specimens, with an emphasis on the parent drug since this is usually the dominant species present in oral fl uid. Confi rmatory techniques are largely based on mass spectrometry (MS) with an emphasis on Liquid Chromatography-Mass Spectrometry (LC-MS), due to low sample volumes and the low detection limits required. Drug testing outside laboratory environments has become widespread and provides presumptive results within minutes of collection of specimens. This review focuses on the developments, particularly over the last 10 years, and outlines the roles and applications of testing for drugs in oral fl uid, describes the diffi culties associated with this form of testing and illustrates applications of oral fl uid testing for specifi c drugs. to the high sensitivity and specifi city of MS.2-4 This has Drug testing has undergone major advances, particularly assisted in the development of drug testing in oral fl uid due to over the last 10 years. The use of alternative specimens to the relatively small sample volumes that are usually collected.
blood or urine for establishing exposure to drugs has become The last decade has also seen a signifi cant development in the a signifi cant direction in clinical and forensic toxicology.1 understanding of the target drugs and their pharmacokinetics These alternative specimens include hair, sweat and oral in oral fl uid. This has applied particularly to the abused fl uid. Oral fl uid has been seen as a non-invasive alternative to drugs and what concentrations need to be targeted, and also blood but also as an alternative to urine when substitution or how these concentrations may or may not relate to blood adulteration is suspected. While these attributes are real, oral concentrations and the likely drug effects on the individual.
fl uid cannot be seen as a substitute for blood or urine drug In addition, the use of initial screening cartridges or devices testing. Each specimen has its own distinct advantages and providing an electronic readout has developed and is now widely used. In particular, kits are now designed for on-site drug detection without the need for sophisticated laboratory The introduction of LC-MS as a routine laboratory technique screening equipment. These are able to provide a preliminary has enabled the benefi ts of High Performance Liquid drug result within minutes.
Chromatography (HPLC) separation techniques to be linked Clin Biochem Rev Vol 27 August 2006 I 147 A number of reviews and major papers currently exist for that is used to collect the oral fl uid is added to a diluent. various aspects of drugs and drug testing in oral fl uid. These Following mixing, the solution is used for drug analysis. include its use as a diagnostic tool,5 workplace applications,6 Other devices involve squeezing absorbed oral fl uid from a applications in drugs in driving,7 legal issues associated pad or foam onto the drug-detection device. The collection with drug testing in oral fl uid,8 and detection times and time is typically one to three minutes, however this can vary as pharmacokinetics of selected drugs.9,10 discussed earlier. A number of the devices have some form of indicator to show that suffi cient oral fl uid has been collected.
This review outlines the roles and applications of testing for drugs in oral fl uid, describes the relative advantages The DrugWipe® only involves swiping a collection pad on and disadvantages of this form of testing and illustrates the tongue or skin, a process that takes only seconds (Table 1).
applications of oral fl uid testing for specifi c drugs.
However, there is no oral fl uid for any confi rmatory assay if the result is positive.
Scope of Review
This paper reviews the developments and applications of Oral fl uid production is stimulated by use of agents such as drug testing in oral fl uid particularly over the last 10 years. citric acid candy, chewing gum or other agents. This will Published peer-reviewed literature and other selected inevitably change the pH and concentration of drug in the references in the English language in humans as sourced by oral fl uid. This has been shown to lower concentrations of Medline and Science Direct (since 1995) are reviewed for codeine by about two- to six-fold,12,13 two- to four-fold for clinical and forensic applications of drug testing in oral fl uid. methamphetamine,14 and about fi ve-fold for cocaine.15 It is Publications before this time are included if pivotal or later likely that similar changes will occur for other drugs.
papers were not available. The term oral fl uid refers to saliva and other secretions in the oral cavity. The focus is on testing A number of drugs are known to affect the secretion of oral for drugs of abuse such as the amphetamines, benzodiazepines, fl uid.11 Most commonly these are amphetamines, including cannabis, cocaine and opioids, but other applications will be the designer forms such as ecstasy (MDMA), and cannabis. discussed in the context of abused drugs.
Other drugs include the sedating antihistamines, antipsychotic drugs, anticholinergic drugs and a number of antidepressants. Source of Oral Fluid
There are less commonly used drugs that increase fl ow and Oral fl uid (saliva) is excreted primarily by three glands: the these include clonidine, pilocarpine and beta-2 stimulants parotid, submaxillary and sublingual and by other smaller (salbutamol, terbutaline etc).
glands. Oral fl uid has low protein content (0.3%) and can vary in fl ow rate from zero to several mL per minute depending on Consequently, there is signifi cant intra- and inter-subject infl uences from various factors, including emotional state and variation in relation to drug concentrations depending on the hunger. Dry mouth syndrome is relatively common and can technique used, the physiology of the person and the infl uence be caused by the anxiety of the collection procedure, or even of factors affecting drug concentration in oral fl uid. by lack of proper hydration of the individual. Dry mouths require much longer collection times; often several minutes to Since the collection of oral fl uid specimen can be viewed by a collect 1 mL. On some occasions this may force the collection second person without infringing privacy it does not suffer from of an alternative specimen if collection is too slow, ie blood the same issues regarding possible adulteration or substitution (unpublished data). Aps and Martens provide an excellent as for urine.13 While this can be a distinct advantage, it must review of physiological and pharmacological issues involved be recognised that methods can be employed to potentially in the production of oral fl uid.11 affect the collection of oral fl uid or the concentration of drugs in oral fl uid. The prior administration of drugs and a Collection Techniques and Adulterants
range of physiological factors covered earlier can affect drug Expectoration (or spitting) provides neat oral fl uid, but this concentration. Foodstuffs, various beverages and various is relatively viscous and can be diffi cult to work with in toothpastes did not affect the concentration of drugs of abuse the laboratory. It may also be contaminated with food and using the Oratect® device 30 min after exposure.16 The use of other debris from the mouth and will therefore require commercial adulterants or other products capable of acting as centrifugation. More often than not, the volume will be less adulterants, such as Clear Choice®, Fizzy FlushTM Spit and than 1 mL requiring the use of sensitive detection techniques. Clean®/™ mouth wash and Cool Mint Listerine® also had no Some of the commercial collectors available use some form substantial effect after 30 min. In an early study using a small of proprietary diluent to mix with the collected oral fl uid number of volunteers, the consumption of beer immediately (Table 1). In this situation, typically the absorbent pad/foam smoking a marijuana joint appeared to lower concentrations 148 I Clin Biochem Rev Vol 27 August 2006 Table 1. Selection of collection devices reported in literature.
Name of collector
Method of operation
Swipe only (tongue or skin) 62, 66, 67, 93, 94 Cozart® collector Absorbent foam pad plus diluent 75, 76, 93, 95, 96 Dräger DrugTest® Absorbent foam pad with diluent Absorbent foam pad with diluent Absorbent foam pad only, drops applied to device Absorbent foam pad, collector squeezed to apply oral fl uid into test cartridge Direct application to oral cavity, or use of other collectors Absorbent directly connected to device Absorbent foam pad plus diluent Absorbent foam pad, drops applied to device Cotton wool swab which is then fi ltered Absorbent bud, oral fl uid squeezed into syringe and applied to device of THC in oral fl uid at 1 h post dose.17 It is likely that a short Recovery of Drugs from Collectors
rinsing effect is seen with these agents and others, including There is no one type of collection device that is clearly water, for a short period after use. Since the oral fl uid in the superior based on design or ease of use. However, recovery mouth is rapidly turned over, a wait of several minutes should studies conducted on some devices suggest that desorption allow re-equilibration of drug in the surrounding tissues. of drugs may limit the usability of some collection materials. For example, the Salivette® has poor recovery for THC but The rinsing effect with drugs is in some way similar to the is reasonable for codeine,12,19 whereas the Cozart® collector contamination of breath alcohol by recently consumed alcohol has good recovery for THC,20 and methamphetamine in that a wait of 15-20 min allows any mouth alcohol to be (unpublished data). The Quantisal® collection device has a removed by normal physiological processes.18 However, more good recovery for THC,21 although another study found lower research is needed to investigate this phenomenon in oral fl uid recoveries for THC.20 for the various collection techniques.
Clin Biochem Rev Vol 27 August 2006 I 149 Clearly more information is required for all drugs likely from the pH of oral fl uid and blood, the protein binding of the to be measured in oral fl uid, for each collection device. drug and its pKa.13,28 For acidic drugs the equilibrium favours Indeed a device should not be used until recovery and blood, hence oral fl uid concentrations are less than for blood, stability studies have been performed and show adequate while for basic drugs higher oral fl uid concentrations occur. performance. However, it should be emphasised that products The average concentration ratio is shown in Table 2. In the are continually being developed and hence published results absorptive phase there are often higher concentrations in the on an earlier design may not bear any resemblance to more oral fl uid due to local absorption in the mucous membranes recent designs.
of the buccal cavity. This local absorption effect is probably highest for THC due to its higher fat solubility and ease of Applications of Oral Fluid Drug Testing
penetration through membranes and the very low partitioning The use of oral fl uid to detect drugs has potentially wide from blood to oral fl uid. However, as discussed later in this applications. To date its main application has been to provide section this effect is also seen for other drugs.
a non-invasive specimen for testing of possible drug-affected drivers.22 It has also been used for workplace testing, The most commonly detected drug toxicologically, alcohol particularly following a safety incident, to check for possible (ethanol), has been subject to much research in terms of its drug use.23,24 Other applications include testing of persons in presence in oral fl uid. The oral fl uid to plasma concentration prisons and other correctional institutions, the monitoring of is similar to that predicted based on the water content of the drug use by drug courts, or testing of detainees suspected of a two fl uids and averages just over unity and has been used to crime who may be under the infl uence of a drug. assess alcohol exposure.25,29-31 Oral fl uid should not be seen as a specimen that replaces the Subjective intoxication and the increase in heart rate in use of other specimens. As discussed later the pharmacokinetic volunteers taking cannabis correlated well with oral fl uid THC characteristics of drugs are more closely aligned to blood concentrations.32 Oral fl uid concentrations also correlated well concentrations than, for example, urine or hair. Urine should with plasma concentrations.33 still be seen as the specimen of choice if evidence of prior exposure to drugs of abuse is sought (eg routine workplace The administration of 30 mg doses of codeine phosphate screening without cause and drug screening of prisoners). showed a good correlation of plasma and oral fl uid Hair will still be much more useful if a longer time frame of concentrations particularly after 2 h following the initial exposure to drugs is sought. However, if evidence of recent contamination of the oral cavity.34 The individual oral fl uid use (or abstinence) of drugs is sought then either blood or oral to plasma concentration ratios varied substantially and was fl uid are preferred specimens. partly due to the pH of the oral fl uid. Some concordance of the physiologic and subjective effects of codeine and oral fl uid Oral fl uid has the advantage over blood in that it can be concentration occurred following single oral codeine doses to obtained non-invasively in a situation where adulteration or substitution is diffi cult. A review of the advantages and Orally administered morphine shows a delay in the appearance disadvantages of specimens is available.25 in oral fl uid compared to its presence in plasma suggesting some rate limiting movement in oral fl uid possibly due to its relatively low lipid solubility.25 6-Acetylmorphine and A recent review of the pharmacokinetics of some drugs in morphine are also present in detectable amounts in oral fl uid oral fl uid has been published.10 As distinct from urine the after use of heroin.27 dominant species in oral fl uid is the parent drug.10 Hence, initial screens and confi rmatory techniques target the parent Buprenorphine is widely used for the treatment of opioid drug. For example, there is almost no carboxy metabolite dependency and is available (amongst other formulations) of THC present in oral fl uid. However, due to the rapid as a sublingual tablet.36,37 The local absorption of drug in the bioconversion of cocaine, benzoylecgonine and ecgonine mucous membranes of the oral cavity produces a depot-like methyl ester they are also detectable in oral fl uid.26,27 Moreover, effect for some hours after administration of drug. In the anhydroecgonine methyl ester is also detected in oral fl uid terminal phase of elimination oral fl uid concentrations were after smoking cocaine.27 similar to plasma.38 Nevertheless, the data suggested that oral fl uid could be used to monitor the use of this opioid. A similar As a general rule there is some similarity between an oral fl uid depot effect occurs with nicotine. Research suggests that concentration and a blood/plasma concentration. In the case measurement of the major metabolite cotinine is more useful of most drugs the oral fl uid concentration can be estimated than nicotine to determine exposure to this drug.39,40 150 I Clin Biochem Rev Vol 27 August 2006 Table 2. Average oral fl uid to blood concentration ratios for selected drugs.
Average oral fl uid to blood concentration Alcohol (ethanol) Methamphetamine (basic) Diazepam (acidic) Methadone (basic) Type refers to physiochemical property of drug, ie acidity, basicity or neutrality. The average ratios are indicative fi gures derived from pharmacokinetic studies and will change depending on a number of factors, including pH of oral fl uid, protein binding and degree of contamination of the membranes in the oral cavity by recently consumed drug.
Slow equilibration between plasma and oral fl uid has also years although it has been increasingly used in a range of been observed for diazepam.41 Since diazepam and other applications. There is some support for its routine application benzodiazepines are highly protein bound and are weakly for some anticonvulsants and theophylline.51,52 Recently acidic they have low oral fl uid concentrations. Mean oral applications for carbamazepine,53 digoxin,51 topirimate,54 fl uid to plasma ratios for diazepam are about 0.01-0.02.42,43 methadone,55 disopyramide,56 and docetaxel and paclitaxel Little data is available for other benzodiazepines, however, have been described.57 they are likely to behave similarly.44,45 Stability problems for the nitrobenzodiazepines (clonazepam, fl unitrazepam and Other clinical applications include testing for HIV- nitrazepam) have been reported in which conversion to the antibodies,58,59 and a number of steroids including cortisol and corresponding 7-amino metabolite also occurs.45,46 The use of dehydroepiandrosterone,60 and 17-hydroxy progesterone.61 fl uoride has been shown to stabilise the drug.45 An often mooted advantage for oral fl uid over serum is the ease of collection and is, of course, less invasive than These data show that the pharmacokinetics of drugs in oral venepuncture. However, care will be needed to avoid changes fl uid is more complex than that of blood. Detection times in in concentration in oral fl uid caused by a variety of factors this specimen will depend on a range of factors including discussed elsewhere in this review.
dose, frequency of use (ie acute versus chronic use) and detection limits of analytical assays. A number of other drugs Initial Testing Techniques
of forensic interest have been measured in oral fl uid. These Initial testing of oral fl uid for drugs can either occur in the include hydromorphone,47 phencyclidine,48 pholcodine,49 and fi eld (other words for this form of testing include on-site or point-of-care testing) or in the laboratory. Therapeutic Drug Monitoring (TDM)
TDM of drugs in oral fl uid has been studied for at least 30 A number of devices are available for fi eld use. These include Clin Biochem Rev Vol 27 August 2006 I 151 instrumental devices providing an electronic readout such Belgium who failed a sobriety assessment but were below the as the Dräger DrugTest® and Orasure Uplink®, Cozart legal limit of alcohol (0.05%). Oral fl uid concentrations were Rapiscan® and Drugread® hand photometer (Securetec) to closely related to plasma concentrations and gave positive hand-held cartridges requiring visual identifi cation such as the predictive values and sensitivities exceeding 90% when DrugWipe® (Securetec AG), iScreen OFDTM (Rapid Detect evaluated against plasma legal cut-offs. The DrugWipe® Inc.), OralScreen® (Avitar Technologies), Oratect® (Branan device through a tongue swipe was not recommended since Medical Corp.), SalivaScreenTM (Craig Medical Inc.) (Table accuracy of more than 90% was only obtained for amphetamine 3). These are optical readers that provide a visual readout of and MDMA.66 The use of Drugread® hand-held photometer intensity of response of the immunoassay signal.
with the DrugWipe® cartridge has been used to assess the detection of MDMA in oral fl uid following a 100 mg single Unfortunately, there is no consistency in the specifi cations dose in volunteers. Detection times ranged up to 10 h post- applied to these devices. For some, cut-off concentrations are dose although it was most reliable to 6 h. The detection limit used to defi ne their detectability, for others concentrations are was about 450 ng/mL.67 The pharmacological effects of the given when drugs can be detected. The apparent sensitivity is drug were most signifi cant up to about 6 h.
often not defi ned in terms of consistency of detection in oral fl uid specimens. This means that at the present moment there A recent study in Victoria, Australia in which over is no objective way to assess performance of these devices or 13,000 motorists were randomly screened for presence of methamphetamines (methamphetamine and MDMA) and THC as part of a campaign to reduce drug affected driving found a In the laboratory, terms such as false positive (FP) and positive rate of 1:40 compared to about 1:100 for alcohol. This false negative (FN) are used. FP refers to a situation when a testing was based on an initial tongue wipe using DrugWipe®, presumptive initial test result is not confi rmed. FN refers to and if positive, repeat testing on the Rapiscan®/™ following a situation when a confi rmation test fi nds a drug present that collection of oral fl uid with the Cozart® collector (unpublished was not detected by the initial test. Sensitivity is often used data). The overall FP rate using both devices was very low in defi ning performance of initial testing kits and refers to the (one cannabis and four methamphetamines). Individually the relative detectability of the kit or device (of positive cases) devices gave more FP for methamphetamines (DrugWipe®) in question over a comparison method. On the other hand, and THC (Rapiscan®/™).
specifi city refers to the percentage of negative results using the kit or device compared to the total number of negative As for urine, immunoassay tests of a drug class will not specimens using a comparison method. The comparison detect all members of the drug class equally. For example, the method is usually a Gas Chromatography-Mass Spectrometry required sensitivities of the initial test for the amphetamine, (GC-MS) or LC-MS method.
opiate and benzodiazepine classes will be different for the various drugs since the concentration of drugs in these classes A number of published studies have evaluated devices either are quite different to that of blood. For example, amphetamines using fortifi ed oral fl uid with known drug concentrations or have higher concentrations in oral fl uid compared to blood and real specimens taken from humans exposed to the drug under benzodiazepines have concentrations only a fraction of those in blood. Hence, it is important that the selection of on-site testing devices has the appropriate sensitivity (and indeed other A number of studies have used the DrugWipe® cartridge. performance characteristics) for the intended applications. These have found a high rate of FN for volunteers given 60 mg codeine over a 24 h period using a limit of quantitation 2. Laboratory Testing of 5 ng/mL.62 The testing for amphetamines, cocaine and There are a number of commercial kits based on ELISA opiates in drivers was reasonably reliable when tested against technology available for laboratory screening of oral fl uids. a GC-MS technique, except for some FN in heroin users.63 Amphetamines and opiates performed better than for benzo- These generally work satisfactorily for amphetamines,68,69 diazepines and THC on the DrugWipe® and Rapiscan®/™ buprenorphine,70 cocaine,65,71-73 methadone,74 and other in a drug driving study in Finland.64 The sensitivity and opioids,72,75,76 and provide a reliable means to screen oral specifi city were all close to 90% for Rapiscan®/™ when fl uid. Cannabis can be more diffi cult particularly if the testing for cocaine using a confi rmation cut-off of 30 ng/ immunoassay has little cross-reactivity to THC. Nevertheless mL.65 In a Belgian study the DrugWipe® performed best for enzyme immunoassay has been successfully used for this amphetamines and cocaine.63 The reliability of DrugWipe® drug.77 The same applies for benzodiazepines despite their was assessed on drivers detained at special roadblocks in low concentrations in oral fl uid.78 152 I Clin Biochem Rev Vol 27 August 2006 The power of MS has been used as a screening system for a offs in urine between countries after over 30 years of testing. range of drugs.79 Other general drug screening systems using Moreover, inadvertent exposure may limit the concentrations GC-MS,80,81 or LC-MS have been published that allow multiple that can be used to prove deliberate use. In the case of drugs to be detected and quantifi ed in oral fl uid.81-84 The choice cannabis, a study has found THC concentrations for a short of method is more to do with availability of instrumentation period following high passive exposure in an unventilated and costs since both forms of MS show suffi cient sensitivity room of up to 26 ng/mL.86 Ingestion of poppy seeds in food for most forms of drug detection in oral fl uid.
can cause a positive test result for morphine and exceed the 40 ng/mL SAMHSA cut-off for about one hour following Confi rmatory Analytical Techniques
Confi rmatory techniques for drugs in oral fl uid are for the most part adaptations of their counterparts in blood or plasma/ Nevertheless, numerous papers exist that provide validated serum. Given the larger water component and lower protein methods for the detection of notable drugs in oral fl uid. A content of oral fl uid compared to blood, recovery of drugs summary of these is shown in Table 4. Predominantly, the is not usually a limiting factor. The smaller sample volume preferred technique is MS due to its high sensitivity and and often lower concentrations in oral fl uid require the most specifi city. Consequently the focus is on the use of this adjustments to analytical techniques. The required detection or quantifi cation limit for drugs in oral MS can be in the form of GC-MS or LC-MS including tandem fl uid depend very much on the application and type of screening mass spectrometry (MS-MS) applications for both forms of test employed. For example, in workplace applications the chromatography. Detection limits are within or less than those Substance Abuse & Mental Health Services Administration in mentioned in Table 3 and use volumes of oral fl uid from 0.1- the USA (SAMHSA) has recommended confi rmation cut-offs 0.5 mL. The majority of methods use LC-MS as distinct from for THC, cocaine, morphine and the amphetamines of 4, 8, 40, GC-MS to cater for the lower sample volumes and required and 50 ng/mL, respectively.22 In contrast, the European Union low detection limits, although a number of GC-MS techniques roadside assessment testing study (ROSITA) for impaired have been reported with adequate sensitivity. A review of the drivers have recommended somewhat different cut-offs,22 as pros and cons of LC-MS methods in oral fl uid drug detection has the Australian Draft Standard for the collection, detection has been published.88 and quantifi cation of drugs of abuse in oral fl uid.85 (Table 3) Whatever technique is used it is important that the detection The variable target minimum concentrations probably refl ect limits applied to confi rmation testing is the same, but preferably the relative embryonic stage of drug testing compared to urine lower, than the initial testing threshold concentration. This drugs of abuse testing. It is possible that some international avoids not being able to confi rm an initial on-site positive agreement may exist in the future regarding minimum result because of insuffi cient sensitivity and to cater for detectable concentrations (or cut-offs). However, this is presence of some metabolites that cross-react with antibodies unlikely since there are still signifi cant differences in cut- used in immunoassays.
Table 3. Recommended minimum detectable concentrations of drugs in oral fl uid.
Standards Australia proposed target concentrations (ng/mL) 6-AM = 6-Acetylmorphine, MDMA = methylenedioxymethamphetamine, THC = ∆9-tetrahydrocannabinol.
Clin Biochem Rev Vol 27 August 2006 I 153 Table 4. Selected mass spectrometric methods used to quantify some common drugs in oral fl uid.
LOQ (volume of specimen) ng/mL (AM, MA, MDMA, MDEA etc) 500 µL (0.1-0.2 ng/mL) 6-AM = 6-acetylmorphine, AM = amphetamine, APCI = atmospheric pressure chemical ionization, MS2 = MS/MS, MA = methamphetamine, MDMA = methylenedioxymethamphetamine, MDEA = methylenedioxyethylamphetamine, LOQ = limit of quantifi cation – note some methods have used limit of detection in validation, * most methods also measure benzoylecgonine, ecgonine methyl ester and other metabolites of cocaine.
(workplace, street, etc) there is a need to ensure a level of It is now expected that laboratories in many parts of the quality assurance takes place to ensure that the devices are world, including the author's own laboratory, are subject to used as recommended by the manufacturer and suffi cient some form of certifi cation or accreditation process. Most quality issues have been addressed to ensure optimum and commonly laboratories conform, or should conform to the consistent performance. This includes the training of staff, ISO/IEC 17025 standard.89,90 This means that laboratories the running of suitable quality controls and participation in testing batches of specimens would also employ blank external profi ciency tests.85 samples, samples with known concentrations (calibrators) and quality controls to ensure the results of each batch of The principles of good laboratory practice do need to be also specimens meet appropriate laboratory performance criteria. considered for on-site testing. In practice this may be more Only results from those batches where performance criteria are diffi cult given that the environmental conditions and location satisfactorily met are therefore accepted. All other results are are much less controlled than a laboratory. Nevertheless, rejected and the analysis repeated. Additionally the methods it is imperative that the collection and testing process is as used must be fully validated and comply with International controlled as is reasonably feasible and the staff performing harmonised guidelines.91 the collection of specimens and the testing are properly trained, otherwise it is likely that initial on-site results will be In addition, most laboratories take part in some form of less reliable. This may produce a higher rate of FP and FN.
profi ciency test to independently assess their ability to detect drugs. One program in oral fl uid has been reported.92 The last decade has seen a revolution in the development of In essence this means that there are checks and balances for alternative specimens for drug analysis. The use of oral fl uid using laboratory-based test kits including their calibrations and has been found to offer signifi cant promise when detection monitoring their performance. When cassettes by themselves of relatively recent use of drugs is sought in a non-invasive or cassettes placed in readers are used in the testing location manner. Technological advances do allow on-site detection of 154 I Clin Biochem Rev Vol 27 August 2006 drugs, but there are technical issues in relation to collection in saliva under stimulated and nonstimulated conditions. of oral fl uid and in the variability of drug concentrations (of J Anal Toxicol 1993;17:338-41.
different drug types) in this fl uid. More research is needed to 16. Wong RC, Tran M, Tung JK. Oral fl uid drug tests: further the detection of drugs present in this fl uid which should effects of adulterants and foodstuffs. Forensic Sci Int allow improved reliability of detection of drugs. Similarly, future technological developments of on-site devices should 17. Maseda C, Hama K, Fukui Y, Matsubara K, Takahashi allow more sensitive and reliable detection of a number of S, Akane A. Detection of delta 9-THC in saliva by capillary GC/ECD after marihuana smoking. Forensic Sci Int 1986;32:259-66.
Competing interests: None declared.
18. Chu M, Wells DL, King RG, Farrar J, Drummer OH. The effect of blood in the oral cavity on breath alcohol References
analysis. J Clin Forensic Med 1998;5:114-8.
1. Kintz P, Samyn N. Use of alternative specimens: drugs 19. Teixeira H, Proenca P, Verstraete A, Corte-Real F, of abuse in saliva and doping agents in hair. Ther Drug Vieira DN. Analysis of Delta9-tetrahydrocannabinol in oral fl uid samples using solid-phase extraction and 2. Marquet P. Progress of liquid chromatography-mass high-performance liquid chromatography-electrospray spectrometry in clinical and forensic toxicology. Ther ionization mass spectrometry. Forensic Sci Int Drug Monit 2002;24:255-76.
3. Maurer HH. Liquid chromatography-mass spectrometry 20. Dickson SJ, Park A, Nolan SL, et al. The recovery of in forensic and clinical toxicology. J Chromatogr B drugs from oral fl uid collection devices. Forensic Sci Int Biomed Sci Appl 1998;713:3-25.
2006;In Press.
4. Van Bocxlaer JF, Clauwaert KM, Lambert WE, 21. Moore C, Vincent M, Rana S, Coulter C, Agrawal A, Deforce DL, Van den Eeckhout EG, De Leenheer AP. Soares J. Stability of Delta(9)-tetrahydrocannabinol Liquid chromatography-mass spectrometry in forensic (THC) in oral fl uid using the Quantisal(TM) collection toxicology. Mass Spectrom Rev 2000;19:165-214.
device. Forensic Sci Int Epub 2006 Jan 16 ahead of 5. Choo RE, Huestis MA. Oral fl uid as a diagnostic tool. Clin Chem Lab Med 2004;42:1273-87.
22. Verstraete AG. Oral fl uid testing for driving under the infl uence of drugs: history, recent progress and 6. Caplan YH, Goldberger BA. Alternative specimens for remaining challenges. Forensic Sci Int 2005;150:143- workplace drug testing. J Anal Toxicol 2001;25:396-9.
7. Walsh JM, de Gier JJ, Christopherson AS, Verstraete 23. Verstraete AG, Pierce A. Workplace drug testing in AG. Drugs and driving. Traffi c Inj Prev 2004;5:241-53.
Europe. Forensic Sci Int 2001;121:2-6.
8. Kadehjian L. Legal issues in oral fl uid testing. Forensic 24. George S. A snapshot of workplace drug testing in the Sci Int 2005;150:151-60.
UK. Occup Med (Lond) 2005;55:69-71.
9. Verstraete AG. Detection times of drugs of abuse in blood, 25. Cone EJ, Jenkins AJ. Saliva Drug Analysis. In: urine, and oral fl uid. Ther Drug Monit 2004;26:200-5.
Wong SHY, Sunshine I, eds. Handbook of Analytical 10. Drummer OH. Review: Pharmacokinetics of illicit drugs Therapeutic Drug Monitoring and Toxicology. Roca in oral fl uid. Forensic Sci Int 2005;150:133-42.
Baton, Florida: CRC Press, 1997, pp303-33.
11. Aps JK, Martens LC. Review: The physiology of 26. Cone EJ, Hillsgrove M, Darwin WD. Simultaneous saliva and transfer of drugs into saliva. Forensic Sci Int measurement of cocaine, cocaethylene, their metabolites, and "crack" pyrolysis products by gas chromatography- 12. O'Neal CL, Crouch DJ, Rollins DE, Fatah AA. The mass spectrometry. Clin Chem 1994;40:1299-305.
effects of collection methods on oral fl uid codeine 27. Jenkins AJ, Oyler JM, Cone EJ. Comparison of heroin concentrations. J Anal Toxicol 2000;24:536-42.
and cocaine concentrations in saliva with concentrations 13. Crouch DJ. Oral fl uid collection: the neglected variable in blood and plasma. J Anal Toxicol 1995;19:359-74.
in oral fl uid testing. Forensic Sci Int 2005;150:165-73.
28. Guidelines for testing drugs under international control 14. Schepers RJ, Oyler JM, Joseph RE, Jr., Cone EJ, in hair, sweat and saliva. Vienna, Austria: United Moolchan ET, Huestis MA. Methamphetamine and Nations, 1998.
amphetamine pharmacokinetics in oral fl uid and plasma 29. Jones AW. Inter- and intra-individual variations in the after controlled oral methamphetamine administration saliva/blood alcohol ratio during ethanol metabolism in to human volunteers. Clin Chem 2003;49:121-32.
man. Clin Chem 1979;25:1394-8.
15. Kato K, Hillsgrove M, Weinhold L, Gorelick DA, 30. Haeckel R, Peiffer U. Comparison of ethanol Darwin WD, Cone EJ. Cocaine and metabolite excretion concentration in saliva and blood from police controlled Clin Biochem Rev Vol 27 August 2006 I 155 persons. Blutalkohol 1992;29:342-9.
special focus on drug-facilitated crimes. Ther Drug 31. Jones AW. Measuring ethanol in saliva with the QED enzymatic test device: comparison of results with blood- 45. Samyn N, De Boeck G, Cirimele V, Verstraete A, Kintz and breath-alcohol concentrations. J Anal Toxicol P. Detection of fl unitrazepam and 7-aminofl unitrazepam in oral fl uid after controlled administration of rohypnol. 32. Menkes DB, Howard RC, Spears GF, Cairns J Anal Toxicol 2002;26:211-5.
ER. Salivary THC following cannabis smoking 46. Hart BJ, Wilting J, de Gier JJ. The stability of correlates with subjective intoxication and heart rate. benzodiazepines in saliva. Methods Find Exp Clin Psychopharmacology (Berl) 1991;103:277-9.
33. Huestis MA, Cone EJ. Relationship of delta-9- 47. Ritschel WA, Parab PV, Denson DD, Coyle DE, Gregg tetrahydrocannabinol concentrations in oral fl uid and RV. Absolute bioavailability of hydromorphone after plasma after controlled administration of smoked peroral and rectal administration in humans: saliva/ cannabis. J Anal Toxicol 2004;28:394-9.
plasma ratio and clinical effects. J Clin Pharmacol 34. O'Neal CL, Crouch DJ, Rollins DE, Fatah A, Cheever ML. Correlation of saliva codeine concentrations with 48. Matin SB, Wan SH, Knight JB. Quantitative plasma concentrations after oral codeine administration. determination of enantiomeric compounds. I-- J Anal Toxicol 1999;23:452-9.
Simultaneous measurement of the optical isomers 35. Kim I, Barnes AJ, Oyler JM, et al. Plasma and oral fl uid of amphetamine in human plasma and saliva using pharmacokinetics and pharmacodynamics after oral chemical ionization mass spectrometry. Biomed Mass codeine administration. Clin Chem 2002;48:1486-96.
36. Mendelson J, Upton RA, Everhart ET, Jacob P 3rd, 49. Chen ZR, Bochner F, Somogyi A. Pharmacokinetics of Jones RT. Bioavailability of sublingual buprenorphine. pholcodine in healthy volunteers: single and chronic J Clin Pharmacol 1997;37:31-7.
dosing studies. Br J Clin Pharmacol 1988;26:445-53.
37. Nath RP, Upton RA, Everhart ET, et al. Buprenorphine 50. Tracqui A, Ludes B. HPLC-MS for the determination of pharmacokinetics: relative bioavailability of sublingual sildenafi l citrate (Viagra) in biological fl uids. Application tablet and liquid formulations. J Clin Pharmacol to the salivary excretion of sildenafi l after oral intake. J Anal Toxicol 2003;27:88-94.
38. Cone EJ, Dickerson SL, Darwin WD, Fudala P, Johnson 51. Drobitch RK, Svensson CK. Therapeutic drug RE. Elevated drug saliva levels suggest a "depot-like" monitoring in saliva. An update. Clin Pharmacokinet effect in subjects treated with sublingual buprenorphine. NIDA Res Monogr 1990;105:569.
52. Liu H, Delgado MR. Therapeutic drug concentration 39. Curvall M, Elwin CE, Kazemi-Vala E, Warholm C, monitoring using saliva samples. Focus on Enzell CR. The pharmacokinetics of cotinine in plasma anticonvulsants. Clin Pharmacokinet. 1999;36:453-70.
and saliva from non-smoking healthy volunteers. Eur J 53. Vasudev A, Tripathi KD, Puri V. Correlation of serum Clin Pharmacol 1990;38:281-7.
and salivary carbamazepine concentration in epileptic 40. Curvall M, Vala EK, Enzell CR, Wahren J. Simulation patients: implications for therapeutic drug monitoring. and evaluation of nicotine intake during passive smoking: Neurol India 2002;50:60-2.
cotinine measurements in body fl uids of nonsmokers 54. Miles MV, Tang PH, Glauser TA, et al. Topiramate given intravenous infusions of nicotine. Clin Pharmacol concentration in saliva: an alternative to serum Ther 1990;47:42-9.
monitoring. Pediatr Neurol 2003;29:143-7.
41. Giles HG, Miller R, Macleod SM, Sellers EM. Diazepam 55. Moolchan ET, Umbricht A, Epstein D. Therapeutic and N-desmethyldiazepam in saliva of hospital drug monitoring in methadone maintenance: choosing inpatients. J Clin Pharmacol 1980;20:71-6.
a matrix. J Addict Dis 2001;20:55-73.
42. DiGregorio GJ, Piraino AJ, Ruch E. Diazepam 56. Sagawa K, Mohri K, Shimada S, Shimizu M, Muramatsu concentrations in parotid saliva, mixed saliva, and J. Disopyramide concentrations in human plasma and plasma. Clin Pharmacol Ther 1978;24:720-5.
saliva: comparison of disopyramide concentrations in 43. de Gier JJ, 't Hart BJ, Nelemans FA, Bergman H. saliva and plasma unbound concentrations. Eur J Clin Psychomotor performance and real driving performance of outpatients receiving diazepam. Psychopharmacology 57. Mortier KA, Renard V, Verstraete AG, Van Gussem A, Van Bel e S, Lambert WE. Development and validation of a 44. Concheiro M, Villain M, Bouchet S, Ludes B, Lopez- liquid chromatography-tandem mass spectrometry assay Rivadulla M, Kintz P. Windows of detection of for the quantifi cation of docetaxel and paclitaxel in human tetrazepam in urine, oral fl uid, beard, and hair, with a plasma and oral fl uid. Anal Chem 2005;77:4677-83.
156 I Clin Biochem Rev Vol 27 August 2006 58. Spielberg F, Critchlow C, Vittinghoff E, et al. Home Clin Chem 2003;49:1498-503.
collection for frequent HIV testing: acceptability of 72. Lachenmeier K, Musshoff F, Madea B. Determination oral fl uids, dried blood spots and telephone results. HIV of opiates and cocaine in hair using automated enzyme Early Detection Study Group. AIDS 2000;14:1819-28.
immunoassay screening methodologies followed by 59. Soldin SJ, Rakhmanina NY, Spiegel HM, Sever gas chromatographic-mass spectrometric (GC-MS) JL. Therapeutic drug monitoring for patients with confi rmation. Forensic Sci Int 2006;159:189-99.
HIV infection: Children's National Medical Center, 73. Cooper G, Wilson L, Reid C, Baldwin D, Hand C, Washington DC experience. Ther Drug Monit Spieher V. Validation of the Cozart microplate EIA for cocaine and metabolites in oral fl uid. J Anal Toxicol 60. Gallagher P, Leitch MM, Massey AE, McAllister Williams RH, Young AH. Assessing cortisol and 74. Cooper G, Wilson L, Reid C, Baldwin D, Hand C, dehydroepiandrosterone (DHEA) in saliva: effects of Spiehler V. Comparison of GC-MS and EIA results for collection method. J Psychopharmacol 2006 Jan 9 ahead the analysis of methadone in oral fl uid. J Forensic Sci of print. Jan 9;[Epub ahead of print] 61. Mylonas PG, Makri M, Georgopoulos NA, et al. Adequacy 75. Kacinko SL, Barnes AJ, Kim I, et al. Performance of saliva 17-hydroxyprogesterone determination using characteristics of the Cozart RapiScan Oral Fluid Drug various collection methods. Steroids 2006;71:273-6.
Testing System for opiates in comparison to ELISA and 62. Kintz P, Cirimele V, Ludes B. Codeine testing in GC/MS following controlled codeine administration. sweat and saliva with the Drugwipe. Int J Legal Med Forensic Sci Int 2004;141:41-8.
76. Barnes AJ, Kim I, Schepers R, et al. Sensitivity, 63. Samyn N, van Haeren C. On-site testing of saliva and specifi city, and effi ciency in detecting opiates in oral sweat with Drugwipe and determination of concentrations fl uid with the Cozart Opiate Microplate EIA and GC- of drugs of abuse in saliva, plasma and urine of suspected MS following controlled codeine administration. J Anal users. Int J Legal Med 2000;113:150-4.
64. Grönholm M, Lillsunde P. A comparison between on- 77. Niedbala RS, Kardos KW, Fritch DF, et al. Detection site immunoassay drug-testing devices and laboratory of marijuana use by oral fl uid and urine analysis results. Forensic Sci Int 2001;121:37-46.
following single-dose administration of smoked and 65. Kolbrich EA, Kim I, Barnes AJ, et al. Cozart RapiScan oral marijuana. J Anal Toxicol 2001;25:289-303.
Oral Fluid Drug Testing System: an evaluation of 78. Kemp P, Sneed G, Kupiec T, Spiehler V. Validation of sensitivity, specifi city, and effi ciency for cocaine a microtiter plate ELISA for screening of postmortem detection compared with ELISA and GC-MS following blood for opiates and benzodiazepines. J Anal Toxicol controlled cocaine administration. J Anal Toxicol 79. Allen KR, Azad R, Field HP, Blake DK. Replacement of 66. Samyn N, De Boeck G, Verstraete AG. The use of oral immunoassay by LC tandem mass spectrometry for the fl uid and sweat wipes for the detection of drugs of abuse routine measurement of drugs of abuse in oral fl uid. Ann in drivers. J Forensic Sci 2002;47:1380-7.
Clin Biochem 2005;42:277-84.
67. Pichini S, Navarro M, Farre M, et al. On-site testing 80. Gunnar T, Ariniemi K, Lillsunde P. Validated of 3,4-methylenedioxymethamphetamine (ecstasy) in toxicological determination of 30 drugs of abuse as saliva with Drugwipe and Drugread: a controlled study optimized derivatives in oral fl uid by long column fast in recreational users. Clin Chem 2002;48:174-6.
gas chromatography/electron impact mass spectrometry. 68. Laloup M, Tilman G, Maes V, et al. Validation of an J Mass Spectrom 2005;40:739-53.
ELISA-based screening assay for the detection of 81. Wylie FM, Torrance H, Anderson RA, Oliver JS. Drugs amphetamine, MDMA and MDA in blood and oral fl uid. in oral fl uid Part I. Validation of an analytical procedure Forensic Sci Int 2005;153:29-37.
for licit and illicit drugs in oral fl uid. Forensic Sci Int 69. Kupiec T, DeCicco L, Spiehler V, Sneed G, Kemp P. Choice of an ELISA assay for screening postmortem 82. Mortier KA, Maudens KE, Lambert WE, et al. blood for amphetamine and/or methamphetamine. J Simultaneous, quantitative determination of opiates, Anal Toxicol 2002;26:513-8.
amphetamines, cocaine and benzoylecgonine in oral 70. De Giovanni N, Fucci N, Scarlata S, Donzelli G. fl uid by liquid chromatography quadrupole-time- Buprenorphine detection in biological samples. Clin of-fl ight mass spectrometry. J Chromatogr B Analyt Chem Lab Med 2005;43:1377-9.
Technol Biomed Life Sci 2002;779:321-30.
71. Kim I, Barnes AJ, Schepers R, et al. Sensitivity and 83. Wood M, Laloup M, Ramirez Fernandez Mdel M, specifi city of the Cozart microplate EIA cocaine oral et al. Quantitative analysis of multiple illicit drugs in fl uid at proposed screening and confi rmation cutoffs. preserved oral fl uid by solid-phase extraction and liquid Clin Biochem Rev Vol 27 August 2006 I 157 chromatography-tandem mass spectrometry. Forensic Sci Int 2005;152:149-55.
Sci Int 2005;150:227-38.
98. Kintz P, Bernhard W, Villain M, Gasser M, Aebi B, 84. Dams R, Murphy CM, Choo RE, Lambert WE, De Cirimele V. Detection of cannabis use in drivers with the Leenheer AP, Huestis MA. LC-atmospheric pressure drugwipe device and by GC-MS after Intercept device chemical ionization-MS/ MS analysis of multiple collection. J Anal Toxicol 2005;29:724-7.
illicit drugs, methadone, and their metabolites in oral 99. Cone EJ, Presley L, Lehrer M, et al. Oral fl uid testing fl uid following protein precipitation. Anal Chem for drugs of abuse: positive prevalence rates by Intercept immunoassay screening and GC-MS-MS confi rmation 85. Draft Standard: Procedures for the collection, detection and suggested cutoff concentrations. J Anal Toxicol and quantitation of drugs in oral fl uid. Standards Australia, 2006.
100. Kintz P, Villain M, Concheiro M, Cirimele V. Screening 86. Niedbala R, Kardos K, Salamone S, Fritch D, Bronsgeest and confi rmatory method for benzodiazepines and M, Cone EJ. Passive cannabis smoke exposure and oral hypnotics in oral fl uid by LC-MS/MS. Forensic Sci Int fl uid testing. J Anal Toxicol 2004;28:546-52.
87. Rohrig TP, Moore C. The determination of morphine in 101. Barrett C, Good C, Moore C. Comparison of point- urine and oral fl uid following ingestion of poppy seeds. of-collection screening of drugs of abuse in oral fl uid J Anal Toxicol 2003;27:449-52.
with a laboratory-based urine screen. Forensic Sci Int 88. Maurer HH. Advances in analytical toxicology: the current role of liquid chromatography-mass spectrometry 102. Cirimele V, Villain M, Mura P, Bernard M, Kintz P. Oral in drug quantifi cation in blood and oral fl uid. Anal fl uid testing for cannabis: On-site Oraline(R) IV s.a.t. Bioanal Chem 2005;381:110-8.
device versus GC/MS. Forensic Sci Int 2006; July 17; 89. National Association of Testing Authorities, see www.
[Epub ahead of print] nata.asn.au (accessed March 27, 2006). 103. Lamey PJ, Nolan A. The recovery of human saliva using 90. de la Torre R, Segura J, de Zeeuw R, Williams J. the Salivette system. Eur J Clin Chem Clin Biochem. Recommendations for the reliable detection of illicit drugs in urine in the European Union, with special 104. Biermann T, Schwarze B, Zedler B, Betz P. On-site attention to the workplace. EU Toxicology Experts testing of illicit drugs: the use of the drug-testing device working Group. Ann Clin Biochem 1997;34:339-44.
"Toxiquick". Forensic Sci Int 2004;143:21-5.
91. Peters FT, Maurer HH. Bioanalytical method validation 105. Cone EJ. Saliva testing for drugs of abuse. Ann N Y and its implications for forensic and clinical toxicology Acad Sci 1993;694:91-127.
- A review. Accred Qual. Assur 2002;7:441-9.
106. Inaba T, Kalow W. Salivary excretion of amobarbital in 92. Clarke J, Wilson JF. Profi ciency testing (external quality man. Clin Pharmacol Ther 1975;18:558-62.
assessment) of drug detection in oral fl uid. Forensic Sci 107. van der Graaff M, Vermeulen NP, Heij P, Boeijinga JK, Int 2005;150:161-4.
Breimer DD. Pharmacokinetics of orally administered 93. Walsh JM, Flegel R, Crouch DJ, Cangianelli L, Baudys hexobarbital in plasma and saliva of healthy subjects. J. An evaluation of rapid point-of-collection oral fl uid Biopharm Drug Dispos 1986;7:265-72.
drug-testing devices. J Anal Toxicol 2003;27:429-39.
108. Navarro M, Pichini S, Farre M, et al. Usefulness of 94. Crouch DJ, Walsh JM, Flegel R, Cangianelli L, Baudys saliva for measurement of 3,4-methylenedioxymetham J, Atkins R. An evaluation of selected oral fl uid point- phetamine and its metabolites: correlation with plasma of-collection drug-testing devices. J Anal Toxicol drug concentrations and effect of salivary pH. Clin 95. Cooper G, Wilson L, Reid C, Main L, Hand C. Evaluation 109. Chikhi-Chorfi N, Pham-Huy C, Galons H, et al. Rapid of the Cozart((R)) RapiScan drug test system for opiates determination of methadone and its major metabolite and cocaine in oral fl uid. Forensic Sci Int 2005;150:239- in biological fl uids by gas-liquid chromatography with thermionic detection for maintenance treatment 96. Moore L, Wicks J, Spiehler V, Holgate R. Gas of opiate addicts. J Chromatogr B Biomed Sci Appl chromatography-mass spectrometry confi rmation of Cozart RapiScan saliva methadone and opiates tests. J 110. Kankaanpaa A, Gunnar T, Ariniemi K, Lillsunde P, Anal Toxicol 2001;25:520-4.
Mykkanen S, Seppala T. Single-step procedure for 97. Toennes SW, Kauert GF, Steinmeyer S, Moeller MR. gas chromatography-mass spectrometry screening Driving under the infl uence of drugs -- evaluation of and quantitative determination of amphetamine-type analytical data of drugs in oral fl uid, serum and urine, stimulants and related drugs in blood, serum, oral fl uid and correlation with impairment symptoms. Forensic and urine samples. J Chromatogr B Analyt Technol Biomed Life Sci 2004;810:57-68.
158 I Clin Biochem Rev Vol 27 August 2006 111. Wood M, De Boeck G, Samyn N, et al. Development of a rapid and sensitive method for the quantitation of amphetamines in human plasma and oral fl uid by LC- MS-MS. J Anal Toxicol 2003;27:78-87.
112. Quintela O, Cruz A, Castro A, Concheiro M, Lopez- Rivadulla M. Liquid chromatography-electrospray ionisation mass spectrometry for the determination of nine selected benzodiazepines in human plasma and oral fl uid. Journal of Chromatography B Analyt Technol Biomed Life Sci 2005;825:63-71.
Clin Biochem Rev Vol 27 August 2006 I 159

Source: http://www.etop.org.tw/dsp/E17.php?c=dspfile&id=YzI3MjhhMzJkMjE0MjdkMWVkNjM0NzRlZjNkZDI2N2IucGRm