Originally Published IVD Technology July/August 2002
Assay Development
Testing urine for drugs of abuse
An overview of testing methods demonstrates opportunities for assay manufacturers in on-site workplace testing.
Alan H. B. Wu
In vitro testing for drugs of abuse has clinical and forensic applications. In clinical toxicology, drug testing is performed when individuals exhibit signs of intoxication or overdose such as agitation, depression of the respiratory or central nervous system, and organ-specific reactions, including hepatic failure, cardiac arrhythmia, and severe metabolic acidosis. The finding of particular drugs can help in determining countermeasures, such as naloxone for opiate poisoning and N-acetylcysteine for acetaminophen overdose. Testing also is used to verify drug abstinence in various rehabilitation and criminal justice programs.
Most drugs of abuse are either illegal or controlled by prescription. Consequently, samples are assayed for the presence of such drugs in various legal and quasilegal circumstances. Drug testing in the workplace is prevalent, particularly in public safety–related occupations. The Substance Abuse and Mental Health Services Administration (SAMHSA) in the U.S. Department of Health and Human Services (HHS) has found that drug-abusing employees are twice as likely as nonabusers to have unexcused absences or voluntarily leave an employer. And while 5.5% of nonabusers are involved in workplace accidents, 7.5% of abusers are.1 Although drug testing is not required in the private sector as it is in the public, some large corporations have instituted testing programs to mitigate the costs involved both in having drug abusers on the job and in replacing them.
Drug testing of federal workers is administered by SAMHSA through laboratories certified by the National Laboratory Certification Program. Mandatory test guidelines apply to employees of the federal government, the military, the transportation industry, and nuclear regulatory agencies. Drugs tested under the guidelines are amphetamine, methamphetamine, codeine, morphine, heroin-specific metabolite (6-acetylmorphine), phencyclidine (PCP), cocaine-specific metabolite (benzoylecgonine), and marijuana metabolite (9-carboxy tetrahydrocannabinol; THC).
Laboratory quality control and quality assurance procedures are strictly governed. Certified laboratories must regularly employ single-blind and double-blind controls to minimize the possibility of errors and inaccurate results. They also have procedures for collecting and testing split samples; if a specimen is confirmed positive, the donor can have the second, unopened aliquot of the original sample analyzed by another certified laboratory.
Mandatory guidelines and the certification program afford drug testing laboratories some protection against legal claims. Litigated challenges to their procedures and methodologies have been unsuccessful, but negligent laboratories are not free from liability. Results of all tests are reviewed by a qualified certifying scientist employed by the laboratory, and also by an independent medical review officer (MRO), a physician trained in interpreting drug test results.
Screening Methods
All SAMHSA-mandated testing is performed in two steps: immunoassay screening followed by gas chromatographic confirmation. For some drugs, cutoff concentrations differ between steps. All samples must be submitted and tested with strict chain-of-custody documentation.
Specimens. The preferred specimen for drug testing now is urine, although alternative specimens such as hair, oral fluids (saliva), and sweat have potential (see Table I). The half-life of drugs in blood is short, making blood specimens less useful for routine drug screening (though a high concentration of active drug will suggest recent usage and, sometimes, likelihood of impairment).
|
Sample
|
Advantages
|
Disadvantages
|
|
Blood |
|
|
| Hair |
|
|
| Saliva |
|
|
| Sweatable |
|
|
| Urine |
|
|
| Table I. A comparison of the types of samples that can be used in drugs-of-abuse testing. | ||
Hair, if long enough, can be a record of drug use over weeks and even months. No convenient screening procedures for the measurement of drugs in hair exist, and hair tests lack the sensitivity of urine tests for occasional or single-time drug use.
Oral fluid and sweat samples are less prone to adulteration than urine but, because drug concentrations are lower in oral and sweat samples, they require more-sensitive test methods. Sweat patches are being considered for use in drug-compliance programs. Applied with tamper-evident tape, and later removed by the program monitor for laboratory analysis, a sweat patch could track an individual over several days. Collecting samples large enough for confirmation testing and retesting would be difficult, however.
Urine specimens are preferred for testing because large sample volumes can be collected noninvasively. Drugs generally remain detectable in urine for two to three days, longer than in blood. THC can remain positive in urine for several weeks after the last use, especially in chronic users. A positive THC urine drug test does not necessarily imply impairment of the donor, however, because THC is usually inactivated by the liver within a few minutes or hours after administration.
) |
|
Figure
1. General structure of the barbiturate family of chemicals. In bartiuric
acid, from which the drugs are derived, a hydrogen atom occupies the positions
R1 and R2.
(click to enlarge) |
Drug screening involves performing tests for a broad spectrum of controlled substances. The tests are relatively simple and inexpensive to perform. Their objective is to quickly rule out the evident use by a donor of the most common drugs of abuse. Drug screening tests are usually directed against a class of drugs rather than individual substances; for example, the barbiturate assay will produce a positive result with most compounds that have a barbituric acid chemical structure (see Figure 1).
Competitive Immunoassays. The most common analytical approach to urine drug screening is the use of competitive immunoassays, in which specific antibodies bind to targeted chemical atoms and functional groups (see Figure 2). A fixed amount of labeled drug material from the test kit (marked with a radioactive substance, enzyme, fluorescent tag, or colored particle) competes for antibody binding sites with the variable amount of unlabeled drug in the urine sample. When the binding sites are saturated, the amount of either free or bound labeled drug is measured. For a bound-labeled drug, low sample-drug concentration will produce a high analytical signal; high concentration will produce a low analytical signal.
) |
|
Figure
2. In a competitive immunoassay, (a) labeled drugs (solid circles) and
unlabeled drugs (open circles) compete for a limited number of antibody
binding sites (y shapes) until (b) the sites are saturated and bound drugs
can be separated from free drugs. When the amount of bound labeled drug
is measured, (c) High analytical signal signifies low drug concentration
in the sample and (d) a low signal indicates high concentration.
(click to enlarge) |
Assays must first be calibrated by measuring the response of a urine sample containing the drug against the response of a calibrator with a known drug concentration. All qualitative drug assays have a threshold or cutoff level. Signals above the cutoff are deemed negative readings, while those below the cutoff are positive. Because minimum thresholds are necessary in drug screening, a urine sample containing the drug of interest at a concentration below the cutoff level is reported as negative despite the drug's presence. Lowering the threshold to reduce the number of false negatives might result in more false positives, a potentially greater problem.
Cutoff levels used for urine drug testing were originally established at a drug concentration that produced an analytical signal some multiplicative factor above the noise level (the signal obtained from drug-free urine). The HHS mandatory drug testing guidelines for federal employees set cutoffs that have largely been adopted by nonmandated programs. These thresholds have been modified by SAMHSA on several occasions. Cutoff concentrations for drug assays not included in mandatory programs, and for new drug tests, are set by the manufacturer, who must justify them to FDA.
) |
|
Figure
3. A calibration curve is constructed from known drug standards.
(click to enlarge) |
The analytical response to drug concentration is sigmoidal; the signal is flat at low and high concentrations and changes rapidly around the cutoff concentration (see Figure 3). Because of this, most immunoassays for drug testing are qualitative.
An exception is the cloned enzyme donor immunoassay, or CEDIA, manufactured by Microgenics (Fremont, CA). Its linear calibration enables this assay to produce quantitative results. A very high drug concentration might imply more-recent drug use than a concentration nearer the cutoff. However, the immunoassay limitations discussed next make interpretation of urine drug concentrations tenuous. Still, quantitative immunoassay analysis is certainly useful for range finding for the second, more-definitive test; that is, for determining the dilution of the urine sample necessary to ensure that the confirmation test result falls within the linear range of the confirmation assay.
Immunoassays are available in two basic formats: liquid-phase assays that allow automated high-throughput laboratory analysis of a large number of samples and drug classes, and solid-phase-reagent single- and multianalyte on-site testing devices (see Table II).2 On-site devices cost substantially more per test but, because results can be obtained immediately by specimen collectors, they save on transportation costs; only the few samples that screen positive need to be sent to the laboratory. Their quick turnaround time can also help emergency room personnel make good patient-management decisions.
|
Attribute
|
Automated
Test
|
On-Site
Test
|
|
Cost of each |
Low (<$1 each) | High ($5–25) |
| Instrumentation | Expensive | None or minimal |
| Sample delivery | Required | Not required |
| Sensitivity | High | Moderate |
| Specificity | Antibody dependent | Antibody dependent |
| Adulteration testing | Available | Separate dipsticks available |
| Table II. A comparison of automated and point-of-care immunoassay screening tests. | ||
) |
|
Figure
4. The chemical structures of the tricyclic antidepressants amitriptyline
and imipramine resemble those of drug analogs without antidepressant effect.
(click to enlarge) |
Recently, totally automated drug-screening systems were approved for on-site testing. The Triage Device Reader from Biosite (San Diego) reads an inserted strip and prints out the results. The Web-enabled eScreen workstation introduced by eScreen Inc. (Overland Park, KS) samples urine from a cup, performs analysis, and transmits results via a secure Internet connection.
Immunoassay Limitations. Antibodies used in immunoassays often cannot recognize subtle differences in chemical structure between the targeted drugs (and their metabolites) and analogous compounds. Analytical specificity is therefore a major challenge for drug testing. Pharmaceuticals and other substances with structures similar to those of the target drugs are often detected by immunoassays even though they are not members of the drug group of interest. The tricyclic antidepressants, for example, have molecular arrangements similar to those of tricyclic and nontricyclic compounds that have no antidepressant properties (see Figure 4). The presence of the analogs in urine may lead to false-positive results for the tricyclic antidepressants.3
Assays for amphetamines cross-react with over-the-counter sympathomimetic amines because the latter are chemically similar to amphetamines.4 Sometimes even totally unrelated chemicals produce unexpected false-positive results (see Figure 5).5 These false-positives are not a widespread problem and seem to be related to reagent manufacture. The interferents do not produce a false-positive result with the chromatographic assays now employed for confirmation—a justification of the two-tiered approach to drug testing.
) |
|
Figure
5. Unexpected false positive results have arisen with (a) oxaprozin in
the oxazepam immunoassay, (b) efavirenz in the 9-carboxy-tetrahydrocannabinol
immunoassay, nad (c) diphenhydramine and venlafaxine in the phencyclidine
(PCP) immunoassay.
(click to enlarge) |
The specificity limitations of some immunoassays have created diagnostic confusion, particularly when the interfering drug is popular. For example, the widely used antihistamine diphenhydramine can interfere with the PCP immunoassay. The extremely low prevalence of PCP abuse in most places means that a positive PCP test result is usually a false-positive.
Physicians and other caregivers tend to assume that all drugs with a particular function will produce the same clinical laboratory result. Thus, they might expect that propoxyphene (Darvon) and meperidine (Demerol) would cross-react with the opiate assay—since all are analgesics—and can be surprised that they do not do so because of the quite different chemical structures of these drugs.
More confusion arises when different cutoff concentrations are used for different purposes. A positive urine test indicates only possible drug use, not impairment or toxicity. In certain cases (e.g., marijuana and cocaine), nonactive metabolites are measured by the immunoassay. Some urine drug assays make available multiple cutoff concentrations to address these issues. A low cutoff of 300 ng/ml for tricyclic antidepressants will capture both therapeutic use and toxic overdoses, whereas a 1000-ng/ml cutoff concentration will more likely correlate with toxic presentations. A high opiate cutoff concentration of 2000 ng/ml is being used in workplace drug testing to avoid false-positives resulting from poppy seed consumption. For clinical toxicology purposes, 300 ng/ml is more appropriate.
) |
|
Figure
6. Gas chromatographic separation of cocaine from other drugs in a urine
sample, with the corresponding mass spectrum.
(click to enlarge) |
A more subtle issue is differential antibody sensitivity to members of the same drug class. For example, in an enzyme-multiplied opiate immunoassay, morphine will produce a positive result above the 300-ng/ml cutoff. The narcotic analgesic oxycodone will cross-react with the antibody but at a rate less than 10% of that of morphine; thus, the oxycodone concentration must be greater than 3000 ng/ml before the drug is detected. But although the assay targets morphine, the cross-reactivity of the antibody for codeine is even higher; a codeine concentration of 240 ng/ml will produce a positive result at the 300-ng/ml morphine cutoff.
Owing to the inherent limitations of immunoassay screening, particularly in the area of specificity, all positive urine drug tests that have forensic purpose must be confirmed by a more definitive and legally defensible procedure. This confirmation test is intended to determine whether the drug or drugs in question are present in the sample and were responsible for the initial test result, or whether that result is traceable to an interfering drug or metabolite. Confirmation analysis must involve retesting a sample that was presumptively positive on the initial test, not collecting and testing a new sample from the donor.
Confirmation
Analysis
Chromatographic procedures are usually employed in confirmation analysis of urine drug samples. Thin-layer (TLC), liquid (LC), and gas chromatography (GC) all separate drugs from a sample according to their inherent physical properties of solubility, volatility, and polarity. The sample is exposed to two different phases, stationary and mobile, for which each compound has a differential affinity. Drugs or substances favoring the stationary phase will lag behind, and therefore separate from, compounds that favor the mobile phase.
For TLC, the polar stationary phase is silica coated onto filter paper or glass and the nonpolar mobile phase is an organic solvent. Detection of compounds is based on their reaction to specific dyes and reagents added after the separation.
) |
|
Figure
7. Electron bombardment of cocaine breaks the drug into fragments whose
number and size are quantified. Unlike the illustrative glass tumbler,
the result of mass fragmentation of cocaine or other chemical compounds
is both predictable and reproducible.
(click to enlarge) |
The stationary phase in LC is a nonpolar solid, and the mobile phase is aqueous. As each compound is removed from the column, LC measures its ultraviolet absorbance spectrum and compares it to a library of known drugs and compounds.
In GC, stationary and mobile phases are liquid and gaseous, respectively. After separation, the individual drugs are detected by mass spectrometry (MS), a technique involving reproducibly fragmenting a drug by bombarding it with electrons (see Figure 6). The determinable size and number of fragments comprise the mass spectrum. Drug identification is based on the fragmentation pattern, which is highly specific for each drug (see Figure 7).
While their principles are similar, each chromatographic procedure has distinctive advantages and disadvantages in drug-testing urine (see Table III). TLC and LC are useful for nonforensic purposes because they are faster and cheaper than GC/MS. Many laboratories use these to support clinical toxicology.
|
Attribute
|
Screening
|
TLC
|
HPLC
|
GC/MS
|
|
Sensitivity |
Excellent (<10 ng/ml) | Fair (1000 ng/ml) | Good (500 ng/ml) | Excellent (<10 ng/ml) |
| Specificity | Often poor | Fair | Fair | Definitive |
| Labor required | Low | Medium | Low | Very high |
| Sensitivity | High | Moderate | ||
| Assay turnaround time | Fast (<1 hour) | Slow (3 hours) | Fast (20 minutes) | Very slow (>8 hours) |
| Menu of tests | Adequate | Wide (>200) | Wide (>200) | Very wide |
| Table III. A comparison of immunoassay screening, thin-layer chromatography (TLC), high-performance liquid chromatography (HPLC) with a rapid-scanning or diode-array ultraviolet detector, and gas chromatography/mass spectrometry (GC/MS) as assays for testing urine for drugs of abuse. | ||||
MS is considered the gold standard for forensic drug testing. Data from immunoassays or chromatographic procedures not backed by GC/MS results will be challenged in criminal and civil court cases because of their inherent limitations. But GC/MS has its own analytical limitations.6 To extract samples from urine and derivatize them to more-volatile compounds prior to injection onto the GC requires expensive equipment and specially trained technicians. Forensic drug testing also requires chain-of-custody documentation of every action taken with the specimen. A breach in the record of the specimen trail could invalidate the analytical test result.
Limitations of Drug Testing
Inadvertent positive
results have been an issue for drug testing programs.7–9 Some
drugs are consumed as food, and some consumer medications contain drugs of abuse
(Tylenol 3 contains codeine as well as acetaminophen). In such cases, analytically
true-positive test results are false-positive indicators of drug abuse by the
individual. Review of prescriptions and dietary habits by an MRO is critical
in clarifying these situations.
Some individuals resort to diluting, substituting, or adulterating urine samples to avoid the consequences of a positive drug test result.10 SAMHSA has established testing criteria to identify these circumstances.
The object of dilution is to reduce the concentration of illicit substance in urine to below the cutoff level, either in vitro, by adding water, saline, or other fluid after collection, or in vivo, by consuming diuretics or plasma volume expanders to stimulate excess water excretion. SAMHSA considers a sample dilute when creatinine is below 20 mg/dl and specific gravity is less than 1.003.
Replacement of donor urine with nonurine fluid is substitution. To succeed as a ruse, the substituted fluid must be warm, as the collector is instructed to check specimen temperature within 3 minutes of collection. According to SAMHSA, a sample is deemed substituted if creatinine is less than 5 mg/dl and the specific gravity is 1.000 or 1.001.
SAMHSA has no criteria for detecting replacement of donor urine by other human or nonhuman urine. In this situation, the determination of substitution can only be accomplished by DNA analysis of renal epithelial cells in the questionable sample and a fresh (monitored) sample using polymerase chain reaction techniques.11 Commercial dot blot methods, such as the PM1 + DQa kit from PerkinElmer Corp. (Foster City, CA), have long been used for forensic identity testing. The degree of discrimination of these tests (the likelihood of urine samples from two donors having the same genotype) exceeds 1 in 100,000. DNA testing has been used largely to resolve claims that samples were mixed up by the collection station or laboratory.
Adulteration involves putting into a specimen a substance designed to mask or destroy the drug or drug metabolite it may contain, or to adversely affect the assay reagent, an act constituting potentially criminal fraud of the drug testing system. Since most immunoassays are designed to be performed under narrow pH and ionic strength ranges, the addition of acids, bases, salts, or detergents will invalidate them. However, these adulterants generally do not affect GC/MS test results.
Commercially available adulterants, however, are oxidizers that change drugs such as marijuana and opiates into entirely different compounds. They can negate confirmation tests. Drug testing laboratories are required to test for the presence of these oxidants. The active ingredients of Urine Aid, Klear, Urine Luck, and Stealth are glutaraldehyde, nitrite, pyridinium chlorochromate, and peroxidase, respectively. Oxidation-reduction indicators using colorimetric end points are available to detect these compounds, though rapid consumption of the oxidizing capability of Stealth makes that adulterant difficult to detect by these means.
Challenges and Opportunities
The Internet provides
drug abusers with an education in common drug testing practices and how to exploit
the limitations of screening assays, as well as a shopping center for sample
adulterants. Laboratories constantly need assays to detect the presence of ever
more deceptive adulterants.
Changing patterns of drug abuse pose another challenge. New drugs like oxycodone, oxymorphone, Ecstasy (3,4-methylenedioxymethamphetamine), and gamma hydroxybutyrate become popular partly because they are not included in current drug testing programs. Regulatory agencies, drug testing laboratories, and manufacturers of diagnostic reagents have to keep up. Mandatory drug testing guidelines are evolving, but slowly. Ecstasy has been proposed for addition to the menu of analytes, but not yet to the semisynthetic opiates.
And screening tests for some existing drugs need improvement. The cocaine metabolite and THC immunoassays are model tests, producing no false positives or false negatives, but the failure rates for benzodiazepines and tricyclic antidepressants are so high that many physicians in emergency room settings recommend not ordering these urine tests because they produce more confusion than clinical value.
More-specific antibodies must be incorporated into next-generation assays. Urine drug testing is an area in which nanotechnology and gene chips have potential for high-volume, low-cost multivariate analysis. Regulatory compliance is a barrier to innovation in this area that the IVD industry wants removed. FDA is under pressure from diagnostics manufacturers to deregulate kits for on-site workplace drug testing, as they are not used to diagnose disease.
Some good news can already be reported. Drug testing accuracy has improved with recent regulatory and research advancements. DNA identity techniques have helped certifying scientists and MROs sort out sample mix-up claims. The high accuracy of DNA technology also is a litigation deterrent. In the matter of resolving claims that poppy seed ingestion caused an opiate assay to be positive, research has discovered that poppy seed eaters' urine contains thebaine, an effective marker in that it is absent in pharmaceutical opiates and street heroin.12 Finding thebaine confirms a claim of poppy seed use; however, it does not rule out the possibility of opiate drug abuse by the donor.
New legislation will help reduce the incidence of inadvertent positives. Noteworthy are bans on the consumption of hemp products that contain THC and the use of growth-enhancing drugs and steroids for farm-raised meat animals.
Conclusion
As analytical technology
continues to improve, on-site drug testing will grow in popularity at the expense
of the central laboratory approach, particularly for workplace applications.
Employers find nearly instant results very attractive. However, the road is
not clear. The serious limitations of immunoassay screening, along with the
persisting need for nonnegative workplace samples to be sent to toxicology laboratories,
raise concerns. The frequency with which an on-site positive result becomes
negative after confirmation analysis in the laboratory must be reduced. This
is to minimize the number of urine donors stigmatized even temporarily during
the period between on-site test administration and laboratory confirmation.
Also, controls and proficiency testing must be in place to ensure the quality
of on-site testing by individuals not trained in laboratory techniques. Finally,
inappropriately administered on-site drug tests could subject the collection
site to legal liability. Stiffer regulation will surely follow if a wrongful
discharge case of landmark stature arises in connection with use of on-site
drugs-of-abuse testing devices.
References
1. JP Hoffman, C Lairson, and A Sanderson, "An Analysis of Worker Drug Use and Workplace Policies and Programs" [on-line] (Rockville, MD: U.S. Department of Health and Human Services, Office of Applied Studies, Substance Abuse, and Mental Health Services Administration, 1997) [cited 9 June 2002]; available from Internet: http://www.samhsa.gov/wkplace/workplac.htm.
2. AHB Wu, "Point-of-Care Testing for Drugs of Abuse," in The Clinical Toxicology Laboratory: Contemporary Practice of Poisoning Evaluation, ed. LM Shaw and TC Kwong (Washington, DC: AACC Press, 2001), 145–156.
3. DS Chattergoon et al., "Carbamazepine Interference with an Immune Assay for Tricyclic Antidepressants in Plasma," Journal of Toxicology–Clinical Toxicology 36 (1990): 109–113.
4. GJ Turner, DL Colbert, and BZ Chowdry, "A Broad Spectrum Immunoassay Using Fluorescence Polarization for the Detection of Amphetamines in Urine," Annals of Clinical Biochemistry 28 (1991): 588–594.
5. PD Camara et al., "False-Positive Immunoassay Results for Urine Benzodiazepine in Patients Receiving Oxaprozin (Daypro)," Clinical Chemistry 41 (1995): 115–116.
6. AHB Wu, "Mechanism of Interferences for Gas Chromatography/Mass Spectrometry Analysis of Urine for Drugs of Abuse," Annals of Clinical Laboratory Science 25 (1995): 319–329.
7. C Meadway, S George, and R Braithwaite, "Opiate Concentrations Following the Ingestion of Poppy Seed Products—Evidence for ‘the Poppy Seed Defence,'" Forensic Science International 96 (1988): 29–38.
8. TZ Bosy and KA Cole, "Consumption and Quantitation of ³9-Tetrahydrocannabinol in Commercially Available Hemp Seed Oil Products," Journal of Analytical Toxicology 24 (2000): 562–566.
9. AT Kieman et al., "Effect on Sports Drug Tests of Ingesting Meat from Steroid (Methenolone)-Treated Livestock," Clinical Chemistry 40 (1994): 2084–2087.
10. AHB Wu, "Integrity of Urine Specimens Submitted for Toxicologic Analysis: Adulteration, Mechanisms of Action, and Laboratory Detection," Forensic Science Review 10 (1998): 47–65.
11. G Tsongalis, DE Anamani, and AHB Wu, "DNA Fingerprinting for Identification of Urine Specimen Donors by Polymerase Chain Reaction Amplification Typing of the HLA DQA Locus," Journal of Forensic Science 41 (1996): 1031–1034.
12. G Cassella et al., "The Analysis of Thebaine in Urine for Detection of Poppy Seed Consumption," Journal of Analytical Toxicology 21 (1997): 376–383.
Alan H.B. Wu, PhD, is director of clinical chemistry in the department of pathology and laboratory medicine at Hartford Hospital (Hartford, CT). He can be reached via awu@harthosp.org.
Copyright ©2002 IVD Technology
