6: Laboratory Principles

Toxicology addresses harm caused by acute and chronic exposures to excessive amounts of a xenobiotic. Detecting the presence or measuring the concentration of toxic xenobiotics is the primary activity of the analytical toxicology laboratory. Such testing is closely intertwined with therapeutic drug monitoring, in which drug concentrations are measured as an aid to optimizing drug dosing regimens. The toxicology laboratory is frequently viewed in much the same way as other clinical laboratories often are—as a “black box” that converts orders into test results. Because toxicology testing volumes are relatively low and menus are extensive, testing is not as highly automated as in other clinical laboratories. Many results may be “handmade the old-fashioned way.” A downside of this may be somewhat longer turnaround times. But the upside is that toxicology laboratory personnel have the incentive and flexibility to develop substantial expertise. Clinicians who understand how toxicology testing is done will be able to order more judiciously and apply the results more effectively.

Despite a common focus, there is remarkable variability in the range of tests offered by analytical toxicology laboratories. Test menus may range from once-daily testing for routinely monitored drugs and common drugs of abuse to around-the-clock availability of a broad array of assays with the theoretical potential to identify several thousand compounds. Consensus statements have recommended tests that should be available to support management of poisoned patients presenting to emergency departments.^21,32 These guidelines make specific recommendations but recognize that no set of recommendations will be universally appropriate and that it is impossible for a clinical laboratory to offer a full spectrum of toxicology testing in real time.

Decisions on the menu of tests to be offered by any specific laboratory should be decided by the laboratory director in consultation with the medical toxicologists and other clinicians who will use the service and should take into account regional patterns of use of licit and illicit drugs and exposure to environmental toxins, as well as resources available and competing priorities.

The recommendations in Table 6–1 were developed by the National Academy of Clinical Biochemists (NACB) from a consensus process that involved clinical biochemists, medical toxicologists, forensic toxicologists, and emergency physicians.³² Although these tests should be readily available in the clinical laboratory, they should not be considered as a test panel for possibly poisoned patients. As with all laboratory tests, they should be selectively ordered based on the patient’s history, clinical presentation, or other relevant factors. Suggested turnaround time for reporting serum concentrations of the drugs listed in Table 6–1 was one hour or less. Quantitative tests for serum methanol and ethylene glycol were also recommended, with the reservations that these tests are not needed in all settings and that a realistic turnaround time is 2 to 4 hours. Serum cholinesterase testing with a turnaround time of less than 4 hours was proposed by some participants but did not achieve a general consensus. In the United Kingdom, the National Poisons Information Service and the Association of Clinical Biochemists have recommended a nearly identical list of tests, omitting the anticonvulsants.²¹

Although the consensus for the menu of serum assays was generally excellent, there was less agreement as to the need for qualitative urine assays. This was largely a result of issues of poor sensitivity and specificity, poor correlation with clinical effects, and infrequent alteration of patient management. Although these were potential issues for all of the urine drug tests, they led to explicit omission of tests for tetrahydrocannabinol (THC) and benzodiazepines from the recommended list despite their widespread use. THC results were thought to have little value in managing patients with acute problems, and tests for benzodiazepines were believed to have an inadequate spectrum of detection. Testing for amphetamines, propoxyphene, and phencyclidine (PCP) were only recommended in areas where use was prevalent. It was also suggested that diagnosis of tricyclic antidepressant (TCA) toxicity not be based solely on the results of a urine screening immunoassay because a number of other drugs may cross-react. The significance of TCA results should always be correlated with electrocardiographic and clinical findings. The only urine test included in the United Kingdom guidelines was a spot test for paraquat.²¹ Paraquat testing was omitted in the NACB guidelines because of a very low incidence of paraquat exposure in North America.³²

The NACB guidelines also recommend the availability of broad-spectrum toxicology testing in addition to the tests in Table 6–1 to be used for selected patients with presentations compatible with poisoning but who remain undiagnosed and who are not improving. In general, such testing should not be ordered until the patient is stabilized and input has been obtained from a medical toxicologist or poison center. This second level of testing may be provided directly by the local laboratory or by referral to a reference laboratory or a regional toxicology center.

Many physicians order a broad-spectrum toxicology screen for a poisoned patient if one is readily available, but only approximately 2% of clinical laboratories provide relatively comprehensive toxicology services (as estimated from proficiency testing data³). Although broad-spectrum toxicology screens can identify most drugs present in overdosed patients, the results of broad-spectrum screens infrequently have altered management or outcomes.^{13,14,20,22,29}

The extent to which the NACB recommendations are being followed may be estimated from the numbers of laboratories participating in various types of proficiency testing. Result summaries from the 2011 series of proficiency surveys administered by the College of American Pathologists suggest that among laboratories that offer routine clinical testing, 50% to 60% offer quantitative assays for APAP, carbamazepine, ethanol, lithium, phenobarbital, salicylates, theophylline, and valproic acid; 60% to 70% offer digoxin, iron, and transferrin or iron-binding capacity; and 70% to 80% offer carboxyhemoglobin and methemoglobin. About half of these laboratories offer screening tests for drugs of abuse in urine.³

About 2% of laboratories participated in proficiency testing for a full range of toxicology services. These full-service laboratories typically offer quantitative assays for additional therapeutic drugs, particularly TCAs, as well as assays that are designated as broad-spectrum or comprehensive toxicology screens. About 80% of these full-service toxicology laboratories offer testing for volatile alcohols other than ethanol, and half offer testing for ethylene glycol.⁴

Although relatively few laboratories offer a wide range of in-house testing, most laboratories send out specimens to reference laboratories that offer large toxicology menus. The turnaround time for such “send-out” tests ranges from a few hours to several days, depending on the proximity of the reference laboratory and the type of test requested.

Even in full-service toxicology laboratories, the test menu may vary substantially from institution to institution. Larger laboratories typically offer one or more broad-spectrum testing choices, often referred to as “tox screens.” There is as much variety in the range of xenobiotics detected by various toxicologic screens as there is in the total menu of toxicologic tests. Routinely available tests are usually listed in a printed or online laboratory manual. Laboratories with comprehensive services may be able to offer ad hoc chromatographic assays for additional xenobiotics that are not listed. Testing that is sent to a reference laboratory is often not listed in the laboratory manual. The best way to determine if a particular xenobiotic can be detected or quantitated is to ask the director or supervisor of the toxicology or clinical chemistry section because laboratory clerical staff may only be aware of tests listed in the manual.

There are many reasons for toxicologic testing. The most common function is to confirm or exclude suspected toxic exposures. A laboratory result provides a level of confidence not readily obtained otherwise and may avert other unproductive diagnostic investigations driven by the desire for completeness and medical certainty. Testing increased diagnostic certainty in more than half of cases,^2,13,14 and in some instances, a diagnosis may be based primarily on the results of testing. This can be particularly important in poisonings with xenobiotics having delayed onset of clinical toxicity, such as APAP, or in patients with ingestion of multiple xenobiotics. In these instances, characteristic clinical findings may not have developed at the time of presentation or may be obscured or altered by the effects of coingestants.

Testing can provide two key parameters that will have a major impact on the clinical course, namely, the xenobiotic involved and the intensity of the exposure. This information can assist in triage decisions and can facilitate management decisions, such as use of specific antidotes or interventions to hasten elimination. Well-defined exposure information can also facilitate provision of optimum advice by poison centers. Finally, positive findings for ethanol or drugs of abuse in trauma patients may serve as an indication for substance use intervention as well as a risk marker for the likelihood of future trauma.¹³

The confirmation of a clinical diagnosis of poisoning provides an important feedback function, whereby the physician may evaluate the diagnosis against a “gold standard.” Another important benefit is reassurance that an unintentional ingestion did not result in absorption of a toxic amount of xenobiotic. This reassurance may allow a physician to avoid spending excessive time with patients who are relatively stable. It may also allow admissions to be made and interventions undertaken more confidently and efficiently than would be likely based solely on a clinical diagnosis. Testing may also be indicated for medicolegal reasons to establish a diagnosis “beyond a reasonable doubt.”

The key to optimal use of the toxicology laboratory is communication. This begins with learning the laboratory’s capabilities, including the xenobiotics on its menus, which can be quantitated and which merely detected, and the anticipated turnaround times. For screening assays, one should know which xenobiotics are routinely detected, which ones can be detected if specifically requested, and which ones cannot be detected even when present at concentrations that typically result in toxicity.

One should know specimen type that is appropriate for the test requested. A general rule is that quantitative tests require serum (red stopper) or heparinized plasma (green stopper) but not ethylenediamine tetraacetic acid (EDTA) plasma (lavender stopper) or citrate plasma (light-blue stopper). EDTA and citrate bind divalent cations that may serve as cofactors for enzymes used as reagents or labels in various assays. Additionally, liquid EDTA and citrate anticoagulants dilute the specimen. Serum or plasma separator tubes (identifiable by the separator gel at the bottom of the tube) are also acceptable, provided that prolonged gel contact before testing is avoided. Some hydrophobic drugs may diffuse slowly into the gel, leading to falsely low results after several hours. A random, clean urine specimen is generally preferred for toxicology screens because the higher drug concentrations usually found in urine can compensate for the lower sensitivity of the broadly focused screening techniques. A urine specimen of 20 mL is usually optimal. Requirements for all specimens may vary from laboratory to laboratory.

When requesting a screening test, an important—and often overlooked—item of communication is specifying any xenobiotics of ­particular concern. This may allow faster results and a greater likelihood of detection.

Most full-service toxicology laboratories welcome consultation on puzzling cases or results that appear inconsistent with the clinical presentation. The laboratory will be familiar with the capabilities and limitations of their testing methods, as well as common sources of discrepant results.

Most tests in the toxicology laboratory are directed toward the identification or quantitation of xenobiotics. The primary techniques used include spot tests, spectrochemical tests, immunoassays, and chromatographic techniques. Mass spectrometry may also be used, usually in conjunction with gas chromatography (GS) or liquid chromatography. Table 6–2 compares the basic features of these methodologies. Other methodologies include ion-selective electrode measurements of lithium, atomic absorption spectroscopy or inductively coupled plasma mass spectroscopy for lithium and heavy metals, and anodic stripping methods for heavy metals. Many adjunctive tests, including glucose, creatinine, electrolytes, osmolality, metabolic products, and enzyme activities, may also be useful in the management of poisoned patients. The focus here is on the major methods used for directly measuring xenobiotics.

Spot Tests

The simplest tests are spot tests. These rely on the rapid reaction of a xenobiotic with a chemical reagent to produce a colored product (eg, the formation of a colored complex between salicylate and ferric ions) that is visually assessed in a semiquantitative manner. Because the reagents may cause precipitation of serum proteins, spot tests are more commonly performed on urine specimens or gastric aspirates. Such tests were once a mainstay of toxicologic testing. Because of the poor selectivity of chemical reagents, as well as substantial variability in visual interpretation, these assays suffer from fairly frequent false-positive results and occasional false-negative results and are rarely used today.

Spectrochemical Tests

Spectrochemical tests rely on measurement of a light-absorbing substance. Some analytes that are intrinsically light absorbing may be directly measured. Cooximetry (also known as hemoximetry) represents a sophisticated application of spectrophotometry to the measurement of various forms of hemoglobin in a hemolyzed blood sample. Measurement of light absorbance at multiple wavelengths allows several hemoglobin species to be simultaneously quantitated. For mathematical reasons, the number of wavelengths used must be greater than the number of different types of hemoglobin present. This is why classic pulse oximetry, which uses only two wavelengths, yields spurious results in the presence of significant amounts of methemoglobin or carboxyhemoglobin (Chaps. 29, 125, and 127). Cooximetry is relatively free of interferences because the concentrations of the hemoglobins are so much higher than other substances in the blood. However, the presence of intensely colored substances (eg, methylene blue) may cause spurious increases or decreases in the apparent percentages of the hemoglobins. Modern instruments are often able to recognize a significantly atypical pattern of absorbance and generate an error message in addition to or instead of a result.

Most analytes are neither as deeply colored nor as highly concentrated as hemoglobin species. Their detection requires a chemical reaction to produce an intensely light-absorbing product that is quantitatively measured at a specific wavelength in a spectrophotometer. Because spectrophotometers can also measure ultraviolet and infrared light, it is not necessary for the product to have a visible color. Early spectrochemical assays typically measured the absorbance after conversion of all of the analyte to the light-absorbing product. Modern assays usually use rate spectrophotometry, taking multiple absorbance measurements over time to determine the rate of change in light absorbance as the reaction proceeds. During the initial phase of the reaction, this rate is constant and proportional to the initial concentration of the analyte. This significantly reduces the time needed to obtain a result because it is not necessary for the reaction to go to completion, and it allows the averaging of multiple measurements, improving precision. Furthermore, it is unaffected by nonreacting substances that absorb light at the test wavelength because the absorbance of the nonreacting substances is constant and does not contribute to the rate of change in the absorbance.

Rate spectrophotometry remains subject to interference by substances that react to produce light-absorbing products, thereby falsely increasing the apparent concentration. Substances that inhibit the assay reaction or that consume reagents without producing a light-absorbing product give falsely low results. For example, ascorbic acid produces negative interference in many spectrophotometric assays that use oxidation reactions to generate colored products.

One way to improve the selectivity of a spectrochemical assay is to increase the selectivity of the reaction that generates the light-absorbing product. Enzymes, which can catalyze highly selective reactions, are often used for this purpose. For example, many assays for ethanol use alcohol dehydrogenase (ADH) to catalyze the oxidation of ethanol to acetaldehyde, with concomitant reduction of the cofactor NAD⁺ (oxidized form of nicotinamide adenine dinucleotide) to NADH (reduced form of nicotinamide adenine dinucleotide). The initial rate of increase in light absorption produced by the conversion of NAD⁺ to NADH is proportional to the concentration of ethanol. Although other alcohols, such as isopropanol and methanol, can also be oxidized by ADH, they are much poorer substrates for ADH with low rates of reaction and correspondingly low levels of interference.

Many other enzymatic assays also rely on measuring the change in light absorption at 340 nm when NAD⁺ is converted to NADH or vice versa. These include enzymatic assays for ethylene glycol, as well as some enzyme-linked immunoassays, such as EMIT (enzyme-multiplied immunoassay technique) assays. All such assays are potentially subject to interference by specimens with high concentrations of lactate. Lactate dehydrogenase, which is naturally present in serum, will oxidize this lactate to pyruvate if NAD⁺ becomes available for simultaneous reduction to NADH. When a serum specimen with high lactate is mixed with assay reagents that contain NAD⁺, oxidation of the lactate contributes to the total rate of NADH production. The increased rate of NADH production results in a false increase in the measured concentration of the target analyte.

Some enzymatic reactions do not produce a colored product. Enzymes such as glucose oxidase or lactate oxidase couple oxidation of the substrate to reduction of oxygen to hydrogen peroxide, which is colorless. A coupled second reaction is then necessary using the peroxide to convert a colorless dye to a colored one. Oxidase-based reactions may be subject to interference by compounds with high structural similarity to the target analyte. For example, glycolate, a toxic metabolite of ethylene glycol, is an excellent substrate for lactate oxidase and will give falsely high lactate results when it is present.

Immunoassays

The need to measure very low concentrations of an analyte with a high degree of specificity led to the development of immunoassays. The combination of high affinity and high selectivity makes antibodies excellent assay reagents. There are two common types of immunoassays: noncompetitive and competitive. In noncompetitive immunoassays, the analyte is sandwiched between two antibodies, each of which recognizes a different epitope on the analyte. In competitive immunoassays, analyte from the patient’s specimen competes for a limited number of antibody binding sites with a labeled version of the analyte provided in the reaction mixture. Because most drugs are too small to have two distinct antibody binding sites, drug immunoassays are usually competitive.

In competitive immunoassays, increasing the concentration of xenobiotic in the specimen results in increased displacement of labeled xenobiotic from the antibodies. The amount of xenobiotic in the specimen can be determined by measuring either the amount of label remaining bound to the assay antibodies or the amount of label free in solution. In the earliest immunoassays, the label was a radioisotope, typically iodine-125, tritium, or carbon-14. Today, radioimmunoassays are relatively uncommon because of problems associated with handling and disposal of radioactivity. Nonisotopic immunoassays are currently the most widely used methodologies for the measurement of drugs. They offer high selectivity and good precision and are readily adapted to automated analyzers, thereby decreasing both the cost and the turnaround time of the assays. The xenobiotics for which immunoassays are available are limited to those for which there is a high demand, such as widely monitored therapeutic drugs and the drugs of abuse included in workplace drug screening. However, because production costs are relatively low, these tests are widely distributed at reasonable prices.

The most widely used nonisotopic drug immunoassays are in the category of homogenous immunoassays. Homogenous immunoassays measure differences in the properties of bound and free labels rather than directly measuring one or the other after their physical separation. Avoiding a separation step allows homogenous immunoassays to be readily adapted to automated analysis. Homogenous techniques that are in wide use include EMIT (Fig. 6–1), kinetic inhibition of microparticles in solution (KIMS), and cloned enzyme donor immunoassay (CEDIA).

Many of the newest automated immunoassays are again using physical separation techniques. In these assays, the detection antibody is physically attached to a solid support, and separation occurs by a simple wash step. This wash step removes the patient’s serum along with many potentially interfering substances. Older assays of this type used antibodies bound to large plastic beads or wells of microtiter plates and required long incubation steps because of substantial times required for diffusion of the reactants to the antibodies. Newer assays typically use latex microparticles that have very high total surface areas, allowing rapid equilibration and short assay times.

Figure 6–2 shows a schematic magnetic microparticle enzyme-labeled chemiluminescent competitive immunoassay. A single enzyme label can generate many photons, allowing high signal amplification. Coupled with a background luminescence that is essentially zero, such assays can measure concentrations below the nanomolar level. Many variations of this approach are in use. Enzyme substrates may be used that result in fluorescent or colored products. Enzymes other than alkaline phosphatase may be used as labels, or nonenzymatic fluorescent, chemiluminescent, or electroluminescent labels may be used. These new techniques are readily automated and have higher sensitivities than homogenous immunoassays.

Microparticle capture assays are a type of qualitative competitive immunoassay that have become very popular, especially for urine drug screening tests. The use of either colored latex or colloidal gold microparticles enables the result to be read visually as the presence or absence of a colored band, with no special instrumentation required. Competitive binding occurs as the assay mixture is drawn by capillary action through a porous membrane. This design feature is responsible for alternate names for the technique: lateral flow immunoassay or immunochromatography.

The simplest microparticle capture design uses an antidrug antibody bound to colored microparticles and a capture zone consisting of immobilized drug (Fig. 6–3). If the specimen is xenobiotic free, the beads will bind to the immobilized analyte, forming a colored band. When the amount of drug in the patient specimen exceeds the detection limit, all of the antibody sites will be occupied by drug from the specimen, and no labeled antibody will be retained in the capture zone. The use of multiple antibodies and discrete capture zones with different immobilized analytes can allow several xenobiotics to be detected with a single device.

A disadvantage of this design is the potential for causing confusion because a positive test result is indicated by the absence of a band. More complex (and more expensive) variations have been developed in which a colored band denotes a positive test result.

Although immunoassays have a high degree of sensitivity and selectivity, they are also subject to interferences and problems with cross-reactivity. Cross-reactivity refers to the ability of the assay antibody to bind to xenobiotics other than the target analyte. Xenobiotics with similar chemical structures may be efficiently bound, which can lead to falsely elevated results. In some situations, cross-reactivity can be beneficially exploited. For example, some immunoassays effectively detect classes of drugs rather than one specific drug. Immunoassays for opioids use antibodies to morphine that cross-react to varying degrees with structurally related substances, including codeine, hydrocodone, and hydromorphone. Oxycodone typically has low cross-reactivity, and higher concentrations are required to give a positive result. The cross-reactivity of non-morphine opiates varies with manufacturer. Consult with your lab for the relative sensitivities of the immunoassay it uses. Structurally unrelated synthetic opioids, such as meperidine and methadone, have little or no cross-reactivity and are not detected by opiate immunoassays. Immunoassays for the benzodiazepine class react with a wide variety of benzodiazepines but with varying degrees of sensitivity.^12,16 Because of the highly variable response of immunoassays to the various opiates and benzodiazepines, methods based on mass spectrometry should be used for definitive results.

Class specificity can be a two-edged sword. Assays for the TCA family have similar reactivity with amitriptyline, nortriptyline, imipramine, and desipramine and can be used to provide an estimate of the total concentration of any combination of these drugs. To account for nonuniform cross-reactivity, such results of these assays are usually reported as concentration ranges (eg, <100 ng/mL, 100–300 ng/mL). A large number of other drugs with tricyclic structures, including carbamazepine, many phenothiazines, and diphenhydramine, also cross-react and generate a signal, particularly at concentrations found in patients who overdose. Qualitative tests, such as microparticle capture assays, may then yield false-positive results if the signal generated by the cross-reacting drug (eg, carbamazepine) exceeds the detection limit of the immunoassay. With quantitative or semiquantitative assays, however, the apparent concentration produced by a cross-reacting drug is generally well below TCA concentrations associated with toxicity.

Even when an antibody is selected to be specific to a single drug, it is common that metabolites of the target drug show some cross-reactivity. This, too, may be beneficial. When the metabolite is an active one (eg, carbamazepine epoxide), the contribution of its cross-reactivity may yield results that correlate better with the drug effect than the true concentration of the parent drug alone.

Immunoassays are also subject to interference by substances that impair detection of the label. Elevated lactate concentrations may lead to spuriously increased drug concentrations in specimens tested by EMIT, as described earlier. Immunoassays that rely on enzyme labels are particularly sensitive to nonspecific interference because enzyme activity is highly dependent on reaction conditions. A number of substances that can inhibit the enzyme reaction in EMIT assays are used to adulterate urine submitted for drug abuse testing with the intent of producing false-negative results (see the discussion of drug-abuse screening tests under “Special Considerations for Drug Abuse Screening Tests” later). Such adulteration may be detected when the rate of reaction is lower than the rate observed with a drug-free control.

Chromatography

Chromatography encompasses several related techniques in which analyte specificity is achieved by physical separation. The unifying mechanism for separation is the partition of the analytes between a stationary phase and a moving phase (mobile phase). In most instances, the stationary phase consists of very fine particles arranged in a thin layer or enclosed within a column. The mobile phase flows through the spaces between the particles. Analytes are in a rapid equilibrium between solution in the mobile phase and adsorption to the surfaces of the particles. They move when in the mobile phase and stop when adsorbed to the stationary phase. The average velocity of the analyte xenobiotics depends on the relative time spent in the moving versus stationary phase. Xenobiotics that partition primarily into the mobile phase have average velocities slightly lower than the mobile phase velocity. Average velocity decreases as the proportion of time adsorbed to the stationary phase increases. Under controlled conditions, these average velocities are highly reproducible. Xenobiotics may be provisionally identified based on their characteristic velocity, as measured by the amount of time required to traverse the length of a chromatography column (retention time). Chromatography is a separation method and must be combined with a detection method to allow identification and measurement of the separated substances.

Chromatographic behavior is sufficiently reproducible that the failure to detect a signal at the retention time characteristic of a compound effectively excludes the presence of that compound in amounts greater than the detection limit. On the other hand, a number of different substances may have migration velocities that are identical or nearly so. A positive finding is therefore not completely specific. Definitive identification depends on having additional information, which may be obtained through selective detection techniques or by confirmatory testing using a second method. The sensitivity of chromatographic methods depends on both the amount of specimen available and the sensitivity of the detection method. A major advantage of chromatographic techniques is that multiple xenobiotics may be detected and measured in a single procedure. Nor is it necessary to know in advance the specific xenobiotic to be looked for. For this reason, chromatographic techniques have a major role in screening for multiple xenobiotics.

Most chromatographic procedures require extraction and concentration of the xenobiotics to be analyzed before the chromatography is done. Extraction results in removal of salts, proteins, and other materials that may exhibit unfavorable interactions with either of the chromatographic phases. Concentration allows the substances to be introduced in a narrow “band,” so that compounds with slightly different relative mobilities become completely resolved, or separated from one another, rather than overlapping. This also results in a more intense signal as a band passes through the detector and increases sensitivity.

Extraction of drugs is most commonly done with organic solvents, but “solid-phase extraction” is also very popular.¹¹ Solid-phase extraction is itself a modified chromatographic procedure in which a urine or serum specimen is passed through a short chromatography column with a hydrophobic stationary phase. Most drugs are sufficiently hydrophobic to partition almost completely into the stationary phase and are retained on the column. Subsequently, the retained xenobiotics are eluted with an organic solvent. The organic solvents from either extraction technique are evaporated to concentrate the extract. The extraction process allows the analyte from a large volume of specimen to be concentrated. Detection sensitivity can thereby be increased provided large-volume specimens can be readily obtained, as is true with urine.

Often a preextraction treatment is used to increase the hydrophobicity of the substances to be extracted. The most common manipulation is pH adjustment, either upward or downward, to convert charged forms of drugs into uncharged, hydrophobic, and therefore extractable ones. In other instances, enzymatic or chemical hydrolysis may be used to convert water-soluble glucuronide metabolites back to their more readily extracted parent compounds, for example, conversion of morphine glucuronide to morphine.

In the technique of high-performance liquid chromatography (HPLC), a stationary phase is packed into a column and the mobile phase is pumped through under high pressure (Fig. 6–4). This allows good flow rates to be achieved even when solid phases with very small particle sizes are used. Smaller particle size increases surface area, decreases diffusion distances, and improves resolution, but the spaces between the particles are also smaller, increasing the resistance to flow. The use of high pressure and small particles allows good separations while keeping assay time short.

HPLC drug assays typically use “reverse-phase” chromatography. Early chromatographic techniques typically used relatively polar silica gel particles as the stationary phase, with organic solvent mobile phases. Reverse-phase chromatography uses stationary phases consisting of silica gel particles that have had hydrocarbon molecules covalently linked to the outer surface. This coats the particles with a permanently bonded oil-like layer. Mobile phases are primarily aqueous with varying amounts of organic solvent. Because of these modifications, hydrophobic xenobiotics are more strongly adsorbed by the stationary phase, but hydrophilic ones tend to remain in the mobile phase. This results in an order of elution from the column that is approximately the reverse of that seen with unmodified silica gel and organic solvents. Thus, the term reverse-phase chromatography is used. HPLC can be done using either “normal-phase” or “reverse-phase” conditions, but reverse-phase conditions are much more commonly used. A variety of hydrocarbons can be used to derivatize the silica gel. By far, the most common reverse-phase columns use an octadecyl hydrocarbon as the outer coating and are often referred to as C-18 columns.

In HPLC, the xenobiotics are detected after they exit the chromatographic column. In this case, they are identified by their retention time (the characteristic time required to traverse the column). Because most xenobiotics absorb ultraviolet light, detection is commonly by ultraviolet spectroscopy using specially designed flow-through cuvettes. Measuring light absorbance at a selected wavelength allows the amount of the xenobiotic to be determined. Accuracy is often enhanced by comparing the absorbance of the target analyte with absorbance of an internal standard (ie, a compound with a different retention time that is added in a fixed amount to all specimens). The ratio of the drug absorbance to the internal standard absorbance is proportional to the drug concentration in the specimen.

Although most HPLC detectors allow a selection of the detection wavelength, only one wavelength is commonly used during a given run. Some detectors, however, allow absorbance at multiple wavelengths to be determined by breaking white light into its component wavelengths after it has passed through the detection cuvette and then making measurements at multiple wavelengths simultaneously using a light-sensitive chip similar to those used in digital cameras. This allows the absorbance spectrum of a compound to be determined as it elutes from the column. This information can supplement the retention time and allow more specific identifications to be made.

HPLC is often the method of choice for measuring serum concentrations of xenobiotics for which no immunoassay is available. However, it is limited by an inability to analyze drugs with a wide range of polarities in a single assay or to fully resolve substances with very similar polarity, both of which limit its usefulness as a broad drug-screening technique.

GC is similar in principle to HPLC except that the moving phase is a gas, usually the inert gas helium but occasionally nitrogen. The schematic illustration of HPLC in Fig. 6–4 is also applicable to GC. The low flow resistance of gas allows high flow rates that make possible substantially longer columns than are used in HPLC. This offers the dual advantages of high resolution and fast analysis. As was true in HPLC, most GC assays incorporate an internal standard to increase precision.

Because the inert carrier gas does not engage in intermolecular interactions, partition of the analytes into the moving gas phase depends primarily on their natural volatility. Elevated column temperatures are required to achieve sufficient volatility for analysis of most xenobiotics. The use of a temperature gradient (the column temperature is programmed to increase throughout the course of the analysis) can allow xenobiotics with a wide range of volatility to be analyzed in a single run. This feature, coupled with excellent resolution, makes GC suitable for screening assays that encompass a broad range of drugs.

GC is limited to xenobiotics that are reasonably volatile at temperatures below 572°F (300°C), above which the stationary phase may begin to break down. Two principal attributes of a xenobiotic limit its volatility: its size and its ability to form hydrogen bonds. Xenobiotics that form hydrogen bonds via amino, hydroxyl, and carboxylate moieties can be made more volatile by replacing hydrogens on oxygen and nitrogen atoms with a nonbonding, preferably large, substituent. (Large substituents sterically hinder access to the acceptor electron pairs on the nitrogen and oxygen atoms.) A number of derivatizing agents can be used to add appropriate substituents. The most common derivatives involve the trimethylsilyl (TMS) group. Although derivatization with TMS substantially increases the molecular weight, the resulting derivative is much more volatile as a consequence of the loss of hydrogen bonding.

In traditional packed-column GC, the packing may consist of inert support particles with a thin coating of nonvolatile, high-molecular-weight oil that comprises the stationary phase. It is increasingly common for the stationary phase to be covalently bonded to the support particles. A highly useful variant of GC is capillary chromatography. A long, thin capillary tube of fused silica is coated on the inside with a covalently bonded stationary phase. The mobile gas phase flows through the tiny channel in the middle. These capillaries are flexible, allowing very long columns (≥10 m) to be coiled into a small space. The long column length, coupled with highly uniform conditions throughout the column, results in extremely high resolution. The small diameter of the column allows rapid thermal equilibration and the use of steep temperature gradients that can speed analysis. The major drawback to capillary chromatography is a very limited column capacity. Special techniques are needed to restrict the amount of material introduced into the column and thereby to avoid overloading it. High-sensitivity detectors are required to measure the small quantities that can be chromatographed.

A number of detectors are available for GC. The most common detector, particularly for packed columns, is the flame ionization detector. This involves directing the outflow of the column into a hydrogen flame. Organic molecules emerging from the column are burned, creating charged combustion intermediates that can be measured as a current. The amount of current flow is largely determined by the mass of carbon that is being burned. Nitrogen–phosphorus detectors are also widely used in drug analysis. In this modification of a flame ionization detector, a heated bead coated with an alkali metal salt is used to selectively generate ions from xenobiotics containing nitrogen or phosphorus. These devices detect broad ranges of substances but do not identify them. The identity of the compounds detected must be inferred from the retention time.

The mass spectrometer can serve as a highly sensitive GC detector and possesses the ability to generate highly characteristic mass spectra from the compounds it is detecting. A special requirement of the mass spectrometer is that it requires a high vacuum to prevent the ionic particles that it creates from interacting with other molecules or ions. This requires removal of the inert carrier gas and is easiest when there is a low total gas flow, such as occurs with capillary GC. The mass spectrometer in turn provides good sensitivity for the small amounts of analyte that can be accommodated in capillary GC.

The detection process begins by generating ions from the analyte. This is usually done using electron impact ionization. The gas phase analyte is separated from the bulk of the carrier gas and introduced into an ionization chamber, where it is bombarded by a stream of electrons. Electron impact can dislodge an electron from the analyte, creating a positively charged ion and frequently imparting sufficient energy to the ion to break it into pieces. If fragmentation occurs, conservation of charge requires that one of the resulting fragments be a positively charged ion. The fragments into which a molecular ion can break are characteristic of the xenobiotic as are the relative probabilities that any given fragment will carry the positive charge.

The mass spectrometer then uses electromagnetic filtering to direct only ions of a specified mass-to-charge (m/z) ratio to a detector. Because most of the ions produced have a single positive charge, the observed peaks generally correspond to the mass of the ions. The detector has sufficient electronic amplification that a single ion could theoretically be detected, accounting for the high sensitivity of mass spectrometric detection. By rapidly scanning through a range of masses that are sequentially allowed to reach the detector, a mass spectrum may be generated. The mass spectrum records the masses of the pieces produced by fragmentation of the parent ion, as well as the relative frequency with which these fragments are produced and detected. The highest mass observed in the spectrum usually corresponds to the mass of intact parent ions generated from collisions that were not energetic enough to cause fragmentation.

Figure 6–5 shows the mass spectrum obtained from a gas chromatograph at a time when the TMS derivative of the cocaine metabolite benzoylecgonine was emerging from the capillary column. The mass spectrum of any compound is highly distinctive and usually unique. The primary exception involves optical enantiomers, both of which have the same mass spectrum. Toxicologically significant examples of enantiomers include d-methamphetamine, a drug of abuse, and l-methamphetamine, which is found in decongestant inhalers. It is also important to distinguish dextrorphan, the major metabolite of the cough suppressant dextromethorphan, and levorphan (levorphanol), a controlled substance.

To avoid the need to scan the full range of masses in a typical mass spectrum, selected ion monitoring is often used. Here, the mass spectrometer is typically programmed to filter and detect only three of the larger and more characteristic peaks in the mass spectrum. In the case of TMS-benzoylecgonine (TMS-BE), the peaks at m/z 240, 256, and 361 are used. The concentration of TMS-BE in the specimen is determined from the ratio of the peak height at m/z 240 to height of a peak at m/z 243 that results from a corresponding fragment of the internal standard, which is TMS-BE that is labeled with three deuterium atoms (d₃-TMS-BE) to give it a mass distinct from that of TMS-BE from the specimen (Fig. 6–5). The specificity of the identification is verified by finding peaks at m/z 256 and m/z 361, with peak height ratios to the peak at m/z 240 comparable to the ratios seen with authentic TMS-BE. The detection at the correct retention time of a xenobiotic producing all three peaks in the correct ratios produces an extremely specific identification.

The high sensitivity and specificity afforded by GS/mass spectrometry is being further extended by the related hybrid technique of liquid chromatography/tandem mass spectrometry, often abbreviated as LC/MS/MS.¹⁸ This technique is being used in an increasing number of toxicology laboratories. In LC/MS/MS, a tandem mass spectrometer is used as the detector for liquid chromatography system. The initial ionization is done under conditions that do not promote fragmentation and is commonly achieved by adding or removing a proton rather than forcefully dislodging an electron. The resulting ions have a mass that differs from that of the parent molecule by one mass unit ([M+H]⁺ or [M–H]⁻). The first mass spectrometer is used to selectively filter only unfragmented ions with the expected molecular mass. As the selected ions exit the first mass spectrometer at high speed, they are allowed to collide with molecules of an inert gas. These collisions cause the ions to break apart to create the fragment mass spectrum that is detected by the second mass spectrometer. The additional selection step provided by the first mass spectrometer greatly enhances specificity and reduces background signal, enhancing sensitivity.

When properly used to guide dosing adjustments, drug concentration measurements improve medical outcomes.⁹ However, many therapeutic drug measurements are drawn at inappropriate times or are made without an appropriate therapeutic question in mind. An essential requirement for interpretation of xenobiotic concentrations is that the relationship between concentrations and effects must be known. Such knowledge is available for routinely monitored xenobiotics and is often encapsulated in published ranges of therapeutic concentrations and toxic concentrations. Concentrations designated as “toxic” are usually higher than the upper end of the therapeutic range and typically represent concentrations at which toxicity is acute and potentially serious (see back cover).

For most xenobiotics, the relationships between toxic concentrations and effects cannot be systematically studied in humans and consequently are often incompletely defined. These relationships are largely inferred from data provided in overdose case reports and case series. The measurement of xenobiotic concentrations in overdose cases in which concentration–effect relationships are not well defined may contribute more to the management of future overdosed patients than to the management of the patient in whom the measurements were made.

For toxicologists, xenobiotic concentrations are especially useful in two ways. For xenobiotics whose toxicity is delayed or is clinically inapparent during the early phases of an overdose, concentrations may have substantial prognostic value and facilitate anticipatory management. These concentrations may also be used to make decisions regarding the use of antidotes or of interventions to hasten drug elimination, such as hemodialysis.

Quantitative xenobiotic measurements are subject to various interferences, but these are less problematic than in qualitative assays. Signals generated by cross-reacting substances are weaker than those from the target analyte and are relatively unlikely to lead to a false diagnosis of toxicity, particularly if the target analyte is absent. Such cross-reactivity can be exploited in some instances to provide confirmatory evidence of a poison for which no specific assay is immediately available. For example, the immunoassay finding of apparent subtoxic concentrations of TCAs can help confirm a diphenhydramine overdose, and the finding of a measurable digoxin concentration in an unexposed patient may suggest poisoning with other cardioactive steroids of plant or animal origin. Negative interferences are much less frequent. Interferences in chromatographic methods usually result from the presence of other compounds with migration rates similar to the target ­analyte. Because the migration rates are rarely exactly the same, the laboratory can often recognize the presence of the interference as an overlapping peak when both compounds are present. In such instances, the interference may impair accurate measurement of the drug concentration. When no target xenobiotic is present, misidentification of the interfering peak as the target becomes much more likely because a single peak is seen at approximately the expected position. Because interferences in chromatographic methods are generally unique to a specific method, information about these interferences should be obtained by asking the laboratory.

Xenobiotic measurements are unlike most other laboratory measurements in that the concentrations are highly dependent on the timing of the measurement. Knowledge of the pharmacokinetics of a xenobiotic can substantially enhance the ability to draw meaningful conclusions from a measured concentration. Some xenobiotics alter their pharmacokinetic behavior at very high concentrations. These changes in pharmacokinetics may be predictable from the mechanisms of drug clearance and the extent of binding to plasma proteins and to tissues (Chap. 9).

Knowledge of the relationship between xenobiotic concentrations and effects, or pharmacodynamics, is also important. Effects depend on local concentrations at the site of action, typically at cell membrane receptors or intracellular locations. Serum or plasma concentrations can be correlated with effects only when these concentrations are in equilibrium with concentrations at the site of action. During the absorption and distribution phases, the concentration ratio will be higher than its equilibrium value, yet often the only xenobiotic concentration measured after an acute overdose is one obtained while absorption and distribution are still ongoing. This effect may explain some observations of apparent poor correlation between measured concentrations and toxic effects.

For xenobiotics that bind significantly to plasma proteins, it is the concentration of xenobiotic that is not bound to proteins (the free-xenobiotic concentration) that is in equilibrium with concentrations at the site of action. For most drugs at therapeutic concentrations, the free-drug concentration is an approximately constant percentage of the total drug concentration. The total concentration is what is usually measured in the laboratory. Under these conditions, the ratio of total concentration to active site concentration is approximately constant, and a reasonable correlation between total concentration and effects can be expected.

A major change in the free fraction occurs after treatment of digoxin toxicity with digoxin specific antibody fragments when the free digoxin concentration falls from approximately 75% of the total concentration to less than 1% as a consequence of digoxin binding by the antidigoxin antibody fragments. At the same time, there is extensive redistribution of digoxin from tissues to plasma, leading to substantial increases in total digoxin concentration. This situation may be further complicated by complex digoxin specific antibody fragments interference in many digoxin immunoassays.

Measurement of free-drug concentrations can clarify such situations.¹⁵ Assays for free phenytoin are available in many laboratories. Assays for other free-drug concentrations may require special arrangements. The availability and expected turnaround time can be provided by the laboratory. For example, for patients treated with digoxin specific antibody fragments, newer immunoassays that use antibodies attached to microparticles or glass fibers give results that can be used to set an upper bound on free-digoxin concentrations and thereby verify adequacy of treatment.²³

Toxicology Screening

A test unique to the toxicology laboratory is the toxicology screen, or “tox screen.” Depending on the laboratory, this term may refer to a single testing methodology with the ability to detect multiple xenobiotics, such as GC/MS; it may refer to a panel of individual tests, such as a drug-abuse screen; or it may be a combination of broad-spectrum and individual tests. The widespread use of the term “tox screen” is unfortunate because this inappropriately implies for many physicians the availability of a test that can confirm or exclude poisoning as a diagnosis. Although there are many more toxic xenobiotics in the world than there are named diseases, most serious poisonings involve a relatively limited number of xenobiotics. Comprehensive drug screens can identify many of these agents but may miss recently introduced therapeutic agents or new designer drugs, such as components of “bath salts.” Moreover, comprehensive toxicology screens typically do not detect ionic substances, including bromide, cyanide, lithium, iron, lead, and other heavy metals, and may miss substances that are toxic at extremely low concentrations, such as digoxin and synthetic cannabinoids. Designer drugs are a problem because for any drug, authentic material must be available to establish its behavior in the assay system. By the time a particular drug is added to the system, it has often been largely supplanted in users by something new. It should be apparent that a negative toxicology screen result cannot exclude poisoning. It is equally true that a positive finding does not necessarily confirm a diagnosis of poisoning. For assays that detect only the presence of a xenobiotic, it is not possible to distinguish benign or therapeutic concentrations from toxic ones. Quantitative tests may falsely suggest toxicity when drug concentrations are measured during the drug’s distribution phase, which may extend for several hours with drugs such as digoxin and lithium. Moreover, the phenomenon of tolerance may allow chronic drug users to be relatively unaffected by concentrations that would be quite toxic to a non-using individual. Because comprehensive drug screens may differ widely among institutions and patterns of exposure also show substantial regional variation, there is limited ability to draw meaningful conclusions from any study of the sensitivity and specificity of such screens for detecting or excluding poisoning.

The predictive power of the result of a toxicology screen depends on a number of factors, including the likelihood of poisoning before receiving the test results (the prior probability or the prevalence), the range of xenobiotics effectively detected, and the frequency of false-positive results. It should be noted that false-positive and false-negative results may be either analytical or clinical in origin. A clinical false-positive result occurs when a xenobiotic is detected that is not contributing to the medical problem (eg, a therapeutic amount of APAP). A clinical false-negative result may occur when the wrong test is ordered (eg, a screen for drugs of abuse for a patient with APAP poisoning).

Although only approximately 2% of laboratories offer comprehensive toxicology screening,⁴ most laboratories offer some sort of testing in response to a request for a toxicology screen. This may vary from a panel of immunoassays for drugs of abuse to a comprehensive screening test performed at a reference laboratory. Other laboratories may offer a focused panel of tests rather than a comprehensive screening analysis. Larger laboratories may have several types of “tox screens” available for use in different situations. Among laboratories that do not limit their tests exclusively to commercially available methods, it is likely that no two will have exactly the same menu of drugs that can be reliably detected. Given the trends toward increasing automation and decreasing personnel in clinical laboratories, it is relevant to ask what benefits may be derived from such testing. Studies show that comprehensive toxicologic screening has the potential to provide significant information, with utility varying with the indication for testing. The prevalence of positive results has ranged from 34% to 86% of specimens submitted for testing. When drug exposure, as predicted from the patient’s history and physical examination, was compared with screening results, clinically unsuspected substances were found in 7% to 48% of the cases, and clinically suspected xenobiotics were not found in 9% to 25%.^13,14,24 However, limited utility is suggested by studies showing that the results of comprehensive screening affect management in fewer than 5% of cases.^{13,14,20,22,28} A survey of emergency physicians found that more than 75% were not fully aware of the range of drugs detected and not detected by their laboratory’s toxicology screen. The majority believed that the screen was more comprehensive than it actually was.¹⁰

One reason for a limited effect on management is the substantial time delay before the results of comprehensive screening are typically available. Generally, more than 3 hours is required for the report of a negative result, and an even longer time may be required for confirmation of a positive finding. By this time, most consequential management decisions have been implemented. Another possible explanation for the limited utility of screening is that comprehensive screening is largely available only in major medical centers, where consultation from a medical or clinical toxicologist is more likely to be available. Such experts may be more able to make correct diagnoses and initiate appropriate management relying on clinical findings alone.

Several point-of-care devices are capable of rapidly screening urine for the presence of drugs of abuse, as well as TCAs. Results are typically available in 5 to 20 minutes. The accuracy of these devices is typically less than that of laboratory testing.³⁰ In a small study of one such device, diagnosis was believed to have been aided in 82% of cases, and clinical management was changed in 25%.² Additional studies are needed to ascertain the utility of point-of-care drug screening in emergency toxicology.³¹

A useful alternative to the toxicology screen is the toxicology hold. This is a set of serum or urine (or both serum and urine) specimens drawn at the time of presentation, when xenobiotic concentrations are likely to be near maximum concentrations, and initially held refrigerated or frozen without testing. This allows a specimen to remain available for subsequent testing if needed. Most laboratories hold such specimens for several days.

Testing for drugs of abuse is a significant component of medical toxicology testing. These tests are widely used in the evaluation of potential poisonings and are assuming an increasing role is assuring the appropriate use of pain medications.⁷ Initial testing is usually done with a screening immunoassay. Although drug-screening immunoassays were initially developed for use in workplace drug-screening programs and are not always optimal for medical purposes, their wide availability, low cost, and ease of use led to their nearly universal adoption in clinical laboratories. Growth of the market for medical drug screening has led to the development of point-of-care tests specifically for medical use, but these devices largely retain the deficiencies of their predecessors. Drug abuse testing for nonmedical reasons is generally considered to be forensic testing, and confirmation of immunoassay results is considered mandatory in such circumstances. Confirmatory testing can compensate for some immunoassay shortcomings but is frequently not done when screening tests are used for medical purposes. Despite the widespread use of drug-screening immunoassays in medical practice, studies suggest that many physicians do not fully understand the capabilities and limitations of these tests.²⁵

The most commonly tested-for drugs are amphetamines, cannabinoids, cocaine, opioids, and PCP. These are often referred to as the NIDA five because they are the five drugs that were recommended in 1988 by the National Institute on Drug Abuse (NIDA) for drug screening of federal employees. (Responsibility for recommendations for federal drug testing now lies with the Substance Abuse and Mental Health Services Administration {SAMHSA}.) Drug-screening immunoassays are also frequently done for methadone and benzodiazepines and less frequently for oxycodone and hydrocodone. Drug-screening devices intended primarily for medical use may also include tests for APAP or TCAs. Table 6–3 lists some of the general characteristics of these tests. Commercial urine immunoassays are also available for buprenorphine, lysergic acid diethylamide, methaqualone, methylenedioxymethamphetamine (MDMA), and oxycodone. Drug-screening immunoassays are available in a number of formats, which may differ in performance. Almost all of them are designed to be used with urine specimens because these can be obtained noninvasively and generally have higher concentrations than serum, enhancing the sensitivity of the test.

The drug-screening tests for cannabinoids and cocaine are directed toward drug metabolites rather than the parent compound. The parent drugs, cocaine and THC, are both short lived and persist for no more than a few hours after use. The metabolites remain present substantially longer. Detection of the metabolites increases the ability to detect any recent drug use. However, this limits the utility of the assays for determining whether a patient is currently under the influence of the drug. Because the metabolites are rapidly formed, a negative test result generally excludes toxicity, but a positive test result indicates only use in the recent past, not current toxicity.

To increase sensitivity for detection of less recent drug use, substrates other than urine can be used for drug screening, including hair and meconium. The latter is used to document intrauterine drug exposure (Chap. 31). SAMHSA is planning to develop regulations governing the use of hair, saliva, and sweat specimens for federal workplace testing after their performance characteristics have been adequately studied. This can be expected to increase the availability of clinical testing using these substrates. However, testing performed on hair and sweat is unlikely to offer advantages over testing of serum and urine for the management of toxicologic emergencies.

The stigma attached to a positive test result for an abused drug requires that special care be exercised in performing and reporting the test results. To protect citizens’ rights, many states have legislated specific requirements for workplace drug screening. In some states, the requirements apply only to screening in the workplace, exempting testing for medical purposes. Laws in other states might apply to all drug screening. Although they are not always legally required, some workplace drug-screening practices have been widely applied to all drug screening.

The use of specific cutoff concentrations is nearly universal. Test results are considered positive only when the concentration of drugs in the specimen exceeds a predetermined threshold. This threshold should be set sufficiently high that false-positive results as a consequence of analytic variability or cross-reactivity are extremely infrequent. They should also be low enough to give consistent positive results in persons who are regularly using drugs. Cutoff concentrations used vary with the drug or drug class under investigation. In some drug-screening immunoassays, the laboratory has the option of selecting from several cutoff values.

The use of cutoff values sometimes creates confusion when a patient who is known to have recently used a drug has a negative result reported on a drug screen. In such instances, the drug is usually present but at a concentration below the cutoff value. A quantitative confirmatory test (see below) is usually able to demonstrate the presence of the drug or its metabolites under such circumstances. Another potential problem occurs when a patient’s drug-screening test result is positive after previously having become negative. This is usually interpreted as indicating renewed drug use, but it may actually be an artifact. Urine drug concentrations are directly proportional to the serum drug concentrations but inversely proportional to the rate of urine production. The rate of the urine flow may vary up to 100-fold, with a resulting possible 100-fold change in the urine drug concentration. This effect is often exploited by individuals who drink large quantities of water before taking a urine drug test to increase urine flow and decrease urine drug concentrations. In contrast, a decrease in the rate of urine production may result in a positive test result after a negative one despite no new drug exposure. A similar effect may be produced by changes in urine pH. Drugs containing a basic nitrogen may demonstrate ionic trapping, with increasing concentrations as urine pH decreases. Similarly, excretion of an anionic drug (eg, phenobarbital) may increase with increasing urine pH. Another widely used practice is the confirmation of positive screening results using an analytical methodology different from that used in the screen, such as an immunoassay screen followed by mass spectrometry based confirmation. The possibility of simultaneous false-positive results by two distinct methods is quite low. Clinical laboratories may differ in their policies with regard to confirmatory testing. Some may confirm all positive results from screening immunoassays, but others may not provide any confirmatory testing unless it is explicitly requested.

The most common confirmatory method is gas chromatography/mass spectrometry (GS/MS). The high specificity afforded by the combination of the retention time and the mass spectrum makes false-positive results extremely unlikely. GC/MS also has greater sensitivity than the screening immunoassays, minimizing failed confirmations because of a drug concentration below the sensitivity of the confirmatory assay. Some states require GC/MS confirmation for workplace drug screening, and it may be legally required for all drug screening.

Immunoassay results can generally be obtained within one hour. Confirmatory testing usually requires at least several hours. This can create a problem when confirmation of initial immunoassay results is considered mandatory. Most laboratories provide a verbal report of a presumptive positive result to facilitate medical management but may not enter the result into a permanent record, such as the laboratory computer, until after confirmation has been completed.

Confirmatory testing is less critical in an emergency department setting because a positive finding infrequently has consequences that extend beyond the medical management of the patient. An exception may occur when results of testing performed on motor vehicle crash victims can be subsequently subpoenaed as evidence in legal proceedings. Confirmatory testing also becomes more important in drug abuse testing associated with chronic pain management programs, in which unexpected findings (whether positive or negative) may result in termination of care.

One workplace drug-screening practice that is not widely followed in medical toxicology is maintenance of a chain of custody. Employers generally insist on chain of custody for workplace testing because actions taken in response to a positive result may be contested in court. A chain of custody provides results that are readily defended in court. Laboratories providing testing for medical purposes rarely keep a chain of custody because it is quite expensive and does not benefit the patient. Additionally, the medical personnel responsible for obtaining the specimens are rarely trained in collection requirements for a chain of custody.

Another practice common in workplace testing but rare in medical laboratories is testing for specimen validity. It is common for individuals to try to “beat” a workplace drug test through a variety of means, including diluting the specimen (either physiologically by water ingestion or by direct addition of water to the specimen), substituting “clean” urine obtained from another individual, or adding various substances that will either destroy drugs in the specimen or inactivate the enzymes or antibodies used in the screening immunoassays. Such substances include acids, bases, oxidizing agents (bleach, nitrite, peroxide, peroxidase, iodine, chromate), glutaraldehyde, pyridine, niacin, detergents, and soap. SAMHSA requires validity testing for all specimens in federal workplace testing, including measurement of urinary pH, specific gravity, and creatinine concentration, as well as tests for the presence of adulterants.⁸ Dipsticks are available that detect the most common adulterants. However, manipulation or adulteration is rarely a problem in specimens obtained for emergency medical management, and clinical laboratories may not provide validity testing.

A rapidly expanding epidemic of misuse of opioid analgesics and accompanying deaths due to unintentional overdoses has led to the issuance of extensive recommendations, and often regulations, for managing chronic opioid therapy for noncancer pain. The use of periodic urine drug tests to monitor compliance with prescribed medications, as well as to detect use of illicit or unprescribed drugs, is consistently a feature of such recommendations, although there is currently very little evidence that this practice reduces misuse or diversion of prescribed drugs or improves outcomes for patients.⁶ There is evidence that the results of urine drug monitoring are inconsistent with the agreed-upon pain management plan in a very substantial fraction of patients so tested.⁶

The objectives of urine drug testing in chronic pain patients are more complex than either testing during possible toxicologic emergencies or workplace drug testing in that negative results could signal misuse or diversion of a prescribed opioid, but positive results may indicate illicit drug use. The opioids screening immunoassay is not capable of this task and must be assisted by follow-up testing with a mass spectrometry–based assay capable of distinguishing and quantitating prescription opiates and opioids, as well as key metabolites. This may be done on specimens with a positive screening result to ensure that it is positive for the right reasons and for an unexpected negative result to address the possibility that the expected opioid or its metabolites are actually present but at concentrations too low to give a signal exceeding the cutoff value.

Because of the possible consequences of a positive result in workplace drug testing, all positive results are interpreted by a Medical Review Officer with special training in the pharmacology and metabolism of the tested drugs, as well as the performance characteristics of both the screening and confirmatory tests. No corresponding review is mandated for chronic pain testing, and the interpretation of the results may be left to the patient’s physician, who may not be well prepared for this task.

Medical toxicologists, toxicology laboratory directors, and practicing physicians may frequently get questions about the significance of drug-screening assays, particularly about the causes of false-positive results. Often these questions come from an individual who recently had a positive test result. Table 6–3 summarizes drug-screening test performance characteristics, which are discussed in more detail below.

There are problems with all summaries of performance characteristics of urine drug tests. There are multiple manufacturers of urine drug immunoassays. Each uses proprietary antibodies in a proprietary design and thus has unique performance characteristics, including reactivity with target drugs and cross-reactivity with interfering substances. Reviews and summaries often list performance characteristics, especially sources of false-positive results, without identifying the specific manufacturer of the assay. This can lead readers to assume that the interference applies to all immunoassays for the drug in question. Recent summaries that cite interferences described in earlier misleading summaries may further obscure the lack of generality for a reported source of false-positive results. Reformulation of assays can remedy problems that previously existed, but this does not stop older reports of false-positive results that are no longer valid from being recited in current articles. The interference of ibuprofen in a widely used assay for cannabinoids was corrected more than 20 years ago but continues to be cited in contemporary literature.²⁴ Adding further confusion, tabular summaries of interferences may use negative to mean either that the substance does not interfere or that the substance falsely lowers results. The best source of current information is the laboratory that is doing the testing or the manufacturer of a point-of-care device.

Immunoassays for opiates are directed toward morphine but have good cross-reactivity with many (but not all) structurally similar natural and semisynthetic opioids. The extent of cross-reactivity may vary among manufacturers. For example, oxycodone exhibits approximately 30% cross-reactivity relative to morphine in a fluorescence polarization immunoassay but less than 5% cross-reactivity in a number of other screening assays.^16,27 A failure to appreciate the poor detection of oxycodone can create problems when opiate-screening immunoassays are used to confirm that patients receiving prescription oxycodone for chronic pain are indeed taking it rather than diverting it for illicit sale. If a low cross-reactivity assay is used, a patient taking oxycodone as prescribed might have a negative result, but another patient who is selling the oxycodone and using the proceeds to buy heroin would have a positive result. To address this problem, assays specific for oxycodone have been introduced. These assays are sensitive to therapeutic amounts of oxycodone but relatively insensitive to other opiates.

Synthetic opioids, such as dextromethorphan, fentanyl, meperidine, methadone, propoxyphene, and tramadol, show little or no cross-reactivity in opiate immunoassays. Urine immunoassays specific for meperidine, methadone, and propoxyphene are available. Given the increasing importance of buprenorphine as maintenance therapy for opioid dependency, it is worth noting that the combination of high potency and low cross-reactivity means that buprenorphine will generally not be detected by opiate immunoassays. Immunoassays for specific detection of buprenorphine have therefore been developed.

A positive immunoassay result may reflect multiple contributions from various opiates and opiate metabolites. Concentrations of morphine glucuronide in the urine may be up to 10-fold higher than the concentrations of unchanged morphine and can contribute substantially to positive results. A positive opiate result after the use of heroin (diacetylmorphine) is primarily a result of the morphine and morphine glucuronide that result from heroin metabolism. Distinguishing heroin from other opioids requires detection of 6-monoacetylmorphine, the heroin-specific metabolite. Small amounts of the metabolite may be detected by GC/MS for up to 24 hours after use. A half-life of 5 minutes means that unchanged heroin can only be found in the urine if sampling is done immediately after use.

The duration of positivity of an opiate immunoassay after last use depends on the identity and amount of the opiate used, the specific immunoassay, the cutoff value, and the pharmacokinetics of the individual. Currently, SAMHSA recommends a cutoff equivalent to 2000 ng/mL of morphine for workplace screening because poppy seeds can rarely produce transient positive results with the previously recommended cutoff of 300 ng/mL. However, most toxicology laboratories continue to use a 300 ng/mL cutoff.

Drug-screening assays for “cocaine” are actually assays for the cocaine metabolite benzoylecgonine, which is eliminated more slowly than cocaine. This extends the duration of positivity after last use from a few hours to 2 days and sometimes to one week or longer after prolonged heavy use. Because the assay is directed toward a metabolite, positive results do not equate with toxicity but merely indicate recent exposure. The assay is highly specific for benzoylecgonine, and false-positive results due to an interfering substance are extremely uncommon (Chap. 78). Point-of-care devices may yield false-positive results because of device failure or operator error. False-positive results may also occur with misidentified specimens or (rarely) intentional specimen adulteration.

Immunoassays for cannabinoids are also directed toward a metabolite, in this case tetrahydrocannabinoic acid. These immunoassays exhibit cross-reactivity with other cannabinoids but little else. Because cannabinoids are structurally unique and occur only in plants of the genus Cannabis, false-positive results are uncommon (Chap. 77). It is unusual, although possible, to become exposed to sufficient “secondhand” or sidestream marijuana smoke to develop a positive urine test result.⁵ Legal hemp products include fiber, oil, and seedcake derived from Cannabis varieties with low concentrations of cannabinoids. Hemp food products contain insufficient amounts of THC to produce psychoactive effects and usually will not increase urinary cannabinoid concentrations above a 50 ng/mL screening threshold.^12,25

Interpretation of a positive result for cannabinoids can be problematic. Urine may be positive for up to 3 days after occasional recreational use. However, with heavy or prolonged use, there may be significant accumulation of cannabinoids in adipose tissue. These stored cannabinoids are slowly released into the bloodstream and can produce positive findings for one month or more. Consequently, little can be concluded from a positive finding in terms of current toxicity. Because positive results in the absence of toxicity are very common and because THC is rarely responsible for serious acute toxicity, NACB guidelines recommend against its routine inclusion in drug screening for patients with acute symptoms.³² Synthetic cannabinoids (found in “spice”) are generally not detected by cannabinoids assays, presumably because of a combination of very low concentrations in urine and limited cross-reactivity.

Amphetamine-screening tests have the greatest problems with false-positive results. A number of structurally related compounds may have significant cross-reactivity, including bupropion metabolites³ and nonprescription decongestants such as pseudoephedrine, as well as l-ephedrine, which is found in a variety of herbal preparations. Some nonprescription nasal inhalers contain l-methamphetamine, the less potent levorotary isomer of d-methamphetamine. It is particularly problematic because it not only cross-reacts in immunoassays but also cannot be distinguished from the d-isomer by mass spectrometry.¹¹ This cross-reactivity is beneficial from the point of view of the medical toxicologist because all of these compounds may produce serious stimulant toxicity. But it is problematic in drug abuse screening because of the widespread legitimate use of cold medications. Assays with greater selectivity for amphetamine or methamphetamine have been developed. Although these assays produce fewer false-positive results caused by decongestant cross-reactivity, they are also less sensitive for the detection of other abused amphetaminelike compounds, including methylenedioxyamphetamine (MDA), MDMA, and phentermine. Cross-reactivity patterns vary from assay to assay.¹⁶ The manufacturer’s literature should be consulted for specific details. A number of designer phenethylamine derivatives (such as those found in “bath salts”) might be expected to cross-react in amphetamine assays because of structural similarity. However, they are typically much more potent than amphetamine or methamphetamine and do not achieve high enough concentrations to give a signal exceeding the cutoff.

Testing for benzodiazepines is complicated by the wide array of benzodiazepines that differ substantially in their potency, cross-reactivity, and half-lives. There are also substantial differences in the detection patterns of the various immunoassays.^12,16 This heterogeneity complicates the interpretation of benzodiazepine-screening assays. Screening results may be positive in persons using low therapeutic doses of diazepam but negative after an overdose of a highly potent benzodiazepine such as clonazepam. To improve the scope of detection, some assays use antibodies to oxazepam, which is a metabolite of a number of different benzodiazepines. These assays may have poor sensitivity to benzodiazepines that are not metabolized to oxazepam. False-negative results may occur for benzodiazepines that are excreted in the urine almost entirely as glucuronides that have poor cross-reactivity with antibodies directed toward an unmodified benzodiazepine. This is one reason for the poor detectability of lorazepam in some screening assays. The latter situation has led to the recommendation that specimens be treated with β-glucuronidase before analysis.¹⁹ Some assays now include β-glucuronidase in the reagent mixture or use antibodies directed toward glucuronidated metabolites. The frequency of false-negative results and the fact that benzodiazepines are relatively benign in overdose have led the NACB guidelines to withhold recommendation for routine screening of urine for benzodiazepines until these problems with the immunoassays are addressed.³²

Barbiturates are comparable to benzodiazepines in their heterogeneity of potency, cross-reactivity, and half-lives, although the differences are less substantial. Specific assays for serum phenobarbital can often help to clarify the significance of a positive barbiturate screen.

Some PCP screening assays may give positive results with dextromethorphan, ketamine, or diphenhydramine but only when these are used in amounts above usual therapeutic quantities. A positive result may serve as a clue to a possible overdose with any of these substances. Furthermore, much of the illicit PCP actually consists of a mixture of various congeners and byproducts of synthesis. The cross-reactivity of these xenobiotics with the assay varies significantly and may result in false-negative assay results in patients who use PCP.

Since 1992, medical laboratory testing has been governed by federal regulations (42 CFR part 405 et seq) issued under the authority of the Clinical Laboratory Improvement Amendments of 1988 (often referred to as CLIA-88 or simply CLIA). These regulations apply to all laboratory testing of human specimens for medical purposes regardless of site. They include the universal requirement for possession of an appropriate certificate to perform even the simplest of tests. The remaining requirements depend on the complexity of the test. These regulations become important to the clinician whenever testing is done at the bedside, whether using spot tests or commercial point-of-care devices such as dipsticks, glucose meters, or urine drug-screening devices.

The regulations divide testing into three categories: waived, moderate complexity, and high complexity. Waived tests include a number of specifically designated simple tests, including urine dipsticks, urine pregnancy tests, urine drug-screening immunoassay devices, and blood glucose measurements with a handheld monitor. The only legal requirement for performing waived testing are the possession of an appropriate CLIA certificate (certificate of waiver or higher) and performance of the test in accordance with the manufacturer’s instructions.

There are substantial additional requirements for both moderate and highly complex testing, most of which simply represent good laboratory practice. Table 6–4 lists the most significant of these requirements. Most assays performed with commercial kits or devices are classified as belonging to the moderately complex category. All tests not specifically classified as waived or moderately complex are considered highly complex. This includes essentially all noncommercial tests, including spot tests, because the testing materials have not been subject to review and approval by the Food and Drug Administration.

These regulations have had a substantial impact in all areas of laboratory testing. Some of the most significant effects have been on bedside testing, including spot tests and point-of-care devices. Although clinical laboratories had been following most required practices before the implementation of the regulations, this was usually not the case for testing done at other sites. Most institutions have now established point-of-care testing programs to facilitate compliance with the regulations, as well as with additional requirements of accrediting agencies, such as The Joint Commission. Any toxicologic or other testing done at the point of care should be set up in consultation with the institutional program. Often, all point-of-care testing is done under a CLIA certificate held by the program. There is frequently a point-of-care testing coordinator who may make recommendations or personally assist in efficiently addressing the assorted requirements.

Table 6–4 lists only the most significant requirements of the regulations implementing CLIA. These regulations continue to evolve. Many states also have laws regulating medical testing. Accrediting agencies such as The Joint Commission may have additional requirements. Consultation with the clinical laboratory or with an institutional point-of-care coordinator is recommended before implementing any testing.

Personnel unaccustomed to quality control and assessment practices may find the CLIA requirements initially burdensome. Nonetheless, compliance is important. Following these practices may lead to a threefold reduction in incorrect results,²⁷ thereby greatly improving the quality of care provided to patients. Moreover, noncompliant testing is illegal under federal law and may also be illegal under state law. Any untoward outcome associated with illegal testing creates a major risk management liability for both the institution and the individual. Additionally, billing for any testing that is not CLIA compliant may be considered fraudulent.

Another area in which the CLIA regulations have impacted toxicology testing is in the provision of infrequently requested tests. Meeting regulatory requirements involves a substantial labor investment even when few patient specimens are being tested. Mounting pressures to reduce laboratory costs make it less likely that laboratories will continue to maintain such assays.

Another important regulation, although not part of CLIA regulations, requires that the medical reason for ordering a test be provided with the order. Federal regulations require that the ordering physician provide the diagnosis that establishes the medical necessity for the test, either by name or by diagnostic code (CPT code). Laboratories may not use a “best guess” to assign codes to undocumented test requests.

• Although broad spectrum toxicology screens can identify most drugs present in overdosed patients, the results of broad spectrum screens infrequently have altered management outcomes.

• When requesting a screening test, an important and often overlooked item of communication is specifying any xenobiotics of particular ­concern.

• Xenobiotic concentrations may be useful to make decisions regarding the use of antidotes or if interventions to hasten drug elimination, such as hemodialysis are indicated.

• The predictive power of the result of a toxicology screen depends on a number of factors, including the likelihood of poisoning before receiving the test results (the prior probability or the prevalence), the range of xenobiotics effectively detected, and the frequency of false positive test results.

Additional information about clinical toxicology laboratories, including topics not covered in this chapter, may be found in contemporary textbooks.^17,26

References