College of American Pathologists
Printable Version

Drug culture—ins and outs of DOA screens

February 2003
Cover Story

William Check, PhD

Screening for drugs of abuse is basically clinical chemistry. Sometimes such testing even takes place in a clinical chemistry laboratory. Yet many features make testing for drugs of abuse unique, such as the primacy of specificity.

"Specificity is the most important analytical criterion in this type of testing," says Wayne R. Markus, MD, staff pathologist at Physicians Laboratory Services in Omaha, Neb. "If someone is taking hydromorphone [Dilaudid] or hydrocodone, you wouldn’t want to accuse them of abusing morphine or heroin."

When it became evident that people who ate poppy seed pastries could have enough morphine in their urine to test positive on an opiate screen, some regulatory agencies and vendors raised their cutoff levels from 300 to 2,000 ng/mL to avoid that possibility.

Whether a laboratory uses an opiate assay with a higher cutoff depends on context. Barbarajean Magnani, PhD, MD, acting chief of the Department of Laboratory Medicine and director of clinical chemistry at Boston Medical Center, prefers an opiate assay with the lower cutoff. "We don’t do workplace drug testing, where you are primarily interested in detecting heroin abusers," she explains. Dr. Magnani may be screening a sample from a person brought to the emergency department with an overdose or from a person treated in a pain or rehabilitation clinic. "I am interested in any opiate that the patient may have taken that may produce symptoms or that they may have been prescribed," she says.

Dr. Magnani’s typical phone call demonstrates the difference between her practice and standard clinical chemistry. "I may get a call from a rehabilitation physician treating a patient who had back surgery who is getting slow-release oxycodone," she says. Such patients typically get prescriptions month-to-month and are screened at the end of each month for compliance. "If the test comes back negative and the patient insists that they are taking the drug, the treating physician faces a dilemma," Dr. Magnani continues. "The physician wonders whether the patient is maybe selling the drug or whether the drug is being diverted by a family member, especially if the patient says the prescription is not adequate and asks for a higher dose. At this point, they call me."

Dr. Magnani explains to the treating physician that opiate screening tests may not detect the amount of oxycodone typically prescribed for pain. And she has to convince the physician not to accuse the patient of nonadherence or of selling the drug, "which can be devastating to the patient." Dr. Magnani sends the sample out for a confirmatory test. "Many times the test comes back positive," she says, "which validates the patient and makes the physician feel better."

Tai Kwong, PhD, professor of pathology and laboratory medicine and director of the regional toxicology laboratory at the University of Rochester Medical Center, specializes in screening urine samples from people in drug-treatment programs. He too spends time clearing up misunderstandings about test results. Dr. Kwong’s typical phone call follows a positive drug test. "The patient claims he has been clean for so many weeks or months and the positive test must be residual excretion," Dr. Kwong says. The patient could be telling the truth—after a person has stopped using a drug, there is a window during which it is still detectable. This period varies by drug, dose, and the extent of the user’s experience.

"First I try to get as much information as possible from the clinician," Dr. Kwong says. "And I have to help the physician or nurse question the patient more closely to determine their credibility. Sometimes I can rule out the patient’s claim, and sometimes I can only say it’s possible."

A different set of issues confronts Jeffrey Jentzen, MD, chief medical examiner for Milwaukee County, Wis. "This is a very challenging field," Dr. Jentzen says, "because of test interpretation problems as well as the difficulty of obtaining histories for dead people." Dr. Jentzen’s difficulty in determining whether a death was due to drug abuse is aggravated by the fact that the age for people using illicit drugs is increasing.

"Just yesterday we had a person 72 years old who died from cocaine toxicity," he says. "Stroke, heart attack, aortic dissection, seizures—you can’t assume any longer that those events are not drug related. They all can be caused by cocaine."

Laboratories doing workplace screening for drugs of abuse face complications not often seen in a standard clinical chemistry laboratory. Patients having urine glucose measurements don’t alter their specimens, but some workers being tested for drugs of abuse do. "Substituted specimens are the hot issue in workplace testing today," says Michael Peat, PhD. Now vice president of operations at Atherotech, he spent many years managing workplace screening for LabOne in Kansas City. Patients may dilute their urine specimens with water before giving them to a collector. Dr. Peat says four to five years ago the government issued a definition of a substituted specimen based on creatinine and specific gravity. "Now the majority of regulated laboratories test for substitution and adulteration, and it is being increasingly done outside of regulated laboratories," he says.

Almost all laboratories screening for drugs of abuse share one important feature with standard clinical chemistry laboratories: the need to define a limited test panel. "Years ago if a person came in with an overdose we would do a comprehensive screen," Dr. Magnani says. "Those are fairly expensive. Most laboratories today don’t have the equipment to do that. And with the technologist shortage, we don’t have the luxury anymore of having technologists who specialize in toxicology and having a separate toxicology laboratory." Now almost all non-reference laboratories that test for drugs of abuse do immunoassay screening and offer a limited menu of tests.

One common panel, Dr. Markus says, is the "NIDA five," mandated for all regulated workplace screening: amphetamines (d-amphetamine and methamphetamine), opiates (morphine and codeine), phencyclidine (PCP), cocaine metabolite (benzoylecgonine), and marijuana metabolite (11-nor-delta-9-tetrahydrocannabinol-9-carboxylic acid, THCA). Barbiturates and benzodiazepines (BDZ) can be added to make a seven-drug panel, while some employers opt for a nine-drug panel with propoxyphene (Darvon) and methadone. Amphetamine assays can be expanded to include MDMA (ecstasy), and specific confirmation assays for the opioids hydromorphone (Dilaudid) and hydrocodone (Vicodan) can be added. (Standard opiate screening tests do not reliably detect other opioids, including oxycodone, methadone, and propoxyphene.)

Outside regulated workplace screening, variations are common. Dr. Kwong substitutes BDZ for PCP. Dr. Magnani includes barbiturates and BDZ in her initial five-drug screen and offers PCP and THCA on an expanded panel. "We don’t see much PCP here in Boston," she says. "I only run that if they ask for it." Benzodiazepines have a low abuse potential but are so heavily prescribed that some abuse still occurs. The BDZ result could help diagnose a patient who comes to the emergency room mildly neurologically impaired, Dr. Magnani notes. "The physician might be wondering, Should I get a head CT or could this be benzodiazepine-induced drowsiness and confusion?"

Screens rarely include methaqualone (Quaalude), which is mostly a holdover from the past. "Methaqualone has not been seen since the 1980s, when it was made a schedule one drug," Dr. Peat says. "But some employer-employee agreements for drug testing negotiated years ago are still in force."

In the hospital setting, clinicians may order tests for two so-called date rape drugs: Rohypnol (flunitrazepam) and GHB (gamma hydroxybutyrate), a derivative of the neurotransmitter GABA. "You only need a very low dose of Rohypnol to knock someone out," Dr. Markus says, "and it can be fairly hard to detect on some standard benzodiazepine assays. I am concerned we won’t be able to pick it up."

Dr. Magnani, too, worries about the adequacy of available tests. "I would like to urge vendors to start making screening products that are appropriate for what we see today," she says. Teenagers can overdose on dextromethorphan, which current screening assays do not pick up. Another example is MDMA. "Some amphetamine assays will detect it," Dr. Magnani says, "but others don’t. One company has come out with a specific assay for MDMA. I would like to see more of that."

Dr. Jentzen cites the opioid oxycodone as an instance of vendors responding to new drugs of abuse. "Oxycodone is not easily detected by routine opiate screening methods," he says. "You need a directed assay. Manufacturers responded very quickly in providing directed assays for oxycodone."

Accurate oxycodone detection has become a major issue. Soon after slow-release oxycodone (Oxycontin) became available, drug users found they could get a big, fairly cheap high by crushing and snorting the pills. So emergency room physicians and drug-treatment programs order screening for oxycodone. And, as Dr. Magnani notes, its use is monitored in pain patients who have a prescription, a practice stimulated in part by JCAHO’s pain control mandate.

Historically, opiate screening assays were designed to detect heroin, and unfortunately they don’t always pick up oxycodone. "This is not well understood by health centers that are affiliated with a hospital rehabilitation program," Dr. Magnani says. When she gets a negative result for oxycodone, she sends it out for confirmatory testing by gas chromatography/mass spectrometry, or GCMS. "I hate to see a patient dropped out of a rehab program for the wrong reason," she says.

Dr. Kwong has a similar perspective. "It is amazing how many physicians don’t realize that an opiate assay does not pick up oxycodone very well," he says. "This is a real concern." When a physician sends a urine specimen from a pain patient being treated with Oxycontin to be tested for the presence of the drug, the physician should tell the laboratory to look specifically for oxycodone, Dr. Kwong says. "Then we can go directly to an HPLC [high pressure liquid chromatography] or GCMS assay."

Similarly, methadone tests are often ordered to check compliance in heroin abusers on methadone maintenance, but many professionals don’t know that opiate assays don’t detect methadone and that a special test must be ordered. Dr. Magnani gets calls from physicians whose patients have tested positive on an opiate test and who assume that means the patient is taking the methadone. "I tell them, No, your heroin abuser is most likely using heroin again.

"The story I really like," she continues, "was the physician who called me about a urine specimen he had sent to a reference laboratory from a heroin abuser in treatment." The physician had sent the sample to find out whether the patient was taking codeine that had been prescribed for postoperative pain. The physician told Dr. Magnani that the test result had come back consistent with codeine. When she asked for details, he told her the test found morphine, codeine, and 6-monoacetyl morphine (6-MAM). Codeine is metabolized to morphine, so those compounds were to be expected. And the clinician assumed that 6-MAM was another metabolite of codeine.

"But 6-MAM is only a metabolite of heroin," Dr. Magnani says. "So I told him that the test result really meant his patient was back on heroin. The reference laboratory just reported what they found, not what it implied or what the metabolic pathways are. That’s why someone has to be available to answer clinicians’ questions."

Dr. Kwong works with methadone maintenance programs for treating heroin addicts. They are interested primarily in whether clients have relapsed into illicit drug use. So, to save money, they may not test for methadone routinely, since most of the patients are taking their methadone dose and should test positive. "Methadone programs don’t order methadone testing routinely," he says. "The only time they want us to check for methadone is when they suspect a patient is not taking it."

Although some clients request urine alcohol in a screen for drugs of abuse, in Dr. Markus’ view this is not prudent. "Alcohol is a legal drug," he points out. People can drink a few hours before going to work and still have alcohol in their urine during work. "So if you do find a small concentration of alcohol, what would you do?" he asks. "Most employers don’t require that you never drink, while they can require that you never illegally use drugs." Breath alcohol testing, used by some bigger companies, is a much better technology than urine testing, Dr. Markus says: "A breath test or blood alcohol concentration reflects your state of intoxication at that moment."

As a programmatic requirement, Dr. Kwong says, some methadone programs test all patients for alcohol when they come to get their methadone dose. They may choose to use a breath alcohol test for an instantaneous result. "A urine alcohol test sent to the laboratory takes too long for the results to come back," he says.

Some drug-treatment programs present special needs. "For some programs, the range of drugs that we may have to help them monitor can be quite extensive. Most are not detectable by commercially available immunoassays," Dr. Kwong says. For example, such programs may treat adolescents abusing dextromethorphan. "We have to set up something special for that," he says.

Another especially challenging type of program treats drug-abusing health professionals. Their treatment may be under the supervision of a regulatory agency with specific laboratory monitoring requirements. "Physicians, nurses, and pharmacists have access to drugs that are not typical street drugs," Dr. Kwong notes. "In addition to the standard drugs, like morphine and codeine, those programs would have specific drugs they want to look for. These would include opiates like hydrocodone, hydromorphone, oxycodone, and nalbuphine, and opioids like fentanyl, meperidine, methadone, and propoxyphene." Everyone in these programs has a fairly specific history of abuse, and the test volume for any given drug is low. "It is hard for a clinical toxicology laboratory to maintain quality and proficiency for so many drugs under these conditions," Dr. Kwong says. For this reason, a few specialty laboratories do all of this testing. "I tell the treatment programs, ’Don’t even send those specimens to me,’" Dr. Kwong says.

He also declines to test for LSD routinely. "There is still a low level of LSD activity out there," he says, "but we are fooling ourselves if we think that we can detect it accurately in a typical clinical laboratory setting. Immunoassays for LSD are notoriously nonspecific and confirmation is very difficult." If a patient’s history or clinical symptoms are strongly suggestive of LSD, Dr. Kwong sends the specimen to a reference laboratory. The long turnaround time means that the result does not contribute to the immediate management plan for the patient.

Dr. Kwong is being asked to do more drugs-of-abuse testing for organ transplantation purposes. "There is a long waiting list for donor organs," he says, "and transplant programs want to evaluate whether potential recipients are substance abusers." Results can affect patients’ candidacy and their ranking on the list.

As a medical examiner, Dr. Jentzen faces unique issues.
"Drugs are becoming an increasingly important part of death investigations," Dr. Jentzen says. "Almost half of all cases we do have something to do with drug toxicity or drug-to-drug interaction." Illicit drugs contribute to many of these cases, but a large portion of the medical examiner’s clientele abuses therapeutic drugs, such as pain medications or antidepressants. "And a high percentage take an intentional overdose," he reports.

Drugs of abuse tend to come in waves, he finds. In the mid- to late ’80s, propoxyphene was popular. Before that it was barbiturates. Further back, BDZs were abused. Now heroin has become more readily available and oxycodone is causing many accidental deaths. "It’s always changing, depending on what drug is in fashion and readily available," Dr. Jentzen says.

Determining which drug was involved in any particular death is challenging. In some cases the patient may have been resuscitated in the emergency room and interviewed before dying. In other cases drug paraphernalia or empty containers or pill vials found at the scene of death indicate what the deceased took. In many cases, however, there is no patient interview and no physical evidence. "So we don’t know what they may have taken," Dr. Jentzen says. Even when there are hints, full testing is still mandatory. That dictates broad-spectrum screening.

"Our work is different from occupational drug testing, which looks at a limited number of drugs," Dr. Jentzen says. "We would not be content to do specific directed assays. We do a broad screen, then directed confirmation." Dr. Jentzen uses Toxi-Lab, a commercial thin-layer liquid chromatography method, to screen urine, and he uses GC or GCMS, or both, to screen blood samples.

Sometimes simply obtaining an adequate sample for postmortem testing is troublesome. "Physicians working in a hospital have the luxury of taking blood from a person alive and talking to you," Dr. Jentzen says. "You get samples in good condition and can even take repeat samples." In postmortem drug testing, the person may have been dead for a prolonged time and recovering sufficient volume for testing may be difficult. On the other hand, blood isn’t the only sample available; there are liver and brain tissue, gastric contents, and vitreous fluid, which has the advantage of not deteriorating as much as other body fluids.

Postmortem drug redistribution complicates the interpretation of test results. "Postmortem drug redistribution has been appreciated only for the last 10 years or so," Dr. Jentzen says. "Before, if we saw a very high concentration of a drug in the blood, we would have said that person took an overdose. Now we know that drug can come out of tissue [after death] and artifactually elevate the amount in blood." A pathologist who looked at a postmortem blood drug level and concluded it was high enough to cause death would be wrong in 40 percent of cases. To minimize this possibility, peripheral samples—arteries and veins of the legs and upper extremities—are taken. "We also need to know which drugs are more prone to postmortem drug redistribution," Dr. Jentzen says, such as antidepressants and digoxin. In those cases, tissue levels are measured.

New long-acting drugs complicate interpretation as well. With Oxycontin, for example, a large number of pills in the stomach doesn’t necessarily translate into a high blood drug level, because the pills break down slowly over time.

"A new concept that we are involved in here is pharmacogenetics," Dr. Jentzen says. "We now know that genetics plays a large role in drug metabolism." In general, cytochrome P450 liver enzymes strongly influence drug levels. "Now in our laboratory we routinely do genetic testing when a drug level doesn’t fit with the history or scene-of-death information," Dr. Jentzen says. In a recent case, a woman’s very high antidepressant drug level did not fit her clinical history. Dr. Jentzen determined that she lacked one of the enzymes needed to metabolize that drug. "So instead of calling it a suicide, we called it an accidental overdose."

Maintaining a chain of custody is an extra burden on any laboratory doing postmortem drug testing or any other kind of forensic drug testing. "Without complete documentation, we might as well throw out the sample," Dr. Jentzen says. "It will not hold up in court."

Forensic issues can arise in workplace testing and in hospitals
as well. Dr. Magnani cites the case of a female known drug abuser who delivers a baby. The baby’s urine or meconium may be sent for a drug screen. "If we are just concerned about treating the baby and the baby’s urine tests positive for cocaine, there is no problem," she says. "But if there is any question that mom and baby might have social services involved and there might be a filing to remove baby from mother, you had better be sure you have established a chain of custody. Otherwise all you will be able to say in court is that you got a positive result on a tube with the baby’s name on it."

Workplace testing for drugs of abuse differs in substantive ways from testing in other settings, Dr. Peat says. Workplace drug testing may be regulated or nonregulated. Regulated testing covers federal government agencies and companies required under government regulation to test certain employees for drugs of abuse. "The principal federal agency that requires drug testing in the workplace is the Department of Transportation," Dr. Peat says.

Testing for drugs of abuse in the nonregulated arena occurs chiefly by union agreement and, more frequently, by company policy. It generally includes preemployment, random, and post-accident testing. "Much less random testing is done in the nonregulated arena," he says.

Positivity rates in workplace drug testing are very low, says Dr. Peat, and have been declining over the past decade, as data on Quest Diagnostics’ Web site demonstrate. In 2001, tests performed by Quest in the general U.S. workforce had the following positivity rates: amphetamines, 0.29 percent; barbiturates, 0.34 percent; BDZ, 0.6 percent; cocaine, 0.69 percent; marijuana, 3.17 percent; opiates, 0.29 percent; PCP, 0.02 percent; propoxyphene, 0.52 percent.

"Another major difference between workplace and non-workplace drug testing is the role played by the medical review officer," Dr. Peat says. A medical review officer, or MRO, is an occupational health physician who can be an independent contractor or an employee of a corporation but not an employee of a laboratory. "A number of large companies have their own occupational health physicians," Dr. Peat says.

An MRO’s most important function is determining whether a positive drug test can be explained by authorized use of a legitimate drug. Use of Tylenol No. 3 containing codeine, for example, results in urine positive for codeine and morphine. If an MRO verifies that use, he or she will report the result to the employer as negative. In principle, the employer never knows the specimen was positive.

Validity testing—detecting specimens that have been adulterated or substituted—poses a major challenge in workplace drug testing. Urine can be adulterated with tampering agents bought over the Web and ingested. "Companies that sell these products claim they clean you up," Dr. Peat says. "But their product inserts also suggest that you drink copious amounts of water, which sounds like flushing drugs out of your system."

Drug users can also buy products to add to urine after it’s collected. "These products come in small tubes that you can hide in your socks," Dr. Peat says. "In today’s world, their main purpose is to destroy THCA." Since these adulterants are basically oxidizing agents—nitrites, bleach, pyridinium chlorochromate—reference laboratories have devised processes to detect them. "Quest Diagnostics in particular has been enthusiastic about detecting oxidants," Dr. Peat says. According to Quest’s Web site, the fraction of urine samples containing oxidants declined from 0.23 percent in 1999 to 0.05 percent in 2001.

People have also used bleach or household soap as adulterants for decades. "Depending on the amount and on the laboratory, those products can often be detected," says Dr. Peat.

Substituting urine specimens has come to the fore in the last few years. As Dr. Peat notes, one method is to dilute urine with water, which can be detected by measuring creatinine and specific gravity. The MRO, in what Dr. Peat calls a "gatekeeper capacity," must review these results to ensure there is no medical explanation.

"Say the laboratory concludes that creatinine and specific gravity are too low and not consistent with human urine and the specimen was substituted," Dr. Peat says. "The MRO can contact the donor and give them a chance to explain the result. The reviewer may suggest that person consult a urologist, who sends a report to the MRO. Then the MRO makes a final determination."

A more sophisticated form of substitution is to purchase clean urine online and carry it under an arm in a bag to keep it near body temperature. "Females sometimes carry negative urine in their vagina in condoms and pierce them when they need to give a specimen," Dr. Peat says. This practice is hard to detect without DNA testing. "Some people believe you should do directly observed urine collection," Dr. Peat says. "Directly observed collection is mandatory in the military but only rarely performed elsewhere."

At the frontier of drug testing, says Dr. Peat, are attempts to use specimens other than urine, especially oral fluid and hair. "Hair has been used for a number of years and oral fluid for the last three years or so," he says. Hair and oral fluid can be tested in immunoassays but not those used for urine. Both are accepted alternative tests to urine but are not currently approved specimens in the regulated arena.

Neither oral fluids nor hair detects anything that urine doesn’t, but each has putative advantages. "People who do hair testing claim it can detect drugs for longer times than urine," Dr. Peat says. "There is some truth in that, but a lot of debate." Hair is unsuitable for post-accident drug testing, for example. If a person took a drug shortly before an accident or took just a single dose, a detectable concentration would not have built up in hair.

Oral fluid, on the other hand, has a window for detection similar to that of urine. Its clinical sensitivity and pickup rate also resemble that of urine. Oral fluid’s main advantage is that it is more suitable than urine for directly observed collection. Dr. Peat says LabOne has done the majority of work on oral fluid to date.

In technology, Dr. Peat says, "probably the biggest change in the last five years outside the government arena has been increased use of point-of-collection testing devices for urine and oral fluids." A large number of handheld products are available with variable pickup rates. "Do thorough market research before buying," Dr. Peat advises. Point-of-collection devices are used primarily in preemployment screening. "You can generally get a person into the workplace sooner with them," Dr. Peat says.

These diverse sites of drugs-of-abuse testing share one mandate—the need for good communication between the treating physician and the laboratory to avoid misunderstandings about the test results. "We pathologists have to be very clear about the capabilities and limitations of the tests we use," Dr. Magnani says, "so that we can advise clinicians when to send for confirmatory testing." Education is one form of communication. "The pathologist’s role is to educate staff," Dr. Magnani says. "Not just laboratory staff but medical staff as well. For example, I have printed up charts available at the hospital that detail what we can and cannot test for."

Dr. Kwong agrees. "We have to educate clinicians, nurses, and drug counselors about what they should and should not expect from a drug test," he says. "My general principle is that as long as the laboratory understands the limitations of its assays, and communicates them to clinicians, then the clinicians know when to ask the laboratory for a special test."

Dr. Kwong’s bottom line: "If clinicians don’t talk to you, you talk to them."

William Check is a medical writer in Wilmette, Ill.