Drug Tests’ Reliability Is Limited, Experts Say

The nation’s rush to screen workers for evidence of drug use rests on the premise that drug tests yield consistently reliable results. But such tests in fact are fraught with limitations and can cause people to be falsely labeled as drug users, experts in the field say.

The most popular screening tests can mistake ordinary over-the-counter medications like Midol, Dristan or Triaminic-DM for illegal drugs. Such tests also can be easily fooled by drug users who dilute their urine sample. In addition, the tests are incapable of detecting certain substances, such as “designer” drugs.

Moreover, experts agree, faulty laboratory equipment or procedures, as well as sloppy work by ill-trained technicians, can easily lead to erroneous test results.

Subjective Readings

Errors also can occur because some types of tests require subjective interpretations, turning on questions such as the location, size and color of spots on a piece of laboratory plastic.

Finally, both critics and proponents agree, the tests are incapable of measuring actual impairment at the time a urine sample was given. The tests merely signal the presence of a drug, not the effect it is having on the user. In some cases, the traces of a drug consumed days earlier will still register on a test. Cocaine and heroin, for example, can be detected for as many as two days after use and marijuana for a month or more--long after any effects have worn off.

“There are some very real problems here with what is being asked of these tests and what they can deliver,” said UCLA psycho-pharmacologist Ronald Siegel. “The widespread testing and reliance on telltale traces of drugs in the urine is simply a panic reaction invoked because the normal techniques (of controlling drug use) haven’t worked very well. The next epidemic will be testing abuse.”

Added William H. Anderson, a physician at the Harvard Medical School: “The misuse of screening tests generally seems to arise not out of malice or insensitivity, but rather because of genuine lack of understanding of the limitations of this technology.”

Many of the problems that can cause unreliable test results, however, can be overcome.

For example, experts say, any urine sample that leads to a positive result should be tested a second time for confirmation--preferably using a different kind of test, to minimize the likelihood of a false reading.

And to avoid a simple mix-up of urine samples, experts further recommend a rigorous “chain of custody” procedure that closely tracks a specimen from the time it is produced through the completion of the analytical process.

Most companies now are requiring confirmation of initial positive results. But some state prison systems are not, or else are not confirming positive results by accepted techniques.

“Would you want your physician making the decision to operate on you on the basis of a single blood or urine test? No. You would want other tests, a second opinion, and so forth,” Siegel said. “But in the workplace, a lot of workers are being ‘operated on’--severed from their jobs, prevented from getting jobs, etc., on the basis of a single chemical test.”

Added Anderson, a medical epidemiologist: “The rule of thumb has to be: Never do anybody harm on the basis of a single test.”

Corrections Being Made

To be sure, some employers are belatedly recognizing the limitations of the existing drug screening tests and are making mid-course corrections in their testing programs. Some such changes are being made only as a result of lawsuits or court verdicts.

In February, for instance, U.S. District Judge Louis F. Oberdorfer ordered the reinstatement of a District of Columbia school bus attendant who was fired after a single positive test for marijuana during an annual physical. He ruled that an employee who tests positive cannot be fired without a second, confirmatory test, citing instructions of the test’s manufacturer that any positive finding should be confirmed by an alternate test.

“A few years ago, when some companies did use unconfirmed tests to discharge employees, there was a serious problem,” according to Robert L. DuPont, president of the Center for Behavioral Medicine in Rockville, Md., and former head of the National Institute on Drug Abuse.

“Today, virtually all . . . positive tests are confirmed using a different method of analysis,” DuPont said. “The possibility of some non-drug substance causing a false positive under such circumstances is virtually zero.”

Because of the inherent limitations of most urine tests, many experts say, such a standard also should be extended to job applicants as well as people already on the payroll.

Drug use screening got its start in the mid-1960s amid concerns that American soldiers were becoming addicted to heroin and other drugs in Vietnam.

At the time, the only available technique for urine screening was thin-layer chromatography, commonly referred to as TLC. (Please see accompanying graphic.)

But TLC was--and still is--time consuming, with each procedure taking 45 minutes to 90 minutes, done manually by an experienced technician. Another drawback to TLC is that it “doesn’t give numbers that you can read out and subject to objective interpretation,” said chemist W. Lee Hearn of Toxicology Testing Service in Miami. “It requires a lot of training and experience to interpret.”

(Some laboratories still use TLC for initial screening, although it is now more commonly used to confirm an initial positive test by some other method.)

In the late 1960s, two new types of tests were developed that are capable of detecting very small amounts of biological materials. Both are widely used in drug screening efforts today as well as in general medicine to measure for levels of prescription drugs, hormones and other substances in a patient’s system.

One is the radioimmunoassay (RIA) and the other, enzyme-multiplied immunoassay technique (EMIT). (Please see graphic.)

Limited Use

RIAs require the use of radioisotopes that emit radiation, and therefore must be performed by a technician in a lab that is licensed by the federal government to handle radioactive materials. Thus many labs--and most employers who do their own on-site testing--prefer EMITs, and they are used about twice as often as RIAs, according to market consultant Louis Frisina of Boston Biomedical Consultants.

Both tests are inexpensive, readily adaptable to automation and have a high claimed accuracy--typically 97% to 99% under ideal conditions and circumstances, which experts agree are not easily attainable.

The overriding concern among experts and critics of drug screening is that RIAs and EMITs are highly susceptible to yielding a positive result by mistake--so-called false positives.

Many different substances can produce false positives.

Occasionally, poppy seeds found in some baked goods may contain sufficient traces of morphine to yield a positive urinalysis. (Not all poppy seeds contain morphine, and those that do vary widely in potency.)

Los Angeles toxicologist Gordon Hisayasu of Pacific Toxicology Laboratories said that, as an experiment, he personally consumed 13.5 grams of poppy seeds--under half an ounce--scraped from 5 large bagels. He found that his urine tested positive for opiates for at least 12 hours afterwards.

Certain Peruvian herbal teas may contain coca leaves and, hence, cocaine traces, according to John McLenegan, vice president of operations at PharmChem Laboratories Inc. in Menlo Park.

Tea Drinkers

“There are many examples in the literature of individuals who drink several cups of tea and produced a positive test (up to a day later),” McLenegan said. (Such herbal teas have been banned in the United States since 1983 because of their cocaine content. But they can still be found occasionally in some shops.)

Hisayasu noted that in the most potent teas, 10 tea bags contain about the same amount of cocaine as a typical dose taken by a drug user.

Although positive results from poppy seeds and coca teas are not frequent, Hisayasu said, they are a problem because they are not “false positives”--the drug is actually present. “Even a confirmatory test cannot tell the difference between morphine from a poppy seed and morphine that has been injected,” he said.

It is also widely recognized that many common prescription drugs and even over-the-counter medications can produce positive urinalysis results--even when taken in recommended doses.

Among the over-the-counter drugs that can cause a positive result for amphetamines are antihistamines, which are used for congestion and allergies; propanolamine, which is found in appetite suppressants; and ephedrine, which is found in many cough syrups.

“During the cold and flu season as many as 25% of our amphetamine tests come out positive,” said toxicologist Thorne Butler of Associated Pathologists Laboratory in Las Vegas.

Backup Kit

In fact, the Syva Co. of Palo Alto, which sells EMIT kits, also markets another kit designed specifically to differentiate among such chemicals if the result of the initial assay is positive.

Many other prescription and over-the-counter medications can also produce false positives (Please see graphic). A 1981 study of 161 prescription and over-the-counter medications conducted by pharmacologists Lloyd V. Allen Jr. and Mary L. Stiles of the University of Oklahoma showed that 65 of the drugs caused false positives in one or more of the EMIT tests.

Most employers say they now ask individuals to identify any over-the-counter or prescription medications they may recently have taken before submitting a urine specimen. But even so, experts concede, many people fail to list over-the-counter drugs.

Finally, it is not only synthetic agents that can produce false positives. A small fraction of the population, for example, have certain enzymes in their urine that can lead to false positives.

Interestingly, the growing awareness of the potential for misleading test results itself has led to several false alarms--in which certain chemicals were reported to produce false positives when, in fact, they do not.

Earlier this year, an Atlanta toxicologist said that marijuana tests discriminate against blacks and Latinos because melanin--a pigment found in high quantities in their skin--also produces a positive result in marijuana tests.

But studies by Mamoud ElSohly, a University of Mississippi pharmacologist, showed that neither melanin nor its metabolites produces positive results on a marijuana test. ElSohly also screened urine from blacks who had not used drugs and observed no unusual results.

Furthermore, according to Richard L. Hawks, head of the research technology branch of the National Institute on Drug Abuse, the structure of melanin is not at all similar to that of marijuana.

Another suspected culprit in producing false positives among non-users is marijuana smoke that is inhaled passively.

Unfounded Concerns

But research shows that such concern is largely unfounded, according to University of North Carolina pathologist Arthur McBay.

In 1983, he and a university colleague, pharmacologist Mario Perez-Reyes, sent a graduate student into a small room, measuring 8-by-8-by-10 feet, in which four people were smoking marijuana.

Even under the severest conditions, when each of the smokers had as many as four marijuana “joints” going, the researchers later found only small quantities of marijuana metabolites in the student’s urine--”right at the level of detection of the most sensitive EMIT tests,” McBay said. “Most laboratories use a higher cutoff level in their assays.” He said many similar studies have reached the same conclusions.

Many companies have recently started using higher cutoff levels for the screening tests, McBay said, because a lower detection level increases the chances for error.

With marijuana tests, for example, “we use 100 nanograms because we really want to pick up the heavy or the moderate user--or the recreational user who has used it in the last 24 hours,” according to toxicologist Thorne Butler of Associated Pathologists Laboratory.

“If you start dropping the cutoff way down,” Butler added, “you start picking up more people, sure, but you also have more false positives. At the lower levels, it is also more difficult to confirm the positives. And the cost goes up.”

RIAs and EMITs are incapable of detecting so-called designer drugs--such as fentanyl derivatives, which are used as a heroin substitute. Because such drugs are so potent--often 1,000 times more potent than heroin, for example--they are consumed in such small doses that detection is virtually impossible, according to pharmacologist Gary R. Henderson of the UC Davis School of Medicine.

In addition, the tests are not designed to detect hallucinogens, such as LSD or psylocibin mushrooms, largely because of the widespread belief that such drugs are rarely used nowadays. Another potential shortcoming of RIAs and EMITs is that they can be fooled if a urine sample is diluted by water to reduce the concentration of drugs in the specimen.

Fooled by Contaminants

Similarly, the tests can be fooled by a small quantity of ammonia, vinegar or even table salt, according to Hawks. These agents can sufficiently interfere with the tests’ chemical processes so as to produce negative results. But such adulteration can usually be detected by simple lab tests, such as measuring the acidity of the urine, he said. The vulnerability of urine samples to such tampering is why many employers insist that a person be watched as he gives a sample.

The very fact that both RIAs and EMITs are extremely sensitive also presents something of a problem in itself.

Improper cleaning of equipment and glassware used in the tests can easily leave residues that will contaminate subsequent test samples.

Technicians also can unwittingly contaminate samples, as chemist Werner A. Baumgartner of the Veterans Administration Wadsworth Medical Center learned.

In the course of performing routine urinalyses for the center, Baumgartner’s laboratory once started obtaining abnormally large numbers of positive cocaine tests.

“After some investigation, we found that a small amount of cocaine had somehow gotten on the handle of the restroom door,” Baumgartner recalled. “If the technicians touched the handle, then accidentally touched the inside of the lid in a specimen jar, enough cocaine was transferred to give a positive result.”

The Human Factor

He added: “There are just so many ways of messing up on these things--mixing up the samples . . . contamination in the laboratory. . . . If you have a drug user in the laboratory, he can contaminate the samples. That’s not farfetched. Basically, it’s the human factor that is a problem.”

Human frailty also is a potentially significant factor behind false positives. One person’s urine sample can easily be mixed up for another’s, for instance. “The most likely error to occur is a clerical (mix-up),” said toxicologist Kenneth Campbell, head of Kern Analytical Laboratories.

To avoid such mix-ups, large laboratories--and even some smaller ones--often label a urine sample with a bar code similar to that found on supermarket products. Every time a technician performs an analysis of the sample, the bar code is read with a light wand that provides an accurate identification.

Campbell notes that in his laboratory the identity of the specimen is checked and rechecked throughout the entire analytical process. “If a sample is positive,” he added, “we go back and check all the paper work again.”

The two most widely recommended confirmatory techniques are TLC and gas chromatography/mass spectrometry, or GC/MS, which was developed in the 1970s. (Please see accompanying graphic.)

GC/MS is perhaps the most sensitive and accurate tool yet for the identification of minuscule amounts of chemicals. “GC/MS is virtually 100% accurate,” according to UCLA’s Anthony Daley, who was medical director of the 1984 Olympics and is also chairman of the Major League Baseball drug abuse committee.

GC/MS requires the use of a $150,000 instrument operated by a highly trained technician.

“GC/MS is the standard against which everything else is matched, but that is not to say that it can’t be done improperly,” said UCLA pharmacologist Don Catlin, who is in charge of substance abuse testing for the National Collegiate Athletic Assn. and the U.S. Olympic Committee.

Weak Link in Chain

“You can use very comprehensive, sophisticated equipment--and still have a bad operator,” added Peter Bensinger, former head of the Drug Enforcement Administration and now president of a Chicago drug test consulting firm.

Among 35 companies contacted by The Times, none would admit to using unconfirmed positive results for hiring decisions or disciplinary actions. “Companies that would use such results are very naive,” Hisayasu said. A dozen laboratories contacted also said that they report only confirmed positive results to their clients.

Critics of drug testing fear that some companies that screen job applicants and employees on site may be relying only on unconfirmed positive test results.

Boston Biomedical’s Frisina estimates that 10,000 U.S. companies are large enough to have a medical department--and therefore equipped to do on-site screening. But he estimates that less than 10% of such firms now do so.

The number of companies that perform tests on site is expected to grow significantly because of the emerging availability of simpler assays designed for use by companies.

Many prison systems also screen urine samples from prisoners on site, and some do not confirm positive results or else confirm one immunoassay with a second.

Prison inmates in New York, Washington and New Jersey, for example, have filed suit against those states alleging that they have been subjected to disciplinary actions on the basis of improperly confirmed positive results on screening tests. New Jersey has recently begun confirming positive results, but the other cases are still pending.

California Prisons

The California Department of Corrections has its analyses performed by Analytitox/Hind of San Francisco, and requires confirmation of all positive results, but a positive Emit result for marijuana is confirmed with an RIA--a procedure most scientists find unacceptable.

Roger Newman, general supervisor of the laboratory at Analytitox/Hind, said in a telephone interview that he thinks Emit and RIAs represent different technologies, “and we wouldn’t feel comfortable using an RIA for confirmation if they weren’t.”

The federal prison system uses commercial laboratories for drug abuse screening and requires the labs to confirm all positives.

Meanwhile, American Drug Screens Inc. of Dallas is preparing to market a $24.95 kit targeted at parents who want to test their children for drug use.

The kit includes a specimen bottle, a disclosure form and a mailing label.

When the company receives a specimen, it will screen the sample for marijuana, PCP, cocaine, barbiturates and Valium--and report the results without confirmation.

The firm is awaiting approval of the kit by the U.S. Food and Drug Administration.

Even if appropriate confirmation is carried out for all positive results, according to Harvard’s Anderson, the confirmatory tests “may get the accuracy only up to 95% to 97%,” leaving 3% to 5% to “fall through the cracks.”

INTERFERING WITH URINE TESTING These over-the-counter and prescription drugs used in normal doses can interfere with the results of urine tests.

DRUG BRAND NAME INTERFERES WITH TIME Dextromethorphan Vicks Formula 44-M Immunoassay 1 Day Triaminic DM opiates Phenobarbital Primatene Immunoassay, MS/GC 1-3 days barbiturate Phenylpropanolamine Dietac Immunoassay 1 day Dexatrim amphetamines Cotylenol Triaminic Diphenhydramine Benadryl Immunoassay 1-2 days methadone Ephedrine Primatene Immunoassay 1-2 days Bronkotabs amphetamines Nyquil Amitriptyline Elavil Immunoassay 3-11 days opiates Meperidine Demerol Immunoassay No data opiates Carisoprodol Soma TLC, meprobromate No data Imipramine Tofranil Immunoassay 1-3 days opiates Perylamine Midol, Premensin Immunoassay No data Primatene-M opiates

Source: There is little in the scientific literature about drugs that interfere with drug use tests. But most toxicologists have compiled lists of such data, based on their own laboratory experiences. This list summarizes their consensus.

DETECTING DRUGS Drugs differ in the length of time they can be detected in the blood:

DRUG TYPE DETECTION PERIOD Amphetamines Stimulant 2-4 days Barbiturates* Sedative 12 hours-3 weeks Cocaine Stimulant 2-4 days Heroin Analgesic 2-4 days Marijuana Euphoric 3-10 days (occasional user) 1-2 months (regular user) Methadone Analgesic 2-4 days PCP Anestheti 1-30 days Source: Consensus of toxicologists.

*Secobarbital, pentobarbital, and phenobarbital: 12-24 hours; butabarbital, amobarbital: at least 2 weeks.