February 2012
Feature Story

Anne Paxton

Among drugs of abuse, opiates tend to play most of the leads, while benzodiazepines are supporting actors, and barbiturates are faded stars that now barely qualify as bit players. But testing for benzodiazepines and barbiturates, as well as less common drugs of abuse such as club drugs, remains an important part of clinical toxicology, presenting challenges and potential pitfalls that laboratories need to help clinicians be aware of.

Dr. Stacy Melanson (left) with Alicia Darragh, senior medical technologist, toxicology. The lab at Brigham and Women’s is formalizing a pain toxicology consultative service. It’s doing five to 10 such consults monthly now, mostly with primary care physicians.

Urine barbiturate and benzodiazepine immunoassays can provide rapid preliminary results, often within a matter of minutes, says Marion L. Snyder, PhD, technical director of laboratories at Greenville (SC) Hospital System University Medical Center. But these screening results are presumptive. The benzodiazepine immunoassays are known to exhibit non-specific antibody binding, and drugs such as oxaprozin, among others, have been shown to generate false-positive responses. On the plus side, many clinicians are aware that immunoassays suffer from false-positives and that testing by a more specific and sensitive method, such as gas chromatography/mass spectrometry (GC-MS) or liquid chromatography-tandem mass spectrometry (LC-MS/MS), is necessary for a definitive result.

However, “I would say physicians aren’t well aware of the potential for false-negatives,” Dr. Snyder says. In her experience, “there is a great lack of understanding regarding the potential for false-negatives with benzodiazepine immunoassays.” Similarly, though barbiturate immunoassays are more reliable, “clinicians are frequently unaware that immunoassay screens for barbiturates will not detect all barbiturates equally and that misleading negative results may be obtained, even for a patient taking relatively high levels of the drug.”

Traditionally, testing for benzodiazepines and barbiturates was performed for the hospital emergency department in cases of suspected toxicity, or for workplace or forensic purposes. Those needs continue, but the increasing prevalence of pain management programs has created a need for compliance monitoring that has altered ordering patterns significantly. Now, the largest demand for testing of benzodiazepines and barbiturates originates with physicians who are gauging therapy compliance among their pain management patients, since these two classes of drugs may be prescribed or abused in conjunction with the opiates prescribed for pain. “It has changed quite dramatically in the past 10 years,” Dr. Snyder says.

Dr. Snyder, who co-authored the chapters on benzodiazepines and barbiturates in the newly published Clinical Toxicology Testing—A Guide for Laboratory Professionals, issued by CAP Press and the CAP Toxicology Resource Committee (edited by Barbarajean Magnani, PhD, MD; Michael G. Bissell, MD, PhD, MPH; Tai C. Kwong, PhD, DABCC; and Alan H.B. Wu, PhD, DABCC), was hired as director of toxicology at Brigham and Women’s Hospital a few years ago to set up a drug testing program at the Boston institution. “The main reason was to support pain management physicians because the volume coming from them had grown so much. And they had different demands than the urgent care physicians had had previously, including more extensive panels and lower cutoffs.”

Qualitative urine screening with immunoassays in the emergent setting is performed primarily to reinforce the clinical symptoms of the patient and the patient’s history, says Stacy Melanson, MD, PhD, associate director of clinical chemistry, Brigham and Women’s Hospital, and a co-author of Clinical Toxicology Testing, for which she co-wrote the chapters on benzodiazepines and barbiturates. “The ED has a panel of eight tests geared toward drugs we see more commonly abused in the Boston area: amphetamines, barbs, benzos, cocaine metabolite, methadone, opiates, oxycodone, and cannabinoids. The patient may have symptoms consistent with an opiate overdose such as respiratory depression, and an empty bottle of phenobarbital in her purse, so we’ll do a barb screen and if it’s positive we won’t send it for confirmation unless the clinician specifically requests it.” If such a request is made in the emergent setting, the test is sent out for GC-MS or LC-MS/MS, which takes a few days. The hospital does perform LC-MS/MS testing for benzodiazepines along with opiates. However, “We only perform the testing two days a week to support the pain management center,” Dr. Melanson says. “Therefore, it is not necessarily useful for the ER doctors.”

Federal guidelines for pain management programs now say that clinicians need to monitor patients being treated for chronic pain for undisclosed and illicit use of other substances, Dr. Melanson says. And there is more than one irregularity that prompts pain management physicians to worry about compliance. “Physicians need to know,” Dr. Snyder notes, “whether their patients are taking all of their medication at the right dose, and not diverting or selling it.” For example, even though benzodiazepines may not have the high street value of opiates, patients may still divert them. Clonazepam, lorazepam, and alprazolam are among the top drugs prescribed in this country and are widely available; they are the second most often abused class of prescribed drugs behind opiates.

Some patients will abstain from water or drink as little as possible to concentrate their urine as much as they can, to make it appear they are taking the drugs as prescribed. Other patients will drink a lot of water to try to dilute out some of the illicit drugs they’re abusing. Creatinine levels can help check on whether this is happening, Dr. Snyder says. “We do measure creatinine, which is excreted at a relatively constant rate, so if they have a very low urine creatinine, this suggests that the patient diluted their urine or consumed a lot of water.”

Dr. Snyder and Dr. Melanson have found the same phenomenon with pain management patients both in the large city of Boston, and the small city in the south, Greenville, where Dr. Snyder works now: Between 30 percent and 40 percent of the patients being treated at these sites are testing positive for benzos. Barbiturates are found more rarely. “When physicians are trying to monitor compliance they also want to test for a wide range of drugs of abuse and that generally includes barbiturates,” Dr. Snyder says. “Since the 1970s, they’ve been included in screening panels. As far as positivity, though, it’s very low. We particularly don’t see any unexpected positives; it just isn’t that common.”

The available benzodiazepine immunoassays work very well with Valium, one of the earliest benzodiazepines, Dr. Snyder says. “As a result, most physicians think the screening tests will detect newer benzodiazepines such as lorazepam, which is a commonly prescribed benzodiazepine. And it’s not going to be detected.” The available screening methods frequently miss lorazepam, even at relatively high concentrations, because it is excreted primarily as a metabolite, lorazepam glucuronide, that is not recognized by the antibodies used in the assays. In a recent study that Dr. Snyder co-directed, relatively high concentrations of lorazepam, in the thousands of ng/mL, were missed by the assays being evaluated.

No one kit of which Dr. Snyder is aware will recognize all benzodiazepines or their relevant metabolites at concentrations encountered with therapeutic use. “The ‘classic’ benzodiazepines that share metabolites with diazepam are best detected by currently marketed benzodiazepine screening immunoassays,” she says. “Barbiturate immunoassays also show variable sensitivity within and across assays for detecting different barbiturates within the class.” She believes that clinicians who manage pain need to carefully interpret the absence of the benzodiazepine or barbiturate class when assessing treatment adherence, keeping in mind that not all benzodiazepines and barbiturates are detectable by all methods.

In the absence of a sufficiently sensitive immunoassay, direct analysis by a more sensitive method such as LC-MS/MS should be considered, Dr. Snyder says, because it can provide highly specific detection of benzodiazepines and barbiturates often at concentrations well below those detected by immunoassay and other available methods, such as thin-layer chromatography. “However, these methods are generally time-consuming and complex, making them impractical for many hospital laboratories. It may take more than a week for the clinician to receive the results if the testing has to be performed at an outside laboratory.”

Laboratory directors’ practices in communicating the characteristics of their drugs-of-abuse assays to clinicians may vary. At Brigham and Women’s Hospital, Dr. Melanson says, “We don’t have a lot of interaction with clinicians in the emergent setting. It’s more of a supportive role. The question we might typically get a couple of times a year is, ‘My patient said they are taking lorazepam and the benzodiazepine screen is negative—does this make sense?’ They worry that the patient is lying because the result should have been positive, so we explain to them the specificity of the assay and stress that they need to send out the test for confirmation to be sure.”

The pain management physicians are getting more extensive panels and automatic confirmations of any positive results. “Seventy percent of the patients are positive for opiates, and about 40 percent are positive for benzos, so it doesn’t really make sense to do a screening test. We go right to LC-MS/MS for opiates, and we recently started doing the same for benzos, so we actually get a lot of consults from them,” Dr. Melanson says, adding that the laboratory is in the process of formalizing a pain toxicology consultative service. This would be similar to a clinician consulting on renal failure or infectious diseases where a formal note would go into the patient chart. “The lab is already doing five to 10 of these consults a month. They’re mostly with the primary care physicians, who don’t do pain management on a day-to day basis but who are managing patients on oxycodone or hydrocodone.”

The opiate assays often have the most complex results, but benzodiazepines too can create confusion, she points out. “For example, diazepam is metabolized to nordiazepam, temazepam, and oxazepam, which are all detected by our LC-MS/MS panel. The three positive results indicate diazepam use, but the clinician might think the patient is taking more than one benzo,” Dr. Melanson says. A lot of the antibodies in the assays are directed to the older benzos and not the newer ones, which are more potent and can be given at lower concentrations, she adds. “So even the cutoff for benzos may be an issue. Typically 200 ng/mL may be too high; we may need to go lower to detect these more potent benzos.”

Clinicians are often unaware that many benzodiazepines are excreted as metabolites, Dr. Snyder says, and that cross-reactivity of the primary metabolite can be far more important than that of the parent drug for detecting its use. “This lack of knowledge can have significant clinical consequences,” she says. “If a patient prescribed lorazepam, for example, screens falsely negative, the clinician may conclude that the patient is diverting their medication and discontinue treatment.” Dr. Snyder says it is critical that the toxicology laboratory communicate with physicians to make them aware of known interferences, and to make sure they know certain drugs within a class may not be detected.

Ideally, says Dr. Melanson, the laboratory would include a comment with every result outlining the most common sources of false- positives and false-negatives observed with each assay. “But we would get a lot of complaints from clinicians if we add long footnotes, so it is impractical. We have to be minimalistic in that sense.” Her laboratory is now working on creating a reference to a dedicated Web page in the electronic health record that would give clinicians the details about their test results along with the caveats.

Given the ever-evolving landscape of clinical drug testing, Dr. Snyder says laboratory directors will need to come up with a way of providing real-time updates regarding the current characteristics of their toxicology testing, including drugs detected, analytical cutoffs, and possible interferences and limitations for each method. “This may be accomplished in the form of regularly scheduled meetings and open dialogue with physicians’ groups, a continually updated Web site, or through individualized interpretations that combine the patient’s medication list with the final results of their urine drug testing.” Several specialty toxicology laboratories have begun offering these types of interpretations, Dr. Snyder notes.

CAP proficiency testing data highlight the significant within and between assay variability in terms of what drugs are detected and at what concentrations, she adds. However, as far as being able to distinguish which assays will work best in any particular laboratory, Dr. Snyder finds it rather hard to tell from CAP Surveys. With benzos specifically, the CAP may have included lorazepam in the challenge and the assays would detect it. “But if they challenge with the metabolites that are found in the urine, detection rates would be much lower. Simulating the metabolic profiles of urine in prepared PT materials is difficult to do,” she says.

Club drugs, notably methylenedioxymetham-phetamine (MDMA, or Ecstasy), dextromethorphan (DXM), and gamma-hydroxybutyric acid (GHB), may be fixtures of newspaper headlines, but are not that common in the toxicology arena, says Donald L. Frederick, PhD, a toxicology specialist with Peoria Tazewell Pathology Group and clinical assistant professor at the University of Illinois College of Medicine at Peoria. Dr. Frederick wrote a chapter on club drugs for the new CAP Clinical Toxicology Testing book.

“We don’t see that much abuse of club drugs. It happens rarely and it’s by some adolescent who may have been exposed to it at a party.” As a result, he says, an order for a club drug test would have to be a special request based on the clinical symptoms a doctor may see in the ER, and the test would likely be referred to one of the large reference labs. “For a community hospital to put in this assay for MDMA or any of the other drugs on a routine basis is not very likely.”

Moreover, the range of available assays is narrow, he says. “There’s an MDMA test that’s been available for a while now, but it’s not very good and has a lot of false-positives. So you’d end up having to go to the specific GC-MS test for the drug because the screening tests are not very accurate.” The vast majority of the tests are negative.

Beyond the low specificity, cross-reactivity can be an issue too. “We don’t really see dextromethor-phan as a club drug, but it does cross-react with the PCP test,” Dr. Frederick says. “So if you have a result positive for PCP, a lot of times that would be just due to large concentrations of cold medicine that someone took accidentally, not purposely getting high.” He would have to warn clinicians that they may get a positive for PCP when the drug is really DXM. “You have to train the physicians to look at what the specificity is of these tests.”

The biggest misunderstanding he sees is that physicians often think when they screen for amphetamines they will get some of the club drugs too. “And that’s not true. They think they’re getting something they’re not. And even if the person had used PCP or a club drug, these are all short-acting drugs and they stay in the system such a short period of time.” MDMA may be gone in a day or so while GHB will be detectable for only a few hours—eight to 12 hours in urine, only four in blood—which is another reason why the test is not done in community hospitals. “Physicians have to realize if the drug wasn’t taken in the last few hours, a negative doesn’t necessarily mean they weren’t exposed to it or didn’t use it.” Blood levels of GHB do not correlate well with patient symptoms, Dr. Frederick notes, as there is a wide range of intra- and inter-patient response to a given dose.

However, new CAP proficiency testing in toxicology began this year, and by early 2013, “we’ll have a lot more data” on testing for these club drugs, he says.

There is an ongoing need to get better data and to increase understanding of all drugs-of-abuse testing, Dr. Melanson believes. “I think it’s important to be aware of the complexities and limitations and advantages of drugs-of-abuse testing, and the importance of the role of the pathologist in communicating those limitations to clinicians and helping them interpret results.”