September 2016—This fall, Brigham and Women’s Hospital, Boston, is taking the mass spec leap.
A plucky PR person might be tempted to refer to it as MassSpec LEAP!™ but Stacy Melanson, MD, PhD, doesn’t have time for such nonsense. As the associate director of clinical laboratories and co-director of chemistry, Dr. Melanson has more important matters to attend to.
She and her colleagues are shifting from using a screening immunoassay for pain management drug testing to up-front definitive testing by LC tandem mass spectrometry.
The decision is based, in part, on following the numbers. Seventy to 80 percent of patients are positive for one of the six common opioids—morphine, codeine, hydromorphone, hydrocodone, oxycodone, and oxymorphone. “Which makes sense—those drugs are the ones that are commonly prescribed to manage chronic pain,” says Dr. Melanson.
“But it didn’t make sense to screen for something that 80 percent of the time will require confirmatory or definitive testing,” Dr. Melanson adds, also noting that the opiates immunoassays have problems with both false-negatives (particularly for oxycodone and oxymorphone) and false-positive results. “So we made the decision to go right to definitive testing when we brought the testing in-house six years ago.”
Dr. Melanson and her colleagues also go straight to definitive testing for six benzodiazepines: temazepam, oxazepam, nordiazepam, alpha-hydroxyalprazolam, 7-aminoclonazepam, and lorazepam. It was a decision made several years ago. “We used to screen with an immunoassay and only confirm the positives,” she says. “But we published a few papers that have shown there’s a high rate of false-negatives, particularly for the newer benzodiazepines that are being prescribed, lorazepam and clonazepam. Thirty, 40 percent of our patients were screening negative by immunoassay but were actually positive with detectable levels of benzodiazepines by mass spectrometry” (Darragh A, et al. Pain Physician. 2014;17:359–366; Snyder ML, et al. Pain Physician. 2016. In press). This fall, Brigham and Women’s Hospital plans to eliminate all immunoassay testing for pain management and switch to definitive testing by LC tandem mass spectrometry for all drugs in its pain management panel, including fentanyl, buprenorphine, methadone, and illicit substances.
The limitations of immunoassays (see “Painstaking process of drug monitoring,” CAP TODAY, August 2016) may tilt other laboratories in the direction of mass spectrometry as well. “Confirmation testing is a vital part of pain management drug testing,” says Tai C. Kwong, PhD, professor of pathology and laboratory medicine and director of the hematology/chemistry laboratory, University of Rochester (NY) School of Medicine and Dentistry. Not only is mass spec useful in this regard, but it can be used directly, without initial screen by immunoassay, says Dr. Kwong.
A confirmation test should be based on an analytical principle that differs from that of the immunoreaction and be more specific than immunoassay (to eliminate false-positives) and of equal or greater sensitivity than immunoassay. Mass spectrometry hits the bull’s-eye for all three.
Mass spectrometry can confirm true positives and eliminate false-positives. The latter interferences are caused by drugs that are structurally unrelated to target drugs, such as ofloxacin and opiates assays, bupropion and amphetamines assays, and oxaprozin and benzodiazepines assays. Since assays from different manufacturers have different profiles of interfering substances, says Dr. Kwong, laboratories need to consult package inserts and technical support services.
A confirmatory test can also provide specific information about which drugs and metabolites are present, unlike the class immunoassays. This is important because a patient may have taken more than just the prescribed drug, says Dr. Kwong, who spoke about mass spectrometry with CAP TODAY as well as at an AACC virtual conference on drug monitoring for pain management. With an opiates immunoassay, he adds, the results will be the same regardless of the drug combination, whether morphine alone, morphine plus hydromorphone, or morphine plus hydrocodone plus hydromorphone. All would yield the same result: positive. “Without specific identification, you can appreciate how much important information is lost.”
Another advantage of mass spec versus immunoassay is its quantitative nature. “You can have very different interpretations based on concentrations,” Dr. Kwong says, for both positive and negative results. “Knowing the drug concentration is particularly helpful when a patient on an opiate also has another opiate detected.” Confirmatory mass spec can help physicians identify the source of an unexpected opiate as well. Is it a metabolite or contaminant of the prescribed drug? Is there nonprescription use?
If a patient is on MS Contin (morphine), for example, four different scenarios will result in the same positive opiates immunoassay result: 1) a morphine concentration of 500 ng/mL; 2) a morphine concentration of 50,000 ng/mL; 3) a morphine concentration of 50,000 ng/mL plus hydromorphone concentration of 500 ng/mL; and 4) a morphine concentration of 50,000 ng/mL plus hydromorphone concentration of 5,000 ng/mL.
Those same scenarios will be interpreted with much more nuance with mass spec quantitative testing. In the first case, the low concentration could be a result of the patient skipping doses of MS Contin or from ingesting poppy seeds. In the second case, the level suggests the patient could have been taking MS Contin as prescribed, while in the third case, testing has identified the unexpected presence of hydromorphone, but at such a low level relative to morphine concentration that it’s likely a metabolite of the morphine. In the last case, the much higher level of hydromorphone suggests the patient could have taken not only MS Contin but also an additional, unprescribed opiate, such as Dilaudid (hydromorphone).
Clearly, the uniform positive immunoassay result “is not very informative,” Dr. Kwong says. Without further digging, the real answers lie hidden, like money in an offshore account.
With its lower reporting cutoffs and better analytical sensitivity, mass spectrometry has better detection rates.
For example, says Dr. Kwong, citing one study, dropping the cutoff for cocaine metabolite from 300 ng/mL (for EIA) to 25 ng/mL (tandem MS) increased positives from 171 (2.3 percent) to 297 (4.1 percent). Put another way, 42.4 percent of samples detected by mass spec were missed by the immunoassay (Mikel C, et al. Ther Drug Monit. 2009;31:746–748). Results for the semisynthetic opiates are even more dramatic, Dr. Kwong says, since they are less reactive than morphine and require higher concentrations for a positive result. For hydromorphone, using the EIA cutoff (300 ng/mL) meant missing 69.3 percent of the positive samples detected by MS/MS (50 ng/mL cutoff).
Recently, notes Dr. Kwong, some laboratories have begun to use liquid chromatography time-of-flight (TOF) mass spec for broad-based drug screening. Since identification is not based on structural information, false-positives by TOF mass spectrometry are possible, he cautions, but confirmation can be done with fragmentation spectra by quadrupole-TOF (also known as Q-TOF) mass spectrometry.
Mass spec assay validation is fairly typical, resembling that of other clinical assays, with one important exception: “We have to do a careful evaluation of specimen matrix effects on ionization,” Dr. Kwong says. “It is not straightforward, but it’s very doable.” Published mass spec practice guidelines can help with this task.
Mass spectrometry may get a boost from National Academy of Clinical Biochemistry laboratory medicine practice guidelines for pain management testing, now in development. “These are the first evidence-based guidelines related to pain management testing co-developed by laboratory professionals and pain management providers,” says Dr. Melanson, who is a CAP representative to the NACB. “The recommendations include what testing to perform and how to effectively monitor patients. Laboratories can use these guidelines to develop and justify a testing program.”
One of the recommendations notes that definitive testing is the preferred method, she says. “That’s why we’re switching to mass spectrometry for all drugs and drug classes in our pain management panel at Brigham and Women’s.”
The mass spectrometry switch is not the only change laboratories might want to consider. Dr. Melanson notes that while most of the drug testing at Brigham and Women’s Hospital is done using urine samples (with the occasional use of serum testing), she’s had a few requests for oral fluid. “In clinics where adulteration is a concern—particularly because collections are not observed—then oral fluid might be a better option as collections are observed and therefore the risk of adulteration is lower.”
“Our colleagues at Massachusetts General Hospital have switched to oral fluid testing primarily because they manage more addiction management patients and are concerned about adulterated urine samples. We are collaborating with MGH to do a study directly comparing the clinical utility of oral fluid versus urine.”
Oral fluid may be a better alternative in certain settings—the detection window is closer to serum and therefore it may be more useful for the acute drug use, she says—but is not appropriate for all substances. “So it’s a balancing act.
“I don’t know where we’re going to end up—maybe a combination—or it may depend entirely on patient population,” Dr. Melanson continues. “If urine adulteration is a big concern, maybe oral fluid is the way to go, with some supplemental urine testing for drugs like benzodiazepines.” The proposed NACB guidelines recommend urine testing but acknowledge oral fluid can be useful as well, she says.
Oral fluid is an intriguing alternative test matrix, says Jennifer Collins, PhD, of MedTox Laboratories, St. Paul, Minn., who spoke about this option as part of the AACC virtual conference, noting that it may provide better insight into compliance with dosing protocols.
She began with some basics of saliva physiology, explaining that the colorless fluid, with an approximate pH of 6.7, is secreted into the oral cavity by the salivary glands (parotid, submandibular, and sublingual). Human saliva is 99 percent water; here, the “one percent,” so to speak, is made up of electrolytes, mucus, glycoproteins, enzymes, immunoglobulins, and biomarkers. Under normal conditions, a healthy adult produces between 0.5 to 1.5 L of saliva daily.
Does oral fluid differ from saliva? The former is sometimes called mixed saliva because it represents all the fluid in the mouth, including gingival crevices, mucus, and cellular debris as well as salivary gland secretions. “Even though the terms are often used interchangeably, ‘oral fluid’ better represents the analytical specimen,” she says.
Drugs appear in saliva shortly after dosing, mostly by passive diffusion down a concentration gradient from the blood. Oral fluid drug concentrations reflect the free, or unbound, fraction of the drug in plasma. Final drug concentrations are determined by physiochemical properties of the drug and saliva pH, she notes. Variables affecting this can include pH, the pKa of the drug, lipid solubility, charge, molecular weight, protein binding, and salivary flow rate (which can be affected by medications and chewing). Parent drugs often predominate in oral fluid, she says. Since they tend to be more lipid soluble, they move more easily through capillary cell membranes.
The partitioning of drug between oral fluid and plasma can be expressed as the saliva/plasma, or S/P ratio. A 1:1 ratio indicates equal drug concentration in both compartments. “Because oral fluid is slightly more acidic than blood, basic drugs”—including many pain management drugs—“ionize after diffusing into oral fluid and become ‘trapped,’ resulting in higher concentrations in the oral fluid than in plasma,” Dr. Collins says. The opposite is true for weakly acidic drugs, as well as for those that are highly protein bound in plasma.
“Opioid compounds prescribed for pain management, including oxycodone, hydrocodone, and tramadol, have high S/P ratios and are easily measured in oral fluid,” Dr. Collins says. “For example, while reported S/P ratios vary depending on route of administration, saliva pH, and other factors, studies have reported S/P ratios for codeine, hydrocodone, oxycodone, and tramadol in excess of 3.0.”
In general, she says, detection times of drugs in oral fluid mirror those in plasma and tend to be shorter than in urine. Detection periods vary based on how the drug is delivered. “Drugs that are smoked or administered sublingually, for example, may exhibit more extended detection periods due to contamination of the oral cavity.”
Dr. Collins notes that interest in using oral fluid in drug screening programs is growing, including in criminal justice programs and in nonregulated workplace drug testing programs, as well as in Driving Under the Influence of Drugs, or DRUID, law enforcement programs. It has also, not surprisingly, captured the interest of those involved in compliance programs, such as pain management.
One of the concerns in use of oral fluid as an alternative to urine is the more limited detection periods and lower drug concentrations. However, Dr. Collins shared data from several published studies that show comparable positive rates between oral fluid and urine populations, even though oral fluid detection periods are shorter on a per-dose basis. One data set was from the 2015 Drug Testing Index published by Quest Diagnostics, which contains results from its testing sites comparing percent positives for oral fluid versus urine. In the data from 2010 to 2014, the overall positive rates in oral fluid were similar to, if not higher than, those in urine. Other studies have shown similar results, she says.
In a 2012 study (Heltsley R, et al. J Anal Toxicol. 36:75–80) comparing oral fluid and urine specimens in chronic pain patients (133 patients, 1,544 paired tests), overall qualitative agreement was 85 percent, Dr. Collins reports, increasing to 89 percent when the comparison involved only HHS drug categories, including hydrocodone and oxycodone.
However, while the data show comparable positive rates in urine and oral fluid, Dr. Collins urges careful interpretation of results. Studies have used different collection and testing methods and devices, as well as different cutoffs. “And currently, equivalence between urine and oral fluid cutoffs has not been clearly established,” she cautions.
The most common collection method for oral fluid testing is with a device containing a collection pad and a transport tube with buffer solution. Buffers stabilize samples, Dr. Collins explains. They also aid with drug recovery from the collection pad and prevent bacterial growth during transport and storage.
Several studies have shown that collection methods and devices are the source of considerable variability in oral fluid testing, she continues. For example, specimen volume can vary depending on whether the device is qualitative or quantitative.
“If you are using a collection device that contains a buffer,” says Dr. Collins, “the actual analytical sample is oral fluid diluted between two and four times by the buffer. And if that’s the case, the assays that are used are designed to compensate for the dilution factor.” A 1-mL sample of oral fluid, for example, collected in a device containing 3 mL of buffer, results in a 4× dilution. “When you’re performing testing, matrix-matched calibrators and controls are generally prepared at the same dilution so that the final results are reported based on drug concentration in neat, or undiluted, oral fluid.”
“In general, the volume of oral fluid collected using these devices is relatively limited,” Dr. Collins says. “This is a particularly important consideration for applications that may have multiple positive results that have to be confirmed, as in pain management programs.” Most currently available quantitative devices collect 1 mL of oral fluid, so testing methods need to be optimized for small sample volumes and lower concentrations (such as mass spec).
Oral fluid collection protocols are a primary driver for the interest in oral fluid as an alternative specimen in drug testing programs, Dr. Collins says. Oral fluid collection is easy and fast and the methods are noninvasive. Most important, the collection is easily observable, which minimizes opportunities for specimen tampering or substitution.
But the disadvantages can’t be overlooked. Depending on the device, the collection volumes can be inconsistent (though most devices now being developed have volume indicators, she says). Drug recovery from collection pads may also vary by device. Medications that affect salivary secretion can extend collection times, and stimulated collections can change pH and resulting drug concentrations in oral fluid. In addition, oral specimens can be more expensive to collect based on added costs for the collection devices.
For screening, commercially available immunoassays include ELISA and homogeneous EIA formats that have been modified for improved cross-reactivity with the parent drug. While the number of FDA-cleared immunoassays is limited, Dr. Collins says multicomponent LC-MS/MS assays have also been developed for screening, similar to comprehensive urine drug screening used in some pain management programs.
For those who decide to perform oral fluid testing using immunoassay screening, “one of the issues you will have to address is the screening cutoff.” Currently available tests do not have standardized cutoffs; among the handful of FDA-cleared tests for opiates, for example, cutoffs were either 30 or 40 ng/mL, while for amphetamines the range was from 50 to 300 ng/mL.
Confirmation testing is generally performed by chromatography with mass spectrometry; HPLC with tandem mass spectrometry is the most prevalent method. These definitive testing procedures have been optimized for the target drugs with minimum detection levels in the low ng/mL range. The methods can be used for either screening or confirmation testing of oral fluid, Dr. Collins says. “There are numerous publications of validated methods for the detection of drugs relevant to pain management in oral fluid.”
Several studies have tried to correlate oral fluid with plasma concentrations for therapeutic or dosing studies. Results have been mixed, says Dr. Collins. Generally speaking, oral fluid results are temporally correlated with plasma; however, concentrations depend on lipophilicity and pKa of the drug.
Buccal cavity contamination from smoked, inhaled, or sublingually administered drugs can skew data, though concentrations generally equilibrate over time. “For compliance purposes,” Dr. Collins says, “numerous studies support the relative equivalence of oral fluid versus urine as the analytical sample.” And, she adds, parent drugs predominate in oral fluid, which may provide a better indication of recent use.
Despite the evolving nature of oral fluid testing, Dr. Collins makes it clear: “Oral fluid is a valid matrix for compliance testing in pain management programs.” For a laboratory that already provides toxicology services, she says, implementation can be fairly straightforward.
While acknowledging the technical challenges of pain management testing—whether by immunoassay or mass spec, screening or confirmatory/definitive, urine or oral fluid—those doing the testing emphasize it’s important not to lose sight of how test results echo through patients’ lives.
“Providers value the lab’s opinion,” says Dr. Melanson. “And they do want to give patients a chance.”
Matthew Krasowski, MD, PhD, agrees. At the University of Iowa Hospital and Clinics, Iowa City, where he’s a clinical professor and vice chair of clinical pathology and laboratory services, “A lot of providers often want to give the patient a second chance, especially if they have just one set of toxicology results that don’t make sense.”
For his part, Dr. Krasowski says he tries to stick closely to what the data reveal, though even that can be tricky at times. “You’ll have cases where the analysis shows no evidence of prescribed medications,” even though confirmatory testing using mass spectrometry often can detect low concentrations of drugs and metabolites. When patients are on pain-type doses, “we can inform the provider that it is unlikely the patient is taking the drug at the prescribed doses.”
“When we discuss findings with clinicians or document interpretations of toxicology analysis in the patient chart,” Dr. Krasowski continues, “we should be careful with wording and be aware that the results may have significant downstream consequences for patients.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.