The AACC Academy, formerly the National Academy of Clinical Biochemistry, has developed a laboratory medicine practice guideline (LMPG)10 for using laboratory tests to monitor drug therapy in pain management patients. The purpose of this guideline was to compile evidence-based recommendations for the use of laboratory and point-of-care (POC) urine drug tests for relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients. The exact process of preparing and publishing the LMPG is shown in Table 1.
Process of preparing and publishing a laboratory medicine practice guideline.
Briefly, a multidisciplinary LMPG committee was established to include clinical laboratory professionals, clinicians practicing in pain management, and other relevant stakeholders, healthcare professionals, and clinical experts. The experts on the committee are listed in the guideline and represented the AACC Academy (L.J. Langman, P.J. Jannetto); Clinical and Laboratory Standards Institute, which is jointly preparing an expert opinion guideline on laboratory testing for pain management (C.A. Hammett-Stabler, L.J. Langman, G.A. McMillin); College of American Pathologists (S.E. Melanson); Evidence-Based Laboratory Medicine Committee (W.A. Clark); clinical laboratories performing pain management testing (L.J. Langman, P.J. Jannetto, C.A. Hammett-Stabler, G.A. McMillin, S.E. Melanson); AACC (C.A. Kassed); American Academy of Pain Medicine (T.J. Lamer, R.J. Hamill-Ruth, N. Bratanow); active pain management clinicians (T.J. Lamer, R.J. Hamill-Ruth, N. Bratanow); and the National Institute of Drug Abuse (M.A. Huestis). Before a systematic literature search, the LMPG committee defined all the key questions that would be addressed in the guideline using the PICO(TS) strategy for construction of the questions. PICO(TS) stands for the (P)atient population, (I)ntervention, (C)omparator, (O)utcome, (T)ime period, and (S)etting. In this guideline, the patient population was acute and/or chronic pain management patients, and the interventions were the laboratory tests (screening or definitive) that were compared with other clinician tools (e.g., physician interview, medical record review, prescription monitoring programs, screener and opioid assessment for patients with pain). In general, screening tests have adequate clinical sensitivity but may not be highly specific. On the other hand, definitive or confirmatory testing (e.g., mass spectrometry- or chromatography-based) can identify a specific drug and/or its associated metabolites.
Outcomes included adherence, diversion, emergency department visits, and others. The period covered from January 2000 to February 2015 in outpatient, inpatient, and community settings. A systematic literature search was performed using the inclusion and exclusion criteria shown in Table 2.
Systematic literature search inclusion and exclusion criteria.
The following databases were searched: PubMed, the National Library of Medicine; Cochrane Database of Systematic Reviews, which includes the full text of regularly updated systematic reviews of the effects of healthcare prepared by the Cochrane Collaboration; the National Guideline Clearinghouse (an initiative of the Agency for Healthcare Research and Quality), a public resource for evidence-based clinical practice guidelines; EMBASE, which emphasizes drug-related literature and toxicology; CINAHL, which covers nursing and allied health disciplines and includes journal articles, healthcare books, nursing dissertations, selected conference proceedings, and standards of professional practice; SCOPUS; Web of Science; and Psych Info. The combined literature search from 2000 to 2015 resulted in 7647 articles being identified and reviewed by at least 2 committee members using the DistillerSR software to document the process. Of the 7647 abstracts reviewed, 2352 were selected for full text review. Committee members then assessed each article and documented the answers to 32 questions in the DistillerSR software, which covered everything from the authors' declarations, study aims, and objectives to their conclusions. The articles were again reviewed for appropriateness, and of the 2352 articles that had a full text review, 562 were ultimately used to formulate the recommendations for the guideline. The strengths of each recommendation were evaluated and graded using an approach described in the 2011 Institute of Medicine report. The approach was a modification of the US Preventive Services Task Force system. The strength of each recommendation was determined to be A, B, C, or I, and the grading of the quality of the evidence was designated I, II, or III (Table 3). Table 4 contains a summary of the evidence-based recommendations, and Table 5 contains a summary of the consensus-based expert opinions.
Strength and grading of the recommendations.
Summary of evidence-based LMPG recommendations.
Summary of consensus-based expert opinions.
BACKGROUND
The use of opioids for pain management has been broadly accepted by regulatory bodies, professional organizations, and clinicians. Compliance monitoring is viewed as necessary for safe opioid prescribing, and chronic opioid prescribing includes “contracts” or treatment agreements, periodic urine drug testing (UDT), and random pill counts. The magnitude of prescription opioid abuse has grown over the past decade, leading the CDC to classify prescription opioid analgesic abuse as an epidemic. This appears to be, in large part, because of individuals using a prescription drug nonmedically, most often an opioid analgesic. The number of drug-induced deaths has rapidly increased and continues to be one of the leading causes of death in Americans. In 2011, the Office of National Drug Control Policy established a multifaceted approach to address prescription drug abuse, including prescription drug monitoring programs that allow practitioners to determine whether patients are receiving prescriptions from multiple providers and the use of law enforcement to eliminate improper prescribing practices.
Over time, multiple guidelines from professional societies and organizations and regulatory bodies have evolved to include standard practices of assessing risk and documenting responsible care in a systematic way. In general, there is agreement that UDT is recommended before the initiation of treatment with opioids and during therapy. Federal regulatory agencies have developed guidelines and policies that support compliance testing. These include the Veterans Administration/Department of Defense Clinical Practice Guidelines for COT: Management of Opioid Therapy for Chronic Pain from May 2010 (http://www.healthquality.va.gov/guidelines/Pain/cot/COT_312_Full-er.pdf). Their recommendations include obtaining a UDT before initiating opioid therapy trial and randomly at follow-up visits to confirm the appropriate use of opioids. The CDC Guideline for Prescribing Opioids for Chronic Pain—United States, 2016 details the use of UDT (http://www.cdc.gov/mmwr/volumes/65/rr/pdfs/rr6501e1.pdf). The recommendations state that before prescribing opioids for chronic pain and periodically during patients' opioid therapy, clinicians should use UDT to assess for prescribed opioids, as well as other controlled substances and illicit drugs that increase risk for overdose when combined with opioids, including nonprescribed opioids, benzodiazepines, and heroin.
Forty-seven states and the District of Columbia also have policies regarding pain management and proper prescribing. In addition, specialty boards have developed guidelines for proper opioid prescribing. The American Academy of Family Practice developed recommendations in 2012, Rational Use of Opioids for Management of Chronic Nonterminal Pain (http://www.aafp.org/afp/2012/0801/p252.html), with recommendations for UDT pretreatment and randomly during treatment. The American Pain Society and American Academy of Pain Medicine also teamed up to develop the landmark APS/AAPM 2009 Guidelines (http://americanpainsociety.org/uploads/education/guidelines/chronic-opioid-therapy-cncp.pdf), which include examination of various aspects of UDT and recommend pretreatment and concurrent monitoring of patients. The American Society of Addiction Medicine released a detailed review of UDT with Drug Testing: A White Paper of the American Society of Addiction Medicine dated October 26, 2013 (http://www.asam.org/docs/default-source/public-policy-statements/drug-testing-a-white-paper-by-asam.pdf). They reviewed the science and practice of drug testing. It explored the wide range of applications for drug testing and its utility in a variety of medical and nonmedical settings. It promoted the use of drug testing as a primary prevention, diagnostic, and monitoring tool in the management of addiction or drug misuse in medical practice.
Although UDT is currently regarded as the standard for adherence monitoring of patients taking controlled substances to manage chronic pain, UDT results are performed/read and interpreted by distinctly different sets of individuals. One group is clinical laboratory physicians and scientists; another group is the clinical providers: clinicians, nurses, pharmacists, and others directly involved in the patient's care. Others may have reason to access or review such data from time to time, such as those in legal or law enforcement, policy, and insurance. Correctly interpreting test results requires that these individuals have the knowledge and experience needed for accurate interpretation, and the skill levels vary considerably within and between each group (1). In the end, the goal of this LMPG guideline for pain management was to address many of these issues and challenges described above and to provide evidence-based recommendations for clinical laboratorians and practicing pain management clinicians.
EVIDENCE-BASED RECOMMENDATIONS/STATEMENTS
The goal of developing specific testing recommendations is to balance the completeness and accuracy of test results with the cost of the testing paradigm. It is critical that a valid specimen is obtained and enough substances evaluated to determine appropriate adherence with the treatment regimen. The testing must also be able to identify polysubstance use, abuse, addiction, and possible diversion before the patient (or recipient of diverted medications) experiences a significant adverse event. Lastly, it is also important to note that studies continue to demonstrate that the administered dosage does not necessarily correlate with the concentration of the drug in an individual's urine.
Evidence-based recommendation 1
Testing biological specimens for drugs and drug metabolites is recommended and effective for detecting the use of relevant over-the-counter products, prescribed and nonprescribed drugs, and illicit substances in pain management patients. Laboratory testing does not specifically identify most other outcomes but should be used in conjunction with additional information to detect other outcomes in pain management patients.
Strength of Recommendation: A
Quality of Evidence: I
Numerous studies looked at outcomes including adherence to the prescribed regimen along with detection of illicit drug use with laboratory drug testing as the tool. Although most reports were looking at urine, other matrices, such as plasma and oral fluid, have also been evaluated and showed some efficacy (2–5).
One other point to consider is the breadth of laboratory testing. Table 6 shows the 3 main tiers of drugs and drug classes that are being recommended to test in pain management patients based on risk. It should be noted that this table is not meant to be a comprehensive list of all drugs that need to be tested for in every pain management patient, but instead should be used as a guideline. Tier I represents the scope of testing that should be done as part of routine monitoring and covers the common classes of drugs of abuse, as well as the drugs commonly prescribed to pain management patients. Tier II testing should also be added to screen for drug use and abuse in patients identified as high risk by the treating clinicians. These could include patients with a known history of abuse for medications in this category. However, it may also include drugs of which the prevalence of use and abuse is endemic to local region. In addition, it applies to patients who have polypharmacy that puts them at an increased risk of adverse drug reactions, or to detect patients with multiple providers. Furthermore, it may also apply to patients who experience a lack of efficacy for 1 of these drugs or who may be experiencing toxicity from them. Tier III tests can also be examined when they are clinically indicated, either by history of use, medication list, or high probability of misuse and abuse, in a specific patient rather than for every patient.
Tiers of drug testing.
FREQUENCY OF LABORATORY TESTING
Consensus-based expert opinion 1
Based on level II evidence, baseline drug testing should be performed before initiation of short- or long-term controlled substance therapy. In addition, random drug testing should be performed at a minimum of 1 or 2 times a year for low-risk patients (based on history of past substance abuse and addiction, aberrant behaviors, and opioid risk screening criteria), with increasing frequency for higher-risk patients prescribed controlled substances.
Strength of Recommendation: A
Quality of Evidence: II
Evidence-based recommendation 2
More frequent laboratory testing is recommended for patients with a personal or family history of substance abuse, mental illness, evidence of aberrant behavior, or other high-risk characteristics.
Strength of Recommendation: A
Quality of Evidence: II
The evidence for specific schedules of drug testing in general is weak, mainly because of the lack of randomized clinical trials comparing the effectiveness of testing schedules or methods specifically in the chronic pain population. Existing practice guidelines make recommendations based on observational studies or expert consensus opinion (4). Existing clinical practice guidelines recommend testing at baseline and randomly, but at minimum annually for low-risk patients (American College of Occupational and Environmental Medicine, American Pain Society and American Academy of Pain Medicine, American Society of Interventional Pain Physicians, University of Michigan Health System, Veterans Administration/Department of Defense). However, for patients with risk factors for misuse and abuse, more frequent monitoring is recommended, but the optimal frequency for these patients has not been determined (3).
LABORATORY TESTING AND ITS ABILITY TO IDENTIFY NONCOMPLIANCE IN PAIN MANAGEMENT REGIMENS
Evidence-based recommendation 3
Laboratory testing is recommended to identify the use of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients. However, it does not effectively identify all noncompliance with the prescribed regimen. No single monitoring approach provides adequate information about the pattern or dose of patient drug use. Safest prescribing habits should include a combination of tools and laboratory test results to correctly detect outcomes.
Strength of Recommendation: A
Quality of Evidence: III (pain management population), II (substance abuse disorder monitoring population)
Studying patient noncompliance with the therapeutic regimen is difficult unless nonprescribed medications or illicit drugs are present in the tested matrix. Generally, testing frequency is low, and the windows of detection in the different matrices (urine, oral fluid, blood/plasma/serum) are usually only a few days. Thus, most of the time between biological testing, the patient is inadequately monitored. Even when the matrix has a longer window of detection, such as for hair, minimum exposure is required to give a positive result, and differences in disposition can occur based on hair color for basic drugs; for meconium, minimum exposure frequency is needed to produce positive test results. Therefore, additional means of monitoring are highly useful to improve the detection of noncompliance, such as pill counts and interviews. Additional research studies are needed in which the collection of other physician tool data (e.g., self-report, pill counts) are directly compared with biological testing data.
LABORATORY TESTING VS OTHER PHYSICIAN TOOLS, PRESCRIPTION MONITORING, AND SELF-REPORT
Evidence-based recommendation 4
Laboratory testing is more effective than other physician tools for the detection of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients and should be used routinely to monitor compliance.
Strength of Recommendation: A
Quality of Evidence: II
Most controlled administration studies of prescription and over-the-counter drugs examined urine, blood, or serum concentrations, providing a scientific database for using these biological fluids in monitoring programs (6, 7). Urine has been the matrix of choice for monitoring pain patients, but other matrices are now being used more frequently (5, 8, 9). In addition, UDT is more effective than self-reporting at revealing recent opioid use (10).
SPECIMEN TYPES
Urine is typically the preferred matrix for pain management drug testing, as it has a longer window of drug detection than blood and an adequate specimen volume for drug screening and confirmation, and drug markers (either parent drug or metabolites) are present in high concentrations. It is also less invasive and does not require a phlebotomist for collection. Disadvantages include a high risk of adulteration of the sample by the patient to avoid detection of noncompliance with the therapeutic regimen. Observed specimen collection is generally not performed and is disliked by patients and collectors. Specialized bathroom facilities may be needed, and specimen collectors should be of the same sex as patients. For these reasons, there is much interest in alternative matrices such as oral fluid or hair for drug testing of pain management patients.
Evidence-based recommendation 5
Urine testing is recommended for the detection of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients.
Strength of Recommendation: B
Quality of Evidence: II
Alternative matrices such as oral fluid, blood/plasma/serum, hair, meconium, and umbilical cord show promise and offer advantages over urine for testing, but the evidence to date is insufficient to assess whether the results are equivalent to urine testing for monitoring patient compliance. Other matrices may also be appropriate in specialized circumstances, but the samples must be properly collected, stored, and transported in the appropriate collection device at the proper temperature, and tested by qualified personnel using a validated method for that matrix.
Consensus-based expert opinion 2
Serum or plasma is an acceptable alternate matrix for the detection of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients with end-stage renal failure (anuria). For dialysis patients, the blood (serum/plasma) should be collected before dialysis. Oral fluid testing can also be used for selected drugs (e.g., amphetamine, benzodiazepines, buprenorphine, tetrahydrocannabinol, cocaine, codeine, hydrocodone, hydromorphone, methadone, morphine, oxycodone, and oxymorphone).
Strength of Recommendation: A
Quality of Evidence: III
As discussed above, blood/plasma/serum are good matrices for biological monitoring of patient compliance in pain management testing; however, no manuscripts were found that specifically detailed the use of these matrices during end-stage renal failure.
Alternative matrices such as oral fluid show promise and have advantages over urine or blood, but the evidence to date is insufficient to assess their benefits in predicting clinical outcomes. Heltsley et al. (11) examined the screening positivity rates for oral fluid in a chronic pain population and compared them with published positivity rates for urine drug screening in the pain population and found that the oral fluid nonnegative screening rate was 83.9% compared with a previously published nonnegative rate of 78% for urine screening. Within those overall positives, they found that 11.5% of the screening positives in oral fluid were for illicit drugs, compared with 10.9% of the urine screening positives from a previous urine study. The authors concluded that oral fluid screening is comparable with urine screening for detecting illicit drug use in a pain management population. In a follow-up study from the same group (12), the authors examined paired oral fluid and urine specimens from a chronic pain population (n = 133). On screening of both specimens for each patient, they found 21.3% of specimens positive in both matrices and 63.7% negative in both matrices, for an overall agreement rate of 85%. Of the 15% that disagreed, 5.4% were positive in oral fluid and negative in urine, and 9.6% were negative in oral fluid and positive in urine. The authors concluded that the Cohen κ statistical test for agreement between the 2 methods was 0.64, documenting substantial agreement, and that the oral fluid screening results were comparable with urine screening results.
In conclusion, although there are some studies that describe the utility of alternate specimens for drug testing in certain populations, there is no evidence that drug testing in alternate matrix specimens is more effective than urine testing for detection of drugs in pain management patients. In the absence of evidence, the committee cannot make a recommendation for or against alternate matrix testing in pain management.
QUALITATIVE AND SEMIQUANTITATIVE SCREENING ASSAYS
Traditionally, UDT for pain management patients has followed a forensic (legal) model and has been based on Department of Health and Human Services guidelines and protocols for drugs-of-abuse testing. As such, immunoassays are typically used as the first-line screening test. These immunoassays can either be run in a qualitative (e.g., positive/negative) or semiquantitative mode. Laboratories often use these assays in the semiquantitative format to assist the laboratory in setting dilutions on concentrated samples up front before downstream confirmatory (e.g., mass spectrometry-based) testing is performed to minimize carryover and avoid repeat testing. Although immunoassays offer several advantages, including ease of use, fast turnaround time, noninvasive collection, and lower costs, they can produce false-positive and false-negative results (5). In a forensic model, positive immunoassay screening tests are followed by a definitive or confirmatory test, such as mass spectrometry, to avoid false-positive results. False-negative results, however, remain problematic with this approach. Furthermore, the Food and Drug Administration-approved immunoassays originally designated by the Mandatory Guidelines for Federal Workplace Drug Testing Programs commonly use higher cutoffs. These cutoffs may not be clinically appropriate for adherence monitoring of pain management patients. For these reasons, modifications to the forensic model of testing in which laboratories use orthogonal testing (e.g., immunoassay screen followed by an LC-MS/MS confirmation assay) to monitor compliance in pain management are necessary.
Evidence-based recommendation 6
Although definitive testing is recommended and preferred, urine immunoassays performed on laboratory-based analyzers offer some clinical utility to detect the use of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients. However, physicians using immunoassay-based tests (especially amphetamine, benzodiazepine, and opiate immunoassays) must reference the package insert if testing in the physician's office or consult with laboratory personnel to evaluate the assay's capabilities and limitations for detecting specific medications within a drug class to prevent incorrect interpretation and to determine when additional testing is necessary.
Strength of Recommendation: B
Quality of Evidence: II
Numerous articles have compared the accuracy of immunoassays with mass spectrometry-based assays. However, many articles do not include pain management patients or specifically correlate results with outcomes. Overall, laboratory-based immunoassays across several populations (e.g., pain management, addiction patients) have been shown to correlate to mass spectrometry-based testing and can be used to detect compliance and adherence to therapy and misuse or abuse of other drugs.
QUALITATIVE DEFINITIVE TESTING
Immunoassays, as described above, have known limitations. Mass spectrometry-based assays have traditionally been considered the gold standard, despite the prevalence and ease of use of laboratory-based immunoassays. Furthermore, many qualitative immunoassays are designed to detect only a class of compounds. Therefore, a positive immunoassay result does not indicate which drug(s) in the class were present in the urine, whereas a definitive result by mass spectrometry provides this information. The specific drugs in urine can help determine compliance, as well as the potential abuse of multiple drugs within a class.
Evidence-based recommendation 7
Qualitative definitive tests should be used over immunoassays because they are more effective at identifying relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients.
Strength of Recommendation: A
Quality of Evidence: II
Evidence-based recommendation 8
Qualitative definitive tests should be used when possible over immunoassays for monitoring use (compliance) to relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients because of their superior sensitivity and specificity.
Strength of Recommendation: A
Quality of Evidence: II
Several articles provide evidence that qualitative definitive assays such as GC-MS and LC-MS/MS are more sensitive and specific than laboratory-based immunoassays. One may infer, therefore, that these assays are superior at detecting adherence and compliance with or diversion or misuse of various drugs and drug classes in pain management. However, none of the studies examined any patient outcomes directly. All articles demonstrate that LC-MS/MS and GC-MS are technically superior to laboratory-based immunoassays. Many of the articles state that targeted screening assays should be used for definitive results and testing with legal implications. However, despite the lack of outcome data, most of the authors conclude that immunoassays are clinically acceptable and should be used to facilitate real-time clinical decisions.
POC TESTING
Urine or oral screening immunoassays are also available at the POC. POC testing can be done in the pain clinic or physician's office using single-use dipstick or cup-based technologies and can provide immediate results for the provider and patient. Negative results are typically used to rule out drug abuse. Positive samples are usually sent for definitive laboratory-based testing to identify the drug(s) present and to determine adherence or identify abuse or diversion. POC immunoassays, similar to laboratory-based screening immunoassays, have lower sensitivity and specificity than definitive assays. In addition, quality control, quality assurance, and result documentation are challenging with POC testing. It should also be mentioned that most, but not all, POC assays indicate a negative test with the presence of a line and a positive test by the absence of a line.
Evidence-based recommendation 9
POC (oral and urine) qualitative presumptive immunoassays offer similar performance characteristics to laboratory-based immunoassays and can detect some over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients. However, physicians using POC testing must reference the POC package insert and/or consult laboratory personnel to accurately determine the assay's capabilities (especially amphetamine, benzodiazepine, and opiate immunoassays) and understand the limitations for detecting specific medications within a drug class to prevent incorrect assumptions or interpretation and to determine when additional testing is necessary.
Strength of Recommendation: B
Quality of Evidence: II
Note: POC devices must be performed exactly according to the manufacturer's instructions. Any deviation from this can significantly alter the POC device's ability to operate correctly and may affect the interpretation of the test result. Lastly, it should also be noted that most devices require additional confirmatory testing, especially when unexpected results are observed.
TIMING OF UDT
Although guidelines recommend UDT as 1 tool to monitor compliance in pain management, the existing guidelines do not recommend how frequently patients should be tested, if baseline testing is indicated, or whether testing should be random or scheduled. This information is critical for both providers and the laboratory to successfully treat patients and predict resource use.
Evidence-based recommendation 10
Qualitative immunoassay drug testing before prescribing controlled substances can be used to identify some illicit drug use and decrease adverse outcomes in pain management patients.
Strength of Recommendation: B
Quality of Evidence: II
Consensus-based expert opinion 3
Random urine testing for relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances is recommended to detect outcomes in pain management patients.
Strength of Recommendation: A
Quality of Evidence: III (pain management population), II (substance abuse disorder monitoring population).
Although laboratory testing once yearly for low-risk patients and twice yearly for higher-risk patients has been recommended, the same recommendations call for POC screening every 6 months for low-risk patients and every 3 months for higher-risk patients (13), which is far more frequent than the previously cited American Society of Interventional Pain Physicians' recommendations (14, 15). Therefore, because of a lack of scientific evidence to suggest that random testing is superior to scheduled testing, the committee recommends random drug testing to better assess compliance and outcomes. If testing is scheduled, patients have an opportunity to adulterate their specimen before or during the visit. Furthermore, patients who know the date of testing may adhere to their prescribed medication(s) immediately before their visit, only to continue abuse or diversion when testing is not scheduled.
COST-EFFECTIVENESS OF UDT
Cost is a concern in all areas of healthcare but particularly with laboratory testing. Providers and laboratorians are under pressure to provide the same level of patient care at a lower cost. Therefore, there is interest in whether qualitative screening immunoassays, either in the laboratory or at the POC, are more cost-effective than mass spectrometry-based assays. Any cost benefits need to be weighed with the clinical benefits and sensitivity and specificity of the most cost-effective testing options.
There is no evidence to suggest that qualitative or semiquantitative urine screening assays are more cost-effective than mass spectrometry-based assays in detecting outcomes in pain management patients. Additional studies are needed.
Evidence-based recommendation 11
Appropriately performed and interpreted urine POC immunoassay testing can be cost-effective for detecting use or inappropriate use of some over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients.
Strength of Recommendation: B
Quality of Evidence: II
There is a lack of evidence to suggest that laboratory-based qualitative or semiquantitative urine screening assays are more cost-effective than mass spectrometry-based assays in detecting outcomes in pain management patients. However, Manchikanti et al. (16, 17) wrote 2 articles that concluded that appropriate use of urine drug screening assays at POC is more cost-effective than LC-MS/MS. The authors report a cost per test of $25 for immunoassay and a cost per test of $600 for mass spectrometry and advocate for a testing algorithm to reduce costly LC-MS/MS use. According to the authors' testing algorithm, mass spectrometry-based assays should be performed only in patients who test negative when prescribed a drug, in patents who test positive when not prescribed the drug, or in patients who test positive for an illicit drug. In the latter 2 scenarios, mass spectrometry-based assays should not be confirmed if the patient admits adherent use. Instead, repeat testing should be performed by POC immunoassay at the patient's next visit or at a random time. As stated earlier in the POC section, it is important that providers understand the limitations of POC assays and consult the laboratory if appropriate so that the lower cost is not compromising patient care, leading to incorrect interpretations.
DEFINITIVE TESTING
Evidence-based recommendation 12
First-line definitive testing (qualitative or quantitative) is recommended for detecting the use of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients.
Strength of Recommendation: A
Quality of Evidence: II
A study by Pesce et al. (18) evaluated the diagnostic accuracy of LC-MS/MS vs immunoassay for drug testing in pain patients. In this study, the authors tested 4200 urine specimens from pain patients for amphetamine, methamphetamine, α-hydroxyalprazolam, lorazepam, nordiazepam, oxazepam, temazepam, cannabinoids, cocaine, methadone, methadone metabolite, codeine, hydrocodone, hydromorphone, morphine, propoxyphene, and norpropoxyphene. The authors compared the immunoassay results with the LC-MS/MS results. Using the drug and metabolites to define a positive result by LC-MS/MS, the authors found the following false-negative results in urine by immunoassay: 9.3% for amphetamines, 22% for benzodiazepines, 10.6% for cannabinoids, 50% for cocaine, 6.1% for methadone, 1.9% for opiates, and 23.4% for propoxyphene. The authors attribute the differences to variance in cross-reactivity for immunoassays, along with lower cutoffs for the LC-MS/MS methods. The authors concluded that the use of LC-MS/MS significantly reduces the risk of false-negative results. Implicit in this study is that the LC-MS results are of higher quality compared with immunoassay results, as it is designated as the gold standard. There is no discussion of the impact of immunoassay or LC-MS methods of measurement on detection of outcomes.
Evidence-based recommendation 13
Recommend definitive testing for any immunoassay (laboratory-based or POC) result that is not consistent with the clinical expectations in a pain management patient.
Strength of Recommendation: A
Quality of Evidence: III
Manchikanti et al. (16) presented a comparative evaluation of a POC immunoassay kit vs LC-MS/MS for detection of UDT opioids and illicit drugs in the urine of pain management patients. In this study, the authors analyze 1000 consecutive urine specimens submitted for analysis. The immunoassay was performed first, followed by LC-MS/MS analysis at a reference laboratory—the LC-MS/MS test was designated as the reference method. Agreement for prescribed opioids was high with the index test (80.4%). The reference test of opioids improved the accuracy from 80.4% to 89.3%. Nonprescribed opioids were used by 5.3% of patients. The index test provided false-positive results for nonopioid use in 44%, or 83 of 120 patients. For illicit drugs, the false-positive rate was 0% for cocaine, 2% for marijuana, 0.9% for amphetamines, and 1.2% for methamphetamines. Overall, the authors suggest that confirmation was required in 32.9% of the samples. They state that POC immunoassay is sufficient for front-line UDT in pain management and suggest that all samples negative for prescribed opiates, positive for nonprescribed opiates, and positive for illicit drugs should be sent for confirmatory testing. There is no discussion of the impact of this testing paradigm on clinical outcomes for pain management patients. Manchikanti et al. (17) also presented data from the same study but focused on the detection of benzodiazepines. They drew the same conclusion for benzodiazepines that they published for opiates and illicit drugs.
QUANTIFICATION VS QUALITATIVE DEFINITIVE TESTS
Evidence-based recommendation 14
Quantitative definitive urine testing is not more useful at detecting outcomes in pain management patients compared with qualitative definitive urine testing. Furthermore, quantitative definitive urine testing should not be used to evaluate dosage of administered drug or adherence to prescribed dosage regimen. However, quantitative urine definitive testing is recommended to identify variant drug metabolism, detect pharmaceutical impurities, or metabolism through minor routes. Quantitative results may also be useful in complex cases to determine the use of multiple opioids, confirm spiked samples, and/or rule out other sources of exposure (e.g., morphine from poppy seeds).
Strength of Recommendations: A
Quality of Evidence: II
A number of different studies by Couto et al. (19–20) assessed the ability of an algorithm applied to urine drug levels of oxycodone or hydrocodone in healthy adult volunteers to differentiate among low, medium, and high doses. The authors concluded that the algorithm normalized urine drug levels for pH, specific gravity, and lean body mass and could differentiate between the different daily doses of oxycodone or hydrocodone. However, there are several important limitations to both studies. The study patients were relatively homogenous with respect to cytochrome P450 2D6—poor, intermediate, and ultrarapid metabolizers were excluded from the study. In addition, they were restricted from any medications or items in their diet that could inhibit or induce the CYP2D6 enzymes. Lastly, a careful observation of the data demonstrates significant overlap between the distributions. Although the medians may be statistically differentiated between the groups, a comparison of an individual result with a population distribution would not likely be able to place the patient in 1 particular group or another.
DETECTION LIMITS
The evidence in the literature is currently insufficient to determine standardized cutoffs or limit of quantifications to determine full compliance, partial compliance, or misuse/abuse of controlled drugs by pain management patients.
Consensus-based expert opinion 4
The use of lower limit-of-detection cutoff concentrations can be more effective to detect use (either partial or full compliance) or the lack of use of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients, especially those taking lower dosages.
Strength of Recommendation: B
Quality of Evidence: II
Crews et al. (21) demonstrated the use of LC-MS/MS to detect 6-acetylmorphine (6-AM) in the absence of morphine in pain management patients. In this study, the authors analyzed 22361 urine specimens from chronic pain patients. From these specimens, 30 tested positive for 6-AM above a cutoff of 10 ng/mL, and 23% of those had a morphine concentration less than the cutoff of 300 ng/mL. The authors suggest that using a standard screening cutoff of 300 ng/mL for morphine as a threshold for confirmation (including 6-AM) will result in a missed diagnosis of heroin use in approximately 25% of the cases. It is important to note that there is no discussion of the impact of this confirmatory testing on clinical outcomes.
A study by West et al. (22) examined the comparison of clonazepam compliance as measured by immunoassay and LC-MS/MS in a pain management population. In this study, the authors selected samples from their database prescribed clonazepam only, while eliminating any patients who were prescribed a second (or more) benzodiazepine drug. From this selection, 180 urine specimens were found that met the criteria and were analyzed using an immunoassay with a cutoff concentration of 200 ng/mL and then analyzed with an LC-MS/MS method using cutoffs of both 200 ng/mL and 40 ng/mL that detected both clonazepam and the primary metabolite 7-aminoclonazepam. The positivity rate for the immunoassay was 21%, and the positivity rates for the LC-MS/MS method were 70% and 87% for the 200 ng/mL and 40 ng/L cutoffs, respectively. The authors attributed the differences in positivity rates to the lack of cross-reactivity of the immunoassay with the clonazepam metabolite. They suggest that a much lower cutoff (e.g., 40 ng/mL) is needed to reliably monitor clonazepam adherence. There was no discussion of the impact of using either the immunoassay or LC-MS/MS assay on clinical outcomes in pain management patients.
PREANALYTICAL HYDROLYSIS (ENZYMATIC/CHEMICAL) OF URINE
The evidence in the literature is inconsistent to support routine use of hydrolysis for all drug classes to more effectively detect outcomes in pain management patients.
Consensus-based expert opinion 5
Recommend clinicians and/or referring laboratories consult with the testing laboratory personnel about the use and efficiency of preanalytical hydrolysis for UDTs, as well as the expected impact on results.
Strength of Recommendation: I (insufficient)
Quality of Evidence: III
Preanalytical hydrolysis is commonly used to liberate glucuronide and sulfate conjugate metabolites of drug analytes in mass spectrometric methods such as GC-MS and LC-MS/MS. This practice is common for urine because many drugs are eliminated in a conjugated form. The consequence of preanalytical hydrolysis is to increase the concentrations of drug analytes and thereby increase the sensitivity of an assay for the associated drug analytes. Hydrolysis reactions can be enzymatic or chemical. Enzymes used include β-glucuronidase from abalone, β-glucuronidase type H-3 from Helix pomatia, β-glucuronidase type L-II from Patella vulgata, and glusulase (23, 24). Recombinant β-glucuronidase is also now available (IMCSzyme from IMCS). A common approach to chemical hydrolysis includes incubation with concentrated hydrochloric acid. Hydrolysis conditions, such as substrate concentrations, temperature, pH, and time, should be evaluated and optimized by the laboratory. One study comparing 3 methods of hydrolysis (2 enzymes and 6 N HCl) with nonhydrolyzed recoveries on efficiency of tapentadol recovery demonstrated different yields for each method (23). The chemical hydrolysis method was preferred over the enzymatic methods because of better compatibility with the associated liquid chromatography columns. As such, chromatography quality and consistency were superior to the enzymatic hydrolysis products.
As suggested above, the efficiency of hydrolysis reactions may be incomplete, despite optimization of conditions. For example, a study using β-glucuronidase demonstrated that between 17% and 27% of morphine-3-glucuronide was not cleaved. Similarly, between 32% and 45% of morphine-6-glucuronide was not cleaved (25). When comparing hydrolyzed and unhydrolyzed urine samples collected from pain management patients prescribed tramadol, no qualitative differences in detection were observed. This study suggests that qualitative drug testing can be performed with unhydrolyzed urine, and that doing so considerably reduces matrix interferences in mass spectrometric methods (26). Unconjugated tapentadol (cutoff, 50 ng/mL) and the N-desmethyltapentadol metabolite (cutoff, 100 ng/mL) were detected when urine was unhydrolyzed. Only 1 of 8 patient samples evaluated required hydrolysis for detection. However, concentrations of tapentadol and metabolite were significantly increased after hydrolysis. It was estimated that the average amount of tapentadol conjugated is 65%, and the metabolite is approximately 20% conjugated (23). However, the inclusion of a known concentration of conjugated metabolites should be included as quality control material to ensure stability and consistency of hydrolysis efficiency.
USE OF CONJUGATED AND UNCONJUGATED DRUG METABOLITES
The evidence in the literature is currently insufficient to make any recommendations at this time regarding the use or superiority of conjugated vs unconjugated drug metabolites in definitive tests for pain management patients.
Consensus-based expert opinion 6
Laboratories ultimately need to measure the appropriate analytes based on the matrix (e.g., serum vs urine). In urine, the conjugated form is most prevalent, and it can either be measured separately or combined with the less abundant unconjugated form after hydrolysis.
Strength of Recommendation: I (insufficient)
Quality of Evidence: III
Direct measurement of glucuronide or other conjugated metabolites will improve detection of drug use with or without use of preanalytical hydrolysis. This approach also overcomes the variation in efficiency of hydrolysis reactions. One study demonstrated that detection of morphine-3-glucuronide, morphine-6-glucuronide, oxymorphone glucuronide, hydromorphone glucuronide, and norbuprenorphine glucuronide significantly increased detection of the associated drugs when evaluating medication adherence in pain management patients. Between 10% and 100% of samples would have been misclassified if glucuronide metabolites were not included (27). The interpretive value of quantitative analysis of conjugated and unconjugated drug metabolites depends on the efficiency of hydrolysis and the cutoff concentration used for detection. Ratios of conjugated metabolites may provide phenotype information, although this finding is controversial (28).
ADULTERANT/SPECIMEN VALIDITY TESTING
For drug testing results to be used appropriately in clinical decision-making, the results must be valid. The goal of drug testing in the pain management population is not only to confirm compliance with appropriate use of prescribed medications but also to identify aberrant behaviors and the risk of adverse outcomes. Noncompliance can include binging, use of nonprescribed and/or nonreported medications and illicit substances, as well as diversion. Press and political attention often focus on overdose deaths, but diversion is also another significant public health issue that contributes indirectly to the overdose statistics.
The ease of urine sample adulteration makes it critical to address the method of collection. In an ideal world, all urine sample collections would be observed, although attempts have been made to foil this approach as well (http://realwhizzinatorxxx.com). Data regarding these more stringent standards for specimen collection can be found in the addiction and occupational screening literature, but no references were found for the pain management population. This method is time-consuming, expensive because of staffing requirements, and often not possible in a busy practice. Alternatives include specialized collection facilities in which the water can be turned off and the toilet water is colored. The risk of an invalid specimen increases as the level of supervision decreases. In addition, announced UDT or testing performed at an off-site laboratory provides the opportunity not only for planned adulteration or urine specimen substitution but also for the patient to take enough medication to have an appropriate test result. This fails to identify potential binging or diversion. Finally, time from request for a urine specimen to time of actual void can affect results. Although review articles may make recommendations for specimen collection methods for pain patients, these guidelines are extrapolated from the addiction literature. Diuretics and excessive fluid intake provide a delayed effect on urine content, which can take ≥1 hour to be seen, so some guidelines go as far as to suggest a 20-min window during which the specimen should be provided. Because of these issues, alternative matrices like oral fluid are proposed as another alternative to get a valid “witnessed” sample.
SPECIMEN VALIDITY TESTING
Evidence-based recommendation 15
Specimen validity testing (e.g., pH, temperature) is recommended because it is an effective tool to ensure outcomes (e.g., use of relevant over-the-counter, prescribed, and nonprescribed drugs) are correctly interpreted in pain management patients. Specimen validity testing determines the suitability of the urine specimen collected and received, which directly affects the ability to correctly identify relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances used by pain management patients.
Strength of Recommendation: A
Quality of Evidence: I (workplace drug testing), II (pain management population)
Evidence-based recommendation 16
For urine specimens, the pH and temperature should be measured within 5 min at the point of collection and be used to determine whether testing should be performed on that sample. In addition, the determination of creatinine and other adulteration tests (e.g., oxidants) should be performed on the urine specimen in the laboratory using federal workplace drug testing cutoffs. In the end, if any of the specimen validity tests fall outside the range of physiological urine values and acceptance criteria, the adulterated sample must not undergo further testing, and the patient should be further evaluated for aberrant drug-taking behavior.
Strength of Recommendation: A
Quality of Evidence: I (workplace drug testing population), III (pain management population)
Evidence-based recommendation 17
Clinicians should consult the laboratory regarding proper collection, storage, and transportation of urine specimens to maintain specimen validity.
Strength of Recommendation: A
Quality of Evidence: III
In an evidence-based analysis looking at methadone compliance testing by the Ontario Medical Advisory Secretariat (29), urine temperature of 32.5 °C to 37.7 °C was shown to be a good indicator that a specimen was just provided by the identified donor. However, it was noted that this specimen validity method could potentially be circumvented by warming substituted urine specimens. As a result, volume collection could be used to increase the validity of temperature readings and ensure specimen validity from the donor. In addition, laboratory analysis of the urine's pH and creatinine could offer enhanced reliability of test result. The absence of drug detected in a concentrated urine specimen was found to be more reliable in terms of nonuse than a negative test result in a diluted sample. pH, in a similar fashion, could affect the amount of drug (e.g., parent methadone) in the urine and be used to better interpret inappropriate negative results in a patient who was taking methadone as prescribed. In the end, it was recommended that pH and creatinine should be determined on all urine specimens (29). Another expert opinion suggested that urinary creatinine, pH, and temperature should be used to assist with result interpretation and increase specimen reliability for pain management patients (30). Further evidence in pain patients, heroin users, and marijuana/cocaine users showed that normalization of drug concentrations to specific gravity and creatinine was an effective way to cope with diluted urine specimens (31). In this study, 10899 urine specimens were used from pain patients receiving long-term treatment with opioids from 31 pain clinics in 6 states where they had concurrent specific gravity and creatinine measurements. Drug and metabolite concentrations were performed by GC-MS. Correlations of corrected drug concentrations and specific gravity and creatinine relationships were high for all 28 drug and metabolite groups. The overall average positivity rates increased (9.8% by specific gravity correction; 4.2% by creatinine correction) and considered a large portion of variation caused by different patterns of fluid intake.
SPECIMEN VALIDITY TESTING VS OTHER PHYSICIAN TOOLS
Evidence-based recommendation 18
Identification of aberrant drug-taking behavior through specimen validity testing is supplemental to other tools at detecting outcomes in pain management patients. Multiple tools, including specimen validity testing, should be used as a component of UDT to more reliably identify use of relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients.
Strength of Recommendation: A
Quality of Evidence: II
There were no papers identified that specifically compared efficacy of adulteration testing with other physician tools. Moore and colleagues (32) compared structured interview with the Screener and Opioid Assessment for Patients with Pain; the Diagnosis, Intractability Risk, and Efficacy inventory; and/or the Opioid Risk Tool. They evaluated a cohort of 48 chronic pain patients who were subsequently discontinued from their opioids for significant aberrant drug-related behaviors. Because the authors did not include drug testing data in their report, no conclusions can be made that are pertinent to this article. However, psychologist interview was most sensitive (0.77), and the Screener and Opioid Assessment for Patients with Pain was the most sensitive of the questionnaires (0.72) for identifying likelihood of aberrant behavior. Combination of the 2 gave a sensitivity of predicting aberrant behavior of 0.90. Hamill-Ruth (33) compared patient report with the medical record, prescription monitoring report, and POC urine drug screening in an anonymous and voluntary quality improvement project evaluating utility of POC UDT in chronic pain patients using a 10-class test cup with temperature and internal adulteration testing. In addition, adulteration test strips were used. In all, 4.2% of specimens had a temperature below the cutoff limit. Less than 1% showed overt adulteration, but confirmatory testing was not allowed because of the anonymity requirements of the quality improvement project. Consequently, the rigor of the adulteration screening was also limited. The authors did find that patient report was frequently inconsistent with the urine screen, the medical record, and/or the prescription monitoring program. The addition of the UDS and prescription monitoring program identified 9 times as many inconsistencies as the combination of the medical record and patient report alone.
TIMING OF SPECIMEN VALIDITY TESTING
Consensus-based expert opinion 7
Specimen validity testing should be performed on every urine drug test for pain management patients.
Strength of Recommendation: A
Quality of Evidence: II
Multiple guidelines recommend UDT before initiation of therapy, then randomly (34), and 2 to 4 times/year for lower-risk patients, although high-risk patients may need more frequent monitoring (35, 36). Guidelines strongly support random drug testing, but none of these addresses the frequency of specimen validity testing. Random specimen validity testing, on the other hand, can be predicted to decrease the number of specimens that would be confidently considered valid. Accurate interpretation of UDT is critical to clinical decisions for continued prescribing. Hence, efforts to maximize the identification of a valid specimen are paramount. Failure to perform validity testing on all specimens could lead to inappropriate and inaccurate interpretation of drug test results.
BROAD VS TARGETED SPECIMEN VALIDITY TESTING
There is no evidence in the literature to support the statement that targeted specimen validity testing is less effective than broad panel specimen validity testing at detecting outcomes in pain management patients.
Evidence-based recommendation 19
At a minimum, it is recommended that pH, temperature, creatinine, and oxidant testing should be performed on all urine drug tests for pain management patients (timing and site of these tests as noted above). It should also be recognized that these tests will not detect all forms of adulteration.
Strength of Recommendation: A
Quality of Evidence: I (workplace drug testing), III (pain management population)
There is no published evidence for or against targeted specimen validity testing vs broad panel specimen validity testing relative to clinical outcomes. In the absence of evidence, the committee cannot make a recommendation for or against targeted specimen validity testing. The recommendation is that routine specimen validity testing be performed as part of a UDT program to improve the likelihood of accurate interpretation of results. At a minimum, it is recommended that pH, temperature, creatinine, and oxidant testing should be performed on all urine drug tests for pain management patients, recognizing that these tests will not detect all forms of adulteration. As noted above, temperature and pH should be checked preferably within 5 min of specimen collection; creatinine and oxidants (which detect pyridinium chlorochromate, nitrite, and glutaraldehyde) should be tested at the laboratory. At a minimum, POC testing should include onsite specimen validity testing, including temperature, pH, creatinine, and oxidant testing, either incorporated in the testing device or with validated adulterant test strips, if it is to be used for clinical decision-making without definitive testing results available. POC UDT results should be interpreted with caution because of incomplete adulterant testing and limitations of this technology.
PHARMACOGENOMIC CONSIDERATIONS
Understanding the details of the human genome supports research designed to identify heritable causes of disease and response to medications. As such, genetic and genomic testing are rapidly evolving tools for achieving personalized precision medicine. In pain and addiction medicine, genomic variation has been studied to identify associations between gene variants and the pathophysiology of pain sensation, rare pain disorders, pain threshold, as well as patterns of response to pain medications and likelihood for drug addiction. Evidence-based outcome studies are currently lacking for routine clinical application of genomic or genetic testing to guide diagnosis, characterization, and management of chronic pain and drug addiction. However, genetic information is sometimes used to guide drug and dose selection; this application of genetic testing is referred to as pharmacogenetics.
USE OF PHARMACOGENETICS TO GUIDE DRUG AND DOSE SELECTION
Drug response requires a coordinated effort between the 2 major processes of pharmacology: pharmacokinetics and pharmacodynamics. Pharmacokinetics describes the absorption, distribution, metabolism, and elimination of a drug, whereas pharmacodynamics describes the mechanisms of both desirable and undesirable drug effects.
Evidence-based recommendation 20
Although the current evidence in the literature does not support routine genetic testing for all pain management patients, it should be considered to predict or explain variant pharmacokinetics and/or pharmacodynamics of specific drugs as evidenced by repeated treatment failures and/or adverse drug reactions and toxicity.
Strength of Recommendation: A
Quality of Evidence: II
Most evidence for pharmacogenetics associations comes from retrospective or observational studies as opposed to randomized prospective clinical trials. One retrospective study evaluated rates of abnormal pharmacogenetics findings in a pain practice for 104 adult patients, with a focus on 4 genes that code for drug-metabolizing enzymes (37). Overall, 42.3% of test results were normal, 25.5% suggested intermediate metabolizer phenotypes, 7% were poor metabolizers, and 7.2% were ultrarapid metabolizers. Only 3 patients had normal metabolizer phenotypes for all 4 genes. The authors acknowledge a need for large prospective studies conducted with diverse populations to evaluate the generalizability of these results. Another study evaluating the effect of pharmacogenetics on opioid therapy outcomes in an outpatient pain clinic found that the frequency of genetic variants was equivalent to average population frequencies, and only modest associations with opioid dose requirements were observed (38). Nonetheless, gene-based dosing guidelines have been published for select gene–drug pairs, many of which are relevant to chronic pain management. A commonly cited source of gene-based dosing guidelines is the Clinical Pharmacogenomics Implementation Consortium (CPIC). CPIC assigns a level of evidence to each gene–drug pair ranging from “A” (highest level of evidence in favor of changing prescribing of an affected drug) to “D,” wherein evidence is limited and may be conflicting. However, CPIC does not advocate or recommend testing. The guidelines provide expert review of associated literature and guidance for translation of results into actions, when testing is performed. All gene–drug pairs represented by a published guideline have achieved the “A” level of evidence. CPIC guidelines are available publicly through its website (https://cpicpgx.org). The Pharmacogenomics Knowledgebase (https://www.pharmgkb.org) provides summaries of many such associations, along with clinical annotations that are categorized based on the level of evidence surrounding the association. Examples of genes that were identified in ≥10 studies to have associations with opioid response and/or dose are summarized in Table 7.
Genes associated with opioid responses and/or dosages.
SUPPORTIVE TESTING OF PHARMACOGENETICS
Results of pharmacogenetic testing could impact optimal dose and dosing of a specific drug for an individual patient. The impact of the variant metabolic phenotype may be characterized and/or illustrated by metabolic ratios determined with quantitative urine or serum drug testing. Recognizing metabolic patterns and how they may be affected by pharmacogenetic variability is important for interpretation of drug testing results and for detecting drug–drug interactions. For example, a poor metabolizer may not generate a metabolite that is common to normal metabolizers and could be viewed as noncompliant because of the lack of metabolite in the urine. Likewise, a rapid metabolizer may not realize the benefit of a drug and may request higher doses because of accelerated elimination; such a patient could be inappropriately viewed as a drug seeker. Drug–drug interactions can produce or change a variant CYP metabolic phenotype, such as by inhibition of CYP enzyme function. As such, directed quantitative urine or serum drug testing and evaluations of metabolic ratios may help evaluate and monitor the effects of abnormal drug metabolism on drug testing results.
Evidence-based recommendation 21
Directed quantitative drug testing (urine, serum) should be performed to verify and characterize variant pharmacokinetics and patient adherence to a prescribed regimen to assist in the interpretation and application of genetic data.
Strength of Recommendation: B
Quality of Evidence: II
Gene–dose guidelines often recommend therapeutic drug monitoring to optimize dose when impaired metabolic phenotypes are predicted (39, 40). For example, therapeutic drug monitoring was used in combination with CYP2D6 genotyping to more quickly attain therapeutic plasma concentrations and metabolic ratios of imipramine to desipramine (41). The plasma concentration ratios of several antidepressants were shown to be higher in CYP2D6 poor metabolizers and often exceeded the therapeutic ranges in a retrospective study of 62 hospitalized psychiatric patients (42). The ultrarapid metabolizer phenotype for CYP2D6 has been associated with steady-state concentrations of methadone, normalized for dose and patient weight, that were 54% of the concentrations observed in poor metabolizers, suggesting that individualization of methadone dose could be based on plasma concentrations (43). Therapeutic drug monitoring with plasma has also been proposed as a complementary tool for optimizing dose of many other drugs when variant metabolic phenotypes are recognized through pharmacogenetics testing (44–47).
Urine testing results may reflect variation in CYP phenotypes as well. For example, hydromorphone is an expected metabolite of hydrocodone. The ratio of hydromorphone to hydrocodone may represent a patient phenotype that could be explained by variation in CYP2D6 activity (48). In a retrospective evaluation of 25200 urine samples that contained both hydrocodone and hydromorphone, the median metabolic ratio calculated with creatinine-corrected concentrations (mg/g creatinine) was 0.162, and the central 50% range (25th and 75th percentiles) was 0.074 to 0.351. The authors suggest that low metabolic ratios could reflect CYP2D6 metabolic phenotype, although this was not specifically tested. Theoretically, CYP2D6 poor metabolizers would not produce hydromorphone, whereas ultrarapid CYP2D6 metabolizers would produce higher than expected amounts of hydromorphone. Monitoring metabolic ratios could identify CYP2D6 metabolic phenotype and could also detect drug–drug interactions that affect the phenotype for an individual patient. Studies have shown that CYP metabolic status is reflected in the urine metabolic ratios for several other opioids, such as meperidine, oxycodone, buprenorphine, fentanyl, methadone, and propoxyphene, and the benzodiazepine drug diazepam (49–52).
REPORTING, INTERPRETATION, AND COMMUNICATION OF LABORATORY RESULTS WITH PHYSICIANS
Standards specified by various regulatory agencies, such as the Centers for Medicare and Medicaid Services through the Clinical Laboratory Improvement Act, the College of American Pathologists, and the Joint Commission, define critical elements that are common to all testing reports (http://www.ecfr.gov/cgi-bin/text-idx?SID=1248e3189da5e5f936e55315402bc38b&node=pt42.5.493&rgn=div5, Laboratory General Checklist, College of American Pathologists, July 28, 2015 Edition). Outside of these required elements, there is not a standard uniform format agreed on or in use specifically related to the reporting of UDT results for clinical purposes. A review of reports from a variety of laboratories conducting this testing shows a range of formats in use, not dissimilar to the variations observed for any other test report. Reporting formats range from the simple to the inclusion of colorful graphics and interpretative “aids.” These variations represent differences in information services support and marketing and do not represent the quality of the laboratory testing.
Despite the variety in reporting, there are some unique features to consider when configuring a format for reporting UDT results:
The test name should clearly identify the test performed by the drug or drug class as well as purpose or methodology to avoid confusion. For example, naming testing for opiates as “opiate class, screening (immunoassay)” or “opiates, confirmation (LC-MS/MS)” reduces confusion as to what the end user should expect. Today's electronic medical records and laboratory information systems should not be encumbered by overly restrictive character limitations.
Reference intervals in the traditional sense are not applicable, and clinical laboratories, to comply with the requirement to provide a reference or “normal” range, use a variety of comments in this field, from “not detected” to “not applicable.” It is important to note that the term “not detected” or “negative” may be appropriate for some drugs, e.g., cocaine, and in some situations, but not universally. Certainly, one expects to detect the prescribed medications discussed in this guideline in the urine specimens of compliant patients.
The cutoff(s) used should be defined. For screening methods, these are typically established by the assay manufacturer, and many manufacturers make several cutoff options available to accommodate the various settings in which these assays are used, e.g., many opiate screening immunoassays have the option of using a cutoff of either 2000 ng/mL or 300 ng/mL. When the testing method is developed by the laboratory, as is typically the case when LC-MS/MS or GC-MS methods are used, the cutoff is based on validation data, such as the limits of detection and limits of quantification. As will be discussed below, there is no evidence that reporting the cutoff improves the accuracy of the interpretation, but the committee believes it is important. Cutoff values may be set high to avoid false-positive results, and providers should take that into account when interpreting a result below the cutoff.
If test(s) are not approved by the Food and Drug Administration, that should be noted with the result. Furthermore, College of American Pathologists-accredited laboratories should clearly state if the method was internally developed and validated by the laboratory.
Additional information regarding the method and testing may be helpful to the interpretation but impractical to provide or maintain as part of the report. This information should be maintained as part of the laboratory formulary, handbook, testing menu, or other similar resource.
Although there is not a standard format in which UDT results for pain management are reported, the committee agrees that the laboratory should use a format that conveys the results in a clear, concise, and understandable manner, and that this is especially important when done through an electronic medical record system. Additional details regarding reporting are found in Clinical and Laboratory Standards Institute C63, Laboratory Support of Pain Management.
REPORTING OF QUALITATIVE OR SCREENING RESULTS
The manner in which results are reported should be considered carefully. The use of the terms positive and negative in the reporting of qualitative results may mislead the reader who sees these terms as definitive, that is, drug is present and drug is absent from the sample. As an alternative, some laboratories have adopted the use of the assay cutoff (<300 ng/mL or ≥300 ng/mL) in hope that such would convey to the reader that a less than result could range from not detected to just below the reported number (e.g., 0–299 ng/mL) and thus facilitate interpretation. Unfortunately, the literature searches revealed that although this is an often-discussed issue, there have been no studies conducted to demonstrate whether either manner of reporting, or an alternative, is effective.
There is no evidence in the literature that the manner in which qualitative results are reported improves the accuracy of interpretation by the healthcare provider for pain management patients. Additional studies are needed.
TURNAROUND TIME OF REPORTING SCREENING RESULTS
There has also been considerable debate as to how quickly screening results are needed and if such results should be held until confirmatory testing is completed. At the center of this debate is the concern that the release of unconfirmed results could lead to a negative patient care outcome, such as inappropriate dismissal from a facility based on a false-positive screening result.
There is no evidence in the literature that the timing of the release of screening results with respect to the completion of confirmative testing reduces or prevents negative outcomes in patient care. Additional studies are needed.
There may be circumstances when reporting presumptive immunoassay results may be clinically useful. Other providers may prefer all testing be complete before reporting. The committee recommends that laboratories and healthcare providers communicate and determine which pattern of reporting is important to their specific clinical setting. When screening results are reported without confirmation or definitive testing, a reminder should be appended that additional, i.e., confirmatory or definitive, testing is available on request when unexpected results are obtained (unexpected results may include both negative and positive screening results).
REPORTING OF QUANTITATIVE RESULTS
The application of mass spectrometry-based methods to UDT permits both the identification of the compounds present in the sample and the quantification of the result. What to report and how to use the data for patient care warrant discussion.
REPORTING PATTERNS OF DRUG AND DRUG METABOLITES TO INFER COMPLIANCE AND NONCOMPLIANCE
The literature readily supports that identification of an excreted drug and/or drug metabolite is useful in detecting recent exposure to a drug. Not all drugs are metabolized, but when metabolites are known, detection of common metabolites ensures that the drug has been processed by the body, which would infer drug ingestion and possibly compliance.
Evidence-based recommendation 22
Quantitative or proportional patterns of some drug and drug metabolites are recommended to explain complex cases and detect the presence of pharmaceutical impurities, simulated compliance (e.g., adding drug directly to urine), and/or the major route of metabolism in a patient.
Strength of Recommendation: I (insufficient) for most drugs, B for some drugs
Quality of Evidence: II
The current evidence in the literature does not support using specific patterns of conjugated and unconjugated drug and drug metabolites to define a patient's metabolic phenotype. Additional studies are needed.
Dickerson et al. found that direct measurement of glucuronide metabolites in urine improved detection of opioids, including codeine, morphine, hydromorphone, oxymorphone, and buprenorphine (27). Of significance, no patients tested positive for buprenorphine parent only, suggesting that either hydrolysis or direct measurement of glucuronides was required to evaluate adherence to buprenorphine in this population. In another retrospective study, 216 urine samples from 70 patients prescribed buprenorphine were evaluated (24). Buprenorphine was found in only 33 samples, whereas norbuprenorphine was found in all samples. There was strong evidence that 9 samples were adulterated. Of the adulterated samples that could be further evaluated (n = 6), the norbuprenorphine/buprenorphine ratio was <0.02 as compared with ratios of >0.99 for typical samples. Four of the samples had buprenorphine concentration in the 10000 to 50000 ng/mL range and naloxone concentrations between 4000 and 15000 ng/mL. Because the expected ratio of the pharmaceutical product suboxone is 4:1, these data suggest that the patients who provided these urine samples had added drug directly to the urine after voiding to mimic compliance.
In a retrospective case-controlled study, the prevalence of hydromorphone as a metabolite of morphine was evaluated relative to morphine dose and sex (53). Patients were included if the urine drug screen and chart review indicated that the patient was taking only morphine. Of the 32 patients meeting the inclusion criteria, 11 patients did not show evidence of hydromorphone and were designated as the controls. The remaining 21 patients showed evidence of both morphine and hydromorphone (prevalence, 66%). The assay reporting limit was 50 ng/mL. Hydromorphone was observed in 87% of positive urine samples collected from women and 47% of positive samples collected from men, but the ratio of hydromorphone to morphine in urine was not significantly different between the sexes. In general, the concentration of hydromorphone was approximately 2% of the total morphine concentration, suggesting that hydromorphone occurs as a minor metabolite of morphine. Detection of hydromorphone is likely to be associated with the detection limit of the assay used. A similar relationship has been described for hydrocodone as a minor metabolite of codeine in both controlled administration and postoperative patient studies (54). The concentration of hydrocodone could appear in urine at up to 11% of the parent (codeine) concentration.
In another retrospective study of UDT results wherein pharmacy history was known, a small amount of codeine was observed in the urine collected from patients prescribed only morphine (55). In all, 15 of 535 samples evaluated were described to contain total morphine concentrations in the range of 10000 to 150000 ng/mL and total codeine concentrations between 20 and 50 ng/mL. Using average concentrations of morphine (94000 ng/mL) and codeine (40 ng/mL), it was estimated that the fraction of codeine nearly approximates the estimated impurity observed in pharmaceutical morphine (0.04%). The authors conclude that their data suggest evidence of pharmaceutical impurity rather than a minor route of metabolism.
APPROXIMATION OF THE TIME OF LAST DOSE
Some laboratories have applied therapeutic drug monitoring principles appropriate to serum, plasma, or blood concentrations to the quantified urine results with claims of such permitting a more effective assessment of the approximate time of the patient's last dose.
Evidence-based recommendation 23
UDT (quantitative or qualitative) is not recommended for approximating the time of last dose.
Strength of Recommendation: B
Quality of Evidence: II
In a retrospective study of 161 patients prescribed transdermal fentanyl, the medial metabolic ratio of norfentanyl to fentanyl in urine was 6, with the central 50% range (25th and 75th percentiles) 3 to 12 (56). However, the study also acknowledged that the metabolic ratio could vary in a single subject by 10-fold and between subjects by 37-fold. No pattern was shown between the total amount of drug excreted and the metabolic ratio, suggesting that metabolic ratio does not correlate with dose. Studies with oxycodone and hydrocodone in urine have suggested that use of a proprietary algorithm can predict dose compliance (19, 20), but these data were challenged based on substantially overlapping distributions of urine concentrations by dose. Misclassification of dose estimates occurred in >25% of patients (57, 58).
NORMALIZATION OF QUANTITATIVE RESULTS TO CREATININE OR SPECIFIC GRAVITY
The reporting of quantitative UDT results normalized to creatinine (ng drug/mg creatinine) or to specific gravity stems from the use of the practice in the testing of other urinary analytes, especially hormones, for which it serves as a means of assessing the completeness of a 24-h collection and accounting for variations between random sample collections.
There is insufficient evidence to support the practice of normalizing quantitative results to creatinine or specific gravity or that doing so is an effective means of detecting compliance or misuse and diversion. Additional studies are needed.
Two articles were identified related to the normalization of results. In the first, Pesce et al. (59) used mathematical modeling to assess the upper limits of drug excretion observed for 8971 patients and to define reference intervals for the measured opiates. The distribution pattern obtained was minimally affected when the excreted drug concentrations were normalized to creatinine. Insufficient data were provided to fully assess the impact of this transformation. In the second, Barakat et al. (48) investigated the utility of the excretion of urinary hydrocodone concentrations and urinary hydromorphone concentrations to assess the variability of hydrocodone metabolism. Concentrations were normalized to creatinine before modeling, but nonnormalized data were not provided.
INTERPRETATION OF RESULTS
Much has been written and discussed about the ability of physicians and other healthcare providers to consistently and correctly interpret UDT results. Urine results for any analyte are among the most complicated to interpret, and those for drug analysis are no exception. One must begin with sound knowledge of the pharmacology of the drugs (including the expected metabolic profiles), appreciate the variation in renal function over the course of the day and between individuals, recognize the inherent limitations of a randomly collected urine sample, and tie these together in light of the limitations and strengths of the analytical methods used to generate the result. Unfortunately, the data show that many clinical providers have insufficient knowledge and expertise to correctly interpret urine laboratory test results for pain management patients.
Evidence-based recommendation 24
Data showed that many clinical providers have insufficient knowledge and expertise to correctly interpret urine laboratory test results in pain management patients. It is recommended that clinicians should contact laboratory personnel for any test result that is inconsistent with the clinical picture and/or prescribed medications to more effectively interpret urine test results in pain management patients.
Strength of Recommendation: A
Quality of Evidence: I
Evidence-based recommendation 25
It is recommended that laboratories provide educational tools and concise detailed reports to guide the interpretation of UDT for pain management patients by clinicians.
Strength of Recommendation: A
Quality of Evidence: III
Evidence-based recommendation 26
It is recommended that clinical laboratories offering pain management testing must also have knowledgeable personnel who can assist clinicians to correctly interpret urine laboratory test results in pain management patients.
Strength of Recommendation: A
Quality of Evidence: III
To assess physician knowledge on UDT interpretation, Reisfield et al. developed a questionnaire consisting of 7 multiple-choice questions (60, 61). The questions included assessment of knowledge regarding the metabolism and excretion patterns expected for codeine, morphine, and heroin, the interpretation of unexpected negative screening results, the effects of poppy seed ingestion, and implications of second-hand exposure to marijuana smoke. The authors administered the assessment to 170 physicians attending 2 conferences: the first an opiate education conference (60) and the second a family medicine conference (61). Of the 114 physicians attending the opioid education conference who completed the questionnaire, 77 reported using UDT as part of their management, whereas 37 did not. None of the physicians achieved a score of 100%, and only 30% answered >50% correctly. The performance of the physicians who performed UDT was the same as those who did not. Of the 60 family medicine physicians who participated in the second assessment, 44 reported using UDT and 16 did not. Again, none achieved a score of 100%, and only 20% answered more than half the questions correctly. For this group, the highest score was 5 of 7 questions correct, or 71%, and those who self-identified as routinely ordering UDT performed better on only 4 of the 7 questions compared with those physicians who indicated they did not routinely order the testing. A new question was added surveying who would consult the laboratory director when abnormal or unexpected findings were reported and found only 23% of physicians indicated they would contact the laboratory director. For each group, the authors concluded that physician knowledge of UDT interpretation is inadequate, that physicians are making important clinical decisions without understanding how to interpret the results, which could have severe consequences for both the patient and physician when tests are misinterpreted, and that efforts should be made to increase physician knowledge and encourage laboratory consultation.
Although there are a few articles that demonstrate that physicians are not proficient in interpreting UDT, there is no evidence that clinical pathology and laboratory medicine consultations are more effective for correct interpretation of urine test results for any drug given to pain management patients. This is most likely because providers are unaware of their knowledge gap and do not currently contact the laboratory director. Therefore, the studies cannot be performed. Despite the lack of evidence of the efficacy of laboratory medicine consultations, we strongly recommend that laboratories offering pain management testing have knowledgeable personnel to assist clinicians. Laboratories, in all aspects of testing, are responsible for providing accurate results, and assisting with interpretation and pain management testing should be no exception. At a minimum, laboratories should provide educational resources and detailed reports, including whom to contact with questions regarding interpretation.
UTILITY OF CLINICAL ALGORITHMS
Quantitative UDT results have been used alone or in combination with clinical data (e.g., drug dose, clinical presentation) to predict drug efficacy and side effects, guide drug dosing, and/or assess compliance. However, the utility and accuracy of these clinical-based algorithms are unclear.
There is insufficient evidence in the literature to determine whether quantitative concentrations of prescribed medications, alone or in combination with a clinical algorithm, improve the use of the testing in terms of identifying compliance, efficacy, or noncompliance. Additional studies are needed.
There are a few articles that describe the use of quantitative testing and clinical algorithms, but none demonstrate how their use improved outcomes. Therefore, there is no evidence that the reporting of quantitative drug concentrations is more effective in facilitating the assessment of any outcomes for pain management patients (25, 50, 56, 59, 62–64).
CONCLUSION
In the end, the purpose of this guideline was to compile evidence-based recommendations for the use of laboratory and POC UDT for relevant over-the-counter medications, prescribed and nonprescribed drugs, and illicit substances in pain management patients. Although these guidelines did find evidence to address several important areas and questions, we also uncovered significant gaps in the literature where additional research studies are needed to provide evidence for future recommendations.
Footnotes
↵† M.A. Huestis is now at NMS Laboratories, Willow Grove, PA.
↵10 Nonstandard abbreviations:
- LMPG
- laboratory medicine practice guidelines
- POC
- point of care
- UDT
- urine drug testing
- 6-AM
- 6-acetylmorphine
- CPIC
- Clinical Pharmacogenomics Implementation Consortium.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form.
Employment or Leadership: P.J. Jannetto, Health Service Advisory Group, Inc.; N.C. Bratanow, American Board of Pain Medicine; W.A. Clarke, AACC Board of Directors. G.A. McMillin, ARUP Laboratories. Consultant or Advisory Role: W.A. Clarke, Thermo Fisher Scientific, Shimadzu. Stock Ownership: None declared. Honoraria: None declared. Research Funding: None declared. Expert Testimony: None declared. Patents: None declared. Other Remuneration: R.J. Hammill-Ruth, AACC, AAPM; M.A. Huestis, NC Governor's Institute on Substance Abuse.
Role of Sponsor: No sponsor was declared.
- Received July 17, 2017.
- Accepted October 12, 2017.
- © 2017 American Association for Clinical Chemistry
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.
- 46.
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.
- 64.↵
- 65.
- 66.
- 67.
- 68.
- 69.
- 70.
- 71.
- 72.
- 73.
- 74.
- 75.
- 76.
- 77.
- 78.
- 79.
- 80.
- 81.
- 82.
- 83.
- 84.
- 85.
- 86.
- 87.
- 88.
- 89.
- 90.
- 91.
- 92.
- 93.
- 94.
- 95.
- 96.
- 97.
- 98.
- 99.
- 100.
- 101.
- 102.
- 103.
- 104.
In this issue
Jump to section
- Article
- BACKGROUND
- EVIDENCE-BASED RECOMMENDATIONS/STATEMENTS
- FREQUENCY OF LABORATORY TESTING
- LABORATORY TESTING AND ITS ABILITY TO IDENTIFY NONCOMPLIANCE IN PAIN MANAGEMENT REGIMENS
- LABORATORY TESTING VS OTHER PHYSICIAN TOOLS, PRESCRIPTION MONITORING, AND SELF-REPORT
- SPECIMEN TYPES
- QUALITATIVE AND SEMIQUANTITATIVE SCREENING ASSAYS
- QUALITATIVE DEFINITIVE TESTING
- POC TESTING
- TIMING OF UDT
- COST-EFFECTIVENESS OF UDT
- DEFINITIVE TESTING
- QUANTIFICATION VS QUALITATIVE DEFINITIVE TESTS
- DETECTION LIMITS
- PREANALYTICAL HYDROLYSIS (ENZYMATIC/CHEMICAL) OF URINE
- USE OF CONJUGATED AND UNCONJUGATED DRUG METABOLITES
- ADULTERANT/SPECIMEN VALIDITY TESTING
- SPECIMEN VALIDITY TESTING
- SPECIMEN VALIDITY TESTING VS OTHER PHYSICIAN TOOLS
- TIMING OF SPECIMEN VALIDITY TESTING
- BROAD VS TARGETED SPECIMEN VALIDITY TESTING
- PHARMACOGENOMIC CONSIDERATIONS
- USE OF PHARMACOGENETICS TO GUIDE DRUG AND DOSE SELECTION
- SUPPORTIVE TESTING OF PHARMACOGENETICS
- REPORTING, INTERPRETATION, AND COMMUNICATION OF LABORATORY RESULTS WITH PHYSICIANS
- REPORTING OF QUALITATIVE OR SCREENING RESULTS
- TURNAROUND TIME OF REPORTING SCREENING RESULTS
- REPORTING OF QUANTITATIVE RESULTS
- REPORTING PATTERNS OF DRUG AND DRUG METABOLITES TO INFER COMPLIANCE AND NONCOMPLIANCE
- APPROXIMATION OF THE TIME OF LAST DOSE
- NORMALIZATION OF QUANTITATIVE RESULTS TO CREATININE OR SPECIFIC GRAVITY
- INTERPRETATION OF RESULTS
- UTILITY OF CLINICAL ALGORITHMS
- CONCLUSION
- Footnotes
- References
- Figures & Data
- Info & Metrics