- Email: [email protected]
Monitoring opioid and benzodiazepine use and abuse: Is oral fluid or urine the preferred specimen type?
Clinica Chimica Acta 481 (2018) 75–82
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Monitoring opioid and benzodiazepine use and abuse: Is oral fluid or urine the preferred specimen type?☆
T
⁎
Athena K. Petridesa,b, , Stacy E.F. Melansona,b, Michalis Kantartjisa,c, Rachel D. Led, Christiana A. Demetrioue,f, James G. Floodb,g a
Department of Pathology, Brigham and Women's Hospital, Boston, MA, United States Harvard Medical School, Boston, MA, United States c Department of Medicine, Brigham and Women's Hospital, Boston, MA, United States d University of Massachusetts Medical School, Worcester, MA, United States e The Cyprus Institute of Neurology and Genetics, Nicosia, Cyprus f The Cyprus School of Molecular Medicine, Nicosia, Cyprus g Department of Pathology, Massachusetts General Hospital, Boston, MA, United States b
A R T I C L E I N F O
A B S T R A C T
Keywords: Oral fluid drug testing Urine drug testing Opioid Benzodiazepine Liquid chromatography-tandem mass spectrometry Pain management
Background: Oral fluid (OF) has become an increasingly popular matrix to assess compliance in pain management and addiction settings as it reduces the likelihood of adulteration. However, drug concentrations and windows of detection are not as well studied in OF as in urine (UR). We compared the clinical utility and analytical performance of OF and UR as matrices for detecting common benzodiazepines and opioids. Methods: OF and UR concentrations of 5 benzodiazepines and 7 opioids were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS) in 263 paired OF and UR specimens. UR creatinine was measured and prescription medications were reviewed. Results: The benzodiazepines 7-aminoclonazepam, lorazepam, and oxazepam exhibited statistically higher detection rates in UR. For opioids, 6-AM was statistically more likely to be detected in OF, while hydromorphone and oxymorphone were statistically more likely to be detected in UR. Chemical properties including glucuronidation explain preferential detection in each matrix, not UR creatinine nor prescription status. Conclusion: We found that OF is the preferred matrix for 6-AM, while UR is preferred for 7-aminoclonazepam, lorazepam, oxazepam, hydromorphone, and oxymorphone. However, OF should be considered if the risk of adulteration is high and use and/or misuse of benzodiazepines, hydromorphone, and oxymorphone is low.
1. Introduction Between 2000 and 2015, half a million deaths were due to drug overdoses, and for the first time, in 2015, drug overdoses were the leading cause of accidental death in the United States, highlighting prescription drug misuse and addiction as a national issue in recent years [1,2]. Prescription opioids and benzodiazepines constituted nearly all (70% and 30%, respectively) of prescription overdose deaths in 2013, with deaths commonly involving both substances [1,3–5].As a consequence, substantial efforts have been made to prevent and treat substance abuse, including opioid-agonist medication-assisted treatment (OA-MAT) [1,2].
Routine and random drug testing as an adjunct to OA-MAT has provided an objective measure of compliance and treatment efficacy in both the pain management and addiction settings [6–8]. Historically, urine (UR) specimens have been used for drug monitoring. UR collection is non-invasive and drugs are present at higher concentrations for longer periods of time compared to serum [6–9]. Numerous studies have demonstrated the effectiveness of UR drug testing for monitoring compliance [6–8]. However, UR can be easily adulterated, particularly if collections are not observed, and patients with shy bladder or anuria may not be able to provide a specimen [10]. For this reason, the utility of oral fluid (OF) has been explored as a tool to assess compliance. OF collection significantly reduces the
Abbreviations: opioid-agonist medication-assisted treatment, (OA-MAT); oral fluid, (OF); urine, (UR); Massachusetts General Hospital, (MGH); liquid chromatography-tandem mass spectrometry, (LC-MS/MS); 6-acetylmorphine, (6-AM); mass spectrometer, (MS); Brigham and Women's Hospital, (BWH); heated electrospray ionization, (HESI); Substance Abuse and Mental Health Services Administration, (SAMHSA) ☆ This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. ⁎ Corresponding author at: Brigham and Women's Hospital, 75 Francis Street, Amory 2, Boston, MA 02115, United States. E-mail address: [email protected] (A.K. Petrides). https://doi.org/10.1016/j.cca.2018.02.034 Received 22 December 2017; Received in revised form 7 February 2018; Accepted 26 February 2018 Available online 27 February 2018 0009-8981/ © 2018 Published by Elsevier B.V.
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
likelihood of adulteration. In recent years, more studies have investigated OF in monitoring drugs of abuse with promising results. Nordal et al. reported that benzodiazepines can be measured qualitatively in OF and suggested OF may be an alternative for detection of clonazepam, diazepam, and alprazolam [11]. Likewise, Conermann et al. concluded that OF may be used for monitoring compliance of opioids and benzodiazepines after analyzing 132 paired specimens [12]. OF has some analytical challenges including low sample volumes and difficult specimen collections in patients with conditions such as dry mouth. Furthermore, collections are typically performed by clinic staff and can be time-consuming as they involve the patient rinsing their mouth and waiting 10–15 min prior to collection. Additionally, OF specimens may become contaminated with mouthwash, toothpaste, foods, and beverages [13,14]. Buccal contamination of OF by placing the drug sublingually immediately prior to collection may also pose a challenge for drug compliance assessment. Although OF has been studied less extensively, current literature demonstrates that OF exhibits lower drug and metabolite concentrations and narrower windows of detection compared to UR for the majority of drugs [9]. There are a limited number of studies directly comparing the performance of paired OF and UR specimens for both opioids and benzodiazepines. Additionally, to our knowledge, there are no published studies reporting the ratio of OF to UR concentrations for opioids and benzodiazepines in matched OF and UR specimens. In this study, we examined the clinical and analytical performance of each matrix and make suggestions on the utility of each matrix.
Table 1 Limit of detection (LOD) for urine (UR) versus oral fluid (OF) liquid chromatographytandem mass spectrometry (LC-MS/MS) Testing. Class
Benzodiazepines
Opioids
Drug or metabolite
7-Aminoclonazepam Alpha-OH-Alprazolam Alprazolam Lorazepam Nordiazepam Oxazepam 6-Acetylmorphine Codeine Hydrocodone Hydromorphone Morphine Oxycodone Oxymorphone
LOD (ng/mL) for LC-MS/MS UR
OF
50 50 N/A 50 50 50 5 50 50 50 50 50 50
1 N/A 2 2 2 2 2 2 2 4 2 1 2
2.3. Urine drug analysis UR was analyzed at the Brigham and Women's Hospital (BWH) Clinical Chemistry Laboratory (Boston, MA) using a laboratory-developed LC-MS/MS method. Samples were prepared by adding an internal deuterated standard to the following drugs or metabolites: 7-aminoclonazepam, alpha-hydroxy-alprazolam, lorazepam, nordiazepam, oxazepam, 6-AM, codeine, hydrocodone, hydromorphone, morphine, oxycodone, and oxymorphone. Samples were diluted and subjected to a hydrolysis step to remove glucuronide and sulfate groups. Chromatographic separation was achieved on a ACQUITY UPLC I-Class (Waters, Milford, MA) using a Kinetex C18 analytical column (Phenomenex Inc., Torrance, CA) and mass spectrometric analysis was performed on a tandem triple quadrupole Xevo TQS (Waters, Milford, MA) preceded by HESI. 6-AM was analyzed with the same equipment, but was not subjected to a hydrolysis step. The limits of detection (LOD) for each drug or metabolite in UR are also described in Table 1.
2. Methods 2.1. Specimen acquisition A total of 263 paired OF and UR specimens were collected consecutively from 140 unique patients at the Massachusetts General Hospital (MGH) addiction-psychiatry clinics during routine visits and processed at the MGH Clinical Chemistry Laboratory (Boston, MA). For OF collection, the Orasure Intercept Sample Collection Device (Orasure Technologies, Bethlehem, PA) was utilized according to the manufacturer's collection instructions. OF and UR pairs were received by the laboratory within 2–10 h of collection. OF was refrigerated overnight and tested the next day. UR was frozen within 8 h of receipt and tested in batches at a later date. The Partners Human Research Committee approved this study.
2.4. Calculation of oral fluid urine ratios UR creatinine was measured using the rate-blanked Jaffe reaction with Roche Diagnostics reagents on a Roche Cobas e501 (Roche Diagnostics, Indianapolis IN). Quantitative UR drug measurements for 7-aminoclonazepam, alprazolam/alpha-hydroxy-alprazolam, nordiazepam, 6-AM, codeine, morphine, oxycodone, and oxymorphone were corrected for creatinine levels using the following formula: [Drug in ng/ mL]/[UR Creatinine in mg/dL]×100. Corrected drug concentrations were used to calculate oral fluid:urine ratios (OF:UR) using the following formula: [OF Drug in ng/mL]/[UR Drug in ng/mL]. Ranges and medians were calculated.
2.2. Oral fluid drug analysis As published previously, OF was analyzed at MGH using a laboratory-developed liquid chromatography-tandem mass spectrometry (LCMS/MS) method [15,16]. The following benzodiazepines and opioids were detected: 7-aminoclonazepam, alprazolam, lorazepam, nordiazepam, oxazepam, 6-acetylmorphine (6-AM), codeine, hydrocodone, hydromorphone, morphine, oxycodone, and oxymorphone. Briefly, OF specimens were mixed with Internal Standard Solution containing deuterated analogues of each analyte. The mixture was injected onto a TLX2 chromatograph (Thermo Scientific, Waltham, MA) where the analytes and internal standards were first isolated on a Cyclone-P turbulent-flow extraction column (Thermo-Fisher, Franklin, MA) and then transferred to an Ascentis Phenyl analytical column (Supelco, Bellefonte, PA). The analytes were then separated using a gradient elution program and detected using a Thermo Quantum Ultra triple quadrupole MS (Thermo Scientific, Waltham, MA) equipped with a heated electrospray interface (HESI-II) operated in the positive ion mode. The limits of detection (LOD) for each drug or metabolite in OF are described in Table 1.
2.5. Medication review The electronic health records of all patients in the study were reviewed for any active prescriptions/medications at the time of specimen collection. In this study, a prescription was considered active if the start date was before specimen collection and the end date was after or within 5 days of specimen collection. 2.6. Statistical analysis The non-parametric McNemar's symmetry test was used to assess agreement between paired OF and UR specimens. A p-value < 0.05 was considered to be statistically significant. To assess for dilution of UR specimens, creatinine concentrations of UR specimens were calculated. The UR creatinine distribution in 76
Clinica Chimica Acta 481 (2018) 75–82 Fig. 1. Positive Benzodiazepines in Urine (UR) versus Oral Fluid (OF). Light gray bars represent the number of paired specimens for which drug/metabolites were positive in both urine (UR) and oral fluid (OF) (i.e., UR+/OF+). Dark gray bars represent specimens for which drug/metabolites were positive in only UR (i.e., UR+/OF−). Black bars represent specimens for which drug/metabolites were positive in only OF (i.e., UR−/OF+). The results of the McNemar's symmetry test are displayed next to each drug/metabolite.
A.K. Petrides et al.
specimens for which opioids and/or benzodiazepines were detected in only OF (i.e, target population) was compared to the UR creatinine distribution of a) samples negative for both opioids and benzodiazepines in both OF and UR (i.e., reference population, n = 123) and b) randomly-selected samples from an independent, healthy population (i.e., control population, n = 263). The creatinine distributions were compared visually by a cumulative frequency distribution graph and numerically by the non-parametric Kruskal-Wallis test. All analyses were performed using STATA version SE12 and GraphPad Prism 6. 3. Results Of the 263 paired specimens, 123 (47%) were negative for benzodiazepines and opioids in both OF and UR (i.e., UR-/OF-). The remaining 140 (53%) were positive for 1 or more benzodiazepines and/or opioids in either OF and/or UR. 3.1. Benzodiazepines The detection rate for benzodiazepines was higher in UR with 33% of samples positive for 1 or more benzodiazepine(s) compared to 21% of OF specimens. Furthermore, 7-aminoclonazepam, lorazepam, and oxazepam were statistically more likely to be detected in UR versus OF (p = 0.002, p = 0.008, p < 0.001, respectively; Fig. 1). Of the specimens positive for 7-aminoclonazepam, lorazepam, and oxazepam, 28% (n = 19), 89% (n = 8), and 82% (n = 14), respectively, were detected only in UR (i.e., UR+/OF−). No specimens positive for lorazepam or oxazepam were positive in OF without also being positive in UR (i.e., UR+/OF+). The number of specimens positive only in OF (i.e., UR−/ OF+) was low for 7-aminoclonazepam (n = 4), alprazolam/alpha-OHalprazolam (n = 1), and nordiazepam (n = 2). Nonetheless, alprazolam and nordiazepam were not statistically more likely to be detected in UR versus OF (p = 0.07, p > 0.999, respectively; Fig. 1). 3.2. Opioids The detection rate for opioids was slightly higher in OF with 32% of samples positive for 1 or more opioids in OF compared to 28% in UR. 6AM was statistically more likely to be detected in OF (p < 0.001), while hydromorphone and oxymorphone were statistically more likely to be detected in UR (p < 0.001, p < 0.004, respectively; Fig. 2). Oxycodone was preferentially detected in OF with 4 UR−/OF+ specimens and was never detected in only UR (i.e., UR+/OF−), though the results were not statistically significant (p = 0.125). Codeine and morphine were not preferentially detected in either specimen types, having been detected in both or either matrix (p = 0.782, p = 0.336, respectively). Hydrocodone was detected in only 2 samples, 1 UR+/OF + and 1 UR+/OF− (p = 0.999; Fig. 2). The detection of morphine, hydromorphone, and oxymorphone can be a result of heroin (6-AM), morphine, and oxycodone metabolism, respectively. 6-AM was preferentially detected in OF with 15 positive specimens detected in only OF (Fig. 2). There was 1 paired specimen in which 6-AM was detected only in UR. Furthermore, 2 UR specimens positive for 6-AM did not contain detectable UR morphine levels (data not shown). Of the 21 UR+/OF− specimens for hydromorphone, 20 were also UR+/OF+ for morphine. Both of the 2 UR+/OF+ for hydromorphone were UR−/OF− for morphine. Lastly, of the 9 UR +/OF− for oxymorphone, 3 were also UR+/OF+ for oxycodone. Likewise, all 5 UR+/OF+ specimens for oxymorphone were also UR +/OF+ for oxycodone. 3.3. Drug concentrations and ratios Concentrations and OF:UR ratios for all drugs and metabolites with 5 or more positive samples are shown in Table 2. Overall, wide ranges in concentrations were seen in both UR and OF. 6-AM, oxycodone, and 77
Fig. 2. Positive Opioids in Urine (UR) versus Oral Fluid (OF). Light gray bars represent the number of paired specimens for which drug/metabolites was positive in both urine (UR) and oral fluid (OF) (i.e., UR+/OF+). Dark gray bars represent specimens for which drug/metabolites were positive in only UR (i.e., UR+/OF−). Black bars represent specimens for which drug/metabolites were positive in only OF (i.e., UR−/OF+). The results of the McNemar's symmetry test are displayed next to each drug/metabolite.
A.K. Petrides et al.
Clinica Chimica Acta 481 (2018) 75–82
78
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
detected in both matrices (i.e., UR+/OF+), 8 were detected in UR only (i.e., UR+/OF−), 2 were detected in OF only (i.e., UR−/OF+) and 3 were not detected in either matrix.
Table 2 Description of opioid and benzodiazepine quantitative findings in urine (UR) versus oral fluid (OF) liquid chromatography-tandem mass spectrometry (LC-MS/MS) testing. Drug or metabolite (n)⁎
Oxymorphone (5) Morphine (31) 7-Aminoclonazepam (45) Alprazolam/Alpha-OHAlprazolam (8) Nordiazepam (6) Codeine (10) Oxycodone (8) 6-Acetylmorphine (10)
⁎
UR LC-MS/MS (ng/ mL) median (ranges)
OF LC-MS/MS (ng/mL) median (ranges)
OF:UR ratio median (ranges)
9731 (1111–45,744) 17,100 (31–83,586) 519 (21–3602) 570 (97–3101) 1045 (290–1304) 468 (198–854) 6887 (309–18,847) 241 (2.4–3404)
5 (2–15) 24 (2–1641) 5 (1–95.7) 9 (2–67) 24 (8–73) 18 (3–34) 476 (27–1201) 88 (2.4–431)
0.0005 (0.0002–0.002) 0.001 (0.0001–0.31) 0.01 (0.0007–0.07) 0.02 (0.006–0.05) 0.024 (0.008–0.1) 0.05 (0.004–0.09) 0.1 (0.03–0.22) 0.5 (0.006–3.6)
4. Discussion We measured the concentrations of 5 benzodiazepines and 7 opioids in 263 paired UR and OF specimens. Our study found that all 5 benzodiazepines were more frequently detected in UR, with 7-aminoclonazepam, lorazepam, and oxazepam statistically more likely to be detected in UR versus OF. Conversely, the overall detection rate for opioids was slightly higher in OF than UR with 6-AM statistically more likely to be detected in OF. Hydromorphone and oxymorphone were statistically more likely to be detected in UR. The chemical properties of drugs including acidity and glucuronidation govern to what extent they are detected in OF and/or UR (Table 4) [17]. Passive diffusion is the most common mechanism of drug transfer to OF. Weakly basic, non-protein bound molecules with lower molecular weights are more likely to be present in OF. Also, OF exhibits a slightly acidic pH and attracts weak bases that ionize and become trapped in the acidic environment. Drugs that are converted to more soluble forms, such as glucuronides, are more likely to be detectable in UR. For the same reason, hydroxylated metabolites, not the primary drug, are predominately found in UR. Consistent with these properties, our results show that concentrations of all drugs were lower in OF than UR (Table 2). In addition, we found that weakly basic, nonglucuronidated analytes like 6-AM and oxycodone were more frequently detected in OF and had the highest OF:UR ratios (Table 4). Non-conjugated analytes with intermediate OF:UR ratios such as codeine, nordiazepam, and alprazolam were equally detected in both matrices. Lastly, highly glucuronidated analytes such as oxymorphone, hydromorphone, oxazepam and lorazepam were most frequently detected in UR and had the lowest OF:UR ratios. Morphine is an exception, as it has a low OF:UR ratio and is highly conjugated, yet it is detected equally well in OF and UR. Morphine had the highest average concentrations of any analyte in both OF and UR. This is likely due to the fact that high doses of morphine are prescribed or ingested, which lead to detectable concentrations in OF despite the high degree of glucuronidation. The range of drug concentrations is infrequently published. Drugs with narrow OF:UR ranges (i.e., 101–102) such as oxymorphone and 7aminoclonazepam showed matrix preference based on their chemical properties as mentioned above. However, 6-AM and morphine were found to have wide ranges of concentrations (i.e., 103) in both matrices, which may imply that the detection of these substances is highly-dependent on the time of drug use and collection. Particularly for 6-AM positive specimens we observed that the relationship between OF and UR morphine is inversely proportional. As the concentration of OF morphine decreases, the concentration of UR morphine increases suggesting more remote use. 6-AM is thought to be detected in OF and UR up to 8 h and 34.5 h after use, respectively [9,18]. However, a study by Vindenes et al. showed that 6-AM was detected in OF 4–8 days after use with undetectable UR and OF morphine levels, thereby challenging these reported detection windows [19]. Our results seem to support Vindenes et al.'s observation. We had 3 patients that seem to match that same profile, with detectable 6-AM levels and very low concentrations of morphine OF and undetectable levels in UR (sp. 26–28, Fig. 3). Additionally, our study corroborates recent literature showing that 6AM is detectable in UR with no detectable morphine [19–21]. Although hydromorphone and oxymorphone were statistically more likely to be detected in UR, the clinical significance of this finding is less relevant in our patient population where only 4 patients were prescribed hydromorphone and none prescribed oxymorphone. Hydromorphone was present in combination with morphine in 87% (20/23) of specimens and oxymorphone with oxycodone in 57% (8/14) of specimens, suggesting that the specimens positive for
Note: Only drugs/metabolites with n ≥ 5 are listed.
codeine had the highest OF:UR ratios at 0.5, 0.1, and 0.05 respectively, oxymorphone and morphine had the lowest ratios of 0.0005 and 0.001, respectively, whereas the three benzodiazepines had intermediate ratios. The largest range in OF:UR ratios was seen in specimens positive for morphine and 6-AM, both to the order of 103 (Table 2). Otherwise, the remaining ranges were to the order of 101 except for 7-aminoclonazepam, which had a range to the order of 102. For the majority of patients, as the OF:UR ratio for morphine decreases, UR 6-AM concentration increases while OF 6-AM concentration decreases (Fig. 3). There were a total of 6 patients with detectable 6-AM in either OF or UR and detectable morphine in OF, but undetectable morphine in UR. 3 of those patients had high concentrations of morphine in OF (SP. 1–3, Fig. 3), whereas the other 3 patients had very low concentrations morphine in OF (SP. 26–28, Fig. 3). 3.4. Urine creatinine To assess whether dilution of UR specimens may have altered drug concentrations, we quantified creatinine concentrations in UR specimens. The UR creatinine distribution in specimens in which benzodiazepines and/or opioids were detected only in OF (i.e., UR−/OF+) was similar to the UR creatinine distribution of both the reference and control populations (p = 0.6232) (Fig. 4). There was only 1 specimen in the target population with a creatinine concentration considered dilute according to SAMSHA guidelines (i.e., < 20 mg/dL). Additionally, only 46.9% (n = 15) of UR−/OF+ specimens had creatinine concentrations < 100 mg/dL. Interestingly, the percent of samples with creatinine levels < 20 mg/dL was higher in the reference population (8.9%) than control (1.9%) and target (3.1%) populations. 3.5. Medication review Lastly, we investigated whether the drugs detected were prescribed. Only 24% (n = 62) of paired specimens were from patients who were prescribed 1 or more of the medications measured. Overall, prescribed drugs were detected at rates between 80 and 100% in UR for all drugs except hydromorphone which was detected at a rate of 50% in UR (Table 3). Moreover, prescribed drugs were detected at higher rates in UR versus OF in all drugs except diazepam, which was detected with equal frequency (80%). Clonazepam was prescribed most frequently. Of the 31 specimens associated with a clonazepam prescription, 18 were 79
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
Fig. 3. Pattern of 6-AM and Morphine Concentrations in Urine (UR) and Oral Fluid (OF). 6-AM concentrations in OF (black line) and UR (shaded gray line) for each paired specimen (SP 1–28) are plotted against their OF/UR morphine concentration ratio. Concentrations are displayed in order of decreasing OF/UR morphine ratios, noted by the arrow below the x-axis. The final 3 specimen pairs (SP 26–28), denoted with an asterisk (*), have very low OF morphine concentrations and undetectable UR morphine concentrations, which does not match the pattern seen with other specimens. All concentrations of drugs measured in UR are creatinine corrected.
hydromorphone and oxymorphone were more like secondary to morphine and oxycodone metabolism. Both morphine and oxycodone were readily detected in both matrices. Hydrocodone was only detected in 2 specimen pairs so we were unable to make a conclusion on the preferred specimen type. Of note, Heltsley et al. indicated that OF is the preferred specimen type for hydrocodone with 40 OF positive specimens versus 32 UR positive specimens out of a total of 112 specimen pairs [22]. Dilution, by means of excessive fluid intake prior to specimen collection or the addition of non-native fluids to the specimen, is a common UR adulteration technique that leads to artificially lower to undetectable drug levels. We identified 21 discrepant UR−/OF+ specimens (target population) and investigated whether these UR specimens did not contain measurable drug or metabolite levels due to dilution by measuring UR creatinine concentrations. Only 1discrepant specimen pair exhibited a urine creatinine level < 20 mg/mL, the accepted cutoff for dilution [23]. With that said, we identified a higher frequency of UR creatinine levels below the accepted cutoff in samples
Fig. 4. Cumulative frequency graph and Kruskal-Wallis comparison of creatinine concentrations in target, reference, and control populations. The creatinine distribution between the 3 populations did not differ significantly, except that a higher frequency of urine creatinine levels < 50 mg/dL was observed in the reference population compared to the target and control populations.
80
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
Table 3 Detection of Prescribed Drugs and Metabolites in Urine (UR) and Oral Fluid (OF). Prescribed drug
Total
Clonazepam Alprazolam Lorazepam Diazepam
31 4 6 5
Hydromorphone Morphine
4 6
Oxycodone
6
Drug or metabolite detected
Detected in UR
7-Aminoclonazepam Alprazolam/Alpha-OH-Alprazolam Lorazepam Nordiazepam Oxazepam Hydromorphone Morphine Hydromorphone Oxycodone Oxymorphone
Detected in OF
n
%
n
%
27 4 5 4 4 2 4 1 4 6
84 100 83 80
21 3 0 4 2 1 4 0 4 3
68 75 0 80
50 83 100
25 67 67
Table 4 Drug Properties and Preferred Urine (UR) versus Oral Fluid (OF) Specimen Type for Detection. Drug or metabolite
OF:UR
Weak base
Glucoronidation
Preferred specimen type
Oxymorphone Hydromorphone
0.0005 2 UR+/OF+ 21 UR+/OF− 0.001 1 UR+/OF+ 8 UR+/OF− 3 UR+/OF+ 14 UR+/OF− 0.01 0.02 0.024 0.05 0.1 0.5
Yes Yes
Significant Significant
UR UR⁎
Yes No
Significant Significant
UR or OF UR
No
Significant
UR
No No No Yes Yes Yes
Negligible Negligible Negligible Negligible Negligible Negligible
UR UR UR UR UR OF
Morphine Lorazepam Oxazepam 7-Aminoclonazepam Alprazolam/Alpha-OH-Alprazolam Nordiazepam Codeine Oxycodone 6-Acetylmorphine
or or or or
OF OF OF OF
⁎ Note: UR is superior for detection of hydromorphone present secondary to morphine metabolism. There were not enough cases of prescribed hydromorphone in the study to determine whether UR or OF is preferred.
15 (UR−, OF+) and the opposite conclusion would be made (i.e.,7aminoclonazepam is statistically more likely to be detected in OF; p = 0.05). Similarly, our study detected lorazepam in both OF and UR due to our lower direct-to-LC-MS/MS cutoffs, and we were able to determine with statistical confidence that UR is the preferred matrix. Despite our lower UR detection limits for 6-AM (5 ng/mL) compared to Vindenes et al. (20 ng/mL) OF still had more positive samples compared to UR (27 to 13, respectively) [24,25]. However, our conclusions could be subject to change with the use of even more sensitive assays for both UR and OF. Finally, our study was the first to report the OF:UR ratios of opioids and benzodiazepines and use them to corroborate our preferred matrix findings. This study has several limitations. First, this study is limited by the number of drugs measured; patients may be using other drugs within those drug classes that may not be detected by the targeted LC-MS/MS assays used in this study. There are many factors that affect measured drug concentrations in UR and OF, including interpersonal variability in drug metabolism due to clinical scenario, drug-drug interactions, and polymorphisms in drug metabolizing enzymes, all of which were not addressed in this study nor taken into account when forming the conclusions of this investigation. Additionally, the device used to collect OF does not include a visual marker for sufficient volume collection; therefore, even though the duration of OF collection was set based on the manufacturer's recommendation, it is possible that in some cases inadequate fluid was collected. This could have decreased our sensitivity for detection of drugs in some oral fluid specimens. Lastly, though UR dilution was evaluated by measuring UR creatinine concentration as a measure of sample adulteration, there are no current biochemical tests to assess potential OF adulteration and thus was not examined.
from non-discrepant pairs (reference population) and the healthy (control) population. Overall, UR creatinine levels were low in all three study populations and there was no statistical difference between the groups. Our review of prescribed medications shows that most of the drugs detected in either matrix were from illicit use of benzodiazepines and/ or opioids not a prescription. When drugs are prescribed, they are more frequently detected in UR. We postulate that this may be because steady states were achieved with chronic use. The number was low for most medications with the exception of clonazepam. Clonazepam prescriptions were distributed throughout the 3 groups (i.e., 18 UR+/OF+, 8 UR+/OF−, 2 UR−/OF+) suggesting that a patient's prescription status could not explain the matrix preference. Our study design was one-patient/simultaneously collected UR and OF. This approach is not utilized as frequently as population-directed studies as it is more challenging. To our knowledge, there are only three other studies similar to ours that used this approach on 100 or more sample pairs, analyzed for ten or more opioids and benzodiazepines, and reported analyte-specific (not class) results using LC-MS/MS. [22,24,25] However, unlike these referenced studies that first screened by immunoassays using 200–300 ng/mL cutoffs, we analyzed specimens upfront by the most sensitive and specific method (i.e. LC-MS/ MS) using cutoffs of ≤50 ng/mL. The direct to definitive method approach used in this study eliminates the bias of the immunoassay screen. For example, we found that 7-aminoclonazepam was statistically more likely to be detected in UR, while the other studies were either equivocal [24,25] or had a very low number of positive samples [22]. If we simulate using an immunoassay with a 200 ng/mL 7-aminoclonazepam cutoff in UR and recalculate the results for 7-aminoclonazepam in Fig. 1, it would yield 35 (UR+, OF+), 6 (UR+, OF−), and
81
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
5. Conclusion Our study showed that all 5 benzodiazepines were more likely to be detected in UR with statistically higher rates seen with 7-aminoclonazepam, lorazepam, and oxazepam. 6-AM was statistically more likely to be detected in OF, while hydromorphone and oxymorphone were statistically more frequently detected in UR. OF should be considered if the risk of adulteration is high and use and/or misuse of benzodiazepines, hydromorphone, and oxymorphone is low. Additional studies are warranted to explore for which drugs and to what extent OF may be used as the preferred matrix for drug testing.
[12] [13] [14] [15]
[16]
References
[17] [18]
[1] Centers for Disease Control, Injury Prevention & Control: Opioid Overdose, Centers for Disease Control and Prevention, Atlanta, GA, 2016. [2] American Society of Addiction Medicine, Opioid Addiction 2016 Facts & Figures, American Society of Addiction Medicine, Chevy Chase, MD, 2016. [3] M.S. Gold, N.S. Miller, K. Stennie, C. Populla-Vardi, Epidemology of benzodiazepine use and dependence, Psychiatr. Ann. 25 (1995) 146–148. [4] C.M. Jones, J.K. McAninch, Emergency department visits and overdose deaths from combined use of opioids and benzodiazepines, Am. J. Prev. Med. 49 (4) (2015) 493–501. [5] R.N. Hansen, G. Oster, J. Edelsberg, G.E. Woody, S.D. Sullivan, Economic costs of nonmedical use of prescription opioids, Clin. J. Pain 27 (3) (2011) 194–202. [6] P.J. Christo, L. Manchikanti, X. Ruan, et al., Urine drug testing in chronic pain, Pain Physician 14 (2) (2011) 123–143. [7] L. Manchikanti, R. Manchukonda, V. Pampati, et al., Does random urine drug testing reduce illicit drug use in chronic pain patients receiving opioids? Pain Physician 9 (2) (2006) 123–129. [8] P.J. Jannetto, N.C Bratanow, W.A. Clarke, et al., Executive summary: American Association of Clinical Chemistry, Laboratory Medicine Practice Guidelines - using clinical laboratory tests to monitor drug therapy in pain management patients, J. Appl. Lab. Med. 02 (04) (2018) 429–526. [9] A.G. Verstraete, Detection times of drugs of abuse in blood, urine, and oral fluid, Ther. Drug Monit. 26 (2) (2004) 200–205. [10] K.R. Allen, Screening for drugs of abuse: which matrix, oral fluid or urine? Ann. Clin. Biochem. 48 (Pt 6) (2011) 531–541. [11] K. Nordal, E.L. Oiestad, A. Enger, A.S. Christophersen, V. Vindenes, Detection times
[19]
[20]
[21]
[22]
[23] [24]
[25]
82
of diazepam, clonazepam, and alprazolam in oral fluid collected from patients admitted to detoxification, after high and repeated drug intake, Ther. Drug Monit. 37 (4) (2015) 451–460. T. Conermann, A.R. Gosalia, A.J. Kabazie, et al., Utility of oral fluid in compliance monitoring of opioid medications, Pain Physician 17 (1) (2014) 63–70. D.J. Crouch, Oral fluid collection: the neglected variable in oral fluid testing, Forensic Sci. Int. 150 (2–3) (2005) 165–173. R.C. Wong, M. Tran, J.K. Tung, Oral fluid drug tests: effects of adulterants and foodstuffs, Forensic Sci. Int. 150 (2–3) (2005) 175–180. S.E. Melanson, D. Griggs, I. Bixho, T. Khaliq, J.G. Flood, 7-aminoclonazepam is superior to clonazepam for detection of clonazepam use in oral fluid by LC-MS/MS, Clin. Chim. Acta 455 (2016) 128–133. J.G. Flood, T. Khaliq, K.A. Bishop, D.A. Griggs, The new substance abuse and mental health services administration oral fluid cutoffs for cocaine and heroin-related analytes applied to an addiction medicine setting: important, unanticipated findings with LC-MS/MS, Clin. Chem. 62 (5) (2016) 773–780. C.R. Baselt, Disposition of Toxic Drugs and Chemicals in Man, 8th edition, 55(12) Biomedical Publications, 2009. A. Reiter, J. Hake, C. Meissner, J. Rohwer, H.J. Friedrich, M. Oehmichen, Time of drug elimination in chronic drug abusers. Case study of 52 patients in a "low-step" detoxification ward, Forensic Sci. Int. 119 (2) (2001) 248–253. V. Vindenes, A. Enger, K. Nordal, U. Johansen, A.S. Christophersen, E.L. Øiestad, Very long detection times after high and repeated intake of heroin and methadone, measured in oral fluid, Scand. J. Forensic Sci. 20 (2) (2015) 33–41. O. Beck, M. Bottcher, Paradoxical results in urine drug testing for 6-acetylmorphine and total opiates: implications for best analytical strategy, J. Anal. Toxicol. 30 (2) (2006) 73–79. M. von Euler, T. Villen, J.O. Svensson, L. Stahle, Interpretation of the presence of 6monoacetylmorphine in the absence of morphine-3-glucuronide in urine samples: evidence of heroin abuse, Ther. Drug Monit. 25 (5) (2003) 645–648. R. Heltsley, A. Depriest, D.L. Black, et al., Oral fluid drug testing of chronic pain patients. II. Comparison of paired oral fluid and urine specimens, J. Anal. Toxicol. 36 (2) (2012) 75–80. Substance Abuse and Mental Health Services Administration, Federal Register, 73 (228) (2008) (Section 3.4). V. Vindenes, H.M. Lund, W. Andresen, et al., Detection of drugs of abuse in simultaneously collected oral fluid, urine and blood from norwegian drug drivers, Forensic Sci. Int. 219 (1–3) (2012) 165–171. V. Vindenes, B. Yttredal, E.L. Oiestad, et al., Oral fluid is a viable alternative for monitoring drug abuse: detection of drugs in oral fluid by liquid chromatographytandem mass spectrometry and comparison to the results from urine samples from patients treated with methadone or buprenorphine, J. Anal. Toxicol. 35 (1) (2011) 32–39.
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Monitoring opioid and benzodiazepine use and abuse: Is oral fluid or urine the preferred specimen type?☆
T
⁎
Athena K. Petridesa,b, , Stacy E.F. Melansona,b, Michalis Kantartjisa,c, Rachel D. Led, Christiana A. Demetrioue,f, James G. Floodb,g a
Department of Pathology, Brigham and Women's Hospital, Boston, MA, United States Harvard Medical School, Boston, MA, United States c Department of Medicine, Brigham and Women's Hospital, Boston, MA, United States d University of Massachusetts Medical School, Worcester, MA, United States e The Cyprus Institute of Neurology and Genetics, Nicosia, Cyprus f The Cyprus School of Molecular Medicine, Nicosia, Cyprus g Department of Pathology, Massachusetts General Hospital, Boston, MA, United States b
A R T I C L E I N F O
A B S T R A C T
Keywords: Oral fluid drug testing Urine drug testing Opioid Benzodiazepine Liquid chromatography-tandem mass spectrometry Pain management
Background: Oral fluid (OF) has become an increasingly popular matrix to assess compliance in pain management and addiction settings as it reduces the likelihood of adulteration. However, drug concentrations and windows of detection are not as well studied in OF as in urine (UR). We compared the clinical utility and analytical performance of OF and UR as matrices for detecting common benzodiazepines and opioids. Methods: OF and UR concentrations of 5 benzodiazepines and 7 opioids were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS) in 263 paired OF and UR specimens. UR creatinine was measured and prescription medications were reviewed. Results: The benzodiazepines 7-aminoclonazepam, lorazepam, and oxazepam exhibited statistically higher detection rates in UR. For opioids, 6-AM was statistically more likely to be detected in OF, while hydromorphone and oxymorphone were statistically more likely to be detected in UR. Chemical properties including glucuronidation explain preferential detection in each matrix, not UR creatinine nor prescription status. Conclusion: We found that OF is the preferred matrix for 6-AM, while UR is preferred for 7-aminoclonazepam, lorazepam, oxazepam, hydromorphone, and oxymorphone. However, OF should be considered if the risk of adulteration is high and use and/or misuse of benzodiazepines, hydromorphone, and oxymorphone is low.
1. Introduction Between 2000 and 2015, half a million deaths were due to drug overdoses, and for the first time, in 2015, drug overdoses were the leading cause of accidental death in the United States, highlighting prescription drug misuse and addiction as a national issue in recent years [1,2]. Prescription opioids and benzodiazepines constituted nearly all (70% and 30%, respectively) of prescription overdose deaths in 2013, with deaths commonly involving both substances [1,3–5].As a consequence, substantial efforts have been made to prevent and treat substance abuse, including opioid-agonist medication-assisted treatment (OA-MAT) [1,2].
Routine and random drug testing as an adjunct to OA-MAT has provided an objective measure of compliance and treatment efficacy in both the pain management and addiction settings [6–8]. Historically, urine (UR) specimens have been used for drug monitoring. UR collection is non-invasive and drugs are present at higher concentrations for longer periods of time compared to serum [6–9]. Numerous studies have demonstrated the effectiveness of UR drug testing for monitoring compliance [6–8]. However, UR can be easily adulterated, particularly if collections are not observed, and patients with shy bladder or anuria may not be able to provide a specimen [10]. For this reason, the utility of oral fluid (OF) has been explored as a tool to assess compliance. OF collection significantly reduces the
Abbreviations: opioid-agonist medication-assisted treatment, (OA-MAT); oral fluid, (OF); urine, (UR); Massachusetts General Hospital, (MGH); liquid chromatography-tandem mass spectrometry, (LC-MS/MS); 6-acetylmorphine, (6-AM); mass spectrometer, (MS); Brigham and Women's Hospital, (BWH); heated electrospray ionization, (HESI); Substance Abuse and Mental Health Services Administration, (SAMHSA) ☆ This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. ⁎ Corresponding author at: Brigham and Women's Hospital, 75 Francis Street, Amory 2, Boston, MA 02115, United States. E-mail address: [email protected] (A.K. Petrides). https://doi.org/10.1016/j.cca.2018.02.034 Received 22 December 2017; Received in revised form 7 February 2018; Accepted 26 February 2018 Available online 27 February 2018 0009-8981/ © 2018 Published by Elsevier B.V.
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
likelihood of adulteration. In recent years, more studies have investigated OF in monitoring drugs of abuse with promising results. Nordal et al. reported that benzodiazepines can be measured qualitatively in OF and suggested OF may be an alternative for detection of clonazepam, diazepam, and alprazolam [11]. Likewise, Conermann et al. concluded that OF may be used for monitoring compliance of opioids and benzodiazepines after analyzing 132 paired specimens [12]. OF has some analytical challenges including low sample volumes and difficult specimen collections in patients with conditions such as dry mouth. Furthermore, collections are typically performed by clinic staff and can be time-consuming as they involve the patient rinsing their mouth and waiting 10–15 min prior to collection. Additionally, OF specimens may become contaminated with mouthwash, toothpaste, foods, and beverages [13,14]. Buccal contamination of OF by placing the drug sublingually immediately prior to collection may also pose a challenge for drug compliance assessment. Although OF has been studied less extensively, current literature demonstrates that OF exhibits lower drug and metabolite concentrations and narrower windows of detection compared to UR for the majority of drugs [9]. There are a limited number of studies directly comparing the performance of paired OF and UR specimens for both opioids and benzodiazepines. Additionally, to our knowledge, there are no published studies reporting the ratio of OF to UR concentrations for opioids and benzodiazepines in matched OF and UR specimens. In this study, we examined the clinical and analytical performance of each matrix and make suggestions on the utility of each matrix.
Table 1 Limit of detection (LOD) for urine (UR) versus oral fluid (OF) liquid chromatographytandem mass spectrometry (LC-MS/MS) Testing. Class
Benzodiazepines
Opioids
Drug or metabolite
7-Aminoclonazepam Alpha-OH-Alprazolam Alprazolam Lorazepam Nordiazepam Oxazepam 6-Acetylmorphine Codeine Hydrocodone Hydromorphone Morphine Oxycodone Oxymorphone
LOD (ng/mL) for LC-MS/MS UR
OF
50 50 N/A 50 50 50 5 50 50 50 50 50 50
1 N/A 2 2 2 2 2 2 2 4 2 1 2
2.3. Urine drug analysis UR was analyzed at the Brigham and Women's Hospital (BWH) Clinical Chemistry Laboratory (Boston, MA) using a laboratory-developed LC-MS/MS method. Samples were prepared by adding an internal deuterated standard to the following drugs or metabolites: 7-aminoclonazepam, alpha-hydroxy-alprazolam, lorazepam, nordiazepam, oxazepam, 6-AM, codeine, hydrocodone, hydromorphone, morphine, oxycodone, and oxymorphone. Samples were diluted and subjected to a hydrolysis step to remove glucuronide and sulfate groups. Chromatographic separation was achieved on a ACQUITY UPLC I-Class (Waters, Milford, MA) using a Kinetex C18 analytical column (Phenomenex Inc., Torrance, CA) and mass spectrometric analysis was performed on a tandem triple quadrupole Xevo TQS (Waters, Milford, MA) preceded by HESI. 6-AM was analyzed with the same equipment, but was not subjected to a hydrolysis step. The limits of detection (LOD) for each drug or metabolite in UR are also described in Table 1.
2. Methods 2.1. Specimen acquisition A total of 263 paired OF and UR specimens were collected consecutively from 140 unique patients at the Massachusetts General Hospital (MGH) addiction-psychiatry clinics during routine visits and processed at the MGH Clinical Chemistry Laboratory (Boston, MA). For OF collection, the Orasure Intercept Sample Collection Device (Orasure Technologies, Bethlehem, PA) was utilized according to the manufacturer's collection instructions. OF and UR pairs were received by the laboratory within 2–10 h of collection. OF was refrigerated overnight and tested the next day. UR was frozen within 8 h of receipt and tested in batches at a later date. The Partners Human Research Committee approved this study.
2.4. Calculation of oral fluid urine ratios UR creatinine was measured using the rate-blanked Jaffe reaction with Roche Diagnostics reagents on a Roche Cobas e501 (Roche Diagnostics, Indianapolis IN). Quantitative UR drug measurements for 7-aminoclonazepam, alprazolam/alpha-hydroxy-alprazolam, nordiazepam, 6-AM, codeine, morphine, oxycodone, and oxymorphone were corrected for creatinine levels using the following formula: [Drug in ng/ mL]/[UR Creatinine in mg/dL]×100. Corrected drug concentrations were used to calculate oral fluid:urine ratios (OF:UR) using the following formula: [OF Drug in ng/mL]/[UR Drug in ng/mL]. Ranges and medians were calculated.
2.2. Oral fluid drug analysis As published previously, OF was analyzed at MGH using a laboratory-developed liquid chromatography-tandem mass spectrometry (LCMS/MS) method [15,16]. The following benzodiazepines and opioids were detected: 7-aminoclonazepam, alprazolam, lorazepam, nordiazepam, oxazepam, 6-acetylmorphine (6-AM), codeine, hydrocodone, hydromorphone, morphine, oxycodone, and oxymorphone. Briefly, OF specimens were mixed with Internal Standard Solution containing deuterated analogues of each analyte. The mixture was injected onto a TLX2 chromatograph (Thermo Scientific, Waltham, MA) where the analytes and internal standards were first isolated on a Cyclone-P turbulent-flow extraction column (Thermo-Fisher, Franklin, MA) and then transferred to an Ascentis Phenyl analytical column (Supelco, Bellefonte, PA). The analytes were then separated using a gradient elution program and detected using a Thermo Quantum Ultra triple quadrupole MS (Thermo Scientific, Waltham, MA) equipped with a heated electrospray interface (HESI-II) operated in the positive ion mode. The limits of detection (LOD) for each drug or metabolite in OF are described in Table 1.
2.5. Medication review The electronic health records of all patients in the study were reviewed for any active prescriptions/medications at the time of specimen collection. In this study, a prescription was considered active if the start date was before specimen collection and the end date was after or within 5 days of specimen collection. 2.6. Statistical analysis The non-parametric McNemar's symmetry test was used to assess agreement between paired OF and UR specimens. A p-value < 0.05 was considered to be statistically significant. To assess for dilution of UR specimens, creatinine concentrations of UR specimens were calculated. The UR creatinine distribution in 76
Clinica Chimica Acta 481 (2018) 75–82 Fig. 1. Positive Benzodiazepines in Urine (UR) versus Oral Fluid (OF). Light gray bars represent the number of paired specimens for which drug/metabolites were positive in both urine (UR) and oral fluid (OF) (i.e., UR+/OF+). Dark gray bars represent specimens for which drug/metabolites were positive in only UR (i.e., UR+/OF−). Black bars represent specimens for which drug/metabolites were positive in only OF (i.e., UR−/OF+). The results of the McNemar's symmetry test are displayed next to each drug/metabolite.
A.K. Petrides et al.
specimens for which opioids and/or benzodiazepines were detected in only OF (i.e, target population) was compared to the UR creatinine distribution of a) samples negative for both opioids and benzodiazepines in both OF and UR (i.e., reference population, n = 123) and b) randomly-selected samples from an independent, healthy population (i.e., control population, n = 263). The creatinine distributions were compared visually by a cumulative frequency distribution graph and numerically by the non-parametric Kruskal-Wallis test. All analyses were performed using STATA version SE12 and GraphPad Prism 6. 3. Results Of the 263 paired specimens, 123 (47%) were negative for benzodiazepines and opioids in both OF and UR (i.e., UR-/OF-). The remaining 140 (53%) were positive for 1 or more benzodiazepines and/or opioids in either OF and/or UR. 3.1. Benzodiazepines The detection rate for benzodiazepines was higher in UR with 33% of samples positive for 1 or more benzodiazepine(s) compared to 21% of OF specimens. Furthermore, 7-aminoclonazepam, lorazepam, and oxazepam were statistically more likely to be detected in UR versus OF (p = 0.002, p = 0.008, p < 0.001, respectively; Fig. 1). Of the specimens positive for 7-aminoclonazepam, lorazepam, and oxazepam, 28% (n = 19), 89% (n = 8), and 82% (n = 14), respectively, were detected only in UR (i.e., UR+/OF−). No specimens positive for lorazepam or oxazepam were positive in OF without also being positive in UR (i.e., UR+/OF+). The number of specimens positive only in OF (i.e., UR−/ OF+) was low for 7-aminoclonazepam (n = 4), alprazolam/alpha-OHalprazolam (n = 1), and nordiazepam (n = 2). Nonetheless, alprazolam and nordiazepam were not statistically more likely to be detected in UR versus OF (p = 0.07, p > 0.999, respectively; Fig. 1). 3.2. Opioids The detection rate for opioids was slightly higher in OF with 32% of samples positive for 1 or more opioids in OF compared to 28% in UR. 6AM was statistically more likely to be detected in OF (p < 0.001), while hydromorphone and oxymorphone were statistically more likely to be detected in UR (p < 0.001, p < 0.004, respectively; Fig. 2). Oxycodone was preferentially detected in OF with 4 UR−/OF+ specimens and was never detected in only UR (i.e., UR+/OF−), though the results were not statistically significant (p = 0.125). Codeine and morphine were not preferentially detected in either specimen types, having been detected in both or either matrix (p = 0.782, p = 0.336, respectively). Hydrocodone was detected in only 2 samples, 1 UR+/OF + and 1 UR+/OF− (p = 0.999; Fig. 2). The detection of morphine, hydromorphone, and oxymorphone can be a result of heroin (6-AM), morphine, and oxycodone metabolism, respectively. 6-AM was preferentially detected in OF with 15 positive specimens detected in only OF (Fig. 2). There was 1 paired specimen in which 6-AM was detected only in UR. Furthermore, 2 UR specimens positive for 6-AM did not contain detectable UR morphine levels (data not shown). Of the 21 UR+/OF− specimens for hydromorphone, 20 were also UR+/OF+ for morphine. Both of the 2 UR+/OF+ for hydromorphone were UR−/OF− for morphine. Lastly, of the 9 UR +/OF− for oxymorphone, 3 were also UR+/OF+ for oxycodone. Likewise, all 5 UR+/OF+ specimens for oxymorphone were also UR +/OF+ for oxycodone. 3.3. Drug concentrations and ratios Concentrations and OF:UR ratios for all drugs and metabolites with 5 or more positive samples are shown in Table 2. Overall, wide ranges in concentrations were seen in both UR and OF. 6-AM, oxycodone, and 77
Fig. 2. Positive Opioids in Urine (UR) versus Oral Fluid (OF). Light gray bars represent the number of paired specimens for which drug/metabolites was positive in both urine (UR) and oral fluid (OF) (i.e., UR+/OF+). Dark gray bars represent specimens for which drug/metabolites were positive in only UR (i.e., UR+/OF−). Black bars represent specimens for which drug/metabolites were positive in only OF (i.e., UR−/OF+). The results of the McNemar's symmetry test are displayed next to each drug/metabolite.
A.K. Petrides et al.
Clinica Chimica Acta 481 (2018) 75–82
78
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
detected in both matrices (i.e., UR+/OF+), 8 were detected in UR only (i.e., UR+/OF−), 2 were detected in OF only (i.e., UR−/OF+) and 3 were not detected in either matrix.
Table 2 Description of opioid and benzodiazepine quantitative findings in urine (UR) versus oral fluid (OF) liquid chromatography-tandem mass spectrometry (LC-MS/MS) testing. Drug or metabolite (n)⁎
Oxymorphone (5) Morphine (31) 7-Aminoclonazepam (45) Alprazolam/Alpha-OHAlprazolam (8) Nordiazepam (6) Codeine (10) Oxycodone (8) 6-Acetylmorphine (10)
⁎
UR LC-MS/MS (ng/ mL) median (ranges)
OF LC-MS/MS (ng/mL) median (ranges)
OF:UR ratio median (ranges)
9731 (1111–45,744) 17,100 (31–83,586) 519 (21–3602) 570 (97–3101) 1045 (290–1304) 468 (198–854) 6887 (309–18,847) 241 (2.4–3404)
5 (2–15) 24 (2–1641) 5 (1–95.7) 9 (2–67) 24 (8–73) 18 (3–34) 476 (27–1201) 88 (2.4–431)
0.0005 (0.0002–0.002) 0.001 (0.0001–0.31) 0.01 (0.0007–0.07) 0.02 (0.006–0.05) 0.024 (0.008–0.1) 0.05 (0.004–0.09) 0.1 (0.03–0.22) 0.5 (0.006–3.6)
4. Discussion We measured the concentrations of 5 benzodiazepines and 7 opioids in 263 paired UR and OF specimens. Our study found that all 5 benzodiazepines were more frequently detected in UR, with 7-aminoclonazepam, lorazepam, and oxazepam statistically more likely to be detected in UR versus OF. Conversely, the overall detection rate for opioids was slightly higher in OF than UR with 6-AM statistically more likely to be detected in OF. Hydromorphone and oxymorphone were statistically more likely to be detected in UR. The chemical properties of drugs including acidity and glucuronidation govern to what extent they are detected in OF and/or UR (Table 4) [17]. Passive diffusion is the most common mechanism of drug transfer to OF. Weakly basic, non-protein bound molecules with lower molecular weights are more likely to be present in OF. Also, OF exhibits a slightly acidic pH and attracts weak bases that ionize and become trapped in the acidic environment. Drugs that are converted to more soluble forms, such as glucuronides, are more likely to be detectable in UR. For the same reason, hydroxylated metabolites, not the primary drug, are predominately found in UR. Consistent with these properties, our results show that concentrations of all drugs were lower in OF than UR (Table 2). In addition, we found that weakly basic, nonglucuronidated analytes like 6-AM and oxycodone were more frequently detected in OF and had the highest OF:UR ratios (Table 4). Non-conjugated analytes with intermediate OF:UR ratios such as codeine, nordiazepam, and alprazolam were equally detected in both matrices. Lastly, highly glucuronidated analytes such as oxymorphone, hydromorphone, oxazepam and lorazepam were most frequently detected in UR and had the lowest OF:UR ratios. Morphine is an exception, as it has a low OF:UR ratio and is highly conjugated, yet it is detected equally well in OF and UR. Morphine had the highest average concentrations of any analyte in both OF and UR. This is likely due to the fact that high doses of morphine are prescribed or ingested, which lead to detectable concentrations in OF despite the high degree of glucuronidation. The range of drug concentrations is infrequently published. Drugs with narrow OF:UR ranges (i.e., 101–102) such as oxymorphone and 7aminoclonazepam showed matrix preference based on their chemical properties as mentioned above. However, 6-AM and morphine were found to have wide ranges of concentrations (i.e., 103) in both matrices, which may imply that the detection of these substances is highly-dependent on the time of drug use and collection. Particularly for 6-AM positive specimens we observed that the relationship between OF and UR morphine is inversely proportional. As the concentration of OF morphine decreases, the concentration of UR morphine increases suggesting more remote use. 6-AM is thought to be detected in OF and UR up to 8 h and 34.5 h after use, respectively [9,18]. However, a study by Vindenes et al. showed that 6-AM was detected in OF 4–8 days after use with undetectable UR and OF morphine levels, thereby challenging these reported detection windows [19]. Our results seem to support Vindenes et al.'s observation. We had 3 patients that seem to match that same profile, with detectable 6-AM levels and very low concentrations of morphine OF and undetectable levels in UR (sp. 26–28, Fig. 3). Additionally, our study corroborates recent literature showing that 6AM is detectable in UR with no detectable morphine [19–21]. Although hydromorphone and oxymorphone were statistically more likely to be detected in UR, the clinical significance of this finding is less relevant in our patient population where only 4 patients were prescribed hydromorphone and none prescribed oxymorphone. Hydromorphone was present in combination with morphine in 87% (20/23) of specimens and oxymorphone with oxycodone in 57% (8/14) of specimens, suggesting that the specimens positive for
Note: Only drugs/metabolites with n ≥ 5 are listed.
codeine had the highest OF:UR ratios at 0.5, 0.1, and 0.05 respectively, oxymorphone and morphine had the lowest ratios of 0.0005 and 0.001, respectively, whereas the three benzodiazepines had intermediate ratios. The largest range in OF:UR ratios was seen in specimens positive for morphine and 6-AM, both to the order of 103 (Table 2). Otherwise, the remaining ranges were to the order of 101 except for 7-aminoclonazepam, which had a range to the order of 102. For the majority of patients, as the OF:UR ratio for morphine decreases, UR 6-AM concentration increases while OF 6-AM concentration decreases (Fig. 3). There were a total of 6 patients with detectable 6-AM in either OF or UR and detectable morphine in OF, but undetectable morphine in UR. 3 of those patients had high concentrations of morphine in OF (SP. 1–3, Fig. 3), whereas the other 3 patients had very low concentrations morphine in OF (SP. 26–28, Fig. 3). 3.4. Urine creatinine To assess whether dilution of UR specimens may have altered drug concentrations, we quantified creatinine concentrations in UR specimens. The UR creatinine distribution in specimens in which benzodiazepines and/or opioids were detected only in OF (i.e., UR−/OF+) was similar to the UR creatinine distribution of both the reference and control populations (p = 0.6232) (Fig. 4). There was only 1 specimen in the target population with a creatinine concentration considered dilute according to SAMSHA guidelines (i.e., < 20 mg/dL). Additionally, only 46.9% (n = 15) of UR−/OF+ specimens had creatinine concentrations < 100 mg/dL. Interestingly, the percent of samples with creatinine levels < 20 mg/dL was higher in the reference population (8.9%) than control (1.9%) and target (3.1%) populations. 3.5. Medication review Lastly, we investigated whether the drugs detected were prescribed. Only 24% (n = 62) of paired specimens were from patients who were prescribed 1 or more of the medications measured. Overall, prescribed drugs were detected at rates between 80 and 100% in UR for all drugs except hydromorphone which was detected at a rate of 50% in UR (Table 3). Moreover, prescribed drugs were detected at higher rates in UR versus OF in all drugs except diazepam, which was detected with equal frequency (80%). Clonazepam was prescribed most frequently. Of the 31 specimens associated with a clonazepam prescription, 18 were 79
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
Fig. 3. Pattern of 6-AM and Morphine Concentrations in Urine (UR) and Oral Fluid (OF). 6-AM concentrations in OF (black line) and UR (shaded gray line) for each paired specimen (SP 1–28) are plotted against their OF/UR morphine concentration ratio. Concentrations are displayed in order of decreasing OF/UR morphine ratios, noted by the arrow below the x-axis. The final 3 specimen pairs (SP 26–28), denoted with an asterisk (*), have very low OF morphine concentrations and undetectable UR morphine concentrations, which does not match the pattern seen with other specimens. All concentrations of drugs measured in UR are creatinine corrected.
hydromorphone and oxymorphone were more like secondary to morphine and oxycodone metabolism. Both morphine and oxycodone were readily detected in both matrices. Hydrocodone was only detected in 2 specimen pairs so we were unable to make a conclusion on the preferred specimen type. Of note, Heltsley et al. indicated that OF is the preferred specimen type for hydrocodone with 40 OF positive specimens versus 32 UR positive specimens out of a total of 112 specimen pairs [22]. Dilution, by means of excessive fluid intake prior to specimen collection or the addition of non-native fluids to the specimen, is a common UR adulteration technique that leads to artificially lower to undetectable drug levels. We identified 21 discrepant UR−/OF+ specimens (target population) and investigated whether these UR specimens did not contain measurable drug or metabolite levels due to dilution by measuring UR creatinine concentrations. Only 1discrepant specimen pair exhibited a urine creatinine level < 20 mg/mL, the accepted cutoff for dilution [23]. With that said, we identified a higher frequency of UR creatinine levels below the accepted cutoff in samples
Fig. 4. Cumulative frequency graph and Kruskal-Wallis comparison of creatinine concentrations in target, reference, and control populations. The creatinine distribution between the 3 populations did not differ significantly, except that a higher frequency of urine creatinine levels < 50 mg/dL was observed in the reference population compared to the target and control populations.
80
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
Table 3 Detection of Prescribed Drugs and Metabolites in Urine (UR) and Oral Fluid (OF). Prescribed drug
Total
Clonazepam Alprazolam Lorazepam Diazepam
31 4 6 5
Hydromorphone Morphine
4 6
Oxycodone
6
Drug or metabolite detected
Detected in UR
7-Aminoclonazepam Alprazolam/Alpha-OH-Alprazolam Lorazepam Nordiazepam Oxazepam Hydromorphone Morphine Hydromorphone Oxycodone Oxymorphone
Detected in OF
n
%
n
%
27 4 5 4 4 2 4 1 4 6
84 100 83 80
21 3 0 4 2 1 4 0 4 3
68 75 0 80
50 83 100
25 67 67
Table 4 Drug Properties and Preferred Urine (UR) versus Oral Fluid (OF) Specimen Type for Detection. Drug or metabolite
OF:UR
Weak base
Glucoronidation
Preferred specimen type
Oxymorphone Hydromorphone
0.0005 2 UR+/OF+ 21 UR+/OF− 0.001 1 UR+/OF+ 8 UR+/OF− 3 UR+/OF+ 14 UR+/OF− 0.01 0.02 0.024 0.05 0.1 0.5
Yes Yes
Significant Significant
UR UR⁎
Yes No
Significant Significant
UR or OF UR
No
Significant
UR
No No No Yes Yes Yes
Negligible Negligible Negligible Negligible Negligible Negligible
UR UR UR UR UR OF
Morphine Lorazepam Oxazepam 7-Aminoclonazepam Alprazolam/Alpha-OH-Alprazolam Nordiazepam Codeine Oxycodone 6-Acetylmorphine
or or or or
OF OF OF OF
⁎ Note: UR is superior for detection of hydromorphone present secondary to morphine metabolism. There were not enough cases of prescribed hydromorphone in the study to determine whether UR or OF is preferred.
15 (UR−, OF+) and the opposite conclusion would be made (i.e.,7aminoclonazepam is statistically more likely to be detected in OF; p = 0.05). Similarly, our study detected lorazepam in both OF and UR due to our lower direct-to-LC-MS/MS cutoffs, and we were able to determine with statistical confidence that UR is the preferred matrix. Despite our lower UR detection limits for 6-AM (5 ng/mL) compared to Vindenes et al. (20 ng/mL) OF still had more positive samples compared to UR (27 to 13, respectively) [24,25]. However, our conclusions could be subject to change with the use of even more sensitive assays for both UR and OF. Finally, our study was the first to report the OF:UR ratios of opioids and benzodiazepines and use them to corroborate our preferred matrix findings. This study has several limitations. First, this study is limited by the number of drugs measured; patients may be using other drugs within those drug classes that may not be detected by the targeted LC-MS/MS assays used in this study. There are many factors that affect measured drug concentrations in UR and OF, including interpersonal variability in drug metabolism due to clinical scenario, drug-drug interactions, and polymorphisms in drug metabolizing enzymes, all of which were not addressed in this study nor taken into account when forming the conclusions of this investigation. Additionally, the device used to collect OF does not include a visual marker for sufficient volume collection; therefore, even though the duration of OF collection was set based on the manufacturer's recommendation, it is possible that in some cases inadequate fluid was collected. This could have decreased our sensitivity for detection of drugs in some oral fluid specimens. Lastly, though UR dilution was evaluated by measuring UR creatinine concentration as a measure of sample adulteration, there are no current biochemical tests to assess potential OF adulteration and thus was not examined.
from non-discrepant pairs (reference population) and the healthy (control) population. Overall, UR creatinine levels were low in all three study populations and there was no statistical difference between the groups. Our review of prescribed medications shows that most of the drugs detected in either matrix were from illicit use of benzodiazepines and/ or opioids not a prescription. When drugs are prescribed, they are more frequently detected in UR. We postulate that this may be because steady states were achieved with chronic use. The number was low for most medications with the exception of clonazepam. Clonazepam prescriptions were distributed throughout the 3 groups (i.e., 18 UR+/OF+, 8 UR+/OF−, 2 UR−/OF+) suggesting that a patient's prescription status could not explain the matrix preference. Our study design was one-patient/simultaneously collected UR and OF. This approach is not utilized as frequently as population-directed studies as it is more challenging. To our knowledge, there are only three other studies similar to ours that used this approach on 100 or more sample pairs, analyzed for ten or more opioids and benzodiazepines, and reported analyte-specific (not class) results using LC-MS/MS. [22,24,25] However, unlike these referenced studies that first screened by immunoassays using 200–300 ng/mL cutoffs, we analyzed specimens upfront by the most sensitive and specific method (i.e. LC-MS/ MS) using cutoffs of ≤50 ng/mL. The direct to definitive method approach used in this study eliminates the bias of the immunoassay screen. For example, we found that 7-aminoclonazepam was statistically more likely to be detected in UR, while the other studies were either equivocal [24,25] or had a very low number of positive samples [22]. If we simulate using an immunoassay with a 200 ng/mL 7-aminoclonazepam cutoff in UR and recalculate the results for 7-aminoclonazepam in Fig. 1, it would yield 35 (UR+, OF+), 6 (UR+, OF−), and
81
Clinica Chimica Acta 481 (2018) 75–82
A.K. Petrides et al.
5. Conclusion Our study showed that all 5 benzodiazepines were more likely to be detected in UR with statistically higher rates seen with 7-aminoclonazepam, lorazepam, and oxazepam. 6-AM was statistically more likely to be detected in OF, while hydromorphone and oxymorphone were statistically more frequently detected in UR. OF should be considered if the risk of adulteration is high and use and/or misuse of benzodiazepines, hydromorphone, and oxymorphone is low. Additional studies are warranted to explore for which drugs and to what extent OF may be used as the preferred matrix for drug testing.
[12] [13] [14] [15]
[16]
References
[17] [18]
[1] Centers for Disease Control, Injury Prevention & Control: Opioid Overdose, Centers for Disease Control and Prevention, Atlanta, GA, 2016. [2] American Society of Addiction Medicine, Opioid Addiction 2016 Facts & Figures, American Society of Addiction Medicine, Chevy Chase, MD, 2016. [3] M.S. Gold, N.S. Miller, K. Stennie, C. Populla-Vardi, Epidemology of benzodiazepine use and dependence, Psychiatr. Ann. 25 (1995) 146–148. [4] C.M. Jones, J.K. McAninch, Emergency department visits and overdose deaths from combined use of opioids and benzodiazepines, Am. J. Prev. Med. 49 (4) (2015) 493–501. [5] R.N. Hansen, G. Oster, J. Edelsberg, G.E. Woody, S.D. Sullivan, Economic costs of nonmedical use of prescription opioids, Clin. J. Pain 27 (3) (2011) 194–202. [6] P.J. Christo, L. Manchikanti, X. Ruan, et al., Urine drug testing in chronic pain, Pain Physician 14 (2) (2011) 123–143. [7] L. Manchikanti, R. Manchukonda, V. Pampati, et al., Does random urine drug testing reduce illicit drug use in chronic pain patients receiving opioids? Pain Physician 9 (2) (2006) 123–129. [8] P.J. Jannetto, N.C Bratanow, W.A. Clarke, et al., Executive summary: American Association of Clinical Chemistry, Laboratory Medicine Practice Guidelines - using clinical laboratory tests to monitor drug therapy in pain management patients, J. Appl. Lab. Med. 02 (04) (2018) 429–526. [9] A.G. Verstraete, Detection times of drugs of abuse in blood, urine, and oral fluid, Ther. Drug Monit. 26 (2) (2004) 200–205. [10] K.R. Allen, Screening for drugs of abuse: which matrix, oral fluid or urine? Ann. Clin. Biochem. 48 (Pt 6) (2011) 531–541. [11] K. Nordal, E.L. Oiestad, A. Enger, A.S. Christophersen, V. Vindenes, Detection times
[19]
[20]
[21]
[22]
[23] [24]
[25]
82
of diazepam, clonazepam, and alprazolam in oral fluid collected from patients admitted to detoxification, after high and repeated drug intake, Ther. Drug Monit. 37 (4) (2015) 451–460. T. Conermann, A.R. Gosalia, A.J. Kabazie, et al., Utility of oral fluid in compliance monitoring of opioid medications, Pain Physician 17 (1) (2014) 63–70. D.J. Crouch, Oral fluid collection: the neglected variable in oral fluid testing, Forensic Sci. Int. 150 (2–3) (2005) 165–173. R.C. Wong, M. Tran, J.K. Tung, Oral fluid drug tests: effects of adulterants and foodstuffs, Forensic Sci. Int. 150 (2–3) (2005) 175–180. S.E. Melanson, D. Griggs, I. Bixho, T. Khaliq, J.G. Flood, 7-aminoclonazepam is superior to clonazepam for detection of clonazepam use in oral fluid by LC-MS/MS, Clin. Chim. Acta 455 (2016) 128–133. J.G. Flood, T. Khaliq, K.A. Bishop, D.A. Griggs, The new substance abuse and mental health services administration oral fluid cutoffs for cocaine and heroin-related analytes applied to an addiction medicine setting: important, unanticipated findings with LC-MS/MS, Clin. Chem. 62 (5) (2016) 773–780. C.R. Baselt, Disposition of Toxic Drugs and Chemicals in Man, 8th edition, 55(12) Biomedical Publications, 2009. A. Reiter, J. Hake, C. Meissner, J. Rohwer, H.J. Friedrich, M. Oehmichen, Time of drug elimination in chronic drug abusers. Case study of 52 patients in a "low-step" detoxification ward, Forensic Sci. Int. 119 (2) (2001) 248–253. V. Vindenes, A. Enger, K. Nordal, U. Johansen, A.S. Christophersen, E.L. Øiestad, Very long detection times after high and repeated intake of heroin and methadone, measured in oral fluid, Scand. J. Forensic Sci. 20 (2) (2015) 33–41. O. Beck, M. Bottcher, Paradoxical results in urine drug testing for 6-acetylmorphine and total opiates: implications for best analytical strategy, J. Anal. Toxicol. 30 (2) (2006) 73–79. M. von Euler, T. Villen, J.O. Svensson, L. Stahle, Interpretation of the presence of 6monoacetylmorphine in the absence of morphine-3-glucuronide in urine samples: evidence of heroin abuse, Ther. Drug Monit. 25 (5) (2003) 645–648. R. Heltsley, A. Depriest, D.L. Black, et al., Oral fluid drug testing of chronic pain patients. II. Comparison of paired oral fluid and urine specimens, J. Anal. Toxicol. 36 (2) (2012) 75–80. Substance Abuse and Mental Health Services Administration, Federal Register, 73 (228) (2008) (Section 3.4). V. Vindenes, H.M. Lund, W. Andresen, et al., Detection of drugs of abuse in simultaneously collected oral fluid, urine and blood from norwegian drug drivers, Forensic Sci. Int. 219 (1–3) (2012) 165–171. V. Vindenes, B. Yttredal, E.L. Oiestad, et al., Oral fluid is a viable alternative for monitoring drug abuse: detection of drugs in oral fluid by liquid chromatographytandem mass spectrometry and comparison to the results from urine samples from patients treated with methadone or buprenorphine, J. Anal. Toxicol. 35 (1) (2011) 32–39.