Can an Immunoassay Become a Standard Technique in Detecting Oxycodone and Its Metabolites?

Transcription

1 Journal of Analytical Toxicology, Vot. 29, November/December 2005 Technical Note ] Can an Immunoassay Become a Standard Technique in Detecting Oxycodone and Its Metabolites? Jude M. Abadie*, Kim H. Allison, David A. Black, James Garbin, Andrew J. Saxon, and Daniel D. Bankson Departments of Pathology and Laboratory Medicine, Veterans Affairs Puget Sound Health Care System and University of Washington Medical Center, Seattle, Washington L Abstract ] Opiate toxicology testing is routinely performed in the hospital setting to identify abusers and/or to determine those patients who are not taking prescribed opiate analgesics such as oxycodone. Commercially available assays for opiate detection in urine have decreased sensitivity for oxycodone, which contributes to a high false-negative rate. Functioning as a beta site, our Veterans Affairs hospital evaluated a new enzyme immunoassay, DRI Oxycodone Assay, for its use in the qualitative and semiquantitative detection of oxycodone in urine. We hypothesize that an immunoassay for oxycodone with superior sensitivity and specificity, when compared to the traditional opiate assays, would reduce the need for more expensive and time-consuming confirmatory testing. We used the new liquid homogenous enzyme immunoassay to determine oxycodone results in a total of 148 urine samples from 4 different sample groups. Gas chromatography-mass spectroscopy was subsequently used to confirm the presence or absence of oxycodone (or its primary metabolite, noroxycodone). We also evaluated within-run, between-run, and linearity studies and conducted a crossover study to establish a cutoff value for oxycodone, in our patient population, we used the new DRI immunoassay to evaluate 17,069 urine samples to estimate oxycodone misuse profiles (patients not taking prescribed oxycodone or taking oxycodone without a prescription) during a 4-month period. The sensitivity and specificity of the new oxycodone immunoassay were 97.7% and 100%, respectively, at the cutoff concentration of 300 ng/ml. The assay linearity was 1250 ng/ml, and the sensitivity was 10 ng/ml. Within-run precision and between-run coefficient of variation were 2.3% and 1.8%, respectively. None of the 15 compounds that we evaluated for interference had crossover significant enough to produce a positive oxycodone result when using 300 ng/ml as the cutoff value. None of the 17,069 oxycodone immunoassays was followed with a request for confirmation. Among patients with positive results (n = 224), 93 (41.5%) were not prescribed oxycodone. The new DRI Oxycodone Assay is a sensitive and specific screening test for the determination of oxycodone. The improved opiate screening results may lead to better patient * Author to whom correspondence should be addressed: University of Washington Medical Center, 1959 NW Pacific Street, Department of Laboratory Medicine (NW 120; Box ), Seattle, WA judeabadie@medscape.com. and prescription management, to decreased laboratory spending, and to the identification of oxycodone abusers, which could result in decreased oxycodone-related mortality. Introduction Oxycodone HCl [14-hydroxydihydrocodeinone] is a thebainederived opium alkaloid p-opioid agonist that has analgesic potency almost equivalent to morphine. Oxycodone has long been available in lower dose formulations ( rag) used in combination with aspirin (Percodan ~) or acetaminophen (Percocet~). In 1995, following the Food and Drug Administration approval, Purdue Pharma introduced a time-release preparation, OxyContin ~, in 10-, 20-, 40-, 80-, and 160-rag doses; however, the 160-mg tablets have been removed from the market (1). Because these formulations allow slow release of oxycodone, it could be dosed every 12 h instead of the more frequent dosing required with other forms. Therefore, OxyContin quickly became the drug of choice for chronic pain management and is currently the most commonly prescribed Schedule II controlled narcotic in the United States (2). The widespread availability of oxycodone quickly contributed to its "street" sale and frequent abuse. Oxycodone has a street value that is approximately 10 times higher than its prescription purchase cost (3). Abusers quickly learn to achieve maximum bioavailability by chewing the tablets, crushing and snorting them, or dissolving and injecting the oxycodone in order to quickly release the entire dose. Activation of p-receptors by oxycodone causes euphoria and analgesia. The primary cause of death in overdose is respiratory depression to the level of respiratory arrest. In 2001, the Drug Enforcement Administration (DEA) reported 146 cases in which OxyContin was the direct cause of death and an additional 318 cases in which OxyContin was a contributing cause of death (4). A 2002 Drug Abuse Warning Network (DAWN) survey reported a 108% increase in emergency room incidences involving oxycodone abuse from the previous year and identified the drug as a major contributor to overdose death throughout the United States (5,6). DEA prosecution of physicians, who may be "freely" pre- Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 825

2 scribing OxyContin, has encouraged health care providers to rely heavily on opiate screening results to direct their prescription writing and management of pain-clinic patients. Some pain clinics are incorrectly interpreting a negative urine drug screen for opiates as an indication that the patient is selling instead of taking their prescribed OxyContin and subsequently refusing further care of the patient (7). In some clinical settings, it would therefore be advantageous for oxycodone assays to provide an accurate measurement of oxycodone levels and correctly reflect prescribed use. The sensitivities and cross-reactivities of oxycodone immunoassays have been known to show significant variability. In 1992, opiate immunoassays showed significant cross-reactivity with oxycodone levels between 5000 and 10,000 ng/ml (8). It is important to note that these assays were created to detect multiple opiates, and the new oxycodone assay was designed to detect only oxycodone and its metabolites. In 1999, a major study concluded that immunoassay technology might have limited effectiveness when monitoring the 6-keto-opiates such as hydrocodone, hydromorphone, oxycodone, and oxymorphone (2). More recent studies that investigated oxycodone-related deaths in the U.S. did not "trust" opiate immunoassay results, citing lack of specificity of accurate data to be the major problem (1). Because of these concerns, opiate immunoassay results that are used to screen oxycodone levels often require confirmation via a gas chromatographic-mass spectrometric (GC-MS) method. Such duplicate testing for confirmation is not appropriate for screening and often leads to increased turnaround times and higher operational costs. Most clinical laboratories use commercially available immunoassay kits to screen for opiates. It is important to note that such kits were designed to detect morphine and codine, but not hydromorphone, hydrocodone, or oxycodone. Cone et al. (8) evaluated the sensitivity and cross-reactivity of opiates with four commercial opiate urine immunoassays. These included the TDx Opiates (Abbott Laboratories), Coat-A-Count Morphine (Diagnostic Products Corp.), Abuscreen Radioimmunoassay for Morphine (Roche Diagnostic Systems), and Emit d.a.u. Opiate Assay (Syva Co.). The authors concluded that each of the 6-keto-opioid compounds had concentrationdependent cross-reactivities with antibodies used in the immunoassay. Furthermore, each had the potential to produce positive urine screening results. Microgenics Corp. (Fremont, CA) developed the first immunoassay with monoclonal antibodies specific for oxycodone. The Veterans Affairs Puget Sound Health Care system Seattle Division was chosen as a beta site for this assay. In this study, we report our method evaluation of the DRI Oxycodone Assay for the qualitative and semiquantitative detection of oxycodone in urine. Furthermore, we describe our experiences using this immunoassay to screen urine samples from our substance abuse program. We hypothesize that this new immunoassay screen for opiates would reduce the need for GC-MS confirmation and therefore provide a more rapid and nearly equally accurate assessment of oxycodone use in patients prescribed OxyContin. However, current oxycodone immunoassays may not be best suited to replace GC-MS confirmation in settings, such as forensic laboratories, that necessitates the gold standard technique. Methods All testing was performed using standard clinical lab method development protocols approved by the Veterans Affairs and University of Washington Medical Center System. This study measured oxycodone via immunoassay using the DRI Oxycodone Assay on a Hitachi 911 (Roche Diagnostics, Indianapolis, IN) with Diagnostic Reagents Inc. DRI (Catalog # 0136). Antibody/substrate Reagent A contains mouse monoclonal anti-oxycodone derivative antibodies, glucose-6-phosphate, and nicotinamide adenine dinucleotide (NAD) in Tris buffer. Enzyme Conjugate Reagent E contains oxycodone labeled with glucose-6-phosphate dehydrogenase (G6PD) in Tris buffer. The low cutoff calibrator, used to distinguish positive from negative results, contains 300 ng/ml oxycodone in human urine. Reagents and calibrators contained sodium azide as a preservative. The new oxycodone assay is a liquid homogenous enzyme immunoassay. The specific antibodies can detect oxycodone, oxymorphone, and noroxycodone without significant crossreactivity to other opiate compounds. The assay is based on competition between a drug-labeled enzyme G6PD and free drug from the urine sample for a fixed amount of specific antibody binding sites. In the absence of free drug from the sample, the specific antibody binds the drug-labeled G6PD, and the enzyme activity is inhibited. This phenomenon creates a direct relationship between drug concentration in urine and enzyme activity. The enzyme activity is determined spectrophotometrically at 340 nm by measuring its ability to convert NAD to NADH. The traditional DRI opiate immunoassay is not specific for oxycodone. Within-run precision studies were run on the Hitachi 911 using a urine sample containing 810 ng/ml of noroxycodone, as measured with our GC-MS (n = 10). Hitachi 911 betweenrun semiquantitative and qualitative precision studies were run over a 10-day period using the 225 and 375 ng/ml calibrators. Linearity studies were run using one blank (0 ng/ml urine) and 11 different known concentrations of oxycodone (75, 125, 225, 300, 375, 500, 1000, 1250, 2500, 5000, and 10,000 ng/ml). The components listed in Table I were combined with drug-free urine and subsequently analyzed to determine interference levels. Method comparison studies were conducted on four sample groups. The first group contained 99 consecutive urine samples from an alcohol and drug abuse treatment program. The second sample group contained 23 urine samples labeled oxycodone positive, and the third sample group contained 23 urine samples labeled oxycodone negative (both supplied by Microgenics). The fourth sample group contained 3 urine samples from 3 different patients who were prescribed 5 mg of oxycodone 3 or 6 times a day. These four groups, containing 148 urine samples, were screened by both the Hitachi DRI Opiate and Oxycodone assays. These samples were subse- 826

3 quently assayed for oxycodone via GC-MS (HP6890 GC, HP5973 MS, Agilent Technologies, Palo Alto, CA) to confirm the presence or absence of oxycodone (or its primary metabolite, noroxycodone). Using the new Hitachi 911 oxycodone immunoassay, we retrospectively evaluated an additional 17,069 urine samples over a 4-month period. We subsequently determined the number of GC-MS oxycodone confirmations that were requested for these samples. Furthermore, we determined the frequency of positive oxycodone immunoassay screening resuits in positive samples from patients who were and who were not being prescribed oxycodone. standard. However, the 2500 and 5000 ng/ml standards were underestimated by the immunoassay by about 20%, and the 10,000 ng/rnl standard had better agreement (11,045 ng/ml) with the immunoassay result. These highest three concentrations (2500, 5000, and 10,000 ng/rnl) are not illustrated in Figure 1. The manufacturer reports a sensitivity of 5.6 ng/ml; in our hands, the assay sensitivity was 10 ng/ml. Figure 2 illustrates the screening results and prescription status profiles among the 17,069 oxycodone immunoassays that we reported during a 4-month period. Over 98% (16,845) of these screening results were negative. Among the 224 positive results, 131 (58.5%) were being prescribed oxycodone, and Results Table II lists the oxycodone assay resuits for 148 urine samples from the 4 sample groups. The new DRI Oxycodone Immunoassay results are compared to the current gold standard GC-MS result. For both assays, 41 out of 148 results were positive and 106 results were negative for oxycodone. There was one sample from the first sample group that resulted negative via the DRI Oxycodone Immunoassay and positive via the GC-MS (false negative). There were no false-positive results. The serniquantitative oxycodone immunoassay result was 278 ng/ml for the false negative sample; 300 ng/ml was our cutoff for positive, and this sample resuited at 334 ng/ml by GC-MS. This result comparison generated a sensitivity [(True Positives * 100)/(True Positives + False Negatives)] and specificity [(True Negatives * 100)/(False Positives + 1Yue Table I. Interference Study for DRI Oxycodone Immunoassay* Hydrocodone 16,000 Codeine 28,290 Zolpidem 500,000 Chlordiazepoxide 200,000 Levorphenol 200,000 Methadone 200,000 Benzoylecogonine 200,000 Oxazepam 200,000 7-Aminoclonazepam 200,000 Hydromorphone 200,000 Morphine-3-glucuronide 200,000 Desipramine 500,000 Promethazine 500,000 Naltrexone 500,000 Morphine 500,000 Oxymorphone* 300 Noroxycodone* 300 Values obtained from the manufacturer. Negatives)] of 97.7% and 100%, respectively. The predictive positive [(True Positives)/(True Positives + False Positives) * 100] and negative [(Ture Negatives)/True Negatives + False Negatives) * 100] values were 99% and 100%, respectively. Table III lists within-run and between-run oxycodone measurement from the oxycodone immunoassay and the corresponding GC-MS result. Within-run precision was calculated using a GC-MS-positive oxycodone urine sample. Because of our 300 ng/ml cutoff, we conducted the between-run measurements using the 225 and 375 ng/ml calibrators. The results of the interference study are presented in Table I. Of the 14 compounds used to assess interference, levorphenol showed the highest contribution (96 ng/rnl) when present at a concentration of 200,000 ng/ml. At concentrations tested, none of the compounds produced a positive result using a 300 ng/ml cutoff. Figure I illustrates the result from our accuracy study. Semiquantitative results from the new DRI oxycodone assay were compared to known oxycodone standards up to 10,000 ng/ml. There was a high degree of correlation up to the 1250 ng/ml Apparent Semiquantitative Concentration of Pure Oxycodone Concentration % Cross-Reactivity Compound Tesled Due to Cross-Reactivity at Indicated (ng/ml) (ng/ml) Concentrations O O0 300 I00 Table II. Oxycodone Results for 148 Urine Samples Tesled by GC-MS and DRI Oxycodone Immunoassay Immunoassay Positive Immunoassay Negative GC-MS Positive 41 I GC-MS Negative Table III. Within-Run and Between-Run Precision DRI Measurement with the Corresponding GC-MS Result Within-Run Precision Between-Run Between-Run (at 225 ng/ml) (at 375 ng/ml) Mean (nglml) Standard deviation % CV GC-MS result (ng/ml)

4 93 (41.5%) were not on prescription oxycodone from our facility. Of the 17,069 oxycodone requests, 67.3% were generated from the alcohol treatment clinic. Forty (43%) of the 93 positive oxycodone results from patients who were not prescribed oxycodone were from the alcohol treatment clinic. We were unable to determine how many negative screening results,-~ ~ looo ~= 600- "~= 400- ~J 200- r f Oxycodone linearity js Sp /J dsd ~ s / j~ y = 1.043x r = = OI 0 2;0 4;0 G;o 8;0 10'00 12'00 1;00 Oxycodone stadard (ng/ml) Figure 1. Accuracy study for DRI Oxycodone immunoassay. / \ Figure 2. Oxycodone result and prescription profiles for a 4-month period. Table IV. Oxycodone Doses Prescribed to Patients Screening Positive on the DRI Oxycodone Immunoassay Dose Doses per Daily Dose Number of (mg) Day (mg) Patient Samples corresponded to patients who were on prescription oxycodone. Table IV lists the dose (in milligrams) by number of does per day and total daily dose for patients who were both on prescription oxycodone and who screened positive by the oxycodone immunoassay. These results are summarized by the number of positive screening results for the 131 ( not being prescribed oxycodone) positive results corresponding to patients who were prescribed oxycodone. Discussion Males represent about 85% of oxycodone abusers, and the mean age has increased from 24.5 years in 1979 to 40.0 years in 2003 (1,9,10). Additionally, death rates are increasing in older populations (11). Many studies indicate that oxycodone abusers concurrently abuse alcohol and/or other drugs (polydrug abuse). Furthermore, the risk of fatal overdose increases significantly when oxycodone is abused with alcohol (12-14). In a study of 1243 oxycodone postmortem overdose cases, it was determined that alcohol presence in combination with oxycodone was the third most prevalent finding after diazepam and hydrocodone (1). Interestingly, a substantial proportion of positive specimens from this study occurred in patients who were being treated for alcohol dependence. In our function as a beta site for the new DRI Oxycodone Immunoassay, we determined that the assay has superior sensitivity and specificity when compared to other screening assays. There were no GC-MS confirmation requests for any of the 17,069 oxycodone immunoassay results generated during the 4-month period. Previously, the observation was made that immunoassay technology was "not well-suited to detect or monitor the prevalence of 6-keto-opioid abuse" (2). Our study indicates that this observation may no longer be true. However, as is often the case with other drug screening assays, each facility should determine their own appropriate and populationspecific cutoff value. The utility of this assay may not be as reliable if the chosen cutoff value is significantly lower than 300 ng/ml. Our study did not include cross-reactivity with norhydrocodone, hydrocodol, or hydromorphol. Such analytes should be included in future evaluations if clinically indicated. We did not perform a formal cost-analysis study. However, discussion of costs is based on our own reagents and instrument operation and do not include other variables such as tech time. Such values will vary in different facilities. It is evident that the DRI Oxycodone Immunoassay saves significant time and expense. Point-of-care testing is less reliable, and the average cost is about US$25 per test. GC-MS confirmation is approximately US$35 per test. Before implementation of the immunoassay at our facility, we used the automated high-performance liquid chromatography (REMEDI) analyzer (BioRad, Hercules, CA). This system used a scanning ultraviolet detector to identify a broad spectrum of drugs. The initial instrument cost was about $100,000, and the oxycodone assay cost about US$20 per test. The cost per test for reagents on the DRI assay is about 45 cents. The improved oxycodone-screening assay may lead to better 828

5 patient and prescription management, decreased cost with increased laboratory efficiency, and may facilitate identification of individuals abusing oxycodone. Perhaps with additional improvements to oxycodone detection, the screening test may one day approximate the reliability of the gold standard GC-MS assay while operating at a fraction of the cost in some clinical settings. References I. E.J. Cone, R.V. Fant, and J.M. Rohay. Oxycodone involvement in drug abuse deaths: a DAWN-based classification scheme applied to an oxycodone postmortem database containing over 1000 cases. J. Anal. Toxicol. 27:57-67 (2003). 2. M.L. Smith, R.O. Hughes, and B. Levine. Forensic drug testing opiates, Vl. Urine testing for hydromorphone, hydrocodone, oxymorphone, and oxycodone with commercial opiate immunoassays and gas chromatography-mass spectrometry. J. Anal. Toxicol. 19:18-26 (1999). 3. E.J. Cone, R.V. Fant, and J.M. Rohay. Oxycodone involvement in drug abuse deaths. II. Evidence for toxic multiple drug-drug interactions. J. Anal. Toxicol. 28: (2004). 4. K.A. Moore, V. Ramcharitar, B. Levine, and D. Flower. Tentative identification of novel oxycodone metabolites in human urine. J. AnaL Toxicol. 27: (2003). 5. Substance Abuse and Mental Health Services Administration, Office of Supplied Studies. Emergency department trends from the Drug Abuse Warning Network, final estimates, , DAWN series D-24, DHHS Publications No. (SMA) , Rockville, MD. Published August Substance Abuse and Mental Health Services Administration, Office of Supplied Studies. Emergency department trends from the Drug Abuse Warning Network, final estimates, , DAWN series D-25, DHHS Publications No. (SMA) , Rockville, MD. Published January R.L. Vin Seggern, C.P. Fitzgerald, L.C. Adelman, and J.U. Adeleman. Laboratory monitoring of OxyContin (oxycodone): clinical pitfalls. Headache 44:44-47 (2004). 8. EJ. Cone, B.D. Dickerson, and J.M. Mitchell. Forensic drug testing of opiates. IV. Analytical sensitivity, specificity, and accuracy of commercial urine opiate assays. J. Anal. Toxicol. 16:72-78 (1992). 9. M. Gossop, D. Stewart, S. Treacy, and J. Marsden. A prospective study of mortality among drug misusers during a 4-year period after seeking treatment. Addiction 97(1): (2002). 10. S. Drake, J. Ross, D. Zador, and S. Sunjic. Heroin-related deaths in New South Wales, Australia, Drug Alcohol Depend. 60:141-t 50 (2000) January S. Drake and D. Zador. Fatal heroin 'overdose': a review. Addition 91: (1996). 13. A. Polettini, A. Groppi, and M. Montagna. The role of alcohol abuse in the etiology of heroin-related deaths. Evidence for pharmacokinetic interactions between heroin and alcohol. ]. Anal. Toxicol. 23: (1999). 14. S. Drake, S. Sunjic, D. Zador, and T. Prolov. A comparison of blood toxicology of heroin-related deaths and current herorin users in Sydney, Australia. Drug Alcohol Depend. 47:45-53 (1997). Manuscript received December 13, 2004; revision received January 25,