Methadone and Metabolite Urine Concentrations in Patients ...

103 downloads 26103 Views 2MB Size Report
detectable in urine from compliant patients. Methadone and EDDP concentrations significantly increased with methadone dose and (in one participant with poor ...
Journal of Analytical Toxicology, Vol. 27, September 2003

Methadone and Metabolite Urine Concentrations in Patients Maintained on Methadone Kenzie t. Preston 1,*, David H. Epstein 1, David Davoudzadeh 2, and Marilyn A. Huestis 1

IClinical Pharmacologyand TherapeuticsResearchBranch, Intramural ResearchProgram, National Institute on Drug Abuse, Baltimore, Maryland 21224 and 2Microgenics Corporation, Fremont, California 94538

normal. Unfortunately, most pharmacokinetic studies of methadone differ from typical clinical practice in at least three

Abstract [ As regulatory control over methadone maintenance relaxes, the need for methodsof monitoring compliance will increase. In community clinics, monitoring would most likely involve immunoassaysof outpatients' trough urine specimens.There are no publishednormsfor such data. Therefore, we determined concentrationsof methadone in 1093 urine specimenscollected thrice weekly in 27 outpatientsduring up to 17 weeks of observed methadone ingestion(35 to 80 mg/day) usinga semiquantitative homogeneousenzyme immunoassay(CEDIA). We used a separate CEDIA assayto measure methadone'smain metabolite, 2-ethylidene-3,3-diphenylpyrrolidine(EDDP), which may help detect compliance in fast metabolizers or patientswho adulterate samplesto simulate compliance. Methadone concentrationswere more variable than those of EDDP. Concentrationsof methadone were < 100 n~mL in one specimen, between 100 and 300 ng/m/ in 27, and _>300 ng/mL in all others. EDPP concentrationswere _>100 ng/mt in all specimens,suggestingthat EDDP should be detectable in urine from compliant patients. Methadone and EDDP concentrationssignificantlyincreasedwith methadone dose and (in one participant with poor clinic attendance) significantlydecreased following missedmethadone doses. Nevertheless,variability was too great to permit estimation of methadone dose (or detect a single missedadministration)from any single specimen.

Introduction

As efforts to increase the availability of maintenance therapy on methadone and other opioid agonists expand and the guidelines for maintenance change, treatment providers are likely to turn to drug monitoring as an aid in optimizing outcome and improving detection of methadone diversion (1). Clinicians wishing to monitor patients' methadone levels may look to the published literature to determine what should be considered 9Author to whom correspondence should be addressed. Kenzie L. Preston, Ph.D., Clinical Pharmacology and Therapeulics Research Branch, Intramural Research Program, National Institute on Drug Abuse, 5500 Nathan Shock Dr., Baltimore, MD 21224. E-mail: kpreston@inlra,nida,nih,gov.

332

ways. First, many of the studies have focused on measurements

made in plasma (2-5). A significant barrier to applying the knowledge gained from these studies is that many patients in substance-abuse treatment have poor venous access and infectious diseases, making blood collection difficult and a safety risk for both patients and staff. In addition, compared with blood sample procedures, collection and analysis of urine specimens require less staff expertise, less staff time, and less costly equipment, and urine testing can more easily be done on-site. Urine is also easier and safer to collect repetitively and, in general, drugs have longer windows of detection in urine. Therefore, urine would be a more practical matrix for methadone monitoring in clinical settings. Second, even in studies in which the matrix analyzed was urine, most of the samples were collected during prolonged stays on closed research wards (6--8), enabling the collection of complete 24-h urine specimens. [In one study whose participants were described as methadone-maintained "outpatients," assays were performed on complete 24-h urine specimens, presumably collected on a closed ward (9).] In a clinical setting, it is much more likely that assays would be performed on single specimens collected just prior to daily dosing. This collection schedule, combined with the likelihood of missed doses and concomitant use of illicit drugs, seems likely to produce results with greater variability than has been observed in the published data. Third, nearly every published study of methadone's elimination profile in urine has been conducted using assays [typically gas chromatography-mass spectrometry (GC-MS)] that are beyond the financial reach of most clinics. If a community clinic were to adopt a policy of monitoring patients' methadone levels, budgetary considerations would probably dictate the use of a semiquantitative immunoassay. Thus, although the body of published GC-MS data constitutes a scientifically important "gold standard," it may be of only limited use as a reference source for clinicians attempting to interpret patients' immunoassay results.

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.

Journal of Analytical Toxicology, Vol. 27, September2003

To our knowledge,the largest published dataset using a standard immunoassay (FPIA) to measure methadone in outpatient urine samples is that of McCarthy (10). Results were given from 401 methadone-positive samples in an unspecified number of patients, and case examples showed that semiquantitative testing could reveal illicit dose supplementation or verify the presence of methadone in samples that were below the standard thin-layer chromatography cutoff of 2000 ng for methadone and 1000 ng for metabolite. Unfortunately,no timecourse data were provided, EDDP concentrations were not reported, and no analyses were presented to relate methadone levels to patient characteristics. Therefore, although the study shows the need for and potential value of quantitative urine drug monitoring, more data are needed. Urine concentrations of methadone can be determined semiquantitatively (i.e., concentrations calculated by interpolation from a nonlinear curve) with CEDIA| DAU,a homogenous immunoassay (MicrogenicsCorporation, Fremont, CA; CEDIAis a registered trademark of Roche Diagnostics, Inc.). Also, a newly developed CEDIA DAU immunoassay can semiquantitatively measure methadone's main metabolite, 2-ethylidene-3,3diphenylpyrmlidine (EDDP) without significant contribution from the parent drug. EDDP is produced by N-demethylation, primarily by cytochmme P450 3A4 in the liverand the intestine; some EDDPis further N-demethyIated to 2-ethyl-5-methyl-3,3diphenylpyraline (EMDP) (Figure 1). EDDPand EMDP (each believed to be pharmacologically inactive) are both detectable in urine (11-17). [Variations in expression or activity of CYP3A4 may occur within and between individuals, leading to substantial variability in methadone metabolism and elimination (18).] Also detectable in urine are small quantities of hydroxylated metabolites, including normethadol (which may be pharmacologically active), along with free and conjugated parent compound (11). To our knowledge, there are no published data on simultaneous determination of the relative concentrations of parent compound, EDDP, EMDP, and normethadol in urine during methadone maintenance treatment, but most investigators have found that concentrations of EDDP equal or exceed those of parent compound, whereas concentrations of EMDP and normethadol are substantially lower with EMDP/methadoneratios ranging from 0.003:1 to 0.08:1 (17) and normethadol/methadone ratios approximately 0.2:1 (11).

Methadol

Nonnethado~

CHs Methadone

~

\

EDDP

EMDP

p-Hydroxylation and Glucuronlde Conjugation

Figure1. Metabolism of methadone.

One group of investigators (19,20) suggested that urine EDDP can be used as a marker of compliance with methadone administration. Assayingfor both methadone and EDDP could be useful in determining compliance among "fast metabolizers" (EDDP but no methadone in the urine) and in identifying urine that has been adulterated ("spiked") with methadone (methadone but no EDDP in the urine). Alternatively, assaying for EDDP alone could be a cost-efficient way to monitor compliance (19,20). In the present study, we evaluated methadone and EDDP concentrations determined by semiquantitative CEDIA DAU immunoassays in sequential single urine specimens from 27 methadone-maintained outpatients. Specimens were collected thrice weekly throughout patients' first 17 weeks of methadone treatment at doses ranging from 35 to 80 rag/day. A semiquantitative immunoassay (CEDIA)was used because (1) the CEDIA assays have been demonstrated to be valid, reliable methods when tested against GC-MS assays, (2) the less expensive and less technically difficult immunoassay is more clinically relevant than GC-MS, and (3) one of the research questions was whether screening for EDDP could or should be substituted for methadone screening, which is most frequently done by immunoassay. Thus, the semiquantitative immunoassay satisfied both the scientific and the application needs of the study.

Methods

Participants Twenty-sevenindividuals (18 male, 9 female; 8 African-American, 19 Caucasian) were enrolled in a behavioral contingencymanagement study as part of a larger methadone-maintenance treatment program. Individuals were eligible for the primary treatment study if they were between the ages of 18 and 65, if they qualified for methadone maintenance, and if they reported histories of intravenous opioid use. Applicants who tested positive for methadone but were not enrolled in a methadonemaintenance program were excluded from the study. All volunteers gave informed written consent prior to participation; the study was approved by the NIDAInstitutional ReviewBoard. All participants received standard methadone-maintenance treatment throughout the study that included daily methadone (35 to 80 rag/day, oral) and weekly individual counseling. Racemic methadone HCl (Mallinckrodt,Inc., St. Louis, MO) was administered daily in a cherry-flavored solution. Each dose was administered under direct observation of a nurse, and the patient was required to speak immediately after taking the dose, reducing the likelihood of the patient's failing to swallow the dose. Methadone was consumed in the clinic between the hours of 11:00 a.m. and 1:00 p.m. or 4:00 p.m. and 6:30 p.m. on weekdays and between 10:00 a.m. and 12:00 p.m. on Saturday and Sunday. Take-home doses were given only on Federal holidays. Dosing records were maintained in each patient's medical chart and in the pharmacy; these records were used to identify missed methadone doses. Unless transferring from another methadone-maintenance program, each patient was started at 333

Journal of Analytical Toxicology, Vol. 27, September2003

an initial dose of 30 mg. For most participants, the dose was raised to 50 mg by the fifth day of treatment. After that point, participants were allowed to request dose increases during the first four weeks of treatment in 10-mg increments to a maximum dose of 80 mg. Dose increases were given no more frequently than once every five days. Patient-requested dose increases were provided if the patient did not show undesirable side effects (e.g., sedation) from the current dose and if the patient continued to provide opioid-positiveurine specimens, or reported heroin use, desire to use heroin, or opioid-specific withdrawal symptoms. No dose increases were given after the fourth week of treatment. Some participants were unable to tolerate the dose escalation schedule described. Dose increases could be stopped at any point, or a decrease in dose in 5-mg increments could be prescribedby the study physician,as deemed clinicallyappropriate. Once the dose escalation was stopped, or the methadone dose was decreased, the patient remained on that dose for the remainder of the study. The first fiveweeks, of which at least one week was at the maintenance dose, are defined in this paper as the stabilization phase. The next 12 weeks of treatment are defined as the maintenance phase. Some participants did not complete 12 weeks of the maintenance phase.

Specimen collection Urine specimens (N = 1093) were collected three times per week, on Monday,Wednesdayand Friday, under direct supervision by trained staff and always before methadone administration (at trough). Urine collection was not initiated until after the first methadone dosing day. The specimens were screened within 24 h of collection for methadone by Enzyme Multiplied Immunoassay Technique (EMIT| Dade-Behring, San Jose, CA) for treatment program use with a cutoff concentration of 300 ng/mL. Specimens were then stored in a -20~ freezer until further analyzed for methadone and EDDP as described.

Analytical methods Methadone and EDDP CEDIA imrnunoassays. At the conclusion of the study, all urine specimens were randomized, blinded, and analyzed separately for methadone and EDDP equivalents by CEDIADAU,homogeneous enzyme immunoassays for the semiquantitative determination of methadone and EDDP. In the CEDIADrugs of Abuse assays, drug in the sample competes with drug conjugated to an inactive fragment of ~-galactosidase for an antibody-binding site. If drug is in the sample, inactive [3-galactosidase fragments are free to reassociate to form active enzyme. The active enzyme interacts with the substrate to produce a color and absorbance change that is proportional to the amount of drug present in the sample. The methadone and EDDP assays are approved for diagnostic use by the Food and Drug Administration (FDA).A Hitachi 717 analyzer (Boehringer Mannheim, Indianapolis, IN) was used according to the manufacturer's recommended procedures. The manufacturer's recommended cutoff concentration for methadone in urine is 300 ng/mL; the cutoff concentration for EDDP is 100 ng/mL. Semiquantitative concentrations also permitted comparison of methadone and EDDP urine test results and determination of EDDP/methadoneratios 334

for all urine specimens. The cross-reactivity of the methadone assay was 100% for methadone, < 0.02% for EDDP, < 0.03% for EMDP, 1.5% for methadol, and 2.65% for ~c-methadol.Therefore,the combined urinary concentrations of methadone metabolites should not contribute to the semiquantitativeconcentration of methadone, permitting independent measurement of urinary methadone and EDDP. Morphine, cocaine metabolites, amphetamines, cannabinoids, and other common drugs of abuse also did not cross-reactat expectedconcentrations. The limit of detection (3 SD of 21 replicates of the negative calibrator) of the semiquantitative assay was 91.5 ng/mL, and the upper limit of linearitywas 1000 ng/mL (Product InformationSheet 2001-2003, Microgenics Corporation). In validation studies submitted by the company to the FDA, the CEDIA methadone assay was compared to EMITmethadone in 820 clinicalurine specimens. There was agreement in positiveand negativetest results at the 300 ng/mL cutoff, except for 16 specimens that screened positive by CEDIAand negative by the other immunoassay.GC-MS results confirmed the presence of methadone at 210 to 299 ng/mL in 12 specimens and 308 to 319 ng/mL in 4 of the discrepant specimens. The cross-reactivityof the EDDPassay was 100% for EDDP, < 0.016% for methadone, r 0.004% for EMDP,and < 0.01% for methadol metabolites. The limit of detection (3 SD of 21 replicates of the negative calibrator) of the semiquantitative assay was 1.95 ng/mL, and the upper limit of linearity was 2,000 ng/mL (Product Information Sheet, November, 1998, Microgenics Corporation). Specimens that contained methadone or EDDP equivalents greater than the upper limit of linearity were diluted with reagent buffer until results fell within the linear range of the assay. The manufacturer compared CEDIA EDDP immunoassay results in 200 urine samples to those obtained with an HPLC reference method (MicrogenicsCorporation data). There was qualitative agreement for 194 of the samples. Four samples were negative by the immunoassay at the 100 ng/mL concentration cutoff but positive by HPLC (105 to 173 ng/mL) and two samples positive by immunoassay were found to contain 70 and 89 ng/mL EDDPby HPLC. Semiquantitative concentrations of methadone and EDDP were obtained by assaying four calibrators for each drug and utilizing a logit-log data transformation. Calibrators of 0, 300, 600, and 1000 ng/mL methadone and 0, 100, 500, and 1000 ng/mL EDDPwere included in each assaycalibration. Quality control samples spiked to contain methadone and EDDP at concentrations above and below the respective qualitative immunoassay cutoff concentrations of 300 and 100 ng/mL were assayed with all analytical runs to assure adequate assay performance. In general, the assay was calibrated once each day; however, if quality control samples did not meet qualitative negative and positive criteria, the assay was recalibrated and specimens reanalyzedwith accompanyingquality control samples. The mean + SD concentrations and CVfor the quality control samplesacross all assays were as follows:methadone below cutoff (N = 14) 226.5 + 7.1 ng/mL CV = 3.1%, above cutoff (N = 14) 373.8 • 9.1 ng/mL CV 2.4%; EDDP below cutoff (N = 13) 75 • 2.1 ng/mL CV 2.7%, and above cutoff (N = 13) 143.0 • 2.3 ng/mL CV 1.6%.

Journal of Analytical Toxicology, Vol. 27, September 2003

Opiate, cocaine, and cannabinoid immunoassays. In order to evaluate the effect of other illicit drug use on methadone and EDDP urine concentrations, urine specimens were also tested for opiates, cocaine and cannabinoid metabolites. Urine specimens were analyzed in a similar manner as described with drug class specificCEDIAreagents. Cutoff concentrations were 300 ng/mL for the opiate and cocaine assays, and 50 ng/mL for the cannabinoid metabolite assay. Creatinine assay. Creatinine concentrations were determined in all specimens; analyses were performed by the Jaffe method with Sigma-Aldrich (St. Louis, MO) reagents on a Hitachi 717 analyzer (Boehringer Mannheim, Indianapolis, IN).

Univariate procedure in SAS V. 8.0. To test the relationship between daily methadone dosage and assay results, doses were categorized into four groups (30-49 rag, 50-59 rag, 60-69 rag, and > 70 rag), and mixed regressions were performed using the Mixed procedure in SAS V. 8.0. This procedure, essentially a repeated-measures regression, enabled us to include subjects with missing data points and to model dose as a time-varying covariate. Two sets of mixed regressions were performed--one including all specimens and one including only specimens from the last 12 weeks of the study (maintenance phase, when all participants had been on their final maintenance doses for at least one week, and no further dose adjustments were made). Data from one participant (Subject 15) were not included in the mixed regressions because her clinic attendance was much more sporadic than that of other subjects: out of 48 urine specimens collected, 15 were collected approximately 48 h after her previous dose of methadone (one dose missed), and four were approximately 72 h after her previous dose (two doses missed). For 23 of the other 26 subjects, none of the urine

Data analysis To correct for fluctuations in urine concentration, all values were normalized to creatinine by dividing methadone and EDDP concentrations by corresponding creatinine concentrations. The ratio of EDDP to methadone was calculated for each specimen. Distributions of values of methadone, EDDP, and EDDP/rnethadone ratio were examined descriptively with the

1.1o

1000O.

~

SOaO ] 7ome

. 25.

.

. .

.

.

. ]aOme

.

.

.

. 26

.

.

. ~

.

11

.

.

.

q ] m Immm

~

27

.

.

.

.

.

.

.

.

.

o 8l l OIO C ol&f l s2o e curine u25t l v3oespecimen 316 4o 48 I I M~mllmw/GnlWbl~l o EDOP~ma~rm

o s Io Is ao zs ~ 8s 4o ,~ so o Consecutive urine specimen

s 1o Is 2o a ~o x 4o 4s so o s io is ~ 2s so as ,io 4s 8o Consecutive urkle specimen Consecutive urkle specimen

Figure 2. Creatinine-corrected semiquantitative urinalysis results (ng methadone or EDDP/mg creatinine) of sequential urine specimens of individual patients maintained on methadone. Methadone (solid squares)and EDDP (open circles) concentrations were determined in up to 51 consecutive specimens collected Mondays, Wednesdays, and Fridays over 17 weeks (5-week stabilization and 12-week maintenance phases). On each graph, participant number is indicated at the top, and methadone maintenance dose is shown in the upper left corner.

335

Journal of Analytical Toxicology, Vol. 27, September 2003

Most of the methadone-negative specimenswere collected from just two participants: 10 negative specimens from Subject 3, who was maintained on 45 rag/day, and 6 from Subject 2, who was maintained on 35 mg/day. The distribution of negative specimens among the other participants was 3 (Subject 18), 2 (Subject 9), 1 (Subjects 8, 13, 14, 15, 19, 21, and 23), and 0 (remaining 16 participants). All but one of these methadone-negative specimens contained methadone concentrations between 100 and 300 ng/mL. The only specimen in which the methadone concentration was below 100 ng/mL came from Subject 3. All 1093 specimens tested positivefor EDDPwith the cutoff set at 100 ng/mL, as suggested by the manufacturer. Resulh Only two specimens, both from Subject 21 (urine specimens 1 and 4), contained EDDP concentrations less than 300 ng/mL. The time courses of methadone and EDDP excretion in urine A total of 1093 specimens were screened with the CEDIA are shown for each of the 27 study participants, arranged in asDAU methadone and EDDP assays: 364 specimens from the cending order of methadone-maintenance dose (Figure 2). As first fiveweeks of treatment (stabilization phase) and 729 specindicated, methadone doses started at 30 mg for all particiimens from the remaining 12 weeks (maintenance phase). The pants and increased to the maintenance dosewithin 4 weeks (by mean numbers of specimens collected from each of the 27 par* urine number 12), then remained constant for the rest of the ticipants were 13.5 (SE 0.4) for the stabilization phase and 27.0 study. Consistent with this schedule, urine methadone and (SE 1.7) for the maintenance phase. Of the 1093 specimens, 28 EDDP concentrations corrected for creatinine concentration (2%) tested negative for methadone using the concentration tended to be low in the first specimen collected and to increase cutoff of ~ 300 ng/mL; 14 (50%) of the methadone-negative over the next few collections. specimens occurred during the five-week stabilization phase. Readily apparent in Figure 2 is the greater variability of methadone as compared to Table I. Effect of Methadone Dose on Methadone and EDDP CreatinineEDDP concentrations within each participant. Normalized Concentrations and EDDP/Methadone Ratios This greater variabilitycan also be seen in the respective heights of the boxes in Figure 3. Dose(mg) StabilizationplusMaintenance' MaintenanceOnlyf Each box shows data for an individual particMethadone* F(3,39) = 9.71, p < .0001w F(3,22) = 1.76, p = .1837w ipant, and the height of the box represents 30-49 3044 (843) 3753 (1810) the interquartile range (25% and 75% per50-59 6459 (617) 7394 (956) centile) for concentration of methadone or 60-69 7541 (710) 9503 (1808) EDDP (corrected for creatinine) during the > 70 7586 (760) 7001 (1052) maintenance phase (after stabilization). The difference in variability reached statistical sigF(3,39) = 42.96 p < .0001~ F(3,22) = 2.58, p = .0793w EDDP* nificance: the mean coefficients of variation 30-49 2366 (455) 2368 (1275) for EDDP and methadone were 57% and 74%, 50-59 4468 (408) 4934 (668) respectively [t(26) = 4.25,p = 0.0002]. Across 60-69 5262 (429) 6787 (1275) all maintenance-phase specimens, the mean, > 70 6370 (443) 5975 (737) median, and 25th and 75th percentile conTotal EDDP+methadone* F(3,39) = 24.04, p < .0001 F(3,22) = 2.98, p = .0533* centrations of methadone were 9509, 6240, 30-49 5339 (1087) 6142 (2463) 3240, and 12,800 ng/mL, respectively. For 50-59 10840 (869) 12291 (1296) EDDP, the mean, median and 25% and 75% 60-69 12886 (953) 16208 (2461) percentile concentrations were 6336, 5250, >_70 14140 (1005) 12904 (1429) 2993, and 8380 ng/mL, respectively. Creatinine-corrected values for methadone were EDDP/methadoneratio F(3,39) =1.67, p = .1863w F(3,22) = 0.90, p = .4593~ 7499, 6111, 3344, and 10,522 ng/mg, respec30-49 1.9 (0.4) 2.4 (0.8) tively. Creatinine-corrected values for EDDP 50-59 1.1 (0.2) 1.2 (0.4) were 5125, 4450, 3157, and 6619 ng/mg, re60-69 1.8 (0.3) 2.1 (0.8) spectively. > 70 1.6 (0.3) 1.5 (0.5) Among the participants in the study, there 9 Least-squaresmean (SE); Ns (specimens/participants) for dose groups: 30-49, 124/19; 50-59, 447/22; were several patterns of relative concentra60-69, 181/16; > 70, 293/9. t Least-squaresmean (SE);Ns (specimens/participants) for dose groups: 30-49, 79/3; 50-59, 286/11; tions of methadone and EDDP.With respect to 60~9, 91/3; > 70, 240/9. the different patterns, the urine specimens of * Corrected for creatinine; ng/mg creatinine. wResultsof mixed regression,with Dose modeled as a time-varying covariate (though it varied in the some participants (e.g., Subjects 5 and 17) Stabilization phase only); N = 26; Subject 15 omitted from analyses because of significant effect of missed consistently contained higher concentrations methadone doses. of methadone than EDDP, whereas concen-

specimens was preceded by a missed dose; for each of the other 3 subjects, 1 specimen was preceded by a missed dose (but never by two missed doses in a row). Data from Subject 15 were entered into separate mixed regressions with two timevarying covariates: Dose and Missed Last Dose (yes/no). Mixed regressions were also performed to test for sex differences, gradual changes across time (independent of dose), and effects of concomitant use of illicit drugs. For ease of presentation, these analyses are described in the Results section.

336

Journal of Analytical Toxicology, Vol. 27, September 2003

trations of methadone were lower than EDDP in the urine of other participants (e.g., Subjects 13 and 18) (Figure 2). In a third group of participants (e.g., Subjects 15 and 16) the relative concentrations of EDDP and methadone varied across time. Consequently, EDDP/methadone ratios during maintenance varied widely both within and across participants (Figure 4) with some extremely high values (e.g., 37.5 for Subject 18 and 24.2 for Subject 3). In several participants, for example Subjects 18 and 26, the mean ratio was greater than the 75 percent quartile EDDP/methadone ratio because of these high values. Across all participants, the mean EDDP/rnethadone ratio (representing 729 values from the maintenance phase) was 1.48, but the median was only 0.71. For 20 out of 27 participants, the median ratio was less than 1.0. The ratio was less than 1.0 in 65% of specimens from the maintenance phase, and in 63% of specimens from both phases combined. As mentioned in the Data analysis section, one participant (Subject 15) missed a substantial number of methadone doses. As shown in Figure 5, missed methadone doses (indicated by the triangles on the x-axis) tended to be associated with low methadone concentrations in her urine specimens. A mixed regression on Subject 15's data showed that creatinine-corrected methadone concentrations were significantly lower in specimens collected the day after a missed methadone dose compared with those collected the day after a methadone dose was taken [least-squares mean • SE: 2970 • 1240 vs. 6520 • 1390 ng/mg; F(1,44) = 5.4, p = .02]. Neither her creatininecorrected EDDP nor EDDP/methadone ratios were significantly altered following a missed methadone dose, F(1,44) = 0.08, p = .78 and F(1,44) = 1.86, p = .18, respectively. In the other 26 participants, analyses of all urine specimens (i.e., from stabilization and maintenance phases) showed that mean methadone, EDDP, and the mean sum of methadone plus EDDP (all corrected for creatinine) increased with methadone dose (Table I). When we included in the analyses only those specimens collected during the maintenance phase, there were no statistically significant relationships between dose and concentrations of either analyte. The mean EDDP/methadone ratio was not significantly affected by dose in either analysis. To determine whether the ratio or concentrations of methadone and EDDP changed with chronic administration of methadone, we compared urine specimens collected in the stabilization phase (the first five weeks of treatment) to those collected in the maintenance phase (next 12 weeks). Because time was partly confounded by increases in dose, we performed a mixed regression controlling for dose as a time-varying covariate. Methadone concentrations increased modestly from stabilization (least-squares mean = 5913 ng/rnL, SE = 575) to maintenance (least-squares mean = 6691 ng/mL, SE = 593), F(1,25) = 4.3; p = 0.048. EDDP concentrations also tended to increase from stabilization (least-squares mean = 4507 ng/mL, SE = 398) to maintenance (least-squares mean = 4754 ng/mL, SE = 399), but the increase did not reach statistical significance, F(1,25) = 2.99; p = 0.096. EDDP/methadone ratios did not significantly change from stabilization (least-squares mean = 1.49, SE = 0.22 ) to maintenance (least-squares mean = 1.71, SE = 0.24, F(1,25) = 0.89,p = 0.36. To evaluate whether the ratio or concentration was affected by

concomitant use of illicit drugs, we performed three separate sets of mixed regressions (for creatinine-corrected EDDP concentration, creatinine-corrected equivalent methadone concentration, and EDDP/methadone ratio) in which the predictor variables were creatinine-corrected concentrations of morphine, benzoylecgonine, and cannabinoid metabolites. We controlled for dose category, and we excluded data from Subject 15. Also excluded were data from three urine specimens that followed missed doses. None of the effects of morphine, benzoylecgonine, or cannabinoid metabolites approached significance, regardless of whether they were entered into the regression models individually or simultaneously (data not shown). We performed additional analyses to determine whether ratios or concentrations differed by gender. Gender was used as a predictor in mixed regressions that controlled for body weight (at study intake) and dose category, again excluding data from Subject 15 and the three urine specimens that followed missed doses in other participants. Also excluded were data from one female participant for whom body weight was not available. Creatinine-corrected EDDP concentrations were higher in women (least-squares mean = 722,537, SE = 60,353) than in men (least-squares mean = 407,615, SE = 40,336), F{1,19) = 21.53, p = 0.0001. In this analysis, EDDP concentration tended to increase with dose, F(3,19) = 4.46, p = 0.094, and body weight was negatively associated with EDDP concentration, F(1,19) = 4.87, p = 0.0398. Creatinine-corrected methadone concentrations also tended to be higher in women (least-squares mean --958,468, SE = 124,667) than in men (least-squares mean = 699,854, SE = 83,633), though the difference was significant only at trend level, F(1,19) = 3.39, p = 0.081. In this analysis, there was no main effect of body weight, and there was only a trend toward a positive association with dose category, F(3,19) = 2,46, p = 0.094. EDDP/methadone ratio did not differ by gender, body weight, or dose.

Discussion The present study, evaluating trough urinary concentrations of methadone and EDDP in 27 outpatients maintained on methadone for up to 17 weeks, documents quite variable methadone and EDDP concentrations over time. Median EDDP/methadone ratios in individual participants were typically less than 1.0. The present data should be useful to clinicians as a comparison for their outpatients' immunoassay results and to scientists as a basis for addressing the potential utility of urine testing for monitoring compliance with methadone maintenance. George and Braithwaite (19) have suggested that urine EDDP can be used as a marker for compliance with methadone administration, under the assumption that any urine specimen testing positive for methadone but negative for EDDP may have been spiked with the parent compound. Our results support the idea insofar as EDDP was present in all 1093 specimens we collected, at concentrations exceeding the cutoff of > 100 ng/mL. (It is important to note that this cutoff, along with the 337

Journal of Analytical Toxicology, Vol. 27, September 2003

methadone cutoff of > 300 ng/mL, was selected by the manufacturer; no official cutoff concentrations for these substances have been set by regulatory bodies.) However, we are not confident that either cutoff is especially sensitive to one or two missed doses because none of the specimens from Subject 15 (who missed substantially more methadone doses than any other patient) were below either cutoff. We cannot directly address the question of methadone-spiked urine specimens because our clinic procedures (monitored dosing, monitored urine collection, and very few take-home doses) made both diMethadone/Creatinine and EDDP/Creatinine 75% Q3 ; mean median , 25% Q1 ' [7 I-II~Methadone I: II

~

2OOO

1500

, HI'=EDOP

10oo

'

I:

version and spiking unlikely. Still, monitoring EDDP in urine may have another advantage over monitoring methadone in urine. In our experience running an 80-patient treatment-research program in which urine specimens are tested by EMIT three times per week (with the data reported here representing only a small proportion of the urine specimens we have collected), we have noticed that patients occasionally test negative for methadone despite having been administered an observed dose 24 h earlier. Indeed, in the present study, we found that 28 (2%) specimens tested negative for methadone using the concentration cutoff of > 300 ng/mL. As might be expected, Ratios half of these occurred early in treatment [

[1 M

": : i

=,'

i

i

'

i

i

i ui

n

5oo0

=

--:

==

I

U:

',I[ i i ,

q 0 . '..:a'.r' . . Subject 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 Methadone 35 45 50 55 60 70 80 dose (rag/d)

Figure 3. Mean, median, and 75th and 25th percentile concentrations (corrected for creatinine) of methadone (thin lines, closed circles, open boxes) and EDDP (thick lines, open circles, gray boxes) in the urine s )ecimens of individual participants maintained on methadone. Methadone and EDDP concentrations were determined in up to 36 consecutive specimens collected Mondays, Wednesdays, and Fridays over 12 weeks of a maintenance phase following a 5-week stabilization phase. Vertical dashed lines separatethe data from participants maintained on different methadone doses (35 to 80 mg/day, indicated at the bottom of the figure).

10

"[1~.21

[,~o]

[2,t~l

[a-;d [ta.d '~

i

9

,

ip j

highest ratio

[d.d

lf-I 75% Q3

mean ~ median

~ 25% Q1 9 lowest ratio

i i i

.o_ 7

6

~s Q m

i i i i =

3

9

0

Methadone Dose (rngld)

35

3 45

4

5

6

7 50

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 55 60 70 80

Figure 4. Mean, median, lowest, highest, and 75th and 25th percentile EDDP/methadone ratios in the urine specimens of individual participants maintained on methadone (35 to 80 mg/day). Methadone and EDDP concentrations were determined in up to 36 consecutive specimens collected Mondays, Wednesdays, and Fridays over 12 weeks of a maintenance phase following a 5-week stabilization phase. Vertical dashed lines separate the data from participants maintained on different methadone doses, as indicated at the bottom of the figure. A horizontal gray bar indicates the EDDP/methadoneratio of one. Up and down arrowheads indicate the value of the highest and lowest, respectively, EDDP/methadone ratios for each patient.

:]38

being increased to final maintenance doses). All but one of these methadone-negative specimens contained methadone concentrations between 100 and 300 ng/mL; only one specimen contained a methadone concentration below 100 ng/mL. The methadone-negative specimens were not equally distributed across patients. Most (16 out of 28) of the methadonenegative specimens were collected from just two participants. These individuals were maintained on relatively low doses (35 and 45 mg/day), which may have contributed to the low urine methadone concentrations. It is also possible that such patients represent rapid metabolizers of methadone. Subject 3, for example, had mean and median EDDP/ methadone ratios greater than one, consistent with rapid metabolism of methadone (Figure 4). If urine tests are to be used to monitor compliance, such patients could be incorrectly classifted as noncompliant unless their urine specimens were also tested for EDDP. Monitoring urine specimens for EDDP as well as methadone could detect specimens spiked with methadone that did not contain EDDP. Monitoring for EDDP rather than methadone should be sufficient to reveal compliance even in fast methadone metabolizers. We found a robust relationship between missed doses and next-day decreases in methadone levels in a single patient. The other 26 participants in our study did not miss a sufficient number of doses to permit examination of the relationship. Future outpatient studies will need to include larger samples so that a broader range of clinic attendance can be represented. In the 26 participants who rarely missed doses, we saw a positive relationship between current methadone dose and urinary concentrations of methadone and EDDP corrected for creatinine. This dose/concentration relationship reached significance only when we included all data (stabilization plus mainten-

Journal of Analytical Toxicology, Vol. 27, September 2003

tually increased rather than decreasing, a finding that did not seem attributable to desiccation. These findings imply that EDDP levels in the present study may be slightly inflated, especially relative to methadone levels. The semiquantitative immunoassays of trough urine specimens in methadone-maintained outpatients produced results that are notably different from those seen in GC-MS assays of 24-h urine specimens collected on closed wards. Because our methodology differed from that of most prior studies in terms of patient sample (outpatient vs. inpatient), specimen-collection schedule (single trough urines versus complete 24-h urines), and type of assay used (CEDIA immunoassay vs. GC-MS), we cannot firmly ascribe the differences to any of these factors. To guide future work, we will contrast our results with those of prior studies that used different methodologies.

ance). Within the maintenance phase, when steady state had presumably been achieved, the relationship no longer reached significance (Table I). It is not clear to what extent this was due to loss of statistical power (with fewer participants in each dose category once stable doses were reached). In any case, despite the positive relationships between methadone dose and urinary concentrations of methadone and EDDP, variability was too great to permit clinically useful prediction of one from the other. We found that urinary EDDP concentrations were higher in women than in men, even after controlling for body weight. Weight did not account for the higher EDDP in women, but, rather, appeared to be an independent factor affecting EDDP urine concentrations (as would be expected because participants were not dosed on a mg methadone per kg weight basis). The published literature appears to contain relatively little information on sex differences in methadone metabolism, though in a sample of 20 methadone-maintained patients, de Vos and colleagues (13) found a trend toward slower methadone elimination in women than in men. Thus, our finding of higher EDDP concentrations and (nonsignificantly) higher methadone concentrations suggests greater drug accumulation in women. Illicit drug use, including cocaine use, had no significant effect on EDDP/methadone ratio. Anecdotally, patients, and some clinic staff, report that cocaine decreases the effects of ("eats") methadone. The present study did not support a pharmacokinetic mechanism for this widely held belief. The specimens analyzed in the present study were collected and frozen at -20~ approximately six years prior to analyses. Although they were only exposed to a single freeze-thaw cycle during that time, it is possible that some degradation had occurred. Moody and colleagues (21) have found that the longterm stability of methadone is good, with an estimated 3950 and 3080 days to reach 15% deviation in single-analyte-spiked urine specimens with nominal concentrations of 100 and 300 ng/mL. EDDP was less stable in urine, with an estimated 153 to 98 days to reach 15% deviation in specimens with nominal concentrations of 25 to 300 ng/mL. For both analytes, concentrations ac-

Inpatient studies using GC-MS In these studies, mean EDDP/methadone ratios were usually greater than 1.0 (6-9). Although the same was true in the

present study, this occurred only because the mean was inflated by a few specimens with especially high ratios. The median ratio in the present study was below 1.0. As can be seen in Figure 4, the median ratio was lower than the mean in nearly every participant, and, for some participants, even the 75th percentile value was lower than the mean. The inflation of the mean was due to a disproportionate arithmetical influence from the lowest methadone values. Thus, for our EDDP/methadone data, the median was a better summary measure. Another finding in two of the inpatient studies (6,7) was that the EDDP/methadone ratio increased during stabilization on methadone. Verebely et al. (7) indicated that the ratio increased up to threefold in six patients during stabilization on 40 or 80 rag/day during 30 days on a closed ward. ~ngg~rd et al. (6) reported a similar threefold increase in six patients as their dose increased from 10 to 80 rag/day during 30 days on a closed ward. ~ngg~rd et al. (6) acknowledged that their design, with 24-h urines collected on two different days at the beginning and end of the protocol, could not distinguish the effect of time from that of dose. We attempted to make that dis8o tinction by testing for a difference across 2000~1 . 9 999 ~'~"~* phases while controlling for dose as a timeso varying covariate, and we found no change in the ratio from stabilization to maintenance. 15~176176 [~'~s~-] ' 4O We also found no relationship between ratio ~= I I* Methadone[ Jl' 30 ~ and dose after doses were stabilized (Table I). In addition, we did not observe excretion increases of the magnitude reported by Verebely 2o ~ et al. (7); in their participants, urinary EDDP lo concentration increased 13-fold from the first dose to day 30, whereas methadone excretion o increased about fourfold. In our participants, Urine specimen although urine methadone and EDDP concenFigure 5. Creatinine-correctedquantitativeurinalysisresultsof sequentialurine specimensand trations increased from stabilization to maindose of methadoneadministeredon eachday for Subject15. Methadone(solidcircles)and EDDP tenance (significantly, in the case of (open squares)concentrations were determined in 48 urine specimenscollected Mondays, methadone), the magnitudes of the increases Wednesdaysand Fridaysover 17 weeks(5-weekstabilizationand 12-weekinterventionphases). were on the order of 10%. It seems unlikely Methadonedosesare indicated by solid triangles;triangleson the x-axis(0 mg) indicate missed that this across-study difference in intrasubject methadonedoses. time-course data could be accounted for by a ~ 411~ii~

O

. . . . . . . . . . .

0

14

- . . . . .

28

,.

42

. . . . . -~

.........

. . . . . .

56

. . . . . .

70

~

~dMd, ~tA A h ~ J ~ t , A ~

, . . . . . . . . . . . .

84

98

AJ~idl A ~

J~, A

, . . . . . . . . . .

112

12

339

Journal of Analytical Toxicology,Vol. 27, September2003

difference in assay technique; it may represent a difference between single-trough urines and complete 24-h urines or between inpatients and outpatients.

Outpatient studies using GC-MS Kreek (22) tested single urine specimens collected 24 h after administration of 100 mg methadone in liquid formulation from nine patients whose routine methadone maintenance doses ranged from 80 to 120 mg/day. The EDDP/methadone ratios in those patients can be calculated from one of the data tables; the median ratio was 0.87, while the mean was 1.55 (SD 1.46, range 0.45-5.07). This is similar to our findings. Another similarity between our findings and those of Kreek (22) was that the coefficient of variation was high for creatinine-corrected EDDP concentrations (58% across subjects in the Kreek sample) and even higher for creatinine-corrected methadone concentrations (73% across subjects in the Kreek sample). In our sample, the coefficients of variation across subjects were lower (44% and 47%, respectively) because of our having collected many samples per subject, but the coefficients of variation within subjects (57% and 74%, respectively) are strikingly similar. The similarities of our results, despite different analytical techniques, suggest that concentrations of methadone and EDDP are more variable in trough urines from outpatients than in complete 24-h urines from inpatients.

Outpatient studiesusingthe CEDIA DAU immunoassay George and colleagues (20) followed up their GC-MS study with a CEDIA study in which they reported assay results from 1381 urine specimens (though no information was given on the number of patients or their clinical characteristics). This study included what is apparently the only published comparison of GC-MS values and semiquantitative CEDIAvalues for concentrations of EDDP. The comparison was conducted for 38 of 108 specimens that were EDDP-positive but methadone-negative by CEDIA. The correlation across assays was high (r = .87) with a range of EDDP concentrations from 52 to 515 mg/L (ng/mL). Because no similar comparison has been published for methadone, we cannot rule out the possibility that CEDIAoverestimates methadone relative to EDDP. If this were the case, it could partly account for the lower EDDP/methadone ratios we observed. (In addition, Gallowayand Bellet [23] have shown that the heat of the GC-MS injection port can convert methadone to EDDP when high concentrations of methadone are present. This could increase the EDDP/methadone ratio in GC-MS studies, although this effect is not likely to be large enough to account for the discrepancy across studies.) We should point out, again, that Kreek (22) found EDDP/methadone ratios similar to ours using GC-MS. One other CEDIA study in outpatients is available as an abstract only (24). In that study, 480 single urine specimens were taken from 106 outpatients on a collection schedule similar to the present study, though more variable in interspecimen interval from day to day, with dosing occasionally preceding specimen collection (personal communication, L.C. Ackers, VA North Texas Health Care System, Dallas, TX, 2002). Quantitative results from 14 urine specimens from one "typical" patient maintained on 50 rag/day collected across 12 weeks showed a

340

median EDDP/methadone ratio of 0.6 (mean 0.7, SD 0.4, range 0.3-2.0). If this patient was indeed typical of the whole sample, then the results are consistent with the results of the present study and with those of Kreek (22). Because most of the published papers report on methadone and EDDP concentrations determined by GC-MS in 24-h urine collections from inpatients, there are fewguidelines as to what concentrations or ratios should be regarded as normal in a clinical setting. The methadone levels in the present study appear roughly comparable to those seen in an outpatient sample whose urine was assayed with FPIA (10). Still, we hesitate to suggest that our findings can be treated as norms. There remains a need for more large-scale descriptive studies of methadone's urinary elimination profile in naturalistic settings, with specimens collected on a clinicallyrealistic schedule. Moreover, there is a need for controlled investigation of the patient-related, collection-schedule-related, and assay-related factors that have led to discrepancies across studies. Such investigations may increase the clinical utility of methadone and EDDP urine concentrations for improving methadone-maintenance treatment and detecting drug diversion.

Acknowledgments This work was supported by the Intramural Research Program of the National Institute on Drug Abuse Intramural Research Program.

References 1. K. Wolff and J. Strang. Therapeutic drug monitoring for methadone: scanning the horizon. Eur. Addict. Res. 5:36-42 (1999). 2. P. Beauverie, A.M. Taburet, M.C. Dessalles, V. Furlan, and D. Touzeau. Therapeutic drug monitoring of methadone in HIVinfected patients receiving protease inhibitors. AIDS 12: 2510-2511 (1998). 3. K.R. Dyer, D.J.R. Foster, J.M. White, A.A. Somogyi, A. Menelaou, and F. Bochner. Steady-state pharmacokinetics and pharmacodynamics in methadone maintenance patients: comparison of those who do and do not experience withdrawal and concentration-effect relationships. Clin. Pharmacol. Ther. 65:685-694 (1999). 4. M. Torrens, C. Castillo, L. San, E. del Moral, M.L. Gonz~lez, and R. de la Torre. Plasma methadone concentrations as an indicator of opioid withdrawal symptoms and heroin use in a methadone maintenance program. Drug AIc. Depend. 52:193-200 (1998). 5. K. Wolff, A.W.M. Hay, D. Raistrick, and R. Calvert. Steady-state pharmacokinetics of methadone in opioid addicts. Eur. J. Clin. Pharmacol. 44:189-194 (1993). 6. E. Angg~rd, L.M. Gunne, J. Homstrand, R.E.McMahon, C.G. Sandberg, and H.R. Sullivan. Disposition of methadone in methadone maintenance. Clin. Pharmacol. Ther. 17:258-266 (1975). 7. K. Verebely, J. Volavka, S. Mule, and R. Resnick. Methadone in man: pharmacokinetic and excretion studies in acute and chronic treatment. Clin. Pharmacol. Ther. 18:180-190 (1975). 8. M.I. Nilsson, E. ~,ngg~rd, J. Holmstrand, and L.M. Gunne. Pharmacokinetics of methadone during maintenance treatment: adaptive changes during the induction phase. Eur. J. Clin. Pharmacol. 22:343-349 (1982).

Journal of AnalyticalToxicology,Vol. 27, September2003 9. G.D. Bellward, RM. Warren, W. Howald, J.E.Axelson, and F.S.Abbott. Methadone maintenance: effect of urinary pH on renal clearance in chronic high and low doses. Clin. Pharmacol. Ther. 22: 92-99 (I 977). 10. J. McCarthy. Quantitative urine drug monitoring in methadone programs: Potential clinical uses. J. Psychoactive Drugs 26: 199-206 (1994). 11. H.R. Sullivan and S.L. Due. Urinary metabolites of d/-methadone in maintenance subjects. J. Medicinal Chem. 16:909-913 (1973). 12. M.J. Kreek, F.A. Bencsath, and F.H. Field. Effectsof liver diseaseon urinary excretion of methadone and metabolites in maintenance patients: quantitation by direct probe chemical ionization mass spectrometry. Biomedical Mass Spectrometry 7:385-395 (1980). 13. J.W. de Vos, P.J. Geerlings, W. van den Brink, J.G. Ufkes, and H. van Wilgenburg. Pharmacokinetics of methadone and its primary metabolite in 20 opiate addicts. Eur. J. Clin. Pharmacol. 48: 361-366 (1995). 14. D.E. Moody, M.E. Alburges, R.J. Parker, J.M. Collins, and J.M. Strong. The involvement of cytochrome P450 3A4 in the N-demethylation of L-~-acetylmethadol (LAAM), norLAAM, and methadone. Drug Metab. Dispos. 25:1347-1353 (1997). 15. Y. Oda and E.D. Kharasch. Metabolism of methadone and levoalpha-acetylmethadol (LAAM) by human intestinal cytochrome P450 3A4 (CYP3A4): potential contribution of intestinal metabolism to presystemic clearance and bioactivation. ]. Pharrnacol. Exp. Ther. 298:1021-1032 (2001). 16. D.J. Foster,A.A. Somogyi, and F. Bochner. Methadone N-demethylation in human liver microsomes: lack of stereoselectivity and involvement of CYP3A4. Br. J. Clin. PharmacoL 47:403-412 (1999). 17. M.E. Alburges, W. Huang, R.L. Foltz, and D.E. Moody. Determination of methadone and its N-demethylation metabolites in biological specimens by GC-PICI-MS. J. Anal ToxicoL 20:362-368

(1996). 18. E.T. Moolchan, A. Umbricht, and D. Epstein. Therapeutic drug monitoring in methadone maintenance: choosing a matrix. J. Addictive Dis. 20:55-73 (2001). 19. S. George and R.A. Braithwaite. A pilot study to determine the usefulness of the urinary excretion of methadone and its primary metabolite (EDDP) as potential markers of compliance in methadone detoxification programs. J. Anal. Toxicol. 23:81-435 (1999). 20. S. George, S. Parmar, C. Meadway, and R.A. Braithwaite. Application and validation of a urinary methadone metabolite (EDDP) immunoassay to monitor methadone compliance. Ann. Clin. Biochem. 37:350-354 (2000). 21. D.E. Moody, K.M. Monti, and A.C. Spanbauer. Long-term stability of abused drugs and antiabuse chemotherapeutic agents stored at -20~ ]. Anal. Toxicol. 23:535-540 (1999). 22. M.J. Kreek. Plasma and urine levels of methadone: comparison following four medication forms used in chronic treatment. N.Y. 5tate J. Med. 73:2773-2777 (1973). 23. F.R. Galloway and N.F. Bellet. Methadone conversion to EDDP during GC-MS analysis of urine samples. ]. Anal. Toxicol. 23: 615~19 (1999). 24. P.J.Orsulak, L.C. Akers, and N. Schuyler. Clinical application of the Cedia EDDP (methadone metabolite) assay. Poster presented at Joint SOFT-TIAFT International Meeting, Albuquerque, New Mexico, 1998. Abstract available at (accessed November 28, 2001 ).

Manuscript received August 23, 2002; revision received January 13, 2003.

341