Journal of Analytical Toxicology 2014;38:135 –142 doi:10.1093/jat/bku001 Advance Access publication February 4, 2014

Article

Urinary Diazepam Metabolite Distribution in a Chronic Pain Population Samantha Luk1, Rabia S. Atayee1,2, Joseph D. Ma1,2 and Brookie M. Best1,3* 1

Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San Diego (UC San Diego), 9500 Gilman Drive, MC 0719, La Jolla, CA, USA, 2Doris A. Howell Pain and Palliative Care Service, La Jolla, CA, USA, and 3UC San Diego Department of Pediatrics, Rady Children’s Hospital, San Diego, CA, USA

*Author to whom correspondence should be addressed. Email: [email protected]

Diazepam is often used as an adjuvant to pain therapy. Cytochrome P450 (CYP) 3A4 and 2C19 metabolize diazepam into the active metabolites: nordiazepam, temazepam and oxazepam. Owing to diazepam’s side-effect profile, mortality risk and potential for drug –drug interactions with CYP 3A4 and/or CYP 2C19 inhibitors, urine drug testing (UDT) could be a helpful monitoring tool. This was a retrospective data analysis that evaluated urine specimens from pain management practices for the distribution of diazepam metabolites with and without CYP 3A4 and 2C19 inhibitors. Intersubject nordiazepam, temazepam and oxazepam geometric mean fractions were 0.16, 0.34 and 0.47, respectively. Intrasubject geometric mean fractions were 0.157, 0.311 and 0.494, respectively. Sex, but not age or urinary pH, had an effect on metabolite fractions. Methadone significantly increased temazepam and oxazepam urinary fractions via CYP3A4 inhibition, whereas fluoxetine and esomeprazole increased nordiazepam fractions via CYP2C19 inhibition. Although more studies are needed, these results suggest the viability of UDT for increased monitoring for therapy and possible drug –drug interactions.

Introduction Diazepam is an adjuvant to opioid therapy to reduce pain-related anxiety and to treat muscle spasms in patients with pain. It is metabolized to nordiazepam by cytochrome P450 (CYP) 2C19 or to temazepam by CYP 3A4 (1). These metabolites are biotransformed to oxazepam by hydroxylation or demethylation (Figure 1) (2, 3). Nordiazepam, temazepam and oxazepam are pharmacologically active, and the latter two metabolites are commercially available medications. Diazepam and its metabolites bind to the gammaaminobutyric acid (GABA) receptor to enhance GABA transmission and affect downstream processes (1, 4). Concern is increasing over the potential for benzodiazepine dependence and the difficulty of weaning patients off of benzodiazepines. Diazepam has a long, variable half-life (20–50 h), which can result in prolonged sedation, ataxia, confusion, psychoses, hypotension, tremor and decreased respiratory rate. Many patients may resist discontinuing benzodiazepines because of the strong anxiolytic effects, high-dependent properties and fear of withdrawal symptoms. Withdrawal symptoms, including nausea, vomiting, anxiety, insomnia and tremor, can occur with medication courses as short as 2 weeks (5). When patients use diazepam with opioids, a common practice in pain patients, the potential for overdose and death is increased (6), and this particular combination was the most common cause of polysubstance overdose death in the USA from 2005 to 2009 (7). Studies conflict regarding the mechanism behind the interaction between benzodiazepines and opioids. Possibly, the pharmacodynamic interaction enhances and reinforces the reward pathway and euphoric feelings activated by

the opioids, which can lead to increased abuse and potential overdose (8). Consequently, benzodiazepine monitoring is highly encouraged. Benzodiazepines are also susceptible to CYP-mediated drug – drug interactions. Polymorphisms in CYP genes and drugs that inhibit CYP have exerted clinical and pharmacokinetic effects on diazepam and related benzodiazepines. Subjects with increased CYP 2C19 metabolism will emerge more quickly from general anesthesia induced by diazepam and sevoflurane (9). CYP 3A4 and 2C19 inhibitors increase plasma concentrations of diazepam and other benzodiazepines when co-administered (10–13). The clinical implication of the increased plasma concentrations is not well characterized and can be contradictory. For example, increased sedation has been reported when diazepam or other benzodiazepines were co-administered with CYP inhibitors (14). In contrast, some report no psychomotor dysfunction with increased plasma concentrations of benzodiazepines (11, 15). Since CYP 3A4 and CYP 2C19 inhibitors are commonly used, concurrent medication effects on diazepam disposition should be monitored. Non-invasive urine testing may be preferred for monitoring diazepam use. Although extensive studies describe diazepam and its metabolites in the blood (5, 16, 17), data on urinary values are limited and conflicting. Some studies report that temazepam is the primary urinary metabolite (18, 19), while others report that oxazepam is the primary metabolite (20, 21). In addition, standard concentrations of diazepam metabolites in the urine are not known (18–21). Furthermore, nordiazepam has been established as the primary metabolite in blood (5, 16). The aim of this retrospective data analysis was to examine urinary distribution of diazepam metabolites in a population of pain patients reported to be on diazepam and not any of its metabolites. The study also investigated the effects of CYP 3A4 and 2C19 substrates and inhibitors on the urinary distribution of diazepam metabolites.

Methods Subject selection The study was granted Institutional Review Board-exempt status by the University of California, San Diego Human Research Protection Program. Specimens collected between March 2008 and May 2011 were de-identified, and 927, 772 specimens with creatinine concentrations .20 mg/dL were screened for inclusion.

Inclusion criteria (i) Specimens with creatinine concentrations .20 mg/dL were used to ensure the validity of the urine specimen (22).

# The Author [2014]. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Figure 1. Diazepam metabolic pathway. Diazepam is metabolized to either nordiazepam by CYP 2C19 or temazepam by CYP3A4. Nordiazepam and temazepam are hydroxylated and demethylated to oxazepam. Oxazepam is then glucuronidated and excreted in the urine.

(ii) Physician reported that the patient was taking diazepam and not any of the diazepam metabolites (nordiazepam, temazepam and/or oxazepam). (iii) The specimen had detectable concentrations of at least one of the three diazepam metabolites. All values below the instrument cut-off were evaluated as zeros. Additional inclusion criterion for intersubject analyses (i) Single specimens from each subject’s first or only visit were used. Additional inclusion criterion for intrasubject analyses (i) The subject had five or more urine specimens from different clinic visits. Additional inclusion criteria for the CYP substrates/ inhibitor analyses (i) Only specimens from each subject’s first or only visit were used. (ii) The subject was also taking a CYP substrate or inhibitor as defined by the Flockhart CYP Drug Interactions Table and the FDA (23, 24). (a) Only strong or moderate CYP3A4 inhibitors were included. All CYP2C19 inhibitors were included (Table I). Specimen collection Urine specimens were collected from physician clinics monitoring patients on opioid therapy in accordance with established guidelines (5, 25). Specimens were shipped at room temperature by overnight courier. In some cases, urine specimens were refrigerated until shipment. Upon arrival, liquid chromatography– tandem mass spectrometry (LC –MS-MS) without prior

136 Luk et al.

immunoassay screening was conducted to detect a wide variety of drugs and drug metabolites.

Analytical procedures An Agilent 1200 series binary pump SL Liquid Chromatography system, well plate sampler, thermostatted column compartment, paired with an Agilent 6410 QQQ mass spectrometer and Agilent Mass Hunter software were used for the analysis of all drugs. An acetonitrile –aqueous formic acid gradient ran at 0.4 mL/min. A 2.1  50 mm, 1.8 mm Zorbax SB C 18 column was used for chromatography. The column temperature was 508C. Mobile phase A ¼ 0.1% formic acid in water and B ¼ 0.1% formic acid in acetonitrile. The Agilent 6410 Triple Quadrupole mass spectrometer was used in the positive electrospray ionization mode. The nitrogen drying gas temperature was 3508C, the flow was 12 L/min, nebulizer gas (nitrogen) pressure was 40 psi and the capillary voltage was 3,000 V. Dwell times were 50 ms. HPLC-grade water, acetonitrile, methanol and HPLC-grade formic acid were obtained from VWR (Westchester, PA, USA). Transition ions can be found in Table II. Ion ratios for the qualifier ions were accepted if the variance was ,20% from the target value. Deuterated internal standards of 100 mg/mL in methanol were obtained from Cerilliant Corporation (Round Rock, TX, USA). Calibration solutions of nordiazepam at 40, 400, 2,560 and 5,120 ng/mL; temazepam at 50, 500, 3,200 and 6,400 ng/mL; and oxazepam at 40, 400, 2,560 and 5,120 ng/mL were prepared by diluting the standards into synthetic urine (Microgenics Corp., Fremont, CA, USA). Deuterated internal standards were added to calibration solutions and to subject specimens to a final concentration of 1,200 ng/mL. Quality control specimens of 100 and 1,000 ng/mL were placed

Table I CYP substrates and inhibitors analyzed CYP3A4

CYP2C19

Substrates included

† † † † † † † † † † † † † † † † † † † † †

Alfentanil Alprazolam Amlodipine Aripiprazole Astemizole Atorvastatin Buspirone Cafergot Cerivastatin Chlorpheniramine Cilostazol Cisapride Codeine Cyclosporine Dapsone Dexamethasone Dextromethorphan Docetaxel Domperidone Eplerenone Estradiol

† † † † † † † † † † † † † † † † † † † † †

Felodipine Fentanyl Finasteride Gleevec Haloperidol Hydrocortisone Irinotecan Lercanidipine Lidocaine Lovastatin Methadone Midazolam Nateglinide Nifedipine Nisoldipine Nitrendipine Ondansetron Pimozide Progesterone Propranolol Quetiapine

† † † † † † † † † † † † † † † † † † † † †

Quinidine Quinine Risperidone Salmeterol Sildenafil Simvastatin Sirolimus Sorafenib Sunitinib Tacrolimus Tamoxifen Taxol Terfenadine Testosterone Torisel Trazodone Triazolam Vincristine Zaleplon Ziprasidone Zolpidem

† † † † † † † † † †

Amitriptyline Carisoprodol Citalopram Clomipramine Clopidogrel Cyclophosphamide Hexobarbital Imipramine Moclobemide Nelfinavir

† † † † † † † † † †

Nilutamide Phenobarbitone Primidone Progesterone Proguanil Propranolol R-mephobarbital S-mephenytoin Teniposide Warfarin

Inhibitors included

† † † † † † † † †

Aprepitant Atazanavir Boceprevir Ciprofloxacin Clarithromycin Conivaptan Darunavir Diltiazem Erythromycin

† † † † † † † † †

Fluconazole Fosamprenavir Grapefruit juice Imatinib Indinavir Itraconazole Ketoconazole Lopinavir Nefazodone

† † † † † † † †

Nelfinavir Posaconazole Ritonavir Saquinavir Telaprevir Telithromycin Verapamil Voriconazole

† † † † † † † † †

Chloramphenicol Cimetidine Esomeprazole Felbamate Fluoxetine Fluvoxamine Indomethacin Ketoconazole Lansoprazole

† † † † † † † † †

Modafinil Omeprazole Oxcarbazepine Pantoprazole Probenicid Rabeprazole Ticlopidine Topiramate Voriconazole

Table II LC –MS-MS standards Compound

Transition ions (amu)

Fragment voltage (V)

Collision energy (V)

Dwell time (ms)

Nordiazepam

271 ! 165 271 ! 140 287 ! 241 287 ! 104 292.1 ! 246.1 301 ! 255 301 ! 177 306 ! 206

160

20

27

94

19

14

94 120

35 35

– 73

120

35

–

Oxazepam Oxazepam D5 Temazepam Temazepam D5

in each run. Upper limits of linearity were 100,000 ng/mL for all the analytes, determined by the dilution of the Cerilliant-certified standards into synthetic urine (Microgenics negative control). Quantitation of specimens with concentrations .100,000 ng/mL was estimated through linear extrapolation. Ion suppression was corrected by use of the deuterated internal standards. Specific experiments in accordance with the experiments in Pesce et al. (26) were conducted to rule out large variances due to ion suppression. Efficiency of the hydrolysis procedure was determined with morphine glucuronide controls set at 1,000 and 100 ng/mL. Morphine was used because this assay tests for the presence of multiple drugs that are glucuronidated. Hydrolysis of the control material was considered acceptable if the value of the recovered morphine was .90% of the nominal concentration. Specimens were prepared for injection by incubating 25 mL of urine with 50 units of b-glucuronidase Type L-II from Patella vulgata (keyhole limpet) Sigma Product number G 8132 (Sigma-Aldrich Corp., Saint Louis, MO, USA) in 50 mL of 0.4 M

( pH 4.5) acetate buffer for 3 h at 458C. Samples were centrifuged for 10 min at 3,000 rpm to remove turbidity. Five microliters of the specimen were injected using a CT-PAL HTS injection device (CTC Analytics AG, Switzerland). The interassay coefficient of variation (% CV) for all the analytes at the low and high ends of the quantification curve was 15%. At 1,000 ng/mL, the coefficients of variation for nordiazepam, oxazepam and temazepam were 10.6, 7.5 and 4.5%, respectively, based on the analysis of .100 runs. Similarly, at 100 ng/mL, the coefficients of variation for nordiazepam, oxazepam and temazepam were 13.7, 8.5 and 6.1%, respectively. The allowed variance for the quantitative determination in these surveys is ,20%. All quantitative data were obtained from calibration curves with R 2 . 0.95. Most were 0.99. The lower limits of quantitation for diazepam, nordiazepam and oxazepam were 40, 50 and 40 ng/mL, respectively.

Statistical analysis Statistical analyses were performed using OriginPro 8.5.1 (OriginLab, Northampton, MA, USA) and Microsoft Excel 2010 (Microsoft Corp., Redmond, WA, USA). Urine data were creatinine-corrected to account for variable body mass, hydration status and daily water intake (27). Owing to the low urinary excretion of diazepam (17), metabolite fractions were used to assess urinary diazepam metabolite excretion. The fraction of each metabolite to the total moles of excreted diazepam metabolites and that of each metabolite to the other metabolites were calculated. Logarithmic values were used to approximate a Gaussian distribution. Two-sided two-sample t-tests were used to determine any statistically significant differences. Urine Diazepam Metabolites in Pain Patients 137

Figure 2. Fraction of diazepam metabolites for the inter-subject population. The fraction of nordiazepam is in white, temazepam in gray and oxazepam in black. The mean fraction of nordiazepam is 0.16, temazepam 0.34 and oxazepam 0.47 (*P , 0.00001 for each pairwise comparison).

Results Intersubject analysis The log distribution of the fraction of nordiazepam, temazepam and oxazepam from 22,509 specimens from unique subjects approximated a Gaussian distribution (Figure 2). The geometric mean fractions of excreted nordiazepam, temazepam and oxazepam were 0.16, 0.34 and 0.47 (P , 0.00001 for each pairwise comparison). Of the total, 92% of the population had a larger percentage of temazepam than nordiazepam, 82% had a larger percentage of oxazepam than nordiazepam and 71% had a larger percentage of oxazepam than temazepam. Most (86.8%) had all three metabolites detectable in urine (Figure 3A and B). Approximately 13.2% did not have all three metabolites in urine. To determine the characteristics of this group, the data were stratified for total moles (Figure 3C). The urine specimens of this population with one or more of the undetectable metabolites also had low total excreted moles. Semi-logarithmic scatterplots of the total excreted moles of diazepam metabolites vs. each metabolite fraction demonstrated no relationship. A linear fit for each metabolite fraction showed that slopes were close to zero and R 2 , 0.15. Age was not related to metabolite concentrations for any of the diazepam metabolites. Urinary pH and the fraction of diazepam metabolite also demonstrated no statistically significant association. Urinary excretion of diazepam metabolites was compared between females (n ¼ 12,026) and males (n ¼ 9,763). The geometric mean fraction of nordiazepam in females was 0.157 (95% confidence interval [95% CI] 0.155–0.158), which was 4.8% lower than those in males, 0.165 (95% CI 0.160–0.164; P , 0.01). The temazepam fraction in females was 0.337 (95% CI 0.334–0.339), which was 7.4% higher than in males, at 0.312 (95% CI 0.309– 0.315; P , 0.00001). The oxazepam fraction in females was 0.452 (95% CI 0.381–0.573), 4.0% lower than the fraction in males, at 0.471 (95% CI 0.397–0.603; P , 0.00001). 138 Luk et al.

Figure 3. Diazepam metabolites detected in urinary specimens. (A) Most of the population on diazepam and not any of the diazepam metabolites have all three diazepam metabolites in the urinary specimen. However, 13.20% of the population do not make one or more of the diazepam metabolites. Of these subjects, 90% of the subjects have oxazepam in the urinary specimen. (B) Most of the population makes all three metabolites. However, some subjects are missing one or more of the metabolites. These subjects are located on the lines where nordiazepam, temazepam or oxazepam equal zero (as indicated on the plot). (C) Fractions are stratified for total moles. Subjects who are missing one or more of the diazepam metabolites tend to be those that have low total moles, suggesting that the subjects are close to the beginning or end of a dosing interval.

Table III Fractions of diazepam metabolites in urine by groups Groups for comparison

N

Fraction of nordiazepam

Fraction of temazepam

Fraction of oxazepam

Overall populationa % CV No CYP 3A4 substrates/inhibitorsa CYP 3A4 substratesa % Difference vs. control (P-value) CYP 3A4 inhibitorsa % Difference vs. control (P-value) No CYP 2C19 substrates/inhibitorsa CYP 2C19 substratesa % Difference vs. control CYP 2C19 inhibitorsa % Difference vs. control Esomeprazolea % Difference vs. control Fluoxetinea % Difference vs. control No CYP 3A4 or 2C19 substrates/inhibitorsa CYP 3A4 and CYP 2C19 substrate/inhibitora % Difference vs. control

21,789

0.164 (0.163, 0.165) 69.7 0.167 (0.163, 0.165) 0.147 (0.144, 0.150) 212.0 (P , 0.00001) 0.167 (0.162, 0.193) 0 (P . 0.05) 0.166 (0.164, 0.167) 0.161 (0.156, 0.166) 23.0 (P . 0.05) 0.150 (0.145, 0.154) 29.6 (P , 0.00001) 0.139 (0.137, 0.153) 216.3 (P , 0.00001) 0.145 (0.131, 0.149) 212.7 (P , 0.00001) 0.169 (0.168, 0.171) 0.183 (0.154, 0.217) þ7.7 (P . 0.05)

0.336 (0.334, 0.338) 48.4 0.336 (0.335, 0.338) 0.338 (0.334, 0.343) þ0.6 (P . 0.05) 0.280 (0.252, 0.292) 216.7 (P , 0.00001) 0.334 (0.332, 0.336) 0.346 (0.339, 0.353) þ3.47 (P , 0.05) 0.352 (0.346, 0.359) þ5.1 (P , 0.00001) 0.355 (0.345, 0.371) þ5.9 (P , 0.00001) 0.358 (0.340, 0.370) þ6.7 (P , 0.00001) 0.335 (0.333, 0.337) 0.273 (0.237. 0.314) 218.5 (P , 0.005)

0.468 (0.465, 0.471) 41.2 0.466 (0.465, 0.471) 0.481 (0.464, 0.488) þ3.1 (P , 0.00001) 0.530 (0.497, 0.557) þ12.1 (P , 0.05) 0.468 (0.465, 0.471) 0.473 (0.461, 0.484) þ1.06 (P . 0.05) 0.469 (0.459, 0.476) þ0.2 (P . 0.05) 0.457 (0.450, 0.488) 21.1 (P . 0.05) 0.469 (0.437, 0.478) þ0.2 (P . 0.05) 0.465 (0.462, 0.468) 0.558 (0.489, 0.626) þ16.7 (P , 0.05)

18,218 3,477 93 18,844 1,175 1,770 471 424 16,138 36

a

Geometric mean (95% CI).

Intrasubject analysis Eight hundred and thirty-six unique subjects with 6,433 specimens had an average of eight visits (range 5 –24). The geometric mean urinary fraction of nordiazepam was 0.157 (95% CI 0.152 – 0.161; % CV ¼ 33.6). The fraction of temazepam was 0.311 (95% CI 0.304–0.318; % CV ¼ 20.8), and that of oxazepam was 0.494 (95% CI 0.486 –0.501; % CV ¼ 18.7).

CYP substrate and inhibitor analyses Subjects taking concurrent diazepam with CYP3A4 substrates (n ¼ 3,477), strong or moderate CYP3A4 inhibitors (n ¼ 93), or no CYP3A4 substrates/inhibitors (n ¼ 18,218) were compared. Significant but very slight differences in nordiazepam and oxazepam fractions were seen for those on CYP3A4 substrates. Significant differences in temazepam and oxazepam fractions were noted for those on CYP3A4 inhibitors (Table III). When individual CYP3A4 substrates and inhibitors were evaluated, only subjects taking methadone were significantly different from controls. For the CYP2C19 and diazepam analyses, 1,175 subjects took CYP2C19 substrates, 1,770 took CYP2C19 inhibitors and 18,844 were not on any CYP2C19 substrates/inhibitors. The geometric mean fraction of temazepam was significantly different for those on CYP2C19 substrates. The nordiazepam and temazepam fractions were significantly different for those on CYP2C19 inhibitors. Of the individual CYP2C19 substrates/inhibitors, only fluoxetine and esomeprazole demonstrated significant differences from controls (Table III). Thirty-six subjects were concurrently on both CYP3A4 and CYP2C19 substrates and/or inhibitors with diazepam and were compared with 16,138 control subjects not on any CYP3A4/CYP2C19 substrates/inhibitors. The temazepam and oxazepam fractions were significantly different in this population (Table III).

Discussion Intersubject analysis In this study, oxazepam comprised the largest fraction of urinary diazepam metabolite, while nordiazepam was the smallest.

These findings support both Schwartz and Smith-Kielland’s data on oxazepam being the major urinary metabolite of diazepam (20, 21) but contrast with Arnold and Chiba, who found the major metabolite to be temazepam (18, 19). The wide variation in prior studies could be due to varied study designs and small sample sizes (from 2 to 27 subjects), which may not represent the general population when compared with the 22,509 subjects in this study. Prior studies included young, healthy subjects, while this study had a wide range of patient ages and co-morbidities. In addition, this study utilized LC–MS-MS to quantify urine concentrations, whereas previous studies utilized liquid and thin-layer chromatography. Specimens were creatinine-corrected in the current study, which was not reported in past studies (18–21). Although the distributions of the three metabolites overlap, the mean fraction of each metabolite was significantly different from each other. The % CV was greatest for nordiazepam and least for oxazepam. The increased variability of the nordiazepam fraction could result from the polymorphic CYP enzymes involved in the formation of nordiazepam (10). Even though log-transformed distributions of metabolites approximate Gaussian distributions, a group of specimens (n ¼ 1,580, 7%) had an oxazepam fraction equal to 1, and contained exclusively oxazepam. When analyzed with Grubb’s outlier test, these data were not significant outliers. The findings were similar whether these specimens were included or excluded. Most of the population had all three metabolites detectable in urine. However, 13.2% did not have one or more of the diazepam metabolites. When stratifying for total moles of excreted metabolite, these specimens appear to have low total excreted moles of metabolites, suggesting that these were collected at the beginning or end of a dosing interval. Most were likely collected near the end of a dosing interval due to the high number that had detectable oxazepam.

Potential factors influencing urinary diazepam and metabolites One possible factor that could affect diazepam metabolism is the total moles of excreted metabolite. However, no relationship Urine Diazepam Metabolites in Pain Patients 139

was seen between the total moles of excreted metabolite and the fraction of each metabolite. Although this does not preclude an effect of diazepam load on the metabolic system, this suggests that at therapeutic doses, the diazepam metabolic pathways are not saturated. Age and urinary pH do not appear to affect the fractions of metabolites. Although kidney function decreases with age (28), the distribution of metabolites in the urine remains unchanged. In addition, urinary diazepam excretion does not have any acid – base effects. Notable differences were seen, though, between females and males. Females had higher fractions of temazepam and lower fractions of nordiazepam and oxazepam compared with males. Females might have increased CYP3A4-mediated diazepam metabolism when compared with males. In prior studies with other drug compounds, females have demonstrated greater CYP3A4 activity, but this observation has not been confirmed with diazepam (29, 30). The magnitude of differences between females and males was slight and no previous evidence has indicated differential clinical effects of diazepam between sexes, suggesting that the difference found in this study is too small to have clinical relevance.

Intrasubject vs. intersubject analyses The intra-subject population had a similar diazepam distribution to the inter-subject population. The mean fractions were distinct from each other as seen from the non-overlapping 95% CIs between the intra-subject fractions of nordiazepam, temazepam and oxazepam. The oxazepam fraction was the largest at 0.50 and the nordiazepam fraction was the smallest at 0.17. The similarity to the inter-subject population suggests that the intersubject trends are also true for individual subjects. The lower intra-subject variability in the intra-subject population suggests that urinary diazepam excretion is more dependent on individual factors, such as CYP polymorphisms, instead of factors such as urinary pH.

CYP substrate and inhibitor analyses With CYP3A4 substrates, the fraction of nordiazepam was lower, while the fraction of temazepam remained unchanged. The fraction of temazepam could theoretically decrease due to competition for binding sites with concurrent administration of a CYP3A4 substrate. However, the fraction of nordiazepam decreased instead with a slight increase in the fraction of oxazepam. Perhaps, the fraction of temazepam remained unchanged due to the high contribution of CYP3A4 in diazepam metabolism. Schmider et al. (31) postulated that CYP3A4 can constitute 60% of all diazepam metabolism in single-dose studies, and the percentage can be even greater in clinical practice with multiple doses. Thus, with a CYP3A4 substrate, the change in temazepam would be negligible due to the higher affinity of diazepam for CYP3A4. Alternatively, urinary fractions of temazepam may not be robust enough to detect subtle perturbations in CYP3A4 activity that would be expected with a CYP3A4 substrate. Another possibility is that, with the increased load on CYP3A4, the body might compensate by shunting the parent diazepam into the CYP2C19 pathway to form nordiazepam and quickly converts the nordiazepam into oxazepam, resulting in the small decrease in nordiazepam and increase in oxazepam. 140 Luk et al.

On the other hand, when a strong or moderate CYP3A4 inhibitor was given, the temazepam fraction was lower and the oxazepam fraction higher. Since the CYP enzyme responsible for the formation of temazepam is inhibited, the end product cannot be formed as readily. Diazepam may be shunted into the CYP2C19 pathway to form more nordiazepam. However, the nordiazepam fraction did not change, suggesting that nordiazepam rapidly converts into oxazepam under conditions of CYP enzymatic stress. Perhaps, under high enzymatic stress, diazepam can also engage alternative metabolic pathways to bypass CYP3A4 inhibition. Minimal differences were found between individual CYP3A4 substrates and inhibitors when compared with the population without any concurrent CYP3A4 substrates or inhibitors, except for methadone, perhaps due to higher CYP affinity for methadone when compared with diazepam. The minor decrease in the temazepam fraction with CYP 2C19 substrates may have been due to the increased load on CYP 2C19. CYP 2C19 may have large capacity and can handle the increased metabolic load of an additional CYP 2C19 substrate. With CYP 2C19 inhibitors, the nordiazepam fraction decreased and the temazepam fraction increased. When the CYP 2C19 enzyme is inhibited, less nordiazepam is made, possibly shifting the parent diazepam compound to the CYP 3A4 pathway instead. CYP 2C19 inhibitors have the greatest inhibition on the formation of temazepam in the blood when compared with that of nordiazepam (32, 33), which appears to be echoed in urine. However, as with the CYP 2C19 substrates, changes in diazepam metabolites are minor, perhaps due to the large capacity of the CYP 2C19 pathway. When evaluating each individual CYP2C19 substrate and inhibitor, only fluoxetine and esomeprazole had significant changes. Fluoxetine and esomeprazole are potent inhibitors of CYP2C19 (34, 35). The changes seen with concurrent fluoxetine and diazepam administration shed additional light on the delirium side effect observed with concurrent administration of both drugs (10). Another interesting finding is that omeprazole concurrently given with diazepam did not have any significant effect on the urinary metabolite disposition, even though it is thought to be a strong CYP 2C19 inhibitor (23), has been shown to increase plasma concentrations of diazepam (10) and can potentially inhibit the CYP 3A family (36). One possibility is that omeprazole’s inhibition is time- and concentration-dependent, and that maximal CYP 2C19 inhibition was not attained. When CYP 3A4 and 2C19 substrates and/or inhibitors were concurrently administered with diazepam, the nordiazepam fraction did not change, the temazepam fraction decreased and the oxazepam fraction increased. The effect seen with this particular medication pattern is similar to the effect seen with CYP 3A4 inhibitors and concurrent diazepam administration, except that the differences are much larger with both a CYP 3A4 and a 2C19 substrate and/or inhibitor. This suggests that CYP 3A4 plays a larger role in diazepam metabolism than CYP 2C19, especially since the inhibition pattern more closely resembles that seen with CYP 3A4 inhibitors. The clinical implications of the shifts in urinary diazepam are not well understood. Plasma concentrations do not directly correlate with urine data in past studies. Furthermore, changes observed in urine may be too slight to have clinical relevance, especially since some previous studies did not observe effects on psychomotor function with increased benzodiazepine concentrations

(11, 15). Nevertheless, shifts in diazepam metabolism that are observed in urine tests with CYP substrates and inhibitors can be very helpful for clinicians, especially since the potential for unknown adverse drug interactions with CYP substrates/inhibitors is still a concern. Limitations One study limitation is the lack of dosage data. Variations seen in the population may be due to external factors, differences in doses or both. There may also be variations due to the timing of the diazepam administration in relation to the CYP substrate/ inhibitor administration. Another limitation is the lack of plasma data to correlate with observed urine concentrations. Urine specimens were taken at variable times within the dosing regimen. A select population may be early or late in the dosing interval, resulting in different metabolite distributions when compared with the larger population. However, without definitive post-dose testing times, whether these subjects have different diazepam metabolite distributions due to the timing of specimen acquisition or inherent differences in diazepam metabolism is unknown. In addition, the medication list was reported by physicians and was not independently verified by a third party. The analysis could have included some subjects who were not on diazepam or the studied CYP substrates/inhibitors. Some subjects included in the study may have also been on diazepam metabolites that were unreported by the patient to the physician. Some subjects who were actually taking diazepam may also have been excluded from the study on the basis of the physician-reported medication. In addition, the CYP genotypes for 3A4 and 2C19 were not known for each subject. Although these limitations are important, several caveats also need to be considered when interpreting these limitations. While the dosing data for diazepam and the co-administered medications are lacking, out of the total population analyzed, only 13.2% of the subjects did not have all three diazepam metabolites detectable in the urine (Figure 3A). Subjects without all three compounds detected in urine would most likely consist of subjects with urine samples drawn early in the dosing interval or late in the dosing interval, or subjects who actually only took one of the diazepam metabolites instead of diazepam. This population is small compared with the larger population with urine samples containing all three metabolites. In addition, diazepam’s long half-life (20–50 h) may diminish the consequence of diazepam administration timing, especially in the context of multiple doses, in the metabolic pathway. Even though the CYP genotype is unknown and has theoretical implications for diazepam metabolism, there also has not yet been any data to indicate the clinical relevance of CYP polymorphisms in diazepam metabolism. Although these limitations must be taken into consideration when interpreting the results, advantages of this study were a large sample size and sensitive, comprehensive drug and metabolite assay results within each subject. In addition, the results may reflect the type of uncontrolled situations in which clinicians would normally practice. Thus, despite the aforementioned limitations, the results still have practical applicability.

Conclusions This study described urinary distribution of diazepam metabolites in a large sample of pain patients. Unlike previous studies,

oxazepam was the largest fraction and nordiazepam was the smallest fraction. These trends were seen in the majority of the inter- and intra-subject population, suggesting that this particular pattern of diazepam excretion can potentially be extrapolated to the general population. In addition, the variability in diazepam urinary excretion may be affected more by inherent subject-specific factors such as genetic polymorphisms in CYP enzymes when compared with changing factors such as urinary pH. However, the inter- and intra-subject mean fractions of metabolites were similar, which suggests that the current guidelines for diazepam administration which do not take genetics, sex or other factors into consideration should be applicable to most patients. Furthermore, although this study was not able to make direct links between the urine data and plasma concentrations, it was possible to observe changes in the diazepam metabolism as a result of co-medications. Thus, urine testing remains viable as a potential tool, in conjunction with other medication monitoring techniques, for clinicians to monitor their patients for potential adverse effects. These results can serve as guideposts for clinicians interpreting urine test results for diazepam. By knowing how their patient compares with the general population, clinicians can better adjust therapy for each patient’s needs, especially in light of the difficulty of weaning patients off of diazepam and the potential for drug –drug interactions in patients on multiple medications.

Acknowledgments The authors acknowledge Amadeo J. Pesce, PhD, DABCC, for his crucial guidance and support throughout the entirety of this study. Urine specimens were tested and provided by Millennium Laboratories. Dr Joseph D. Ma is a paid consultant of Millennium Laboratories, Inc.

Funding This work was supported in part by an educational grant provided by the University of California, San Diego Skaggs School of Pharmacy and Pharmaceutical Sciences from an unrestricted gift from the Millennium Research Institute (to S.L.).

References 1. Di Piero, V., Ferracuti, S., Sabatini, U., Tombari, D., Di Legge, S., Pantano, P. et al. (2001) Diazepam effects on the cerebral responses to tonic pain: a SPET study. Psychopharmacology (Berlin), 158, 252–258. 2. Andersson, T., Miners, J.O., Veronese, M.E., Birkett, D.J. (1994) Diazepam metabolism by human liver microsomes is mediated by both S-mephenytoin hydroxylase and CYP3A isoforms. British Journal of Clinical Pharmacology, 38, 131–137. 3. Greenblatt, D.J., Ochs, H.R., Lloyd, B.L. (1980) Entry of diazepam and its major metabolite into cerebrospinal fluid. Psychopharmacology (Berlin), 70, 89 –93. 4. Licata, S.C., Rowlett, J.K. (2008) Abuse and dependence liability of benzodiazepine-type drugs: GABA(A) receptor modulation and beyond. Pharmacology Biochemistry, and Behavior, 90, 74– 89. 5. Committee on the Review of Medicines. (1980) Systematic review of the benzodiazepines. Guidelines for data sheets on diazepam, chlordiazepoxide, medazepam, clorazepate, lorazepam, oxazepam, temazepam, triazolam, nitrazepam, and flurazepam. Committee on the Review of Medicines. British Medical Journal, 280, 910–912.

Urine Diazepam Metabolites in Pain Patients 141

6. Kuryshev, Y.A., Bruening-Wright, A., Brown, A.M., Kirsch, G.E. (2010) Increased cardiac risk in concomitant methadone and diazepam treatment: pharmacodynamic interactions in cardiac ion channels. Journal of Cardiovascular Pharmacology, 56, 420– 430. 7. Calcaterra, S., Glanz, J., Binswanger, I.A. (2013) National trends in pharmaceutical opioid related overdose deaths compared to other substance related overdose deaths: 1999–2009. Drug and Alcohol Dependence, 131, 263– 270. 8. Jones, J.D., Mogali, S., Comer, S.D. (2012) Polydrug abuse: a review of opioid and benzodiazepine combination use. Drug and Alcohol Dependence, 125, 8–18. 9. Inomata, S., Nagashima, A., Itagaki, F., Homma, M., Nishimura, M., Osaka, Y. et al. (2005) CYP2C19 genotype affects diazepam pharmacokinetics and emergence from general anesthesia. Clinical Pharmacology and Therapeutics, 78, 647– 655. 10. Desta, Z., Zhao, X., Shin, J.G., Flockhart, D.A. (2002) Clinical significance of the cytochrome P450 2C19 genetic polymorphism. Clinical Pharmacokinetics, 41, 913–958. 11. Kosuge, K., Jun, Y., Watanabe, H., Kimura, M., Nishimoto, M., Ishizaki, T. et al. (2001) Effects of CYP3A4 inhibition by diltiazem on pharmacokinetics and dynamics of diazepam in relation to CYP2C19 genotype status. Drug Metabolism and Disposition: The Biological Fate of Chemicals, 29, 1284– 1289. 12. Ozdemir, M., Aktan, Y., Boydag, B.S., Cingi, M.I., Musmul, A. (1998) Interaction between grapefruit juice and diazepam in humans. European Journal of Drug Metabolism and Pharmacokinetics, 23, 55 –59. 13. Zhou, S.F., Xue, C.C., Yu, X.Q., Li, C., Wang, G. (2007) Clinically important drug interactions potentially involving mechanism-based inhibition of cytochrome P450 3A4 and the role of therapeutic drug monitoring. Therapeutic Drug Monitoring, 29, 687– 710. 14. Dresser, G.K., Spence, J.D., Bailey, D.G. (2000) Pharmacokineticpharmacodynamic consequences and clinical relevance of cytochrome P450 3A4 inhibition. Clinical Pharmacokinetics, 38, 41–57. 15. Kato, K., Yasui-Furukori, N., Fukasawa, T., Aoshima, T., Suzuki, A., Kanno, M. et al. (2003) Effects of itraconazole on the plasma kinetics of quazepam and its two active metabolites after a single oral dose of the drug. Therapeutic Drug Monitoring, 25, 473–477. 16. Greenblatt, D.J., Harmatz, J.S., Friedman, H., Locniskar, A., Shader, R.I. (1989) A large-sample study of diazepam pharmacokinetics. Therapeutic Drug Monitoring, 11, 652– 657. 17. Zingales, I.A. (1973) Diazepam metabolism during chronic medication unbound fraction in plasma, erythrocytes and urine. Journal of Chromatography, 75, 55 –78. 18. Arnold, E. (1975) A simple method for determining diazepam and its major metabolites in biological fluids: application in bioavailability studies. Acta Pharmacologica Toxicology (Copenhagen), 36, 335–352. 19. Chiba, K., Horii, H., Chiba, T., Kato, Y., Hirano, T., Ishizaki, T. (1995) Development and preliminary application of high-performance liquid chromatographic assay of urinary metabolites of diazepam in humans. Journal of Chromatography B: Biomedical Sciences and Applications, 668, 77– 84. 20. Schwartz, M.A., Koechlin, B.A., Postma, E., Palmer, S., Krol, G. (1965) Metabolism of diazepam in rat, dog, and man. The Journal of Pharmacology and Experimental Therapeutics, 149, 423–435. 21. Smith-Kielland, A., Skuterud, B., Olsen, K.M., Morland, J. (2001) Urinary excretion of diazepam metabolites in healthy volunteers and

142 Luk et al.

22.

23.

24.

25.

26.

27.

28. 29.

30.

31.

32.

33.

34.

35.

36.

drug users. Scandinavian Journal of Clinical and Laboratory Investigation, 61, 237–246. Cook, J.D., Caplan, Y.H., LoDico, C.P., Bush, D.M. (2000) The characterization of human urine for specimen validity determination in workplace drug testing: a review. Journal of Analytical Toxicology, 24, 579–588. United States Food and Drug Administration. (2006) Drug Development and Drug Interactions: Table of Substrates, Inhibitors and Inducers. US Food and Drug Administration. Silver Spring, MD. Flockhart, D.A. (2007) Drug Interactions: Cytochrome P450 Drug Interaction Table. Indiana University School of Medicine. http:// medicine.iupui.edu/clinpharm/ddis/clinical-table/ (Accessed Jul 29, 2011). Washington State Agency Medical Director’s Group. (2010) Interagency Guideline on Opioid Dosing for Chronic Non-cancer Pain: An Educational Aid to Improve Care and Safety with Opioid Therapy. Olympia, WA: AMDG. Pesce, A., West, C., West, R., Latyshev, S., Masters-Moore, D., Friel, P. et al. (2012) Analytical considerations when monitoring pain medications by LC–MS/MS. Journal of Analytical and Bioanalytical Techniques S5(003), 1–55. Cone, E.J., Caplan, Y.H., Moser, F., Robert, T., Shelby, M.K., Black, D.L. (2009) Normalization of urinary drug concentrations with specific gravity and creatinine. Journal of Analytical Toxicology, 33, 1–7. Weinstein, J.R., Anderson, S. (2010) The aging kidney: physiological changes. Advances in Chronic Kidney Disease, 17, 302– 307. Tanaka, E. (1999) Gender-related differences in pharmacokinetics and their clinical significance. Journal of Clinical Pharmacy and Therapeutics, 24, 339– 346. Scandlyn, M.J., Stuart, E.C., Rosengren, R.J. (2008) Sex-specific differences in CYP450 isoforms in humans. Expert Opinion on Drug Metabolism and Toxicology, 4, 413– 424. Schmider, J., Greenblatt, D.J., von Moltke, L.L., Shader, R.I. (1996) Relationship of in vitro data on drug metabolism to in vivo pharmacokinetics and drug interactions: implications for diazepam disposition in humans. Journal of Clinical Psychopharmacology, 16, 267–272. Jones, H.M., Hallifax, D., Houston, J.B. (2004) Quantitative prediction of the in vivo inhibition of diazepam metabolism by omeprazole using rat liver microsomes and hepatocytes. Drug Metabolism and Disposition: The Biological Fate of Chemicals, 32, 572–580. Zomorodi, K., Houston, J.B. (1996) Diazepam-omeprazole inhibition interaction: an in vitro investigation using human liver microsomes. British Journal of Clinical Pharmacology, 42, 157–162. Frelinger, A.L. III., Lee, R.D., Mulford, D.J., Wu, J., Nudurupati, S., Nigam, A. et al. (2012) A randomized, 2-period, crossover design study to assess the effects of dexlansoprazole, lansoprazole, esomeprazole, and omeprazole on the steady-state pharmacokinetics and pharmacodynamics of clopidogrel in healthy volunteers. Journal of the American College of Cardiology, 59, 1304– 1311. Jeppesen, U., Gram, L.F., Vistisen, K., Loft, S., Poulsen, H.E., Brøsen, K. (1996) Dose-dependent inhibition of CYP1A2, CYP2C19 and CYP2D6 by citalopram, fluoxetine, fluvoxamine and paroxetine. European Journal of Clinical Pharmacology, 51, 73 –78. Funck-Brentano, C., Becquemont, L., Lenevu, A., Roux, A., Jaillon, P., Beaune, P. (1997) Inhibition by omeprazole of proguanil metabolism: mechanism of the interaction in vitro and prediction of in vivo results from the in vitro experiments. The Journal of Pharmacology and Experimental Therapeutics, 280, 730–738.