Kinetic and dynamic interaction study of zolpidem with ketoconazole, itraconazole, and fluconazole

Background: Azole antifungal agents may impair hepatic clearance of drugs metabolized by cytochrome P450-3A isoforms. The imidazopyridine hypnotic agent zolpidem is metabolized in humans in part by P450-3A, as well as by a number of other cytochromes. Potential interactions of zolpidem with 3 com- monly prescribed azole derivatives were evaluated in a controlled clinical study. Methods: In a randomized, double-blind, 5-way, crossover, clinical pharmacokinetic-pharmacodynamic study, 12 volunteers received (A) zolpidem placebo plus azole placebo, (B) 5 mg zolpidem plus azole placebo (C) zolpidem plus ketoconazole, (D) zolpidem plus itraconazole, and (E) zolpidem plus fluconazole. Results: Mean apparent oral clearance of zolpidem when given with placebo was 422 mL/min, and elimi- nation half-life was 1.9 hours. Clearance was significantly reduced to 250 mL/min when zolpidem was given with ketoconazole, and half-life was prolonged to 2.4 hours. Coadministration of zolpidem with itraconazole or fluconazole also reduced clearance (320 and 338 mL/min), but differences compared to the zolpidem plus placebo treatment did not reach significance. Zolpidem-induced benzodiazepine ago- nist effects (increased electrocardiographic beta activity, digit-symbol substitution test impairment, and delayed recall) during the first 4 hours after dosage were enhanced by ketoconazole but not by itracona- zole or fluconazole. Conclusion: Coadministration of zolpidem with ketoconazole impairs zolpidem clearance and enhances its benzodiazepine-like agonist pharmacodynamic effects. Itraconazole and fluconazole had a small influence on zolpidem kinetics and dynamics. The findings are consistent with in vitro studies of differentially impaired zolpidem metabolism by azole derivatives. (Clin Pharmacol Ther 1998;64:661-71.)

David J. Greenblatt, Lisa L. von Moltke, Jerold S. Harmatz, Polyxane Mertzanis, Jennifer A. Graf, Anna Liza B. Durol, Molly Counihan, Barbara Roth-Schechter, and Richard I. Shader Boston and Dover, Mass

From the Department of Pharmacology and Experimental Therapeu- tics and the Division of Clinical Pharmacology, Tufts University School of Medicine and New England Medical Center Hospital. Boston, and Boston Research and Science Consulting, Dover.

Supported in part by grants MH-34233 and DA-05258 from the Department of Health and Human Services, by grant RR-00054 supporting the General Clinical Research Center, Tufts University School of Medicine and New England Medical Center Hospital, and by a grant-in-aid from Lorex Pharmaceuticals, Chicago, 111. Dr. van Moltke is the recipient of a Scientist Development Award (K21-MH-01237) from the Department of Health and Human Services.

Received for publication June 30, 1998; accepted Sept 18, 1998. Reprint requests: David J. Greenblatt. MD, Department of Pharma-

cology, and Experimental Therapeutics, Tufts University School of Medicine, 136 Harrison Ave. Boston, MA 021 1 I. E-mail: Dgreenblatt@infonet.tufta.edu

Copyright 0 1998 by Mosby, Inc. 0009-9236/98/$5.00 + 0 13/l/94703

The imidazopyridine derivative zolpidem is exten- sively prescribed in clinical practice for the treatment of sleep disorders.t-4 Its mechanism of action involves GABA-benzodiazepine receptor agonism, with relative selectivity for the Type 1 (BZ- 1) class of central benzo- diazepine receptors.” The elimination half-life of zolpi- dem is in the range of 1% to 4 hours, and it therefore is classified as a short half-life hypnotic.6,7 The initial meta- bolic transformation of zolpidem in humans involves hydroxylation at 3 different sites on the molecule.@ These reactions are mediated mainly but not entirely by cytochrome P450-3A isoforms.8.9 Available evidence suggests that P450-2C9 may make a significant contri- bution to metabolic clearance of zolpidem at therapeuti- cally relevant substrate concentrations, with additional possible contributions of 1 A2, 2D6, and 2C 19.9

661

662 Greenblatt et al. CLINICAL PHARMACOLOGY & THERAPEUTICS

DECEMHER 1998

Pharmacokinetic drug interactions involving inhibition of cytochrome P450 activity have received considerable recent attention. Antifungal agents of the azole class are of importance as inhibitors of human cytochrome P450- 3A isoforms.loJ1 Coadministration of 3A substrate drugs with azole derivatives such as ketoconazole, itraconazole, and fluconazole can result in impairment of clearance of such drugs which, in some cases, can be quantitatively large and clinically important. This is particularly true for high-extraction compounds given orally, since the azoles may inhibit both the gastrointestinal and hepatic compo- nents of presystemic extraction. l&l3 In the case of orally administered triazolam and midazolam (both relatively “pure” substrates for P450-3At4-16), their oral clearance may be impaired by 85% or more when given together with ketoconazolel6-‘9; significant, although somewhat smaller, interactions are caused by itraconazole and flu- conazole~l7,182@22

Since zolpidem is metabolized partially, but not exclu- sively, by P450-3A isoforms, the possibility exists that azole antifungal agents have a lesser inhibiting effect on zolpidem clearance in vivo, resulting in a smaller clini- cal interaction, compared with their effect on triazolam or midazolam clearance. In studies of human liver micro- somes in vitro, ketoconazole inhibition constants (Ki) in the low nanomolar range were observed for inhibition of midazolam or triazolam hydroxylation.1sv16 In this study we compared the susceptibility of zolpidem hydroxyla- tion and triazolam hydroxylation to inhibition by keto- conazole in vitro. Possible interactions of zolpidem with ketoconazole, as well as with itraconazole and flucona- zole, were tested in a controlled clinical pharmacoki- netic-pharmacodynamic study.

METHODS In vitro studies. Procedures for acquisition and stor-

age of human liver tissue samples and for preparation of microsomal fractions have been described previ- ously. 15,16~2s~24 Zolpidem and its principal hydroxylated metabolite (termed the M-3 metab~life)~ were provided by Synthelabo Recherche, Bagneux, France. Triazolam and its principal hydroxylated metabolites (a-hydroxy- triazolam and 4-hydroxytriazolam)r6,25 were kindly provided by Pharmacia and Upjohn Co, Kalamazoo, Mich. Other chemical reagents and drug entities were purchased from commercial sources.

Zolpidem and triazolam were prepared in methanol solution. Aliquots were added to incubation tubes to pro- duce a final concentration of 10 pmol/L zolpidem in one series, and 100 pmol/L triazolam in another series. For zolpidem, 10 pmol/L is below the concentration (K,) cor- responding to 50% of the maximum velocity of M-3

metabolite formation; for triazolam, 100 FmoliL is slightly above the K, for a-hydroxytriazolam formation and below the K, for 4-hydroxytriazolam formation. Ketoconazole in methanol solution was co-added to incu- bation tubes to yield final concentrations ranging from 0 to 2.5 pmol/L. The solvent was evaporated to dryness at 40°C under mild vacuum. To each tube was then added incubation buffer, appropriate cofactors, and an NADPH- regenerating system as described previously. 15J6J.W The mixture was heated to 37°C and reactions initiated by addition of microsomal protein (up to 0.5 mg/mL). After 20 minutes at 37°C reactions were stopped by cooling on ice and addition of acetonitrile. Internal standard was added, the mixtures centrifuged, and supernatants trans- ferred to autosampling vials for HPLC analysis. All indi- vidual incubations were done in duplicate. Studies of zolpidem and triazolam each were performed using microsomal preparations from 4 human liver samples.

The HPLC mobile phase consisted of acetonitrile, methanol, and 50 ymollL phosphate buffer. For analy- sis of the M-3 metabolite of zolpidem, mobile phase component proportions were 30: 10:60, the analytical column was Cl, Bondapak (30 cm x 3.9 mm), and the ultraviolet absorbance wavelength was 242 nm. For analysis of a-hydroxytriazolam and 4-hydroxytriazo- lam, mobile phase proportions were 22.5:10:67.5, the column was Cl, NovaPak (15 cm x 3.9 mm), and ultra- violet absorbance was at 220 nm.

Design of clinical study. The protocol was reviewed and approved by the Human Investigation Review Committee serving Tufts University School of Medi- cine and New England Medical Center Hospital. Twelve healthy volunteers (8 men and 4 women), aged 20 to 40 years, participated after giving written informed consent. All were active ambulatory non- smoking adults, with no evidence of medical disease and taking no other medications. Female subjects were not taking oral contraceptives and did not have contra- ceptive implants.

The study had a double-blind, randomized, 5-way crossover design, with at least 7 days elapsing between treatments. Medications were separately and identically packaged in opaque capsules and administered orally. The 5 treatments were as follows:

A. Zolpidem placebo plus azole placebo B. Zolpidem (5 mg) plus azole placebo C. Zolpidem (5 mg) plus ketoconazole, 200 mg

twice daily D. Zolpidem (5 mg) plus itraconazole, 100 mg twice

daily E. Zolpidem (5 mg) plus fluconazole, 100 mg twice

daily

Greenblntt et al. 663

Table I. Summary of analytical methods for analysis of azole antifungal agents in plasma

Ketoconazolr Itmconc~zole/h?,dr~x~itrclconLlznle

Internal standard Extraction solvent

Mobile phase composition

Mobile phase flow rate Analytical column

Ultraviolet absorbance wavelength Lower limit of quantitation Assay variance (Y&V) within-day;

between-day

Dextromethorphan Ethyl acetate/isoamyl alcohol (98.5: 1 S) CH,CN/CH,OH/SO mmol/L NH3H,P0, (405%) I.5 mL/min C,, uBondaPak*(30 cm length, IO urn particle size) 220 nm 0.05-o. 1 ug/mL 110%‘: <lO%

Clotrimazole Phenacetin Ethyl acetate Ethyl acetate

CH,CN/lO mmol/L CH,OH/IO mmol/L KH,PO, (50:50) Na acetate (40:60) I .5 mL/min I .O mL/min C I 8 NovaPak” ( I5 cm length, C,, yBondaPak* (30 cm 5 urn particle size) length, 10 pm particle size) 263 nm 261 nm 0.03-0.05 yg/mL 0.1 uLg/mL <90/o; <IX% <7%: <3%

CV, Coeffcienl of variance. ‘Waters A\\ocinte\, Milfurd. Man

Doses and dosage schedules for the azoles were cho- sen based on recommendations in approved labeling.

Procedures. At 8 AM on study day 1, subjects entered the outpatient Clinical Psychopharmacology Research Unit where they received the initial dose of azole (or placebo) and remained under observation for 30 min- utes. Subjects took a second dose of azole (or placebo) at home at 4 PM on day 1. On the morning of day 2, after ingesting a standardized light breakfast with no caffeine-containing food or beverages and no grapefruit juice, they returned to the Research Unit at approxi- mately 7:30 AM. They fasted until 12 noon, after which they resumed a normal diet (without grapefruit juice or caffeine-containing food or beverages). The third dose of azole (or placebo) was given at 8 AM, and the single dose of zolpidem or placebo was given at 9 AM. A final azole (or placebo) dose was given at 5 PM.

Venous blood samples were drawn from an indwelling cannula into heparinized tubes before zolpi- dem or placebo dosage and at the following postdosage times: i/l, 1. I Z, 2, 2X, 3, 4, 5, 6, 8, and 24 hours. Sam- ples were centrifuged, and the plasma separated and frozen until the time of assay.

The electroencephalogram (EEG) was recorded using a 6-electrode montage, with instrumentation and methodology described previously.t6,‘9.‘6-‘8 At 2 pre- dosage times and during 8 hours after dosage at times corresponding to blood sampling, the EEG was quanti- fied in 4-second epochs for as long as necessary to ensure at least 2 minutes of artifact-free recording. Data were digitized over the power spectrum from 4.0 to 3 1.75 cycles per second (Hz,), then fast-Fourier trans- formed to determine activity over the 4.0 to 3 1.75 Hz spectrum, and in the “beta” (13.0 to 3 1.75 Hz) band.

Subjects’ self-ratings of sedative effects and mood

state were obtained on a series of loo-mm visual ana- log scales, 16,19.%-30 R atings of sedation were also per- formed by trained observers, using the same rating instrument, without knowledge of the treatment condi- tion. Self- and observer-ratings were obtained twice before medication administration and at postdosage times indicated above.

The digit symbol substitution test (DSST) was administered twice before dosing and at times corre- sponding to rating scales. t6,19%3a Subjects were asked to make as many correct symbol-for-digit substitutions as possible within a 2-minute period. Subjects com- pleted equivalent DSST variants. with no individual taking the same test more than once.

Acquisition and recall of information were evaluated using a word-list free recall proceduret6.t”.‘“-“) that was administered at 1 Z hours after zolpidem or placebo administration. Sixteen words, taken from 4 different categories, were read in random order in “shopping- list” fashion. Recall was tested immediately after pre- sentation of each list. Subjects wrote down items imme- diately after lists were presented in random order. List presentation and recall were repeated a total of 6 times at 1% hours after dosage. At 24 hours after dosing, sub- jects were asked to remember as many words as possi- ble from the previous day’s list (delayed or “free” recall). Thereafter, the same lists were read in the same sequence in which they were presented on the previous day, to assess whether residual effects of drug admin- istration on immediate recall were detectable.

AnaZysis of data. In vitro rates of formation of metabolites of zolpidem or triazolam with coaddition of ketoconazole were expressed as a percentage ratio versus the control velocity with no inhibitor present. The relation of ketoconazole concentration to velocity

664 Greenblatt et al. CLINICAL PHARMACOLOGY & THER4PEUTICS

DECEMBER 1998

3 \ ZOLPIDEM M-3 METABOLITE

(ZOLPIDEM - 1OpM)

0.1 0.2 0.5 1 2 3

KETOCONAZOLE (JAM)

Figure 1. Rates of formation of the zolpidem M-3 hydrox- ylated metabolite from zolpidem (10 umol/L) or of a-hydroxy- triazolatn (a-OH-triazolam) from triazolarn (100 umoUL) by human liver microsomes in vitro. Reaction velocities with co- addition of the inhibitor ketoconazole are expressed as a per- centage ratio versus the control velocity with no ketocona- zole present. Each point is the mean + SE ratio from 4 sepa- rate human liver microsomal preparations. Mean 50% inhibitory concentrations (IC,,) for ketoconazole were 0.61 umol/L versus the zolpidem M-3 metabolite and 0.053 umol/L versus a-hyclroxytriazolam. The IC,, value versus 4- a-hydroxytriazolam (data not shown) was 0.046 umol/L.

ratio was analyzed by nonlinear regression to determine the ketoconazole concentration corresponding to a velocity ratio of 50% of the control value (IC5n).19,32,33

Plasma concentrations of zolpidem from the clinical study were determined by HPLC with fluorescence detec- tion.34 The sensitivity limit was 1 to 2 ng/mL, and the vari- ance between replicate samples did not exceed 8%. The slope (beta) of the terminal log-linear phase of each zolpi- dem plasma concentration versus time curve was deter- mined by linear regression analysis. This slope was used to calculate the apparent elimination half-life. Area under the plasma concentration versus time curve from time zero until the last detectable concentration was determined by the linear trapezoidal method. To this area was added the residual area extrapolated to infinity, calculated as the final concentration divided by beta, yielding the total area under the plasma concentration versus time curve (AUC). The peak plasma concentration and the time of peak concen- tration represented the rate of appearance of drug in sys- temic circulation. Apparent oral clearance was calculated as the administered dose divided by the total AUC.

Plasma concentrations of ketoconazole, itraconazole (and its metabolite hydroxyitraconazole), or fluconazole were determined by HPLC (Table I).

For self- and observer-ratings on visual analog scales, the 2 pre-dose baseline ratings were averaged, and post-dosage scores were expressed as the increment or decrement relative to the mean pre-dose value. Scores on the DSST were analyzed similarly. The word- list memory test was analyzed as the mean absolute number of words correctly remembered for delayed recall and as mean number of words remembered after 6 trials for immediate recall.

For each EEG recording session, the relative beta amplitudes (beta divided by total, expressed as percent) were calculated, and values from the left and right frontotemporal leads were averaged. The means of the relative beta amplitudes in the pre-dose recordings were used as baseline, and all post-dosage values were expressed as the increment or decrement over that treat- ment’s mean pre-dose baseline value.

For each pharmacodynamic variable, the area under the 4-hour plot of effect change score versus time was calculated to obtain a single integrated measure of phar- macodynamic action during the period of greatest drug effect. Also evaluated were pharmacodynamic effects at individual time points.

Statistical procedures included linear and nonlinear regression, ANOVA, the Student-Newman-Keuls pro- cedure, and Dunnett’s t test. One of the subjects did not ingest zolpidem as instructed during treatments D and E; accordingly mean values for treatments A, B, and C represent n = 12, and for treatments D and E, n = 11. For analysis of variance procedures, n = 11 was used throughout.

RESULTS In vitro results. The mean f SE (n = 4) ICsn value

for ketoconazole versus formation of the M-3 metabo- lite of zolpidem was 0.61 + 0.27 l.tmol/L (Figure 1). In contrast, the ketoconazole IC,, was 0.053 f 0.009 pmol/L versus a-hydroxytriazolam formation, and 0.046 -r- 0.007 l.tmol/L versus 4-hydroxytriazolam for- mation (Figure 1).

Clinical study: Plasma concentrations of azobs. Plasma levels of the 3 azole derivatives were consistent with those anticipated during administration of usual therapeutic doses35 (Figure 2). As reported previ- ously,s6,37 plasma concentrations of hydroxyitracona- zole exceeded those of itraconazole. Since values of elimination half-life for itraconazole and fluconazole generally exceed 20 hours,35 these 2 agents may not have actually reached steady state.

FLUCONAZOLE

Gwenblutt et al. 665

KETOCONAZOLE ‘i--d

OH-ITRACONAZOLE

ITRACONAZOLE

0.02 / I I L / I ( I I -1 o 1 2 3 4 5 6 7 8 24

HOURS AFTER ZOLPIDEM DOSE Figure 2. Mean + SE plasma concentration of ketoconazole, itraconazole, hydroxyitraconazole, or fluconazole on study day 2. The third azole dose was given 1 hour before zolpidem.

5 i

I - ZOLPIDEM + PLACEBO(B)

- - n ZOLPIDEM + KETOCONAZOLE (C)

LPIDEM + ITRACONAZOLE (D) ti - - - - -A ZOLPIDEM + FLUCONAZOLE (E)

I / I I ,

0 1 2 3 4 5 6 7 8

HOURS AFTER DOSE

Figure 3. Mean plasma zolpidem concentrations following 5.0 mg zolpidem alone (treatment B) and with coadministration of ketoconazole, itraconazole, or fluconazole (treatments C, D, and E). See Table II for kinetic and statistical analysis.

Clinical pharmacokinetics of zolpidem. Coadminis- B; Figure 3 and Table II). ‘Zolpidem AUC during treat- tration of zolpidem with ketoconazole (treatment C) sig- ment C was increased by a factor of 1.83 + 0.24 (mean nificantly prolonged zolpidem elimination half-life, + SE) compared to treatment B values, and clearance dur- increased total AUC, and decreased apparent oral clear- ing treatment C was reduced to 64% * 7% of treatment ante when compared to zolpidem plus placebo (treatment B values. However, the elimination half-life of zolpidem

666 Greenblatt et al. CLINICAL PHARMACOLOGY & THERAPEUTICS

DECEMREK 199X

‘- - 0 PLACEBO + PLACEBO (A) - ZOLPIDEM + PLACEBO (6)

- - . ZOLPIDEM + KETOCONAZOLE (C)

- - -0 ZOLPIDEM + ITFWONAZOLE (D)

b - - - - -A ZOLPIDEM + FLUCONAZOLE (E)

mmmmxx I * x I I I I I I I I J -20 0 1 2 3 4 5 6 7 6 0 1 2 3 4 5 6 7 6

HOURS AFTER ZOLPIDEM DOSE

Figure 4. Left panel, Changes over predose baseline in percentage of electroencephalographic (EEG) amplitude falling in the beta frequency range; values are the mean of left and right frontal leads. Each point is the mean for all subjects at the corresponding time. Right panel, Changes over predose baseline in scores on the digit-symbol substitution test (DSST) over a 2-minute period. Each point is the mean for all subjects at the corresponding time. Asterisks (*) along the x-axis indi- cate time points at which the 5 conditions differed significantly (P < .05) based on ANOVA. See Table III for individual tests of each treatment condition versus treatment A.

Table II. Effects of azole coadministration on zolpidem pharmacokinetics

Kinetic Treatment B: Treatment C: Treatment D: Treatment E: variables for zolpidem + zolpidem + zolpidem + zolpidem + zolpidem placebo ketoconazole itraconazole ~uconazole

Value of F (ANOVA)

C,,, 0-d~) 72+ 10 97 + 12” 91 It 11 84+ 10 2.57 (P < .08) tmax (h after dose) 1.5 f 0.2 1.7 f 0.3 1.8 + 0.3 1.6 f 0.2 1.16 (NS) Elimination tK (h) 1.86 + 0.13 2.41 f 0.15* 1.95 & 0.11 1.98 ir 0.11 8.06 (P < .005) Total AUC (ng/mL h) 254 f 43 424 + 70* 335 + 61 333 k 60 4.16 (P < .02) Apparent CL,, (mL/min) 422 + 55 250 f 35” 320 f 45 338 -r- 54 4.71 (P < .Ol)

Data are mean values + SE. C max, Peak plasma concentration;&,,,. time to reach C,,; q t half-life; AUC, area under the plasma concentration versus time curve; CL,,,,, oral clearance *Indicates significant individual difference (P < .05) compared with treatment B value based on the Dunnett 2-tailed r test.

during treatment C still averaged only 2.4 hours, and zolpidem was essentially undetectable in the 24-hour samples for treatment C, as well as for the other treat- ments. Coadministration of zolpidem with itraconazole (treatment D) and fluconazole (treatment E) caused less impairment of zolpidem clearance than did ketoconazole, and clearance values for treatments D and E individually were not significantly different from the treatment B value based on Dunnett’s 2-tailed t test (Figure 3 and Table IT).

Pharmacodynamic results. ANOVA indicated no significant difference among the 5 predose treatment means for any pharmacodynamic variables.

Figure 4 shows mean changes in frontal EEG beta amplitude and in DSST score at corresponding times for the 5 treatment conditions.

Based on 4-hour pharmacodynamic effect areas as well as individual time points during the first 4 hours after dosage, zolpidem plus placebo (treatment B) pro-

Geenblatt et cd. 667

(PLL., (Z+Bpl.) (z+EET) (Z+PTRA) (z+k”c) (P&L, (Z+BpL) (Z+kT) R+PTRA) (Z+A”C, Figure 5. Mean + SE 4-hour effect areas for changes over baseline in EEG beta activity in frontal leads (left panel) or DSST scores (right panel). For EEG effect areas (left panel), the Student- Newman-Keuls procedure showed that treatment A and treatment C were significantly different (P < .OS) from each other and from all other treatments; treatments B. D, and E were not signifi- cantly different from each other. For DSST effect areas (right panel), the Student-Newman-Keuls procedure showed that treatments A and C were significantly different from each other. There was no significant difference among A, B, D, and E, nor among B, C, D. and E. PL, Placebo: Z, zolpi- dem; KET, ketoconazole: ITRA, itraconazole; FLUC. tluconazole.

duced increases in frontal EEG beta amplitude and impairments of DSST performance compared to dou- ble placebo (treatment A; Figures 4 and 5; Table III). Coadministration of zolpidem with ketoconazole (treat- ment C) caused significant enhancement of EEG and DSST effects of zolpidem. Coadministration of zolpi- dem with itraconazole (treatment D) or fluconazole (treatment E) did not enhance EEG effects of zolpidem, and slightly (but not significantly) enhanced zolpidem effects on DSST performance. By 8 hours after dosage, treatments A, B, D, and E were not significantly differ- ent on either EEG or DSST measures (Figure 4 and Table III). Treatment C (zolpidem plus ketoconazole) still dif- fered from the other 4 treatments at X hours after dosage; the difference between treatments A and C was signifi- cant for the EEG measure but not significant for the DSST.

The test of information acquisition and immediate recall showed no significant difference among the 5 treatments in immediate recall of words learned at 1% hours after dosage (Figure 6). Zolpidem plus placebo (treatment B) impaired delayed recall of information at 24 hours after dosage compared to treatment A. Co- administration of zolpidem with any of the 3 azoles caused a small and nonsignificant enhancement of

impaired delayed recall associated with zolpidem. There was no significant difference among the 5 treat- ments in relearning of the same word list at 24 hours after dosage.

Differences among the 5 treatments in subjective self-ratings and in observer-rated sedation were quali- tatively similar to those observed with the EEG and DSST measures. However, differences among treat- ments did not reach significance due to high individual variability.

DISCUSSION Human cytochrome P450-3A isoforms are present in

both liver and in gastrointestinal tract mucosa.t’.” The activity of P450-3A is inhibited both in vitro and in vivo by ketoconazole, itraconazole, and fluconazole, with ketoconazole being the most potentt0.t 1.ts.25 When rel- atively “pure” substrates for P450-3A (such as triazo- lam or midazolam) that undergo extensive presystemic extraction are given orally, coadministration of keto- conazole can reduce apparent oral clearance to less than 15% of control, yielding increases of total AUC that may exceed tenfold.te-19 Such interactions are of clear clinical importance and potentially hazardous. Zolpi- dem, in contrast, is metabolized by P450-3A as well as

668 Greenbhtt et al. CLINICAL P HARMACOLOGY & THERAPEVT? I(:S

DECEMAER 1998

16 -

fi 14-

z! 5 12-

ii i ii2 10 -

t5 8- 3

t"o 6- a

E 4-

9 2 -

IMMEDIATE RECALL (1.5 hours after dosage)

f-Eixq T

T T

(z+FL”O)

DELAYED RECALL (FREE RECALL) (24 hours after dosage)

16 r

12 -

10 -

a -

6 -

4-

2-

(r&L) &I.)

Figure 6. Left panel, Mean f SE number of words recalled immediately after 6 learning trials at 1% hours after zolpidem (or placebo) dosage. Differences among treatments were not significant. Right panel, Results of delayed recall (free recall) at 24 hours after zolpidem (or placebo) dosage. The Student-Newman-Keuls procedure indicated that treatment A differed significantly from all others; treatments B, C, D and E were not different from each other.

Table III. Statistical analysis of change scores at individual time points for frontal EEG beta amplitude and DSST score

Hours after zolpidem or placebo dosage

% 1 1% 2 2% 3 4 5 6 8

Frontal EEG beta amplitude: Individual d@erences versus treatment A Treatment B (zolpidem + placebo) * * * * * - - Treatment C (zolpidem + ketoconazole) * * * * * * * * * * Treatment D (zolpidem + itraconazole) * * * * * * * - - - Treatment E (zolpidem + fluconazole) * * * * * - * - - -

DSST score: Individual d@erences versus treatment A Treatment B (zolpidem + placebo) - - * - - - - - - - Treatment C (zolpidem + ketoconazole) * * * * * * * - * - Treatment D (zolpidem + itraconazole) * * - * Treatment E (zolpidem + fluconazole) - * * * - - - - - -

EEG, Electroencephalogram; DSST, digit symbol substitution test. *Indicates significant difference (P 1.05) for individual comparison versus treatment A (placebo + placebo) based on the Dunnett l-tailed test. Hyphen (-) indicates

no significant difference.

a number of other human cytochromes,*j9 suggesting tion of triazolam metabolism.16325 In the present in vitro that ketoconazole, relatively specific for 3A isoforms, study of ketoconazole inhibition of zolpidem and tria- would produce less impairment of zolpidem clearance zolam hydoxylation by human liver microsomes, the compared to triazolam. In vitro studies with human IC,, values for ketoconazole inhibition of a-hydroxy- liver microsomes have in fact demonstrated low triazolam and 4-hydroxytriazolam formation were more nanomolar Ki or IC,, values for ketoconazole inhibi- than an order of magnitude lower than for inhibition of

Geoablatt et al. 669

3

TRIAZOLAM, 0.125 mg p.o. ZOLPIDEM, 5 mg p.o.

- z -I I KETOCONAZOLE

I 1 / I 0 12 3 4 5 6 7 8

5 E 80 B E I # 60

ii

"0 N 40

%

HOURS AFTER DOSE HOURS AFTER DOSE

Figure 7. Left panel, Mean r SE (n = 9) plasma triazolam concentrations in a previous study in which 0.125 mg oral triazolam was given in the control condition and with coadministration of ketoconazole. See von Moltke et alI6 for further explanation. Right panel, Mean + SE (n = 12) plasma zolpidem concentrations in the present study after 5 mg oral zolpidem given in the control condition (treatment B) and with coadministration of ketoconazole (treatment C).

formation of the M-3 metabolite of zolpidem. IC,, val- ues for itraconazole and fluconazole versus zolpidem metabolism in vitro were in turn an order of magnitude higher than that of ketoconazole.

The outcome of the clinical pharmacokinetic study was consistent with in vitro data. Coadministration of usual therapeutic doses of ketoconazole, producing typ- ical steady-state plasma ketoconazole concentra- tions,r6%ty,35 together with zolpidem significantly pro- longed zolpidem elimination half-life, increased total AUC, and decreased apparent oral clearance. On aver- age, ketoconazole reduced zolpidem clearance to 64% of control values; this is a much smaller effect than the reduction of triazolam clearance to 15% or less of con- trol values produced by coadministration of ketocona- zole under similar experimental conditions (Figure 7).16,t7,ty Compared to ketoconazole, itraconazole and fluconazole produced smaller reductions in zolpidem clearance. In a previous study,j7 coadministration of itraconazole produced a quantitatively small decrement in zolpidem clearance that was similar in magnitude to what we observed. The findings emphasize the poten- tial value of in vitro models in providing a prospective estimate of whether specific pharmacokinetic drug interactions in vivo are probable, possible, or unlikely,

thereby allowing a more informed plan for clinical drug interaction studies. iu.i I

Several pharmacodynamic measures were used to evaluate the time course and intensity of benzodiazepine- like agonist effects during each of the 5 treatments. These measures included the fraction of EEG amplitude falling in the “beta” frequency range, performance on the DSST, and immediate and delayed recall of information. The pharmacodynamic results were highly consistent with the pharmacokinetics. Zolpidem plus placebo (treatment B) produced benzodiazepine-like agonist effects of short duration, whereas minimal changes were produced by double placebo (treatment A). The pharmacodynamic effects of zolpidem were enhanced by coadministration of ketoconazole (treatment C), consistent with the increase in zolpidem plasma concentrations. Zolpidem given with itraconazole or fluconazole, in contrast to ketoconazole, produced minimal enhancement of zolpi- dem’s benzodiazepine agonist effects. Our study design did not include trials in which azoles alone were admin- istered. In previous studies we have demonstrated that ketoconazole is a neutral ligand at the CABA-benzodi- azepine receptor38 and produces no evident benzodi- azepine agonist (or inverse agonist) effects when admin- istered alone.‘h.tg

670 Greenblatt et al.

Coadministration of triazolam with ketoconazole pro- duces a very large decrement in triazolam clearance, increase in plasma triazolam concentrations, and enhancement of pharmacodynamic action.t6,17,19 Large and clinically important interactions of triazolam with itraconazole and fluconazole are also reported. l7,*1 In contrast, ketoconazole produced a statistically significant but only quantitatively modest reduction in zolpidem clearance, with modest enhancement of pharmacody- namic action. Zolpidem coadministered with itraconazole or fluconazole produced only small and nonsignificant changes in zolpidem kinetics and dynamics. The findings may have implications for choice of hypnotic drugs in patients receiving azole antifungal agents or other 3A inhibitors, such as the viral protease inhibitors including ritonavir,23,39 or the reverse transcriptase inhibitor delavir- dine,40341 that may be administered to patients with human imrnunodeticiency virus infections.

References

1.

2.

3.

4.

5.

6.

7.

8.

9.

Roth T, Puech AJ, Paiva T. Zolpidem: place in therapy. In: Freeman H, Puech AJ, Roth T, editors. Zolpidem: an update of its pharmacological properties and thereapeu- tic place in the management of insomnia. Paris: Elsevier; 1996. p. 215-30. Langtry HD, Benfield P. Zolpidem: a review of its phar- macodynamic and pharmacokinetic properties and thera- peutic potential. Drugs 1990;40:291-313. Hoehns JD, Perry PJ. Zolpidem: a nonbenzodiazepine hypnotic for treatment of insomnia. Clin Pharm 1993;12:814-28. Nowell PD, Mazumdar S, Buysse DJ, Dew MA, Reynolds CF, Kupfer DJ. Benzodiazepines and zolpidem for chronic insomnia: a meta-analysis of treatment efficacy. JAMA 1997;278:2170-6. Sanger DJ, Benavides J, Perrault G, Morel E, Cohen C, Joly D, et al. Recent developments in the behavioral phar- macology of benzodiazepine receptors: evidence for the functional significance of receptor subtypes. Neurosci Biobehav Rev 1994;18:355-72. Durand A, Thtnot JP, Bianchetti G, Morselli PL. Compar- ative pharmacokinetic profile of two imidazopyridine drugs: zolpidem and alpidem. Drug Metab Rev 1992;24:239-166. Salva P, Costa J. Clinical pharmacokinetics and pharma- codynamics of zolpidem: therapeutic implications. Clin Pharmacokinet 1995;29:142-53. Pichard L, Gillet G, Bonfils C, Domergue J, ThCnot JP, Maurel P. Oxidative metabolism of zolpidem by human liver Cytochrome P45Os. Drug Metab Dispos 1995;23: 1253-62. Greenblatt DJ, von Moltke LL, Harmatz JS, Counihan M, Graf JA, Mertzanis P, et al. Interaction of zolpidem with ketoconazole, itraconazole, and fluconazole: kinetic and dynamic consequences [abstract]. J Clin Pharmacol 1998;38:865.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

CLINICAL PHARMACOLOGY & THERAPEUTICS DECEMBER 199X

Bert RJ, Granneman GR. Use of in vitro and in vivo data to estimate the likelihood of metabolic pharmacokinetic interactions. Clin Pharmacokinet 1997;32:210-58. von Moltke LL, Greenblatt DJ, Schmider J, Wright CE, Har- matz JS, Shader RI. In vitro approaches to predicting drug interactions in vivo. Biochem Pharmacol 1998;55:113-22. Thummel KE, Kunze KL, Shen DD. Enzyme-catalyzed processes of first-pass hepatic and intestinal drug extrac- tion. Adv Drug Deliv Rev 1997;27:99-127. Wilkinson GR. Cytochrome P4503A (CYP3A) metabo- lism: prediction of in vivo activity in humans. J Pharma- cokinet Biopharm 1996;24:475-90. Kronbach T, Mathys D, Umeno M, Gonzalez FJ, Meyer UA. Oxidation of midazolam and triazolam by human liver cytochrome P45OIIIA4. Mol Pharmacol 1989;36:89-96. von Moltke LL, Greenblatt DJ, Schmider J, Duan SX, Wright CE, Harmatz JS, et al. Midazolam hydroxylation by human liver microsomes in inhibition by fluoxetine, norfluoxetine, and by azole antifungal agents. J Clin Phar- macol 1996;36:783-91. von Moltke LL, Greenblatt DJ, Harmatz JS, Duan SX, Harrel LM, Cotreau-Bibbo MM, et al. Triazolam bio- transformation by human liver microsomes in vitro: effects of metabolic inhibitors, and clinical confirmation of a predicted interaction with ketoconazole. J Pharma- co1 Exp Ther 1996;276:370-9. Varhe A, Olkkola KT, Neuvonen PJ. Oral triazolam is potentially hazardous to patients receiving systemic antimycotics ketoconazole or itraconazole. Clin Pharma- co1 Ther 1994;56:601-7. Olkkola KT, Backman JT, Neuvinen PJ. Midazolam should be avoided in patients receiving the systemic antimycotics ketoconazole or itraconazole. Clin Pharma- co1 Ther 1994;55:481-5. Greenblatt DJ, Wright CE, von Moltke LL, Harmatz JS, Ehrenberg BL, Harrel LM. Ketoconazole inhibition of tri- azolam and alprazolam clearance: differential kinetic and dynamic consequences. Clin Pharmacol Ther 1998;64:237-47. Olkkola K, Ahonen J, Neuvonen P. The effect of the sys- temic antimycotics, itraconazole and fluconazole, on the pharmacokinetics and pharmacodynamics of intravenous and oral midazolam. Anesth Analg 1996;82:51 l-6. Varhe A, Olkkola KT, Neuvonen PJ. Fluconazole, but not terbinafine, enhances the effects of triazolam by inhibit- ing its metabolism. Br J Clin Pharmacol 1996;41:319-23. Ahonen J, Olkkola K, Neuvonen l? Effect of itraconazole and terbinafine on the pharmacokinetics and pharmaco- dynamics of midazolam in healthy volunteers. Br J Clin Pharmacol 1995;40:270-2. von Moltke LL, Greenblatt DJ, Grassi JM, Granda SW, Duan SX, Fogelman SM, et al. Protease inhibitors as inhibitors of human cytochromes P450: high risk associ- ated with ritonavir. J Clin Pharmacol 1998;38:106-11. von Moltke LL, Greenblatt DJ, Cotreau-Bibbo MM, Duan SX, Harmatz JS, Shader RI. Inhibition of desipramine

hydroxylation in vitro by serotonin-reuptake-inhibitor antidepressants, and by quinidine and ketoconazole: a model system to predict drug interactions in vivo. J Phar- macol Exp Ther 1994;268: 1278-83.

25. von Moltke LL, Greenblatt DJ, Duan SX, Harmatz JS. Shader RI. Inhibition of triazolam hydroxylation by keto- conazole, itraconazole, hydroxyitraconazole and flucona- zole in vitro. Pharm Pharmacol Commun 1998;4:443-5.

26. Greenblatt DJ, Harmatz JS, Gouthro TA, Locke J, Shader RI. Distinguishing a benzodiazepine agonist (triazolam) from a non-agonist anxiolytic (buspirone) by electroen- cephalography: kinetic-dynamic studies. Clin Pharmacol Ther 1994;56: 100. I 1.

27. Kaplan GB. Greenblatt DJ, Ehrenberg BL, Goddard JE, Cotreau-Bibbo MM, Harmatz JS, et al. Dose-dependent pharmacokinetics and psychomotor effects of caffeine in humans. J Clin Pharmacol 1997;37:693-703.

28. Greenblatt DJ, von Moltke LL, Harmatz JS, Counihan M, Graf J, Durol ALB, et al. Inhibition of triazolam clear- ance by macrolide antimicrobial agents: in vitro corre- lates and dynamic consequences. Clin Pharmacol Ther 199X:64:278-85.

29. Greenblatt DJ, Harmatz JS, Shapiro L. Engelhardt N, Gouthro TA, Shader RI. Sensitivity to triazolam in the elderly. N Engl J Med 199 1;324: 169 l-8.

30. Scavone M, Greenblatt DJ, Harmatz JS, Engelhardt N, Shader RI. Kinetics and dynamics of diphenhydramine 25 mg in young and elderly volunteers. J Clin Pharmacol 1998;38:603-9.

3 I. Shader RI, Dreyfuss D, Gerrein JR, Harmatz JS. Allison SJ, Greenblatt DJ. Sedative effects and impaired learning and recall following single oral doses of lorazepam. Clin Pharmacol Ther 1986;39:526-0.

32. Venkatakrishnan K, Greenblatt DJ, von Moltke LL, Schmider J, Harmatz JS, Shader RI. Five distinct human cytochromes mediate amitriptyline N-demethylation in vitro: dominance of CYP 2C 19 and 3A4. J CIin Pharma- co1 199X:38:1 12-21.

33. Venkatakrishnan K. von Moltke LL. Duan SX. Fleishaker JC, Shader RI, Greenblatt DJ. Kinetic char- acterization and identification of the enzymes responsi- ble for the hepatic biotransformation of adinazolam and N-desmethyladinazolam. J Pharm Pharmacol 199X: 501265-74.

34. Durol ALB, Greenblatt DJ. Analysis of zolpidem in human plasma by high-performance liquid chromatogra- phy with fluorescence detection: application to single- dose pharmacokinetic studies. J Anal Toxic01 I Y97:2 I : 388-92.

35. Como JA, Dismukes WE. Oral azole drugs as systemic antifungal therapy. N Engl J Med 1994:330:263-72.

36. Barone JA, Koh JG, Bierman RH, Colaizzi JL. Swanson KA. Gaffar MC. et al. Food interaction and steady-state pharmacokinetics of itraconazole capsules in healthy male volunteers. Antimicrob Agents Chemother 1993: 37:778-X4.

37. Luurila H. Kivisto KT, Neuvonen PJ. Effect of itracona- zole on the pharmacokinetics and phsrmacodynamics of zolpidem. Eur J Clin Pharmacol 1998;54:336-40.

3X. Fahey JM, Pritchard GA. von Moltke LL, Pratt JS, Grassi JM, Shader RI, et al. The effects of ketoconazole on tria- zolam pharmacokinetics, pharmacodynamics and benzo- diazepine receptor binding in mice. J Pharmacol Exp Ther 1998;285:271-6.

39. Merry C, Barry MG, Mulcahy F, Ryan M, Heavey J. Tjia JF, et al. Saquinavir pharmacokinetics alone and in com- bination with ritonavir in HIV-infected patients. AlDS 1997;l l:F29-33.

40. Cheng CL, Smith DE, Carver PL, Cox SR, Watkins PB, Blake DS, et al. Steady-state pharmacokinetics of delavir- dine in HIV-positive patients: effect on erythromycin breath test. Clin Pharmacol Ther 1997;h I :53 l-43.

4 I. Ferry JJ, Herman BD, Care1 BJ, Carlson CF. Batts DH. Pharmacokinetic drug-drug interaction study of delavir- dine and indinavir in healthy volunteers. J Acquir Immune Detic Syndr Hum Retrovirol 199X; 18:252-Y.