A kinetic and dynamic study of oral alprazolam with and without erythromycin in humans: In vivo evidence for the involvement of CYP3A4 in alprazolam metabolism

Objective: To assess the possible involvement of CYP3A4 in the metabolism of alprazolam in vivo. Method: Twelve healthy male volunteers were randomly allocated to one of the two different treatment sequences, placebo-erythromycin or erythromycin-placebo, with an at least 6-week washout period be- tween the two trial phases. Each volunteer received 400 mg erythromycin or matched placebo given orally three times a day for 10 days and an oral dose (0.8 mg) of alprazolam on the posttreatment day 8. Plasma concentration of alprazolam was measured up to 48 hours after the administration, and psychomotor function was assessed at each time of blood samplings with use of the Digit Symbol Substitution Test, visual analog scale, and Udvalg for kliniske undersogelser side effect rating scale. Resulz~ Erythromycin significantly (p < 0.001) increased the area under the plasma concentration-time curves (200 -C 43 versus 322 + 49 ng + hr/ml fr om 0 to 48 hours and 229 + 52 versus 566 f 161 ng * lx/ml from 0 hour to infinity), decreased the apparent oral clearance (1.02 + 0.31 versus 0.41 2 0.12 ml/n&/kg), and prolonged the elimination half-life (16.0 + 4.5 versus 40.3 + 14.4 hours) of alprazolam. However, any psychomotor function variables did not ditfer significandy between the erythromycin and placebo trial phases. Conclusion: This study suggests that erythromycin, an inhibitor of CYP3A4, inhibits the metabolism of alprazolam, providing an in vivo evidence for the involvement of CYP3A4 in its metabolism. However, the kinetic change of alprazolam by erythromycin does not result in the pharmacodynamic change of this triazolobenzodiazepine, at least after single dosing. (Clin Pharmacol Ther 1996;59:514-9.)

Norio Yasui, MD, Koichi Otani, MD, PhD, Sunao Kaneko, MD, PhD, Tadashi Ohkubo, PhD, Takako Osanai, BS, Kazunobu Sugawara, PhD, Kan Chiba, PhD, and Takashi Ishizaki, MD, PhD Hirosaki, Goshogawara, and Tokyo, Japan

The triazolobenzodiazepine alprazolam is exten- sively used for the treatment of anxiety and panic disorders.‘-3 Previous reports4,’ have suggested that the pharmacokinetics of alprazolam is linear (i.e.,

From the Department of Neuropsychiatry and Department of Pharmacy, Hirosaki University Hospital, Hirosaki; the Depart- ment of Pharmacy, Goshogawara City Hospital, Goshogawara; and the Department of Clinical Pharmacology, Research In- stitute, International Medical Center of Japan, Tokyo.

Supported by a grant-in-aid from Hirosaki University Hospital, Hirosaki, Japan.

Received for publication Aug. 22, 1995; accepted Nov. 21, 1995. Reprint requests: Koichi Otani, MD, PhD, Department of Neu-

ropsychiatry, Hirosaki University Hospital, Hirosaki 036, Japan.

Copyright 0 1996 by Mosby-Year Book, Inc. 0009-9236/96/$5.00 + 0 13/l/70773

dose independent) in humans, demonstrating a lin- ear dose-plasma concentration relationship for al- prazolam at doses up to 10 mgjday. In addition, a significant concentration-response relationship for the drug has been suggested in the treatment of panic disorder; optimal reduction of anxiety occurs in the plasma concentration range of 20 to 40 rig/ml? Mean- while, the central nervous system-depressant side ef- fects increase progressively at higher concentrations.5 Thus information on the kinetic disposition of al- prazolam interacted with other drugs is of clinical importance.

Alprazolam is metabolized primarily by the he- patic microsomal oxidation, yielding 4- and a-hydroxyalprazolam as its principal metabo- lites.3,6,7 The specific cytochrome subfamily respon- sible for the biotransformation of alprazolam has, to

514

CLINICAL PHARMACOLOGY & THERAPEUTICS VOLUMF 5Y, NUMBER 5 Yasui et al. 515

our knowledge, not been identified. However, there has been in vitro’ and in vivo9,io evidence that sug- gests that two triazolobenzodiazepines, triazolam and midazolam, closely related to alprazolam in the structure and metabolic pathways6 are metabolized largely by the cytochrome P4503A4 (CYP3A4). A recent in vitro study with human liver microsomes” has indicated that ketoconazole, a relatively selec- tive inhibitor of CYP3A4,12 inhibits the 4- and a-hydroxylation of alprazolam, suggesting that al- prazolam may also be metabolized by way of CYP3A4. However, there has been no in vivo data that suggest the possible involvement of this isozyme in the metabolism of alprazolam in humans.

We had two study aims. First, we intended to assess the possible involvement of CYP3A4 in the in vivo metabolism of alprazolam in humans. Thus we studied the effect of erythromycin, an inhibitor of CYP3A4,i2,i3 on the single oral dose pharmacoki- netics of alprazolam in an unrelated healthy volun- teer group. Second, we intended to test the hypoth- esis that if erythromycin would alter the kinetics of alprazolam, any pharmacodynamic parameters of this triazolobenzodiazepine might differ between the placebo and erythromycin trial phases, using a double-blind, randomized, two-way crossover study design as described below.

METHODS Subjects. Twelve unrelated, healthy male subjects

(age range, 25 to 41 years; weight range, 52 to 68 kg; height range, 161 to 179 cm) participated in the study. No subject had a history of significant medical illness or hypersensitivity to any drugs. Their healthy normal status was judged on the basis of a physical examination with screening blood chemistries, in- cluding a complete blood count and hepatic function test, urinalysis, and electrocardiogram before the study. The study protocol was approved by the Eth- ics Committee of Hirosaki University Hospital, and each subject gave his written consent before the study.

Protocol. The study was conducted in a double- blind, randomized crossover manner, with at least a 6-week washout period. Erythromycin (400 mg) as the tablet formulation (Ilotysin, Shionogi Pharma- ceutical Co., Osaka, Japan) or matched placebo (as the tablet formulation with the same appearance and size of that of erythromycin) was given orally three times a day (at 7 AM, 3 PM, and 11 PM) for 10 days. Six volunteers each as a group were allocated to either of the different drug sequences: placebo-

erythromycin or erythromycin-placebo. The kinetic and dynamic study of alprazolam was conducted on an inpatient basis after overnight fast on the post- treatment day 8; at 9 AM, a single oral 0.8 mg dose of alprazolam (Solanax, Japan Upjohn Ltd., Tokyo, Japan) as the tablet formulation was given with a cup of tap water. Blood samples (10 ml each) were collected into heparinized tubes from an antecubital vein before and at l/2, 1, 2, 3, 4, 6, 8, 10, 12, 24, 36, and 48 hours after dosing. No food was allowed until 3 hours after drug administration. At the same times as blood samplings, psychomotor function status was evaluated with use of the Digit Symbol Substitution Test (DSST) adapted from the Wechsler Adult In- telligence Scale in 3 minutes, the visual analog scale (VAS) of mood and subjective states that was used in a pharmacodynamic assessment study on chlordi- azepoxide,14 and the item “sleepiness” of the Ud- valg for kliniske undersogelser (UKU) side effect rating scale.15

hsuy. Plasma alprazolam concentrations were measured in duplicate by an HPLC method devel- oped in our laboratory. Alprazolam and estazolam as an internal standard were supplied by the Upjohn Company (Kalamazoo, Mich.). All solvents used were of HPLC grade (Wako Pure Chemical Indus- tries, Osaka, Japan). All reagents were purchased from Waco Pure Chemical Industries or Nakarai Tesque (Kyoto, Japan). We added 10 ng of estazo- lam in 10 l~,l of methanol to 1 ml of plasma sample. The plasma sample was diluted with 5 ml of 1 mol/L sodium chloride, and the solution was briefly mixed. The mixture was applied to a Sep-Pak CN cartridge (Waters Chromatography, Milford, Mass.), which had previously been activated with 5 ml acetonitrile and water. The cartridge was then washed with 10 ml water. The fraction desired was eluted with 5 ml of 20% acetonitrile. The eluate was evaporated to dryness in vacuum at 60” C. The residue was dis- solved in 50 ~1 methanol and 100 l~,l mobile phase, and the sample was injected onto the HPLC system. The HPLC system consisted of a Rheodyne Model 7120 injector (Rheodyne Inc., Cotati, Calif.), a stainless-steel column (150 mm X 4.6 mm internal diameter) packed with Develosil Cs-5 stationary phase (5 km, Nomura Chemical, Seto, Japan), a Jasco model PU-880 chromatography pump (Jasco, Tokyo, Japan), and a Jasco Uvidec 980 ultraviolet detector (Jasco, Tokyo, Japan). The wavelength was set at 230 nm. The mobile phase consisted of 0.5% (pH 4.5) monobasic potassium phosphate and ace- tonitrile (70:30, vol/vol). The flow rate was 1 mlimin

5 16 Yusui et al. CLIKICAL PHhtM.4COLOW’ & THtlUP!iLTICS

hlAY 1996

, . . .

0 12 24 36 48

Time(hours)

Fig. 1. Mean plasma concentration-time data after a sin- gle oral dose of 0.8 mg alprazolam during the treatment with placebo or erythromycin. Open and solid circles indi- cate the mean data observed during the treatment with placebo and erythromycin, respectively, in the 12 subjects. Asterisks indicate statistically significant differences com- pared with the placebo trial phase: *p < 0.05; **p < 0.01; ***p < 0.001.

at ambient temperature. Retention times for alpra- zolam and estazolam were 17.5 and 14.3 minutes, respectively. The lowest limit of detection was 0.5 rig/ml, and the coefficient of variation (both intra- assay and interassay) was less than 7.8%.

Data analysis. The elimination rate constant (K) of alprazolam was estimated from the nonlinear least-squares regression analysis of the terminal log- linear concentration-time data, and the elimination half-life (t,,,) was calculated from 0.693/K. The area under the plasma concentration-time curve from 0 to 48 hours [AUC(O-48)] was calculated by the trap- ezoidal rule. The area under the plasma concentra- tion-time curve from 0 hour to infinity [AUC(O-m)], or total AUC, was calculated from AUC(O-48) + C,$K, in which C,, is the plasma concentration of alprazolam at 48 hours after dosing. The apparent oral clearance (CL,,,,) was calculated as follows: Dose/total AUC. This clearance term is influenced by absolute oral bioavailability, which was not de- termined in this study. The peak plasma concentra- tion (C,,,) and the time to C,,, (tmax) were read from the observed plasma concentration-time data

in each of the individuals and used to estimate the rate of drug absorption.

The data are given as mean values 2 SD. Statis- tical analyses were performed by a paired t test and Wilcoxon singed-rank test, when appropriate. A p value of <0.05 was considered statistically signif- icant.

RESULTS Volunteers reported some sleepiness or sedation

during several hours after the administration of al- prazolam regardless of the treatment. Otherwise, no serious adverse experience was reported throughout the two trial periods. All subjects completed each of the two trials according to the study protocol.

Phurmucokinetic assessment. The mean plasma concentration-time data of alprazolam observed during the two treatment periods are shown in Fig. 1. The mean (&SD) kinetic data are summarized in Table I.

Erythromycin increased significantly (u < 0.05 to 0.001) the mean plasma alprazolam concentrations at 6 to 48 hours after dosing (Fig. 1). Thus erythro- mycin increased significantly (p < 0.001) the mean AUC(O-48) and AUC(O-m), decreased the CL,,,,, and prolonged the t,,, of alprazolam compared with placebo (Table I). In all subjects the AUC(O-48) and AUC(O-m) increased, CLoral decreased, and t,,, be- came longer during the treatment with erythromycin than during treatment with placebo. Erythromycin did not change the mean C,,,, but it delayed signif- icantly (p < 0.05) the mean t,, (Table I).

Pharmucodynumic assessment. The mean psy- chomotor function parameters assessed by DSST, VAS, and the item “sleepiness” of the UKU rating scale are shown in Fig. 2. Erythromycin did not alter any pharmacodynamic parameters assessed in the study compared with placebo, though the antibiotic treatment increased the plasma concentrations of alprazolam at 6 to 48 hours after dosing (Fig. 1) and the systemic drug exposure (i.e., AUC terms, Table I).

DISCUSSION CYP3A4 is the major isozyme in the CYP3A sub-

family, which accounts for up to about 25% to 30% of the total cytochrome P450 in adult human liver.r2,r6 However, the CYP3A4 activity is known to show marked interindividual variability.16 Representative drugs metabolized by this isozyme include triazolam,8~9 midazolam,8~‘0 cyclosporine,12,‘7 terfenadine,12,17,‘8 erythromycin,‘2,17,19 and nifedipine.12,17 Erythromycin

CLINICAL PHARMACOLOGY 8r THERQ’EUTICS VOLUME 59, NUMBER 5

A. DSST

0 12 24 36 46

B. VAS C. UKU Scale

-20 I. 0 12 24 36 46

Time (hours)

Yaszzi etal. 517

0 12 24 36 46

Fig. 2. Mean psychomotor function variables after a single oral dose of 0.8 mg alprazolam during the treatment with placebo or erythromycin in the 12 subjects. Open and solid circles indicate the mean data observed during the treatment with placebo and erythromycin, respectively. Scores on Digit Symbol Substitution Test (DSST), visual analog scale (VAS) (only the item “thinking speed” for brevity), and UKU side effect rating scale (the item “sleepiness”) are shown in A, B, and C, respectively. For DSST and VAS, each point is the mean increase or decrease over the predose baseline score for the all subjects at the corresponding postdose times, whereas for UKU scale it is the mean raw score. Erythromycin- placebo differences were not significant at any time points.

Table I. Pharmacokinetic parameters of alprazolam after a single oral 0.8 mg dose during the treatment with placebo or erythromycin

Parameters Placebo EIythromycin Significance

Cm,, Wml) 12.0 -c 3.2 14.2 2 3.5 NS L,, ON 1.5 + 0.9 3.8 k 3.0 p < 0.05 AUC(O-48) (ng . hriml) 200 -c 43 322 -c 49 p < 0.001 AUC(O-=) (ng * hriml) 229 5 52 566? 161 p < 0.001 CL,,,, (ml/min/kg) 1.02 ? 0.31 0.41 ? 0.12 p < 0.001 Elimination t,,, (hr) 16.0 ? 4.5 40.3 L 14.4 p < 0.001

Data are mean values + SD. C maxr Peak plasma concentration; NS, not significant: t,,,, time to reach C,,,; AUC, area under the plasma concentration-time curve; CL,,,,, apparent oral

clearance; t,,,, half-life.

is demethylated into a nitrosoalkane, which forms a stable and inactive complex with CYP3A4.” There- fore erythromycin is not only a substrate of but also an inhibitor of CYP3A4.12 Indeed, it has been shown that this macrolide antibiotic inhibits the metabolism of substrates toward CYP3A4 (e.g., triazolam,’ midazo- lam,“’ and cyclosporine’3).

We observed that erythromycin increased plasma alprazolam concentrations during the elimination phase (Fig. 1) and the AUC(O-m) and decreased the oral clearance and prolonged the elimination tin in

all individuals, indicating that erythromycin inhibits the metabolism of alprazolam. Therefore the me- tabolism of this triazolobenzodiazepine is mediated, at least in part, by way of CYP3A4, confirming the result obtained from a recent in vitro human liver microsomal study.”

Previous studies20V21 have shown that CYP3A en- zymes are present not only in the liver but also in enterocytes of the gut wall, and these enzymes me- tabolize cyclosporine in the gut wall to a greater extent than in the liver. Therefore the inhibition of

518 Yusui etal.

alprazolam metabolism by erythromycin may occur in the gut wall to a certain extent. We cannot ex- clude this possibility from the present study. How- ever, another triazolobenzodiazepine, midazolam, closely related to alprazolam in structure and met- abolic pathways,6 has been suggested to be metab- olized by CYP3A4 in the liver.8,10 Furthermore, midazolam can be used to assess the intraindividual and interindividual hepatic CYP3A4 activities in pa- tients receiving liver transplantation.22 Therefore we are tempted to assume that the inhibition of alpra- zolam metabolism by erythromycin is mediated mainly through the hepatic CYP3A4 activity.

Erythromycin is known to enhance the gastric emp- tying rate because of its motilin-agonistic propertyt3 suggesting the possibility that it accelerates the absorp- tion of other drugs. Indeed, it has been reported that erythromycin increases the C,, of triazolam’ or mid- az01am’~ and hastens the absorption rate of zopi- clone.24 We observed that the C,, of alprazolam was not changed, but its t,, was rather delayed by eryth- romycin (Table I). In the present study no food was allowed until 3 hours after the dosing of alprazolam during the two trial phases. We cannot offer any ex- planation for the delayed absorption of alprazolam during the treatment with erythromycin. However, the observation that the C,, was not increased by eryth- romycin may be ascribable to the much greater bio- availability of alprazolam (about 92%3,6) than that of triazolam (about 44%h) or of midazolam (about 33%6,10) because any further increase in the absorp- tion of alprazolam by erythromycin, even if it might occur, would not have been detected due to its own greater bioavailability.

We observed that any psychomotor function param- eters assessed did not differ between the two trial periods (Fig. 2), though erythromycin increased plasma alprazolam concentrations significantly during the elimination phase compared with placebo (Fig. 1). This observation is difficult to interpret because the pharmacokinetic interaction effect of erythromycin on alprazolam was not reflected by any significant change in the psychomotor function parameters we assessed in the study. One possible explanation for this observa- tion is that our pharmacodynamic assessment methods may not be sensitive enough to detect any change in the psychomotor function status. We did not use other psychomotor function tests, which have successfully been used to detect the pharmacokinetic-dynamic re- lationship of several benzodiazepines in humans.Z,26 Nevertheless, it has been shown in a cimetidine- diazepam interaction study that, despite an increase in

diazepam concentration by 62% during the treatment with cimetidine, only minimal changes were observed in the clinical pharmacodynamic effects.27 In addition, because benzodiazepines have a wide therapeutic in- dex 28,29 an increase of about 60% in the AUC(O-48) of alpr&olam by erythromycin (Table I) is unlikely to lead to any significant change in the psychomotor func- tion variables assessed, which is apparently compatible with the cimetidine-diazepam interaction study.27

In conclusion, this study provides in vivo evidence for the involvement of CYP3A4 in the metabolism of alprazolam. Nevertheless, the possibility is not excluded that an isozyme(s) other than CYP3A4 is involved in the metabolism. Fluvoxamine, an inhib- itor of CYPlA2,30,31 has been reported to inhibit the 4- and ar-hydroxylation of alprazolam in vitro32 and to increase plasma alprazolam concentrations in vivo in humans.33 In addition, erythromycin inhibits the metabolism of theophylline,‘3,34 which is a sub- strate of CYPlA2.31,35 These findings suggest that the possibility that the inhibitory effect of erythro- mycin on the metabolism of alprazolam is mediated partly through CYPlA2 is not totally negated. Our single-dose study of alprazolam should not give any guarantee that a pharmacodynamic erythromycin- alprazolam interaction does not occur during the long-term therapy of this triazolobenzodiazepine co- administered with erythromycin or other macrolide antibiotics.

References 1.

2.

3.

4.

5.

6.

7.

Fawcett JA, Kravitz HM. Alprazolam: pharmacoki- netics, clinical efficacy, and mechanism of action. Pharmacotherapy 1982;2:243-54. Dawson GW, Jue SG, Brogden RN. Alprazolam: a re- view of its pharrnacodynamic properties and efficacy in treatment of anxiety and depression. Drugs 1984;27:132- 47. Greenblatt DJ, Wright CE. Clinical pharmacokinetics of alprazolam: therapeutic implications. Clin Pharma- cokinet 1993;24:453-71. Ciraulo DA, Antal EJ, Smith RB, et al. The relationship of alprazolam dose to steady-state plasma concentra- tions. J Clin Psychopharmacol 1990;10:27-32. Greenblatt DJ, Harmatz JS, Shader RI. Plasma alpra- zolam concentrations: relation to efficacy and side effects in the treatment of panic disorder. Arch Gen Psychiatry 1993;50:715-22. Garzone PD, Kroboth PD. Pharmacokinetics of the newer benzodiazepines. Clin Pharmacokinet 1989;16: 337-64. von Moltke LL, Greenblatt DJ, Harmatz JS, Shader RI. Alprazolam metabolism in vitro: studies of hu-

CLINICAL PHARMACOLOGY & THER4I’EL’TICS VOLUME 59, NUMBER 5

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

man, monkey, mouse and rat liver microsomes. Phar- macology 1993;47:268-76. Kronbach T, Mathys D, Umeno M, Gonzalez FJ, Meyer UA. Oxidation of midazolam and triazolam by human liver cytochrome P4503A4. Mol Pharmacol 1989;36:89-96. Phillips JP, Antal EJ, Smith RB. A pharmacokinetic drug interaction between erythromycin and triazolam. J Clin Psychopharmacol 1986;6:297-9. Olkkola KT, Aranko K Luurila H, et al. A potentially hazardous interaction between erythromycin and mida- zolam. Clin Pharmacol Ther 1993;53:298-305. von Moltke LL, Greenblatt DJ, Cotreau-Bibbo MM, Harmatz JS, Shader RI. Inhibitors of alprazolam me- tabolism in vitro: effect of serotonin-reuptake- inhibitor antidepressants, ketoconazole and quini- dine. Br J Clin Pharmacol 1994;38:23-31. Watkins PB. Noninvasive tests of CYP3A enzymes. Pharmacogenetics 1994;4:171-84. Periti P, Mazzei T, Mini E, Novelli A. Pharmacoki- netic drug interactions of macrolides. Clin Pharmaco- kinet 1992;23:106-31. Greenblatt DJ, Shader RI, Harmatz JS, Franke K, Koch-Weser J. Absorption rate, blood concentrations, and early response to oral chlordiazepoxide. Am J Psychiatry 1977;134:559-62. Lingjaerde 0, Ahlfors UG, Beth P, Dencker SJ, El- gen K. The UKU side effect rating scale: a new com- prehensive rating scale for psychotropic drugs and a cross-sectional study of side effects in neuroleptic- treated patients. Acta Psychiatr Stand 1987;76(suppl 334):11-100. Shimada T, Yamazaki H, Miura M, Inui Y, Guengerich FP. Interindividual variation in human liver cytochrome P-450 enzymes involved in the oxi- dation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians. J Pharmacol Exp Ther 1994;270:414-23. Brockmoller J, Roots I. Assessment of liver metabolic function: clinical implications. Clin Pharmacokinet 1994;27:216-48. Yun CH, Okerholm RA, Guengerich FP. Oxidation of the antihistaminic drug terfenadine in human liver microsomes: role of cytochrome P-450 3A(4) in N-dealkylation and C-hydroxylation. Drug Metab Dis- pos 1993;21:403-9. Hunt CM, Watkins PB, Saenger P, et al. Heterogene- ity of CYP3A isoforms metabolizing erythromycin and cortisol. Clin Pharmacol Ther 1992;51:18-23. Webber IR, Peters WHM, Back DJ. Cyclosporine metabolism by human gastrointestinal mucosal micro- somes. Br J Clin Pharmacol 1992;33:661-4. Watkins PB, Wrighton SA, Schuetz EG, Molowa DT, Guzelian PS. Identification of glucocorticoid- inducible cytochromes P-450 in the intestinal mucosa of rats and man. J Clin Invest 1987;80:1029-36.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Yasui et al. 519

Thummel KE, Shen DD, Pod011 TD, et al. Use of midazolam as a human cytochrome P4503A probe; II: characterization of inter- and intraindividual hepatic CYP3A variability after liver transplantation. J Phar- macol Exp Ther 1994;271:557-66. Janssens J, Peeters TL, Vantrappen G, et al. Im- provement of gastric emptying in diabetic gastropa- resis by erythromycin. N Engl J Med 1990;322:1028- 31. Aranko K, Luurila H, Backman JT, Neuvonen PJ, Olkkola KT. The effect of erythromycin on the phar- macokinetics and pharmacodynamics of zopiclone. Br J Clin Pharmacol 1994;38:363-7. Greenblatt DJ, Harmatz JS, Engelhardt N, Shader RI. Pharmacokinetic determinants of dynamic differ- ences among three benzodiazepine hypnotics: fluraz- epam, temazepam, and triazolam. Arch Gen Psychia- try 1989;46:326-32. Greenblatt DJ, Ehrenberg BL, Gunderman J, et al. Kinetic and dynamic study of intravenous lorazepam: comparison with intravenous diazepam. J Pharmacol Exp Ther 1989;250: 134-40. Greenblatt DJ, Abernethy D, Morse D, Harmatz JS, Shader RI. Clinical importance of the interaction of diazepam and cimetidine. N Engl J Med 1984;310: 1639-43. Mandelli M, Tognoni G, Garattini S. Clinical phar- macokinetics of diazepam. Clin Pharmacokinet 1978; 3:72-91. Shader RI, Greenblatt DJ. Use of benzodiazepines in anxiety disorders. N Engl J Med 1993;328:1398- 405. Brosen K, Skjelbo E, Rasmussen BB, Poulsen HE, Loft S. Fluvoxamine is a potent inhibitor of cytochrome P45OlA2. Biochem Pharmacol 1993;45:1211-4. Rasmussen BB, Maenpaa J, Pelkonen 0, et al. Selective serotonin reuptake inhibitors and theophylline metabo- lism in human liver microsomes: potent inhibition by fluvoxamine. Br J Clin Pharmacol 1995;39:151-9. von Moltke LL, Greenblatt DJ, Court MH, Duan SX, Harmatz JS, Shader RI. Inhibition of alprazolam and desipramine hydroxylation in vitro by paroxetine and fluvoxamine: comparison with other selective seroto- nin reuptake inhibitor antidepressants. J Clin Psycho- pharmacol 1995;15:125-31. Fleishaker JC, Hulst LK. A pharmacokinetic and pharmacodynamic evaluation of the combined admin- istration of alprazolam and fluvoxamine. Eur J Clin Pharmacol 1994;46:35-9. Jonkman JHG. Therapeutic consequnences of drug interaction with theophylline pharmacokinetics. J Al- lergy Clin Immunol 1986;78:736-42. Sarkar MA, Jackson BJ. Theophylline N-demethylation as probes for P4501Al and P45OlA2. Drug Metab Dis- pos 1994;22:827-34.