Ito K, Chiba K, Horikawa M, Ishigami M, Mizuno N, Aoki J, Gotoh Y, Iwatsubo T, Kanamitsu S, Kato M, Kawahara I, Niinuma K, Nishino A, Sato N, Tsukamoto Y, Ueda K, Itoh T and Sugiyama Y Which Concentration of the Inhibitor Should Be Used to Predict In Vivo Drug Interactions From In Vitro Data? AAPS PharmSci 2002; 4 (4) article 25 ( https://www.aapspharmsci.org/scientificjournals/pharmsci/journal/ps040425.htm ).

Which Concentration of the Inhibitor Should Be Used to Predict In Vivo Drug Interactions From In Vitro Data?

Submitted: June 25, 2001; Accepted: July 10, 2002; Published: October 7, 2002

Kiyomi Ito ¹ , Koji Chiba ² , Masato Horikawa ³ , Michi Ishigami ⁴ , Naomi Mizuno ⁵ , Jun Aoki ⁶ , Yasumasa Gotoh ⁷ , Takafumi Iwatsubo ⁸ , Shin-ichi Kanamitsu ⁹ , Motohiro Kato ¹⁰ , Iichiro Kawahara ¹¹ , Kayoko Niinuma ¹² , Akiko Nishino ¹³ , Norihito Sato ¹⁴ , Yuko Tsukamoto ¹⁵ , Kaoru Ueda ¹⁶ , Tomoo Itoh ¹ and Yuichi Sugiyama ¹⁷

¹ School of Pharmaceutical Sciences, Kitasato University, Tokyo 108-8641

² Clinical Pharmacology, Pharmacia, Tokyo 163-1448

³ Shiraoka Research Station of Biological Science, Nissan Chemical Industries, Ltd, Saitama 349-0294

⁴ New Drug Development Division, Sankyo Co, Ltd, Tokyo 140-8710

⁵ Pharmacokinetics Laboratory, Mitsubishi Pharma Corporation, Chiba 292-0818

⁶ Institute of Biological Science, Mitsui Pharmaceuticals Inc, Chiba 297-0017

⁷ Basic Discovery Research, R&D;, Kissei Pharmaceutical Co, Ltd, Nagano 399-8304

⁸ Drug Development Division, Yamanouchi Pharmaceutical Co, Ltd, Tokyo 174-8511

⁹ Division of Pharmacology, Drug Safety and Metabolism, Otsuka Pharmaceutical Factory, Inc, Tokushima 772-8601

¹⁰ Pre-Clinical Research Dept. I, Chugai Pharmaceutical Co, Ltd, Shizuoka 412- 8513

¹¹ Department of Pharmacokinetics, Bayer Yakuhin Ltd, Kyoto 619-0216

¹² Drug Metabolism & Physicochemical Property Research Laboratory, Daiichi Pha rmaceutical Co, Ltd, Tokyo 134-8630

¹³ Kawanishi Pharma Research Institute, Nippon Boehringer Ingelheim Co, Ltd, Hyogo 666-0193

¹⁴ Developmental Research Laboratories, Shionogi & Co, Ltd, Osaka 561-0825

¹⁵ Preclinical Science, Nippon Roche KK, Kanagawa 247-8530

¹⁶ Pharmacokinetics Research Department, Teikoku Hormone Mfg Co, Ltd, Kanagawa 213-8522

¹⁷ Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo 113-0033

Abstract

When the metabolism of a drug is competitively or noncompetitively inhibited by another drug, the degree of in vivo interaction can be evaluated from the [I] _u /K _i ratio, where [I] _u is the unbound concentration around the enzyme and K _i is the inhibition constant of the inhibitor. In the present study, we evaluated the metabolic inhibition potential of drugs known to be inhibitors or substrates of cytochrome P450 by estimating their [I] _u /K _i ratio using literature data. The maximum concentration of the inhibitor in the circulating blood ([I] _max ), its maximum unbound concentration in the circulating blood ([I] _max,u ), and its maximum unbound concentration at the inlet to the liver ([I] _in,max,u ) were used as [I] _u , and the results were compared with each other. In order to calculate the [I] _u /K _i ratios, the pharmacokinetic parameters of each drug were obtained from the literature, together with their reported K _i values determined in in vitro studies using human liver microsomes. For most of the drugs with a calculated [I] _in,max,u /K _i ratio less than 0.25, which applied to about half of the drugs investigated, no in vivo interactions had been reported or "no interaction" was reported in clinical studies. In contrast, the [I] _max,u /K _i and [I] _max /K _i ratio was calculated to be less than 0.25 for about 90% and 65% of the drugs, respectively, and more than a 1.25-fold increase was reported in the area under the concentration-time curve of the co-administered drug for about 30% of such drugs. These findings indicate that the possibility of underestimation of in vivo interactions (possibility of false-negative prediction) is greater when [I] _max,u or [I] _max values are used compared with using [I] _in,max,u values.

Introduction

As fractions of human tissues such as human liver microsomes and human hepatocytes have become more easily available for in vitro studies, attempts have been made to quantitatively predict in vivo drug metabolism or drug interactions in humans from in vitro data. ^1-3 In the case of a competitive or noncompetitive inhibition of drug metabolism, the degree of in vivo interaction can be evaluated from the [I] _u /K _i ratio, where [I] _u is the unbound concentration around the enzyme and K _i is the in vitro inhibition constant of the inhibitor. In clinical situations, the substrate concentration is usually much lower than Km and the maximum degree of interaction (R = area under the curve [AUC] (+ inhibitor) / AUC (control)) is expressed as R = 1 + [I] _u /K _i , assuming that the substrate is eliminated from the body only by the inhibited pathway. ¹ Although K _i values can be determined by kinetic analyses of in vitro data using human liver microsomes or recombinant enzymes, it is usually impossible to directly measure [I] _u in humans.

In the case of drugs that are transported into the liver by passive diffusion, [I] _u may be assumed to be equal to the unbound concentration in the liver sinusoid at steady-state. In order to avoid a false-negative prediction due to underestimation of [I] _u , we have proposed the use of the maximum unbound concentration at the inlet to the liver, where the blood flow from the hepatic artery and portal vein meet ([I] _in,max,u ), as the maximum value of [I] _u . ^1,4 We have succeeded in semiquantitative prediction of several cases of drug interactions using this [I] _in,max,u value as [I] _u . ^1,5 This method was thus proposed to be useful for predicting the maximal degree of inhibition. ^1,4

In the present study, the metabolic inhibition potential of a number of drugs known to be inhibitors or substrates of cytochrome P450 (CYP) has been evaluated by estimating their [I] _u /K _i ratio using literature data. The maximum concentration in the circulating blood ([I] _max ), the maximum unbound concentration in the circulating blood ([I] _max,u ), and [I] _in,max,u have been used as [I] _u and the results have been compared with each other.

Materials and Methods

Using the values of K _i and pharmacokinetic data reported in the literature, the [I] _u /K _i ratios were calculated for a number of drugs known to be inhibitors or substrates of human CYP enzymes. Inhibitors and substrates of various human CYP isozymes were selected mainly from the reviews or textbooks. ^3,6,7

K _i values evaluated using human liver microsomes (or recombinant human CYP isozymes) were collected together with the inhibition type and the CYP isozyme involved. To avoid the underestimation of in vivo interactions, the smallest K _i value was used in the analysis if more than 2 values of K _i were reported for a drug. If no information of K _i was available, Km (Michaelis constant) was used instead of K _i . Calculation was not conducted if the inhibition type was other than competitive or noncompetitive inhibition.

According to the perfusion model shown in Figure 1 , the maximum concentration of inhibitor at the inlet to the liver ([I] _in,max ) can be calculated as follows ^1-4 :

where ka is the absorption rate constant, Fa is the fraction absorbed from gut to the portal vein, and Qh is the hepatic blood flow rate. Using the unbound fraction in the blood (fu), [I] _in,max,u can be obtained as follows:

The value of [I] _max after a single oral administration of a therapeutic dose was estimated from reported plasma concentrations assuming a linear pharmacokinetics. To avoid underestimation of interactions, the maximum value of [I] _max or fu was used if the parameter was reported with a range. The values of ka, Fa, Qh, and R _B (blood-to-plasma concentration ratio) were assumed to be 0.1 min ^-1 (a first-order rate constant reported for the gastric emptying ⁸ ), 1, 1610 mL/min, and 1, respectively.

In the case of drugs that inhibit more than 2 CYP isozymes, the [I] _u /K _i ratios were calculated for each CYP isozyme. When any metabolites of the drug were also reported to inhibit the enzyme, the [I] _u / K _i ratio was calculated for both the parent drug and the metabolite(s), and the largest value was used in the analysis.

Reports of in vivo interactions in clinical studies involving the investigated drugs were also examined, and the degree of the interaction was expressed as percentage increase in the area under the concentration-time curve (AUC) of a co-administered substrate of the corresponding CYP isozyme.

Results and Discussion

Table 1 shows the calculated values of [I] _in,max,u /K _i , [I] _max,u /K _i , and [I] _max /K _i ratios and the parameters used in the calculation. The calculated values of the absorption term (the second term in Equation 1) were found to be much larger than the [I] _max for most of the drugs. Although this may result from using the maximum possible values of 0.1 min ^-1 and 1 for ka and Fa, respectively, which may vary among the drugs, this finding suggests that the absorption term should not be ignored to avoid the underestimation of drug interactions.

In Table 2 , the calculated [I] _in,max,u /K _i ratios were classified according to the CYP isozyme involved. Figures in this table indicate the number of the inhibitor-CYP isozyme combinations. CYP3A4 and CYP2D6 were found to be inhibited by the largest number of drugs, followed by CYP1A2 and CYP2C9. No clear difference was observed in the distribution of [I] _in,max,u /K _i ratios among the CYP isozymes involved.

Table 3 shows the relationship between the calculated [I] _in,max,u /K _i ratio and the degree of in vivo interaction, indicated as the increase in the AUC of the co-administered substrates. Nine of the investigated drugs were reported not to cause any interactions with the substrates of the corresponding CYP isozymes in vivo. Reports of in vivo interactions were not found for more than half of the drugs investigated, but the [I] _in,max,u /K _i ratios of most of the drugs in this group were calculated to be less than 0.25. As shown in Table 3 , the [I] _in,max,u /K _i ratio was calculated to be less than 0.25 for about half of the drugs investigated and no in vivo drug-drug interactions have been reported or "no interaction" was reported in clinical studies for about 80% of such drugs. In contrast, the [I] _in,max,u /K _i ratio was calculated to be greater than 2.0 for about 16% of the drugs investigated, and about 80% of such drugs were reported to cause in vivo drug-drug interactions in which more than a 2-fold increase was observed in the AUC of the co-administered substrate of the corresponding CYP isozyme. Moreover, the [I] _in,max,u /K _i ratio was calculated to be greater than 2.0 for all of the drugs reported to cause in vivo interactions with more than a 5-fold increase in the AUC. These findings indicate that the degree of in vivo interaction can be predicted based on the [I] _in,max,u /K _i ratio.

The relationship between the [I] _max,u /K _i ratio and the degree of in vivo interaction is summarized in Table 4 . The calculated [I] _max,u /K _i ratios were less than 0.25 for about 90% of the drugs and about 30% of the drugs in this category were reported to cause in vivo interactions with more than a 1.25-fold increase in the AUC. Furthermore, the [I] _max,u /K _i ratio was calculated to be less than 0.25 for half of the drugs reported to cause in vivo interactions with more than a 5-fold increase in the AUC. Therefore, predictions using [I] _max,u instead of [I] _in,max,u may cause substantial underestimation of in vivo interactions.

Finally, Table 5 shows the relationship between the [I] _max /K _i ratio and the degree of in vivo interaction. The calculated [I] _max /K _i ratios were less than 0.25 for about 65% of the drugs, and about 30% of the drugs in this category were reported to cause in vivo interactions with more than a 1.25-fold increase in the AUC. Results in Table 5 indicate that predictions based only on [I] _max can lead to both overestimation and underestimation of in vivo interactions.

Figure 2 shows the comparison of predictions using [I] _in,max,u , [I] _max,u , and [I] _max . The inhibitor-CYP isozyme combinations were classified according to the [I] _u /K _i ratios and the reported degree of in vivo interactions. When the [I] _max,u was used, the number of combinations for which the [I] _u /K _i ratio was calculated to be less than 0.25 or 0.1 was larger than when the [I] _in,max,u was used. Furthermore, as shown in the closed column, the possibility of false-negative predictions was greatly increased when the [I] _max,u was used in the prediction.

In conclusion, the findings in the present study indicate that the [I] _in,max,u /K _i ratios can be used to predict the possibility of drugs causing in vivo drug-drug interactions, in addition to [I] _max,u /K _i or [I] _max /K _i ratios. The possibility of underestimation of in vivo interactions (possibility of false-negative prediction) is greater when the [I] _max,u /K _i or [I] _max /K _i ratios are used in the prediction compared with using the [I] _in,max,u /K _i ratio.

References

1. Ito K, Iwatsubo T, Kanamitsu S, Ueda K, Suzuki H, Sugiyama Y. Prediction of pharmacokinetic alterations caused by drug-drug interactions. Pharmacol Rev. 1998;50:387-411.

2. Lin JH, Lu AYH. Inhibition and induction of cytochrome P450 and the clinical implications. Clin Pharmacokin. 1998;35:361-390.

3. Bertz RJ, Granneman GR. Use of in vitro and in vivo data to estimate the likelihood of metabolic pharmacokinetic interactions. Clin Pharmacokin. 1997;32:210-258.

4. Kanamitsu S, Ito K, Sugiyama Y. Quantitative prediction of in vivo drug-drug interactions from in vitro data based on physiological pharmacokinetics: use of maximum unbound concentration of inhibitor at the inlet to the liver. Pharm Res. 2000;17:336-343.

5. Ito K, Iwatsubo T, Kanamitsu S, Nakajima Y, Sugiyama Y. Quantitative prediction of in vivo drug clearance and drug interactions from in vitro data on metabolism together with binding and transport. Annu Rev Pharmacol Toxicol. 1998;38:461-499.

6. Kobayashi S. Drug interactions. In: Japanese Society of Clinical Pharmacology and Therapeutics, ed. Rinsho Yakurigaku. Tokyo: Igakushoin; 1996:188-199.

7. Parkinson A. Biotransformation of xenobiotics. In: Klaassen CD, ed. Casarett & Doull's Toxicology, The Basic Science of Poisons. 5th ed. New York, NY: McGraw-Hill, Inc; 1996:113-186.

8. Oberle RL, Chen T-S, Lloyd C, et al. The influence of the interdigestive migrating myoelectric complex on the gastric emptying of liquids. Gastroenterology. 1990;99:1275-1282.

9. Tartini R, Kappenberger L, Steinbrunn W, Meyer UA. Dangerous interaction between amiodarone and quinidine. Lancet. 1982;1:1327-1329.

10. Saal AK, Werner JA, Gross BW, et al. Interaction of amiodarone with quinidine and procainamide. Circulation. 1982;66(suppl 2):224.

11. Sato J, Nakata H, Owada E, Kikuta T, Umetsu M, Ito K. Influence of usual intake of dietary caffeine on single-dose kinetics of theophylline in healthy human subjects. Eur J Clin Pharmacol. 1993;44:295-298.

12. Breen KJ, Bury R, Desmond PV, et al. Effects of cimetidine and ranitidine on hepatic drug metabolism. Clin Pharmacol Ther. 1982;31:297-300.

13. Kirch W, Spahn H, Kohler H, Ohnhaus EE, Mutschler E. Interaction of metoprolol, propranolol and atenolol with concurrent administration of cimetidine. Klin Wochenschr. 1982;60:1401-1407.

14. Pourbaix S, Desager JP, Hulhoven R, Smith RB, Harvengt C. Pharmacokinetic consequences of long term coadministration of cimetidine and triazolobenzodiazepines, alprazolam and triazolam, in healthy subjects. Int J Clin Pharmacol Ther Toxicol. 1985;23:447-451.

15. Raoof S, Wollschlager C, Khan FA. Ciprofloxacin increases serum levels of theophylline. Am J Med. 1987;82:115-118.

16. Syvälahti EKG, TaiminenT, Saarijärvi S, et al. Citalopram causes no significant alterations in plasma neuroleptic levels in schizophrenic patients. J Int Med Res. 1997;25:24-32.

17. Lee BL, Medina I, Benowitz NL, Jacob P, Wofsy CB, Mills J. Dapsone, trimethoprim, and sulfamethoxazole plasma levels during treatment of Pneumocystis pneumonia in patients with the acquired immunodeficiency syndrome (AIDS): evidence of drug interactions. Ann Intern Med. 1989;110:606-611.

18. Kondo T, Tanaka O, Otani K, et al. Possible inhibitory effect of diazepam on the metabolism of zotepine, an antipsychotic drug. Psychopharmacology Berl. 1996;127:311-314.

19. Kunzendorf U, Walz G, Brockmoeller J, et al. Effects of diltiazem upon metabolism and immunosuppressive action of cyclosporine in kidney graft recipients. Transplantation. 1991;52:280-284.

20. Kunzendorf U, Walz G, Brockmoeller J, et al. Effects of diltiazem upon metabolism and immunosuppressive action of cyclosporine in kidney graft recipients. Transplantation. 1991;52:280-284.

21. Glue P, Banfield CR, Perhach JL, Mather GG, Racha JK, Levy RH. Pharmacokinetic interactions with felbamate: in vitro-in vivo correlation. Clin Pharmacokinet. 1997;33:214-224.

22. Haefeli WE, Bargetzi MJ, Follath FF, Meyer UA. Potent inhibition of cytochrome P450IID6 (debrisoquin 4-hydroxylase) by flecainide in vitro and in vivo. J Cardiovasc Pharmacol. 1990;15:776-779.

23. Black DJ, Kunze KL, Wienkers LC, et al. Warfarin-fluconazole. II. A metabolically based drug interaction: in vivo studies. Drug Metab Dispos. 1996;24:422-428.

24. Varhe A, Olkkola KT, Neuvonen PJ. Effect of fluconazole dose on the extent of fluconazole-triazolam interaction. Br J Clin Pharmacol. 1996;42:465-470.

25. Bergstrom RF, Peyton AL, Lemberger L. Quantification and mechanism of the fluoxetine and tricyclic antidepressant interaction. Clin Pharmacol Ther. 1992;51:239-248.

26. Transon C, Leemann T, Vogt N, Dayer P. In vivo inhibition profile of cytochrome P450TB (CYP2C9) by (+/-)-fluvastatin. Clin Pharmacol Ther. 1995;58:412-417.

27. Sperber AD. Toxic interaction between fluvoxamine and sustained release theophylline in an 11-year-old boy. Drug Saf. 1991;6:460-462.

28. Ereshefsky L, Riesenman C, Lam YWF. Antidepressant drug interactions and the cytochrome P450 system (The role of cytochrome P450 2D6). Clin Pharmacokinet. 1995;29s:10-19.

29. Fleishaker JC, Hulst LK. A pharmacokinetic and pharmacodynamic evaluation of the combined administration of alprazolam and fluvoxamine. Eur J Clin Pharmacol. 1994;46:35-39.

30. Islam SI, Masuda QN, Bolaji OO, Shaheen FM, Sheikh IA. Possible interaction between cyclosporine and glibenclamide in posttransplant diabetic patients. Ther Drug Monit. 1996;18:624-626.

31. Kakuda TN, Struble KA, Piscitelli SC. Protease inhibitors for the treatment of human immunodeficiency virus infection. Am J Health-Syst Pharm. 1998;55:233-254.

32. Varhe A, Olkkola KT, Neuvonen PJ. Oral triazolam is potentially hazardous to patients receiving systemic antimycotics ketoconazole or itraconazole. Clin Pharmacol Ther. 1994;56:601-607.

33. von Moltke LL, Greenblatt DJ, Harmatz JS, et al. Triazolam biotransformation by human liver microsomes in vitro: effects of metabolic inhibitors and clinical confirmation of a predicted interaction with ketoconazole. J Pharmacol Exp Ther. 1996;276:370-379.

34. Joeres R, Richter E. Mexiletine and caffeine elimination. N Engl J Med. 1987;317:117.

35. Preston KL, Griffiths RR, Cone EJ, Darwin WD, Gorodetzky CW. Diazepam and methadone blood levels following concurrent administration of diazepam, methadone. Drug Alcohol Depend. 1986;18:195-202.

36. Klintmalm G, Sawe J. High dose methylprednisolone increases plasma cyclosporin levels in renal transplant recipients. Lancet. 1984;1(8379):731.

37. O'Reilly RA, Goulart DA, Kunze KL, et al. Mechanisms of the stereoselective interaction between miconazole and racemic warfarin in human subjects. Clin Pharmacol Ther. 1992;51:656-667.

38. Greene DS, Barbhaiya RH. Clinical pharmacokinetics of nefazodone. Clin Pharmacokinet. 1997;33:260-275.

39. Robinson DS, Roberts DL, Smith JM, et al. The safety profile of nefazodone. J Clin Psychiatry. 1996;57(suppl 2):31-38.

40. Cantarovich M, Hiesse C, Lockiec F, Charpentier B, Fries D. Confirmation of the interaction between cyclosporine and the calcium channel blocker nicardipine in renal transplant patients. Clin Nephrol. 1987;28:190-193.

41. Hippius M, Henschel L, Sigusch H, Tepper J, Brendel E, Hoffmann A. Pharmacokinetic interactions of nifedipine and quinidine. Pharmazie. 1995;50:613-616.

42. Gugler R, Jansen JC. Omeprazole inhibits oxidative drug metabolism. Gastroenterology. 1985;89:1235-1241.

43. Özdemir V, Naranjo CA, Herrmann N, Reed K, Sellers EM, Kalow W. Paroxetine potentiates the central nervous system side effects of perphenazine: contribution of cytochrome P4502D6 inhibition in vivo. Clin Pharmacol Ther. 1997;62:334-347.

44. Bonnabry P, Desmeules J, Rudaz S, Leemann T, Veuthey JL, Dayer P. Stereoselective interaction between piroxicam and acenocoumarol. Br J Clin Pharmacol. 1996;41:525-530.

45. Ptachcinski RJ, Venkataramanan R, Burckart GJ, et al. Cyclosporine_high-dose steroid interaction in renal transplant recipients: assessment by HPLC. Transplant Proc. 1987;19:1728-1729.

46. Kowey PR, Kirsten EB, Fu CJ, Mason WD. Interaction between propranolol and propafenone in healthy volunteers. J Clin Pharmacol. 1989;29:512-517.

47. Schellens JHM, Ghabrial H, van der Wart HHF, Bakker EN, Wilkinson GR, Breimer DD. Differential effects of quinidine on the disposition of nifedipine, sparteine, and mephenytoin in humans. Clin Pharmacol Ther. 1991;50:520-528.

48. von Moltke LL, Greenblatt DJ, Duan SX, Daily JP, Harmatz JS, Shader RI. Inhibition of desipramine hydroxylation (cytochrome P450-2D6) in vitro by quinidine and by viral protease inhibitors: relation to drug interactions in vivo. J Pharm Sci. 1998;87:1184-1189.

49. Hsu A, Granneman GR, Cao G, Carothers L, et al. Pharmacokinetic interactions between two human immunodeficiency virus protease inhibitors, ritonavir and saquinavir. Clin Pharmacol Ther. 1998;63:453-464.

50. Nemeroff CB, DeVance CL, Pollock BG. Newer antidepressants and the cytochrome P450 system. Am J Psychiatry. 1996;153:311-320.

51. Veronese ME, Miners JO, Randles D, Gregov D, Birkett DJ. Validation of the tolbutamide metabolic ratio for population screening with use of sulfaphenazole to produce model phenotypic poor metabolizers. Clin Pharmacol Ther. 1990;47:403-411.

52. Toon S, Low LK, Gibaldi M, et al. The warfarin-sulfinpyrazone interaction: stereochemical considerations. Clin Pharmacol Ther. 1986;39:15-24.

53. Lamberg TS, Kivisto KT, Neuvonen PJ. Lack of effect of terfenadine on the pharmacokinetics of the CYP3A4 substrate buspirone. Pharmacol Toxicol. 1999;84:165-169.

54. Yasui N, Tybring G, Otani K, et al. Effects of thioridazine, an inhibitor of CYP2D6, on the steady-state plasma concentrations of the enantiomers of mianserin and its active metabolite, desmethylmianserin, in depressed Japanese patients. Pharmacogenetics. 1997;7:369-374.

55. Appel S, Rufenacht T, Kalafsky G, et al. Lack of interaction between fluvastatin and oral hypoglycemic agents in healthy subjects and in patients with non-insulin-dependent diabetes mellitus. Am J Cardiol. 1995;76:29A-32A.

56. Amchin J, Zarycranski W, Taylor KP, Albano D, Klockowski PM. Effect of venlafaxine on the pharmacokinetics of alprazolam. Psychopharmacol Bull. 1998;34:211-219.