Effect of fluconazole dose on the extent of fluconazole-triazolam interaction

Transcription

1 Br J Clin Pharmacol 1996; 42: Effect of fluconazole dose on the extent of fluconazole-triazolam interaction ANU VARHE, KLAUS T. OLKKOLA & PERTTI J. NEUVONEN Department of Clinical Pharmacology, University of Helsinki and University Central Hospital, Helsinki, Finland 1 Azole antimycotics interact with the short acting hypnotic triazolam. The effect of fluconazole dose on the extent of fluconazole-triazolam interaction was investigated in a double-blind, randomized cross-over study of four phases. 2 Eight healthy volunteers received either 5 mg, 1 mg or 2 mg (4 mg on day 1) of fluconazole or placebo orally once a day for 4 days. On day 4, they took a.25 mg oral dose of triazolam, after which plasma samples were collected and pharmacodynamic effects measured for h. 3 The mean area under the triazolam concentration-time curve (AUC) was increased 1.6-, 2.1- and 4.4-fold (P<.1) by fluconazole 5 mg, 1 mg and 2 mg, respectively. The increase in the elimination half-life of triazolam (t ) varied from 1.3-fold (fluconazole 5 mg, P<.5) to 2.3-fold (fluconazole 2 mg, P<.1). The peak concentration of triazolam was also increased significantly during all fluconazole phases; more than twofold by fluconazole 2 mg (P<.1). 4 The pharmacodynamic effects of triazolam were increased significantly (P<.5) by fluconazole 1 mg and 2 mg. 5 Even a small 5 mg daily dose of fluconazole can interact with triazolam and the extent of the interaction increases with increasing fluconazole dose. When triazolam is used concomitantly with fluconazole 5 2 mg, the dose of triazolam should be reduced, accordingly. Simultaneous use of triazolam with higher fluconazole doses should be avoided. Keywords triazolam fluconazole interaction dose dependent Introduction Methods The short acting benzodiazepine triazolam is metabolized Subjects and study design during its first-pass and elimination phases by CYP3A4 [1]. Azole antimycotics are inhibitors of The study protocol was approved by the Ethics CYP3A4 and can markedly increase the hypnotic effect Committee of the Department of Clinical Pharmacology, of triazolam. Fluconazole affects the metabolism of University of Helsinki, and Finnish National Agency triazolam less than either ketoconazole or itraconazole for Medicines. Eight healthy volunteers (five females [2, 3]. and three males; age range 2 32 years; weight range The present study aimed to investigate the effect of kg) signed written informed consent to participate different doses of fluconazole on the pharmacokinetics in the investigation. None of the subjects was receiving and pharmacodynamics of triazolam. continuous medication. Volunteers who had received medication within 3 weeks before the commencement of the study were excluded. A randomized, double-blind, cross-over study design of four phases was used. The phases were separated by Correspondence: Dr Anu Varhe, Department of Clinical Pharmacology, University of Helsinki, Paasikivenkatu 4, FIN-25 Helsinki, Finland 1996 Blackwell Science Ltd 465

2 466 A. Varhe et al. an interval of at least 2 weeks. The subjects were given (DSST), measuring the number of digits correctly orally a 4-day-premedication with either fluconazole substituted in 3 min, the Maddox wing test measuring 5 mg, fluconazole 1 mg, fluconazole 4 mg on the the coordination of extraocular muscles and the critical first day and 2 mg on the next 3 days (referred as flicker fusion test (CFF) measuring discrimination of fluconazole 2 mg in the text) (Diflucan, Pfizer, the fusion of flickering red light, as described earlier Amboise, France) or placebo once a day at 13. h. On [3]. The subjects were properly trained to do the tests day 4, a.25 mg tablet of triazolam (Halcion, Upjohn, before the trial. For each pharmacodynamic variable, Kalamazoo, Michigan, USA) was ingested with 15 ml the areas under the response-time curves were deterwater at 14. h i.e., 1 h after the last dose of fluconazole mined by the trapezoidal rule for to 7 h [AUC(,7 h)]. or placebo. The volunteers fasted for 3 h before administration The imum effects were also recorded. of triazolam and had a light standard meal 4 h afterwards. Alcohol, coffee, tea or cola and tobacco were forbidden during the test period. Statistical analysis Blood sampling and determination of plasma triazolam The results are expressed as mean values with s.e. mean and fluconazole and 95% confidence intervals also given (C, AUC values and t ), or as median and range (t ). The On day 4, a forearm vein was cannulated with a plastic pharmacokinetic and pharmacodynamic data during the cannula and blood samples (1 ml) were collected into placebo and fluconazole phases were compared using EDTA-tubes immediately before the administration of the analysis of variance and a posteriori testing with the triazolam and.5, 1, 1.5, 2, 3, 4, 5, 6, 7 and h after it. Tukey s test. If necessary to ensure the validity of the Plasma was separated within 3 min and stored at modelling assumptions, the data were transformed 4 C until assay of triazolam and fluconazole. Plasma before the analysis. The t values were compared with triazolam concentrations were determined by gas chromnon-parametric Friedman s two-way ANOVA and the atography with use of a modified method of de Kroon Wilcoxon signed-rank test. Pearson s product-moment et al. [4] and Gaillard et al. [5]. A Hewlett-Packard correlation coefficient was used in correlation studies. (Hewlett-Packard Company, Little Falls, Del.) fused The statistical program Systat for Windows, version 5. silica capillary column (5% phenyl), with a length of (Systat, Evanston, Illinois) was used for the analysis. 3 m and an internal diameter of.53 mm, was used. The chosen significance level was P<.5. Column temperature was raised up to 285 C. Methoxydiazepam was used as an internal standard and an electron-capture detector for detection. The sensitivity of the method was.1 ng ml 1 and the coefficient of day-to-day variation was 6.1% at.28 ng ml 1 (n=15) Results and 3.6% at 1.88 ng ml 1 (n=16). Fluconazole concentrations were also determined by gas chromatography, Pharmacokinetics using nitrogen-phosphorus-detector [6]. The sensitivity Plasma concentrations of triazolam and fluconazole of the fluconazole method was 1 ng ml 1 and the during various phases are shown in Figure 1. The C coefficient of day-to-day variation was 2.1% at of triazolam was raised in seven of the eight subjects by 5.3 mg ml 1(n=12). fluconazole 5 mg (P<.5) and fluconazole 1 mg Pharmacokinetics (P<.5) and in every subject by fluconazole 2 mg (P<.1), when compared with the placebo phase. Peak concentrations (C ) and concentration peak Fluconazole 2 mg increased the mean C more times (t ) were registered directly from the original than twofold. data. The elimination rate constant (l ) was determined The mean AUC of triazolam was increased 1.6-fold z for each subject from the log-linear part of the individual (P<.1) by fluconazole 5 mg, 2.1-fold (P<.1) by plasma triazolam curve by regression analysis. The fluconazole 1 mg and 4.4-fold (P<.1) by flucona- elimination half-life (t ) was calculated from t = zole 2 mg when compared with the placebo phase. ln2/l. Area under the triazolam plasma concentrationfrom 1.3-fold to 4.3-fold by fluconazole 1 mg and The individual increases in the AUC of triazolam ranged z time curves were calculated by use of the trapezoidal rule [AUC(, h)] and with extrapolation to infinity from 2.5-fold to 16.-fold by fluconazole 2 mg. (AUC) by dividing the h concentration by l. The The t of triazolam was increased in seven of the z AUC(, h) of fluconazole on the day 4 refers to the eight subjects by fluconazole 5 mg, but in every subject time between and h after the ingestion of triazolam, by fluconazole 1 mg and 2 mg. The individual i.e., between 1 and 19 h after the last dose of fluconazole. increase in the t of triazolam caused by fluconazole 1 mg ranged from 1.4-fold to 2.8-fold and by fluconazole Pharmacodynamic measurements 2 mg from 2.-fold to 3.5-fold (Figure 2). The mean t of triazolam was prolonged from a control The pharmacodynamic effects of triazolam were measured value of 3.1±.4 h 1.3-fold (to 4.±.3 h, P<.5), immediately after each blood sample by four tests: 1.8-fold (to 5.5±.7 h, P<.1), and 2.3-fold (to subjective drowsiness using a 1 mm long visual 7.±.4 h, P<.1) by fluconazole 5 mg, 1 mg and analogue scale (VAS), the digit symbol substitution test 2 mg, respectively (Table I) Blackwell Science Ltd British Journal of Clinical Pharmacology 42,

3 Plasma triazolam (µg ml 1 ) Plasma fluconazole (µg ml 1 ) a b Time (h) T he interaction between fluconazole and triazolam 467 Discussion The present results indicate that the extent of the fluconazole-triazolam interaction is clearly dependent on the dose of fluconazole. The 5 mg daily dose of fluconazole caused significant changes in the pharmacokinetics of triazolam, without any effect on the pharmacodynamics. On the other hand, higher doses of 1 and 2 mg resulted in more pronounced pharmacokinetic changes and also increased the effects of triazolam in a dose dependent manner. There were large interindividual differences in the extent of the interaction at each fluconazole dose level, but the correlation between, for example, the increase of the triazolam AUC and the AUC of fluconazole was good. In our earlier study fluconazole 1 mg changed the pharmacokinetics of triazolam to the same extent as fluconazole 1 mg in the present study, i.e. the AUC of triazolam was more than doubled and the t was nearly doubled by fluconazole 1 mg, when compared with the placebo phase [2]. Fluconazole has also been shown to interact with other substrates of CYP3A, e.g. cyclosporine and terfenadine [ 7 9]. The reports about these interactions are controversial, however, and the discrepancies may reflect the dose-dependent nature and large interindividual variation of the fluconazole interactions. In some case reports, fluconazole 1 mg and 2 mg changed the pharmacokinetics of cyclosporine, while in other reports no such change was noticed [7]. In one patient there was a sharp rise in cyclosporine trough concen- Figure 1 a) The mean plasma concentrations (±s.e. mean) of triazolam after an oral dose of.25 mg triazolam following pretreatment with fluconazole 5 mg ($), fluconazole 1 mg (+), fluconazole 2 mg (&) or placebo (#) once daily for tration when the patient was taking fluconazole 3 mg, 4 days in eight healthy volunteers. b) Plasma concentrations but no such increase occurred with fluconazole 1 mg of fluconazole on day 4. The time zero refers to the [7]. administration of triazolam (i.e., 1 h after administration of The antihistamine terfenadine has a clinically signififluconazole). cant interaction with the antimycotics ketoconazole and itraconazole [1, 11]. In the study of Honig et al. [8], patients were given fluconazole 2 mg day 1 simul- The plasma concentrations of fluconazole on the day of administration of triazolam are shown in Figure 1. taneously with terfenadine. Unmetabolized terfenadine The highest plasma concentrations of fluconazole on was not detected in plasma in any subject, and cardiac day 4, after administration of triazolam, were 2.2± depolarisation was not significantly changed. However,.1 mg l 1 for fluconazole 5 mg, 4.4±.3 mg l 1 for high dose fluconazole, 8 mg day 1, increased mean fluconazole 1 mg and 1.6±.4 mg l 1 for fluconazole AUC of terfenadine by 52% [9]. Fluconazole 8 mg 2 mg. The highest fluconazole concentrations, the day 1 was also associated with an increase in Qt. Our c concentrations of fluconazole at the time of adminis- present and previous results indicate that triazolam and tration of triazolam, and the AUC of fluconazole midazolam may be more sensitive than terfenadine to correlated significantly (P<.1) with the changes of interact with fluconazole [12]. AUC values of triazolam caused by fluconazole. In vitro fluconazole is a competitive inhibitor of CYP2C whereas the inhibition of CYP3A mediated metabolism (of midazolam) did not appear to be Pharmacodynamics competitive. Fluconazole is a considerably less potent inhibitor of CYP3A mediated metabolism than either The pharmacodynamic effects of triazolam during the ketoconazole or itraconazole in vitro, with a K greater i different fluconazole phases and during the placebo than 8 mm ( ketoconazole 3 mm, itraconazole 6 mm) [13]. phase are shown in Figure 3. The imum effects and In vivo therapeutic plasma concentrations of fluconazole the AUC(,7 h) of the CFF, subjective drowsiness are about 1-fold higher than those of itraconazole. ( VAS), the Maddox wing test and the DSST during the Furthermore, about 9% of fluconazole in plasma is fluconazole 2 mg phase differed significantly (P<.5) present as free drug, whereas the free fraction of from the placebo phase. Fluconazole 1 mg increased itraconazole is of the order of 1% [14, 15]. With daily significantly (P<.5) the imum effects and the doses of fluconazole 1 mg, the steady-state plasma AUC(,7 h) of DSST and CFF. Fluconazole 5 mg did concentration of fluconazole is about 5 1 mg l 1 not change significantly any of the tests. (15 35 mmol l 1), i.e. less than the K of 8 mm. The i 1996 Blackwell Science Ltd British Journal of Clinical Pharmacology 42,

4 468 A. Varhe et al AUC of triazolam (ng ml 1 h) 4 Half-life of triazolam (h) Fluconazole (mg) Figure 2 The individual AUC and t values of triazolam during the control phase and three fluconazole phases (i.e. D fluconazole 5, 1 and 2 mg) in eight subjects. Table 1 The pharmacokinetic variables of triazolam.25 mg (mean±s.e. mean or median and range) following pretreatment with different daily doses of fluconazole or placebo for 4 days in eight healthy volunteers. 95% confidence intervals for the ratio of the fluconazole and placebo phases are presented Placebo Fluconazole Fluconazole Fluconazole Variable (Control ) 5 mg 1 mg 2 mg C (ng ml 1) (95% CI) 1.5±.2 2.2±.3a,d 2.1±.1a,d 3.5±.4b (%, 27%) (94%, 234%) (133%, 444%) t (h) 1.3 (.5, 3.)* 1.5 (.5, 4.)* 2.5 (1., 4.)* 2. (.5, 3.)* AUC(, h) (ng ml 1 h) 9.2± ±1.6b,d.3±1.4c,d 33.6±4.c (95% CI) (125%, 235%) (138%, 347%) (195%, 798%) AUC (ng ml 1 h) 9.2± ±1.7b,d.9±1.6c,d 4.7±4.8c (95% CI) (13%, 234%) (142%, 357%) (238%, 949%) t (h) 3.1±.4 4.±.3a,d 5.5±.6c,d 7.±.4c (95% CI) (12%, 17%) (14%, 229%) (167%, 323%) * Median (range). asignificantly (P<.5) different from the placebo phase. bsignificantly (P<.1) different from the placebo phase. csignificantly (P<.1) different from the placebo phase. dsignificantly (P<.5) different from the fluconazole 2 mg phase. concentration of fluconazole in the gut wall and liver In conclusion, although fluconazole has a smaller risk can be considerably higher than in plasma, especially than either itraconazole or ketoconazole to interact with during absorption; this increases the probability of a triazolam, daily fluconazole doses of 5 mg or more significant interaction between fluconazole and triazolam. may cause a significant interaction. There are interindi- Liver:plasma partition coefficients for fluconazole vidual differences between subjects in the extent of the in rat are concentration dependent ranging from 3 to interaction, but this increases along with the dose and 2 [16]. At fluconazole concentrations of mmol plasma concentrations of fluconazole. If used simultaneously l 1 the partition coefficient is approximately 2 4. with fluconazole 1 mg or 2 mg day 1, the 1996 Blackwell Science Ltd British Journal of Clinical Pharmacology 42,

5 T he interaction between fluconazole and triazolam DSST (symbols/3 min) VAS: drowsiness (mm) Maddox (diopters) CFF (Hz) Time (h) 21 Figure 3 Results of the digit symbol substitution test (DSST), subjective drowsiness (VAS), the critical flicker fusion test (CFF) and the Maddox wing test after an oral.25 mg dose of triazolam following pretreatment with fluconazole 5 mg ($), 1 mg (+), 2 mg (&) or placebo (#). Mean values of eight subjects. dose of triazolam, for example, should be reduced by phosphorus detection after solid-phase extraction. 5 75%. The concomitant use of triazolam and high J Chromatogr 1993; 622: doses of fluconazole cannot be recommended. 6 Harris SC, Wallace JE, Foulds G, Rinaldi MG. Assay of fluconazole by Megabore Capillary Gas-Liquid Chromatography with Nitrogen-Selective Detection. We thank Mr Jouko Laitila, Mrs Eija Mäkinen-Pulli and Mrs Lisbet Partanen for determining plasma triazolam and flucon- Antimicrobial Agents Chemother 1989; 33: azole concentrations. 7 Lopez-Gill JA. Fluconazole-cyclosporine interaction: a dose dependent effect? Ann Pharmacother 1993; 27: References 8 Honig PK, Wortham DC, Zamani K, Mullin JL, Conner DP, Cantilena LR. The effect of fluconazole on the steadystate pharmacokinetics and electrocardiographic pharma- 1 Kronbach T, Mathys D, Umeno M, Gonzales FJ, Meyer codynamics of terfenadine in humans. Clin Pharmacol T her UA. Oxidation of midazolam and triazolam by human 1993; 53: liver cytochrome P45IIIA4. Mol Pharmacol 1989; 36: 9 Cantilena LR, Sorrels S, Wiley T, Wortham D. Fluconazole alters terfenadine pharmacokinetics and electrocardio- 2 Varhe A, Olkkola KT, Neuvonen PJ. Fluconazole, but not graphic pharmacodynamics. Clin Pharmacol T her 1995; 57: terbinafine, interacts with triazolam. Br J Clin Pharmacol 5. [Abstract] 1996; 41: Honig PK, Wortham DC, Zamani K, Conner DP, Mullin 3 Varhe A, Olkkola KT, Neuvonen PJ. Oral triazolam is JL, Cantilena LR. Terfenadine-ketoconazole interaction: potentially hazardous to patients receiving systemic antimy- pharmacokinetic and electrocardiographic consequences. cotics ketoconazole or itraconazole. Clin Pharmacol T her JAMA 1993; 269: ; 56: Pohjola-Sintonen S, Viitasalo M, Toivonen L, Neuvonen 4 De Kroon IFI, Langendijk PNJ, De Goede PNFC. PJ. Itraconazole prevents terfenadine metabolism and Simultaneous determination of midazolam and its three increases risk of torsades de pointes ventricular tachycardia. hydroxy metabolites in human plasma by electron-capture Eur J Clin Pharmacol 1993; 45: gas chromatography without derivatization. J Chromatogr 12 Olkkola KT, Ahonen J, Neuvonen PJ. The effect of the 1989; 491: systemic antimycotics, itraconazole and fluconazole, on the 5 Gaillard Y, Gay-Montchamp J-P, Ollagnier M. pharmacokinetics and pharmacodynamics of intravenous Simultaneous screening and quantitation of alpidem, and oral midazolam. Anesth Analg 1996; 82: zolpidem, buspirone and benzodiazepines by dual-channel 13 Hargreaves JA, Jezequel S, Houston JB. Effect of azole gas chromatography using electron-capture and nitrogen- antifungals on human microsomal metabolism of diclofenac 1996 Blackwell Science Ltd British Journal of Clinical Pharmacology 42,

6 47 A. Varhe et al. and midazolam. Br J Clin Pharmacol 1994; 38: 175P. 16 Ervine CM, Houston JB. Disposition of azole antifungal [Abstract] agents. III. Binding of fluconazole and other azoles in rat 14 Grant SM, Clissold SP. Fluconazole: a review of its liver. Pharm Res 1994; 11: pharmacodynamic and pharmacokinetic properties, and therapeutic potential in superficial and systemic mycoses. Drugs 199; 39: (Received 12 February 1995, 15 Arredondo G, Calvo R, Marcos F, Martínez-Jordá R, accepted 12 June 1996) Suarez E. Protein binding of itraconazole and fluconazole in patients with cancer. Int J Clin Pharmacol T her 1995; 33: Blackwell Science Ltd British Journal of Clinical Pharmacology 42,