The effect of bosentan on the pharmacokinetics of digoxin in healthy male subjects

Abstract

Aims

To investigate the effect of multiple oral dose treatment with the endothelin receptor antagonist bosentan on the pharmacokinetics of digoxin in healthy subjects.

Methods

This was an open-label, randomized, two-way crossover study in 18 evaluable young male subjects. They received, on two occasions which were separated by at least 2 weeks washout period, 0.375 mg digoxin once daily for 13 days following a loading dose of 0.375 mg given twice on the day before the once daily dosing regimen started. On one occasion treatment with 500 mg bosentan twice daily was started on the eighth day of digoxin treatment and continued for 1 week. Serum concentrations of digoxin were determined up to 24 h postdose on day 8 (first day of bosentan treatment) and day 14 (last day of bosentan treatment) of the digoxin treatment period. Plasma concentrations of bosentan were measured at two time points after the first bosentan dose and up to 12 h after the last morning dose of bosentan. Safety was assessed by adverse events, clinical laboratory tests, blood pressure and pulse rate measurements and ECG recordings.

Results

Steady-state of digoxin was always achieved after 7 days of treatment. Serum concentrations of digoxin were within the usual therapeutic range. Average steady-state C_max and C_tr were 2–2.1 μg l⁻¹ and 0.65–0.69 μg l⁻¹, respectively, when given alone. Bosentan did not lead to statistically significant changes in C_max and C_tr of digoxin. AUC (0,24h) of digoxin, however, was slightly reduced after 1 week of treatment with bosentan. The reduction was 12% on average with a narrow 95% confidence interval of 0-23%. Bosentan pharmacokinetic parameters after 1 week of treatment were as expected with a mean C_max of 3260 μg l⁻¹ and a mean AUC (0,12h) of 12600 μg l⁻¹ h.

Conclusions

Treatment with bosentan 500 mg twice daily for 1 week did not show clinically relevant effects on the pharmacokinetics of digoxin in healthy human subjects

Introduction

Endothelin-1 (ET-1), a 21-amino acid peptide, is the most potent and longest-lasting vasoconstrictor known to date. Since its first description and isolation in 1988 [1], evidence has been collected that the endothelin system plays a significant role in the pathogenesis of a variety of diseases, including hypertension, heart failure, renal failure and pulmonary diseases [2]. The potential of various endothelin antagonists as therapeutic interventions in some of these indications has been investigated in a number of small clinical trials [3].

Bosentan is a potent nonpeptide endothelin-receptor antagonist [4] that is in clinical development for the treatment of patients with chronic heart failure (CHF). The drug was shown to produce systemic and pulmonary vasodilatation in patients with chronic heart failure after both, acute intravenous doses [5] and 14 days of oral therapy [6]. Whether this translates into clinical benefit with respect to improvement of symptoms and reduction of mortality in long term is being investigated in large scale clinical trials.

The single and multiple dose pharmacokinetics in healthy volunteers have been described earlier [7, 8]. The drug has an oral bioavailability of approximately 50% at doses up to and including 600 mg which is reduced if the drug is taken on an empty stomach. Bosentan exhibits a low systemic plasma clearance of approximately 7–8 l h⁻¹ and a moderate volume of distribution of approximately 0.2 l kg⁻¹ at intravenous doses of 250–500 mg. Its apparent elimination half-life in man was between 5 and 8 h. In animals (rat and dog), the drug was eliminated mainly through hepatic metabolism followed by biliary excretion of the metabolites [9]. During multiple dosing regimens decreases in plasma levels of bosentan and its metabolites have been observed in human subjects which can be explained by a parallel approximately two-fold increase in systemic plasma clearance. At least part of this change in clearance is due to induction of cytochrome P4503A4 (CYP3A4) [8].

Digoxin is the most commonly used cardiac glycoside for the treatment of atrial fibrillation in patients with CHF. This drug has a narrow therapeutic window. Serum levels need to stay between 0.5 and 2.0 μg l⁻¹ for therapeutic efficacy and to avoid side effects such as certain forms of arrhythmia [10]. Digoxin has a long elimination half-life of approximately 36–48 h and is eliminated mainly by urinary excretion which occurs by both glomerular filtration and active tubular secretion [11]. Changes in renal blood flow due to the concomitant use of other drugs have been associated with changes in the elimination of digoxin [12]. Since endothelin antagonism may potentially lead to changes in renal blood flow [13], coadministration of bosentan could theoretically change the pharmacokinetics of digoxin. Another hypothetically possible mechanism of an interaction could be a bosentan induced change in the activity of P-glycoprotein which plays a role in the tubular secretion of digoxin [14].

Therefore and because bosentan will be given concomitantly with digoxin in many patients with CHF and because digoxin has a very narrow therapeutic window, the effect of bosentan on the pharmacokinetics of digoxin was investigated. For this reason, a drug-drug interaction study was performed in healthy volunteers who received usual therapeutic dosages of digoxin until steady-state, once without and once together with bosentan in order to evaluate the effect of bosentan multiple dose treatment on the steady-state pharmacokinetics of digoxin.

Methods

Subjects

Eighteen healthy young male subjects were planned to participate in this study and completed the treatment schedule. A total of 20 subjects were actually enrolled because two subjects dropped out for adverse events and were replaced. The study was approved by the Ethics Committee in Strasbourg, France (Comite Consultatif de Protection des Personnes dans la Recherche Biomedical d’Alsace No. 1). Written informed consent was obtained from each subject before enrolment. Physical examinations were performed and medical histories, routine laboratory tests, ECGs and vital signs were recorded before and after the course of drug treatment. Subjects with clinically relevant deviation from normal or any major illness within 1 month prior to the screening examination or smoking more than 10 cigarettes a day were not included. During the study no concomitant medication was allowed with the exception of medications to treat adverse events. The subjects were asked not to drink alcoholic beverages or smoke during the stays in the clinic.

Study design

This was an open-label study designed for 18 subjects. After eligibility screening the subjects were randomly allocated to the two treatment sequences AB and BA whereby A stands for treatment with digoxin alone and B for treatment with digoxin and bosentan together. Both treatments were separated by a washout period of at least 2 weeks. Digoxin treatment was started with a loading dose of 0.375 mg (one 0.125 mg and one 0.25 mg Lanoxin® tablet, Glaxo Wellcome) twice daily for 1 day followed by 13 days of 0.375 mg once daily. Bosentan 500 mg tablets were given twice daily for 1 week during treatment B, starting on day 8 of the digoxin treatment period. Both drugs were always taken within 10 min following food intake. Bosentan doses were taken at 12 h intervals. On days 1, 8 and 14 subjects took their medications in the clinic and stayed until the next morning. Blood for the pharmacokinetic assessments of digoxin was sampled from a forearm vein via an intravenous catheter before dosing and at 1, 2, 4, 6, 8, 10, 12, and 24 h after dosing on days 8 and 14 in each treatment period (A and B). In addition trough serum concentrations of digoxin on the second day of digoxin treatment were determined for safety reasons to ensure that the drug concentrations were within the usual therapeutic window. Blood for the pharmacokinetic assessments of bosentan was collected predose and at 4 and 12 h after dosing on day 8 (after the first dose of bosentan) and predose and at 1, 2, 4, 6, 8, 10 and 12 h after the morning dose on the last treatment day (day 14 treatment B, only). Twelve-lead ECGs were recorded every morning before the digoxin dose, and in addition at 12 h after the morning dose on day 1, and at 4 and 10 h after the morning dose on days 8 and 14 in each treatment period. Vital signs (blood pressure and heart rate) were measured at the same time points as ECG recordings were done, on days 1, 8 and 14 in each treatment period (A and B). Clinical laboratory tests were performed before and after the drug treatment phase in each subject. Standardized light meals were supplied 4 h and 12 h after the morning dose on days 8 and 14 in the clinic. Adverse events were recorded throughout the study.

Two subjects dropped out due to the occurrence of adverse events and were replaced by two new subjects who received the two treatments in the same order as was foreseen for the replaced ones.

Sample collection and drug assay

Blood samples for plasma concentration measurements of bosentan were collected into Vacutainers® containing EDTA as anticoagulant and centrifuged at 4° C. Plasma was separated and stored at −20° C until assayed. Blood samples for serum concentration measurements of digoxin were collected into plain Vacutainers®. They were centrifuged and serum was stored as indicated for bosentan plasma samples.

Plasma concentrations of bosentan were determined by a specific h.p.l.c.-u.v. assay [15]. The limit of quantification was 5 ng ml⁻¹. Serum concentrations of digoxin were determined using a commercially available radio immunoassay (Kodak Amerlex RIA). The limit of quantification of this method was 0.15 ng ml⁻¹. The performance of these analytical methods was monitored by simultaneous analysis of independently prepared quality control samples of various concentrations. The interassay precision was ≤10% for the digoxin assay and ≤6% for the bosentan assay.

Pharmacokinetic evaluation

Pharmacokinetic evaluation was performed using model independent methods. The pharmacokinetic parameters calculated were peak plasma/serum concentrations (C_max), time to reach C_max (t_max), trough plasma/serum concentration (C_tr), area under the plasma concentration time curve (AUC (0,12h) for bosentan and AUC (0,24h) for digoxin), oral plasma clearance (CL/F) and apparent terminal elimination half-life (t_1/2) if feasable.

C_max, C_tr and t_max values were taken directly from the observed plasma concentration time data. The area under the curve was estimated using the linear trapezoidal rule up to the last measured concentration value. The terminal elimination constant λ_z was estimated by performing standard unweighted log-linear least squares regression analysis of the terminal phase. The t_1/2,z was calculated by dividing ln2 by λ_z. Several subjects had serum concentrations of digoxin below the limit of quantification before the 24 h time point in one or both treatment periods. Their AUC (0,24h) was calculated by extrapolation from the last measurable time point up to 24 h using the last measurable serum concentration and the average estimated apparent elimination half-life for that particular treatment group and treatment day. Oral serum clearance of digoxin was calculated by dividing dose by AUC (0,24h).

Safety and tolerability assessments

At regular intervals clinical laboratory tests were performed, vitals signs (blood pressure and heart rate) were measured and 12-lead ECGs were recorded. Adverse events were monitored continously throughout the study.

Statistical methods

Pharmacokinetic parameters were descriptively analysed, calculating mean values and their standard deviation (s.d.) or, in case of t_max, calculating median and range. The primary parameters for the assessment of an interaction were AUC (0,24h) and C_max of digoxin on day 8 and on day 14 in each period. In addition, the trough serum concentrations (C_tr) of digoxin were evaluated but considered a secondary parameter. A 3-way ANOVA with the factors subject, period and treatment was applied to the log-transformed parameters, for day 8 and day 14, separately. From these results the ratios of the effects of treatment B to treatment A were estimated and 95% confidence intervals were calculated. The 95% confidence intervals were used for a judgement of the clinical relevance of the interaction. Only data from the 18 subjects who completed treatment were included in the pharmacokinetic evaluation.

Safety and tolerability were assessed using descriptive methods including all subjects randomized to treatment (n = 20).

Results

Pharmacokinetics

Steady-state of digoxin was already achieved after 7 days of treatment. Serum concentrations of digoxin when given alone were not different between day 8 and day 14 (Table 1). Peak serum concentrations within the usual therapeutic range were achieved within 1 and 2 h after dosing. These were 2.01±0.56 μg l⁻¹ (mean±s.d.) on day 8 and 2.12±0.31 μg l⁻¹ on day 14. Trough serum levels of digoxin (C_tr) were 0.69±0.27 and 0.65± 0.26 μg l⁻¹ on day 8 and 14, respectively. The average apparent elimination half-life was 44 and 42 h on day 8 and 14, respectively.

Table 1.

Pharmacokinetic parameters of digoxin. Data are shown as mean (s.d.), except for t_max where median (range) is reported.

There was no relevant difference in C_max and C_tr of digoxin between the treatment with bosentan and the treatment without bosentan (Figure 1). AUC (0,24h) of digoxin on day 8 was also not different between the two treatment periods. However, after 1 week of treatment with bosentan (day 14), AUC (0,24h) of digoxin was slightly reduced. This reduction was small, i.e. 12% on average, and within a narrow 95% confidence interval of 0–23% (Table 2). For C_max on day 14 one subject was considered a possible outlier. Excluding this subject from the analysis did not change the overall result. Therefore, the results presented in this manuscript include the data of this subject.

Figure 1.

Mean serum concentration time plots of digoxin at steady-state with and without concomitant treatment with bosentan 500 mg twice daily. Figure 1a: Day 8. Figure 1b: Day 14.

Table 2.

95% confidence intervals for the effect of bosentan on AUC (0,24h), C_max and C_tr of digoxin.

Mean C_max and AUC (0,12h) of bosentan on day 14 were 3260±1040 (mean±s.d.) μg l⁻¹ and 12 600± 3630 μg l⁻¹ h, respectively. The 95% confidence intervals for C_max and AUC (0,12h) were 2740–3780 μg l⁻¹ and 10 800–14 400 μg l⁻¹, respectively. The average trough plasma concentrations (12 h following dosing) decreased from day 8 (first bosentan dose) until day 14 from 991±1150 μg l⁻¹ to 177±171 μg l⁻¹. The plasma concentrations measured at 4 h after dosing also decreased, from 6995±2878 μg l⁻¹ to 2988±1230 μg l⁻¹.

Safety and tolerability

Two out of the 20 subjects enrolled were prematurely withdrawn and replaced by new subjects who received the same treatment sequence as was foreseen for the dropouts. One subject was withdrawn because of a 2nd degree atrioventricular block after having received the first dose of digoxin. Another subject was discontinued after having received the 13 days of digoxin treatment because of moderate bronchitis. In both cases bosentan treatment was not started yet when subjects were withdrawn.

There was no apparent difference between the two treatments A and B with respect to the occurrence of adverse events. Fourteen subjects (74%) on digoxin alone reported a total of 33 adverse events, 16 subjects (84%) on digoxin plus bosentan reported a total of 34 adverse events. The most frequently reported event was headache with mild to moderate severity which accounted for 30% of all adverse events in the digoxin treatment period and for 38% of all adverse events in the digoxin plus bosentan treatment period. There was no pattern of change in any of the clinical laboratory parameters and no marked abnormality. All ECGs with the one exception mentioned above, the atrioventricular block in one subject under digoxin, were normal. Vital signs did not change from baseline in any apparent pattern.

Discussion

The pharmacokinetics of digoxin when given alone were as expected from the literature [11]. The peak serum levels of 2 and 2.1 μg l⁻¹ (day 8 and 14 respectively) were within the usual therapeutic range. Steady-state was achieved before start of bosentan treatment. There was no relevant effect of bosentan on C_max or C_tr of digoxin. Although steady-state AUC of digoxin was decreased by 12% (95% CI 0–23%) after 1 week of treatment with bosentan it was concluded that this slight change would be of no clinical relevance. There was an apparent difference in elimination half-life on day 14 between the treatments that could be responsible for the difference in AUC. The smaller average C_tr during treatment with bosentan and digoxin on day 14—although not statistically significant—supports this finding. However, because of the relative long half-life of digoxin, an appropriate estimation of this parameter would have been possible only when collecting samples for longer than 24 h. Therefore no definitive conclusion with respect to a difference in elimination half-life can be drawn from this finding.

Since digoxin is mainly excreted via the kidney and bosentan mainly via the liver, it was not surprising that no relevant pharmacokinetic drug-drug interaction was observed in this study. However, theoretically an influence of bosentan on the renal elimination of digoxin could not be excluded for the following reasons.

ET-1 has an important role in the regulation of systemic and renal haemodynamics [2, 3, 16]. Elevated levels of ET-1 have been demonstrated in many pathological conditions including renal failure. Acute and chronic renal failure are both characterized by renal vasoconstriction leading to decreased renal blood flow and glomerular filtration rate. The renal vasculature in pigs was found to be particularly sensitive to the vasoconstricting effect of ET-1 [17]. Renal vasculature in man expresses mainly ET_A-receptors [18]. Systemic infusion of ET-1 in normal humans led to increases in renal vascular resistance with a consequent decrease in renal blood flow [19]. Such ET-1 induced effects could be ameliorated by an ET_A-specific ET-1 receptor antagonist in an animal species which has similar ET-receptor distribution than man [13]. It is therefore postulated that ET_A-specific ET-1 receptor antagonists would have beneficial effects in acute and chronic renal failure in man and probably other renal or hepatorenal diseases by decreasing pre- and postglomerular vasoconstriction leading to increased renal blood flow [16]. Increased renal blood flow could lead to an increased elimination of digoxin by changing its secretory clearance. Tubular secretion appears to be the major route of renal elimination of digoxin [20]. Some drug-interactions with digoxin have been shown to be due to a change of renal blood flow by the concomitantly given drug leading to a change of digoxin clearance. This was e.g. reported for sodium nitroprusside or hydralazine in patients with CHF [12]. Such a drug-interaction was not observed in the reported study with bosentan. However, the effects of bosentan on renal blood flow may differ between normal healthy volunteers with normal systemic and peripheral ET-1 levels and patients with renal diseases which are accompanied by elevated ET-1 levels and reduced renal blood flow. Therefore, studies in patients with renal disease or reduced renal blood flow may be necessary to be able to exclude a drug-interaction between digoxin and bosentan in these patients.

An important role in the tubular secretion of digoxin has been demonstrated for P-glycoprotein [21, 22]. Some of the drug-interactions observed with digoxin could be ascribed to an inhibition of this active transport protein. For example, cyclosporine A was shown to inhibit the P-glycoprotein mediated transport of digoxin leading to increased digoxin blood levels [14]. In vitro experiments with Caco2-cells to study transmembrane transport processes involved in the absorption and biliary excretion of bosentan, have shown that bosentan is a substrate for P-glycoprotein (Alsenz J, Roche Basel, personal communication). Other in vitro experiments using cultures of the human intestinal cell line LS-180 and primary human hepatocytes have revealed that bosentan is able to induce P-glycoprotein expression (Frank K, Roche Nutley, personal communication). Whether bosentan has any effect on renal P-glycoprotein is unknown. From the results in the described clinical study, it can be concluded that bosentan treatment for one week does not exhibit effects on renal or intestinal P-glycoprotein which would be of clinical significance with respect to an in vivo interaction with digoxin. The small effect on AUC of digoxin with no change in C_max could point towards an effect of bosentan on the elimination of digoxin. Whether this is due to induction of renal P-glycoprotein and whether this effect would become larger during a longer treatment duration needs further investigation.

The pharmacokinetic parameters of bosentan estimated in this study agree well with those from earlier studies [8] suggesting that digoxin most probably did not alter the pharmacokinetics of bosentan in this study.

In summary, treatment with bosentan 500 mg twice daily for 1 week did not show clinically relevant effects on the pharmacokinetics of digoxin in healthy human subjects.