Introduction

Ticagrelor (Fig. 1a) is a reversibly binding P2Y12 receptor antagonist [1] approved for the prevention of atherothrombotic events in patients with acute coronary syndrome (ACS). During the clinical development program, several studies showed that ticagrelor is rapidly absorbed and exhibits linear and predictable pharmacokinetics over a wide dose range [2–6]. Unlike thienopyridine compounds, ticagrelor does not require metabolic activation for antiplatelet activity [7, 8]. However, ticagrelor is metabolized to the active metabolite, AR-C124910XX (Fig. 1b), by cytochrome P450 3A (CYP3A [9]). AR-C124910XX is formed rapidly [3], and is approximately equipotent to its parent compound with respect to inhibition of platelet aggregation (AstraZeneca, data on file). Ticagrelor is also a substrate for P-glycoprotein (ABCB1; AstraZeneca, data on file), an efflux transporter protein found in the apical membrane of the intestinal mucosa [10] and in the liver and kidney [11].

Fig. 1
figure 1

Chemical structure of a ticagrelor and b its active metabolite AR-C124910XX

Full size image

Patients with ACS often exhibit multiple comorbidities and may receive a number of drugs concomitantly [12]. Digoxin is a substrate of renal [13] and intestinal P-glycoprotein [14, 15], but not cytochrome P450 enzymes. Therefore, digoxin is a suitable model compound for assessing potential interactions with drugs that interact with P-glycoprotein [16, 17]. In earlier preclinical studies, ticagrelor and AR-C124910XX inhibited the transport of 3H-digoxin in MDR1-MDCK monolayers with an IC50 of 7.8 ± 2.6 μM and 9.9 ± 5.1 μM, respectively [18]. Assuming a steady state maximum plasma concentration (Cmax) for ticagrelor of 1.5 μM [9] and 0.5 μM for AR-C124910XX, corresponding I/IC50 values were approximately 0.19 for ticagrelor and 0.05 for AR-C124910XX. Based on the FDA criteria [16], as the I/IC50 of 0.19 for ticagrelor was ≥ 0.1, an interaction with digoxin was deemed likely and a clinical interaction study necessary. Such an interaction could be clinically relevant for compounds that have a narrow therapeutic range, such as digoxin [19]; the therapeutic range for digoxin is a serum concentration of 0.5–1.0 ng/ml [20, 21]. Furthermore, there is a possibility that digoxin and antiplatelet agents may be co-administered [22, 23]. Current guidelines include recommendations for the use of digoxin in patients with atrial fibrillation, heart failure, or both conditions [24, 25]. Atrial fibrillation is estimated to occur in 2–21 % of patients with ACS [26], and digoxin may be administered to slow a rapid ventricular response in patients with both ACS and atrial fibrillation associated with heart failure [25].

The primary objective of the present study was to compare the pharmacokinetics of digoxin when administered alone and in combination with ticagrelor. Secondary objectives included assessment of the effect of digoxin on the pharmacokinetics of ticagrelor, and evaluation of the safety and tolerability of digoxin and ticagrelor co-administration.

Materials and methods

Study population

The study was conducted in healthy volunteers, who were required to be 18–65 years old with a body mass index (BMI) 18–30 kg/m2; females were to be post-menopausal or surgically sterile. Key exclusion criteria are included in the online supplement.

All volunteers provided written, informed consent prior to study start and the final protocol was approved by an Independent Ethics Committee (Southern Institutional Review Board, Miami, FL, USA; reviewed 18 September 2003). The study was conducted in accordance with the ethical principles established in the Declaration of Helsinki, and was consistent with International Conference on Harmonisation/Good Clinical Practice guidelines, AstraZeneca Bioethics Policy, and regulatory requirements.

Study design and treatment

This study (AstraZeneca Trial Identifier: D5130C05265) was a randomized, double-blind, two-period crossover study conducted in a single center (Fig. 2). Volunteers were randomized to one of the two treatment sequences. Both staff and volunteers were blinded to treatment with ticagrelor and placebo, whereas digoxin administration was open label.

Fig. 2
figure 2

Study design

Full size image

Following randomization, volunteers received the allocated treatment during two 16-day periods, separated by a washout of at least 2 weeks. Ticagrelor at 400 mg once daily (qd) was administered on days 1–16. Digoxin (Lanoxin®, GlaxoSmithKline) at 0.25 mg was given twice daily (bid) on day 6 and at 0.25 mg qd on days 7–14 (Fig. 2). Both ticagrelor and digoxin tablets were administered with 240 ml of room temperature water 1 h before breakfast, and the day 6 evening dose of digoxin was given approximately 12 h after the morning dose. On days 5 and 14, volunteers fasted for 10 h overnight before, and for 4 h after, receiving study medication; water was also prohibited 2 h before and after dose administration.

Key on-study restrictions are included in the online supplement.

Pharmacokinetic assessments

For pharmacokinetic analyses of ticagrelor and AR-C124910XX, blood samples were collected pre-dose on days 1, 4, 5, 13, and 14, and 0.5, 1, 2, 3, 4, 6, 8, 10, 12, 18, and 24 h post-dose on days 5 and 14. Additional samples were collected at 36, 48, and 72 h post-dose on day 14. For digoxin pharmacokinetic analyses, blood samples were collected pre-dose on days 1, 12, 13, and 14 and 0.5, 1, 2, 3, 4, 6, 8, 10, 12, 18, and 24 h post-dose on day 14. Urine samples were also collected pre-dose on day 1 and between 0 and 6 h; 6 and 12 h; and 12 and 24 h post-dose on day 14.

Venous blood (2 ml) was collected into lithium-heparin tubes and immediately placed on ice. Samples were centrifuged at 1,500 g and 4 °C for 10 min and stored frozen at −20 °C until analyzed.

Ticagrelor and AR-C124910XX plasma concentrations were quantified by a validated method using liquid chromatography and tandem mass spectrometry. Lower limits of quantification (LLOQ) were 1 ng/ml for ticagrelor and 2.5 ng/ml for AR-C124910XX [27]. The inter-batch precision and mean accuracy of this method were 3.5–10.7 % and 1.0–12.7 %, respectively, for ticagrelor; corresponding values for AR-C124910XX were 4.7–10.5 % and −3.0–3.2 %, respectively. Digoxin plasma and urine concentrations were measured by PPD Development (Richmond, VA, USA) using validated radioimmunoassays with LLOQs of 0.15 ng/ml (plasma; method ICD 11.1) and 1 ng/ml (urine; method ICD 11.2). Inter-batch precision and mean accuracy were 5.1–10.4 % and −1.5–5.6 %, respectively for the plasma assay; corresponding values for the urine assay were 4.9–13.7 % and 0.65–3.3 %, respectively.

Safety and tolerability

Safety and tolerability were evaluated by monitoring the incidence, severity, and causality of adverse events (AEs) during the study, until follow-up. Clinical laboratory parameters (clinical chemistry, hematology, coagulation, urinalysis), 12-lead electrocardiogram (ECG), and vital signs were assessed throughout the study, and a complete physical examination was performed at screening and at the follow-up visit.

Data analyses

Pharmacokinetic parameters (see online supplement) were estimated by non-compartmental methods using WinNonlin Professional (Pharsight Corporation, Mountain View, CA, USA).

Data were analyzed statistically using SAS version 8 (SAS Institute Inc., Cary, NC, USA). Following log-transformation, Cmax, AUCτ and Cmin (digoxin only) were analyzed by analysis of variance; treatment, sequence, and period were fixed terms, and volunteer within sequence was treated as a random effect. Comparisons were expressed as the geometric mean (Gmean) ratio between treatments (ticagrelor plus digoxin versus placebo plus digoxin).

Results

Baseline characteristics and disposition

Twenty healthy volunteers (ten males and ten females) were enrolled and randomized. The mean (range) age was 44 (22–59) years and the mean (range) BMI was 25.9 (21.8–29.5) kg/m2. Sixteen volunteers were classified as ‘other’ race (mainly Hispanic), three were Caucasian, and one was Black.

All 20 volunteers received ticagrelor and 16 completed the study. Of the four volunteers who did not complete the study, two volunteers completed all of the study procedures except the follow-up visit, one was unwilling to continue, and one discontinued because of persistent pruritus (discontinued on day 9). One volunteer missed a ticagrelor dose on day 5, and was excluded from the ticagrelor and AR-C124910XX pharmacokinetic analyses.

Digoxin pharmacokinetic parameters

Digoxin was rapidly absorbed following co-administration with either ticagrelor or placebo (Table 1, Fig. 3). However, plasma levels of digoxin were higher with co-administration of ticagrelor compared with placebo. The median t1/2 of digoxin was slightly prolonged in the presence of ticagrelor (52.1 h) compared with placebo (41.8 h), whereas the median tmax of digoxin remained unchanged (Table 1). Relative increases in AUCτ (28 %), Cmax (75 %), and Cmin (31 %) were observed in the presence of ticagrelor compared with placebo (Table 1). A greater than 50 % increase in these variables occurred in four, twelve, and five volunteers, respectively. The maximum increase observed in these parameters was 2–2.5 fold. In the presence of ticagrelor, digoxin Ae was increased by 24 %, and renal clearance was unchanged, versus placebo (Table 1).

Table 1 Pharmacokinetic parameters of digoxin in the presence or absence of ticagrelor
Full size table
Fig 3
figure 3

Mean (± SD) plasma concentrations of digoxin, on day 14, in the presence and absence of ticagrelor

Full size image

Ticagrelor pharmacokinetic parameters

Plasma concentrations of ticagrelor were comparable upon co-administration of digoxin or placebo (Table 2, Fig. 4a). Gmean AUCτ and Cmax values were not altered when co-administered with digoxin, with Gmean ratios for AUCτ and Cmax of 1.06 (90 % confidence interval [CI], 0.97–1.16) and 1.02 (90 % CI, 0.94–1.11), respectively (Table 2). Median tmax for ticagrelor alone was 2.2 h and was comparable (3.2 h) in the presence of digoxin (Table 2).

Table 2 Pharmacokinetic parameters of ticagrelor and AR-C124910XX
Full size table
Fig 4
figure 4

Mean (± SD) plasma concentrations of a ticagrelor and b AR-C124910XX, in the presence (day 14) and absence (day 5) of digoxin

Full size image

AR-C124910XX pharmacokinetic parameters

AR-C124910XX plasma concentrations were comparable in the presence and absence of digoxin (Table 2, Fig. 4b). No change in Gmean AUCτ and Cmax values occurred upon co-administration with digoxin, with Gmean ratios for AUCτ and Cmax of 1.11 (90 % CI, 1.01–1.23) and 1.06 (90 % CI, 0.97–1.16), respectively (Table 2). Median tmax values for AR-C124910XX occurred at 3.2 h both in the presence and absence of digoxin (Table 2). Additionally, metabolite:parent ratios for AUCτ and Cmax were unaffected by the presence of digoxin.

Safety and tolerability

Most AEs reported with co-administration of ticagrelor and digoxin were mild and self-limiting. One volunteer receiving ticagrelor plus digoxin discontinued the study on day 9 due to a treatment-related pruritis that was persistent and required treatment with Benadryl.

Overall, 20 AEs were noted in ten out of 20 volunteers receiving ticagrelor plus digoxin, and 16 AEs were recorded in ten out of 19 volunteers receiving placebo plus digoxin. The most common AEs were an increased tendency to bruise, stomatitis, headache, and arthralgia. Seventeen treatment-related AEs occurred in eight volunteers during ticagrelor plus digoxin administration; 15 treatment-related AEs occurred in nine volunteers during placebo plus digoxin administration.

There were no clinically significant changes in other clinical chemistry parameters, hematology, urinalysis, ECGs, vital signs, or physical findings in either group. However, increases in mean serum uric acid values over time were evident for both treatments; the mean (SD) increase from baseline on day 15 was 30 (64) and 13 (53) μmol/l for the ticagrelor plus digoxin and placebo plus digoxin treatments, respectively.

Discussion

Digoxin is a renal and intestinal P-glycoprotein substrate [13–15], and does not undergo metabolism by cytochrome P450 enzymes. Thus, this drug is a useful model compound with which to assess the effect of co-administered drugs on the exposure to P-glycoprotein substrates [16]. The pharmacokinetics of digoxin in the presence of placebo in our study were comparable to those observed in previous studies [28, 29]. In vitro studies have shown that ticagrelor is a substrate and inhibitor of P-glycoprotein [18]. The present findings confirm an interaction between digoxin and ticagrelor, as predicted by the results of earlier in vitro studies. This study showed that compared with placebo, ticagrelor co-administration increased the Cmax and AUCτ of digoxin by approximately 75 % and 28 %, respectively. Moreover, the renal clearance of digoxin from the plasma was not changed in the presence of ticagrelor versus placebo. Collectively, these findings show that the increased exposure to digoxin observed with co-administration of ticagrelor could be explained by the inhibition of intestinal P-glycoprotein by ticagrelor, thereby reducing the efflux of digoxin into the intestine. As in vitro studies have previously shown the parent compound and active metabolite AR-C124910XX to have similar effects with respect to P-glycoprotein inhibition, both may contribute to the inhibition of intestinal P-glycoprotein observed in the present study. Our results also suggest that interaction of ticagrelor with renal P-glycoprotein is of minimal significance for digoxin. Together with data from previous in vitro studies [18], the present results confirm that ticagrelor is a weak inhibitor of intestinal P-glycoprotein.

The 400 mg qd ticagrelor dose used in the present study is higher than the approved 90 mg bid maintenance dose used clinically [30, 31]. The pharmacokinetics of ticagrelor in the present study were broadly comparable to those observed in previous studies in healthy volunteers [2], and in patients with ACS [6] or atherosclerosis [5]. In our study, digoxin had no effect on the pharmacokinetics of ticagrelor or AR-C124910XX. The lack of effect of digoxin on ticagrelor pharmacokinetic parameters is not surprising, as digoxin does not interact with CYP3A and is not an inhibitor of P-glycoprotein [16].

Patients with ACS often present with cardiovascular comorbidities [26] and may receive multiple drugs concomitantly [12]. Moreover, there are a number of clinical scenarios where digoxin could be co-administered with ticagrelor. For example, current treatment guidelines recommend the use of digoxin monotherapy for heart rate control in patients with permanent atrial fibrillation who are predominantly sedentary [24]. Additionally, digoxin is an appropriate alternative for patients with ACS associated with severe left ventricular dysfunction (LVD) and heart failure [25], and is also recommended for use in symptomatic LVD with and without atrial fibrillation [32, 33]. Indeed, in patients with both heart failure and atrial fibrillation, digoxin is recommended as an adjunctive therapy for heart rate control when monotherapy with β-blockers proves inadequate [25]. Therefore, based on the results of this study, close clinical and laboratory monitoring is recommended in the event of co-administration of digoxin and ticagrelor. As with other narrow therapeutic drugs such as warfarin [34], therapeutic drug monitoring will be useful in maintaining digoxin within the therapeutic range (serum concentration of 0.5–1.0 ng/ml [20]) and avoiding adverse effects associated with higher exposures. Higher serum digoxin concentrations are associated with toxicity. For example, in patients with heart failure and/or atrial fibrillation, toxicity occurred in some patients with digoxin serum concentrations in the range 1.4–2.9 ng/ml, whereas toxicity was consistently observed in patients who had digoxin serum concentrations of 3 ng/ml or greater [35].

With respect to efficacy, early high plasma concentrations of digoxin may not give an accurate measurement of the concentration of digoxin at the site of action [36]. However, with chronic digoxin use, steady-state plasma concentrations reach equilibrium with the surrounding tissue concentrations and correlate with specific pharmacologic effects. Therefore, digoxin plasma concentrations measured after the tissue distribution phase may more accurately reflect the likely effect on its efficacy because of the relatively slow equilibration of digoxin within the myocardium. In our study, a 31 % increase in digoxin Cmin was observed in the presence of ticagrelor versus placebo.

The magnitude of the overall increase in digoxin exposure observed in the present study (28 %) is similar to that reported for other P-glycoprotein inhibitors. For example, mibefradil increased digoxin AUC by 31 % [37].

Co-administration of ticagrelor and digoxin was well tolerated. No safety issues were identified during the present study. AEs were generally mild, although one volunteer discontinued from the study due to persistent pruritus that was considered to be treatment related.

In summary, ticagrelor 400 mg qd and digoxin 0.25 mg qd showed a pharmacokinetic interaction. At steady-state, digoxin Cmax, Cmin, and AUCτ were increased by 75 %, 31 %, and 28 %, respectively, in the presence of ticagrelor versus placebo. No effect was seen on the pharmacokinetics and safety of ticagrelor when co-administered with digoxin. These data provide further evidence that ticagrelor is a weak inhibitor of intestinal P-glycoprotein. Based on these findings, it is recommended that serum concentrations of drugs like digoxin (i.e., P-glycoprotein transporter substrates with a narrow therapeutic range) are monitored with initiation of, or any change in, ticagrelor therapy.