New oral anticoagulants (NOAC), also termed direct oral anticoagulants (DOAC), such as the thrombin inhibitor dabigatran etexilate or the factor Xa inhibitors rivaroxaban and apixaban show similar efficacy for stroke prevention in patients with atrial fibrillation as the vitamin K antagonist (VKA) warfarin [1, 2, 3]. One of the advantages of the NOACs, compared with VKAs, is said to be the lack of the necessity for laboratory monitoring and its lower rate of drug–drug and drug–food interactions; however, NOAC absorption is dependent on the intestinal P-glycoprotein system [4, 5, 6]. The aim of this review is to give an overview of P-glycoprotein-associated drug–drug and drug–food interactions with special reference to the data on the NOACs dabigatran, rivaroxaban, and apixaban in patients with atrial fibrillation.

Search strategy and selection criteria

A literature search was carried out by systematically screening MEDLINE for publications with the keywords dabigatran, rivaroxaban, apixaban, P-glycoprotein, and atrial fibrillation from 1998 to 2013. Reference lists and older references generated from initial papers were also considered. Randomized clinical trials, longitudinal studies, case series, and case reports were included.

Function of P-glycoprotein

P-glycoprotein 1 (P-gp) also known as multidrug-resistance protein 1 (MDR1), ATP-binding cassette sub-family B member 1 (ABCB1), or cluster of differentiation 243 (CD243) is a glycoprotein that in humans is encoded by the ABCB1 gene. It is a well-characterized ABC transporter that transports a wide variety of substrates across extracellular and intracellular membranes [7] and is a 160-kDa ATP-dependent drug efflux pump for xenobiotic compounds with broad substrate specificity. It is responsible for decreased drug accumulation in multidrug-resistant cells and often mediates the development of resistance to anticancer drugs. P-gp is expressed in the intestinal epithelium, hepatocytes, renal proximal tubular cells, adrenal glands, and capillary endothelial cells comprising the blood–brain and blood–testis barrier [8]. P-gp also functions as a transporter in the blood–brain barrier and is expressed in many cell types in the brain including the choroid plexus, astrocytes, microglia, and capillary endothelium. The most important physiological effect associated with P-gp expression appears to be in the luminal surface of capillary endothelial cells where the protein prevents the passage of drugs and toxins across the capillary membrane into the brain [9].

Substrates of P-gp

Substrates of P-gp comprise anticancer drugs, human immunodeficiency virus (HIV) protease inhibitors, H2-receptor antagonists, anti-gout agents, antidiarrheal agents, antiemetics, calcium channel blockers, cardiac glycosides, immunosuppressive agents, glucocorticoids, pesticides, anthelmintics, antibiotics, antidepressants, antithrombotic drugs, NOACs, and diagnostic dyes (Tab. 1, [10, 11, 12, 13, 14]).

Tab. 1 Drugs reported to be substrates or modulators of P-gp activity (from[10, 11, 12, 13, 14, 16])
Full size table

Modulators of P-gp activity

The activity of P-gp is modulated by several drugs, herbs, and food components. Generally, P-gp inhibitors increase the serum concentration of a substrate, whereas P-gp inducers decrease the serum concentration of a substrate. As shown in Tab. 1, several drugs are substrates as well as modulators of P-gp activity [14, 15].

Studies on pharmacokinetics and pharmacodynamics of drugs affecting P-gp are impeded by the fact that there is considerable overlap in drug specificity for P-gp and CYP3A. Furthermore, many of the data are derived from in vitro studies or experiments with P-gp knockout mice, thus the clinical relevance of the drug–drug interaction in humans is not completely clarified [15]. Several drugs have been identified as strong P-gp inhibitors as listed in Tab. 1 [14].

Food components, dietary supplements, and herbal drugs, such as phenolic acids and analogues, flavonoids, tannins, stilbenes, curcuminoids, coumarins, lignans, and quinones affect the P-gp system (Tab. 2, [16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]) These food compounds are frequently associated with beneficial health effects and are components of many over-the-counter drugs. Similarly with drug–drug interactions, many of these P-gp modulating effects have only been investigated in vitro or in animal models and the clinical relevance in humans is widely unknown.

Tab. 2 Food components, herbal drugs, and dietary supplements reported to modulate P-gp activity
Full size table

Additionally, P-gp displays considerable genetic heterogeneity. To date, more than 1,000 single-nucleotide polymorphisms for P-gp have been identified and the impact of these polymorphisms on drug metabolism is under investigation [32].

P-gp and the blood–brain barrier

Drugs modulating P-gp do not only influence the serum concentration of a drug but also the entry of substrates into several organs. That increased drug delivery to the brain by a P-gp inhibiting drug can be clinically relevant has been shown for the first time in humans for the antidiarrheal drug loperamide, which does not produce central nervous system effects at normal doses. After administration of quinidine, a known P-gp inhibitor, respiratory depression occurred that was not explained by increased plasma loperamide concentrations [33]. However, not all substances identified as a substrate of intestinal P-gp are also a substrate of brain P-gp, as has been shown with methadone [34]. Moreover, there are indications of differences between mice and humans in substrate specificity of P-gp, leading to differential P-gp-mediated transport efficiency [35].

Clinical consequences of P-gp modulation

Modulation of P-gp activity in order to improve the delivery of drugs to the target organs is of clinical interest in oncology and neurology, such as in epilepsy [36]. However, P-gp modulation can also become relevant as an unwanted side effect due to drug–drug or food–drug interactions [15].

Several drugs used in patients with cardiovascular disorders including atrial fibrillation are either substrates or modulators of P-gp activity (Tab. 1). A survey in hospitalized patients with atrial fibrillation showed that one or more P-gp-modulating drugs were prescribed to 42 % of the patients [37].

P-gp and oral anticoagulant drugs

The NOACs dabigatran etexilate, rivaroxaban, and apixaban are P-gp substrates, whereas the pharmacokinetics and pharmacodynamics of vitamin K antagonists do not seem to be affected by the P-gp system [38]. The clinical impact of drug–drug interactions between NOACs and P-gp-modulating substances is at present uncertain, mainly due to the lack of data. Most studies investigating drug–drug interactions of NOACs with P-gp-modulating drugs are from laboratories of the NOAC manufacturers [4, 5, 6, 39, 40, 41] and only few pharmacological studies from independent sources [11, 12]. No data about the relevance of food–drug interactions between NOACs and P-gp-modulating food components, herbal drugs, or dietary supplements have been published to date.

Bleeding and embolic complications that occurred during the three large clinical trials investigating NOACs in atrial fibrillation patients have not been analyzed with respect to comedications with P-gp-modulating drugs [1, 2, 3]. There are several case reports of bleeding complications, mainly under the NOAC dabigatran in patients with renal failure, which listed comedications of P-gp-inhibiting drugs (Tab. 2, [42]).

Dabigatran etexilate and P-gp

Dabigatran etexilate is an oral reversible direct thrombin inhibitor that is rapidly absorbed and converted to its active form, dabigatran. Dabigatran has been shown to be a potent, competitive, and reversible inhibitor of thrombin, inhibiting both activity and generation of thrombin. The precursor drug dabigatran etexilate, but not dabigatran, is a P-gp substrate and the bioavailability of dabigatran may be altered by P-gp inhibitors or inducers.

Three studies performed by the manufacturer of dabigatran investigated the effect of P-gp-modulating drugs on the bioavailability of dabigatran: Concomitant medication with proton pump inhibitors, amiodarone and verapamil increased the bioavailability of dabigatran [40]. A further study confirmed that verapamil increased dabigatran bioavailability [4]. A decrease in the bioavailability of dabigatran was found with concomitant intake of the P-gp inducer rifampicin [39]. In all these studies the effects of the changes in dabigatran bioavailability on the coagulation system were not investigated. A further independent study found that the P-gp inhibitor clarithromycin increased dabigatran bioavailability from 6.5 % to 10.1 % which resulted in a proportional prolongation of coagulation as measured by the ecarin clotting time [12]. As dabigatran is mainly excreted by the kidneys, concomitant renal insufficiency might further increase the bioavailability of dabigatran when given together with P-gp inhibitors. This assumption is substantiated by the frequent renal impairment in patients with bleeding complications and P-gp-inhibiting drugs [42].

No data are available about a potential influence of P-gp on dabigatran entry into the brain. Hypothetically, changes in the brain penetration of dabigatran by drugs affecting P-gp may not be expected as it is assumed that only the precursor drug dabigatran etexilate and not dabigatran is a P-gp substrate.

Rivaroxaban and P-gp

Rivaroxaban is an oral direct factor Xa inhibitor. Regarding the effects of drugs affecting P-gp on the pharmacokinetics of rivaroxaban, the manufacturer performed a study testing various P-gp-inhibiting drugs in healthy volunteers. In that study coadministration of erythromycin, clarithromycin, and fluconazole led to a 34–54 % increase in rivaroxaban plasma concentrations, whereas coadministration of ketoconazole and ritonavir led to an increase of 153–158 % [41]. The influence of these increases on blood coagulation has not been investigated.

The influence of P-gp on rivaroxaban entry into the brain has been investigated by two studies: An animal experiment performed by the manufacturer in wild-type and P-gp double-knockout mice demonstrated a slight increase in brain concentrations in P-gp double-knockout mice and decreased excretion into the gastrointestinal tract compared with wild-type mice [5]. Another independent in vitro and in vivo study using knockout mice showed that rivaroxaban is a shared substrate of P-gp and a further efflux transporter, breast cancer resistance protein (BCRP), and that these transporter proteins function synergistically. They appear to be particularly relevant for limiting rivaroxaban central nervous system entry [11]. Whether P-gp inhibitors increase the risk for cerebral bleeding by affecting the blood–brain barrier in rivaroxaban-treated patients has to be established. So far, there has only been one case reported with cerebral bleeding occurring under P-gp-inhibiting therapy [43].

Apixaban and P-gp

Apixaban is an oral direct factor Xa inhibitor. Few data are available about the influence of P-gp on apixaban. Apixaban has been reported as a substrate for P-gp [44]. Animal experiments in rats found a low brain concentration of apixaban suggesting that apixaban has limited penetration through the blood–brain barrier. Again, the efflux transporters P-gp and BCRP are constituents of the blood–brain barrier and are assumed to prevent or reduce drug entry [6].

Conclusion

This review shows that knowledge regarding the relevance of P-gp-modulating substances for drug–drug and drug–food interactions with NOACs is very limited. However, NOACs are already widely recommended and approved in many countries for stroke prevention in patients with atrial fibrillation or primary or secondary prevention of venous thromboembolism [45]. In view of the limited knowledge about drug–drug and drug–food interactions between NOACs and P-gp-modulating substances, there is an urgent need to investigate the role of P-gp-modulating substances as potential sources for drug–drug and drug–food interactions. We suggest a thorough analysis of the data accumulated in the large NOAC trials regarding the role of P-gp-modulating drugs in bleeding and embolic events, especially cerebral bleeding. Furthermore, pharmacological studies should be carried out to investigate the influence of P-gp-modulating drugs and food on NOAC plasma concentrations and coagulation parameters. When prescribing NOACs, patients should be informed about the potential interactions with drugs and herbal drugs. Patients who develop bleeding or embolic events under treatment with NOACs should be investigated with respect to comedications as well as over-the-counter drugs and dietary habits. In post-marketing surveillance of NOACs, the association with drugs or food intake with complications, bleeding, and embolic events should be registered.