Drug–Drug Interactions in the Treatment of Cancer-Associated Venous Thromboembolism with Direct Oral Anticoagulants

Abstract

Venous thromboembolism (VTE) is a frequent complication of cancer, and management of cancer-associated thrombosis (CAT) is challenging due to increased risks of bleeding and recurrent VTE. Recent trials have shown an acceptable efficacy and safety of direct oral anticoagulants (DOACs) in the treatment of CAT compared to low-molecular weight heparin. Although DOACs provide an effective and convenient treatment option in CAT, the need to assess the risk of drug–drug interactions (DDI) with antineoplastic therapies poses a barrier to their use in clinical practice. With the aim of supporting the assessment of CAT patients for treatment with DOAC, this review provides a comprehensive overview of the compatibility of antineoplastic therapies with the individual DOACs (apixaban, dabigatran, edoxaban, and rivaroxaban). Using several data sources, we characterized 100 widely used antineoplastic agents with regard to their effect on p-glycoprotein and cytochrome P450, both important in the transport and elimination of DOACs. This enabled us to evaluate 400 “DOAC-antineoplastic agent”-pairs regarding their likelihood to interact (unlikely, potential, or likely), ultimately leading to clinical recommendations on the appropriateness of concomitant use for each pair. A potential or likely DDI was identified for 12% of the evaluated pairs. For nearly all antineoplastic agents, at least one DOAC was considered compatible.

Keywords

anticoagulants - cancer - drug interactions - venous thromboembolism

Venous thromboembolism (VTE) and cancer are closely related. In incident VTE events, cancer will be present in around 20% of patients.[1] VTE is an independent predictor of mortality in patients with cancer,[2] and not surprisingly, cancer-associated venous thromboembolism (CAT) is associated with reduced quality of life and places a significant burden on both the patient and health care system.[3] Therapeutic management of CAT is challenging because of high risks of recurrent VTE and anticoagulant-related bleeding.[4] Importantly, treatment of CAT is evolving. During recent years, considerable trial evidence showing an acceptable efficacy and safety of three direct oral anticoagulants (DOACs: rivaroxaban, apixaban and edoxaban) compared to low-molecular weight heparin (LMWH) in CAT, has emerged.[5] [6] [7] Accordingly, DOACs have now been included as treatment options in selected CAT patients in several international guidelines.[8] [9] [10] While DOACs provide a more convenient and potentially less expensive treatment alternative to LMWH, they carry a higher risk of minor bleeding episodes[11] as well as pharmacokinetic drug interactions with several antineoplastic agents and supportive therapies. Due to the latter, guidelines specify that the risk of drug interactions should be carefully considered when deciding whether a patient with CAT is eligible for DOAC treatment.[8] [9] [10] Unfortunately, there is limited clinical knowledge of drug–drug interactions (DDI) between DOACs and drugs used in cancer treatment. Therefore, assessments of potential DDIs will, in most cases, rely on theoretical considerations of the compatibility of pharmacological properties of the DOAC with those of the antineoplastic agent or supportive therapy. Such assessments are, however, often challenging in clinical practice, and will likely result in unnecessary avoidance of DOAC treatment in patients with CAT. Also, the complexity may result in clinically important DDIs being overlooked, potentially compromising the safety or effectiveness of DOAC therapy.

With this work, we provide specific and clinically applicable insight into the potential DDIs between DOACs and antineoplastic agents. Our aim is to facilitate clinical guidance when assessing CAT patients for eligibility of DOAC use. We did this by assessing a total of 400 “DOAC-antineoplastic agent”-pairs for their potential to cause clinically relevant DDIs. We reviewed available clinical and non-clinical data on the pharmacological properties of each DOAC and each antineoplastic agent. For each potential drug pair, we considered the compatibility of properties, as commonly done when evaluating the potential for DDIs.[12] As part of the process, we developed a framework for evaluation of the interaction potential for any agent with DOACs, enabling easy assessment of future antineoplastic agents.

Methods

The assessment of the potential for DDIs between the individual DOACs and each of the included antineoplastic agents included four steps. First, we defined the drug properties of relevance when evaluating the potential for a given drug to interact with DOACs. Second, we collected data on these specific properties for the included antineoplastic agents. Third, we combined steps one and two to evaluate the potential for each “DOAC-antineoplastic agent”-pair regarding their likelihood to interact. Finally, for each pair we provided a clinical recommendation on the appropriateness of concomitant use.

Included DOACs and Antineoplastic Agents

To ensure completeness, we included all four DOACs approved for the treatment of VTE, i.e., dabigatran etexilate, edoxaban, rivaroxaban, and apixaban. Of note, only the three latter DOACs have been tested in large, randomized trials of patients with CAT. We included a total of 100 antineoplastic agents, collectively representing most of the agents currently used in clinical oncology. The list comprised most agents assessed in the 2021 European Hearth Rhythm Association (EHRA) Practical Guide on non-vitamin K antagonist oral anticoagulant use in patients with atrial fibrillation[13] as well as tyrosine kinase inhibitors not included in the guide (e.g., due to recent market entry). We did not include antibody-based therapies (e.g., monoclonal antibodies and immunotherapies) as, although widely used in cancer therapy, their potential to result in clinical important pharmacokinetic DDIs is generally considered to be low.[14]

Pharmacological Properties of Relevance in DOAC-Related Drug Interactions

While DOACs are not expected to affect the pharmacokinetics of other drugs,[15] [16] [17] [18] other drugs can affect the pharmacokinetics of DOACs leading to fluctuations in DOAC plasma concentration, potentially changing the effectiveness and safety of DOAC treatment. Based on knowledge from in vivo and/or in vitro preclinical DDI studies, the Summary of Product Characteristics (SmPC) of each of the DOACs[15] [16] [17] [18] specify the drug properties expected to interfere significantly with DOAC pharmacokinetics, and provides recommendations for limitations in the use of DOACs in combination with other drugs ([Table 1]). The currently identified pharmacokinetic DDIs relevant to DOACs and mentioned in the DOAC SmPCs involve drugs affecting the activity of the efflux transporter P-glycoprotein (P-gp) and/or the Cytochrome P450 (CYP) enzyme CYP3A4. While all DOACs, and especially dabigatran etexilate (dabigatran), are substrates of P-gp, only rivaroxaban and apixaban are clinically relevant substrates of CYP3A4. Although other structures may be involved in the transport and elimination of DOACs, e.g., breast cancer resistance protein,[19] these have, to our knowledge, not been identified as primary mediators of clinically relevant DDIs involving DOACs,[20] and have therefore not been included in this evaluation.

Table 1 Pharmacokinetic properties of direct oral anticoagulants (DOAC) and summary of the manufacturers' recommendations on concomitant drug use as stated in the European Summaries of Product Characteristics

	Apixaban	Rivaroxaban	Edoxaban	Dabigatran etexilate
Pharmacokinetic properties[a]
CYP3A4	Substrate (minor) Eliminates 15% of the active drug	Substrate (minor) Eliminates 18% of the active drug	Substrate (negligible) Eliminates 4% of the active drug	–
P-glycoprotein	Substrate (moderate)	Substrate (moderate)	Substrate (moderate)	Substrate (extensive)
Recommendations on concomitant drug use
Drugs that should be avoided[b]	Strong inhibitors of both CYP3A4 and P-gp Strong inducers of both CYP3A4 and P-gp	Strong inhibitors of both CYP3A4 and P-gp Strong inducers of CYP3A4	Not mentioned	Strong P-gp-inhibitors All P-gp-inducers
Drugs that should lead to DOAC dose reduction	Not mentioned	Not mentioned	Strong P-gp-inhibitors	Moderate P-gp- inhibitors
Drugs that should be used with caution	Not mentioned	Strong and moderate inhibitors of CYP3A4	All P-gp-inducers	Mild P-gp-inhibitors
Drugs that can be used with DOAC	Mild and moderate inhibitors of P-gp and/or CYP3A4	Not mentioned	Mild P-gp-inhibitors	Not mentioned

Abbreviations: CYP3A4, cytochrome P450 3A4; P-gp, p-glycoprotein.

^a The classifications of substrates are derived from Foerster et al 2020.[20]

^b Including drugs where concomitant use is not recommended.

Based on the recommendations from the DOAC SmPCs on concomitant drug use, we developed a framework enabling assessment of the potential for pharmacokinetic DDIs with all DOACs for any drug ([Table 2]). The framework classifies the likelihood of a given drug to interact significantly with each of the DOACs as either “unlikely”, “potential”, or “likely”. Combinations classified as “unlikely” are combinations considered to have no potential to cause clinically relevant DDIs. Combinations that could lead to minor fluctuations in DOAC plasma concentrations are classified as “potential”. These combinations are not expected to infer clinically relevant DDIs, but in the context of other risks factors, e.g., polypharmacy or decreased renal function, they may contribute to clinically relevant changes in DOAC plasma concentrations. As such, these combinations do not need to be avoided, but should be used with caution. Combinations classified as “likely” included contraindicated/not recommended combinations as well as those where the DOAC should be administered in a reduced dose. Due to the overall lack of knowledge on dose reduction of DOACs in the context of DDIs, we recommend avoidance of all combinations classified as “likely.”

Table 2 Framework for the classification of the likelihood for drug interaction with direct oral anticoagulants (Apixaban (A), rivaroxaban (R), edoxaban (E) and dabigatran etexilate (D)) according to a drug's effect on CYP3A4 and P-glycoprotein (P-gp)

Note: Based on the recommendations on concomitant drug use in the summary of product characteristics of the individual direct oral anticoagulants. Interpretation: green: interaction unlikely, yellow: potential interaction/use with caution, red: likely interaction/avoid combination.

Pharmacological Properties of Antineoplastic Agents

We collected information on the drug properties of antineoplastic agents of relevance to their potential for DOAC interactions, that is, their ability to modify (i.e., inhibit and induce) the activity of drug transporter P-gp and the metabolizing enzyme CYP3A4.[20] This information was gathered for each agent using several data sources; drug-specific regulatory documents of relevance, e.g., the SmPCs and labels,[21] [22] [23] the IBM Micromedex database,[24] a comprehensive review of modulation of P-gp by drugs,[25] as well as published DDI studies referred to in the Micromedex database or in the Danish Drug Interaction Database.[26] If a classification regarding effect on P-gp and/or CYP3A4 (described below) could not be reached from these data sources, or in case of conflicts between data sources, we searched MEDLINE for papers reporting on the pharmacokinetics of the specific antineoplastic agent as well as on DDI-studies involving the drug. Data were extracted and evaluated by two of the authors, both physicians and specialists in clinical pharmacology (M.H. and J.N.H.). To ensure consistency, all drugs classified as inhibitors or inducers of P-gp and/or CYP3A4 were evaluated by both reviewers.

Based on the extracted data, antineoplastic agents were classified as substrates, inhibitors, and/or inducers of P-gp and CYP3A4, or neither of these. In vivo DDI-studies using a relevant probe drug (e.g., digoxin in the context of P-gp) were preferred. When in vivo studies were lacking, we considered in vitro studies as well, but only if conducted in agreement with the regulatory recommendations on the performance of DDI studies from European Medicines Agency[12] and the U.S. Food and Drug administration.[27] For drugs where in vitro and in vivo data disagreed, in vivo data was prioritized. For some drugs assessed in vitro alone, the assessment had been qualified with an extrapolation of the drug concentration needed to modify P-gp or CYP3A4 in vivo. If this concentration was markedly higher than the concentration usually obtained during clinical use, we did not consider the drug to be a clinically relevant inhibitor/inducer. As an example, when only considering in vitro data, afatinib is an inhibitor of P-gp. However, as the drug concentration of afatinib needed to obtain an P-gp inhibiting effect in vivo is 20 times higher than the drug concentration normally seen during clinical use,[28] we did not classify afatinib as a clinically relevant P-gp inhibitor.

When possible, agents classified as inhibitors or inducers of P-gp and/or CYP3A4 were further classified according to the degree of inhibition or induction (i.e., mild, moderate, or strong). For classification, we used the regulatory defined cut-offs for relevant changes in the area under the curve (AUC) of standard probe drugs, when administered concomitantly with the perpetrator drug, as described in [Appendix A] (available in the online version only).[12] [27] These cut-offs are intended for the classification of inhibitors and inducers of CYP3A4. However, as no corresponding definition exist for P-gp inhibition and induction, we chose to apply the same cut-offs in the classification of drugs affecting P-gp-activity. Agents were only classified according to the degree of inhibition and/or induction if an in vivo DDI-study using a standard probe drug was available. In case of disagreement between relevant DDI-studies regarding the level of inhibition or induction, the drug was, from a safety point of view, classified with the most potent of the levels proposed. When the degree of inhibition or induction could not be reached, we considered the antineoplastic agent as a moderate inhibitor or inducer when evaluating the interaction potential. This assumption was subjected to a sensitivity analysis, in which the agents were considered strong inhibitors/inducers.

Evaluation of the Potential for DDIs in “DOAC-Antineoplastic Agent”-Pairs

To evaluate the risk of a clinically relevant DDI to occur when combining a DOAC with an antineoplastic agent, we applied the information gathered on the pharmacological properties of the antineoplastic agent to the framework shown in [Table 2]. Using this, each “DOAC-antineoplastic agent”-combination was classified according to its likelihood to interact significantly as either “unlikely”, “potential”, or “likely” as described above.

Statistics

Results were analyzed using descriptive statistics.

Results

Among the 100 antineoplastic agents reviewed, 38 (38%) were classified as influencing P-gp and/or CYP3A4 ([Table 3]). All of the nine agents classified as influencing P-gp alone, were inhibitors of P-gp. Similarly, inhibition was a more common property than induction among the 13 agents classified as modulators of CYP3A4 alone; 10 versus three agents, respectively. Sixteen agents had an effect on both CYP3A4 and P-gp, with dual-inhibition being more common (n = 9) than dual induction (n = 2) and mixed inhibition and induction (n = 5).

Table 3 Effect of antineoplastic agents on the activity of cytochrome P450 3A4 (CYP3A4) and P-glycoprotein (P-gp)

Antineoplastic agent	CYP3A4		P-gp		Antineoplastic agent	CYP3A4		P-gp
	Inh.	Ind.	Inh.	Ind.		Inh.	Ind.	Inh.	Ind.
Abemaciclib					Ifosfamide
Abiraterone	(√)				Imatinib	√**		(√)
Afatinib					Irinotecan
Alectinib			(√)		Lapatinib	√*		√*
Acalabrutinib					Larotrectinib	√*
Anastrozole					Lenvatinib
Apalutamide		√***		√*	Letrozole
Axitinib					Leuprolide/leuprorelin
Bendamustine					Lomustine
Bicalutamide	√*				Lorlatinib		√**		√**
Binimetinib					Mephalan
Bleomycin					Mercaptopurine
Bortezomib	(√)				Methotrexate
Bosutinib					Mitomycin
Brigatinib		√*	(√)		Mitotane		√***
Busulfan					Mitoxantrone
Cabozantinib			(√)		Neratinib			√*
Capecitabine					Nilotinib	√**		(√)
Carboplatin					Nintedanib
Carfilzomib					Niraparib
Carmustine					Olaparib	(√)	(√)	(√)
Ceritinib	√***		(√)		Osimertinib			√*
Chlorambucil					Oxaliplatin
Ciclosporin	√***		√***		Paclitaxel
Cisplatin					Palbociclib	√*
Cobimetinib					Pazopanib	√*
Crizotinib	√**		(√)		Pemetrexed
Cyclophosphamide					Ponatinib			(√)
Cytarabine					Prednisone
Dabrafenib		√**			Procarbazine
Dacarbazine					Regorafenib
Dactinomycin					Ribociclib	√**
Dasatinib	√*				Ruxolitinib			(√)
Daunorubicin					Sirolimus
Dexamethasone		√**			Sorafenib			(√)
Docetaxel					Sunitinib
Doxorubicin					Talazoparib
Encorafenib	(√)	(√)	(√)		Tacrolimus	(√)		(√)
Enzalutamide		√***	√*		Tamoxifen	(√)
Erlotinib					Temsirolimus	(√)		(√)
Etoposide					Temozolomide
Everolimus	√*		(√)		Tivozanib
Fluorouracil					Topotecan
Flutamide					Trametinib
Fulvestrant					Vandetanib
Gefitinib					Vemurafenib		√*	√*
Gilteritinib			(√)		Venetoclax
Ibrutinib			(√)		Vinblastine
Idarubicin					Vincristine
Idelalisib	√**				Vinorelbine

Abbreviations: Inh., inhibition; Ind., induction.

Interpretation of classifications (elaborated in Appendix A):

√*: mild inhibition/induction.

√**: moderate inhibition/induction.

√***: strong inhibition/induction.

(√): classification based on in vitro data alone, degree of inhibition/induction not applicable.

Note: Empty cell: No known effect; Based on a combined assessment of the European Summary of Product Characteristics, in vitro and in vivo drug–drug interaction studies and drug databases. For interpretation see below the Table.

The degree of inhibition or induction could be classified according to the methods in [Appendix A] (available in the online version only) for 30 (54%) of the 56 individual drug properties identified; 15 were classified as mild, nine as moderate and six as strong inhibitors or inducers. For the remaining 26 drug properties, no formal in vivo DDI studies allowing for a classification of the degree of inhibition/induction were identified. According to the methods of our analysis, these were all considered as moderate inhibitors/inducers in the primary analysis and as strong inhibitors/inducers in the sensitivity analysis ([Supplementary Table S1], available in the online version only). The degree of P-gp inhibition could be classified for 26% (6/23) of the agents categorized as P-gp inhibitors. The corresponding proportions for P-gp induction, CYP3A4 inhibition, and CYP3A4 induction were 100, 67, and 80%, respectively.

Among the total of 400 “DOAC-antineoplastic agent”-pairs evaluated, a potential for interaction was found in 12% (46 pairs). Specifically, we identified 20 potential and 26 likely DDIs distributed on 31 different antineoplastic agents ([Table 4]). For most of these agents (58%), only one DOAC was susceptible to interact, i.e., the remaining three DOACs were considered as unlikely to interact with the agent. Only one agent, ciclosporin, was classified as likely to interact with all four DOACs. Of note, ciclosporin can be combined with edoxaban in a reduced dosing regimen according to the edoxaban SmPC.[18] In around half (24/46, 52%) of the classifications as likely or potential DDIs, knowledge on the specific degree of P-gp and/or CYP3A4 inhibition/induction of the antineoplastic agent was missing (marked as dotted in [Table 4]).

Table 4 Recommendations on combinations of antineoplastic agents and direct oral anticoagulants

Note: Based on the pharmacokinetic properties of the involved drugs, each drug combination is classified with regard to the assumed potential to cause a clinically relevant drug–drug interaction; unlikely (green – can be combined), potential (yellow – combine with caution) or likely (red – avoid combination). In case of incomplete pharmacokinetic data (i.e., in vitro data alone and/or unknown degree of inhibition/induction), the classification is based on the assumption of moderate inhibition/induction; these classifications are marked as dotted cells and are, thus, less certain than undotted classifications. Antineoplastic agents included in Table 3 but not in this table can be combined with DOACs (all other combinations were classified as unlikely to interact).

^a According to the edoxaban SmPC, the combination of cyclosporine and edoxaban can be used but requires dose reduction of edoxaban to 30 mg once daily.

Dabigatran etexilate was classified as potential or likely to interact with 25% of the 100 included antineoplastic agents; for apixaban this was the case for 1% of the agents. For rivaroxaban and edoxaban the corresponding proportions were 17 and 3%. Accordingly, more than half (n = 25) of the 46 individual DDIs identified involved dabigatran etexilate, and one-third (n = 17) involved rivaroxaban. Similarly, the vast majority (20/26, 77%) of likely DDIs involved dabigatran etexilate. For potential DDIs, rivaroxaban was the most commonly involved DOAC (13/20, 65%).

Sensitivity Analysis

In the sensitivity analysis presented in [Supplementary Table S1] (available in the online version only), inhibition and induction of P-gp and CYP3A4 of unknown degree was considered as strong. As expected, this resulted in a higher number of “DOAC-antineoplastic agent”-pairs classified as potential or likely to interact (68 pairs, corresponding to 17% of all pairs evaluated), and a higher proportion (69%) of uncertain classifications. The analysis identified the same 31 individual antineoplastic drugs as potential or likely to interact with one or more DOACs; six of these (ceritinib, ciclosporine, encorafenib, olaparib, tacrolimus, and temsirolimus) were considered potential or likely to interact with all four DOACs in this analysis. The number of likely DDIs was doubled compared to the primary analysis (53 vs. 26 pairs), whereas the number of potential DDIs was slightly lower (15 vs. 20 pairs). For apixaban and edoxaban, the change in definitions led to a markedly higher proportion of antineoplastic agents considered as potential or likely to interact with the two DOACs (from 1 to 6% for apixaban and from 3% to 20% for edoxaban).

Discussion

This study provides timely and clinically applicable guidance on the potential for pharmacokinetic DDIs during treatment of CAT with DOACs in cancer patients receiving antineoplastic therapy. The evaluation of 400 “DOAC-antineoplastic agent”-pairs revealed that, from a DDI perspective, nearly all the antineoplastic agents evaluated can be appropriately combined with at least one of the DOACs. Apixaban had the lowest potential to interact with antineoplastic agents, whereas dabigatran etexilate had the highest potential.

The major strength of this study is the translational approach. By combining results from basic pharmacokinetic studies with the label recommendations on DOAC DDIs an immediate, clinically applicable classification of combinations of DOACs and antineoplastic agents could be attained. Furthermore, the focus on DOACs as four individual drugs with different potentials to involve in DDIs, rather than a uniform drug group, provides the reader with insight likely widening the application of DOACs in the context of CAT. Other strengths include the transparency of the process and classifications, as well as the use of several data sources to ensure valid information on the pharmacological properties of the included antineoplastic agents. Also, we consider the high number of potential DDIs assessed and the development of a framework for future DDI evaluations as noteworthy elements of our study. Nevertheless, our study had some limitations of importance for the interpretation of the results and recommendations. First and most importantly, the true clinical importance of a potential DDI between an antineoplastic agent and a DOAC can only be evaluated in a controlled clinical study of patients, with end points reflecting the safety and effectiveness of DOAC therapy. In the absence of such evidence, the process leading to the recommendations of this study had to include some assumptions and extrapolations. As an example, we assumed that an effect (i.e., CYP3A4 inhibition) observed in a non-clinical study could be translated into a similar effect during clinical use. Also, we extrapolated the expected change in DOAC AUC inferred by a given antineoplastic agent from knowledge on DOAC AUC changes inferred by drugs with similar pharmacological characteristics as the antineoplastic agent (i.e., strong P-gp inhibition). While the method applied in this study, including the use of AUC changes as surrogates of clinical relevance, is a common and acknowledged approach to the evaluation of potential DDIs,[12] it does not inform us about the true clinical importance. Second, the definition of pharmacological properties associated with unlikely, potential, and likely DDIs, respectively, were derived from each of the DOAC SmPCs. We chose this method, as it ensures consistency between the recommendations given in this paper and the DOAC SmPCs. It may, however, have introduced some inconsistency in the evaluation of the DDI potential across DOACs, due to variations in the thresholds for clinical relevance of possible DDIs between the different DOAC SmPCs. Of note, alternative methods of evaluating the limited available data have been applied by other authors. As an example, the 2021 EHRA Practical Guide on non-vitamin K antagonist oral anticoagulant use in patients with atrial fibrillation[13] reach more restrictive recommendations for several DOAC-antineoplastic pairs compared to the present analysis. The method used to reach the classification in the practical guide is, however, not transparent to the reader, thus challenging further comparison of methodologies. Third, the available evidence to classify an antineoplastic agent with respect to P-gp and CYP3A4 inhibition and induction was often incomplete. For these agents, we had to make assumptions of importance to the evaluation of the clinical relevance, e.g., classifying inhibitors of unknown degree as moderate inhibitors. Such assumptions have likely led to an overestimation of the DDI potential of some antineoplastic agents, especially agents classified as P-gp inhibitors, with some DOACs, especially dabigatran etexilate and edoxaban. Finally, although relevant for the assessment of the potential for DDIs in the individual patient, we did not consider the potential for DDIs with DOACs in the context of combinations of more than one antineoplastic agent, concomitant use of other potentially interacting drugs, or in the setting of comorbidities including decreased renal function. While we still believe that the classifications as unlikely, potential, and likely DDIs may be used to inform clinical decision making in such complex patients, an individualized approach is needed, taking into account the additional before mentioned factors.

Few studies have evaluated the importance of DDIs between antineoplastic agents and DOACs in a clinical setting. The Caravaggio study found comparable safety and efficacy of apixaban and dalteparin in patients with active cancer and VTE.[7] A recently published subgroup analysis of this randomized trial[29] found no modification of the comparative risk of major bleeding or recurrent VTE by concurrent antineoplastic therapy, neither overall nor for agents with a known effect on P-gp and/or CYP3A4. Thus, the analysis supports the notion that antineoplastic therapy, overall, does not compromise the safety and efficacy of apixaban in cancer patients. Similar results were derived from a Taiwanese registry-based study including cancer patients with atrial fibrillation treated with DOACs.[30] The authors compared the risk of major bleeding in patients treated with antineoplastic agents with a known effect on P-gp and/or CYP3A4 with patients not treated with these agents, and found no increased bleeding risk. Collectively, we consider the available evidence consistent with the overall results from our analysis; that is, DDIs are not expected to be of major importance for the clinical safety and effectiveness of DOACs in cancer patients despite concurrent use of antineoplastic therapies. However, knowledge on the relevance of DDIs between antineoplastic agents and DOACs in a clinical setting remains very sparse. Such studies, preferably in cancer patients, are highly warranted. Furthermore, as revealed by this review, the specific effects on P-gp and CYP3A4 remain inadequately investigated for a large proportion of current antineoplastic agents, thus challenging assessment of the agents' potential to involve in DDIs with DOACs. Therefore, we strongly encourage regulators and manufactures, as a minimum, to include in vivo DDI studies using standard probe drugs in the SmPCs of existing and future antineoplastic agents.

Another important result of our study was that for nearly all antineoplastic agents examined, at least one of the DOACs was assessed as unlikely to interact with the agent. Thus, if applying an individualized approach to the treatment of CAT, DOAC therapy will, from a DDI-perspective, be a treatment option in the majority of patients. Importantly, dabigatran etexilate has not been specifically investigated in a setting of CAT, and dabigatran etexilate also has the highest potential for clinically relevant DDIs, thus stressing the importance of thoroughly evaluating the choice of anticoagulant treatment in patients receiving dabigatran etexilate who are diagnosed with cancer.

Other factors than the potential for DDIs are of relevance when deciding between LMWH and DOAC therapy, as well as between the individual DOACs, in CAT patients. These factors include cancer localization,[31] gastrointestinal side effects to antineoplastic therapy potentially compromising the absorption of DOACs, as well as patient preferences. Regarding the latter, qualitative studies have reported that cancer patients prefer the oral over the parental route of administration of anticoagulant therapy.[32] [33] Other aspects of high importance to patients with CAT are the effectiveness of anticoagulant therapy and a low risk of interference with the cancer therapy.[33]

Conclusion

When combining recommendations from DOAC SmPCs on concomitant drug use with knowledge on the pharmacological properties of antineoplastic agents, the frequency of potential and likely “DOAC-antineoplastic agent” DDIs was low. From a DDI perspective, nearly all of the 100 included agents could be combined with at least one DOAC. While awaiting the emergence of clinical evidence in this field, our findings do not support the perception of DDIs as a major obstacle to the use of DOACs in patients with CAT receiving antineoplastic therapy.

Conflict of Interest

M.H. has received speaker honoraria fee from Bayer. J.N.H. has received speaker honoraria fees from Pfizer. M.I.H. has received speaker honoraria or consultancy fees from Bayer, Pfizer, and LEO Pharma and has received support for meeting attendance from Pfizer. E.L.G. has received speaker honoraria or consultancy fees from AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, MSD, Organon, and Lundbeck Pharma. He is an investigator in studies sponsored by AstraZeneca and Bayer and has received unrestricted research grants from Boehringer Ingelheim.

• References

• 1 Jensvoll H, Severinsen MT, Hammerstrøm J. et al. Existing data sources in clinical epidemiology: the Scandinavian Thrombosis and Cancer Cohort. Clin Epidemiol 2015; 7: 401-410
• 2 Kuderer NM, Ortel TL, Francis CW. Impact of venous thromboembolism and anticoagulation on cancer and cancer survival. J Clin Oncol 2009; 27 (29) 4902-4911
• 3 Kourlaba G, Relakis J, Mylonas C. et al. The humanistic and economic burden of venous thromboembolism in cancer patients: a systematic review. Blood Coagul Fibrinolysis 2015; 26 (01) 13-31
• 4 Prandoni P, Lensing AWA, Piccioli A. et al. Recurrent venous thromboembolism and bleeding complications during anticoagulant treatment in patients with cancer and venous thrombosis. Blood 2002; 100 (10) 3484-3488
• 5 Raskob GE, van Es N, Verhamme P. et al; Hokusai VTE Cancer Investigators. Edoxaban for the treatment of cancer-associated venous thromboembolism. N Engl J Med 2018; 378 (07) 615-624
• 6 Young AM, Marshall A, Thirlwall J. et al. Comparison of an oral factor Xa inhibitor with low molecular weight heparin in patients with cancer with venous thromboembolism: results of a randomized trial (SELECT-D). J Clin Oncol 2018; 36 (20) 2017-2023
• 7 Agnelli G, Becattini C, Meyer G. et al; Caravaggio Investigators. Apixaban for the treatment of venous thromboembolism associated with cancer. N Engl J Med 2020; 382 (17) 1599-1607
• 8 Khorana AA, Noble S, Lee AYY. et al. Role of direct oral anticoagulants in the treatment of cancer-associated venous thromboembolism: guidance from the SSC of the ISTH. J Thromb Haemost 2018; 16 (09) 1891-1894
• 9 Farge D, Frere C, Connors JM. et al; International Initiative on Thrombosis and Cancer (ITAC) advisory panel. 2019 International clinical practice guidelines for the treatment and prophylaxis of venous thromboembolism in patients with cancer. Lancet Oncol 2019; 20 (10) e566-e581
• 10 Key NS, Khorana AA, Kuderer NM. et al. Venous thromboembolism prophylaxis and treatment in patients with cancer: ASCO clinical practice guideline update. J Clin Oncol 2020; 38 (05) 496-520
• 11 Sabatino J, De Rosa S, Polimeni A, Sorrentino S, Indolfi C. Direct oral anticoagulants in patients with active cancer: a systematic review and meta-analysis. JACC CardioOncol 2020; 2 (03) 428-440
• 12 European Medicines Agency. Guideline on the investigation of drug interactions. Published online 2012. Accessed February 01, 2021, at: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-investigation-drug-interactions-revision-1_en.pdf
• 13 Steffel J, Collins R, Antz M. et al; 2021 European Heart Rhythm Association Practical Guide on the use of non-vitamin k antagonist oral anticoagulants in patients with atrial fibrillation. Europace 2021; 23 (10) 1612-1676
• 14 Ferri N, Bellosta S, Baldessin L, Boccia D, Racagni G, Corsini A. Pharmacokinetics interactions of monoclonal antibodies. Pharmacol Res 2016; 111: 592-599
• 15 European Medicines Agency. Eliquis: Summary of Product Characteristics (updated September 2022). Published online 2022. Accessed 15 September 2022 at: https://www.ema.europa.eu/en/medicines/human/EPAR/eliquis#product-information-section
• 16 European Medicines Agency. Xarelto: Summary of Product Characteristics (updated December 2021). Published online 2021. Accessed 15 September 2022 at: https://www.ema.europa.eu/en/medicines/human/EPAR/xarelto
• 17 European Medicines Agency. Pradaxa: Summary of Product Characteristics (updated July 2022). Published online 2022. Accessed 15 September 2022 at: https://www.ema.europa.eu/en/medicines/human/EPAR/pradaxa#product-information-section
• 18 European Medicines Agency. Lixiana: Summary of Product Characteristics (updated April 2021). Published online 2021. Accessed 15 September 2022 at: https://www.ema.europa.eu/en/medicines/human/EPAR/lixiana#product-information-section
• 19 Hodin S, Basset T, Jacqueroux E. et al. In vitro comparison of the role of P-glycoprotein and breast cancer resistance protein on direct oral anticoagulants disposition. Eur J Drug Metab Pharmacokinet 2018; 43 (02) 183-191
• 20 Foerster KI, Hermann S, Mikus G, Haefeli WE. Drug-drug interactions with direct oral anticoagulants. Clin Pharmacokinet 2020; 59 (08) 967-980
• 21 European Medicines Agency. European Medicines Agency. Accessed February 2, 2021, at: https://www.ema.europa.eu/en
• 22 U.S. Food and Drug Administration. . FDA. Published August 23, 2021. Accessed August 24, 2021, at: https://www.fda.gov/home
• 23 Danish Medicines Agency. . Produktresuméer. Accessed August 24, 2021, at: http://www.produktresume.dk/AppBuilder/search
• 24 Micromedex IBM. . (R). Accessed February 2, 2021, at: https://www.micromedexsolutions.com/micromedex2/librarian/PFDefaultActionId/evidencexpert.DoIntegratedSearch/ssl/true?navitem=headerLogout#
• 25 Lund M, Petersen TS, Dalhoff KP. Clinical implications of p-glycoprotein modulation in drug-drug interactions. Drugs 2017; 77 (08) 859-883
• 26 Danish Medicines Agency. . Interaktionsdatabasen.dk. Accessed February 2, 2021, at: http://www.interaktionsdatabasen.dk/
• 27 U.S. Food and Drug Adminstration. Clinical Drug Interaction Studies - Cytochrome P450 Enzyme- and Transporter-Mediated Drug Interactions. Guidance for Industry. Published online 2020. Accessed 15 September 2022 at: https://www.fda.gov/media/134581/download
• 28 Wind S, Schnell D, Ebner T, Freiwald M, Stopfer P. Clinical pharmacokinetics and pharmacodynamics of afatinib. Clin Pharmacokinet 2017; 56 (03) 235-250
• 29 Verso M, Munoz A, Bauersachs R. et al. Effects of concomitant administration of anticancer agents and apixaban or dalteparin on recurrence and bleeding in patients with cancer-associated venous thromboembolism. Eur J Cancer 2021; 148: 371-381
• 30 Wang CL, Wu VC, Tu HT. et al. Risk of major bleeding associated with concomitant use of anticancer drugs and direct oral anticoagulant in patients with cancer and atrial fibrillation. J Thromb Thrombolysis 2022; 53 (03) 633-645
• 31 Bikdeli B, Jiménez D. Safety of apixaban for cancer-associated thrombosis. Thromb Haemost 2021; 121 (05) 547-551
• 32 Picker N, Lee AY, Cohen AT. et al. Anticoagulation treatment in cancer-associated venous thromboembolism: assessment of patient preferences using a discrete choice experiment (COSIMO study). Thromb Haemost 2021; 121 (02) 206-215
• 33 Noble S, Matzdorff A, Maraveyas A, Holm MV, Pisa G. Assessing patients' anticoagulation preferences for the treatment of cancer-associated thrombosis using conjoint methodology. Haematologica 2015; 100 (11) 1486-1492

Publikationsverlauf

Artikel online veröffentlicht:
02. Februar 2023