Drug-drug interactions between direct oral anticoagulants and anticonvulsants and clinical outcomes: A systematic review

Background Direct oral anticoagulants (DOACs) are widely used in patients with atrial fibrillation and venous thromboembolism. However, DOACs have important potential drug-drug interactions (DDIs) with several classes of drugs. In particular, antiepileptic (AE) drugs may induce cytochrome P450 3A4 or P-glycoprotein. Co-administration of DOACs and AE drugs may result in lower DOAC drug levels and reduced DOAC efficacy. However, the clinical significance of such DDIs is uncertain. Objectives The aim of this systematic review was to generate an updated review of these DDIs and their clinical relevance, given the rapidly evolving knowledge relating to DOAC and AE DDIs. Methods We searched the MEDLINE and Embase databases for studies reporting clinical adverse outcomes (thrombotic events, bleeding events, and all-cause mortality) in patients concomitantly taking DOACs and AE drugs. Results We retrieved 874 studies of which 15 were deemed eligible for this review, including 4 congress abstracts, 3 case reports, 2 letters to the editor, 5 retrospective cohorts, and 1 prospective cohort study. No randomized clinical trials were found. Most of the included studies reported thrombotic events, 3 studies reported major bleeding, and one study reported all-cause mortality associated with DOAC and AE drug administration. Substantial differences in the study designs did not allow for a meta-analysis to be performed. Conclusion The current literature assessing these adverse clinical outcomes from DOAC and AE drug co-administration is limited. Although the available data point to a possible increased risk of thrombotic events, they are insufficient to draw definitive conclusions. Well-designed clinical studies are of utmost importance.

DOACs and other oral anticoagulants with AE drugs highlighted a marked increase in concomitant use between 2011 and 2018 [7].
These findings highlight the need to clarify the existence and clinical significance of DDIs between DOACs and AE drugs.
Taha et al. [8] previously performed a systematic review assessing DOAC and AE DDIs, concluding that enzyme-inducing AE drugs reduced the effectiveness of DOACs, although the overall quality of the records was poor and data relating to edoxaban were lacking. The omission of edoxaban-related data is clinically important because this DOAC is increasing in use and has minimal dependence on CYP3A4 for metabolism and therefore may be a safer DOAC when combined with AE drugs that affect this metabolic pathway [1].
Close monitoring of patients who are concomitantly taking DOAC and AE drugs is recommended; however, there still is no consensus on whether co-prescription of DOACs and AE drugs in clinical practice is appropriate or should be avoided [1]. Given the rapidly evolving knowledge relating to DOAC DDIs, the aim of this study was to provide an expanded and updated systematic review and summarize possible clinical implications of DOAC and AE drug co-administration.

| M E T H O D S
The MEDLINE and Embase databases were systematically searched by title, abstract, and full-text for studies focusing on DOAC and AE DDIs and related clinical adverse events (thrombosis, major bleeding event, or mortality) from March 2019 until December 2022 to build on a previous meta-analysis [8]. Case reports, case series, observational studies, and randomized clinical trials were included. The quality of the included cohort studies was evaluated using the Newcastle-Ottawa scale for observational studies [9]. Reviews, meta-analyses, and articles in a language other than English were excluded. Additionally, records with insufficient information regarding the type of DOAC and/or the type of AE administered, as well as those studies reporting only anti-Xa serum concentration were excluded. Moreover, articles that included patients affected by COVID-19 were excluded to avoid additional confounding.

Essentials
• Direct oral anticoagulants (DOACs) and antiepileptic drugs are often co-administered.
• Co-administration of DOACs and antiepileptic drugs may result in lower DOAC efficacy.
• These drug-drug interactions (DDIs) may result in adverse clinical events such as thrombosis.
• Current data are insufficient to draw conclusions about these DDIs and adverse clinical events.
The search was independently performed by 2 authors (M.C. and S.C.), and conflicts were resolved through discussion and consensus.

| R E S U L T S
The literature search strategy is described in Supplementary Table 1, and the article screening process is shown in the Figure. The interrater percentage of agreement was high (95% for the title and abstract screening and 92% for the full-text screening).
The quality was found to be low or moderate across the cohort studies included in the review (Supplementary Table 2). The "selection of the nonexposed cohorts" was determined to be at a high risk of bias in 83% of the cohort studies included (Supplementary Table 2).
No information on patients' race/ethnicity was reported in the studies included in the review. The results for each DOAC are summarized below.

| Dabigatran
In a case series by Barbar et al. [10], among 14 patients taking a DOAC, 3 were on treatment with dabigatran and a concomitant AE drug. None experienced an adverse outcome (cerebrovascular event and/or venous thrombosis) during an observation period of 10 months (range, 1-27 months) [10]. However, neither the dabigatran dose nor the concomitant AE drug or the dose was specified. Moreover, although 9 of the 14 patients were taking a DOAC for VTE, the DOAC's indication for the remainder was not specified [10]. In a case report from Paskaleva et al. [11], a 30-year-old man with a new deep vein thrombosis (DVT) in the context of heterozygous prothrombin gene mutation who was receiving valproic acid (1000 mg daily) was switched to dabigatran (150 mg twice daily) one month after the diagnosis of DVT after a lack of clinical improvement and low levels on rivaroxaban. There was a persistent lack of clinical improvement, and owing to lower-than-expected dabigatran levels, anticoagulant therapy was changed to a vitamin K antagonist that eventually led to clinical improvement [11]. In a congress abstract, Shah et al. [12] reported the case of a 25-year-old man on treatment with phenytoin for recurrent seizures. The patient developed cerebral venous thrombosis, and therapy with dabigatran was started. Neither the dabigatran nor the phenytoin dose was reported. After an unspecified time, the patient experienced worsening cerebral venous thrombosis [12].

| Rivaroxaban
In the study by Perlman et al. [22], the authors found an increased reporting rate of thromboembolic and ischemic events in patients cotreated with rivaroxaban (no doses were specified); one of phenytoin,  [17]. In another letter to the editor, a case of an 80- year-old woman with AF on treatment with rivaroxaban 20 mg and primidone who developed an ischemic stroke was reported [18]. In an observational study, the co-prescription of rivaroxaban and phenytoin was associated with at least one thrombotic event among 4924 patients (RR 2.39; 95% CI, 1.33-3.29) [19]. In Giustozzi et al. [23], among 25 patients with AF taking rivaroxaban, 8 patients were also taking levetiracetam, 5 were on valproic acid, and 2 were on carbamazepine.
In this study, 4 patients experienced a stroke-one patient was receiving rivaroxaban 15 mg with phenobarbital 100 mg, the second rivaroxaban 15 mg with phenobarbital 100 mg and levetiracetam 250 mg twice daily, the third rivaroxaban 15 mg and levetiracetam 500 mg twice daily, and the last rivaroxaban 20 mg and carbamazepine 400 mg twice daily. Two of these strokes were fatal [23]. In the study by

| Apixaban
In a case report from the study by Di Gennaro et al. [16], an 88-yearold man on treatment with carbamazepine (dose not reported) for chronic essential trigeminal neuralgia was started on apixaban 5 mg twice daily for nonvalvular AF. After one month, the patient developed a transient ischemic attack (TIA) [16].
Perlman et al. [22] found a reporting OR of 1.88 (95% CI, 1.41-2.48) of thromboembolic or ischemic events for patients co-treated with apixaban and one of carbamazepine, phenytoin, or phenobarbital compared with patients co-treated with apixaban and one of valproic acid, lamotrigine, or levetiracetam.
In the case series by Barbar [23].
In our study, among 18 patients taking apixaban (of whom 10 patients were on carbamazepine and 8 were on phenytoin), one experienced a stroke (apixaban 5 mg twice daily + carbamazepine 600 mg once daily) and another had an intracranial bleed (apixaban 5 mg twice daily + phenytoin 750 mg once daily). The outcome of all-cause mortality was assessed in our study. There was only one death in a patient receiving apixaban 5 mg twice daily in combination with phenytoin, and the death was secondary to sepsis [20].
Finally, in a case report from Sulagna et al. [13], a 60-year-old man with a history of DVT and pulmonary embolism post-inferior vena cava filter on treatment with apixaban 5 mg twice daily and phenytoin (dose not reported) experienced a recurrent DVT.

| Edoxaban
In the case report from the study by Di Gennaro et al. [16], after experiencing a TIA while taking apixaban in combination with carbamazepine, the patient was switched to edoxaban 60 mg once daily.
After a follow-up period of 18 months after edoxaban initiation, no further thrombotic events were reported [16].
In the study by Barbar  with DOACs, limiting the ability to attribute any adverse clinical outcomes to the DOAC AE DDI specifically and must be accounted for [25]. Dexamethasone, for example, is commonly co-prescribed in patients with brain tumors receiving DOACs and AEs and could amplify the DOAC AE DDI because it also induces P-gp and CYP3A4 [1,26].
In conclusion, there are several potential DDIs between DOACs and AE drugs, but current evidence is limited and statistically underpowered to draw definitive conclusions regarding clinical significance.
Well-designed clinical studies are of utmost importance.

FUNDING
No funding was obtained for this study.