Fitting the right non-vitamin K antagonist oral anticoagulant to the right patient with non-valvular atrial fibrillation: an evidence-based choice

Abstract

Atrial fibrillation (AF) is the most prevalent arrhythmia and is associated with an increased risk of ischemic stroke (IS) and systemic embolism (SE). Stroke prevention is a key element for the overall management of AF patients. The non-vitamin K antagonist oral anticoagulants (NOACs), such as dabigatran, rivaroxaban, apixaban and edoxaban, are at least as effective as warfarin in reducing IS/SE with a lower rate of major bleeding. Various analyses from the large Phase III randomized trials demonstrated different efficacy and safety of NOACs in specific subgroups of patients. The randomized trials are supplemented by effectiveness and safety data from real-world observational cohorts following the availability of these drugs for use in everyday clinical practice. Given the clinical heterogeneity of AF patients, the available data from trials and real-world studies allow us to fit the right NOAC to the particular patient’s characteristics, with the aim of optimizing outcomes for the individual patient. This review article aims to provide a summary of the evidence on the performance of NOACs in AF patients with specific clinical characteristics. Evidence-based suggestions are presented to provide a simple and viable strategy for clinicians for the choice of a particular NOAC.

KEY MESSAGE

• Given the different performance of the new-oral anticoagulants in patients with the different clinical situation, evidence-based choice of fitting the right new-oral anticoagulants to the patients is provided in this review article.

Keywords:

• Atrial fibrillation
• stroke prevention
• new oral anticoagulant
• non-vitamin K oral anticoagulants
• clinical situations

View correction statement:

Correction to: Li et al., Fitting the right non-vitamin K antagonist oral anticoagulant to the right patient with non-valvular atrial fibrillation: an evidence-based choice

Introduction

Atrial fibrillation (AF) is the most commonly encountered arrhythmia worldwide, increasing the risk of ischemic stroke (IS) and systemic embolism (SE) up to 5-fold [1]. Approximately 20–30% of ischaemic strokes are caused by AF, which are associated with more severe disability, higher morbidity and mortality [2,3]. Therefore, stroke prevention is a key priority in the management of patients with AF [4,5].

Vitamin K antagonists (VKAs) are remarkably effective in AF-related stroke prevention, with a reduction in IS/SE of 64%, and a 26% reduction in all-cause mortality, compared to placebo or control [6,7]. However, the disadvantages of warfarin, such as the need for frequent laboratory monitoring and a narrow therapeutic range, have encouraged the development of the non-vitamin K antagonist oral anticoagulants (NOACs). These include an inhibitor of factor II (thrombin) namely, dabigatran, and three-factor Xa inhibitors, i.e. edoxaban, apixaban and rivaroxaban. All NOACs have some favourable common features such as rapid onset of action and less need for laboratory monitoring.

Various analyses from the large phase III randomised trials demonstrated different efficacy and safety of NOACs in specific subgroups of patients. The randomised trials are supplemented by effectiveness and safety data from real-world observational cohorts following the availability of these drugs for use in everyday clinical practice [8–13]. In addition, patients on NOACs may have better compliance than those on warfarin, which is important for clinical outcome [14–17]. Findings from the GLORIA-AF (Global Registry on Long-Term Oral Antithrombotic Treatment in Patients with Atrial Fibrillation) Registry (n = 15,641) demonstrated that NOACs have been highly adopted into clinical practice, with their use becoming even more frequent than VKAs (47.6% versus 32.3%) in Europe and North America [18].

Given the increasing number of AF patients started on NOACs, prescription of the right drug according to patient’s characteristics and preference is a challenge for clinicians. With the clinical heterogeneity of AF patients, the available data from trials and real-world studies allow us to fit the right NOAC to the particular patient’s characteristics, with the aim of optimizing outcomes for the individual patient.

This review article aims to provide a summary of the evidence on the performance of NOACs in AF patients with specific clinical characteristics. Based on this, evidence-based suggestions are presented to provide a simple and viable strategy for clinicians for the choice of a particular NOAC.

Patients with previous IS/transient ischemic attack (TIA)

Patients with AF who already experienced a cerebrovascular accident, such as an IS or transient ischemic attack (TIA), are a very high-risk category for recurrent cerebrovascular events [19,20]. A recent ancillary analysis from the ENGAGE AF-TIMI 48 (Effective Anticoagulation with factor Xa Next Generation in Atrial Fibrillation-Thrombolysis In Myocardial Infarction 48) trial (n = 5973), showed that patients with previous IS/TIA were at higher risk of experiencing a new thromboembolic events compared to patients without IS/SE (2.83% versus 1.42% per year, respectively, p < .001) [20]. In the Apixaban Versus Acetylsalicylic Acid to Prevent Stroke in Atrial Fibrillation Patients who have Failed or are Unsuitable for Vitamin K Antagonist Treatment (AVERROES) trial, previous IS/TIA was associated with significantly increased risk of another IS/SE (HR: 2.38; 95%CI: 1.66–3.34; p < .0001) [21]. Hence, secondary prevention is pivotal for these high-risk AF patients.

The efficacy and safety of NOACs in this subgroup of high-risk AF patients has been analysed in several post-hoc analyses from phase III clinical trials. In the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial, 3623 patients had previous IS/TIA. Of these, 1195 patients were on dabigatran 110 mg twice daily, 1233 on dabigatran 150 mg twice daily and 1195 on dose-adjusted warfarin. An IS/SE reoccurred in 2.78% of patients on warfarin compared with 2.32% of patients on dabigatran 110 mg and 2.07% on dabigatran 150 mg [22]. In this subgroup of patients, the rate of major bleeding was lower in patients on dabigatran 110 mg, and similar in those receiving dabigatran 150 mg compared with warfarin (Table 1) [22]. Moreover, both 110 and 150 mg dabigatran significantly reduced intracranial haemorrhage (ICH) compared to warfarin (Table 1). Therefore, the efficacy and safety of both doses of dabigatran were comparable to warfarin in patients with previous IS/TIA, along with a better safety profile.

Table 1. Evidence from randomized clinical trials for efficacy and safety of NOACs in specific subgroups of patients with AF.

Subgroup of patients	NOACs	Dosage	Study	N	HR (95%CI) IS/SE	HR (95%CI) Major bleeding	HR (95%CI) ICH	Comments
Previous IS/TIA	Dabigatran	150 mg BID	RE-LY [22]	2482	0.75 (0.52–1.08)	1.01 (0.77–1.34)	0.41 (0.21–0.79)	Both doses of dabigatran were non-inferior to warfarin for IS/SE and better in reducing ICH.
		110 mg BID		2399	0.84 (0.58–1.20)	0.66 (0.48–0.90)	0.20 (0.08–0.47)
	Rivaroxaban	20 mg OD	ROCKET AF [23]	7468	0.94 (0.77–1.16)	0.97 (0.79–1.19)	0.74 (0.47–1.15)	Rivaroxaban was similar to warfarin in efficacy and safety
		15 mg OD	J-ROCKET AF [24]	813	0.50 (0.23–1.14)	0.64 (0.32–1.28)	–
	Apixaban	5 mg BID	ARISTOTLE [25]	3436	0.76 (0.56–1.03)	0.73 (0.55–0.98)	0.37 (0.21–0.67)	Apixaban had lower rates of major bleeding and ICH
	Edoxaban	60 mg OD	ENGAGE AF-TIMI 48 [20]	5973	0.86 (0.67–1.09)	0.84 (0.67–1.06)	0.57 (0.36–0.92)	HDER had lower rate of ICH compared with warfarin
Elderly (≥75 years)	Dabigatran	150 mg BID	RE-LY [26]	7258	0.67 (0.49–0.90)	1.18 (0.98–1.42)	0.42 (0.25–0.70)	Both doses reduced risk of ICH. Dabigatran 150 mg reduced IS/SE.
		110 mg BID		7258	0.88 (0.66–1.17)	1.01 (0.83–1.23)	0.37 (0.21–0.64)
	Rivaroxaban	20 mg OD	ROCKET AF [27]	6229	0.80 (0.63–1.02)	1.11 (0.92–1.34)	0.80 (0.50–1.28)	Rivaroxaban and warfarin had similar safety and efficacy
	Apixaban	5 mg BID	AVERROES [28]	1898	0.33 (0.18–0.56)	1.21 (0.69–2.12)	–	Apixaban more effective than aspirin
	Edoxaban	60 mg OD	ENGAGE AF-TIMI 48 [29]	8474	0.83 (0.66–1.04)	0.83 (0.70–0.99)	0.40 (0.26–0.62)	Edoxaban is safer than warfarin
Well controlled warfarin	Dabigatran	150 mg BID	RE-LY [30]	3023	0.95 (0.61–1.48)	1.16 (0.88–1.54)	0.39 (0.18–0.84)	Dabigatran reduced ICH compared to TiTR >72.6%
		110 mg BID		2991	0.92 (0.59–1.45)	0.90 (0.67–1.21)	0.27 (0.11–0.66)
	Rivaroxaban	20 mg OD	ROCKET AF [31]	3502	0.87 (0.58–1.32)	1.25 (1.10–1.41)	–	Increased risk of bleeding compared to TiTR ≥65.7%
	Apixaban	5 mg BID	ARISTOTLE [32]	4552	0.88 (0.57–1.35)	0.75 (0.58–0.97)	–	Lower major bleeding compared to TiTR ≥71.2%
Chronic kidney disease	Dabigatran	150 mg BID	RE-LY [33]	50–80 mL/min (n = 2803)	0.68 (0.50–0.92)	0.91 (0.75–1.11)	0.45 (0.27–0.74)	Both doses of dabigatran reduced ICH. Dabigatran 150 mg bid reduced IS/SE
				<50 mL/min (n = 1196)	0.56 (0.37–0.85)	1.01 (0.79–1.30)	0.31 (0.14–0.66)
		110 mg BID		50–80 mL/min (n = 2803)	0.93 (0.70–1.23)	0.76 (0.62–0.94)	0.29 (0.16–0.52)
				<50 mL/min (n = 1196)	0.85 (0.59–1.24)	0.99 (0.77–1.28)	0.40 (0.20–0.80)
	Rivaroxaban	10 mg OD	J-ROCKET AF [34]	30–49 mL/min (n = 141)	0.82 (0.25–2.69)	0.89 (0.36–2.18)	–	Similar efficacy and safety of warfarin
		15 mg OD	ROCKET AF [35]	30–49 mL/min (n = 1474)	0.84 (0.57–1.23)	0.95 (0.72–1.26)	0.81 (0.41–1.60)
	Apixaban	Variable	ARISTOTLE [36]	>50–80 mL/min (n = 5272)	0.74 (0.56–0.97)	0.77 (0.62–0.94)	–	Apixaban reduces major bleeding and IS/SE in patients with mild CKD
				≤50 mL/min (n = 2067)	0.79 (0.55–1.14)	0.50 (0.38–0.66)	–
		Variable	AVERROES [37]	<60 mL/min (n = 1697)	0.32 (0.18–0.55)	1.2 (0.65–2.1)	–	Apixaban reduced IS/SE in stage III CKD patients
	Edoxaban	HDER	ENGAGE AF-TIMI 48(56)	>50–95 mL/min (n = 4047)	0.91 (0.72–1.14)	0.89 (0.75–1.04)	0.43 (0.29–0.63)	HDER reduced ICH in both mild and moderate CKD
				30–50 mL/min (n = 1372)	0.93 (0.67–1.30)	0.76 (0.58–0.98)	0.46 (0.26–0.82)
Stable CAD	Rivaroxaban	20/15 mg OD	ROCKET AF [38]	1182	0.61 (0.37–0.99)	1.32 (0.96–1.81)	0. 56 (0.21–1.50)	Rivaroxaban reduced IS/SE
	Apixaban	5 mg BID	ARISTOTLE [39]	6639	0.95 (0.71–1.27)	0.78 (0.62–0.99)	0.36 (0.20–0.66)	Apixaban reduced major bleeding and ICH
Concomitant antiplatelet	Dabigatran	150 mg BID	RE-LY [40]	4630	0.80 (0.59–1.08)	0.93 (0.76–1.12)	0.47 (0.28–0.80)	Dabigatran had similar efficacy but better safety than warfarin
		110 mg BID		4648	0.93 (0.70–1.25)	0.82 (0.67–1.00)	0.23 (0.12–0.47)
	Edoxaban	HDER	ENGAGE AF-TIMI 48 [41]	4912	0.70 (0.50–0.98)	0.82 (0.65–1.04)	0.46 (0.27–0.79)	HDER reduced IS/SE and ICH
		LDER			1.03 (0.76–1.39)	0.51 (0.39–0.66)	0.18 (0.08–0.37)	LDER similar efficacy and reduced major bleeding and ICH
ACS	Dabigatran	150 mg BID + DAPT	RE-DUAL PCI [42]	2725	1.09 (0.42–2.83)	0.64 (0.43–0.94)	0.12 (0.20–0.98)	Dabigatran + SAPT had similar efficacy and better safety than warfarin
		110 mg BID + DAPT		2725	1.30 (0.63–2.67)	0.52 (0.37–0.74)	0.30 (0.08–1.07)
	Rivaroxaban	15 mg OD + SAPT	PIONEER-AF [43]	1415	1.07 (0.39–2.96)	0.66 (0.33–1.31)	–	Rivaroxaban + SAPT/DAPT was as effective and safe as warfarin-based triple therapy
		2.5 mg BID + DAPT	PIONEER-AF [43]	1415	1.36 (0.52–3.58)	0.57 (0.28–1.16)	–
Cardioversion	Dabigatran	150 mg BID	RE-LY [44]	1336	0.49 (0.09–2.69)	0.99 (0.25–3.93)	–	Similar efficacy and safety of both doses of dabigatran
		110 mg BID		1311	1.28 (0.35–4.76)	2.82 (0.90–8.82)	–
	Rivaroxaban	20 mg OD	ROCKET AF [45]	321	1.38 (0.61–3.11)	1.51 (1.12–2.05)	–	Similar efficacy of Rivaroxaban with an increased risk of major bleeding
		20 mg OD	X-VeRT [46]	1504	0.50 (0.15–1.73)	0.76 (0.21–2.67)	–
	Edoxaban	60 mg OD	ENSURE-AF [47]	2199	0.67 (0.06–5.88)	0.61 (0.09–3.13)	–	Similar efficacy and safety of warfarin
Asia	Dabigatran	150 mg BID	RE-LY [48]	1859	0.45 (0.28–0.72)	0.57 (0.38–0.84)	0.40 (0.18–0.92)	Both doses of dabigatran reduced risk of major and ICH. Dabigatran 150 mg reduced IS/SE
		110 mg BID		1859	0.81 (0.54–1.21)	0.57 (0.39–0.85)	0.20 (0.07–0.60)
	Rivaroxaban	15 mg OD	J-ROCKET AF [12]	1280	0.49 (0.24–1.00)	0.85 (0.50–1.43)	0.73 (0.16–3.25)	Rivaroxaban reduced IS/SE with similar safety
		20 mg OD	ROCKET AF [49]	932	0.78 (0.44–1.39)	0.24 (0.08–0.71)	0.40 (0.13–1.27)	Rivaroxaban 20 mg reduced major bleeding
	Apixaban	5 mg BID	ARISTOTLE [50]	1993	0.74 (0.50–1.10)	0.53 (0.35–0.80)	0.36 (0.18–0.71)	Apixaban is safer than warfarin in Asian patients
	Edoxaban	60 mg OD	ENGAGE AF-TIMI 48 [51]	1290	0.53 (0.31–0.90)	0.61 (0.41–0.89)	0.31 (0.15–0.66)	Both doses of edoxaban reduced risk of major and ICH. Edoxaban 60 mg reduced IS/SE
	Edoxaban	30 mg OD		1297	0.98 (0.63–1.54)	0.34 (0.21–0.54)	0.24 (0.11–0.56)

BID: twice daily; CAD: coronary artery disease; CI: confidence interval; DAPT: dual antiplatelet therapy; HDER: high-dose edoxaban regimen; HR: hazard ratio; ICH: intracranial haemorrhage; IS: ischemic stroke; LDER: low-dose edoxaban regimen; MI: myocardial infarction; NOAC: non-vitamin K antagonists oral anticoagulants; OD: once a day; PCI: percutaneous coronary intervention; RCT: randomized controlled trial; SAPT: single antiplatelet therapy; SE: systemic embolism; TIA: transient ischemic attack; TiTR: time in therapeutic range. Significant differences in the relative effects between a pair of drugs are given in bold.

In a subgroup analysis of the Rivaroxaban versus Warfarin in Non-valvular Atrial Fibrillation (ROCKET-AF) trial, 7468 patients with AF and previous IS/TIA were randomized into rivaroxaban 20 mg group (n = 3754) or dose-adjusted warfarin group (n = 3714). Rivaroxaban had a similar risk of IS/SE, major bleeding and ICH compared with warfarin (Table 1) [23].

A subgroup analysis of the Rivaroxaban versus Warfarin in Japanese Patients With Atrial Fibrillation (J-ROCKET AF) trial (n = 1274) that was conducted in Japan investigated the efficacy and safety of rivaroxaban in the secondary prevention of IS/SE [24]. The rate of the IS/SE was 1.66%/year in rivaroxaban-treated and 3.25%/year in warfarin-treated patients, while the rate of primary safety endpoints (composite of major bleeding and non-major clinically relevant bleeding) in this subgroup was 17.02%/year in the rivaroxaban group and 18.26%/year in warfarin-treated patients (Table 1) [24].

Data from the Apixaban versus Warfarin in Patients with Atrial Fibrillation (ARISTOTLE) trial on patients with previous IS/SE, reported a risk of a new IS/SE of 2.46 per 100 patient-years in the apixaban group and 3.24 per 100 patient-years in the warfarin group. Furthermore, the absolute reduction of IS/SE was 0.77 per 100 patient-years for apixaban compared with warfarin [25]. The relative efficacy was similar in the secondary prevention cohort, and the risks of major bleeding and ICH were also lower with apixaban compared to warfarin (Table 1) [25]. In the AVERROES trial, among 5599 AF patients 764 had previous IS/TIA [21]. Apixaban reduced the annual risk of IS/SE by 73% compared with aspirin in patients with previous IS/TIA [21]. The rate of major bleeding and ICH was also comparable between apixaban and aspirin [21].

In the ENGAGE AF-TIMI 48 trial, in a 2.8-year follow-up, in patients with previous IS/TIA (n = 5973), edoxaban was associated with lower rate of ICH compared with warfarin (Table 1). In addition, edoxaban reduced the risk of fatal bleeding by 60%, any bleeding by 16%, and cardiovascular death by 21% [20]. The combined endpoint including all-cause death/IS/SE rate was significantly reduced by the high dose edoxaban regimen (HDER) treatment (HR, 0.83; 95%CI, 0.71–0.96; p = .015) [20].

A meta-analysis of the RE-LY, ROCKET-AF and ARISTOTLE trials including 14,527 patients with AF and prior IS/TIA demonstrated that NOACs use decreased the risk of IS/SE (Odd Ratio [OR], 0.85; 95%CI, 0.74–0.99; p = .03) [52] with a similar risk of major bleeding (OR, 0.84; 95%CI, 0.69–1.03; p = .10) but reduced risk of ICH (OR, 0.42; 95%CI, 0.25–0.70; p < .0009) compared with warfarin [52]. Another recent meta-analysis including also the ENGAGE AF-TIMI 48 trial demonstrated that NOACs reduced the risks of IS/SE (Relative Risk [RR], 0.85; 95%CI, 0.74–0.97; p value not given) and ICH (RR, 0.42; 95%CI, 0.27–0.66; p value not given). The risk of major bleeding was comparable between NOACs and warfarin (RR, 0.85; 95%CI, 0.71–1.03; p value not given) [53].

Few studies have addressed the effect of early restarting or initiation of anticoagulation among patients presenting with acute IS or TIA [54]. Stroke <14 days was an exclusion criteria for most of the randomized trials, except for ARISTOTLE where the cutoff time was stroke <7 days [10]. One of the larger studies was the Trends in oral anticoagulation choice for acute stroke patients with non-valvular atrial fibrillation (SAMURAI-NVAF) study, where anticoagulation restarting was reported among 1192 patients with acute IS/TIA [55]. In this study, warfarin was used in 650 patients, dabigatran for 203, rivaroxaban for 238 and apixaban for 25 subjects [55]. NOAC were started at a median of 4 days after stroke. The risk of stroke was milder and ischemic-risk indices were lower in patients taking NOACs than in those taking warfarin [55].

Elderly patients

Ageing is a risk factor for IS/SE and bleeding in patients with AF [27]. Thus, attention should be paid to elderly patients, in whom the balance between bleeding and thrombosis could be challenging.

In the RE-LY trial, over 7200 elderly patients (≥75 years) were enrolled. Dabigatran 110 mg bid was associated with a lower risk of major bleeding in patients aged <75 years (1.89% versus 3.04%, p < .001) and a similar risk in those aged ≥75 years (4.43% versus 4.37%, p = .89) compared with warfarin (Table 1) [26]. Conversely, dabigatran 150 mg bid was associated with a lower risk of major bleeding in those aged <75 years (2.12% versus 3.04%, p < .001), whereas a non-significant trend toward higher risk of major bleeding was observed in patients aged ≥75 years (5.10% versus 4.37%, p = .07). Of note, the better safety profile of dabigatran in relation to ICH was evident with both doses of Dabigatran (Table 1) [26]. These results were confirmed by a “real-world” study including 47,192 AF patients, which compared dabigatran (150 and 110 mg) with matched warfarin users to investigate the incidence of IS and bleeding events [56]. In elderly patients (≥75 years), dabigatran was associated with lower rates of ICH (HR, 0.60; 95%CI, 0.47–0.76; p value not given) [56].

In a subgroup analysis of the ROCKET-AF trial, 6229 elderly AF patients treated with rivaroxaban or warfarin were compared, and the efficacy and safety of rivaroxaban in elderly patients were comparable to warfarin (Table 1) [27].

In the ARISTOTLE study, a subgroup analysis in relation to age strata by dividing patients into three groups, i.e. age <65 (n = 5471), age 65–74 (n = 7052), age ≥75 years (n = 5678) was made [10]. The efficacy of apixaban in reducing the incidence of IS/SE was more evident in elderly (i.e. ≥65 years) compared to young patients. While in patients aged <65 years, the annual rate of IS/SE was similar for apixaban and warfarin (0.9% versus 1%), it was 1.3%/year versus 1.7%/year, respectively in patients age 65–74 years, and 1.6%/year versus 2.2%/year in patients ≥75 years. Similarly, the safety of apixaban was more evident by ageing, with a rate of major bleeding of 2.0%/year versus 2.8%/year in patients 64–75 years and 3.3%/year versus 5.2%/year in patients ≥75 years compared with warfarin [10].

In the AVERROES trial, there were 1898 patients aged ≥75 years, and 366 patients aged ≥85 years [28]. The efficacy of apixaban was evident also in elderly patients compared to aspirin (Table 1). In particular, in patients aged ≥75 years, apixaban was associated with a significant reduction of IS/SE compared with younger patients (1.0%/year versus 1.7%/year, p = .05) [28]. The risk of major bleeding was comparable between apixaban and aspirin regardless of age strata [28].

In a pre-specified analysis of the ENGAGE AF-TIMI 48 trial, elderly patients (age ≥75 years, n = 8474) were more likely to be female, with lower body weight and impaired renal function, leading to a higher rate of edoxaban dose reduction (30 mg daily). Elderly patients had a higher risk of IS/SE (HR, 1.83; 95%CI, 1.32–2.54; p value not given), major bleeding (HR, 2.68; 95%CI, 2.04–3.52; p value not given) and ICH (HR, 3.77; 95%CI, 1.94–7.30; p value not given) compared with patients ≤65 years [29]. After 2.8 years of follow-up, edoxaban resulted in a similar efficacy and reduced risk of major bleeding, haemorrhagic stroke and ICH in elderly patients (Table 1). Furthermore, edoxaban significantly reduced the rate of fatal bleeding by 54% among elderly patients, compared with 28% in young patients [29]. Hence, the treatment of elderly patients with edoxaban provided a greater absolute reduction in bleeding events over warfarin compared to young patients [29].

A new study on elderly Japanese AF patients has been recently launched, the Edoxaban Low-Dose for EldeR CARE AF patients (ELDERCARE-AF) study [57]. This study is a phase III, randomized, double-blind, placebo-controlled, parallel-group, multicentre study comparing the safety and efficacy of edoxaban 15 mg versus placebo in Japanese patients with AF aged ≥80 years who are ineligible for treatment with VKAs. The study plans to include a total of 800 patients (n = 400 in each treatment group).

Patients with well-controlled warfarin

Optimal time in therapeutic range (TiTR) is needed to guarantee a good efficacy and safety of VKAs [58]. Thus, VKAs-anticoagulated AF patients with TiTR ≥70% have a lower risk of IS/SE, bleeding and cardiovascular events compared to those with poor anticoagulation control [59,60]. The role of NOACs in preventing IS/SE in AF patients according to different levels of anticoagulation quality has been studied.

A subgroup analysis of the RE-LY trial demonstrated that dabigatran 150 mg reduced IS/SE in patients with low TiTR (<65.5%), while broadly similar efficacy was found compared to patients with well-managed warfarin (TiTR >65.5%) (Table 1). Conversely, dabigatran 110 mg had similar efficacy to warfarin across all quartiles of TiTR [30]. Regarding bleeding outcomes, both dosages of dabigatran reduced major bleeding compared to poor-controlled warfarin (TiTR <57.1%), while the safety was similar to warfarin in the other groups of TiTR [30]. The rate of ICH bleeding was lower in both groups of dabigatran compared to warfarin with TiTR >72.6% (Table 1). Hence, dabigatran was superior to warfarin in patients with poorly controlled anticoagulation, while it has similar efficacy and better safety in those patients with well-managed warfarin.

In the ROCKET-AF trial [31], the primary efficacy outcomes of rivaroxaban were similar compared with warfarin across quartiles of cTiTR values [31]. Rivaroxaban demonstrated a reduction in major and non-major clinically relevant bleeding in the lowest quartile of TiTR (<51%), whilst rivaroxaban had an increased bleeding risk in patients with well-controlled warfarin (TiTR >66%) (Table 1) [31].

In the ARISTOTLE trial, the median centre average TiTR of AF patients on warfarin was 66% [32]. Apixaban reduced major bleeding by 36–50% in patients with low-quality warfarin therapy (i.e. TiTR <66%). This result was also evident in patients with TiTR >71.2%, apixaban had a 25% reduced rate of major bleeding (Table 1), and a non-significant different rate of primary outcomes (IS/SE) and all-cause death (HR, 0.91; 95%CI, 0.71–1.16; p for interaction = .34) [32].

In the ENGAGE-AF trial, the efficacy of edoxaban, both with HDER and low-dose edoxaban regimen (LDER) was independent of the level of TiTR (> or ≤66.4%, respectively) [11]. LDER reduced major bleeding both in patients with low and high TiTR, while HDER only in patients with low TiTR (non-inferior in patients with high TiTR).

In a retrospective registry study (n = 12,694), propensity matching analysis was made between NOACs (dabigatran, rivaroxaban, and apixaban) and well-controlled warfarin (mean TiTR 70%) [61]. The efficacy in reducing IS/SE was similar in the two groups. Patients in NOACs had lower rates of major bleeding (HR, 0.78; 95%CI, 0.67–0.92; p = .003), ICH (HR, 0.59; 95%CI, 0.40–0.87; p = .008) and haemorrhagic stroke (HR, 0.49; 95%CI, 0.28–0.86; p = .01), confirming that NOACs are safer also than well-controlled VKAs [61].

Patients with chronic kidney disease

Chronic kidney disease (CKD) is an important risk factor for both IS/SE and bleeding in AF patients with anticoagulation [35–37,62]. Moreover, the decline of renal function is an important aspect to be considered in AF patients, as up to 30% of patients may have a rapid annual worsening of renal function [63–65]. All NOACs have some degree of renal excretion, with dabigatran (≈80%) being the most excreted by the kidney, followed by edoxaban (≈35%), rivaroxaban (≈30%) and apixaban (≈27%). Thus, the assessment of renal function before and during treatment with NOACs is important for clinicians to choose/modify the dosage of NOACs [66,67].

All the four NOACs trials excluded patients with severe CKD, such as those with creatinine clearance (CrCl) < 30 mL/min (<25 mL/min for apixaban in the ARISTOTLE trial) [68]. Currently approved label allowing the use of some NOACs down to CrCl 15 mL/min is based on pharmacokinetic modelling, given that the trials excluded patients with CrCl <25–30 mL/min.

In the RE-LY trial, renal function was evaluated in 17,951 patients, with 10,697 patients having a CrCl 50–80 mL/min, and 3374 patients a CrCl <50 mL/min [33]. Dabigatran 150 mg bid significantly reduced both IS/SE and ICH independently of the severity of kidney disease (Table 1). Patients on dabigatran 110 mg had similar rates of IS/SE and major bleeding, while a significantly lower rate of ICH (Table 1) [33].

In a subgroup analysis of ROCKET-AF study, AF patients with moderate CKD (CrCl, 30–49 mL/min) were randomized to receive rivaroxaban 15 mg (n = 1474) or warfarin (n = 1476) [35]. The primary efficacy (IS/SE) and safety (major and clinically relevant non-major bleeding) endpoints were comparable between the two treatments (Table 1) [35]. In the intention-to-treat analysis of CKD patients, the rate of ICH or fatal bleeding occurred significantly less frequently in the rivaroxaban-treated group [9]. Similar results were obtained in Japanese patients from the J-ROCKET AF trial which included 141 AF patients with moderate CKD (CrCl 30–49 mL/min, Table 1) [34].

In the ARISTOTLE trial, 7587 patients have mild CKD (CrCl >50–80 mL/min), and 3017 with a moderate CKD (CrCl ≤50 mL/min) [36]. In patients with mild CKD, apixaban was associated with a reduced rate of both IS/SE and major bleeding (Table 1). In patients with moderate CKD, apixaban was non-inferior to warfarin in reducing IS/SE, but had a 50% reduction of major bleeding (Table 1) [36]. In the AVERROES trial, among 1697 AF patients with stage III CKD (eGFR <60 mL/min estimated by the CKD-EPI equation), apixaban reduced the rate of IS/SE by 68% with a similar incidence of major bleeding (Table 1) [37].

In the ENGAGE AF-TIMI 48 study, 2740 (19.5%) patients had a moderate CKD, and 8208 patients (58.3%) had a mild CKD [69]. In this study, patients on HDER had similar rates of IS/SE, while HDER reduced major bleeding in patients with moderate CKD, and significantly reduced ICH in all CKD groups [69].

Patients with coronary artery disease

Prevalent and incident coronary artery disease (CAD), both stable CAD or acute coronary syndrome (ACS), is common in AF patients [70,71]. Up to 30% of AF patients in the four NOACs trials presented with concomitant CAD [8–11]. Patients with CAD need to be treated with antiplatelet agents for a period ranging from 3 to 6 months. However, combination therapy of anticoagulants and antiplatelet results in an increased risk of bleeding (HR, 1.50; 95%CI, 1.23–1.82 for aspirin plus VKA; p value not given; HR 1.84, 95%CI 1.11–3.06 for clopidogrel plus VKA; p value not given) [72]. Balancing the risk of IS/SE or CAD events and bleeding is a clinical challenge.

The three most common clinical scenarios presenting to clinician prescribing NOACs in this setting may be: patients with stable CAD (with or without antiplatelet therapy [APT]), patients with long-term APT (for ischemic heart disease or other conditions such as peripheral artery disease or carotid stenosis) and patients presenting with ACS (with or without stent placement).

Stable coronary artery disease

In the RE-LY trial, there were 1886 (31%) patients with prior CAD/MI patients in the dabigatran 110 mg group, 1915 (31%) in the dabigatran 150 mg group and 1849 (31%) in the warfarin group [73]. The rate of stroke/SE was similar between patients with (1.55%/year) and without (1.53%/year) previous CAD/MI treated both with dabigatran 110 or 150 mg twice daily [73]. The bleeding rate was also comparable between the two groups (3.94 versus 2.39%/year, respectively) [73]. Noteworthy, in this subgroup analysis, dabigatran was associated with a non-significant increased risk of total MI (HR, 1.28; 95%CI, 0.98–1.67; p = .07) [73].

A recent meta-analysis of observational studies including 539,559 patients interestingly found that dabigatran use was associated with a lower risk of MI in OAC-naive patients treated with dabigatran 150 mg compared with VKA, and a higher risk of MI in patients switching from VKA to dabigatran 110 mg [74]. Further studies, specifically designed to investigate MI occurrence in patients treated with Dabigatran will clarify these findings.

In the ROCKET AF trial, 1182 AF patients presented with a history of prior MI [38]. These patients tended to be older, more frequently male, and with more atherosclerotic risk factors and used antiplatelet agents in 48.1% of cases compared to 33.9% in the rivaroxaban arm without previous MI. Patients with prior MI were more likely to have SE, MI, vascular diseases and higher all-cause mortality compared to those without. Rivaroxaban reduced the risk of IS/SE and had similar safety of warfarin in this subgroup of patients (Table 1).

In the ARISTOTLE trial, a total of 6639 patients had previous CAD (documented CAD, prior MI and history of coronary revascularization) and 42% of whom were receiving aspirin at baseline [39]. Apixaban performed similarly in reducing IS/SE compared to warfarin in patients with a prior CAD, and significantly reduced major bleeding and ICH both in patients with and without CAD (Table 1) [39].

Concomitant use of antiplatelet therapy

In the RE-LY trial, in patients with concomitant APT therapy, dabigatran 110 mg twice daily was non-inferior to warfarin in reducing IS/SE and was associated with fewer major bleeding and ICH than warfarin [40]. Similarly, dabigatran 150 mg twice daily had comparable efficacy and safety for major bleeding endpoints as warfarin, but it significantly reduced ICH [40].

The 4912 patients who received single APT (SAPT) (aspirin 92%) in the ENGAGE AF-TIMI 48 trial were more frequently male, with a personal medical history more complicated by CAD and diabetes [41]. Safety and efficacy outcomes were analysed for HDER or LDER (Table 1) [41]. HDER in addition to SAPT significantly reduced the rate of IS/SE and ICH, while the LDER had a similar efficacy of warfarin, but reduced the rate of both major bleedings and ICH (Table 1). There was no significant interaction between edoxaban and use of APT, with HDER associated to a reduced the rate of IS/SE and ICH, while the LDHER showed the similar efficacy of warfarin, but reduced the rate of both major bleedings and ICH [41]. Overall, both doses of edoxaban had a better clinical benefit compared to warfarin in this subgroup of patients [41].

The relative safety profile of Apixaban in the ARISTOTLE trial was not significantly different between the two groups [10]. No post-hoc analysis for interaction between Rivaroxaban and antiplatelet drugs has been published.

Acute coronary syndrome

In this high-risk group of patients, we have to balance the risk of stroke, stent thrombosis and serious bleeding. Triple therapy (OAC plus dual antiplatelet) followed by OAC plus aspirin and/or clopidogrel should be used for 1–3 months, respectively [75]. After 12 months, OAC monotherapy is reasonable for long-time treatment [75]. Two recent randomized clinical trials investigated the safety and efficacy of NOACs in combination with SAPT and dual APT (DAPT) in patients with AF.

The Randomized Evaluation of Dual Antithrombotic Therapy With Dabigatran versus Triple Therapy With Warfarin in Patients with Non-valvular Atrial Fibrillation Undergoing Percutaneous Coronary Intervention (RE-DUAL PCI) trial included 2725 patients with AF and undergoing PCI. In these patients, dabigatran (110 mg or 150 mg bid) plus P2Y₁₂ inhibitor was compared to triple therapy (dose-adjusted warfarin plus DAPT) [42]. Both doses of dabigatran reduced the risk of major bleeding and dabigatran 150 mg was associated with lower risk of ICH, while the risks of stroke were comparable between dabigatran plus P2Y12 group and triple therapy group (Table 1) [42]. The risks of MI (HR, 1.51; 95%CI, 0.94–2.41; p = .09 for dabigatran 110 mg; HR,1.16; 95%CI, 0.66–2.04; p = .61 for dabigatran 150 mg) and stent thrombosis (HR,1.86; 95%CI, 0.79–4.40; p = .15 for dabigatran 110 mg; HR,0.99; 95%CI, 0.35–2.81; p = .98 for dabigatran 150 mg) were similar between dual and triple therapy, suggesting the effectiveness of dabigatran plus single APT in reducing myocardial ischemic events [42].

The role of rivaroxaban in patients with AF and undergone PCI was investigated in the Open-Label, Randomized, Controlled, Multicentre Study Exploring Two Treatment Strategies of Rivaroxaban and a Dose-Adjusted Oral Vitamin K Antagonist Treatment Strategy in Subjects with Atrial Fibrillation Who Undergo Percutaneous Coronary Intervention (PIONEER AF-PCI) trial (n = 2124) [43]. Patients were randomized into three groups: rivaroxaban 15 mg once daily plus SAPT for 12 months, rivaroxaban 2.5 mg twice daily plus DAPT for a pre-specified duration of 1, 6 or 12 months or triple therapy with dose-adjusted warfarin plus DAPT for a pre-specified duration of 1, 6 or 12 months.

Both doses of rivaroxaban had lower risk of the composite safety endpoint of clinically relevant bleeding, compared to warfarin plus DAPT. However, when considered separately, the major bleeding rate was similar among the three groups (Table 1) [43]. Regarding efficacy outcomes, both rivaroxaban-based groups were associated with similar risks of IS compared with triple therapy with warfarin (Table 1) [43].

Patients undergoing cardioversion

Cardioversion, electrical, pharmacological or ablation, have an inherited risk of thromboembolism [76,77], which can be reduced by optimal anticoagulation [78]. Ancillary analyses from the randomized trials of NOACs have demonstrated the safety and efficacy of NOACs in preventing stroke in AF patients undergoing cardioversion [44,45,79]. In patients undergoing cryoballoon ablation, NOACs were associated with lower risk of serious bleeding (p < .05) [80]. Some meta-analysis of the real-world studies also demonstrated the safety and efficacy of NOACs in AF patients undergoing catheter ablation, with a similar risk of IS and ∼20–50% reduced the risk of bleeding events [81–84].

A total of 1983 cardioversions were performed in 1270 AF patients in the RE-LY trial [44]. Over 3 weeks of anticoagulation with dabigatran 110 mg, dabigatran 150 mg or warfarin were carried before cardioversion. The IS/SE rates at 30 days were compared among three groups (0.8, 0.3 and 0.6%, respectively, p > .1), indicating that dabigatran is an effective alternative to warfarin in patients requiring cardioversion [44]. In the Randomized Evaluation of Dabigatran Etexilate Compared to Warfarin in Pulmonary Vein Ablation: Assessment of an Uninterrupted Periprocedural Anticoagulation Strategy (RE-CIRCUIT) trial, dabigatran 150 mg twice a day was associated with significantly lower major bleeding events peri-ablation (1.6% versus 6.9%, respectively, p < .001) [85]. A real-world retrospective cohort study (n = 374) showed lower minor haemorrhage (5.7% versus 15%, p = .012) in the dabigatran compared with warfarin group in patients receiving AF ablation [86].

In the post-hoc analysis of the ROCKET AF trial, 285 patients experienced 321 cardioversions (including 79 AF ablation) [45]. The incidence of IS/SE were similar in the rivaroxaban and warfarin-treated groups (Table 1) [45], while the incidence of major bleeding or non-major clinically relevant bleeding was increased in the rivaroxaban group (Table 1) [45]. This has been tested in the prospective eXplore the efficacy and safety of once-daily oral riVaroxaban for the prEvention of caRdiovascular events in patients with non-valvular aTrial fibrillation scheduled for cardioversion trial (X-VeRT) trial [46], which included 1504 AF patients randomized to rivaroxaban or warfarin for 6 weeks after cardioversion. In a 30-day follow-up after treatment, the rate of IS and major bleeding were non-significantly different between the two treatment groups (Table 1) [46]. In the ActiVe-controlled multi-cENTer stUdy with blind-adjudication designed to evaluate the safety of uninterrupted Rivaroxaban and uninterrupted vitamin K antagonists in subjects undergoing cathEter ablation for non-valvular Atrial Fibrillation (VENTURE-AF) trial (n = 250), rivaroxaban 20 mg showed similar efficacy and safety compared to uninterrupted warfarin [87].

In a post-hoc analysis of the ARISTOTLE trial, a total of 743 cardioversions were performed in 540 patients (265 patients in apixaban group [5 mg twice a day or 2.5 mg twice day in patients who had 2 or more of the following criteria: age ≥80 years, body weight ≤60 kg or serum creatinine level ≥1.5 mg/dL] and 275 patients in warfarin group) [79]. A low number of efficacy and safety outcomes were registered in both groups, making it underpowered for any comparison [79]. The Apixaban compared with parenteral heparin and/or vitamin K antagonist in patients with non-valvular atrial fibrillation undergoing cardioversion (EMANATE) trial (n = 1500) is a prospective, open-label study comparing apixaban to warfarin in AF patients undergoing cardioversion [88]. This trial demonstrated that apixaban had lower risk of IS/SE compared with heparin/warfarin (0% versus 0.8%, p = .016). The risks of major bleeding were also numerically lower in apixaban compared with heparin/warfarin (0.4% versus 0.8%) [89]. Recently, the randomized, prospective multi-centre Apixaban During Atrial Fibrillation Catheter Ablation: Comparison to Vitamin K Antagonist Therapy (AXAFA-AFNET 5) trial (n = 650) has explored the efficacy and safety of apixaban in patients undergoing AF catheter ablation. The rate of the primary outcome (composite of stroke, major bleeding and all-cause mortality up to 3 months after ablation) was similar between patients randomized to apixaban or warfarin (22 and 23 patients, respectively). Only two-stroke events were observed in the trial (0.3%) [90]. It demonstrated that apixaban is a safe alternative to warfarin during catheter ablation of AF.

The A prospective evaluation of edoxaban compared to warfarin in subjects undergoing cardioversion of atrial fibrillation: the edoxaban versus warfarin in subjects undergoing cardioversion of atrial fibrillation (ENSURE-AF) trial was a multicentre, prospective, randomized trial in 19 countries with over 200 sites comparing edoxaban with enoxaparin-warfarin in AF patients undergoing cardioversion [47]. AF patients were assigned to receive edoxaban (n = 1095) or enoxaparin/warfarin (n = 1104) and followed for 30 days after completing or discontinuing the treatment. The rates of primary endpoint and safety outcome were low and similar between the two groups (<1% for both; Table 1) [47]. Thus, edoxaban is an effective and safe alternative to conventional anticoagulation (enoxaparin/warfarin) and may ensure prompt cardioversion for AF patients.

Asian patients

Ethnic differences between Asian and non-Asian populations may affect the optimal dosing of the NOACs. Performance of the NOACs has been addressed among Asian patients in various trials and observational studies. Subgroup analysis of the NOACs trials demonstrated greater benefits of NOACs among Asians (OR, 0.65; 95%CI, 0.52–0.83) compared to non-Asians (OR, 0.85; 95%CI, 0.77–0.93) in reducing IS/SE (p for interaction = .045) [91]. The NOACs performed better in Asians than in non-Asians regarding major bleeding (OR, 0.57; 95%CI, 0.44–0.74 versus OR, 0.89; 95%CI, 0.76–1.04, p for interaction = .004), haemorrhagic stroke (OR, 0.32; 95%CI, 0.19–0.52 versus OR, 0.56; 95%CI, 0.44–0.70, p for interaction = .046) [91]. Furthermore, the increased risk of gastrointestinal (GI) bleeding in NOAC trials was not observed in Asian patients (OR, 0.79; 95%CI, 0.48–1.32 versus OR, 1.44; 95%CI, 1.12–1.85, p for interaction = .041) [91].

In the subgroup analysis from the RE-LY trial, where 2782 patients were from 10 Asian countries, Hori et al. [48] found a higher rate of haemorrhagic stroke on warfarin in Asians versus non-Asians (HR, 2.40; 95%CI, 1.30–4.70; p = .007), despite lower international normalized ratio (INR) values [48]. Both doses of dabigatran significantly reduced the risk of major bleeding and ICH compared with warfarin (85% for dabigatran 110 mg, 78% for dabigatran 150 mg) (Table 1). Furthermore, dabigatran 150 mg reduced the rate of efficacy endpoint (IS/SE) (Table 1) [48]. In the subgroup analysis on Japanese population only from the RE-LY trial (n = 326), the efficacy and safety profiles of dabigatran for Japanese patients with AF were comparable to the RE-LY trial population overall [92].

In a subgroup analysis of East Asia participants of the ROCKET-AF trial (n = 932), rivaroxaban had a similar efficacy of warfarin, along with a significant reduction of major bleeding (Table 1) [49]. Furthermore, the J-ROCKET AF trial compared the safety of a Japan-specific rivaroxaban dose with warfarin administered according to Japanese guidelines [12]. Thus, among 1280 Japanese patients with non-valvular AF were enrolled, the INR range was 2.0–3.0 in patients aged <70 years and 1.6–2.6 in patients aged ≥70 years. The dose of rivaroxaban was 15 mg daily, reduced to 10 mg daily in patients with moderate renal impairment [12]. The J-ROCKET AF demonstrated a borderline efficacy with less stroke/SE and a similar rate of major and ICH compared with warfarin (Table 1) [12].

In the ARISTOTLE trial, 1993 patients were recruited from East Asia. Compared with warfarin, apixaban had consistent reduction in major bleeding in East Asian and non-East Asian patients [50], and between East Asian and non-Asian patients (51% versus 29%, p for interaction =0.03) [50]. Compared with non-Asian patients, the East Asian patients had a higher risk of ICH (1.27%/year versus 0.48%/year, p < .0001) [50]. The effect of apixaban in ICH reduction was similar in East Asian and non-East Asian patients (64% per year versus 57% per year, p for interaction = .60) [50]. Overall, apixaban was non-inferior to warfarin for stroke prevention but resulted in a greater reduction in major or clinically relevant non-major bleeding in East Asian patients (Table 1) [50].

In the ENGAGE AF-TIMI 48 study, 1943 AF patients from Asia were randomized to receive HDER, LDER or warfarin groups [51]. In the pre-specified sub-analysis, both doses of edoxaban resulted in decreased risks of major bleeding and ICH (Table 1) [51]. HDER also significantly reduced the risk of IS/SE by 47% (Table 1) [51]. Furthermore, LDER reduced the risk of haemorrhagic stroke to a great extent (HR, 0.22; 95%CI, 0.07–0.64; p < .01) [51]. Importantly, HDER significantly reduced the risks of cardiovascular-cause death (HR, 0.46; 95%CI, 0.25–0.84; p = .01) and all-cause death (HR, 0.63; 95%CI, 0.40–0.98; p = .04) [51].

Observational data also support the efficacy and safety of NOACs in Asians. In Japan, the Japanese Rhythm Management Trial for Atrial Fibrillation (J-RHYTHM) Registry (n = 6616), the beneficial effect of NOACs for Asian AF patients was further confirmed [93]. The patients treated with NOAC (n = 923) had lower rates of thromboembolism (2.1% versus 4.9%), major haemorrhage (2.4% versus 5.9%) and all-cause death (1.4% versus 5.8%) compared with warfarin (p <.01 for each) [93].

In Taiwan National Health Insurance Research Database, 19,853 patients with non-valvular AF were analysed, in which 9940 were treated with dabigatran (1168 with 150 mg and 8772 with dabigatran 110 mg) and 9913 with warfarin [94]. Dabigatran 150 mg showed similar efficacy (HR, 0.61; 95%CI, 0.37–1.00; p = .0505) and better safety (HR for major bleeding, 0.43; 95%CI, 0.19–0.99; p = .0471; HR for ICH, 0.22; 95%CI, 0.06–0.76; p = .0167) of warfarin. Dabigatran 110 mg was associated with less IS/SE (HR, 0.62; 95%CI, 0.52–0.75; p < .0001), major bleeding (HR, 0.61; 95%CI, 0.47–0.78; p = .0001) and ICH (HR, 0.47; 95%CI, 0.34–0.65; p < .0001) [94].

The Korean National Database, including 11,611 AF patients treated with the NOACs (n = 3741 on dabigatran, n = 5681 on rivaroxaban and n = 2189 on apixaban), demonstrated that NOAC treatment was associated with similar risk of IS (HR, 0.98; 95%CI, 0.78–1.22; p = .5708) and lower risk of ICH (HR, 0.50; 95%CI, 0.36–0.68; p < .0001) compared with warfarin [95].

The Hong Kong Atrial Fibrillation Project (n = 2099) demonstrated that the incidence of IS was 2.82%/year in NOAC (dabigatran and rivaroxaban) group and 4.53%/year in warfarin group. Patients on dabigatran had the lowest incidence of IS (1.89%/year), followed by those on rivaroxaban (3.74%/year) [96]. The incidence of ICH was 0.46%/year in NOAC group (0.39%/year for dabigatran and 0.52%/year for rivaroxaban) and 0.89%/year in warfarin group [96].

In a national retrospective cohort study involving 15,088 Asian patients, both dabigatran and rivaroxaban significantly decreased the risk of IS/SE [97]. However, dabigatran had a lower risk for hospitalization for gastrointestinal bleeding than rivaroxaban (p = .04) although the on-treatment analysis did not show a statistical difference between the two groups (p = .58) [97].

Conclusion

A summary of NOACs with regard to the best efficacy and safety compared to warfarin according to each subgroup is provided in Table 2. Real-world post-marketing studies play a substantial role in supporting the use of NOACs in everyday clinical practice [99,100].

Table 2. Summary of the performance of NOACs in subgroups of patients (based on hazard ratio for each outcome compared to warfarin).

Subgroups of patients	Better efficacy	Better safety (HR for major bleeding)	Better safety (HR for ICH)
Previous IS/TIA	All NOACs non-inferior	Dabigatran 110 (0.66) Apixaban 5 (0.73)	Dabigatran 110 (0.20) Apixaban 5 (0.37) Dabigatran 150 (0.41) Edoxaban 60 (0.57)
Elderly (≥75 years)	Dabigatran 150 (0.67) Apixaban (0.71)^d	Apixaban (0.74)^d Edoxaban 60 (0.83)	Apixaban (0.34)^d Dabigatran 110 (0.37) Edoxaban 60 (0.40) Dabigatran 150 (0.42)
Well-controlled TiTR	All NOACs non-inferior	Apixaban 5 (0.75) Edoxaban 30 (0.48)^b	Dabigatran 110 (0.27) Dabigatran 150 (0.37) Apixaban 5 (0.75)^a
Moderate CKD	Dabigatran 150 (0.56)	Apixaban 5 (0.50) Edoxaban 60 (0.76)	Dabigatran 150 (0.31) Dabigatran 110 (0.40) Edoxaban 60 (0.46)
Concomitant antiplatelet	Apixaban (0.67)^c Edoxaban 60 (0.70)	Edoxaban 30 (0.51) Apixaban (0.78)^c Dabigatran 110 (0.82)	Edoxaban 30 (0.18) Dabigatran 110 (0.23) Edoxaban 60 (0.46) Dabigatran 150 (0.47)
Stable CAD	Rivaroxaban (0.61)	Edoxaban 30 (0.42)^b Apixaban 5 (0.78)	Apixaban 5 (0.36)
ACS	Rivaroxaban and Dabigatran non-inferior	Dabigatran 110 (0.52) Dabigatran 150 (0.64)	Dabigatran 150 (0.12)
Cardioversion	Apixaban	Dabigatran and Edoxaban non-inferior	N/A
Asian Patients	Rivaroxaban 15 (0.40) Dabigatran 150 (0.45) Edoxaban 60 (0.53)	Rivaroxaban 20 (0.24) Edoxaban 30 (0.34) Apixaban 5 (0.53) Dabigatran 110/150 (0.57) Edoxaban 60 (0.61)	Dabigatran 110 (0.20) Edoxaban 30 (0.24) Edoxaban 60 (0.31) Apixaban 5 (0.36) Dabigatran 150 (0.40)

Number in bracket stands for hazard ratio.

^a For major bleedings.

^b From the Supplementary Appendix 6: Figure S4 of ENGAGE-AF [11].

^c From Figure 2 of ARISTOTLE trial [98].

^d HR from subgroup analysis from ARISTOTLE [101]. For abbreviations see Table 1 footnote.

Despite new evidence coming from recent trials, further research is necessary for patients experiencing MI or undergoing PCI, or for those requiring long-term antiplatelet therapy in addition to anticoagulation. Similarly, the choice of NOAC in some settings, such as low-risk patients or post-acute stroke (with or without thrombolysis), is still limited.

Disclosure statement

None directly related to this paper. G. Y. H. L.: Consultant for Bayer/Janssen, BMS/Pfizer, Medtronic, Boehringer Ingelheim, Novartis, Verseon and Daiichi-Sankyo. Speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer Ingelheim, and Daiichi-Sankyo. No fees are directly received personally. Other authors: None declared.