Real-World Outcomes of Patients With Non-Valvular Atrial Fibrillation on Anticoagulants Ineligible for Phase III Trials of Direct Oral Anticoagulants　― A Retrospective Cohort Study ―

Abstract

Background: Many patients with non-valvular atrial fibrillation (NVAF) on direct oral anticoagulants (DOACs) in real-world practice were ineligible for DOAC phase III trials. We aimed to determine the proportion of Japanese patients with NVAF eligible for these trials and compare the characteristics and outcomes of ineligible and eligible patients to determine the generalizability of the trial results.

Methods and Results: This retrospective cohort study included 7,826 Japanese NVAF patients on warfarin from 71 hospitals. We assessed trial eligibility and analyzed outcomes (major bleeding, stroke/systemic embolism, all-cause mortality) using Cox proportional hazards models. Nearly half (48.2%; n=3,772) of the patients were ineligible for DOAC phase III trials. Ineligible patients were older with more comorbidities and exhibited significantly higher risks of death (unadjusted hazard ratio [HR] 2.84; 95% confidence interval [CI] 2.36–3.43; P<0.0001), stroke/systemic embolism (unadjusted HR 1.53; 95% CI 1.17–1.98; P=0.0016), and major bleeding (unadjusted HR 2.00; 95% CI 1.63–2.44; P<0.0001) compared with eligible patients.

Conclusions: Half of the NVAF patients receiving anticoagulant therapy in real-world practice were ineligible for phase III DOAC trials, primarily due to safety concerns. This population differs substantially from eligible patients in characteristics and outcomes. The generalizability of phase III results to real-world patients remains uncertain, warranting additional assessment.

Central Figure

Direct oral anticoagulants (DOACs) have been increasingly prescribed to patients with non-valvular atrial fibrillation (NVAF) following regulatory approval and guideline recommendations based on successful phase III trials, demonstrating their non-inferiority and superiority to warfarin.^1–5 However, the results from pivotal trials should be interpreted with caution because of the strict inclusion and exclusion criteria to ensure participant safety and provide clear results for regulatory approval. The Food and Drug Administration recommends that enrolling participants with a wide range of baseline characteristics may create a trial population that more accurately reflects patients who are likely to take the drug if it is approved, and may allow evaluation of the impact of these characteristics on the safety and efficacy of the trial drug.⁶ The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Steering Committee also stated that as drug development progresses, the study population should be expanded to adequately reflect the target population.⁷

These recommendations aim to make the results of phase III trials applicable to a wider range of patients. Conversely, some studies have attempted to assess the generalizability of the results of DOAC phase III trials using patient eligibility in real-world clinical practice as an indicator. The proportion of real-world patients eligible for trials varied widely, ranging from 35% to 72%.^8–13 There were trends towards low eligibility for ROCKET-AF and high eligibility for ARISTOTLE and RE-LY, presumably because of the CHADS₂ score as an inclusion criterion. It is possible to prescribe DOACs to these patients, although their efficacy and safety have not been formally confirmed in clinical trials in accordance with drug labels. However, until now, there has been little discussion on how the ineligibility for these trials affects outcomes, which may also be relevant to the generalizability of the trial results. Thus, this study aimed to determine the proportion of real-world Japanese patients with NVAF who were eligible for DOAC phase III trials, and compare the characteristics and outcomes of ineligible and eligible patients in order to assume the generalizability of the trial results.

Methods

Study Design

We conducted a historical registry study of patients with NVAF who were taking warfarin at 71 centers in Japan. Patients were registered on February 26, 2013, and followed up until February 25, 2017.

Ethics

The ethics committees of all 71 participating centers approved this study, following the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan. The requirement for written informed consent was waived.

Patients

Consecutive patients with NVAF who had already received warfarin on February 26, 2013, were retrospectively registered. Patients with mechanical heart valves or previous pulmonary or deep vein thrombosis were excluded from the study.

Data Collection and Definitions

Trained clinical research coordinators reviewed and collected relevant clinical information, including patient characteristics, laboratory data, risk of ischemic stroke (CHADS₂ score), major bleeding (HAS-BLED score) at baseline, and relevant medications directly from medical charts, at baseline and at 1, 2, 3, and 4 years of follow up. Definitions of baseline characteristics have been described previously.¹⁴ The CHADS₂ score was developed to estimate the stroke risk in patients with NVAF,¹⁵ and the HAS-BLED score was developed to predict major bleeding in patients receiving anticoagulation therapy.¹⁶

Assessment of Registered Patient Eligibility for DOAC Phase III Trial

We scrutinized the most frequently used inclusion and exclusion criteria for the 4 DOAC phase III trials (Table 1).^1,2,4,5 Baseline data collected at registration for all patients (medical history, background, and laboratory values) were checked against the developed inclusion and exclusion criteria to assess eligibility. Regarding the CHADS₂ score, 2 trials used ≥2 as a criterion, and the others used ≥1. We used ≥2 as our criterion. Therefore, ineligible patients were those with a CHADS₂ score <2. Additionally, we checked the baseline data against the inclusion and exclusion criteria for each phase III trial.

Table 1.

Inclusion/Exclusion Criteria Developed for the Assessment of Eligibility of the Registered Patients and Criteria of 4 DOAC Phase III Trials

	Present study	RE-LY1	ROCKET-AF2	ARISTOTLE4	ENGAGE TIMI485
Inclusion criteria
Age	≥18 years	≥18 years	≥18 years	≥18 years	≥21 years
AF	Non-valvular AF	AF	Non-valvular AF	AF or atrial flutter	Paroxysmal, persistent, or permanent AF
CHADS2 score	CHADS2 score ≥2	One of the following: (1) prior stroke, TIA or SE; (2) EF<40%; (3) HF; (4) age ≥75 years; (5) age ≥65 and <75 years, and diabetes, HT or CAD	CHADS2 score ≥2	CHADS2 score ≥1	CHADS2 score ≥2
Exclusion criteria
Bleeding risk	History of major bleeding	History of major bleeding	History of major bleeding	N/A	History of major bleeding
Uncontrolled hypertension	SBP >180 mmHg, or DBP >100 mmHg	SBP >180 mmHg, and/or DBP >100 mmHg	SBP ≥180 mmHg, or DBP ≥100 mmHg	SBP >180 mmHg, or DBP>100 mmHg	SBP >170 mmHg, or DBP >100 mmHg
Renal function	CCr <30 mL/min	CCr ≤30 mL/min	CCr <30 mL/min	Serum creatinine >2.5 mg/dL or CCr <25 mL/min	CCr <30 mL/min
Hepatic function	ALT or AST >2 × ULN, or TBIL ≥1.5 × ULN	Active liver disease, including but not limited to persistent ALT, AST, or ALP >2 × ULN	Known significant liver disease, or ALT >3 × ULN	ALT or AST >2 × ULN, or TBIL ≥1.5 × ULN	Active or persistent liver disease, positive hepatitis B or C test, in: ALT or AST ≥2 × ULN; TBIL ≥1.5 × ULN
Hemoglobin and platelet count	Hb <10 g/dL, or platelet count <100,000 cells/mL	Hb <10 g/dL, or platelet count <100,000 cells/mL	Hb <10 g/dL, or platelet count <90,000 cells/mL	Hb <9 g/dL, or platelet count ≤100,000 cells/mL	Hb <10 g/dL, or platelet count <100,000 cells/mL, or WBC <3,000 cells/mL
Antiplatelet therapy	Aspirin in combination with thienopyridines	N/A	Aspirin >100 mg/day, or aspirin in combination with thienopyridines	Aspirin >165 mg/day, or aspirin in combination with thienopyridines	Aspirin in combination with thienopyridines
Anti-inflammatory agents	Chronic treatment with NSAIDs	N/A	Anticipated need for chronic treatment with NSAIDs	N/A	Chronic treatment with NSAIDs
Concomitant therapy		N/A	CYP3A4 inducer; CYP3A4 inhibitor	CYP3A4 inhibitor; macrolide antibiotics	Cox-2 inhibitor
Drug or alcohol dependence	Alcohol dependence	N/A	Drug addiction, or alcohol dependence	Drug addiction, or alcohol dependence	Drug addiction, or alcohol dependence

AF, atrial fibrillation; ALP, alkaline phosphatase; ALT, alanine aminotransferase; ARISTOTLE, the apixaban for the prevention of stroke in subjects with atrial fibrillation; AST, aspartate aminotransferase; CAD, coronary artery disease; CCr, creatinine clearance; Cox-2, cyclooxygenase-2; CYP3A4, cytochrome P450 3A4; DBP, diastolic blood pressure; EF, ejection fraction; ENGAGE TIMI48; effective anticoagulation with factor Xa next generation in atrial fibrillation; Hb, hemoglobin; HF, heart failure; HT, hypertension; N/A, not available; NSAIDs, non-steroidal anti-inflammatory drugs; RE-LY, the randomized evaluation of long-term anticoagulation therapy; ROCKET-AF, the rivaroxaban once daily oral direct factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and embolism trial in atrial fibrillation; SBP, systolic blood pressure; SE, systemic embolism; TBIL, total bilirubin; TIA, transient ischemic attack; ULN, upper limit of normal; WBC, white blood cell.

Outcome Measure

Outcomes measured in this cohort study were as follows: major bleeding, defined according to the International Society on Thrombosis and Hemostasis criteria;^17,18 ischemic stroke, including transient ischemic attack with systemic embolism; hemorrhagic stroke; and all-cause mortality.

Comparison of Outcomes Between Eligible and Ineligible Patients With CHADS₂ Score ≥2

We compared the outcomes between eligible and ineligible patients after excluding patients with a CHADS₂ score <2.

Variables Associated With All-Cause Death

Cox regression was used to derive hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) to investigate the associations between mortality and several clinically relevant variables in eligible and ineligible patients with NVAF.

Proportion of Patients Who Switched From Warfarin to DOACs

The proportion of patients who switched anticoagulants, from warfarin to DOACs, during the observation period was also calculated.

Statistical Analysis

Descriptive Statistics The eligibility proportions of all patients were determined. The proportion of patients who met each inclusion criterion and conflicted with each exclusion criterion were also determined.

Clinical characteristics were described and compared between the eligible and ineligible patients with CHADS₂ scores ≥2. Categorical variables are presented as numbers and percentages. Continuous variables are expressed as means±standard deviations (SDs), median value, or interquartile range. Student’s t-test or Wilcoxon rank-sum test was performed for continuous variables to evaluate differences in characteristics between the 2 groups; the chi-squared test was used for categorical variables. Values with P<0.05 were considered statistically significant.

Survival Analysis The incidence rates of outcomes are presented as cases per 100 patient-years. The cumulative incidence of outcomes was estimated using the Kaplan-Meier method. Outcome differences between eligible and ineligible patients with CHADS₂ scores ≥2 were assessed using the log-rank test. The effects of patient ineligibility on outcomes were estimated using Cox proportional hazards models to calculate the HR and 95% CI. Statistical significance was set at P < 0.05 for all analyses. We present only unadjusted HR because the effects of ineligibility on outcomes reflect the integrated impact of many variables in ineligible patients. Variables associated with all-cause mortality were estimated using Cox proportional hazards models with clinically relevant variables.

Results

We registered 7,826 Japanese patients with NVAF and followed them for 4 years; the median age was 74 (66–80) years; 3,755 (48.0%) patients were aged at least 75 years, 2,552 (33%) were female, and the mean CHADS₂ score was 2.44.

Eligibility of All Registered Patients

Following our inclusion/exclusion criteria derived from the DOAC phase III trials (Table 1), 4,054 (51.8%) patients with NVAF were classified as the eligible group, whereas the remaining patients (n=3,772; 48.2%) were classified as the ineligible group, mainly because of CHADS₂ score <2 (26.4%), renal dysfunction (9.5%), anemia (6.4%), and chronic treatment with non-steroidal anti-inflammatory drugs (NSAIDs) (4.0%) (Table 2).

Table 2.

Proportion of Eligible and Ineligible Patients According to Inclusion and Exclusion Criteria

	DOACs criteria	Eligible n (%)	Ineligible n (%)
Inclusion criteria
Age	Age ≥18 years	7,826 (100)
AF	Non-valvular AF	7,826 (100)
CHADS2	CHADS2 score ≥2	5,763 (73.6)	2,063 (26.4)
Exclusion criteria
Hemorrhage risk	History of major bleeding	7,547 (96.5)	279 (3.5)
Uncontrolled hypertension	SBP >180 mmHg, or DBP >100 mmHg	7,701 (98.4)	125 (1.6)
Renal dysfunction	CCr <30 mL/min	7,085 (90.5)	741 (9.5)
Hepatic function	ALT or AST >2×ULN, or TBIL ≥1.5×ULN	7,608 (97.2)	218 (2.8)
Hemoglobin and platelet count	Hb <10 g/dL, or platelet count <100,000cells/mL	7,324 (93.6)	502 (6.4)
Antiplatelet therapy	Aspirin >100 mg daily, or aspirin in combination with thienopyridines	7,550 (96.5)	276 (3.5)
Anti-inflammatory agents	Chronic treatment with NSAIDs	7,513 (96)	313 (4.0)
Drug or alcohol dependence	Alcohol dependence	7,783 (99.5)	43 (0.5)
Overall		4,054 (51.8)	3,772 (48.2)

Abbreviations as in Table 1.

Comparisons of Baseline Characteristics Between Eligible and Ineligible Patients With CHADS₂ Score ≥2

Figure 1 presents a flow diagram of the study. We excluded patients with CHADS₂ <2 (n=2,063; 26.4% of all registered patients) from the comparison of outcomes. Table 3 presents the baseline characteristics of the eligible and ineligible patients with CHADS₂ ≥2. Ineligible patients were more likely to be elderly, female, and have comorbidities such as stroke, heart failure, coronary artery disease, chronic liver disease, and kidney disease compared with eligible patients. The CHADS₂ score, after excluding patients with <2, and the HAS-BLED score were significantly higher in the ineligible patients than in the eligible patients.

Figure 1.

Flow diagram of patients.

Table 3.

Baseline Characteristics of Eligible and Ineligible Patients With CHADS₂ Score ≥2

	Eligible (n=4,054)	Ineligible with CHADS2 ≥2 (n=1,709)	P value
Age, median [interquartile range] (years)	76 [69–81]	79 [73–85]	<0.0001
Female sex, no. (%)	1,320 (32.6)	708 (41.4)	<0.0001
Body weight, mean (kg)	62.9	57.5	<0.0001
BMI, mean (SD)	24.7 (4.04)	23.43 (4.31)	<0.0001
SBP, median (mmHg)	128	126	0.196
Paroxysmal AF, n (%)	1,304 (36.9)	467 (32.0)	0.0011
CHADS2 score, mean	2.96	3.24	<0.0001
CHADS2 score 2, n (%)	1,711 (42.2)	508 (29.7)	<0.0001
CHADS2 score ≥3, n (%)	2,343 (57.8)	1,201 (70.3)	<0.0001
HAS-BLED score	1.72	2.43	<0.0001
Prior stroke or TIA, n (%)	1,270 (31.4)	599 (35.2)	0.0047
Age ≥75 years, n (%)	2,324 (57.3)	1,198 (70.1)	<0.0001
HF, n (%)	2,020 (50.1)	1,074 (63.6)	<0.0001
Diabetes, n (%)	1,642 (40.5)	707 (41.4)	0.546
HT, n (%)	3,552 (87.7)	1,497 (87.7)	0.989
CAD, n (%)	1,111 (27.5)	653 (38.4)	<0.0001
ACS, n (%)	422 (10.6)	308 (18.5)	<0.0001
PCI, n (%)	286 (7.2)	318 (19.1)	<0.0001
CABG, n (%)	125 (3.1)	101 (6.0)	<0.0001
COPD, n (%)	177 (4.4)	93 (5.5)	0.076
Chronic liver disease, n (%)	290 (7.2)	183 (10.7)	<0.0001
History of cancer, n (%)	453 (11.2)	254 (14.9)	<0.0001
History of RFCA, n (%)	261 (6.5)	70 (4.1)	0.0004
CCr, mean (SD)	63.68 (24.7)	44.59 (27.5)	<0.0001
PT-INR, mean (SD)	1.91 (0.38)	1.87 (0.41)	0.0006
TTR, mean (SD)	74.69 (33.34)	68.91 (35.51)	<0.0001
ASA, n (%)	805 (19.9)	612 (35.9)	<0.0001

ACS, acute coronary syndrome; ASA, acetylsalicylic acid; BMI, body mass index; CABG, peripheral artery bypass graft; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; PT-INR, prothrombin time/international normalized ratio; RFCA, radiofrequency catheter ablation; SD, standard deviation; TTR, time in therapeutic range. Other abbreviations as in Table 1.

Comparisons of Outcomes Between Eligible and Ineligible Patients With CHADS₂ Score ≥2

At 4 years, 59% of patients completed the follow up, and the median follow-up duration was 3.9 years. The cumulative incidences of major bleeding at 4 years were 7.0% and 13.4% in the eligible and ineligible groups, respectively (Figure 2A). The risk of major bleeding in ineligible patients with CHADS₂ score ≥2 was almost double that of eligible patients and significantly higher (Table 4). The 4-year cumulative incidence rates of ischemic stroke and systemic embolism in the eligible and ineligible groups were 5.5% and 7.5%, respectively (Figure 2B). Notably, both groups had a CHADS₂ score ≥2; however, the risk of ischemic stroke/systemic embolism was significantly higher in ineligible patients than in eligible patients (Table 4). At 4 years, the cumulative incidences of death from any cause in the eligible and ineligible groups were 7.3% and 17.6%, respectively (Figure 2C). The risk of death from any cause in the ineligible patients was approximately 3 times higher than that in the eligible patients (Table 4).

Figure 2.

Cumulative incidence of major bleeding (A), stroke/systemic embolism (B), and death from any cause (C) in eligible and ineligible patients with CHADS₂ score ≥2.

Table 4.

Outcomes of Eligible and Ineligible Patients With CHADS₂ Score ≥2

Outcome	Eligible (n=4,054)		Ineligible and CHADS2 ≥2 (n=1,709)		Hazard ratio (95% CI)	P value
	No. events	Event rate (%/year)	No. events	Event rate (%/year)
Major bleeding	224	1.83	163	3.67	2.00 (1.63–2.44)	<0.0001
Stroke/SE	156	1.27	87	1.93	1.53 (1.17–1.98)	0.0016
Hemorrhagic stroke	36	0.29	18	0.39	1.36 (0.77–2.40)	0.283
Death from any cause	213	1.70	221	4.82	2.84 (2.36–3.43)	<0.0001

CI, confidence interval; SE, systemic embolism.

Variables Associated With All-Cause Death

Table 5 shows results from a multivariate analysis of all-cause death in eligible and ineligible patients. In both groups, reduced renal function, a history of cancer, and heart failure were significantly associated with death. As shown in Table 3, patients with these comorbidities were more prevalent in the ineligible group. In particular, severe renal function impairment (i.e., creatinine clearance [CCR] <30 mL/min), which was only included in the ineligible group, was strongly associated with death. In addition, COPD was significantly associated with death only in the eligible group.

Table 5.

Multivariate Analysis of All-Cause Death in Eligible and Ineligible Patients With CHADS₂ Score ≥2

	P value	Hazard ratio	95% CI
Multivariate analysis of death in eligible patients with CHADS2 score ≥2
Sex, male	0.0068	1.59	1.14–2.23
SBP ≥140 mmHg	0.71	0.93	0.65–1.34
CCr <60 mL/min	<0.0001	2.87	2.05–4.02
Diabetes	0.83	1.03	0.76–1.39
HF	0.0429	1.37	1.01–1.85
CAD	0.06	1.34	0.98–1.81
Prior stroke or TIA	0.68	0.94	0.68–1.29
COPD	0.0006	2.31	1.43–3.73
History of cancer	0.0012	1.83	1.27–2.63
Chronic liver disease	0.11	1.46	0.92–2.34
Multivariate analysis of death in ineligible patients with CHADS2 score ≥2
Sex, male	0.61	1.08	0.81–1.43
SBP ≥140 mmHg	0.053	0.70	0.48–1.00
CCr <60 mL/min	<0.0001	3.27	2.08–5.13
Diabetes	0.65	0.94	0.71–1.24
HF	0.0003	1.81	1.32–2.50
CAD	0.56	1.09	0.82–1.44
Prior stroke or TIA	0.42	1.12	0.85–1.49
COPD	0.46	1.22	0.72–2.07
History of cancer	0.002	1.68	1.21–2.34
Chronic liver disease	0.69	1.09	0.71–1.68

Other abbreviations as in Tables 1,3,4.

Proportion of Patients Who Switched From Warfarin to DOACs

The proportion of eligible patients who switched from warfarin to DOACs was significantly higher than that of the ineligible patients at the end of the study period (30.5% vs. 21.8%; P<0.0001).

Discussion

We examined the eligibility of Japanese patients with NVAF for phase III DOAC trials using a registry. Approximately 50% of the real-world registered patients with NVAF were ineligible following the inclusion and exclusion criteria derived from reports of the phase III trials. Further comparative analysis of outcomes clearly demonstrated significantly worse outcomes in the ineligible group; that is, they had a 1.5-fold increased risk of stroke even while receiving anticoagulant treatment, and the risk of major bleeding associated with anticoagulant use was 2-fold higher, raising concerns about safety. It should be emphasized that the risk of death was approximately 3-fold higher.

Eligibility

Our inclusion and exclusion criteria, which were derived from the most commonly used criteria of the 4 DOAC phase III trials, were considered valid for assessing eligibility and comparing outcomes between eligible and ineligible patients because the proportion of eligible patients using our criteria did not differ significantly from that using the criteria of the individual phase III trials (Supplementary Table).

In a UK general practice database study,⁸ 68% of patients were eligible for RE-LY, compared with 65% and 51% for ARISTOTLE and ROCKET-AF, respectively. A retrospective cross-sectional database analysis at the University Hospital Stroke Unit in Belgium¹¹ found that 47.6% of the patients were eligible for RE-LY, 45.5% for ARISTOTLE, and 39.3% for ROCKET-AF. A study of patients with a discharge diagnosis of atrial fibrillation in a large public hospital network in Melbourne, Australia, showed that 60.5%, 52.6%, and 35.8% of patients would have been eligible for the ARISTOTLE, RE-LY, and ROCKET-AF trials, respectively.¹² Of the patients with NVAF in the MAQI2 registry in Michigan, USA, 54.5% met the selection criteria used in RE-LY, 39.1% in ROCKET-AF, and 59.9% in ARISTOTLE.¹³ The reported proportion of patients eligible for DOAC phase III trials in real-world practice, including that in our results, was consistently approximately 50%.

The ineligible patients in the present study were mainly characterized by renal dysfunction, anemia, and chronic NSAID use, which were the exclusion criteria. These are well-known risk factors for major bleeding in patients with NVAF treated with anticoagulants. However, elderly patients with these risk factors are often encountered in real-world clinical practice, and anticoagulation therapy should be considered in the presence of atrial fibrillation. The studies cited above also reported a high risk of bleeding, poor renal function, and concomitant use of aspirin and antiplatelet agents as reasons for not enrolling patients in the phase III trials, despite the patients taking anticoagulants in actual practice.^12,13

Comparison of Outcomes Between Eligible and Ineligible Patients

We found that the risk of major bleeding, stroke and systemic embolism, and all-cause death in ineligible patients with a CHADS₂ score ≥2 was significantly higher than that in eligible patients. The poor prognosis of these ineligible patients may be explained by more comorbidities, such as anemia, heart failure, coronary artery disease, renal insufficiency, liver disease, history of cancer, or more concomitant medications, such as antiplatelet agents or NSAIDs, compared with those in eligible patients. Some of them are not necessarily variables that were included but are related to the exclusion criteria. In fact, a history of heart failure, cancer, or renal dysfunction, which were more prevalent in ineligible patients, was significantly associated with all-cause mortality in both groups. Also, it is known that NVAF patients with anemia are associated with a worse prognosis of death and heart failure.¹⁹ Therefore, phase III trial ineligibility can be considered a variable that summarizes these factors and indicates worse outcomes.

Generalizability of Results of Phase III Trials in Ineligible Patients

Based on the results from DOAC phase III trials showing their non-inferiority or superiority to warfarin with regard to stroke, major bleeding, and all-cause death, DOACs are the preferred anticoagulants in patients with NVAF. Efficacy and safety of DOACs have been established in Asian countries, including Japan.²⁰ The risks of ischemic stroke/systemic embolism and major bleeding are reduced in NVAF patients with CrCl ≥30 mL/min due to the widespread use of DOACs.²¹ Although such a generalization of the trial results may be valid for eligible patients, given the substantial differences in the characteristics and outcomes shown in this study, this is not easily extended to ineligible patients. Physicians are more likely to avoid DOACs in patients with characteristics of ineligible patients, such as those with reduced renal function.²² Our study showed that a significantly lower proportion of patients switched from warfarin to DOACs. Anticoagulant treatment in ineligible patients with NVAF, characterized by advanced age and multiple comorbidities, including reduced renal function, has not been well established;²³ this includes selection and doses, which may be unmet medical needs outside of a typical phase III trial. The contribution of anticoagulants themselves to improving prognosis is presumed to be limited in ineligible patients with a very high absolute risk of non-stroke-related death,²⁴ as we have shown.

Study Limitations

Our retrospective cohort study has certain limitations that are distinctive to this type of study. The data used in this cohort were initially obtained from medical records and not collected according to a study protocol. Furthermore, this study had a significant amount of missing data, particularly regarding relevant confounding factors, and several patients were lost to follow up.

The criteria developed from the 4 trials are generally considered valid; however, it is essential to note that we were unable to assess specific exclusion criteria, such as patients who were planning to undergo ablation or major surgery, or those with a life expectancy of <1 year at the time of registration. In addition, all patients in the cohort received warfarin at the time of registration, and some transitioned to DOACs. The effects of such transitions on the outcomes and their potential relationship with eligibility at registration were not evaluated.

Conclusions

Among patients with NVAF receiving anticoagulant therapy in real-world clinical practice, half were ineligible for phase III DOAC trials, mainly owing to safety concerns, and this population differs substantially from eligible patients in terms of characteristics and outcomes. The generalizability of the phase III results in real-world patients was uncertain and additional assessment for such patients is warranted.

Acknowledgments

The authors are grateful to Ms. Kaori Une for her assistance with data management, clinical research coordinators Ms. Yuko Fujita, Ms. Akane Kikuchi, Ms. Sayumi Mekaru, and Ms. Hitomi Zukeran for data collection from medical records, and Ms. Takako Okumura and Ms. Kayo Chinen for study management. We are also indebted to the data managers (Ms. Makiko Ohtorii, Ms. Ai Sunagawa, Ms. Kaori Yamamoto, Ms. Sachiko Kitamura, Ms. Hirono Saito, and Ms. Saeko Nagano) of the Institute for Clinical Effectiveness for their assistance with data management and statistical analyses.

Disclosures

S.U. received a research grant from Bristol Myers Squibb and a lecturer’s fee from Kowa. T.M. received lecturer fees from AstraZeneca, Bristol-Myers Squibb, Daiichi Sankyo, Japan Lifeline, Kowa, Toray, and Tsumura, manuscript fees from Bristol-Myers Squibb and Kowa, and serves on the advisory boards for Novartis and Teijin.

IRB Information

The ethics committees of all 71 participating centers approved this study, following the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan. The requirement for written informed consent was waived.

Data Availability

The deidentified participant data will be shared on a request basis. Please contact the corresponding author directly to request data sharing. The data includes all deidentified participant data with clinical information used in the analysis. Additional documents including the study protocol and statistical analysis plan will be available. The data will be available beginning 9 months and ending 36 months following article publication. Data will be shared with researchers who provide a methodologically sound proposal to achieve the aims presented in the approved proposal. Proposals should be directed to [e-mail]. To gain access, data requestors will need to sign a data access agreement.

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circrep.CR-25-0079

References

• 1.   Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: 1139–1151.
• 2.   Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: 883–891.
• 3.   Hori M, Matsumoto M, Tanahashi N, Momomura S, Uchiyama S, Goto S, et al. Rivaroxaban vs. warfarin in Japanese patients with atrial fibrillation: The J-ROCKET AF study. Circ J 2012; 76: 2104–2111.
• 4.   Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: 981–992.
• 5.   Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2013; 369: 2093–2104.
• 6.   U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER). Enhancing the diversity of clinical trial populations---Eligibility criteria, enrollment practices, and trial designs guidance for industry. 2020. https://digirepo.nlm.nih.gov/master/borndig/9918249008406676/9918249008406676.pdf (accessed February 8, 2022).
• 7.   ICH Steering Committee. ICH harmonised tripartite guideline: General considerations for clinical trials. 1997. https://www.pmda.go.jp/files/000156703.pdf (accessed February 8, 2022).
• 8.   Lee S, Monz BU, Clemens A, Brueckmann M, Lip GY. Representativeness of the dabigatran, apixaban and rivaroxaban clinical trial populations to real-world atrial fibrillation patients in the United Kingdom: A cross-sectional analysis using the General Practice Research Database. BMJ Open 2012; 2: e001768.
• 9.   Yoon CH, Park YK, Kim SJ, Lee MJ, Ryoo S, Kim GM, et al. Eligibility and preference of new oral anticoagulants in patients with atrial fibrillation: Comparison between patients with versus without stroke. Stroke 2014; 45: 2983–2988.
• 10.   Hägg L, Johansson C, Jansson JH, Johansson L. External validity of the Aristotle trial in real-life atrial fibrillation patients. Cardiovasc Ther 2014; 32: 214–218.
• 11.   Desmaele S, Steurbaut S, Cornu P, Brouns R, Dupont AG. Clinical trials with direct oral anticoagulants for stroke prevention in atrial fibrillation: How representative are they for real-life patients? Eur J Clin Pharmacol 2016; 72: 1125–1134.
• 12.   Fanning L, Ilomäki J, Bell JS, Dārziņš P. The representativeness of direct oral anticoagulant clinical trials to hospitalised patients with atrial fibrillation. Eur J Clin Pharmacol 2017; 73: 1427–1436.
• 13.   Hughey AB, Gu X, Haymart B, Kline-Rogers E, Almany S, Kozlowski J, et al. Warfarin for prevention of thromboembolism in atrial fibrillation: Comparison of patient characteristics and outcomes of the ‘Real-World’ Michigan Anticoagulation Quality Improvement Initiative (MAQI2) registry to the RE-LY, ROCKET-AF, and ARISTOTLE trials. J Thromb Thrombolysis 2018; 46: 316–324.
• 14.   Morimoto T, Uchida K, Sakakibara F, Kinjo N, Ueda S. Effect of concomitant antiplatelet therapy on ischemic and hemorrhagic events in patients taking oral anticoagulants for nonvalvular atrial fibrillation in daily clinical practice. Pharmacoepidemiol Drug Saf 2021; 30: 1321–1331.
• 15.   Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: Results from the National Registry of atrial fibrillation. JAMA 2001; 285: 2864–2870.
• 16.   Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJGM, Lip GYH. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: The Euro Heart Survey. Chest 2010; 138: 1093–1100.
• 17.   Schulman S, Kearon C, Subcommittee on Control of Anticoagulation of the Scientific, Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005; 3: 692–694.
• 18.   Chesebro JH, Knatterud G, Roberts R, Borer J, Cohen LS, Dalen J, et al. Thrombolysis in Myocardial Infarction (TIMI) Trial, Phase I: A comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Clinical findings through hospital discharge. Circulation 1987; 76: 142–154.
• 19.   Tsuda T, Hayashi K, Kato T, Kusayama T, Nakagawa Y, Nomura A, et al. Effects of longitudinal changes in anemia status on clinical outcomes in patients with non-valvular atrial fibrillation: Analysis from the Hokuriku-Plus AF Registry. Circ J 2025; 89: 164–173.
• 20.   Hara T, Sata M. Current real-world status of oral anticoagulant management in Japanese patients. Circ J 2025; 89: 150–152.
• 21.   Masunaga N, Ishii M, Oka K, Okamoto K, Yoshida Y, Minami K, et al. 10-Year trends of antithrombotic therapy status and clinical outcomes in patients with atrial fibrillation and renal dysfunction: The Fushimi AF Registry. Circ J 2025; 89: 174–183.
• 22.   Takagi Y, Ueda S. Factors related to the choice of warfarin for treating newly diagnosed nonvalvular atrial fibrillation are associated with safety outcomes during anticoagulation: A new-user, active-comparator, retrospective cohort study. J Arrhythm 2024; 40: 1408–1424.
• 23.   Bejjani A, Khairani CD, Assi A, Piazza G, Sadeghipour P, Talasaz AH, et al. When direct oral anticoagulants should not be standard treatment: JACC State-of-the-Art Review. J Am Coll Cardiol 2024; 83: 444–465.
• 24.   Gómez-Outes A, Lagunar-Ruíz J, Terleira-Fernández AI, Calvo-Rojas G, Suárez-Gea ML, Vargas-Castrillón E. Causes of death in anticoagulated patients with atrial fibrillation. J Am Coll Cardiol 2016; 68: 2508–2521.