Severity, Outcomes, and their Secular Changes in 33,870 Ischemic Stroke Patients with Atrial Fibrillation in a Hospital-Based Registry: Japan Stroke Data Bank

Abstract

Aim: Severity, functional outcomes, and their secular changes in acute atrial fibrillation (AF)-associated stroke patients were determined.

Methods: Acute ischemic stroke patients with AF in a hospital-based, multicenter, prospective registry from January-2000 through December-2020, were compared with those without AF. The co-primary outcomes were the initial severity assessed by the NIH Stroke Scale (NIHSS) score and favorable outcome assessed by the modified Rankin Scale scores 0-2 at hospital discharge.

Results: Of the 142,351 patients studied, 33,870 had AF. AF patients had higher NIHSS scores (median 9 vs. 3, adjusted coefficient 5.468, 95% CI 5.354-5.582) than non-AF patients. Favorable outcome was less common in AF patients than in non-AF patients in the unadjusted analysis (48.4% vs. 70.4%), but it was more common with adjustment for the NIHSS score and other factors (adjusted OR 1.110, 95% CI 1.061-1.161). In AF patients, the NIHSS score decreased throughout the 21-year period (adjusted coefficient -0.088, 95% CI -0.115 – -0.061 per year), and the reduction was steeper than in non-AF patients (P＜0.001). In AF patients, favorable outcome became more common over the period (adjusted OR 1.018, 95% CI 1.010-1.026), and the increase was steeper than in non-AF patients (P＜0.001); the increase was no longer significant after further adjustment by reperfusion therapy.

Conclusions: Initial stroke severity became milder and functional outcomes improved in AF patients over the 21-year period. These secular changes were steeper than in non-AF patients, suggesting that AF-associated stroke seemed to reap more benefit of recent development of stroke care than stroke without AF.

See editorial vol. 32: 304-305

Introduction

Atrial fibrillation (AF) increases the risk of stroke 5-fold¹⁾. The incidence of AF has increased with global aging, causing an increase in stroke occurrence in the elderly²⁾. In addition, stroke patients with AF show relatively severe neurological deficits and poor outcome^3-8). Thus, AF-associated stroke seems to be the most worrisome type of ischemic stroke. These years, development of oral anticoagulants (OACs), especially direct OACs (DOACs), has been effective for stroke prevention. Several nationwide registries in Europe reported a reduction of stroke occurrence in the late 2010s, probably due to increased OAC use^9-11), whereas a Danish study reported an increase in the proportion of AF-associated stroke¹²⁾. The severity of stroke was also reportedly decreased by prior anticoagulation^13-19). Acute reperfusion therapy, including intravenous thrombolysis and mechanical thrombectomy, is an established strategy for decreasing disability after ischemic stroke, especially embolic stroke^20-21). Possible long-term changes in initial neurological severity and independency after stroke in AF patients should be ascertained using the long-standing registry.

Recently, we reported the improvement of functional outcomes in ischemic stroke patients during the past 20 years using the nationwide registry named the Japan Stroke Data Bank (JSDB)²²⁾. Such yearly improvement may be especially evident for AF-associated stroke due to the above development of DOACs and reperfusion therapy.

Aims

The present study aimed to compare the initial severity, functional outcomes, and their secular changes in ischemic stroke patients with and without AF using the JSDB.

Methods

Study Design and Setting

The JSDB is an ongoing, hospital-based, multicenter, prospective registry of hospitalized patients with acute stroke or transient ischemic attack based on a web database from 130 stroke centers distributed evenly throughout Japan (Supplementary Table 1)^22-24). The unique aspects of this nationwide registry include standardized clinical information, detailed diagnosis, and acute management by stroke specialists. Patients’ data were prospectively recorded by the study physicians or clinical report coordinators in each institute using a standardized database form on the web. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline. The study protocol was approved by the institutional ethics board (National Cerebral and Cardiovascular Center Research Ethics Committee: M27-090-14). Due to the anonymous nature of the data, individual consent for entry into the database was waived by the institution. Instead, opt-out consent method was used.

Supplementary Table 1.List of the Japan Stroke Data Bank Investigators

Advisers	Shotai Kobayashi, MD; Kazuo Minematsu, MD
Steering committee	Kazunori Toyoda, MD (Chair); Fusao Igawa, MD; Koji Iihara, MD; Ryo Itabashi, MD; Kenji Kamiyama, MD; Takanari Kitazono, MD; Masatoshi Koga, MD; Yasuhiro Miyamoto, MD; Kuniaki Ogasawara, MD; Shigeru Nogawa, MD; Yoshiki Yagita, MD
Central office in the National Cerebral and Cardiovascular Center	Sohei Yoshimura, MD; Kaori Miwa, MD; Junpei Koge, MD; Akiko Ishigami, MD; Tomohide Yoshie, MD; Ai Ito (Department of Cerebrovascular Medicine) Yoshitaka Iwanaga, MD; Michikazu Nakai, PhD; Junji Miyazaki, MD; Yusuke Sasahara, PhD; Shinichi Wada, MD, Yoko Sumida (Department of Medical and Health Information Management)

Participants

For the present analyses, patients with ischemic stroke with known AF or AF diagnosed after stroke onset, who were registered into JSDB within 7 days after symptom onset from January 2000 through December 2020, were eligible for inclusion. Demographics, baseline characteristics, and therapeutic processes listed in Table 1 were collected from the database.

Table 1.Baseline characteristics and therapeutic processes

	Without atrial fibrillation	With atrial fibrillation	P
Number of patients	108481	33870
Female	41435 (38.2%)	15265 (45.1%)	＜0.001
Age, y	73 [64–81]	79 [72–85]	＜0.001
≥ 75 years old	49311 (45.5%)	22613 (66.8%)	＜0.001
Current smoking	19932 (18.4%)	3498 (10.3%)	＜0.001
Drinking ≥ 8 units/week	7039 (6.5%)	1832 (5.4%)	＜0.001
Family history of stroke	14039 (12.9%)	3777 (11.2%)	＜0.001
Medical history
Hypertension	75500 (69.6%)	23256 (68.7%)	0.088
Dyslipidemia	38495 (35.5%)	8206 (24.2%)	＜0.001
Diabetes mellitus	31198 (28.8%)	7454 (22.0%)	＜0.001
Cancer	4298 (4.0%)	1402 (4.1%)	0.160
History of stroke＊	28159 (26.0%)	10422 (30.8%)	＜0.001
Anticoagulation＊	4230 (3.9%)	10972 (32.4%)	＜0.001
Antiplatelet medication＊	25659 (23.7%)	7966 (23.5%)	0.613
Subtype of index stroke			＜0.001
Cardioembolism	12361(11.4%)	27744 (81.9%)
Large artery atherosclerosis	41803(38.5%)	2928 (8.6%)
Small vessel occlusion	37247 (34.3%)	2007(5.9%)
Other stroke	16775(15.5%)	1178(3.5%)
Acute reperfusion therapy
Intravenous thrombolysis alone	4115 (3.8%)	2864 (8.5%)	＜0.001
Endovascular recanalization alone	3001 (2.8%)	1750 (5.2%)	＜0.001
Both	1141 (1.1%)	1074 (3.2%)	＜0.001
Days of hospitalization	16 [10–28]	21 [12–37]	＜0.001

Number (%), median [interquartile range]

^＊Just prior to the index stroke

Outcomes

The co-primary outcome measures were the initial neurological severity at the emergent visit assessed by the National Institutes of Health Stroke Scale (NIHSS; range 0–42, with a higher score indicating more severe stroke) and favorable functional outcome assessed by modified Rankin Scale (mRS; range 0 [no symptoms] – 6 [death]) scores of 0–2 at hospital discharge. The secondary outcome was in-hospital death.

Statistical Analysis

Continuous data are reported as medians (interquartile range), and categorical data are presented as numbers (%). Mann-Whitney U and chi-squared tests were used to test the significance of differences between two groups for continuous and categorical variables, respectively.

To determine secular changes in the outcomes, multilevel mixed-effect regression and logistic regression using the institutes as random intercepts were performed for AF patients relative to non-AF patients. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were calculated. Multivariable analysis was performed by adjustment for 12 baseline characteristics listed in Table 1 (sex, age, current smoking, drinking, family history of stroke, hypertension, dyslipidemia, diabetes mellitus, cancer, history of stroke, prestroke anticoagulation, and prestroke antiplatelet medication: model 1), additionally for the index stroke severity (the initial NIHSS score: model 2), and furthermore for therapeutic process (receiving acute reperfusion therapy: model 3). The variables in the models were pre-specified.

Subgroup analyses for initial NIHSS score ≥ 10 and favorable functional outcome were performed on the basis of patients’ characteristics listed in Table 1. NIHSS score ≥ 10 was indicative of the presence of large-vessel occlusion^25-27).

To evaluate secular changes in the outcomes, multilevel mixed-effect multivariable regression and logistic regression using the institutes as random intercepts were performed. Significance was defined as a P value less than 0.05. Statistical analyses were performed with STATA ver16 (Stata Corp, College Station, TX, USA).

Results

Of a total of 154,440 ischemic stroke patients registered in the JSDB between January 2000 and December 2020, 3,414 registered ≥ 8 days after onset and 8,675 with unavailable data on demography or the presence of AF were excluded from the study (Supplementary Fig.1), and the remaining 142,351 patients (median age 75 years, women 39.8%) were studied. Of these, 33,870 had AF (79 years, women 45.1%), with non-valvular AF (NVAF) in 31,925 patients (94.3%) and valvular AF (VAF) in the other 1,945, and the remaining 108,481 did not have AF (73 years, women 38.2%).

Supplementary Fig.1.

Flow diagram

Baseline characteristics and the therapeutic processes of the patients are shown in Table 1. Patients with AF included more female patients, were older and less commonly had dyslipidaemia, diabetes, a current habit of smoking and drinking, and a family history of stroke, and more commonly had a history of stroke and anticoagulant medication just prior to the index stroke, and received either or both of intravenous thrombolysis and endovascular recanalization therapy after the index stroke, and had longer days of hospitalization than patients without AF (all P＜0.001). Baseline characteristics and the therapeutic processes of patients with NVAF and those with VAF are shown in Supplementary Table 2. There were several significant differences in the characteristics between the two AF groups, such as more female in patients with VAF.

Supplementary Table 2.Baseline characteristics and therapeutic processes of patients with atrial fibrillation

	Nonvalvular atrial fibrillation	Valvular atrial fibrillation
Number of patients	31925	1945
Female‡	14149 (44.3%)	1116 (57.4%)
Age, y	79 [72–85]	79 [72–85]
≥ 75 years old	21306 (66.7%)	1307 (67.2%)
Current smoking‡	3334 (10.4%)	164 (8.4%)
Drinking ≥ 8 units/week‡	1751 (5.5%)	81 (4.2%)
Family history of stroke	3562 (11.2%)	215 (11.1%)
Medical history
Hypertension	21942 (68.7%)	1314 (67.6%)
Dyslipidemia	7716 (24.2%)	490 (25.2%)
Diabetes mellitus‡	7083 (22.2%)	371 (19.1%)
Cancer	1335 (4.2%)	67 (3.4%)
History of stroke＊	9818 (30.8%)	604 (31.1%)
Anticoagulation＊‡	10052 (31.5%)	920 (47.3%)
Antiplatelet medication＊†	7501 (23.5%)	465 (23.9%)
Acute reperfusion therapy†	5405 (16.9%)	283 (14.6%)
Intravenous thrombolysis alone	2726 (8.5%)	138 (7.1%)
Endovascular recanalization	1653 (5.2%)	97 (5.0%)
Both	1026 (3.2%)	48 (2.5%)
Days of hospitalization	21 [12–37]	20 [12–38]

Number (%), median [interquartile range]

^＊Just prior to the index stroke

^†P＜0.05, ^‡P＜0.01, ^§P＜0.001 between the group

Initial NIHSS Scores

The AF group had a median overall NIHSS score of 9 (IQR 3–20); the non-AF group had the score of 3 (1–7) (Fig.1). The AF group had higher NIHSS scores than the non-AF group in the crude analysis of all participants and for any subgroups of sex, age, and 5-year categories of registration (all P＜0.001). The AF group also had higher NIHSS scores than the non-AF group after adjustment for 12 baseline characteristics (model 1, coefficient 5.468, 95% CI 5.354–5.582, Table 2). Both patients with NVAF and patients with VAF had higher NIHSS scores than the non-AF group in model 1 (Table 2).

Fig.1. National Institutes of Health Stroke Scale scores at the emergent visit

Boxes represent the interquartile range. Lines across the boxes indicate median values.

Whiskers represent 10th percentile and 90th percentile values.

P＜0.001 for all pairs of atrial fibrillation (AF) (−) and AF (+). P=0.266 between non-valvular AF (NVAF) and valvular AF (VAF).

Table 2.Outcomes

	Without AF	With AF	Crude	Model 1	Model 2	Model 3
NIHSS score	3 [1-7]	9 [3-20]	6.364 (6.259-6.469)	5.468 (5.354-5.582)	–	–
Favorable outcome (mRS scores 0–2)	60,617 (70.4%)	11,854 (48.4%)	0.401 (0.389-0.413)	0.495 (0.478, 0.512)	1.133 (1.083,1.185)	1.110 (1.061, 1.161)
In-hospital death	2,996 (2.8%)	3,294 (9.8%)	3.745 (3.556-3.943)	2.857 (2.699, 3.026)	1.415 (1.323, 1.514)	1.436 (1.342, 1.537)
	Without AF	With NVAF	Crude	Model 1	Model 2	Model 3
NIHSS score	3 [1–7]	9 [3–20]	6.353 (6.246,6.459)	5.459 (5.343, 5.574)	–	–
Favorable outcome	60,617 (70.4%)	11,256 (48.5%)	0.402 (0.390,0.414)	0.496 (0.479,0.514)	1.138 (1.087, 1.192)	1.115 (1.065,1.168)
In-hospital death	2,996 (2.8%)	3,036 (9.6%)	3.666 (3.478,3.865)	2.804 (2.645,2.972)	1.380 (1.288,1.478)	1.399 (1.306,1.499)
	Without AF	With VAF	Crude	Model 1	Model 2	Model 3
NIHSS score	3 [1–7]	9 [3–20]	6.612 (6.273,6.951)	5.160 (4.814,5.505)	–	–
Favorable outcome	60,617 (70.4%)	598 (47.1%)	0.381 (0.340,0.427)	0.518 (0.459,0.585)	1.101 (0.933,1.299)	1.095 (0.938,1.293)
In-hospital death	2,996 (2.8%)	258 (13.4%)	5.103 (4.441,5.863)	3.304 (2.834,3.850)	1.902 (1.589,2.279)	1.909 (1.593, 2.287)

Median [IQR] and coefficient (95% CI) for NIHSS score.

Number (percent) and OR (95% CI) for favorable outcome and in-hospital death.

Model 1: adjusted by sex, age, current smoking, drinking, family history of stroke, hypertension, dyslipidemia, diabetes mellitus, cancer, history of stroke, prestroke anticoagulation, and prestroke antiplatelets

Model 2: adjusted by NIHSS score and variables used in Model 1

Model 3: adjusted by acute reperfusion therapy and variables used in Model 2

86,055 patients without AF and 24,471 patients with AF [23,202 with non-valvular AF (NVAF) and 1,269 with valvular AF (VAF)] both with prestroke mRS scores 0–2 were included in the analysis of favorable outcome.

Severe stroke, corresponding to NIHSS score ≥ 10, was identified in 16,393 patients with AF (48.4%) and 18,550 patients without AF (17.1%, aOR 3.799, 95% CI 3.689–3.913). Fig.2 shows the results of subgroup analysis for NIHSS score ≥ 10. There was a significant AF-by-subgroup interaction between subgroups of sex, age, hypertension, history of cancer, history of stroke, and prestroke anticoagulation, and prestroke antiplatelet medication. Of these, prestroke anticoagulation showed an especially evident AF-by-subgroup interaction (aOR: 2.045 in patients with anticoagulation vs. 4.187 in those without).

Fig.2. Initial National Institutes of Health Stroke Scale scores ≥ 10 by subgroup

Adjusted by sex and age.

Favorable Functional Outcome Assessed by mRS Scores of 0–2 at Discharge

Fig.3 shows the mRS scores at discharge. The distribution of the scores differed between patients with and without AF and between patients with NVAF and those with VAF (both P＜0.001). The differences in the distributions between patients with and without AF were also significant in the subgroups depending on sex, age, and 5-year categories of registration (all P＜0.001).

Fig.3. Modified Rankin Scale scores at discharge

P＜0.001 for all the pairs of atrial fibrillation (AF) (−) and AF (+). P＜0.001 between non-valvular AF (NVAF) and valvular AF (VAF).

Favorable outcome was identified in 11,854 patients with AF (48.4%) and 60,617 patients without AF (70.4%) when 110,526 patients with prestroke mRS scores of 0–2 were analyzed. After adjustment for 12 baseline characteristics (model 1), favorable outcome was less common in the AF group than in the non-AF group (aOR 0.495, 95% CI 0.478–0.512, Table 2). However, favorable outcome was more common in the AF group than in the non-AF group after further adjustment for NIHSS score (model 2) and further adjustment for acute reperfusion therapy (model 3) (aOR 1.110, 95% CI 1.061–1.161 in model 3). Patients with NVAF showed similar results relative to patients without AF (Table 2).

Fig.4 shows the results of the subgroup analysis for favorable outcome. There was a significant AF-by-subgroup interaction between subgroups of age, initial NIHSS scores, and acute reperfusion therapy.

Fig.4. Favorable outcomes (mRS scores 0–2) at discharge by subgroup

110,526 patients with prestroke mRS scores 0–2 were only included. Adjusted by sex, age, and the NIHSS score.

In-Hospital Death

In-hospital death was more common in the AF group (9.8%) than in the non-AF group (2.8%) on crude analysis and in models 1–3 (aOR 1.436, 95% CI 1.342–1.537 in model 3, Table 2). Both patients with NVAF and patients with VAF showed similar results relative to patients with NVAF (Table 2).

Secular Changes

Supplementary Fig.2 shows the percentage of patients with AF to all IS patients; it was stable at ≈24% until 2016 but was between 21% and 23% thereafter.

Supplementary Fig.2.

Percentage of patients having atrial fibrillation to all ischemic stroke patients

Table 3 shows secular changes in outcomes. In model 1, the NIHSS score decreased significantly in both the AF group (adjusted coefficient -0.088, 95% CI -0.115 – -0.061 per year) and the non-AF group (-0.039, -0.049 – -0.029). There was a significant AF-by-year interaction; the AF group showed a steeper secular reduction than the non-AF group. Both the NVAF and VAF groups showed similar results.

Table 3.Secular changes in outcomes

	Crude	Model 1	Model 2	Model 3
NIHSS score	Coefficient per 1 year, (95% confidence interval)
Without AF	-0.023 (-0.033 – 0.013)	-0.039 (-0.049,-0.029)	–	–
With AF	-0.052 (-0.078 – -0.025)‡	-0.088 (-0.115, -0.061)‡	–	–
With nonvalvular AF	-0.051 (-0.078 – -0.024)	-0.088 (-0.116, -0.061)	–	–
With valvular AF	-0.056 (-0.142 – 0.031)	-0.075 (-0.166, 0.016)	–	–
Favorable outcome	OR per 1 year (95% confidence interval)
Without AF	1.006 (1.003-1.010)	1.014 (1.011, 1.018)	1.005 (1.001,1.009)	1.002 (0.997, 1.006)
With AF	1.011 (1.006-1.017)†	1.025 (1.019, 1.031)‡	1.018 (1.010, 1.026)‡	1.006 (0.998, 1.014)
With nonvalvular AF	1.012 (1.006, 1.018)	1.026 (1.019, 1.032)	1.019 (1.011, 1.026)	1.007 (0.999, 1.015)
With valvular AF	0.994 (0.973, 1.016)	1.010 (0.986, 1.034)	1.002 (0.972, 1.033)	0.995 (0.964, 1.026)
In-hospital death	OR per 1 year (95% confidence interval)
Without AF	0.973 (0.966-0.981)	0.957 (0.950, 0.965)	0.963 (0.954, 0.972)	0.963 (0.953, 0.972)
With AF	0.972 (0.965-0.980)	0.958 (0.950, 0.966)	0.964 (0.955-0.973)＊	0.969 (0.960-0.979)＊

Model 1: adjusted by sex, age, current smoking, drinking, family history of stroke, hypertension, dyslipidemia, diabetes mellitus, cancer, history of stroke, prestroke anticoagulation, and prestroke antiplatelets

Model 2: adjusted by NIHSS score and variables used in Model 1

Model 3: adjusted by acute reperfusion therapy and variables used in Model 2

110,526 patients with prestroke mRS scores 0–2 were only included in the analysis of favorable outcome.

^＊P-interaction ＜0.05, ^†＜0.01, ^‡＜0.001 compared to “without AF”

Both the AF and non-AF groups showing favorable outcome increased significantly in models 1 and 2 (aOR 1.018, 95% CI 1.010–1.026 in model 2 in the AF group), but not in model 3 (Table 3). The AF group showed a steeper secular increase than the non-AF group in models 1 and 2 (P-interaction ＜0.001). In the AF group, patients receiving acute endovascular recanalization therapy showed a steeper secular increase in the proportion of favorable outcome than those without receiving the therapy (aOR 1.051, 95% CI 1.020–1.083 versus 1.004, 0.996–1.012 in model 2).

The AF and non-AF groups who died decreased significantly in models 1–3 (aOR 0.969, 95% CI 0.960–0.979 in model 3 for the AF group), and secular reduction was steeper in the AF group in models 2 and 3. In the AF group, patients with endovascular therapy showed a steeper secular reduction in the proportion of death than those without (aOR 0.931, 95% CI 0.902–0.961 versus 0.970, 0.960-0.981 in model 2).

Discussion

A total of 33,870 ischemic stroke patients with AF were compared with 108,481 ischemic stroke patients without AF using the 21-year-long nationwide JSDB registry. The first major finding of this study was that the initial NIHSS score was higher in patients with AF than those without AF with any adjustment. Second, favorable functional outcome was less common in the AF group than in the non-AF group in the crude analysis but was more common with multivariable adjustment including the initial NIHSS score. Third, in-hospital death was more common in the AF group with multivariable adjustment. Fourth, both the NIHSS score and in-hospital death decreased over the 21-year period in both the AF and non-AF groups with a significant AF-by-year interaction; the secular reduction was steeper in the AF group for all of these outcomes. In contrast, favorable outcome increased over time in the AF group, and the increase was steeper than in the non-AF group. However, the increase was no longer significant after further adjustment by receiving reperfusion therapy. To the best of our knowledge, this is the largest registry study that determined stroke severity and outcome with their secular change in ischemic stroke patients having AF.

Several registry studies reported that patients with AF developed more severe ischemic strokes than patients without AF; higher age, larger infarcts, and more common hemorrhagic transformation were often indicated as the reason for stroke severity in patients with AF^3-8). Greater volumes of more severe baseline hypoperfusion in patients with AF seemed to cause higher NIHSS scores in the combined Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET) and the Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution (DEFUSE) dataset²⁸⁾. Most subgroups in the present analysis showed a similar impact compared to each counterpart subgroup on the frequency of severe stroke in patients with AF than in those without AF, although some showed a significant AF-by-subgroup interaction. Exceptionally, anticoagulated patients showed a much smaller aOR for severe stroke than non-anticoagulated patients. Severity of stroke was reportedly weakened by prior anticoagulation, especially by DOAC^13-19). As a reason, well-controlled anticoagulation might decrease the size of cardiac thrombi. Otherwise, non-cardioembolic stroke might be more common in the well-anticoagulated population even with AF.

Initial stroke severity is strongly related to functional outcome. Thus, a poor functional outcome and high mortality after AF-associated stroke were repeatedly reported^3-8). Most previous studies did not examine stroke outcome adjusted by the initial severity. In a study from the Canadian Stroke Network, the associations of AF with discharge mRS scores 4–5 and 30-day and 1-year mortality were substantially reduced, but not reversed, by additionally adjusting for the initial Canadian Neurological Scale score⁶⁾. In a study from the Danish Stroke Registry, the association of AF with 30-day mortality was insignificant with additional adjustment by the categorized initial stroke severity⁸⁾. In the present analysis, in-hospital mortality was still significantly associated with AF with additional adjustment by the initial NIHSS score, although the association was much reduced. In contrast, the associations of AF with favorable outcome were totally reversed by adjusting by the NIHSS score, partly due to a very strong confounding effect between NIHSS and mRS scores. More severe stroke patients had higher chance to receive thrombolysis or endovascular reperfusion therapy that could bring drastic recovery. In fact, this reverse association was slightly weakened by further adjusting by reperfusion therapy. Thus, patients with AF might have a higher chance to have independency after stroke than patients without AF following acute care for stroke. Bleeding complications might exacerbate clinical outcome. In the present study, major bleeding occurred in 0.16% of patients without AF and 0.32% of patients with AF (crude OR 2.020, 95% CI 1.586-2.573, aOR 2.314, 96% CI 1.722-3.109 after adjustment by prestroke anticoagulation and reperfusion therapy, data not shown in Results). Thus, major bleeding was a cause of poorer outcome in patients with AF that cannot be overlooked.

The lower percentage of patients with AF to all IS patients after 2017 as compared to up to 2016 in the present cohort might be partly caused by a reduction of stroke occurrence in the AF population probably due to increased OAC use^9-11). The Athens Stroke Registry showed an interesting finding, that the proportion of newly diagnosed AF among all AF-associated stroke patients increased annually²⁹⁾. A secular decrease in the initial NIHSS score and a secular increase in the proportion of favorable outcome in the overall ischemic stroke patients were already reported from the JSDB²²⁾. We had discussed that the former reduction was due to improvement in preventive therapy and changes in case mix (for example, increasing chance of detecting minor stroke by development of brain imaging modalities), and the latter increase was due to development of acute care, such as reperfusion therapy²²⁾. A newly added finding in the present study was that both the rates of the former reduction and the latter increase were higher in patients with AF than in those without. This suggests that preventive therapy was more evidently developed in patients with AF, for example, by the spread of DOACs, and acute stroke care was more effective in patients with AF than in those without. Trends of a secular increase in the proportion of favorable outcome disappeared after adjustment by reperfusion therapy, suggesting that implementation of reperfusion therapy itself or ambulance transport to reperfusion therapy-capable hospitals partly caused the improvement of stroke outcome. A secular increase in the proportion of favorable outcome was clearer in AF patients with endovascular reperfusion therapy than those without. Since initial devices for the therapy were officially approved in 2010, potent stent retrievers were approved in 2014, and the therapy gradually spread thereafter, the benefit of the therapy to functional outcome would be larger with time. In the Athens Stroke Registry study, all of the recurrent stroke, cardiovascular events, and death were annually reduced in ischemic stroke patients with AF²⁹⁾.

The strengths of the present study were its duration and the population size of the database. The JSDB belongs to the oldest group that commenced patient registration around 2000 among 28 national registries listed in a systematic review³⁰⁾. Web-based data accumulation from the beginning with thorough cleansing has enabled accurate and well-preserved data. The present study included an overwhelmingly larger number of AF-associated stroke patients (33,870) than the other studies with the same theme; few of those exceeded 10,000 as the number of enrolled patients⁸⁾. AF detected during acute hospitalization was fully documented, although AF detected after hospital discharge was not documented.

The limitations of the JSDB have been described elsewhere²²⁾. Briefly, they included: estimated registration of only ≈3% (recently ≈6%) of stroke patients across Japan; changes in concepts of some of the collected data according to the currents of the times over the 20-year period; and the lack of collection of longer-term outcomes after discharge. In addition, 8,657 patients (5.7%) were excluded due to unavailable data on demographics or the presence of AF. Another limitation of this sub-analysis was that patients with known AF prior to stroke onset could not be distinguished from those with AF diagnosed after onset in our dataset³¹⁾.

In conclusion, patients with AF more commonly had favorable clinical outcome after ischemic stroke than patients without AF with adjustment by the NIHSS score and other factors. Initial stroke severity became milder and functional outcomes improved in patients with AF during the past 21 years, and these changes were more evident than in patients without AF, unless receiving acute reperfusion therapy was added as an adjusting factor. These results support our hypothesis that development of stroke prevention using OACs and acute care by reperfusion therapy weakened the initial severity and improved clinical outcomes after AF-associated stroke.

Acknowledgement

The authors thank Ms. Ai Ito for secretary assistant.

Study Funding

Supported by the Japan Agency for Medical Research and Development (AMED: JP24lk0221186, JP24lk0221171) and JSPS KAKENHI (JP21K07472, JP 23K27522).

Conflicts of Interest

All of the following conflicts are outside the submitted work. Toyoda reports honoraria from Daiichi-Sankyo, Otsuka, Janssen, Bayer Yakuhin, and Bristol-Myers-Squibb,. Kamiyama reports honoraria from Daiichi-Sankyo and Bristol-Myers-Squibb. Miyamoto reports honoraria from Kowa, Meiji Yasuda Research Institute, Softbank, Tokio Marine & Nichido Fire Insurance, Saraya and Bristol-Myers Squibb. Minematsu reports honoraria from Bayer and Pfizer. Koga reports honoraria from Bayer Yakuhin, and Daiichi-Sankyo and research supports from Daiichi-Sankyo, NBI, and Shionogi. None of the other authors have any conflicts of interest to declare.

References

• 1)  Wolf PA, Abbott RD, and Kannel WB: Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke, 1991; 22: 983-988
• 2)  Di Carlo A, Bellino L, Consoli D, Mori F, Zaninelli A, Baldereschi M, Cattarinussi A, D'Alfonso MG, Gradia C, Sgherzi B, Pracucci G, Piccardi B, Polizzi B, and Inzitari D; National Research Program: Progetto FAI. La Fibrillazione Atriale in Italia: Prevalence of atrial fibrillation in the Italian elderly population and projections from 2020 to 2060 for Italy and the European Union: the FAI Project. Europace, 2019; 21: 1468-1475
• 3)  Lin HJ, Wolf PA, Kelly-Hayes M, Beiser AS, Kase CS, Benjamin EJ, and D'Agostino RB: Stroke severity in atrial fibrillation. The Framingham Study. Stroke, 1996; 27: 1760-1764
• 4)  Steger C, Pratter A, Martinek-Bregel M, Avanzini M, Valentin A, Slany J, and Stöllberger C: Stroke patients with atrial fibrillation have a worse prognosis than patients without: data from the Austrian Stroke registry. Eur Heart J, 2004; 25: 1734-1740
• 5)  Kimura K, Minematsu K, and Yamaguchi T: Atrial fibrillation as a predictive factor for severe stroke and early death in 15,831 patients with acute ischaemic stroke. J Neurol Neurosurg Psychiatry, 2005; 76: 679-683
• 6)  McGrath ER, Kapral MK, Fang J, Eikelboom JW, Conghaile Aó, Canavan M, and O'Donnell MJ; Investigators of the Ontario Stroke Registry: Association of atrial fibrillation with mortality and disability after ischemic stroke. Neurology, 2013; 81: 825-832
• 7)  Saposnik G, Gladstone D, Raptis R, Zhou L, and Hart RG: Atrial fibrillation in ischemic stroke: predicting response to thrombolysis and clinical outcomes. Stroke, 2013; 44: 99-104
• 8)  Vinding NE, Kristensen SL, Rørth R, Butt JH, Østergaard L, Olesen JB, Torp-Pedersen C, Gislason GH, Køber L, Kruuse C, Johnsen SP, and Fosbøl EL: Ischemic Stroke Severity and Mortality in Patients With and Without Atrial Fibrillation. J Am Heart Assoc, 2022; 11: e022638
• 9)  Cowan JC, Wu J, Hall M, Orlowski A, West RM, and Gale CP: A 10 year study of hospitalized atrial fibrillation-related stroke in England and its association with uptake of oral anticoagulation. Eur Heart J, 2018; 39: 2975-2983
• 10)  Forslund T, Komen JJ, Andersen M, Wettermark B, von Euler M, Mantel-Teeuwisse AK, Braunschweig F, and Hjemdahl P: Improved stroke prevention in atrial fibrillation after the introduction of non-Vitamin K antagonist oral anticoagulants. Stroke, 2018. 49: 2122-2128
• 11)  Díaz-Guzmán J, Freixa-Pamias R, García-Alegría J, Pérez Cabeza AI, Roldán-Rabadán I, Antolin-Fontes B, Rebollo P, Llorac A, Genís-Gironés M, and Escobar-Cervantes C: Epidemiology of atrial fibrillation-related ischemic stroke and its association with DOAC uptake in Spain: first national population-based study 2005 to 2018. Rev Esp Cardiol (Engl Ed), 2022; 75: 496-505
• 12)  Larsen LK, Ovesen C, Frost L, Damgaard D, Christensen H, and Johnsen SP: Trends in preadmission oral anticoagulant use and clinical outcome in atrial fibrillation patients admitted with acute stroke in Denmark. Eur Heart J Qual Care Clin Outcomes, 2020; 6: 112-120
• 13)  Hylek EM, Go AS, Chang Y, Jensvold NG, Henault LE, Selby JV, and Singer DE: Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation. N Engl J Med, 2003; 349: 1019-1026
• 14)  O’Donnell M, Oczkowski W, Fang J, Kearon C, Silva J, Bradley C, Guyatt G, Gould L, D'Uva C, Kapral M, and Silver F; Investigators of the Registry of the Canadian Stroke Network: Preadmission antithrombotic treatment and stroke severity in patients with atrial fibrillation and acute ischaemic stroke: an observational study. Lancet Neurol, 2006; 5: 749-754
• 15)  Johnsen SP, Svendsen ML, Hansen ML, Brandes A, Mehnert F, and Husted SE: Preadmission oral anticoagulant treatment and clinical outcome among patients hospitalized with acute stroke and atrial fibrillation: a nationwide study. Stroke, 2014; 45: 168-175
• 16)  Xian Y, O'Brien EC, Liang L, Xu H, Schwamm LH, Fonarow GC, Bhatt DL, Smith EE, Olson DM, Maisch L, Hannah D, Lindholm B, Lytle BL, Pencina MJ, Hernandez AF, and Peterson ED: Association of preceding antithrombotic treatment with acute ischemic stroke severity and in-hospital outcomes among patients with atrial fibrillation. JAMA, 2017; 317: 1057-1067
• 17)  Macha K, Marsch A, Siedler G, Breuer L, Strasser EF, Engelhorn T, Schwab S, and Kallmünzer B: Cerebral ischemia in patients on direct oral anticoagulants. Stroke, 2019; 50: 873-879
• 18)  Jung YH, Kim YD, Kim J, Han SW, Oh MS, Lee JS, and Lee KY: Initial stroke severity in patients with atrial fibrillation according to antithrombotic therapy before ischemic stroke. Stroke, 2020; 51: 2733-2741
• 19)  Meinel TR, Branca M, De Marchis GM, Nedeltchev K, Kahles T, Bonati L, Arnold M, Heldner MR, Jung S, Carrera E, Dirren E, Michel P, Strambo D, Cereda CW, Bianco G, Kägi G, Vehoff J, Katan M, Bolognese M, Backhaus R, Salmen S, Albert S, Medlin F, Berger C, Schelosky L, Renaud S, Niederhauser J, Bonvin C, Schaerer M, Mono ML, Rodic B, Tarnutzer AA, Mordasini P, Gralla J, Kaesmacher J, Engelter S, Fischer U, and Seiffge DJ; Investigators of the Swiss Stroke Registry: Prior anticoagulation in patients with ischemic stroke and atrial fibrillation. Ann Neurol, 2021; 89: 42-53
• 20)  Emberson J, Lees KR, Lyden P, Blackwell L, Albers G, Bluhmki E, Brott T, Cohen G, Davis S, Donnan G, Grotta J, Howard G, Kaste M, Koga M, von Kummer R, Lansberg M, Lindley RI, Murray G, Olivot JM, Parsons M, Tilley B, Toni D, Toyoda K, Wahlgren N, Wardlaw J, Whiteley W, del Zoppo GJ, Baigent C, Sandercock P, and Hacke W; Stroke Thrombolysis Trialists' Collaborative Group: Effect of treatment delay, age, and stroke severity on the effects of intravenous thrombolysis with alteplase for acute ischaemic stroke: a meta-analysis of individual patient data from randomised trials. Lancet, 2014; 384: 1929-1935
• 21)  Goyal M, Menon BK, van Zwam WH, Dippel DW, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CB, van der Lugt A, de Miquel MA, Donnan GA, Roos YB, Bonafe A, Jahan R, Diener HC, van den Berg LA, Levy EI, Berkhemer OA, Pereira VM, Rempel J, Millán M, Davis SM, Roy D, Thornton J, Román LS, Ribó M, Beumer D, Stouch B, Brown S, Campbell BC, van Oostenbrugge RJ, Saver JL, Hill MD, and Jovin TG; HERMES collaborators: Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet, 2016; 387: 1723-1731
• 22)  Toyoda K, Yoshimura S, Nakai M, Koga M, Sasahara Y, Sonoda K, Kamiyama K, Yazawa Y, Kawada S, Sasaki M, Terasaki T, Miwa K, Koge J, Ishigami A, Wada S, Iwanaga Y, Miyamoto Y, Minematsu K, and Kobayashi S; Japan Stroke Data Bank Investigators: Twenty-Year Change in Severity and Outcome of Ischemic and Hemorrhagic Strokes. JAMA Neurol, 2022; 79: 61-69
• 23)  Ikawa F, Ichihara N, Uno M, Shiokawa Y, Toyoda K, Minematsu K, Kobayashi S, Yamaguchi S, and Kurisu K; Japanese Standard Stroke Registry Study Group: Visualisation of the non-linear correlation between age and poor outcome in patients with aneurysmal subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry, 2021; 92: 1173-1180
• 24)  Miwa K, Koga M, Nakai M, Yoshimura S, Sasahara Y, Koge J, Sonoda K, Ishigami A, Iwanaga Y, Miyamoto Y, Kobayashi S, Minematsu K, and Toyoda K; Japan Stroke Data Bank Investigators: Etiology and outcome of ischemic stroke in patients with renal impairment including chronic kidney disease: Japan Stroke Data Bank. Neurology, 2022; 98: e1738-e1747
• 25)  Nakajima M, Kimura K, Ogata T, Takada T, Uchino M, and Minematsu K: Relationships between angiographic findings and National Institutes of Health Stroke Scale score in cases of hyperacute carotid ischemic stroke. AJNR Am J Neuroradiol, 2004; 25: 238-241
• 26)  Fischer U, Arnold M, Nedeltchev K, Brekenfeld C, Ballinari P, Remonda L, Schroth G, and Mattle HP: NIHSS score and arteriographic findings in acute ischemic stroke. Stroke, 2005; 36: 2121-2125
• 27)  Broderick JP, Palesch YY, Demchuk AM, Yeatts SD, Khatri P, Hill MD, Jauch EC, Jovin TG, Yan B, Silver FL, von Kummer R, Molina CA, Demaerschalk BM, Budzik R, Clark WM, Zaidat OO, Malisch TW, Goyal M, Schonewille WJ, Mazighi M, Engelter ST, Anderson C, Spilker J, Carrozzella J, Ryckborst KJ, Janis LS, Martin RH, Foster LD, and Tomsick TA; Interventional Management of Stroke (IMS) III Investigators: Endovascular therapy after intravenous t-PA versus t-PA alone for stroke. N Engl J Med, 2013; 368: 893-903
• 28)  Tu HT, Campbell BC, Christensen S, Desmond PM, De Silva DA, Parsons MW, Churilov L, Lansberg MG, Mlynash M, Olivot JM, Straka M, Bammer R, Albers GW, Donnan GA, and Davis SM; EPITHET-DEFUSE Investigators: Worse stroke outcome in atrial fibrillation is explained by more severe hypoperfusion, infarct growth, and hemorrhagic transformation. Int J Stroke, 2015; 10: 534-540
• 29)  Ntaios G, Sagris D, Gioulekas F, Galanis P, Pardali C, Vemmou A, Koroboki E, Papavasileiou V, Vassilopoulou S, Manios E, Makaritsis K, Spengos K, Mitsikostas DD, Milionis H, and Vemmos K: 20-year trends of characteristics and outcomes of stroke patients with atrial fibrillation. Int J Stroke, 2018; 13: 707-716
• 30)  Cadilhac DA, Kim J, Lannin NA, Kapral MK, Schwamm LH, Dennis MS, Norrving B, and Meretoja A: National stroke registries for monitoring and improving the quality of hospital care: A systematic review. Int J Stroke, 2016; 11: 28-40
• 31)  Tokunaga K, Toyoda K, Kimura S, Minematsu K, Yasaka M, Okada Y, Yoshimura S, and Koga M; SAMURAI-NVAF and RELAXED investigators: Association of the timing of atrial fibrillation detection and insular involvement with the risk of embolic events after acute ischemic stroke. Ann Neurol, 2024; 95: 338-346