Association of statin therapy on acute ischemic stroke patients with atrial fibrillation: insights from a nationwide cohort study

Abstract

Although statins are well-established in improving outcomes after ischemic stroke, their effectiveness in patients with acute ischemic stroke (AIS) and atrial fibrillation (AF) without conventional indications remains uncertain. This study investigates the association between statin therapy and vascular outcomes in patients with AIS and AF without previously diagnosed atherosclerotic cardiovascular disease (ASCVD). This retrospective, nationwide cohort study utilized the Korean National Health Insurance Service database (2011–2023). Patients aged ≥ 20 years with AIS or transient ischemic attack and concurrent AF were included if they underwent neuroimaging, had no prior statin use at admission, and no documented history of ASCVD. Propensity score matching and inverse probability of treatment weighting (IPTW) were used to adjust for potential confounders. The primary outcome was a composite of all-cause death, ischemic stroke or systemic embolism, intracranial hemorrhage (ICH), and myocardial infarction within 1 year. Hazard ratios (HR) were estimated using Cox regression. Subgroup analyses assessed the association of statin dose (high vs. standard), ezetimibe combination therapy, and statin type. A total of 64,190 patients (mean age; 73.76, male 55.3%) were finally analyzed and categorized into statin users (n = 37,033) and non-users (n = 27,157). In the IPTW analysis, statin-user was associated with lower risk for primary vascular outcomes (HR 0.821 [95% CI 0.808–0.835]), mortality (HR 0.746 [0.729–0.763]), ischemic stroke and systemic embolism (HR 0.904 [0.884–0.925]), and ICH (HR 0.725 [0.660–0.796]) compared to non-users. Neither statin dose, statin type, nor ezetimibe combination was significantly associated with differences in the primary outcome. In patients with AIS and AF without previously diagnosed ASCVD, statin therapy was associated with a lower risk of major vascular events. These findings suggest a potential benefit of statins in this population and highlight the need for confirmation through prospective randomized trials.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-40042-3.

Introduction

Acute ischemic stroke (AIS) is a leading cause of death and disability worldwide, and approximately 20% of AIS cases are associated with atrial fibrillation (AF). Patients with AIS and AF are at higher risk of mortality, recurrence, and severe functional impairment, making them a critical population for therapeutic intervention and prognostic management^1–3.

Statins are known to exert broad cardiovascular protective effects not only by lowering low-density lipoprotein cholesterol (LDL-C) and inhibiting the progression of atherosclerosis, but also through their pleiotropic effects, including anti-inflammatory, antithrombotic, endothelial function-enhancing, and vascular-stabilizing properties^4,5. The SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels) trial demonstrated that high-intensity statin therapy reduced the risk of recurrent stroke and coronary events in patients with AIS⁶, while the TST (Treat Stroke to Target) trial showed similar benefits through LDL-C target achievement⁷.

Although AF can be considered pathophysiologically distinct from atherosclerosis⁸, previous studies have shown that AF is frequently accompanied by atherosclerotic conditions such as coronary artery disease and carotid artery stenosis^9,10. These findings imply that atherosclerosis may aggravate vascular risk in AF, supporting the potential benefit of statins in stroke patients with both conditions.

Although several observational studies have suggested that statin therapy may reduce mortality and vascular events in patients with cardioembolic stroke, most were limited by their non-randomized design, potential confounding, and heterogeneous study populations^11–15. In particular, few studies have clearly defined stroke related to atrial fibrillation or adequately accounted for the role of concomitant atherosclerosis¹⁶. Moreover, variations in statin type, intensity, and timing of initiation complicate the interpretation of statins’ independent effects¹⁷. Therefore, further validation is warranted in large-scale, real-world datasets.

Using the Korean National Health Insurance Service (K-NHIS) database, this study investigates the association of statin therapy on major vascular outcomes in patients with AIS and AF. By focusing on statin-naïve patients without established diagnosed indications, we aim to provide insights into real-world treatment effects that may inform future therapeutic strategies.

Methods

This study retrospectively analyzed data obtained from the K-NHIS, a nationwide collection of a mandatory health insurance system that covers approximately 97% of the Korean population¹⁸. The K-NHIS database includes diagnoses, treatments, procedures, hospitalizations, discharges, and prescription records. Diagnostic information was identified using International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) codes. Data access requires K-NHIS review committee approval, and only analysis results—not raw data—are provided. Using this database, we assessed the demographic and clinical factors necessary for this study.

Study population

Between 2011 and 2023, 333,846 patients were diagnosed with AIS or transient ischemic attack (TIA) accompanied by AF. A diagnosis of AF was defined as either pre-existing or newly detected at the time of the stroke event. Among them, to improve diagnostic reliability, we included only patients who were hospitalized for at least one admission and underwent computed tomography or magnetic resonance imaging were included (n = 173,982). Patients aged ≥ 20 years at admission were identified (n = 173,764). To eliminate potential confounding effects of prior statin use, patients who had been administered statins at time of the stroke event were excluded (n = 118,248). Following clinical guidelines¹⁹, we also excluded patients with conditions mandating statin use, such as atherosclerotic cardiovascular disease (ASCVD), including acute coronary syndrome, AIS, TIA, or peripheral arterial disease (n = 65,424). Additionally, to accurately assess the effects of acute-phase statin therapy and to consider the potential legacy effect of early treatment initiation, we excluded those classified as statin users who did not receive statins within one week of stroke onset. (n = 64,190). For patients who died during the first week of hospitalization, exposure classification was based on whether they received statin during their hospital stay: those who received statin were classified as statin users, while those who did not were classified as non-users. The detailed patient selection process and numbers at each step are presented in sFigure 1.

Data collection

Information on demographics, comorbidities, and concomitant medications was obtained from the NHIS database. Definition of study covariates including outcomes are provided in sTable 1. We assessed the mean CHA₂DS₂-VASc score and the proportion of individuals with a score of 2 or higher, calculated at baseline during the index stroke hospitalization. Each component of the score was defined using ICD-10-CM codes and corresponding medication records. Statin users were defined as patients who initiated statin therapy within one week of admission and continued treatment after discharge. All patients were categorized as either statin users or non-users. Statin users were defined as those who initiated statin therapy within one week of admission and continued treatment after discharge, while non-users were those who did not receive any statin prescription during hospitalization or at discharge. For patients who died during hospitalization, classification was based on statin use within the first week of admission. Patients who did not receive statins within one week of hospitalization were excluded from the analysis. High-dose statins were defined according to clinical guidelines and classified, based on the agents available in Korea, as atorvastatin 40/80 mg or rosuvastatin 20 mg, while all other statin regimens were categorized as standard-dose statins¹⁹. For combination therapy, patients who received statins with ezetimibe, regardless of statin dosage, were categorized into the combination therapy group. Statin types were classified into atorvastatin, rosuvastatin, and other statins, which are commonly used in Korea²⁰. Definition related to statin prescriptions are provided in sTable 2.

Outcome and follow-up

The primary outcome was the composite vascular outcome at 1 year after stroke, including all-cause death, ischemic stroke or systemic embolism, intracranial hemorrhage (ICH), and myocardial infarction (MI). Each of these events was also analyzed separately as secondary outcomes. Patients were followed for 1 year to determine whether an event occurred. Patients were censored at the end of the observation period, upon the occurrence of an event outcome, or at death. Outcomes were identified based on ICD-10 codes from inpatient and outpatient claims and linked mortality data from the national registry.

Ethics statements

The current study was approved by the local institutional review boards of Chonnam National University Hospital (CNUH-2023-370). The requirement for informed consent was waived because the study used de-identified data from the K-NHIS database.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation, while categorical variables were presented as numbers (%). Student’s t-test and the chi-square test were used to compare continuous and categorical variables between the statin and non-statin groups. Non-normally distributed continuous variables were reported as median with interquartile range and compared using the Mann–Whitney test.

To reduce baseline differences, 1:1 propensity score matching (PSM) was performed using the greedy nearest neighbor method with a caliper of 0.01, and inverse probability of treatment weights (IPTW) were calculated. The propensity score was derived from a logistic regression model that included age, sex, medical history (hypertension, diabetes, dyslipidemia, heart failure, chronic kidney disease, chronic obstructive pulmonary disease), CHA₂DS₂-VASc score (dichotomized at 2), and acute thrombolytic treatment. Baseline characteristics were assessed before and after matching, with covariates considered well-balanced if the standardized mean difference was less than 0.1.

Event rates were reported per 100 person-year, and hazard ratios with 95% confidence intervals (CI) were estimated using Cox regression analysis, presented as Forest plots. In the pre-matching cohort, three models were applied: unadjusted (Model 1), covariate-adjusted (Model 2), and IPTW-adjusted (Model 3). Model 4 was unadjusted in the matched cohort. Kaplan–Meier curves were generated, and differences between groups were assessed using the log-rank test. The proportional hazards assumption was evaluated through visual inspection of log-minus-log survival curves and Schoenfeld residual analysis. For subgroup analyses (e.g., high-dose vs standard-dose statins and statin monotherapy vs statin plus ezetimibe combination therapy), separate propensity score models (both PSM and IPTW) were constructed to ensure appropriate balance within each comparison.

All analyses were performed using SAS version 9.4, with p < 0.05 indicating statistical significance.

Results

General characteristics

A total of 64,190 patients were analyzed, with a mean age of 73.76 years, and 55.3% were male. Hypertension was present in 78.2% of patients, diabetes in 20.5%, and 77.8% had a CHA₂DS₂-VASc score of 2 or more. Dyslipidemia was diagnosed in 34.1% of patients, but they were not receiving statin therapy at the time of admission. Statin users accounted for 37,033 patients (57.69%). Compared to non-users, statin users were younger and had lower rates of hypertension, diabetes, and heart failure. They also had a lower proportion of CHA₂DS₂-VASc scores of 2 or more. In addition, statin users were more likely to receive acute thrombolytic treatment (Table 1). Although there were notable differences in baseline characteristics between the two groups, propensity score weighting balanced the distribution of confounding variables, as indicated by the SMD values converging towards zero (sFigure 2).

Table 1.

Baseline characteristics of Statin users and non-users before and after propensity score matching.

	Before IPTW					After IPTW
Variables	All patients	Statin non-user	Statin-User	P	SMD	All patients	Statin non-user	Statin-User	P	SMD
Number	64,190	27,157	37,033
Age (years, mean ± SD)	73.76 ± 11.91	74.09 ± 12.03	73.51 ± 11.81	< 0.0001	− 0.0493	73.77 ± 16.87	73.77 ± 18.58	73.76 ± 15.5	0.8995	0.0135
Age group
65 years <	13,830(21.55)	5506(20.27)	8324(22.48)	< 0.0001	0.0630	27,600(21.50)	13,786(21.47)	13,813(21.52)	0.9728	0.0000
65–74 years	15,875(24.73)	6588(24.26)	9287(25.08)			31,777(24.75)	15,897(24.76)	15,880(24.74)
75 years ≥	34,485(53.72)	15,063(55.47)	19,422(52.45)			69,016(53.75)	34,530(53.77)	34,486(53.73)
Male (n, %)	35,521(55.34)	14,266(52.53)	21,255(57.39)	< 0.0001	0.0979	71,070(55.35)	35,548(55.36)	35,522(55.35)	0.9640	0.0063
Comorbidity
Hypertension (n, %)	50,172(78.16)	21,612(79.58)	28,560(77.12)	< 0.0001	− 0.0598	100,416(78.21)	50,236(78.23)	50,180(78.19)	0.8399	0.0053
Diabetes (n, %)	13,131(20.46)	5732(21.11)	7399(19.98)	0.0005	− 0.0279	26,254(20.45)	13,134(20.45)	13,121(20.44)	0.9655	0.0000
Dyslipidemia (n, %)	21,879(34.08)	9281(34.18)	12,598(34.02)	0.6784	− 0.0033	43,816(34.13)	21,921(34.15)	21,894(34.11)	0.8902	− 0.0008
Heart failure (n, %)	20,495(31.93)	9491(34.95)	11,004(29.71)	< 0.0001	− 0.1121	40,960(31.90)	20,480(31.89)	20,480(31.91)	0.9485	− 0.0066
Chronic Kidney disease (n, %)	4121(6.42)	1847(6.80)	2274(6.14)	0.0007	− 0.0269	8230(6.41)	4111(6.40)	4118(6.42)	0.9185	− 0.0089
COPD (n, %)	32,207(50.17)	14,087(51.87)	18,120(48.93)	< 0.0001	− 0.0589	64,411(50.17)	32,216(50.17)	32,195(50.16)	0.9835	0.0026
CHA₂DS₂-VASc (median, IQR)	3 (2–4)	3 (2–4)	3 (2–4)	< 0.0001	− 0.1393	3 (2–4)	3 (2–4)	3 (2–4)		0.0062
2 ≥ (n, %)	49,968(77.84)	21,868(80.52)	28,100(75.88)	< 0.0001	-0.1127	99,980(77.87)	50,013(77.89)	49,967(77.86)	0.8935	0.0041
IV tPA (n, %)	3710(5.78)	1072(3.95)	2638(7.12)	< 0.0001	0.1392	7454(5.81)	3739(5.82)	3715(5.79)	0.7888	− 0.0054
EVT (n, %)	9128(14.22)	3572(13.15)	5556(15.00)	< 0.0001	0.0532	18,427(14.35)	9252(14.41)	9175(14.30)	0.5655	− 0.0023

IPTW, inverse probability treatment weighting; SMD, standard mean difference; SD, standard deviation; EVT, endovascular treatment.

Values are presented as n (%) unless otherwise indicated. Numbers in parentheses represent percentages.

1-year vascular outcomes by statin therapy

In this study using the K-NHIS database, all patients without events were followed for a full 365 days. We analyzed vascular events according to statin use before and after PSM, as well as using IPTW. In the IPTW-based analysis, statin users showed significantly lower 100 person-year incidence rates for composite vascular outcomes (60.62 vs. 74.65), all-cause mortality (25.37 vs. 34.04), ischemic stroke and systemic embolism (33.88 vs. 37.85), and ICH (1.49 vs. 2.05) compared to non-users (sTable 3).

Statin use was associated with an 18% reduction in composite vascular outcomes (HR 0.821, 95% CI [0.808–0.835]), a 25% reduction in all-cause mortality (HR 0.746, [0.729–0.763]), a 10% reduction in ischemic stroke and systemic embolism (HR 0.904, [0.884–0.925]), and a 28% reduction in ICH (HR 0.725, [0.660–0.796]) compared to non-users. There was no significant difference between the two groups for MI (Table 2).

Table 2.

Comparison of vascular outcomes between Statin users and non-users.

Main	Crude HR	P	Adjusted HR	P	IPTW HR	P	PSM HR	P
Composite vascular outcome
Statin Non-users	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin Users	0.807(0.788–0.826)	< 0.0001	0.821(0.801–0.840)	< 0.0001	0.821(0.808–0.835)	< 0.0001	0.830(0.807–0.853)	< 0.0001
All-cause of death
Statin Non-users	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin Users	0.713(0.690–0.736)	< 0.0001	0.745(0.721–0.770)	< 0.0001	0.746(0.729–0.763)	< 0.0001	0.758(0.730–0.786)	< 0.0001
Ischemic stroke & Systemic embolism
Statin Non-users	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin Users	0.908(0.879–0.937)	< 0.0001	0.906(0.877–0.935)	< 0.0001	0.904(0.884–0.925)	< 0.0001	0.916(0.883–0.951)	< 0.0001
intracranial hemorrhage
Statin Non-users	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin Users	0.702(0.615–0.802)	< 0.0001	0.728(0.637–0.832)	< 0.0001	0.725(0.660–0.796)	< 0.0001	0.696(0.595–0.814)	< 0.0001
Myocardial infarction
Statin Non-users	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin Users	0.943(0.862–1.033)	0.2070	0.951(0.869–1.042)	0.2826	0.953(0.894–1.015)	0.1320	0.986(0.888–1.095)	0.7951

Statin use was associated with favorable outcomes for all-cause mortality, ischemic stroke and systemic embolism, and ICH—except for MI—across unadjusted HR, multivariable HR, PSM, and IPTW analyses. These results are presented as a Forest plot (Fig. 1). Statin users showed significantly better 1-year outcomes, including lower risks of composite vascular events, all-cause mortality, ischemic stroke or systemic embolism, and ICH (Fig. 2). Vascular outcome according to statin dose, ezetimibe combination and type.

Fig. 1.

Primary outcomes according to statin treatment. IPTW, inverse probability treatment weighting.

Fig. 2.

Vascular outcomes with Kaplan–Meier survival plot according to statin treatment. A Composite vascular outcomes, B All-cause of death, C ischemic stroke and systemic embolism, D Intracranial hemorrhage.

High-dose statins were prescribed in 39.34% of patients, while ezetimibe combination therapy was used in only 6.7%. Among statins, atorvastatin was the most commonly prescribed agent, accounting for 70.97% of all prescriptions (sTable 2).

In the IPTW-based analysis, high-dose statins were associated with an increased risk of composite vascular outcomes compared to standard-dose statins (HR 1.072, [1.047–1.096]), as well as a higher incidence of all-cause of death (HR 1.073, [1.039–1.108]) and systemic embolic events (HR 1.065, [1.033–1.097]) (Table 3 and Fig. 3A).

Table 3.

Comparison of vascular outcomes between high-dose and standard-dose Statin therapy.

	Crude HR	P	adjusted HR	P	IPTW HR	P	PSM HR	P
Composite vascular outcome
Standard dose statin (n = 22,355)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
High dose statin (n = 14,500)	1.027(0.994–1.062)	0.1076	1.084(1.048–1.120)	< 0.0001	1.072(1.047–1.096)	< 0.0001	1.073(1.034–1.113)	0.0002
All-cause of death
Standard dose statin (n = 22,355)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
High dose statin (n = 14,500)	0.975(0.931–1.022)	0.2985	1.107(1.056–1.160)	< 0.0001	1.073(1.039–1.108)	< 0.0001	1.078(1.022–1.136)	0.0055
Ischemic stroke & Systemic embolism
Standard dose statin (n = 22,355)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
High dose statin (n = 14,500)	1.059(1.014–1.106)	0.0092	1.067(1.021–1.114)	0.0038	1.065(1.033–1.097)	< 0.0001	1.055(1.005–1.107)	0.0301
Intracranial hemorrhage
Standard dose statin (n = 22,355)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
High dose statin (n = 14,500)	0.905(0.745–1.099)	0.3142	0.950(0.781–1.156)	0.6101	0.945(0.827–1.080)	0.4079	0.936(0.754–1.162)	0.5487
Myocardial infarction
Standard dose statin (n = 22,355)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
High dose statin (n = 14,500)	1.025(0.908–1.157)	0.6924	1.057(0.936–1.194)	0.3688	1.049(0.965–1.141)	0.2623	1.085(0.947–1.244)	0.2387

HR, hazard ratio; IPTW, inverse probability treatment weighting; PSM, propensity score matching.

Numbers in parentheses indicate the unweighted number of patients in each group; IPTW analyses were performed on weighted samples.

Fig. 3.

Primary outcomes according to statin dose, ezetimibe combination, and statin type with IPTW analysis. A Statin dose, B Ezetimibe combination, C Statin type. IPTW, inverse probability treatment weighting.

In the IPTW-based analysis, combination with ezetimibe was associated with a reduced risk of mortality compared to statin monotherapy (HR 0.871, [0.842–0.901]), but it was associated with an increased risk of ischemic events and embolism (HR 1.201, [1.166–1.237]) as well as ICH (HR 1.296, [1.140–1.472]). Consequently, statin monotherapy was associated with a significantly lower risk of composite vascular outcomes (HR 1.045, [1.022–1.070]) compared to combination therapy (Table 4 and Fig. 3B).

Table 4.

Comparison of vascular outcomes between Statin monotherapy and Statin plus Ezetimibe combination therapy.

	Crude HR	P	adjusted HR	P	IPTW HR	P	PSM HR	P
Composite vascular outcome
Statin monotherapy (n = 34,402)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin + Ezetimibe (n = 2,453)	1.047(0.980–1.118)	0.1719	1.049(0.982–1.120)	0.1527	1.045(1.022–1.070)	0.0001	1.036(0.948–1.133)	0.4350
All-cause of death
Statin monotherapy (n = 34,402)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin + Ezetimibe (n = 2,453)	0.853(0.772–0.943)	0.0018	0.873(0.790–0.965)	0.0078	0.871(0.842–0.901)	< 0.0001	0.842(0.739–0.959)	0.0099
Ischemic stroke & Systemic embolism
Statin monotherapy (n = 34,402)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin + Ezetimibe (n = 2,453)	1.211(1.116–1.314)	< 0.0001	1.199(1.105–1.301)	< 0.0001	1.201(1.166–1.237)	< 0.0001	1.219(1.087–1.367)	0.0007
Intracranial hemorrhage
Statin monotherapy (n = 34,402)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin + Ezetimibe (n = 2,453)	1.266(0.892–1.797)	0.1862	1.324(0.933–1.881)	0.1164	1.296(1.140–1.472)	< 0.0001	1.343(0.810–2.227)	0.2526
Myocardial infarction
Statin monotherapy (n = 34,402)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Statin + Ezetimibe (n = 2,453)	1.067(0.843–1.351)	0.5893	1.075(0.849–1.362)	0.5488	1.040(0.956–1.131)	0.3632	0.966(0.703–1.329)	0.8331

HR, hazard ratio; IPTW, inverse probability treatment weighting; PSM, propensity score matching.

Numbers in parentheses indicate the unweighted number of patients in each group; IPTW analyses were performed on weighted samples.

The results for composite vascular outcome, all-cause of death, ischemic events, and ICH did not differ between the two commonly used statins, atorvastatin and rosuvastatin (Table 5 and Fig. 3C). Kaplan–Meier curves for vascular outcomes according to statin dose, ezetimibe combination therapy, and statin type are presented in sFigures 3–5.

Table 5.

Comparison of vascular outcomes between Atorvastatin and rosuvastatin.

	Crude HR	P	adjusted HR	P	IPTW HR	P	PSM HR	P
Composite vascular outcome
Atorvastatin (n = 26,156)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Rosuvastatin (n = 8,947)	1.015(0.978–1.055)	0.4293	1.007(0.969–1.046)	0.7200	1.009(0.986–1.033)	0.4543	1.020(0.974–1.069)	0.4035
All-cause of death
Atorvastatin (n = 26,156)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Rosuvastatin (n = 8,947)	0.982(0.930–1.037)	0.5181	0.968(0.917–1.022)	0.2352	0.974(0.942–1.007)	0.1156	1.017(0.952–1.088)	0.6117
Ischemic stroke & Systemic embolism
Atorvastatin (n = 26,156)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Rosuvastatin (n = 8,947)	1.026(0.976–1.078)	0.3135	1.022(0.972–1.074)	0.3980	1.021(0.990–1.053)	0.1903	1.015(0.955–1.078)	0.6402
Intracranial hemorrhage
Atorvastatin (n = 26,156)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Rosuvastatin (n = 8,947)	0.909(0.725–1.138)	0.4043	0.916(0.731–1.147)	0.4440	0.917(0.800-1.051)	0.2144	0.812(0.622–1.059)	0.1237
Myocardial infarction
Atorvastatin (n = 26,156)	1 (Ref.)		1 (Ref.)		1 (Ref.)		1 (Ref.)
Rosuvastatin (n = 8,947)	1.138(0.994–1.304)	0.0619	1.129(0.986–1.294)	0.0792	1.134(1.042–1.235)	0.0037	1.024(0.869–1.207)	0.7792

HR, hazard ratio; IPTW, inverse probability treatment weighting; PSM, propensity score matching.

Numbers in parentheses indicate the unweighted number of patients in each group; IPTW analyses were performed on weighted samples.

Discussion

In the analysis of 64,000 patients with AIS or TIA and AF using a large-scale K-NHIS database, we found that initiation of statin therapy after stroke was associated with approximately an 18% reduction in the 1 year risk of composite vascular outcomes compared to non-users, including lower risks of mortality, ischemic stroke or systemic embolism, and ICH. However, the use of high-dose statins, combination therapy with ezetimibe, or different statin types was not associated with additional benefit in terms of composite vascular outcomes in this population.

Although AF has traditionally been regarded as a condition associated primarily with embolic rather than atherosclerotic risk⁸, several studies have demonstrated that the burden of atherosclerosis is not negligible in patients with AF and stroke¹⁰. According to previous studies, a higher CHA₂DS₂-VASc score in patients with AIS and AF has been associated with a greater burden of atherosclerosis, which is thought to reflect the shared vascular risk factors incorporated in the score components^21,22. Previously, cardioembolic stroke was reported to have a higher recurrence rate compared to atherosclerotic stroke²³. However, recent studies suggest that this difference has diminished with the appropriate use of non-vitamin K antagonist oral anticoagulants^17,24. As a result, recurrence associated with concomitant atherosclerotic lesions may now have a relatively greater impact on clinical outcomes. In this context, statin therapy in patients with AF and AIS may contribute to more favorable clinical outcomes.

Our findings align with previous studies demonstrating the vascular protective effects of statins in patients with stroke and AF and may contribute to a broader understanding of their potential in reducing vascular risk in this population. A previous investigation involving approximately 2,800 patients with cardioembolic stroke without guideline-based indications for statin use showed that statin therapy was associated with an approximately 60% reduction in major vascular events compared to non-users¹⁵. Similarly, in another study of 2,100 patients with AIS and AF, statin use was associated with a 36% lower risk of vascular events compared to non-user¹². Furthermore, an analysis of prognosis based on statin adherence in AF-related stroke showed about a 40% reduction in recurrent stroke²⁵. In addition, a meta-analysis of observational studies involving over 12,000 patients with cardioembolic stroke also suggested that statin use was associated with reduced mortality and improved functional recovery¹⁶. Taken together, these studies support the vascular benefits of statin.

Vascular events, including recurrent stroke, are known to occur more frequently during the early phase after ischemic stroke^26,27. In our study, Kaplan–Meier analysis showed that the divergence in cumulative event rates between the statin and non-statin groups appeared to emerge from the early phase of hospitalization. This association may be partly explained by the pleiotropic effects of statins, including anti-inflammatory actions, improvement of endothelial function, and plaque stabilization, which are known to exert beneficial effects in the acute phase of vascular injury^28,29. In contrast, for systemic embolism, the difference between statin users and non-users appeared to widen progressively over time rather than during the acute phase. This suggests that the protective effects of statins related to embolic mechanisms may manifest more gradually, while their early benefits are more likely associated with the modulation of atherosclerotic risk. Over time, this divergence in outcomes tended to become more pronounced.

Despite previous concerns that statin use might increase the risk of ICH, our study showed that statins were associated with a reduced risk of ICH even in stroke with AF patients^6,7, who have a relatively high incidence of HT^30,31. This finding aligns with several meta-analyses suggesting that statins do not elevate ICH risk^32,33. In our cohort, where 77% of patients had a CHA₂DS₂-VASc score more than 2, indicating higher vascular risk, these results suggest that statin therapy may also provide benefits in this population from an ICH perspective.

In interpreting the outcomes, it is important to consider that the higher proportion of patients with a CHA₂DS₂-VASc score of 2 or more among statin non-users suggests that this group may have had greater baseline severity. In real-world practice, statins are less likely to be prescribed in patients at high risk of HT or with severe stroke, which could explain these findings. This potential bias may have influenced the results. However, our study addressed this limitation by analyzing a large patient cohort and adjusting for these variables using propensity score weighting, thereby revealing the protective effects of statin therapy. This aligns with previous studies that demonstrated the protective effects of statins regardless of stroke severity¹⁵.

Our study provides insights by conducting a detailed analysis of statin dose, type, and combination therapy with ezetimibe. In Korea, atorvastatin and rosuvastatin are predominantly used, and initiating high-dose statins or combination therapy from the outset is uncommon. In patients with AIS and AF, high-dose statin did not provide additional benefits for composite vascular outcomes, in contrast to findings from previous studies¹². This discrepancy may partly reflect residual confounding by indication, as patients with greater baseline severity or higher bleeding risk may be less likely to be prescribed standard-dose statins and may be more likely to receive high-intensity regimens. Such unmeasured differences could have potentially contributed to the observed association. However, the possibility that high-dose therapy itself may not confer additional benefit—or could even be unfavorable—in this specific population cannot be excluded. These considerations highlight the need for prospective studies to clarify the optimal intensity of statin therapy in AF-related stroke. Regarding early combination therapy with ezetimibe, while it was associated with a mortality benefit, it showed unfavorable outcomes in terms of systemic embolism and ICH, and no significant difference in composite vascular outcomes. This suggests that current guidelines recommending initiation with statin monotherapy should be upheld, as evidence supporting early combination therapy remains limited. However, because only 6.7% of patients in our cohort received combination therapy, the small sample size may have limited the reliability of these estimates, and the findings should therefore be interpreted with caution. In addition, no significant differences in composite vascular events were observed based on the type of statin used.

Our study has several limitations due to the use of K-NHIS data. First, the lack of detailed clinical information made it difficult to accurately assess stroke severity using measures such as the modified Rankin Scale or the National Institutes of Health Stroke Scale. However, baseline characteristics suggest that statin non-users had higher underlying vascular disease burden and CHA₂DS₂-VASc scores, along with lower rates of acute thrombolytic treatment (typically administered for moderate-to-severe strokes), implying that non-users may have had more severe strokes. This pattern suggests our findings may underestimate rather than overestimate the true protective effect of statins, although these differences were balanced through PSM. In addition, although our study focused on AF-related stroke, the possibility remains that some patients with concomitant large artery atherosclerosis or other stroke mechanisms were included due to limitations of diagnostic coding. Second, evaluating medication adherence is challenging. The NHIS database only includes prescription records, so it is not possible to determine how consistently patients actually took their medications. Third, we aimed to exclude patients with indications for statin use. Although some of these indications are determined based on LDL-C levels, our dataset did not contain laboratory test results. Therefore, we excluded patients based on diagnostic codes for relevant conditions; however, it is possible that some patients with missing codes or those prescribed statins based on clinical judgment were inadvertently included. Fourth, while AF is the leading cause of cardioembolic stroke, other sources exist. Diagnostic code limitations prevented complete inclusion or exclusion of all cardioembolic cases. Fifth, the study included only a relatively ethnically homogeneous Korean population, which may limit generalizability. Finally, due to the predefined structure of the claims database, we were unable to conduct sensitivity analyses for several stratification factors (e.g., CHA₂DS₂-VASc score, use of anticoagulant therapy, exclusion of early deaths, and age). For example, anticoagulant therapy, which is a crucial determinant of outcomes in AF-related stroke, could not be evaluated in our dataset. Differences in anticoagulant use between statin users and non-users may have influenced vascular outcomes, particularly ischemic stroke and systemic embolism. However, we believe this confounding is likely limited, as CHA₂DS₂-VASc scores (the primary anticoagulation indication) were well-balanced after IPTW, and the greater protective effect on ICH compared to ischemic events is inconsistent with differential anticoagulation as the primary driver.

Despite these limitations, our study represents one of the few large-scale, nationwide investigations focusing on statin-naïve patients with AIS or TIA and AF. To more clearly evaluate the effect of statins in this population, we excluded patients who were already on statin therapy at the time of admission and those with a prior diagnosis of ASCVD, for whom statin use is strongly indicated. Additionally, to ensure that a stroke event had definitively occurred, only patients who had undergone neuroimaging were included. We further minimized potential confounding by employing propensity score weighting, thereby enhancing the validity of our findings. These methodological strengths allow for a more accurate assessment of statin effects in patients without established indications and provide meaningful insights into treatment patterns and real-world clinical outcomes. Our findings support current guidelines recommending statin therapy in AIS, regardless of stroke subtype³⁴.

Currently, the clinical trial by Clinical Research Collaboration for Stroke in Korea (CRCS-K) investigators is underway to assess the effects of statins in cardioembolic stroke, and its findings are eagerly anticipated (https://cris.nih.go.kr; KCT0006806).

In conclusion, in patients with AIS or TIA and AF but without diagnosed ASCVD, the use of acute-phase statin therapy was associated with improved major vascular outcomes, including reductions in mortality, ischemic stroke or embolism, and ICH. In this population, high-intensity statin therapy, combination therapy with ezetimibe, and the choice of statin type were not associated with a reduction in major vascular events. Given the observational nature of this study, our results should be interpreted as hypothesis-generating and warrant confirmation through future randomized controlled trials.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Study concept and design: H Kim, JT KimAcquisition of data: H Kim, JM Kim, JT Kim, MS Park, KH Choi, SH Min, HL LeeAnalysis and interpretation of data: J Lee, M KimDrafting of the manuscript: H Kim, JT KimAll authors read and approved the fi nal manuscript.

Funding

This study was supported by a grant (BCRI25031) of Chonnam National University Hospital Biomedical Research Institute. This work was supported by Establishment of K-Health National Medical Care Service and Industrial Ecosystem funded by the Ministry of Science and ICT(MSIT, Korea) Balanced National Development Account. [Project Name:Establishment of K-Health National Medical Care Service and Industrial Ecosystem / Project Number : ITAH0603230110010001000100100].

Data availability

Data used in this study are available upon reasonable request to corresponding author.

Declarations

Competing interests

The authors declare that they have no competing interests.

Ethical approval

The current study was approved by the local institutional review boards at Chonnam National University Hospital (CNUH-EXP-2023-370), and the requirement for informed consent was waived because this big-data–based research posed no potential risk to patients.