Variations in Risk Factors across Different Periods of Stroke and Transient Ischemic Attack Recurrence

Stroke is a disabling disease that causes high mortality rates and a significant decline in quality of life worldwide [1‒3]. Stroke exhibits a frequent recurrence compared to other diseases [4, 5]. Understanding recurrent strokes and transient ischemic attacks (TIAs) is important because they cause severe neurological damage, are difficult to treat, and have a higher risk of death, readmission, and long-term disability than the first stroke or TIA [6]. Therefore, secondary prevention after the first stroke is critical to reducing stroke recurrence.

Previous studies have investigated important variables for stroke recurrence, including age, diabetes, smoking, arterial hypertension, hyperlipidemia, peripheral arterial disease, hypercoagulable states, depression, and the National Institutes of Health Stroke Scale (NIHSS) score status [7‒10]. The cumulative incidence of stroke recurrence in the first 5 years is 16–30% [11‒13]. However, the stroke recurrence rate varies significantly over time, particularly after the initial stroke. The timing of recurrence shows significant heterogeneity depending on the risk factors [3]. The risk of recurrence ranged from 3.54 to 24.5% in the first year after the first stroke [14, 15]; the 2-year recurrence rate was approximately 30% [6]; and in the following 5 years, it was in the range of 9.4–22.90% [16, 17].

Precision in identifying the heightened risk of stroke or TIA recurrence and managing modifiable risk factors is crucial for minimizing the overall recurrence risk. The existing literature has overlooked early recurrence in the subacute phase, which occurs within 3 months after cerebral infarction. Consequently, our study aimed to address this research gap by identifying the risk factors and characteristics associated with early recurrence, particularly within the initial 3-month period. These findings provide valuable insights for clinical practice, aiding in the reduction of early recurrence rates.

This is a prospective institutional database from the Korea University Stroke Registry. The design of the database has been described in detail previously [18]. Briefly, we retrospectively selected consecutive patients with ischemic stroke TIA (diagnosed by a neurologist within 30 days of the event) who were admitted to the Stroke Center of Korea University Ansan Hospital between March 2014 and December 2021 due to neurological problems corresponding to stroke or TIA. Patients with hemorrhagic strokes were excluded from the study (shown in Fig. 1). Stroke was defined as focal clinical signs of central nervous system dysfunction of vascular origin that lasted for at least 24 h according to the World Health Organization (Geneva) criteria. TIA was defined as losing cerebral or ocular function for less than 24 h. Both ischemic stroke and TIA were classified as recurrences, determined based on neurological symptoms or ischemic changes observed in imaging tests. Therefore, if the initial event was a TIA, both subsequent strokes and TIAs were considered recurrences. Similarly, if the initial event was a stroke, both subsequent strokes and TIAs were regarded as recurrences.

Information on demographics, clinical evaluations, neurological examinations, stroke or TIA characteristics, and outcomes was obtained. The diagnosis required brain computed tomography (CT) and/or magnetic resonance imaging (MRI) to exclude hemorrhage and other causes of symptoms. Each patient required at least one vascular imaging scan, including conventional angiography, arterial imaging using MR angiography, CT angiography, or duplex ultrasound imaging. Standard systemic investigations were performed for every patient, including routine laboratory tests, chest radiography, and 12-lead electrocardiography. Routine laboratory tests included a complete blood count, electrolyte, glucose, renal function, liver function, lipid profile, and homocysteine levels. Transcranial Doppler, carotid duplex sonography, transthoracic echocardiography, transesophageal echocardiography, and 24-h Holter electrocardiography monitoring were performed in selected patients. This study was approved by the Institutional Review Board of Korea University Ansan Hospital (Approval No. AS0213). Informed consent was not required owing to the retrospective design of the study.

For demographics and comorbidities, age at admission was categorized into three groups (<60, 60–74, and ≥75 years). Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m²) and was categorized into four groups (underweight, <18.5; normal weight, 18.5–22.9; overweight, 23.0–24.9; obese, ≥25.0) by modified previous criteria [19]. The weight and height required for BMI measurements were measured within 24 h of admission. The smoking status was classified as current smoker or nonsmoker. Hypertension was defined as systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or the presence of hypertension when patients were previously diagnosed or were receiving antihypertensive medication. Diabetes mellitus was defined as fasting serum glucose ≥126 mg/dL, non-fasting serum glucose ≥200 mg/dL, hemoglobin A1c ≥6.5%, or a history of insulin therapy and/or oral hypoglycemic drugs. Dyslipidemia was defined as total cholesterol ≥200 mg/dL [20]. Blood samples were obtained from all patients after at least 8 h of fasting on the morning of admission. Stroke or TIA is associated with abnormal biochemical parameters such as anemia [21], leukocytosis [22], elevated high-sensitivity C-reactive protein (CRP) levels [22, 23], high homocysteine levels [24, 25], and reduced kidney function [26]. In this study, these biochemical cutoff points were used to define and categorize abnormal levels: anemia was defined as hemoglobin <12 g/dL in females and <13 g/dL in males; leukocytosis was defined as a white blood count >12,000/μL; hyperhomocysteinemia ≥15 μmol/L; CRP elevation >2 mg/L; and abnormal kidney function as estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m².

Congestive heart failure was defined as a reduced left ventricular ejection fraction (<50%). Atrial fibrillation (AF) was defined as persistent atrial arrhythmia with irregular R-R intervals and no clear repetitive P waves and was diagnosed with an electrocardiogram, 24-h Holter, or continuous electrocardiogram monitoring during hospitalization.

White matter changes (WMCs) in the hemisphere contralateral to the area affected by the acute stroke were assessed using the proposed visual rating scale [27, 28]. If the WMCs were higher on one side, the rating was based on the less involved or uninvolved side with the assumption of symmetry. In this study, the WMC rating was modified using a dichotomous method (WMCs, grade 3 vs. no WMCS, grades 1–2). Cerebral microbleeds (CMBs) were assessed as rounded or ovoid hypointense lesions on a T2-gradient recalled echo-weighted MRI sequence [29]. CMBs measured 10 mm or less in diameter and were surrounded by brain parenchyma over at least half the circumference of the lesion. The severity of neurological deficits at admission was rated using the NIHSS score [30] on the day of admission, defined as poor initial NIHSS (≥5), which was evaluated by the certified neurologists who were blind to this study.

For subtypes of stroke, ischemic stroke was classified according to the Trial of Org 10172 in the Acute Stroke Treatment (TOAST) classification system: large artery disease, cardioembolism, small vessel occlusion (SVO), stroke of other determined etiology, and stroke of undetermined etiology [31]. The stroke of undetermined etiology category included three heterogeneous groups: “two or more causes,” “negative evaluation,” and “incomplete evaluation.” Recurrent stroke or TIA was defined as a new neurological deficit or deterioration of the previous deficit and fit the definitions for ischemic stroke, or TIA. We monitored patients for 1 year following a stroke or TIA, assessing outpatient visits and neurological changes at 3, 6, 9, and 12 months. Patients who exhibited symptoms between these intervals underwent an immediate follow-up MRI. Additionally, all patients (regardless of symptom recurrence) received a follow-up MRI at 12 months to evaluate for silent infarction. Recurrence was classified according to the recurrent interval as early recurrence (≤3 months) or late recurrence (>3 months). The extent of the diagnostic workup, stroke, and TIA were determined primarily by the stroke neurologists in charge of the patients, and the types of stroke and TIA were confirmed at a monthly stroke registry meeting.

All participants were categorized according to stroke or TIA recurrence (recurrent group and first-ever group), and the recurrent group was further subcategorized into early and late recurrent groups. Categorical variables are described in terms of frequencies and percentages. Differences in characteristics among groups were analyzed using Pearson’s χ² (χ²) test or Fisher’s exact test, as applicable. A pairwise z-test with the Bonferroni correction was used to determine the significance of the contribution of each subgroup of variables.

To identify potential predictors of recurrence, we selected variables with p < 0.2 in a bivariable analysis. We used a modified Nightingale-Rose diagram to show the distribution of potential predictors of recurrence. These potential variables of recurrence were used to generate a predictive model for each recurrence by multivariable binary and multinomial logistic regression models using bootstrapping methods with a statistically significant association at p < 0.05 compared to those of first-ever stroke or TIA. Goodness-to-fit and pseudo-R² statistics were used to evaluate the model fit and predictive strength. The final models were used to calculate the adjusted odds ratios (ORs) and 95% confidence intervals (CIs).

Statistical significance was declared when the two-tailed p value <0.05. Statistical analyses were performed using R version 4.2.1 software (R Foundation for Statistical Computing, Vienna, Austria) and SPSS (version 20.0; IBM Armonk, NY, USA).

Between March 2014 and December 2021, 4,508 individuals with ischemic stroke or TIA were admitted to the Stroke Center at Korea University Ansan Hospital. Of these, 862 patients were excluded from the final analysis, comprising 709 patients who refused the omission of tests for the stroke registry and 153 patients who were lost during the evaluation period for stroke or TIA recurrence. Consequently, the final analysis included 3,646 patients, of whom 255 experienced recurrent stroke or TIA and 3,391 had their first stroke or TIA (shown in Fig. 1).

Table 1 outlines the demographic and clinical features of the participants according to the occurrence of recurrence. The recurrent patient group had a significantly higher proportion of men than the first-ever stroke group (71.4 vs. 60.4%, p = 0.001). Variations in BMI across categories, such as normal, underweight, overweight, and obese, showed that patients with the underweight category (BMI 18.5–22.9) showed a significantly higher rate of recurrence (9.1 vs. 4.4%, p = 0.005).

Concerning stroke or TIA risk factors, higher recurrence rates were observed in individuals with hypertension, diabetes mellitus, and congestive heart failure. Stratifying participants based on stroke subtype according to the TOAST classification revealed a significantly lower recurrence rate in cases of SVO (14.9 vs. 24.8%, p = 0.005).

Laboratory findings indicated a higher prevalence of anemia, elevated CRP levels exceeding 2 mg/dL, and a low eGFR in recurrent cerebral infarctions. In terms of radiological findings, WMCs and CMBs were associated with recurrence.

Figure 2 summarizes the analysis of the differences between the initial recurrent stroke or TIA, recurrence rate, and first stroke or TIA using a diagram. For variables such as male sex, diabetes mellitus, TOAST classification, anemia, and CMBs, patients with early and late recurrence showed similar patterns of significant differences when compared with patients with first-ever stroke or TIA.

Regarding BMI (p = 0.002), congestive heart failure (p = 0.024), leukocytosis (p = 0.029), and WMCs (p < 0.001), significant differences were observed in comparison to first-ever stroke or TIA patients only in the early recurrent group. However, regarding dyslipidemia (p = 0.047), significant differences in patients with first-ever stroke or TIA were observed only in the late recurrent group. A comparison of early and late recurrent strokes showed significant differences in TOAST classification and CRP levels (shown in online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000540571).

A multinomial logistic regression analysis was used to analyze the early and late recurrent stroke variables (shown in Fig. 3). For overall recurrence, male sex (OR: 2.05, 95% CI: 1.45–2.88), hypertension (OR: 1.54, 95% CI: 1.06–2.22), diabetes mellitus (OR: 1.59, 95% CI: 1.15–2.18), other determined etiologies in TOAST classification (OR: 1.98, 95% CI: 1.04–3.78), TIA (OR: 2.01, 95% CI: 1.18–3.43), anemia (OR: 3.96, 95% CI: 2.81–5.57), WMCs (OR: 1.62, 95% CI: 1.08–2.44), and CMBs (OR: 1.88, 95% CI: 1.29–2.72) showed significant results.

In the context of early recurrence, significant associations were observed for diabetes mellitus (OR 1.98, 95% CI: 1.25–3.15), other determined etiology in TOAST classification (OR 3.00, 95% CI: 1.37–6.61), and WMCs (OR 1.97, 95% CI: 1.17–3.33). Conversely, in the case of late recurrence, a significant correlation was noted with TIA (OR 2.95, 95% CI: 1.52–5.71) and CMBs (OR 2.22, 95% CI: 1.32–3.75). Additionally, for both early and late recurrence, significant associations were observed with male sex (OR: 1.98, 95% CI: 1.25–3.15; OR: 2.14, 95% CI: 1.32–3.48, respectively) and anemia (OR: 4.53, 95% CI: 2.84–7.20; OR: 3.40, 95% CI: 2.10–5.51, respectively).

In this study, we focused on understanding the factors influencing stroke or TIA recurrence and distinguishing patients experiencing a recurrence from those who experienced a first-ever stroke or TIA. Given the dynamic nature of stroke or TIA recurrence, we anticipated variations in causative factors over time. Consequently, we categorized our analysis into early and late periods based on a 3-month timeframe.

Our investigation aligns with previous studies that explored stroke recurrence. Lee et al. [32] discovered that the risk of 1-year stroke recurrence was significantly elevated in men, older adults, and those with a history of ischemic stroke. A meta-analysis by Zheng and Yao suggested that hypertension, diabetes, AF, and coronary heart disease are associated with a heightened risk of stroke recurrence [33]. In a study by Hillen et al. [34], diabetes and AF emerged as significant contributors to outcomes in the first year after the index stroke. These risk factors are associated with stroke recurrence, with significant differences. Reviewing multiple prior studies, the most firmly established risk factors linked to stroke recurrence were diabetes and AF [35‒38]. This is consistent with the current research showing that diabetes is a crucial factor. Regarding AF, our study suggests a tendency for a higher prevalence in patients experiencing recurrent cerebral infarction. However, the results no longer reached statistical significance. This observation could be attributed to the widespread clinical use of non-vitamin K antagonist oral anticoagulants and the heightened early AF detection rate facilitated by tools such as injectable cardiac monitors and long-term continuous ambulatory electrocardiography monitors.

Previous research results on hypertension were more heterogeneous, and several studies reported that hypertension was associated with the highest risk of stroke recurrence. However, few studies reported a significant association with stroke recurrence, which may be related to different criteria and methods for measuring hypertension and issues with antihypertensive management [9, 34, 35, 38, 39]. Although this study demonstrated a correlation with overall stroke or TIA recurrence, no significant correlation was observed with early recurrence.

In our study, a high initial recurrence rate was observed in the category of “other determined etiology” in the TOAST classification, consistent with findings from previous research. Specifically, within this classification, the majority of patients exhibited dissection, moyamoya disease, and cancer-related coagulopathy, which is consistent with past studies that have identified these conditions as being associated with elevated recurrence rates. For instance, Weimar et al. [40] highlighted the heightened risk of early recurrence following acute ischemic stroke, or TIA, attributed to carotid artery dissection. Moreover, a study by Chiu et al. [41] indicated that moyamoya disease entails a high risk of stroke recurrence, with an 18% recurrence rate within the first year after diagnosis, decreasing to approximately 5% over the subsequent 5 years. This trend may be influenced by the gradual development of angiogenesis and collaterals following the initial cerebral infarction. The recurrence rate was higher in patients with cancer-related strokes than in patients with inactive cancer or controls. The estimated 1-year stroke recurrence rate in patients with cancer and embolic stroke of undetermined source ranges from 14% to 29%, which is approximately 3 times higher than that in embolic stroke of undetermined source patients without cancer [35]. WMCs indicate stroke recurrence up to 5 years after the first ischemic stroke or TIA. This shows that WMCs can be considered an SVO marker that summarizes the impact of several classical risk factors on small vessel brain networks [42]. We observed that WMC had a particularly strong impact on early recurrence. Conversely, the association between TIA and late recurrence can be attributed to the possibility that symptoms are not observed for an extended duration during the treatment process. The risk factors are effectively managed, and clinicians may opt to discontinue antithrombotic drugs in clinical practice. Considerably, the risk of recurrence in these cases being detected after a long period of time is higher compared to patients with cerebral infarction who consistently adhere to antithrombotic therapy.

Our study has several limitations. First, this was a single-hospital investigation. Inherent in utilizing clinical registry data is a potential bias in patient selection, recording practices, data completeness, and outcome assessment. Therefore, generalizing the findings of this study may be challenging. To bolster the robustness of these results, it is imperative to validate them through prospective multicenter studies with larger sample sizes. Second, a challenge in examining stroke or TIA recurrence is the difficulty of identifying such recurrences, particularly if they manifest immediately after the index stroke. Detecting new neurological signs in unconscious, paralyzed, or bedridden patients with a modified Rankin Scale (mRS) score of 4 or higher is markedly more challenging than in individuals who have recovered. Third, while we categorized stroke or TIA recurrence into early (within 3 months) and late (beyond 3 months) groups, a more nuanced classification of the late recurrence group could have yielded more detailed insights. Lastly, during the initial 3 months, the loss of follow-up was minimal. However, beyond this period, patients exhibited increased loss of follow-up, potentially introducing bias to our findings.

In conclusion, our findings indicated that certain factors were closely associated with stroke or TIA recurrence within the first 3 months. Patients with diabetes mellitus, other determined etiologies according to the TOAST classification, and WMCs were particularly prone to early recurrence. In contrast, individuals with TIA and CMBs exhibited a higher likelihood of delayed relapse. Regardless of the timing, male sex and anemia were identified as consistent risk factors for recurrence. Our study highlighted the substantial differences in the causative factors of stroke or TIA recurrence based on the timing of recurrence. To effectively mitigate the risk of recurrence in clinical practice, it is necessary to accurately identify periods when the risk of stroke or TIA recurrence is heightened, with stringent control of modifiable risk factors.

This study was approved by our Institutional Ethical Review Board of the Korea University (Approval No. AS0213) and was in accordance with the ethical standards and with the 1964 Helsinki Declaration and its later amendments. This study was approved and the requirement for informed consent was waived by the Institutional Review Board of Korea University Ansan Hospital.

The authors have no conflicts of interest to declare.

This work was supported by a Korea University Ansan Hospital Grant (No. K2316101).

Dr. S.H.L. contributed to the study conception and design, data analysis, acquisition of clinical and imaging data, statistical analysis, and manuscript drafting and revision. Dr. M.H.P. contributed to the study conception and design, analysis, and interpretation of the imaging and clinical data, manuscript drafting and revision, and study supervision. Dr. J.M.J. contributed to the data analysis and manuscript revision. Dr. J.C.R. contributed to the study conception and data statistical analysis.

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.