Abstract
Mendelian randomization studies have identified that apolipoprotein B (ApoB) is the primary genetic determinant of ischemic stroke, rather than other lipid markers. However, its association with recurrent non-cardioembolic acute ischemic stroke (NCAIS) remains unclear. This study aimed to investigate this association. This study recruited 578 patients with acute ischemic stroke, excluding those with cardiogenic embolism. After a 3-year follow-up, a total of 428 patients completed the prospective cohort study. A Cox regression model was used to evaluate the association between ApoB levels at admission and the recurrence rate. Additionally, a nested case-control study was conducted by comparing blood samples collected at the time of recurrence from recurrent patients with those from non-recurrent patients. Binary logistic regression and ROC analysis were used to assess the association between serum ApoB, low-density lipoprotein cholesterol (LDL-C), and recurrent stroke at the time of recurrence. The Cox regression model demonstrated that ApoB levels at admission were independently associated with an increased risk of recurrent NCAIS (HR=6.697; 95%CI 2.581–17.374, P < 0.001). Recurrent stroke patients had significantly higher serum ApoB levels at admission than non-recurrent ones [0.85 g/L (IQR 0.21) vs. 0.63 g/L (IQR 0.15)]. In ROC analysis, ApoB (AUC = 0.732) showed a greater discriminatory ability for recurrent stroke than LDL-C (AUC = 0.685). Higher serum ApoB levels increased the risk of recurrence in patients with NCAIS, and ApoB demonstrated better discriminatory ability than LDL-C after therapy. These findings suggest that routine ApoB measurement may help improve secondary stroke risk assessment.
Supplementary Information
The online version contains supplementary material available at 10.1007/s12975-025-01367-9.
Keywords: Apolipoprotein B, Ischemic stroke, Recurrent, Non-cardiogenic, Prediction
Introduction
Stroke is recognized as a significant contributor to global mortality and disability [1]. A recent systematic analysis of the Global Burden of Disease study reported that ischemic stroke accounts for 65.3% (62.4–67.7) of all stroke cases worldwide [2]. Notably, the burden of ischemic stroke is even more pronounced in China [3]. In comparison to high-income countries, the incidence and mortality rates of ischemic stroke are higher in middle- and low-income countries, with a significantly increased recurrence rate [4]. Recent meta-analyses focusing on the Chinese population reveal a recurrence rate surpassing 20% within 3 years for individuals aged 50 and above who have experienced ischemic stroke [5]. Patients with recurrent stroke often experience more severe brain damage and neurological impairments, leading to worse functional outcomes or higher mortality rates. A Swedish study further highlights shifts in the pathogenesis of first and recurrent strokes [6]. Traditional risk factors, such as age, conventional lipid abnormalities, and smoking, do not fully account for recurrent stroke, prompting a reassessment of recurrent stroke risk factors.
While traditional lipid abnormalities, which include elevated total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and triglycerides (TG), alongside reduced high-density lipoprotein cholesterol (HDL-C) levels, contribute to atherosclerosis and the onset of both initial and recurrent strokes [7], an experimental study on atherosclerosis demonstrated that the progression of atherosclerotic lesions could not be effectively halted even when the most representative lipid marker, LDL-C, was well controlled [8].
Previous studies have demonstrated that apolipoprotein B (ApoB) and the apolipoprotein B/apolipoprotein A-I ratio (ApoB/ApoA-I ratio) serve as precursors for ischemic stroke [9]. However, a Mendelian randomization study indicated that after adjusting for interaction, the association of reduced ApoA-I levels with ischemic stroke diminished to zero compared to HDL-C. In contrast, in the same model, ApoB maintained a robust effect compared to LDL-C and triglycerides even after interaction adjustment [10]. This suggests that ApoB might be part of the etiological basis for ischemic stroke, particularly in cases of large artery and small vessel stroke. The cardioembolic stroke (CE), as classified by the TOAST system, is primarily caused by emboli originating from the heart and is clearly distinct from these stroke subtypes; therefore, CE cases were excluded from the present study. ApoB is present in various lipoproteins. Elevated ApoB levels can increase the quantity of small and low-density lipoproteins, including LDL and VLDL, which may promote plaque growth, inhibit the fibrinolytic system, and stimulate cytokine production, thereby increasing the risk of atherosclerotic thrombosis [11].
However, few studies have examined whether serum ApoB levels affect recurrent stroke. Hence, this prospective cohort study, with up to 36 months of follow-up, aims to assess the association between ApoB levels and recurrence in non-cardiogenic acute ischemic stroke (NCAIS).
Materials and Methods
Sample Size Calculation
According to documented literature, the 2-year recurrence rate of ischemic stroke in the Chinese population ranges from 8.71% to 20.52% [12]. Based on previous research and clinical parameters of ApoB, the adjusted HR values range from 1.94 to 3.69 [9]. Assuming a one-sided type I error (α) of 0.05 and a type II error (β) of 0.1, a sample size of 471 was calculated using a survival analysis sample size method based on the log-rank test, as described in Statistics in Clinical Trials. To account for potential data loss, the sample size was increased by 10%, resulting in a planned enrollment of 517 cases.
Patient Selection
This study enrolled acute ischemic stroke patients who sought medical attention at the Neurology and Neuro-interventional Departments of the First People’s Hospital in Lianyungang City, Jiangsu Province, China, between January 2019 and December 2020. Informed consent was obtained from all participants prior to their inclusion in the study. The research adhered to the Helsinki Declaration principles and was approved by the Ethics Committee of the First People’s Hospital in Lianyungang (20,200,310).
Inclusion Criteria
(1) Surviving patients with acute ischemic stroke; (2) serum ApoB levels measured within 24 h of hospital admission; (3) diagnosis of acute ischemic stroke confirmed by magnetic resonance imaging (MRI) or computed tomography (CT) imaging within 24 h of onset, following the “Chinese Guidelines for the Diagnosis and Treatment of Acute Ischemic Stroke” (Neurology and Society 2018); (4) age ≥ 60 years; (5) availability of complete clinical data.
Exclusion Criteria
(1) Lack of initial imaging diagnosis and biochemical parameter tests within 24 h of admission; (2) presence of atrial fibrillation detected by electrocardiogram at admission, a history of atrial fibrillation, or a record of atrial fibrillation during the follow-up period; (3) history of traumatic brain injury, brain tumor, moyamoya disease, or cerebral vascular malformation; (4) presence of severe systemic diseases (such as heart disease, liver or kidney diseases, blood disorders, cancer); (5) severe psychiatric disorders; (6) patient refusal to participate in the study; (7) during the follow-up period, the patient underwent brain surgery.
The patient selection procedure for this study is as follows: From all acute ischemic stroke patients, those with available ApoB values were initially included (n = 578). Patients with cardioembolic ischemic stroke (n = 53), those who died from non-stroke causes (n = 9), and those lost to follow-up (n = 88) were subsequently excluded. After a 3-year follow-up, 428 patients were available for analysis (Fig. 1).
Fig. 1.
Flowchart for selecting study participants. Bidirectional cohort study: Among 428 NCAIS patients completing 36-month follow-up, we conducted a cohort study. Based on the baseline ApoB levels at admission, patients were divided into four quartile groups: Q1 (≤ 0.64 g/L), Q2 (0.65–0.77 g/L), Q3 (0.78–0.92 g/L), and Q4 (> 0.92 g/L). Cox regression analysis was used to assess the data. Case–control study: A case–control study was performed on 151 recurrent patients and 76 non-recurrent patients. The data was analyzed using logistic regression. NCAIS, non-cardiogenic acute ischemic stroke
Baseline Survey
Participant demographic data were obtained through inpatient medical records and face-to-face interviews. Clinical features were collected by trained physicians and included sex, age, smoking status, alcohol consumption, systolic blood pressure (SBP), diastolic blood pressure (DBP), fasting blood glucose (FBG), medical history, and medication taken after discharge. The modified Rankin Scale (mRS), ranging from 0 (asymptomatic) to 6 (death), was used to categorize patients into two groups: favorable outcomes (0–2) and unfavorable outcomes (3–6). Hypertension was diagnosed as SBP > 140 mmHg and/or DBP > 90 mmHg, or based on a self-reported history of hypertension medication use. Hyperglycemia was diagnosed as fasting blood glucose > 6.2 mmol/L, or a self-reported history of medication for hyperglycemia. Smokers were defined as individuals who smoked at least one or more cigarettes daily for a continuous year, with smoking cessation of less than 5 years. Alcohol consumption was defined as drinking at least one alcoholic beverage per week or more. Other medical histories were primarily obtained through patient history collection.
Clinical Data
Blood samples were obtained from all patients within 24 h of stroke onset. Fasting venous blood samples were collected to determine hematological and biochemical parameters, including high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), total cholesterol (TC), apolipoprotein A (ApoA), apolipoprotein B (ApoB), fasting blood glucose (FBG), serum albumin (ALB), glycated hemoglobin (HbA1c), uric acid (UA), total homocysteine (tHcy), complete blood count, coagulation profile, and liver and kidney function. Glycated hemoglobin was measured using a glycated hemoglobin analyzer (Bio-Rad D-100, France), while other blood samples were measured using a fully automated biochemical analyzer (Beckman Coulter AU5831-2, USA).
Patient Follow-up
All patients were monitored by trained clinical neurologists for 3 years post-discharge. Hospitalization information for patients with recurrent strokes were obtained, including recurrence status, time of recurrence, type of recurrence, survival status or cause of death, and duration of hospitalization if applicable. Recurrent ischemic stroke was defined as the onset of new neurological deficits accompanied by newly identified lesions on CT or MRI. Lesions of uncertain origin were not considered recurrent events. All recurrence events are verified through readmission medical records. Lipid samples were collected from recurrent patients, and, for comparison, samples were also collected from non-recurrent patients during the follow-up. We analyzed the outpatient data, ultimately establishing a roughly 1:2 ratio for the second measurement of blood lipid levels. Non-recurrent patients were from outpatient follow-up or hospitalization, which was randomized.
Patients who were prescribed medications were considered to have received relevant treatments, including: (1) antiplatelet therapy (e.g., aspirin, clopidogrel); (2) antihypertensive treatment (e.g., angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB)); (3) antidiabetic treatment (e.g., insulin, metformin); (4) anticoagulant therapy (e.g., heparin, warfarin); (5) lipid-lowering medication (e.g., statins, niacin, and derivatives).
Statistical Analysis
The primary outcome measure of this study is recurrent stroke. Categorical variables such as sex and medical history are presented as frequencies and percentages, while continuous variables, like ApoB and tHcy, are expressed as mean ± standard deviation (SD) or median [interquartile range (IQR)]. When comparing baseline characteristics, chi-square tests are utilized for categorical variables, while t-tests/Kruskal–Wallis tests (Mann–Whitney U test for two independent samples) are employed for continuous numerical variables. For the analysis of the univariate correlation between cerebrovascular risk factors and recurrent stroke, categorical variables are assessed using the Kaplan–Meier test, while continuous variables are analyzed using Cox regression models. In the multivariate analysis, the Cox regression model is used to analyze the longitudinal correlation of ApoB levels and their ratio with the risk of recurrent NCAIS, adjusting for potential confounders. Using the Cox proportional hazards model, the risk ratio (HRs) and their 95% confidence interval (CIs) for the risk of recurrent stroke are calculated, referencing the ApoB bottom quartile. The comparison of ApoB and LDL-C levels, including at admission and recurrence, between the recurrent and non-recurrent patient groups is conducted using the Mann–Whitney U rank-sum test. For the difference in continuous variables between two groups in a paired design, the Wilcoxon signed-rank test was used. The association between serum ApoB and LDL-C levels at admission and at recurrence with recurrent NCAIS was determined using receiver operating characteristic (ROC) curves. The differences in ROC curve values were assessed using the DeLong test. The optimal cutoff value was determined by calculating sensitivity (true positive rate, TPR) and specificity (true negative rate, TNR) at different thresholds. The Youden index (J) was computed for each threshold using the formula Youden J = TPR + TNR − 1, and the threshold with the highest Youden J value was selected as the optimal classification cutoff.
All statistical analyses are conducted using SPSS 26.0 (IBM, New York), and all tests are two-tailed, with an alpha level set at 0.05. Statistical significance is defined by a P-value < 0.05.
Nested Case–Control Design
This study initially employed a prospective cohort design to evaluate the recurrence of NCAIS over a 3-year period, applying the Cox regression model (sample size 427). To enhance the statistical power and efficiency of our research, a nested case–control design was incorporated within the prospective cohort. Subsequently, a case–control analysis was conducted by selecting cases (recurrent patients) and controls (non-recurrent patients) from the original cohort, focusing on lipid changes. This analysis employed conditional logistic regression and the Wilcoxon signed-rank test (sample size 227). This selection process was randomized but controlled to achieve an approximate 1:2 ratio, ensuring temporal consistency and minimizing bias. Matching criteria included age, sex, and comorbidities to control for potential confounding factors. Additionally, paired non-parametric tests (Wilcoxon signed-rank test) were conducted to compare lipid profiles obtained at the first admission and at recurrence within the same patients.
By leveraging the prospective follow-up cohort, the relationship between biomarkers and recurrent stroke was initially established. The subsequent nested case–control design enabled a more focused comparison of key biomarkers, such as serum ApoB levels, between recurrent and non-recurrent stroke patients, thereby significantly improving the study’s efficiency.
Results
Baseline Levels and Cox Regression Assessment
This study recruited 578 patients with acute ischemic stroke, of whom 428 completed the study. The average follow-up period was 31 months. During the first year of follow-up, there were 36 recurrent cases (8.41%). The cumulative number of recurrent cases increased to 71 (16.59%) in the second year and to 84 (19.63%) in the third year. Within the 3 years post-discharge, there were 9 deaths (2.10%). To control for variables, the deceased patients were excluded from the analysis. The average age of all patients was 68.4 (SD = 10.6) years. Among them, there were 243 male patients (56.78%) and 185 female patients (43.22%). The average ApoB level for all patients was 0.79 g/L (IQR 0.14). Thirteen (3.01%) patients received intravascular intervention treatment.
Baseline characteristics were stratified by quartiles of ApoB levels at admission (Table 1). Nearly all patients received statin therapy, predominantly atorvastatin and rosuvastatin, with no significant differences between groups (95.4% of patients). Risk factors deemed suspicious were selected through univariate analysis combined with those validated by relevant clinical studies, and subsequently evaluated using the Cox risk model (Table 2). Among all patients, the suspicious risk factors HbA1c, ALB, Hcy, and ApoB were independently associated with the recurrence of NCAIS. The continuous variables ALB (HR = 1.044; 95%CI 0.996–1.095, P < 0.1), HbA1c (HR = 1.11; 95%CI 0.984–1.252, P < 0.05), and Hcy (HR = 1.05; 95%CI 0.994–1.091, P = 0.001) were correlated with the risk of recurrent NCAIS. After adjusting for other indicators, ApoB levels at admission were independently associated with an increased risk of recurrent stroke, when considered as a continuous variable (HR = 6.697; 95%CI 2.581–17.374, P < 0.001).
Table 1.
Baseline characteristics of patients stratified by ApoB quartiles at admission
| Variables n (%) |
The quartiles of plasma ApoB levels (g/L) | P | ||||
|---|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | |||
| (≤ 0.64) | (0.65–0.77) | (0.78–0.92) | (> 0.92) | |||
| n = 107 | n = 111 | n = 111 | n = 109 | |||
| Age (y) | ≦60 | 14 (14.7) | 24 (25.3) | 29 (30.5) | 28 (29.5) | 0.080 |
| 61–70 | 33 (24.3) | 32 (23.5) | 38 (27.9) | 33 (24.3) | ||
| 71–80 | 42 (28.4) | 39 (26.4) | 37 (25) | 30 (20.3) | ||
| > 80 | 18 (12.2) | 16 (10.8) | 7 (4.7) | 8 (5.4) | ||
| Sex | Male | 64 (26.3) | 64 (26.3) | 66 (27.2) | 49 (20.2) | 0.407 |
| Female | 43 (23.2) | 47 (25.4) | 45 (24.3) | 50 (27.0) | ||
| Hypertension | No | 13 (22.4) | 13 (22.4) | 15 (25.9) | 17 (29.3) | 0.658 |
| Yes | 94 (25.4) | 98 (26.5) | 96 (26.0) | 82 (22.2) | ||
| Diabetes | No | 63 (22.6) | 81 (29.0) | 72 (25.8) | 63 (22.6) | 0.175 |
| Yes | 44 (29.5) | 30 (20.1) | 39 (26.2) | 36 (24.2) | ||
| History of cerebral infarction | No | 81 (23.3) | 89 (25.7) | 92 (26.5) | 85 (24.5) | 0.286 |
| Yes | 26 (32.1) | 22 (27.2) | 19 (23.5) | 14 (17.3) | ||
| History of cerebral hemorrhage | No | 102 (24.8) | 104 (25.3) | 108 (26.3) | 97 (23.6) | 0.365 |
| Yes | 5 (29.4) | 7 (41.2) | 3 (17.7) | 2 (11.8) | ||
| Smoking | No | 85 (25.6) | 87 (26.2) | 79 (23.8) | 81 (24.4) | 0.273 |
| Yes | 22 (22.9) | 24 (25) | 32 (33.3) | 18 (18.8) | ||
| Alcohol drinking | No | 94 (25.5) | 100 (27.1) | 91 (24.7) | 84 (22.8) | 0.400 |
| Yes | 13 (22.4) | 11 (19.0) | 19 (32.8) | 15 (25.9) | ||
| mRs | Good | 80 (25.2) | 82 (25.9) | 82 (25.9) | 73 (23.0) | 0.634 |
| Poor | 11 (19.0) | 14 (24.1) | 16 (27.6) | 17 (29.3) | ||
| lipid-lowering drug therapy | Atorvastatin | 64 (24.6) | 62 (23.9) | 65 (25) | 69 (26.5) | 0.188 |
| Rosuvastatin | 37 (25) | 44 (29.7) | 41 (27.7) | 26 (17.6) | ||
| Antihypertensive therapy | No | 12 (2.89) | 15 (3.61) | 11 (2.53) | 22 (5.05) | 0.247 |
| Yes | 88 (20.58) | 93 (21.66) | 105 (24.55) | 82 (19.13) | ||
| Antidiabetic therapy | No | 66 (15.52) | 68 (15.88) | 63 (14.80) | 56 (13.00) | 0.368 |
| Yes | 34 (7.94) | 40 (9.39) | 53 (12.27) | 48 (11.19) | ||
| Antiplatelet therapy | No | 8 (1.81) | 8 (1.81) | 3 (0.72) | 8 (1.81) | 0.538 |
| Yes | 93 (21.73) | 100 (23.36) | 113 (26.35) | 96 (22.38) | ||
| Anticoagulation therapy | No | 92 (21.66) | 101 (23.47) | 114 (26.71) | 102 (23.83) | 0.110 |
| Yes | 8 (1.81) | 9 (1.87) | 2 (0.36) | 1 (0.23) | ||
Modified Rankin Scale (mRS): evaluated at discharge, good: 0–2 points, poor: 3–5 points (patients with a score of 6 are excluded)
Table 2.
Multifactor analysis of risk factors associated with the recurrence of NCAIS
| Variables | HR | HR 95.0% CI | P | |
|---|---|---|---|---|
| Age (y) | 1.007 | 0.984 | 1.032 | 0.544 |
| Sex (female) | 0.943 | 0.568 | 1.566 | 0.822 |
| History of cerebral infarction | 0.74 | 0.443 | 1.238 | 0.252 |
| Smoking | 1.088 | 0.573 | 2.066 | 0.796 |
| Hcy (μmol/L) | 1.047 | 1.019 | 1.077 | 0.001 |
| Hba1c (%) | 1.151 | 1.017 | 1.304 | 0.026 |
| ALB (g/L) | 1.042 | 0.994 | 1.091 | 0.085 |
| UA (μmol/L) | 1 | 0.998 | 1.002 | 0.994 |
| ApoB (g/L) | 6.6 | 2.546 | 17.106 | < 0.001 |
All data from admission samples. HRs were adjusted for age, hypertension, diabetes, and smoking. Smoking and history of cerebral infarction are categorical variables, while the reference group has none. The reference group for sex (female) is male
Hcy homocysteine, HbA1c glycosylated hemoglobin, type A1C, ALB albumin, UA uric acid, ApoB apolipoprotein B, HR hazard ratio, CI confidence interval
Comparison of Serum ApoB Levels Measured at Admission Between Recurrent and Non-Recurrent NCAIS
Characteristics of patients with and without recurrent NCAIS are depicted in Fig. 1A, comprising 344 non-recurrent and 84 recurrent cases. Patients in the recurrent group exhibited significantly higher serum ApoB levels at admission than those in the non-recurrent group (0.84 g/L (IQR 0.21) vs. 0.63 g/L (IQR 0.15); Z = − 3.691, P < 0.001; Fig. 2A).
Fig. 2.
A Comparison of serum ApoB levels at admission between recurrent and non-recurrent NCAIS patients. All data are presented as median and interquartile range (IQR). P-values are derived from the Mann–Whitney U test for between-group differences. B Univariate association between ApoB quartiles and outcome events (recurrence)
Dividing ApoB into Four Groups for Comparison Based on Quartiles
To observe the impact of ApoB levels on non-cardiogenic NCAIS more significantly, we calculated the cumulative number of recurrences in each ApoB quartile group (Fig. 2B). With increasing ApoB levels, the risk of recurrent stroke significantly increases (Table 3). The cumulative recurrence rate of stroke in patients in the fourth (highest) quartile of ApoB is significantly higher than that in the first (lowest) quartile (P < 0.05). Patients with an ApoB level greater than 0.92 g/L have a 3.43-fold higher risk of developing the disease compared to patients with a level less than or equal to 0.64 g/L.
Table 3.
Associations between ApoB quartiles at admission and risk of recurrent NCAIS: a biological gradient analysis
| Apo B, g/L | Subjects, n | Strokes, n | Adjusted HR | (95% CI) | P |
|---|---|---|---|---|---|
| Q1 (≦0.64) | 107 | 11 | Reference | ||
| Q2 (0.65–0.77) | 111 | 20 | 1.732 | 0.766–3.914 | 0.187 |
| Q3 (0.78–0.92) | 111 | 23 | 2.566 | 1.203–5.475 | 0.015 |
| Q4 (> 0.92) | 99 | 30 | 3.430 | 1.617–7.275 | 0.001 |
HRs were adjusted for age, hypertension, diabetes, and smoking
Apo apolipoprotein, HR hazard ratio, CI confidence interval
ROC Curve of Serum ApoB and LDL-C Levels at Admission for Differentiating Between Recurrent and Non-Recurrent NCAIS
To visually assess the discriminate performance of ApoB, this study employed ROC curves and compared ApoB with the traditional lipid marker, LDL-C. According to the ROC curve analysis, the area under the curve (AUC) at admission was 0.630 for ApoB (95%CI 0.565–0.694, P = 0.002) and 0.641 for LDL-C (95%CI 0.577–0.705; P < 0.001; Fig. 3).
Fig. 3.
Receiver operating characteristic (ROC) curves: Sensitivity for the diagnosis of recurrent NCAIS based on ApoB and LDL-C as a function of 1-specificity. ApoB, apolipoprotein B; LDL-C, low-density lipoprotein cholesterol. Evaluated at admission: Average ApoB level was 0.79 g/L (IQR, 0.14), and average LDL-C level was 2.41 mmol/L (IQR, 0.5). Recurrent group: 84 individuals, control group: 344 individuals. The ROC values at admission were 0.630 for ApoB (95%CI, 0.565–0.694, P = 0.002), and 0.641 for LDL-C (95%CI, 0.577–0.705; P < 0.001). The difference between the two ROC curves was evaluated using the DeLong test: Z = − 0.865 (95%CI 0.035–0.014; P = 0.387)
The results of the DeLong test showed no statistically significant difference in the AUC between the two models (95%CI 0.02–0.18; P = 0.387), indicating that both serum ApoB and LDL-C measured at admission have discriminatory ability for the recurrence of NCAIS, but no significant difference was observed between the two.
Serum ApoB and LDL-C Levels Measured at Admission and Recurrence: Comparison Between Recurrence and Non-Recurrence Groups
Comparisons of serum ApoB and LDL-C levels between recurrent and non-recurrent groups at admission and recurrence were analyzed using the Mann–Whitney U rank-sum test, while within-group differences (admission vs. recurrence) were assessed via the Wilcoxon signed-rank test.
For LDL-C, significant differences were observed between the recurrent and non-recurrent groups at both admission and recurrence (P < 0.05, Fig. 4A), and within-group analysis demonstrated a significant decrease in LDL-C levels from admission to recurrence in both recurrent and non-recurrent groups (P < 0.05, Fig. 4C). Similarly, ApoB levels differed significantly between groups at both time points (P < 0.05, Fig. 4B). However, within-group analysis revealed a marked reduction in ApoB levels in the non-recurrent group (P < 0.05), whereas no significant change was detected in the recurrent group (P > 0.05, Wilcoxon signed-rank test), indicating that the persistent presence of ApoB is a risk factor (Fig. 4C).
Fig. 4.
A, B Comparison of serum ApoB and LDL-C levels between recurrent (n = 151) and non-recurrent (n = 76) groups at admission and recurrence (Mann–Whitney U test) C, D Trends in serum ApoB and LDL-C levels from admission to recurrence in both groups. C Trends in serum ApoB and LDL-C level changes (admission to recurrence) in both recurrent and non-recurrent groups (Wilcoxon signed-rank test). D Percentage decrease in lipid levels (admission to recurrence) in recurrent vs. non-recurrent groups, calculated as: Percentage Decrease = (Level at Admission − Level at Recurrence)/Level at Admission × 100%
In the non-recurrent group (151 patients), the ApoB level decreased by 16.67%, from 0.72 g/L (IQR 0.26) at admission to 0.60 g/L (IQR 0.31). The LDL-C level decreased by 13.89%, from 2.16 mmol/L (IQR 0.93) at admission to 1.86 mmol/L (IQR 0.98). In the recurrent group (76 patients), the ApoB level decreased by 5.81%, from 0.86 g/L (IQR 0.26) at admission to 0.81 g/L (IQR 0.43). The LDL-C level decreased by 10.41%, from 2.64 mmol/L (IQR 0.94) at admission to 2.37 mmol/L (IQR 1.30) (Fig. 4C, D).
ROC Curve of Serum ApoB and LDL-C Levels at the Time of Recurrence for Discriminating Between Recurrent and Non-Recurrent NCAIS
Based on the ROC curve, the optimal cutoff value of the serum ApoB level for identifying recurrent NCAIS at the time of recurrence was estimated to be 675 mg/L, with a sensitivity of 71.1%, specificity of 63.6%, and an AUC of 0.732 (95% CI: 0.662–0.801; P < 0.01). Furthermore, with an AUC for LDL-C at 0.685 (95%CI 0.612–0.759; P < 0.01)—with a sensitivity of 56.6% and a specificity of 71.5%—ApoB significantly outperforms LDL-C in predicting recurrent stroke (Fig. 5). When ApoB and LDL-C are combined for prediction, the specificity increases to 80.8% (Table 4). Although the 95% confidence intervals of the AUCs for ApoB and LDL-C overlapped, the DeLong test indicated that the difference between the two models was statistically significant (95%CI 0.02–0.18; P = 0.012), indicating that serum ApoB and LDL-C levels measured at the time of recurrence are significantly associated with recurrent stroke, with ApoB having better discriminatory ability.
Fig. 5.
Receiver operating characteristic (ROC) curves: Sensitivity for the diagnosis of recurrent NCAIS based on ApoB and LDL-C as a function of 1-specificity. Evaluated at recurrence: Average ApoB level was 0.67 g/L (IQR 0.15), and average LDL-C level was 1.99 mmol/L (IQR 0.49). Recurrent group: 76 individuals, control group: 151 individuals. ApoB, apolipoprotein B; LDL-C, low-density lipoprotein cholesterol; Joint, Joint Factor L, calculated as: L = (ApoB × 100) + (LDL-C × (0.01/0.04) × 100). The AUC for predicting recurrence was 0.732 (95%CI 0.662–0.801, P = 0.003) for ApoB and 0.685 (95%CI 0.612–0.759, P = 0.008) for LDL-C
Table 4.
Association of serum ApoB and LDL-C levels at recurrence with recurrent NCAIS: logistic regression and ROC analysis
| Variables | *B | OR | OR 95% CI | P | AUC | AUC 95% CI | Sensitivity% | Specificity% | ||
|---|---|---|---|---|---|---|---|---|---|---|
| ApoB*100 | 0.040 | 1.040 | 1.026 | 1.055 | 0.000 | 0.732 | 0.662 | 0.801 | 71.1 | 63.6 |
| LDL-C*100 | 0.010 | 1.010 | 1.006 | 1.014 | 0.000 | 0.685 | 0.612 | 0.759 | 56.6 | 71.5 |
| *Joint | 0.021 | 1.021 | 1.013 | 1.029 | 0.000 | 0.761 | 0.020 | 0.180 | 64.5 | 80.8 |
*B: Regression coefficient, reflecting the degree to which the independent variable affects the outcome. *Joint: Joint Factor L (L = ApoB*100 + 0.01/0.04 × LDL-C*100), where ApoB and LDL-C represent the measured values of apolipoprotein B and low-density lipoprotein cholesterol, respectively
Discussion
In this study, the 3-year recurrence rate of acute ischemic stroke was 19.63%, consistent with previous research results [5]. Univariate analysis also identified platelets as a potential influencing factor for the recurrence of acute ischemic stroke. Given that 93.46% of the participants were on antiplatelet medications, this variable was excluded from the multivariate analysis. Meanwhile, homocysteine (Hcy), serum albumin levels [13], and glycated hemoglobin (HbA1c) levels [14] were associated with the recurrence of acute ischemic stroke. These variables were included in the multivariate analysis to ensure a more reliable and interpretable relationship. This cohort study demonstrated that ApoB levels were independently associated with a 3-year recurrence risk in NCAIS patients, and elevated levels increased the recurrence risk.
In the collection of lipid samples from recurrent patients, in order to increase statistical power, enhance robustness, and reduce the likelihood of type II errors, we adopted a ratio of recurrent patients to non-recurrent patients of 1:2, while controlling for time variables. Figure 3 shows that the changes in serum LDL-C (13.89%) and ApoB (16.67%) were almost identical in the non-recurrent group, while the changes in serum ApoB (5.81%) were much smaller in the recurrent group than in LDL-C (10.41%). When combined with the ROC curve analysis, the admission values of ApoB and LDL-C demonstrated similar discriminatory performance. However, after treatment, the advantages of ApoB are more prominent when statins are more effective against LDL-C (95%CI 0.02–0.18; P = 0.012).
In routine clinical practice, LDL-C has become a standard lipid indicator for assessing the risk of vascular diseases [15, 16], but LDL cannot be measured directly; it is usually calculated [17]. Additionally, in the trial by Schwart et al., many people continued to experience cardiovascular events or disease progression, despite using maximum tolerated doses of newer statins to reduce LDL-C [18]. This suggests that more sensitive biomarkers may be needed to reduce the recurrence of subsequent vascular events.
Traditionally, attention has been focused on the role of ApoB in the prevention of coronary heart disease [19], and research related to stroke has mostly been centered around LDL. The 2019 European Society of Cardiology/European Atherosclerosis Society guidelines on lipid management explicitly state that apolipoprotein B is the optimal marker for atherosclerotic lipoproteins [20]. The study by Holme et al. also showed that the use of ApoB as a marker of dyslipidemia was superior to traditional indicators for predicting ischemic stroke [21]. Compared to LDL, ApoB is not only a better predictor of vascular disease risk but also can be directly and accurately measured, enhancing the reliability of the data [22].
Approximately 90% of the protein component of LDL cholesterol consists of ApoB, a key structural protein of all major atherogenic lipoproteins [23]. The retention of ApoB100-containing lipoproteins beneath the endothelium represents an early step in the formation of atherosclerosis [24]. Lipoproteins rich in cholesterol and ApoB are retained and accumulate in the intima of the artery wall [24, 25], leading to the immune response and inflammatory mechanism that promote the formation and development of atherosclerotic plaques [26]. Elevated levels of ApoB may also increase the number of small, dense LDL particles, which are susceptible to oxidation and thereby contribute to the growth of plaques [21]. Furthermore, elevated ApoB levels may lead to impaired reverse cholesterol transport by HDL, further exacerbating the progression of atherosclerosis [23].
Based on numerous studies, the recommendation to measure ApoB in clinical work does not mean it can replace traditional lipid indicators. This study showed that ApoB, although significantly more sensitive in assessing recurrent stroke, was less specific than LDL-C. For example, cardiovascular disease [27] and peripheral vascular disease [28] have abnormal levels of apolipoprotein. Notably, when ApoB and LDL-C were combined in the joint model, the performance improved substantially, with the AUC reaching 0.761 and specificity increasing to 80.8%, indicating enhanced predictive efficacy. Related studies have also indicated that ApoB, together with traditional lipids, can more accurately diagnose all clinically significant dyslipidemia [29]. Given the advantage of combining ApoB with LDL-C in assessing recurrent NCAIS, future studies could consider incorporating ApoB measurement into existing stroke management guidelines, alongside LDL-C, to optimize individual risk stratification and treatment strategies.
Atorvastatin, rosuvastatin, and other statins are first-line drugs for lipid-lowering, and their reduction of LDL-C is significantly greater than the reduction in ApoB levels [30]. Statins mainly lower LDL-C by reducing cholesterol in LDL-C particles. Prior studies show statins significantly reduce the cholesteryl ester transfer protein (CETP)-mediated transfer rate of cholesteryl esters (CEs), especially the 21% rate reduction in CEs transferred from HDL to ApoB-containing lipoproteins, mainly triglyceride-rich very low-density lipoprotein 1 (VLDL1). VLDL1, the main LDL-C precursor, becomes LDL-C after hepatic secretion and further processing such as esterase hydrolysis. In our study, the serum levels of LDL-C and total cholesterol (TC) measured at admission showed strong collinearity with stroke recurrence and also provided reference value (Supplementary File 1).
However, ApoB itself is neither reduced nor cleared in this process. ApoB reflects the number of lipoprotein particles, not the amount of cholesterol [31]. Our trial corroborated this, demonstrating the superior discriminatory ability of ApoB over LDL-C at the time of recurrence. Additionally, while commonly used statin drugs affect ApoB levels to varying degrees, there are currently no specific clinical medications targeting apolipoprotein B [32]. Therefore, committed to the latest drug development, this study has clinical reference value. It is also worth noting that ApoB levels are, to some extent, regulated by genetic factors. Certain genetic variations may influence an individual’s response to lipid-lowering medications, thereby altering the risk of recurrent stroke. Future research could leverage different drug characteristics in conjunction with genomic data to develop more tailored and rational individualized interventions.
This study had several strengths. Firstly, it is pioneering in assessing the impact of ApoB levels on the recurrence of NCAIS. Secondly, to enhance the rigor of this study, we compared ApoB and LDL-C levels within the same patient at admission and at recurrence, as well as between recurrent and non-recurrent patients at the same time points. Lastly, we collected data on various potential confounding risk factors, making our assessment of the independent impact of ApoB on recurrent stroke more reliable.
However, this study has limitations. Firstly, the low mortality rate during follow-up hinders confirming the correlation between mortality rate and ApoB levels. Secondly, data on ApoB levels were not collected regularly during the follow-up period in patients with NCAIS. Further research is needed to evaluate how ApoB levels change over time after stroke and their impact on recurrent stroke. Thirdly, ApoB levels were measured at the onset of NCAIS, which may not accurately reflect pre-stroke exposure. However, previous studies have reported related findings [9]. Lastly, this study only included patients from a specific area, so its applicability to other populations or races remains unverified. While it shows the prognostic value of apolipoprotein B, it does not prove its causal effect. Future multi-center randomized controlled trials are needed to investigate whether interventions targeting ApoB can effectively reduce recurrent stroke.
Conclusion
ApoB levels were independently associated with an increased risk of 3-year recurrence in non-cardiogenic acute ischemic stroke, and the stroke recurrence rate increased with higher ApoB levels. Furthermore, following pharmacological intervention, ApoB demonstrated stronger association with recurrent stroke compared to LDL-C. In conjunction with traditional lipids, ApoB enhances the precision of clinical assessments.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to express our gratitude to Dr. Xie Kang from the Neuroscience Research Institute of Lianyungang First People's Hospital for providing English language editing and proofreading services.
Abbreviations
- ACEI
Angiotensin converting enzyme inhibitor
- AIS
Acute Ischemic Stroke
- ALB
Albumin
- ApoB
Apolipoprotein B
- ApoA-I
Apolipoprotein A-I
- ARB
Angiotensin receptor blockers
- AUC
Area under the curve
- CE
Cardiogenic embolism
- CETP
Cholesteryl Ester Transfer Protein
- CI
Confidence interval
- CT
Computed Tomography
- FBG
Fasting blood glucose
- HbA1c
Glycosylated hemoglobin, type A1C
- Hcy
Homocysteine
- HDL-C
High-density lipoprotein cholesterol
- HR
Hazard ratio
- LDL-C
Low-density lipoprotein cholesterol
- MRI
Magnetic Resonance Imaging
- mRS
Modified Rankin Scale
- NCAIS
Non-cardiogenic Acute Ischemic Stroke
- ROC
Receiver Operating Characteristic
- SBP/DBP
Systolic pressure/Diastolic pressure
- SD
Standard deviation
- UA
Uric acid
- TC
Total cholesterol
- TG
Triglyceride
- VLDL1
Very Low-density Lipoprotein 1
Author Contributions
H.F and W.R. designed this study and were responsible for the completeness of the data and the accuracy of data analysis; Z.J. and H.T.made contributions to data collection; Z.Y.has made contributions to statistical analysis; Z.X. made contributions to manuscript writing and charting; S.Y. and L.A. revised the manuscript.
Funding
This work was supported by the 2020 Medical Research Project of the Jiangsu Provincial Health Commission (No. ZDA2020018), the fifth "333 High level Talents" funding project of Jiangsu Province (No. BRA2019247), 2023 Lianyungang Science and Technology Plan Project (No. JCYJ2304) and 2023 Lianyungang Aging Health Research Project (No. L202301).
Data Availability
No datasets were generated or analysed during the current study
Declarations
During the preparation of this work, the authors used ChatGPT in order to improve readability and language. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Ethics Approval
The research adhered to the principles of the Helsinki Declaration and received approval from the Ethics Committee of the First People's Hospital in Lianyungang (20200310).
Consent to Participate
All human participants have provided their consent.
Competing interests
The authors declare no competing interests.
Disclaimer
The sponsor has no role in research design, data collection and analysis, publication decisions, or manuscript preparation.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Fangbo Hu and Rongjie Wu contributed equally to this work.
Contributor Information
Aimin Li, Email: liaimin_6529@126.com.
Yong Sun, Email: Sunyong@njmu.edu.cn.
References
- 1.Pandian JD, Gall SL, Kate MP, Silva GS, Akinyemi RO, Ovbiagele BI, et al. Prevention of stroke: a global perspective. Lancet. 2018;392(10154):1269–78. [DOI] [PubMed] [Google Scholar]
- 2.Global, regional, and national burden of stroke and its risk factors, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet Neurol. 2024;23(10):973–1003. [DOI] [PMC free article] [PubMed]
- 3.Wu S, Wu B, Liu M, Chen Z, Wang W, Anderson CS, et al. Stroke in China: advances and challenges in epidemiology, prevention, and management. Lancet Neurol. 2019;18(4):394–405.> [DOI] [PubMed] [Google Scholar]
- 4.Katan M, Luft A. Global burden of stroke. Semin Neurol. 2018;38(2):208–11. [DOI] [PubMed] [Google Scholar]
- 5.Lin B, Zhang Z, Mei Y, Wang C, Xu H, Liu L, et al. Cumulative risk of stroke recurrence over the last 10 years: a systematic review and meta-analysis. Neurol Sci. 2021;42(1):61–71. [DOI] [PubMed] [Google Scholar]
- 6.Aked J, Delavaran H, Lindgren AG. Survival, causes of death and recurrence up to 3 years after stroke: a population-based study. Eur J Neurol. 2021;28(12):4060–8.> [DOI] [PubMed] [Google Scholar]
- 7.Musunuru K. Atherogenic dyslipidemia: cardiovascular risk and dietary intervention. Lipids. 2010;45(10):907–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fonseca L, Paredes S, Ramos H, Oliveira JC, Palma I. Apolipoprotein B and non-high-density lipoprotein cholesterol reveal a high atherogenicity in individuals with type 2 diabetes and controlled low-density lipoprotein-cholesterol. Lipids Health Dis. 2020;19(1):127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chou YC, Chan PC, Yang T, You SL, Bai CH, Sun CA. Apolipoprotein B level and the apolipoprotein B/apolipoprotein A-I ratio as a harbinger of ischemic stroke: a prospective observation in Taiwan. Cerebrovasc Dis. 2020;49(5):487–94. [DOI] [PubMed] [Google Scholar]
- 10.Yuan S, Tang B, Zheng J, Larsson SC. Circulating lipoprotein lipids, apolipoproteins and ischemic stroke. Ann Neurol. 2020;88(6):1229–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bhatia M, Howard SC, Clark TG, Neale R, Qizilbash N, Murphy MF, et al. Apolipoproteins as predictors of ischaemic stroke in patients with a previous transient ischaemic attack. Cerebrovasc Dis. 2006;21(5–6):323–8. [DOI] [PubMed] [Google Scholar]
- 12.He Q, Wu C, Guo W, Wang ZY, Zhao YF, Lu J, et al. Hospital-based study of the frequency and risk factors of stroke recurrence in two years in China. J Stroke Cerebrovasc Dis. 2017;26(11):2494–500. [DOI] [PubMed] [Google Scholar]
- 13.Zhang Q, Lei YX, Wang Q, Jin YP, Fu RL, Geng HH, et al. Serum albumin level is associated with the recurrence of acute ischemic stroke. Am J Emerg Med. 2016;34(9):1812–6. [DOI] [PubMed] [Google Scholar]
- 14.Wu S, Shi Y, Wang C, Jia Q, Zhang N, Zhao X, et al. Glycated hemoglobin independently predicts stroke recurrence within one year after acute first-ever non-cardioembolic strokes onset in a Chinese cohort study. PLoS ONE. 2013;8(11): e80690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2017;38(32):2459–72. [DOI] [PMC free article] [PubMed]
- 16.Baigent C, Blackwell L, Emberson J, Holland LE, Reith C, Bhala N, et al. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376(9753):1670–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Siri-Tarino PW, Krauss RM. Diet, lipids, and cardiovascular disease. Curr Opin Lipidol. 2016;27(4):323–8. [DOI] [PubMed] [Google Scholar]
- 18.Schwartz GG, Steg PG, Szarek M, Bhatt DL, Bittner VA, Diaz R, et al. Alirocumab and cardiovascular outcomes after acute coronary syndrome. N Engl J Med. 2018;379(22):2097–107. [DOI] [PubMed] [Google Scholar]
- 19.Walldius G, Jungner I. Apolipoprotein B and apolipoprotein A-I: risk indicators of coronary heart disease and targets for lipid-modifying therapy. J Intern Med. 2004;255(2):188–205. [DOI] [PubMed] [Google Scholar]
- 20.Mach F, Baigent C, Catapano AL, Koskinas KC, Casula M, Badimon L, et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41(1):111–88. [DOI] [PubMed] [Google Scholar]
- 21.As S, Sahukar S, Murthy J, Kumar K. A study of serum apolipoprotein A1, apolipoprotein B and lipid profile in stroke. J Clin Diagn Res. 2013;7(7):1303–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sniderman A, Shapiro S, Marpole D, Skinner B, Teng B, Kwiterovich PO Jr. Association of coronary atherosclerosis with hyperapobetalipoproteinemia [increased protein but normal cholesterol levels in human plasma low density (beta) lipoproteins]. Proc Natl Acad Sci U S A. 1980;77(1):604–8. [DOI] [PMC free article] [PubMed]
- 23.Behbodikhah J, Ahmed S, Elyasi A, Kasselman LJ, De Leon J, Glass AD, et al. Apolipoprotein B and cardiovascular disease: biomarker and potential therapeutic target. Metabolites. 2021;11(10). 10.3390/metabo11100690 [DOI] [PMC free article] [PubMed]
- 24.Goldstein JL, Brown MS. A century of cholesterol and coronaries: from plaques to genes to statins. Cell. 2015;161(1):161–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Skalen K, Gustafsson M, Rydberg EK, Hulten LM, Wiklund O, Innerarity TL, et al. Subendothelial retention of atherogenic lipoproteins in early atherosclerosis. Nat. 2002;417(6890):750–4. [DOI] [PubMed] [Google Scholar]
- 26.Libby P, Loscalzo J, Ridker PM, Farkouh ME, Hsue PY, Fuster V, et al. Inflammation, immunity, and infection in atherothrombosis: JACC Review Topic of the Week. J Am Coll Cardiol. 2018;72(17):2071–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ference BA, Kastelein JJP, Ginsberg HN, Chapman MJ, Nicholls SJ, Ray KK, et al. Association of genetic variants related to CETP inhibitors and statins with lipoprotein levels and cardiovascular risk. JAMA. 2017;318(10):947–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Forte F, Calcaterra I, Lupoli R, Orsini RC, Chiurazzi M, Tripaldella M, et al. Association of apolipoprotein levels with peripheral arterial disease: a meta-analysis of literature studies. Eur J Prev Cardiol. 2022;28(18):1980–90. [DOI] [PubMed] [Google Scholar]
- 29.Sniderman AD, Navar AM, Thanassoulis G. Apolipoprotein B vs low-density lipoprotein cholesterol and non-high-density lipoprotein cholesterol as the primary measure of apolipoprotein B lipoprotein-related risk: the debate is over. JAMA Cardiol. 2022;7(3):257–8. [DOI] [PubMed] [Google Scholar]
- 30.Barter PJ, Ballantyne CM, Carmena R, Castro Cabezas M, Chapman MJ, Couture P, et al. Apo B versus cholesterol in estimating cardiovascular risk and in guiding therapy: report of the thirty-person/ten-country panel. J Intern Med. 2006;259(3):247–58. [DOI] [PubMed] [Google Scholar]
- 31.Guerin M, Lassel TS, Le Goff W, Farnier M, Chapman MJ. Action of atorvastatin in combined hyperlipidemia: preferential reduction of cholesteryl ester transfer from HDL to VLDL1 particles. Arterioscler Thromb Vasc Biol. 2000;20(1):189–97. [DOI] [PubMed] [Google Scholar]
- 32.Rader DJ, Kastelein JJ. Lomitapide and mipomersen: two first-in-class drugs for reducing low-density lipoprotein cholesterol in patients with homozygous familial hypercholesterolemia. Circulation. 2014;129(9):1022–32. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
No datasets were generated or analysed during the current study