Abstract
Traditional screening for stroke risk often focuses on carotid stenosis—the degree of arterial narrowing. However, unstable carotid plaques can trigger strokes even in patients with less than 50% stenosis. This study developed a predictive model to identify unstable carotid plaques in a high-risk population. Analyzing 2,841 individuals from the China National Stroke Screening and Prevention Project (CNSSPP), researchers identified five independent risk factors: male gender, advanced age, marital status, high LDL-C, and low HDL-C. A 10-point scoring system was created with a 6.5-point cutoff, yielding an Area Under the Curve (AUC) of 0.738. This tool allows clinicians to prioritize high-risk patients for carotid ultrasonography and early intervention, potentially preventing strokes before significant arterial blockage occurs.
Introduction
Stroke remains a leading cause of mortality and long-term disability globally. A significant portion of ischemic strokes is attributed to carotid atherosclerosis. While clinical guidelines traditionally emphasize the degree of carotid stenosis (typically >50%) as the primary indicator for intervention, recent evidence suggests that plaque morphology—specifically "instability" or vulnerability—is a critical predictor of embolic events. Unstable plaques, characterized by a thin fibrous cap, large lipid core, or intraplaque hemorrhage, can rupture and cause a stroke regardless of the severity of the narrowing.
Current predictive models often overlook individuals with low-level stenosis who may still harbor these dangerous plaques. To bridge this gap, this study aimed to develop a practical scoring system to identify unstable carotid plaques in individuals already flagged as having a high risk for stroke based on standard clinical factors.
Study Design and Methodology
The researchers conducted a community-based study using data from the China National Stroke Screening and Prevention Project (CNSSPP) in Nanjing.
As shown in Figure 1, the study initially screened 34,227 residents, eventually narrowing the focus to 2,841 high-risk individuals. Participants were categorized based on carotid duplex ultrasonography. Plaques were defined as "unstable" if they exhibited hypoechoic or mixed echoes, or showed signs of ulceration. The cohort was split into a Derivation Set (n=1,894) to build the model and a Validation Set (n=947) to test its accuracy.
Identification of Risk Factors
Using multivariable logistic regression, the study identified five independent predictors for the presence of unstable carotid plaques:
- Gender: Males were nearly twice as likely to have unstable plaques (OR 1.966).
- Age: Risk increased exponentially with age. Compared to the 40–49 age group, those aged 70+ had an Odds Ratio (OR) of 31.267.
- Marital Status: Being married was associated with higher risk (OR 1.780), potentially due to shared lifestyle factors or dietary habits in the study region.
- LDL-C: High levels of low-density lipoprotein cholesterol were a strong predictor (OR 2.015).
- HDL-C: Low levels of high-density lipoprotein cholesterol also significantly increased risk (OR 2.130).
The Predictive Scoring System
To translate these findings into a clinical tool, the researchers developed a 10-point scoring system. The model's performance was evaluated using Receiver Operating Characteristic (ROC) curves.
- Optimal Cutoff: 6.5 points.
- Performance: The AUC was 0.738 for the derivation set and 0.737 for the validation set, demonstrating consistent and fair-to-good discriminatory power.
- Clinical Utility: A score above 6.5 indicates a high probability of unstable plaques, suggesting that these patients should undergo immediate carotid ultrasonography even if they do not meet traditional stenosis criteria.
Discussion and Conclusion
This study shifts the focus of primary stroke prevention from simple arterial narrowing to plaque vulnerability. By identifying high-risk individuals before a major embolic event occurs, clinicians can initiate early interventions such as intensive statin therapy or antiplatelet agents.
While the model is robust, it has limitations. The data is derived from a single center in Nanjing, and the "marriage factor" may reflect specific regional socio-economic conditions that might not apply globally. Furthermore, the cross-sectional design means the model predicts the presence of plaques but does not yet track long-term stroke outcomes. Future research will focus on longitudinal follow-up to validate the impact of this scoring system on actual stroke incidence.
In conclusion, this 10-point model provides a simple, cost-effective method for screening high-risk populations, ensuring that those with vulnerable carotid plaques receive the diagnostic attention and treatment they need to prevent a stroke.
Summary updated