Article Impact Level: HIGH Data Quality: STRONG Summary of Heart. https://doi.org/10.1136/heartjnl-2024-324705 Dr. Mayinuer Yusufu et al.
Points
- The research analyzed retinal vascular parameters in 45,161 UK Biobank participants over 12.5 years, identifying 749 incident strokes and evaluating 118 vascular indicators to predict stroke risk.
- Changes in retinal vascular parameters, particularly density, caliber, complexity, and twistedness, were significantly associated with increased stroke risk, with alterations linked to a 10% to 19.5% higher risk.
- Incorporating retinal parameters into traditional risk models improved predictive accuracy, increasing the ROC curve area from 0.739 to 0.752 (p<0.001), demonstrating their added value in stroke risk assessment.
- Retinal vascular analysis offers a noninvasive, cost-effective tool for stroke risk stratification. Fundus photography is accessible, making it particularly suitable for primary care and low-resource settings.
- The observational design and sample homogeneity, with a predominantly white participant cohort, limit the findings. This highlights the need for further validation in diverse populations.
Summary
This study aimed to evaluate the utility of retinal vascular parameters as predictors of stroke risk, utilizing data from the UK Biobank cohort. A total of 45,161 participants were included, with a median follow-up of 12.5 years. During this period, 749 incident strokes were recorded. The researchers employed the Retina-based Microvascular Health Assessment System to analyze retinal fundus images, focusing on 118 vascular indicators across five categories: caliber, density, twistedness, branching angle, and complexity. After adjusting for traditional cardiovascular risk factors, 29 retinal parameters were significantly associated with stroke risk. Notably, changes in density indicators were linked to a 10% to 19% increase in stroke risk, while alterations in caliber indicators were associated with a 10% to 14% increased risk. Decreases in complexity and twistedness indicators correlated with a 10.4% to 19.5% higher risk of stroke.
The study’s results highlighted the potential of retinal vascular analysis as a non-invasive tool for stroke risk prediction. Incorporating retinal parameters into traditional risk models increased the area under the receiver operating characteristic (ROC) curve from 0.739 to 0.752 (p<0.001). This suggests that retinal vascular parameters offer additional predictive power in assessing stroke risk. This makes it a viable tool for use in primary healthcare, particularly in low-resource settings. Furthermore, these findings underscore the importance of density and complexity parameters in predicting stroke risk, specifically emphasizing their potential applications in routine clinical assessments.
The study concluded that retinal vascular parameters provide valuable insights into stroke risk, outperforming traditional risk models and offering a practical approach to risk stratification. Given the accessibility of fundus photography, this method could be a cost-effective, widely implementable tool for stroke risk assessment, especially in primary care settings. However, the researchers acknowledged that the findings may be limited by the study’s observational design and the homogeneity of the sample, predominantly composed of white participants.
Link to the article: https://heart.bmj.com/content/early/2025/01/03/heartjnl-2024-324705
References Yusufu, M., Friedman, D. S., Kang, M., Padhye, A., Shang, X., Zhang, L., Shi, D., & He, M. (2025). Retinal vascular fingerprints predict incident stroke: Findings from the UK Biobank cohort study. Heart. https://doi.org/10.1136/heartjnl-2024-324705
