Association Between Retinal Vascular Occlusions and Stroke in Individuals With Atrial Fibrillation

The CHA₂DS₂‐VASc (congestive heart failure, hypertension, age ≥75 years, diabetes, previous stroke or transient ischemic attack, vascular disease, age 65 to 74 years, and sex category) score is a tool used to guide anticoagulation decisions in patients with atrial fibrillation (AF). ¹ Identifying additional risk factors may help improve stroke risk stratification. Retinal artery occlusion (RAO) and retinal vein occlusion (RVO) have been linked to stroke and established risk factors, including age, hypertension, and ischemic heart disease. ² , ³ Here, we investigated whether RAO or RVO are significantly associated with stroke among individuals with AF, after adjusting for individual risk factors that comprise the CHA₂DS₂‐VASc score. Our goal was to determine whether these retinal vascular events offer novel prognostic information that could refine stroke‐risk stratification in this high‐risk population.

This is a cross‐sectional study approved by Yale’s institutional review board with a waiver of informed consent due to the retrospective nature of the study. The study was conducted in accordance with Health Insurance Portability and Accountability Act regulations and ethical standards consistent with the Declaration of Helsinki. The data that support the findings of this study are available from the corresponding author on reasonable request. Individuals 30 years and older, who underwent an evaluation by an ophthalmologist or optometrist as well as by a general practitioner at Yale University or Yale New Haven Health System between January 2015 and December 2023 and had a diagnosis of AF, were included. Diagnostic records were obtained through the electronic medical record tool, Epic SlicerDicer, and using the International Classification of Diseases, Tenth Revision (ICD‐10), diagnostic codes.

The primary variables of interest were RAO and RVO, and the main outcome was stroke. Multivariable logistic regression models were used to evaluate the relationship between RAO or RVO and stroke in individuals with AF. The differences between groups were evaluated using χ² test for categorical variables and independent t test for continuous variables. Comparisons between models were made using the likelihood ratio test. Given that stroke was the outcome measure in this cross‐sectional analysis, it was not included in the modified CHA₂DS₂‐VASc score calculation. Statistical analyses were conducted using R version 4.4.2 (R Foundation for Statistical Computing) and GraphPad Prism 10.4.1.

A total of 10 134 individuals with AF met the inclusion criteria. We identified 152 and 217 individuals in the cohort with a diagnosis of RAO or RVO, respectively, and 23 individuals with a diagnosis of both. Those with retinal vascular occlusions tended to be older (79 years versus 75 years, P<0.001) compared with those without and there was no significant difference in the distribution of sex between the 2 groups. Individuals with retinal vascular occlusions were more likely to have cardiovascular risk factors including hyperlipidemia (91.3% versus 88.0%, P=0.044), hypertension (97.2% versus 93.1%, P=0.002), diabetes (55.4% versus 48.8%, P=0.011), coronary artery disease (73.7% versus 62.7%, P<0.001), peripheral artery disease (59.7% versus 46.9%, P<0.001), and congestive heart failure (61.7% versus 52.1%, P<0.001).

In patients with AF, both RAO (odds ratio [OR], 5.04 [CI, 3.71–6.90], P<0.001) and RVO (OR, 1.67 [CI, 1.27–2.17], P<0.001) were each significantly associated with stroke in univariable logistic regression. The relationship remained significant in multivariable regression analyses for both RAO (OR, 4.60 [CI, 3.36–6.35], P<0.001) and RVO (OR, 1.47 [CI, 1.11–1.93], P=0.006), adjusting for age, sex, anticoagulation therapy, and the aforementioned cardiovascular comorbidities and risk factors (Table1).

Table 1.

Multivariable Logistic Regression Models Evaluating the Relationship Between Retinal Vascular Occlusion and a Diagnosis of Stroke in Patients with Atrial Fibrillation

Predictor	Multivariable OR (95% CI)	P value
RAO*	4.60 (3.36–6.35)	<0.001
RVO*	1.47 (1.11–1.93)	0.006
RAO†	4.80 (3.51–6.60)	<0.001
RVO†	1.49 (1.13–1.94)	0.004

OR indicates odds ratio; RAO, retinal artery occlusion; and RVO, retinal vein occlusion.

^*Multivariable logistic regression adjusted for age (years), sex, cardiovascular risk factors (hypertension, hyperlipidemia, peripheral arterial disease, coronary artery disease, congestive heart failure, diabetes), warfarin, and direct oral anticoagulation use (dabigatran, rivaroxaban, apixaban, edoxaban).

^† Multivariable logistic regression adjusted for modified CHA₂DS₂‐VASc score.

We then tested whether both retinal vascular occlusions would remain significantly associated with stroke after adjusting for a modified CHA₂DS₂‐VASc score. Both RAO (OR, 4.80 [CI, 3.51–6.60], P<0.001) and RVO (OR, 1.49 [CI, 1.13–1.94], P=0.004) remained significantly associated with stroke, after adjusting for the modified CHA₂DS₂‐VASc score (Table1). In a sensitivity analysis restricted to individuals with no recorded anticoagulation therapy, both RAO (OR, 5.36 [CI, 2.66–11.12], P<0.001) and RVO (OR, 2.02 [CI, 1.16–3.42], P=0.010) were significantly associated with stroke.

We then assessed whether the incorporation of both RAO and RVO enhances the performance of the modified CHA₂DS₂‐VASc score. A logistic regression model incorporating the presence of retinal vascular occlusions with the modified CHA₂DS₂‐VASc score was superior to the modified score alone (likelihood ratio test, P<0.001).

Our results are consistent with previous work. A study by Christiansen et al demonstrated that a history of RVO/RAO is associated with an increased risk of stroke/thromboembolism/transient ischemic attack in patients with AF. ⁴ Our study includes a more diverse patient population and outpatient data and explicitly tested associations with stroke after adjusting for the overall CHA₂DS₂‐VASc score.

Retinal vascular occlusions, specifically RAO, have been shown to be independently associated with AF, highlighting the potential for retinal findings to uncover occult cardiovascular disease. ⁵ Although RAO and RVO arise from distinct causes, they both confer an increased risk of stroke in patients with AF, suggesting a cardiovascular component possibly via mechanisms of increased microemboli formation and vascular disease, respectively. The presence of RAO and RVO in patients with AF may assist in accurate stroke risk stratification and aid clinicians in decisions on anticoagulation and appropriate lifestyle counseling.

As our study is cross‐sectional, we cannot distinguish temporality between stroke and retinal vascular occlusions or infer causation. In addition, use of electronic medical records and ICD‐10 coding introduces potential for misclassification. Moreover, the single‐center setting limits generalizability. Further prospective studies are needed to validate these findings.

Sources of Funding

Christina Jayaraj is funded by the James G. Hirsch Endowed Medical Student Research Fellowship and Taylor Opportunity Student Research Fellowship. Research reported in this publication was supported by the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NHLBI or the NIH under award number T35HL007649. Christine Bakhoum is funded by American Heart Association Award 857722 and K23 DK129836 from the National Institute of Diabetes and Digestive and Kidney Diseases. Mathieu F. Bakhoum is supported by NIH research grant P30CA016359 from the National Cancer Institute, R21EY035090 from the National Eye Institute, by the Office of the Assistant Secretary of Defense for Health Affairs through the Melanoma Research Program under award number HT9425‐23‐1‐1070, and a grant from the Connecticut Lions Eye Research Foundation.

Disclosures

The authors have no disclosures to report.