Correlation of RETINAL Artery Diameter with Coronary Artery Disease: The RETINA CAD Pilot Study—Are the Eyes the Windows to the Heart?

Cardiovascular disease (CVD) is the leading cause of mortality globally, with an estimated 18 million deaths annually [1]. This significant disease burden requires intensive research and resource allocation into screening and risk stratification to enhance early detection and facilitate therapeutic intervention.

CVD was primarily ascribed to epicardial coronary artery disease (CAD); however, coronary microvascular dysfunction (CMD) also plays a crucial maladaptive role [2]. Current techniques to assess coronary microvascular perfusion are limited and fraught with challenges, such as being invasive and requiring the administration of stress agents or continuous monitoring [3].

Retinal microvasculature, however, can be easily visualized with a noninvasive technique, optical coherence tomography–angiography (OCT–A). Two recent studies have alluded to a bidirectional relationship between retinal microvasculature and incipient cardiovascular risk [6; 5]. Several scoring systems have been developed to angiographically estimate coronary artery disease (CAD) burden [6]. The Synergy between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery (SYNTAX) score was designed through expert consultation, and integrates previous angiographic scores that assess lesion complexity, and it is widely utilized [7, 8].

This study aimed to determine whether there was any correlation between CAD and retinal artery diameter at an academic tertiary medical center in Trinidad and Tobago.

The sample size was calculated using the Massachusetts General Hospital Biostatistics Calculator based on the study performed by Arnould et al., which enrolled 237 patients [4, 5, 13]. The sample size was calculated as 77 patients based on a paired-proportion sample, an alpha (α) value of 0.05, a power (β) of 80%, and a correlation coefficient of 0.8. The study’s biostatistician performed routine statistical analyses with IBM SPSS Statistics version 27.0. Descriptive methods included percentage, frequency, and summary statistics. Continuous variables such as age, SYNTAX score, and retinal artery diameter were presented as means ± standard deviation. The SYNTAX score and retinal artery diameter were not normally distributed and showed skewed data; hence, the median values were used for analysis. Spearman’s rank correlation coefficient was used for correlations, and Mann–Whitney U-tests were done to compare medians between groups. All p-values were two-tailed, and a p-value of < 0.05 was accepted as statistically significant.

A total of eighty (80) patients were approached for participation in the study, and seventy-seven (77) of them agreed to participate (Table 1). The average age of the sample was 57.8 years, ranging from 23 to 88 years (median age 60 years). There were more males [n = 55 (71.4%)] and South Asians [n = 53 (68.8%)]. The most common comorbidity was preexisting cardiovascular disease (CVD) [n = 64 (83.1%)] followed by hypertension (HTN) [n = 46 (59.7%)], dyslipidemia (HLD) [n = 36 (46.8%)], and diabetes mellitus (DM) [n = 33 (42.9%)]. SYNTAX scores ranged from 0 to 36, with a mean score of 12.0 (± 9.3) and a median of 11 (interquartile range 4.5–11). Right retinal artery diameter ranged from 146 to 313 μm with a mean of 196.83 μm (± 37.142 μm) and median of 188 μm. Left retinal artery diameter ranged from 24 to 471 μm with a mean of 193.45 μm (± 53.465 μm) and a median of 187 μm. A statistically significant negative correlation was observed between retinal artery diameter and SYNTAX score for both eyes among all participants (Fig. 1 and Table 2). This was also demonstrated for the female gender (Table 2). DM, insulin and oral hypoglycemic therapy and SYNTAX score were also found to have a significant relationship (p = 0.005, 0.015, and 0.006, respectively) (Table 3). No other comorbidities, such as hypertension and dyslipidemia, displayed a statistically significant relationship.

The pathophysiology of acute coronary syndromes (ACS) was traditionally attributed to obstructive CAD, whereas CMD primarily results from architectural changes and endothelial dysfunction [2]. They can both co-occur due to shared incipient cardiovascular risk factors. CMD may accentuate inducible ischemia from epicardial CAD with sequelae of increased major adverse cardiovascular events (MACE) [2]. Emerging data support that CMD may affect other vascular territories due to similar caliber size [14]. Retinal artery atherosclerosis has been shown to correlate with CAD [15]. Previous studies have implied that viewing retinal artery microvasculature could be a looking glass for latent CAD, and be utilized in assessing cardiovascular risk [4, 5, 16]. The eye offers direct access to the microvasculature and through high-fidelity OCT–A, the retinal artery may be evaluated with high reproducibility [17, 18]. Several studies have demonstrated that changes in retinal vessel diameter reflect an accentuated risk of CVD and cerebrovascular events (CVE)[15, 19, 20]. It is pertinent to study the role of microvasculature because of the established link between small vessels and diseases such as T2DM [21].

OCT–A is a novel imaging technique that rapidly generates volumetric angiographic images (within seconds), which are integral in evaluating retinal and systemic vascular diseases. Compared with gold-standard fluorescein angiography (FA) and indocyanine green angiography (ICGA), advantages of OCT–A are that it is noninvasive, can attain specific depths, uses motion contrast instead of intravenous dye, provides accurate size and localization information, visualizes both the retinal and choroidal vasculature, and shows structural and blood flow information in tandem. The disadvantages include a limited field of view, inability to view leakage, increased artifacts, and inability to detect blood flow below the slowest detectable flow [22]. An Israeli study in which the age of the patients was 25–47 years without ocular pathology or comorbidities, revealed a mean diameter of the retinal artery of 135.73 ± 15.64 μm [23]. The diameter of retinal vessels measured from OCT–A scans are generally wider than other modalities such as adaptive optics ophthalmoscopy, likely attributed to different scanning protocols [24]. The OCT–A retinal imaging community is urged to define and harmonize best practices to enhance accuracy, reliability, and wider collaboration [25].

Several studies demonstrated conflicting results between retinal vasculature and CAD, albeit with the major caveat of employing different CAD scoring and ocular assessment modalities, excluding OCT–A [26]. In a large Danish study, retinal vessel diameters were classically associated with traditional cardiovascular risk factors; however, they were not associated with all-cause mortality despite adjusting for confounding factors [27]. Allon et al. reported in a systematic review that a narrower central retinal artery was associated with an increased incidence of ACS but not SYNTAX or coronary artery calcium (CAC) scores [5; 5]. More recent studies have evaluated this relationship using OCT–A, and the majority revealed an association between vascular density and CAD. Xu et al. established that retinal artery lumen diameter, retinal artery outer diameter, and arteriovenous ratio had a statically significant negative correlation with the severity of CAD [28]. The EYE-Myocardial Infarction (EYE-MI) study concluded that retinal vasculature density was associated with higher cardiovascular risk factor burden, and was associated with a higher SYNTAX score. Wang et al. found that reduced retinal vascular density and choroidal blood flow were associated with CAD [29]. Our study evaluated retinal artery diameter via OCT–A and determined the severity of CAD, calculated by the SYNTAX score. Retinal artery diameter was negatively correlated with the SYNTAX score, with p-values of 0.0038 and 0.003 for the left and right eyes, respectively. This result was consistent with previous studies utilizing OCT–A [4, 5].

There was also a statistically significant negative correlation between retinal artery diameter and SYNTAX score in women but not men. This trend was not noted in other OCT–A studies; however, more extensive population studies assessing alternative retinal microvasculature metrics utilizing other assessment modalities alluded to a similar signal [4, 5, 28, 30, 31]. The Atherosclerosis Risk in Communities Study (ARIC) study demonstrated that narrower retinal arterioles had a higher risk of CAD and cerebrovascular disease in females but not males [31, 32]. A meta-analysis performed in 2010, which evaluated over 20,000 patients, found that retinal disease was independently associated with an increased risk of CAD in women but not men [33]. Conversely, Kromer et al. employed OCT–A to evaluate retinal microvasculature in male-only patients with ACS and found no difference compared with controls [34]. There is a disparity between cardiovascular mortality between sexes, with women receiving less guideline-recommended pharmacotherapy and invasive strategies [35]. Although most women experience obstructive CAD as their culprit etiology for ACS, they also have a higher incidence of myocardial infarction with nonobstructive CAD (MINOCA), spontaneous coronary artery dissection (SCAD), and CMD compared with men [35, 36].

There is a compelling link between DM and CAD [37, 38]. Our study displayed a statistically significant relationship between DM and right eye retinal artery diameter, concordant with the EYE-MI study [4, 5]. Retinal atherosclerosis was associated with DM; however, hypertension, smoking, and age did not demonstrate an association, likewise in our study [39]. In DM, a complex milieu of proinflammatory cytokines, VEGF, activated protein C, platelet-derived growth factor, and advanced glycation end-products result in endothelial dysfunction, which is crucial in developing CAD and CMD [37]. Subsequently, macrophage and vascular smooth muscle cell proliferation lead to atheroma formation. As this process advances, the lumen becomes compromised, detected on CAG. Several studies have revealed that although men may have higher absolute rates of MACE and sudden cardiac death (SCD), relative rates were substantially higher for women across all age ranges [40‐42]. Diabetic angiopathy can lead to negative remodeling and, as sequelae, smaller caliber vasculature [43].

This study is novel in that it is the first in the Caribbean to demonstrate a statistically significant negative correlation between SYNTAX score and retinal artery diameter. The literature is not replete about potential ethnic variations and retinal artery diameter and CAD [44]. Trinidad has a population of 1.4 million and is considered to be multi-ethnic and diverse. The vast majority of adult mortality is attributed to CVD [45‐47]. The large burden of CVD in low- and middle-income countries (LMICs) requires opportunistic screening, and OCT–A provides a potential avenue as it is relatively rapid and inexpensive to perform. OCT–A is an emerging and powerful technology that has tremendous potential to shift the paradigm in detecting retinal disease.

This study has several limitations. This monocentric study is based at an academic medical center, the only one that can perform CAG and OCT–A in Trinidad. As a result, these clinically informative data may not be externally valid to other subpopulations.

Potentially, there can be elements of observer, measurement, reporting, and misclassification bias; for example, the interventional cardiologist inaccurately calculating a SYNTAX score or the ophthalmologist inaccurately calibrating retinal artery diameter, despite being fully trained and qualified by their respective specialty boards. This could have been improved upon with team consensus; however, Trinidad is limited with respect to available specialist personnel. Data entry by the study team could have also been compromised due to factors such as fatigue or clinical caseload, despite being quality controlled and assured by a biostatistician.

This study displayed a similar prevalence of DM compared with other local studies (circa 45–55%) in Trinidad; however, there was a preponderance of male and South Asian patients (both approximately 70%), which may infer an inherent selection bias [45, 48]. As mentioned, DM is implicated in negative remodeling and may confound results in an unadjusted analysis [49].

This study did not include a comprehensive assessment of other baseline retinal metrics or parameters such as retinal venule dimensions, arteriovenous ratio and arteriovenous nicking, vessel tortuosity, or retinopathy [4, 5]. Additionally, it has been demonstrated that Trinidadian South Asian patients have accentuated platelet reactivity and a smaller caliber of coronary arteries, which may contribute to decreased retinal artery diameter [45]. Most studies, such as EYE-MI, are predominantly based on Caucasian populations, and these results may not be applicable if ethnic-related disparities occur with respect to the prevalence of comorbidities and retinal artery diameter [4, 5, 50].

A significantly negative correlation was observed between retinal artery diameter and SYNTAX score. This study alludes to the practical use of optical coherence tomography–angiography (OCT–A) as a noninvasive diagnostic modality for patients with cardiovascular disease (CVD). Further large-scale, multicentric studies are required to confirm these exploratory findings.