Patients of this observational, case-control, single centre study were assessed in the Cardiology and Ophthalmology Departments of the Virgen de la Victoria University Hospital of Malaga between the months of January-March 2019. Inclusion criteria for the study were suspicion of having a coronary heart disease based on presentation of oppressive chest pain at rest or on with an increase in activity or stress, with or without the presence of dyspnoea at the time of the consultation. Exclusion criteria were the existence of advanced kidney failure, defined as a glomerular filtration rate lower than 30 ml/min, suffering from any disease that reduced life expectancy to less than one year, having presented another cardiovascular event before this study, suffering from diabetic retinopathy or amblyopia in either eye or having undergone retinal photocoagulation treatment in the past.
To confirm the existence of CAD, the selected patients underwent coronary angiogram and/or a computerised tomography scan of the coronary arteries. Depending on the angiogram results, we defined CAD as such with findings of a minimum involvement of 70% in any major epicardial artery (right coronary, anterior descending or marginal circumflex) or more than 50% in the left main coronary artery. Based on the computerised tomography of coronary arteries, we defined CAD as such where there was involvement of at least 50% of any coronary artery. After these results, patients we distributed into two study subject groups: patients without coronary lesions and patients with coronary lesions. After their clinical stabilisation, patients underwent a full ophthalmological examination within the first ten days after the onset of the clinical symptoms that were the reason for the cardiovascular study.
The same ophthalmologist conducted the ophthalmological examinations at the same time of day (15 pm to 18 pm). Before carrying out examinations, we collected data about the CRF for the onset of cardiovascular disease for each patient, such as sex, age, presence of diabetes, high blood pressure, hypercholesterolemia, smoking habit or obesity. The ophthalmological examination consisted of evaluating different clinical parameters, such as best-corrected visual acuity (BCVA) with numerical eye charts, intraocular pressure (IOP) using a Perkins Mk3 tonometer (Haag-Streit, Essex, UK), Schirmer's test (ST) using Schirmer-Plus® strips (GECIS, Neung-sur-Beuvron, France), central corneal thickness (TCCT and PCCT), axial length (AL) and several variables obtained by means of optical coherence tomography (OCT) and OCT-angiography. In addition, as laboratory variables, we studied the presence of cytokines and other inflammatory mediators in samples of tears.
Central corneal thickness was automatically measured with the Orbscan® IIz (Bausch & Lomb, Rochester, USA) topographer and, after administration of tetracaine and oxybuprocaine anaesthetic eye drops, manually with the OcuScan® RxP (Alcon Laboratories, Texas, USA) ultrasound pachymeter, with calculation of the arithmetic mean of 10 consecutive measurements taken from each eye with the patient seated. AL was automatically obtained with partial coherence interferometry using the IOLMaster® 500 (Carl Zeiss Meditec AG, Jena, Germany) optical biometer. Cirrus™ (Carl Zeiss Meditec AG, Jena, Germany) high-definition (HD)-OCT was used for OCT, with the equipment providing automatically obtained values for central macular thickness (CMT), macular cube volume (MCV), mean macular thickness (MMT), retinal nerve fibre layer thickness (RNFLT), ganglion cell layer thickness (GCLT) and ganglion cell layer minimum thickness (GCLMT). Choroidal thickness (CT) was calculated manually using the device's ruler from the images it produced, and was defined as the distance between the external hyper-reflective band of the retinal pigment epithelium and the internal hyper-reflective band of the sclera. This was expressed as the arithmetic mean of 10 measurements: 5 on the horizontal axis and 5 on the vertical axis, with a separation of 500 micra between them to either side of subfoveal area. For inclusion in this study, the minimum acceptable image quality was 6 out of 10. The OCT-angiography was performed with DRI OCT Triton™ plus (Topcon Medical Systems, Oakland, USA) and the device software automatically provided the values for central vascular density in the different layers of the retina (ILM, RNFL-GCL, GCL-IPL, IPL-INL). In the case of prior corneal surgery or a clinically significant eye cataract, we excluded the values obtained for these eyes for BCVA, IOP, TCCT, PCCT and the OCT and OCT-angiography.
Conducting ST served also to collect tear samples from each patient. The paper strip was placed to rest in the lower base of each eye, without previous application of topical anaesthetic, and left for 5 minutes. The study excluded samples from any eyes where moisture on the strip measured less than 6 mm after this period. Valid samples were immediately frozen at -80º C after collection, and remained at that temperature until analysis. For protein elution each paper sample was cut into small pieces that were introduced into 100ml of PBS with Tween® 20 at 0.3%, bovine albumin serum at 0.5% and protease inhibitor, for overnight incubation 4º C before collection of supernatant. Measurement of the total amount of protein in each sample used the NanoDropTM One (ThermoFischer Scientific, Waltham, MA, USA) spectrophotometer with absorbance measured at 280 nm. Detection of cytokines and inflammatory mediators followed the instructions of the Bio-Plex Pro TM kit Human Cytokine 27-Plex Assay (Bio-Rad Laboratories, Hercules, CA, USA). The 27 cytokines and inflammatory mediators analysed were interleukin (IL)-1b, IL-1RA, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8,, IL-9, IL-10, IL-12p70, IL-13, IL-15, IL-17, basic fibroblast growth factor (FGF), eotaxin (EO), granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon (IFN)-g, chemokine (IP)-10, monocyte chemoattractant protein (MCP)-1, macrophage inflammatory protein (MIP)-1a, platelet-derived growth factor (PDGF)-bb, MIP-1b, chemoquine ligand 5 (RANTES), tumour necrosis factor (TNF)-a and vascular endothelial growth factor (VEGF).
Data were analysed with Stata® 17 (StataCorp LLC, Texas, USA). Characteristics of participants were described as a mean and standard deviation for continuous variables and percentage for categorical variables. Group comparisons were carried out using a Student’s t-test or Chi-squared test as appropriate.
For each variable associated with the presence of coronary lesions in the bivariate analysis, a logistic regression model was estimated that included the CRF: sex, age, diabetes, high blood pressure, hypercholesterolemia, smoking habit and obesity. We assessed the statistical significance of each variable using the likelihood-ratio test based on the models with and without the variable (only with CRF). For variables that were statistically significant, we estimated the area under curve (AUC) associated with the logistic regression model. To estimate the predictive capacity of these variables, we compared the areas under curve using the Chi-squared test, taking the AUC of the logistic regression model that only included CRF. Finally, a logistic regression prediction model was constructed that included the CRF and all variables whose logistic regression models showed an AUC higher than the only-CRF model.