Retinal microvascular parameters and cardiovascular health, obesity and inflammation in children and mid-life adults

Abstract

Background The microcirculation makes up approximately three-quarters of the circulatory system but has attracted little attention as a target for cardiovascular disease (CVD) prevention, particularly in early life. Retinal microvascular parameters allow noninvasive assessment of the microcirculation, with adverse parameters (e.g., narrower retinal arterioles, wider venules) predicting CVD events in adults. In children, it is largely unknown when such microvascular changes become evident, their determinants and the underlying mechanisms.

Aims In a large population-based sample of Australian children and mid-life adults (their parents), I aimed to determine whether: 1) adverse retinal vascular calibre and preclinical large arterial phenotypes covary; 2) body mass index (BMI) and waist-to-height ratio (WHtR; child only) over the preceding decade predict retinal vascular calibre; 3) inflammation mediates the association between obesity and retinal vascular calibre; 4) traditional CVD risk factors and large arterial phenotypes are related to retinal geometric parameters.

Methods Participants: 1,288 11–12 year-olds (51% girls) and 1,264 parents (88% mothers). Cross-sectional measures: Retinal vascular calibre and (child only) geometric parameters; large arterial phenotypes (pulse wave velocity, PWV; carotid arterial elasticity; carotid intima-media thickness, CIMT); blood pressure (BP); BMI; WHtR; and metabolites (e.g., LDL, HDL cholesterol and the inflammation marker Glycoprotein Acetyls, GlycA). Biennial longitudinal measures: BMI and (child only) WHtR. Analysis: In all aims, linear regression models were used, and Aim 3 used causal mediation analysis.

Results Aim 1: In children, narrower retinal arteriolar and wider venular calibre showed modest associations with faster PWV and lower elasticity but not with CIMT. In adults, results were stronger, and there was weak evidence of an association between wider venular calibre and higher CIMT. Aim 2: In children, higher BMI from age 4–5 years onwards was increasingly associated with adverse arteriolar calibre, but not venular calibre at 11–12 years. Effects were similar in adults. Less favourable BMI and WHtR trajectories predicted narrower arteriolar calibre. Aim 3: Compared to children with normal BMI, those with obesity had: (i) narrower arteriolar calibre, which was not mediated by inflammation; and (ii) wider venular calibre, which was partly mediated by inflammation. In adults with obesity, this association between obesity and wider venular calibre was fully mediated by inflammation. Findings were similar for WHtR. Aim 4 (in children only): BMI, systolic BP and PWV showed modest associations with lower arteriolar fractal dimensions, whereas only systolic BP was associated with arteriolar tortuosity. There was less evidence of associations with venular geometric parameters.

Conclusion Preclinical phenotypes of large arteries and microcirculation have some shared, but mainly unique pathways to CVD, with shared pathways becoming more evident across the life course. The adverse impact of obesity on retinal microvasculature begins early in life, more markedly on arteriolar than venular parameters, with modest venular effects largely mediated by inflammation. In children, retinal geometric parameters did not add novel information about CVD risk beyond the vascular calibre. The microvascular parameter, retinal vascular calibre, may offer additional value to CVD prevention from childhood onwards, especially when combined with other risk assessment measures.