Abstract
Purpose: :
Central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE) have been reported to have high heritabilities (72 and 78%, respectively) and are strongly negatively associated with diastolic and systolic blood pressure (DBP and SBP, respectively). We aimed to investigate shared genetic or environmental factors underlying the relationship between CRAE and CRVE and their correlation with blood pressure.
Methods: :
Non-simultaneous 30° stereoscopic images, taken with a Kowa camera, were digitized and analysed to determine CRAE and CRVE for 226 monozygotic and 290 dizygotic Caucasian female twin pairs aged between 49 and 79 (mean 62) years. Maximum likelihood variance component modeling compared covariance between MZ and DZ twin pairs. In a univariate heritability analysis, total phenotypic variance is partitioned into additive (A) and dominant (D) genetic and common (C) and unique (E) environmental components. In a bivariate heritability, the covariance between two phenotypes is also partitioned into these variance components with the significance of the component(s) assessed using nested models and a likelihood ratio test.
Results: :
The ACE model was best-fitting for all models fitted (CRAE with CRVE, CRAE with DBP then SBP and CRVE with DBP then SBP). Removing the 3 shared environmental components from the ACE model caused a significant decline in model fit (ΔΧ23=68.2, p=0.000). The phenotypic correlation between CRAE and CRVE (r=0.7) was explained by an additive genetic component of -21% (ΔΧ21 = 3.53, p=0.06), by common environmental factors of 81% and by unique environmental factors of 39%. Most of the covariance between the vessel traits and blood pressure was estimated to be due to environmental factors with minimal additive genetic influence.
Conclusions: :
The phenotypic correlation between blood pressure and retinal vessel diameter and between venules and arterioles seems primarily due to environmental factors, despite each of the traits individually having high univariate heritabilities. There was modest evidence of a small negative genetic covariance between venular and arteriolar diameters, which seems to act in an antagonistic direction to the overall positive phenotypic correlation. This implies that in future gene mapping studies, we might expect most associated genetic variants to specifically influence arterioles or venules, but not both; any influencing both are likely to act in opposite directions.
Keywords: blood supply • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology