Objective To investigate the relation of circulating concentrations of vascular endothelial growth factor (VEGF) for the risk of developing cardiovascular disease (CVD) in a large community-based sample.

Methods We prospectively assessed the relation of circulating VEGF concentrations with the incidence of CVD among 3041 Framingham Heart Study participants (mean age 63.4±11.1 years, 59% women). Multivariable Cox proportional hazards models were estimated adjusting for standard risk factors to VEGF quartiles to incident CVD. Restricted cubic splines were used to examine the linearity of the association.

Results After a mean follow-up of 8.8 (±2.8) years, 527 individuals experienced a first CVD event. Compared with participants in the first VEGF quartile, individuals in the second VEGF quartile had a 34% increased risk for future CVD (HR 1.34, 95% CI 1.03 to 1.74; p value=0.03) and individuals in third quartile had a 59% higher risk (HR 1.59; 95% CI 1.23 to 2.05, p value=0.0003). Individuals in the highest VEGF quartile had a similar cardiovascular risk as compared with those in the lowest VEGF quartile (HR 1.18, 95% CI 0.91 to 1.53, p value=0.21). Evaluation of restricted cubic splines confirmed the nonlinear, inverted U-shaped relation of serum VEGF and CVD events (p<0.0001 for model fit, p=0.006 for non-linearity).

Conclusions Circulating VEGF concentrations exhibit a complex non-linear (inverted U-shaped) relation with the risk of developing CVD events, with the lowest risk experienced at the lower and upper end of the distribution. The underlying pathophysiological mechanisms remain to be elucidated.