Background: several studies have been pointing towards a non-linear relationship between serum 25(OH)D and cardiovascular disease. Next to vitamin D deficiency, also higher levels of 25(OH)D have been reported to be associated with increased cardiovascular risk. We aimed to investigate the nature of the relationship between serum 25(OH) D and measures of arterial stiffness and arteriosclerosis in an elderly population. Design: cross-sectional. Setting/subjects: a subgroup of the B-PROOF study was included to determine associations between serum 25(OH)D and arterial stiffness and atherosclerosis (n = 567, 57% male, age 72.6 +/- 5.6 years, mean serum 25(OH)D 54.6 +/- 24.1 nmol/l). Methods: carotid intima media thickness (IMT) was assessed using ultrasonography and pulse wave velocity (PWV) was determined with applanation tonometry. Associations were tested using multivariable restricted cubic spline functions and stratified linear regression analysis. Results: the associations between serum 25(OH)D and carotid IMT or PWV were non-linear. Spline functions demonstrated a difference between 25(OH)D deficient and sufficient individuals. In serum 25(OH)D sufficient participants (a parts per thousand yen50 nmol/l; n = 287), a positive association with IMT and serum 25(OH)D was present (beta 1.24; 95%CI [0.002; 2.473]). PWV levels were slightly lower in vitamin D deficient individuals, but the association with 25(OH)D was not significant. Conclusion: our study demonstrates that associations of serum 25(OH)D and PWV and IMT in an elderly population are not linear. In particular from serum 25(OH)D levels of 50 nmol/l and up, there is a slight increase of IMT with increasing 25(OH)D levels.