Background: severe hyperphosphatemia is a risk factor for vascular calcification (VC) among chronic kidney disease patients, but whether serum phosphate (P) is associated with VC in the general population is unclear.

Objective: to assess the relation between P and VC in the coronary arteries in men and women from the Rotterdam Study.

Methods: 1877 subjects with fasting P levels underwent coronary artery calcification (CAC) measurements through electron-beam tomography with assessment of Agatston’s scores. Linear models between P and CAC were adjusted for age, BMI, smoking (basic model) and subsequently for 25OHD, calcium, C-reactive protein, glucose and cholesterol:HDL ratio (full model). We tested for causality via Mendelian Randomization (MR) applying a genetic P score composed of 05 SNPs of a previous genome wide association study. MR assumptions were tested.

Results: P was positively related to CAC scores (basic model) in sex-combined analysis (β: 0.71(0.49-0.92)) but a significant sex-interaction was found (p _interaction=0.005) with a stronger association in men (β: 1.03(0.74-1.32), n=871) than women (β: 0.41(0.09-0.73), n=1006). Adjustments in the full model did not substantially change results (sex-combined β: 0.66(0.44-0.89), neither exclusion of subjects with GFR<60 mL/min (sex-combined β: 0.67(0.43-0.91)) or restriction to normal P (sex-combined β: 0.73(0.47-0.99)). MR results (P genetic score F-statistic >10; not related to age, BMI, smoking) support a causal role of P in CAC (sex-combined: unweighted score β: 2.05(0.04-4.05), weighted score β: 2.12(0.14-4.10); n=1693).

Conclusions: P seems to be causally related to CAC scores in a population-based setting. Replication and further research into sex differences is warranted.