Relation of vascular growth factors with CT-derived measures of body fat distribution: the Framingham Heart Study

Abstract

Background: Visceral adiposity is associated with metabolic risk. Given that angiogenesis is a key feature of adipogenesis, variation in the association of levels of circulating vascular growth factors (and their soluble receptors) with distinct body fat compartments may explain differences in the systemic pathogenicity of regional fat depots.

Methods and results: Four body fat compartments [visceral adipose tissue (VAT), sc adipose tissue (SAT), thoracic periaortic fat, and pericardial fat] derived from computed tomography were related to serum concentrations of vascular endothelial growth factor (VEGF), the soluble VEGF receptor (fms-like tyrosine kinase-1), hepatocyte growth factor (HGF), and angiopoietin-2 and its soluble receptor (soluble tyrosine kinase with immunoglobulin-like and EGF-like domains 2 sTie-2) in 1806 Framingham Heart Study participants (mean age 44.9 yr, 44.5% women). In multivariable models, we observed positive associations between several fat compartments and VEGF and HGF levels. The magnitude of the associations were similar for VAT, SAT, and periaortic fat. We observed effect modification by sex. A stronger association was observed between VAT and HGF levels in women; higher VAT and periaortic fat were jointly associated with higher HGF concentrations (P=0.02 and P=0.051, respectively). In women within the highest tertile of VAT, HGF levels significantly increased with higher periaortic fat (P=0.0005).

Conclusions: In our large community-based sample, greater adiposity was associated with higher circulating growth factor levels in general. Additional studies are warranted to confirm the stronger association of VAT and periaortic fat with HGF in women and to examine its potential contribution to the sex-related differences in cardiometabolic risk.

Fig. 1.

Mean HGF levels in women partitioned by tertiles of VAT and tertiles of periaortic fat. Subgroup sizes for each periaortic fat tertile within VAT tertile were as follows: 89, 89, and 90 in VAT tertile 1; 89, 90, and 89 in VAT tertile 2; and 89, 90, and 89 in VAT tertile 3. Error bars represent se. P values are given for linear trend in means of log-transformed HGF levels.