The clinical and biochemical characteristics of the study sample are given in . Overall 1,543 women and 1,680 men were available for the current analyses. The mean age was 51.8 years in women and 49.5 years in men and the median VAT/SAT ratio was 0.39 in women and 0.84 in men.
Characteristics of the study sample
Pearson correlation coefficients of log-transformed VAT/SAT ratio with cardiometabolic traits are given in . The VAT/SAT ratio was positively correlated with age in both sexes (r=0.42 and r=0.37). In women, VAT/SAT ratio was weakly correlated with increasing BMI (r=0.06) and waist circumference (r=0.10), whereas in men there were inverse correlations of VAT/SAT ratio with BMI and waist (r=–0.14 and r=–0.16). In women, the VAT/SAT ratio showed a stronger correlation with VAT (r=0.53) than with SAT (r=–0.11), whereas in men correlations were of similar magnitude for VAT and SAT (r=0.42 and r=–0.43). In both women and men, a higher VAT/SAT ratio was correlated with more adverse levels of the risk factors examined.
Age-adjusted, sex-specific Pearson correlation coefficients between log-transformed VAT/SAT ratio and cardiometabolic risk factors
Age-adjusted means or prevalence of cardiovascular risk factors, partitioned by sex and tertiles of VAT/SAT ratio, are provided in . In women, the prevalence of hypertension, impaired fasting glucose, insulin resistance (defined by HOMA-IR), diabetes, the metabolic syndrome and current smoking increased significantly with tertiles of VAT/SAT ratio (all p<0.0001 for trend, except diabetes: p=0.001). In men, trends were similar, but generally weaker.
Age-adjusted, sex-specific cardiometabolic risk factors stratified by tertiles of VAT/SAT ratio
We next explored the association of VAT/SAT ratio with cardiometabolic traits in multivariable adjusted models (). In women, higher VAT/SAT ratio was associated with higher levels of all assessed cardiometabolic traits (or with lower levels for HDL-cholesterol) reflecting blood pressure, glucose homeostasis and dyslipidaemia (all p< 0.0003). In men, findings were similar, although associations were generally weaker and did not reach significance (p>0.05) for fasting glucose, impaired fasting glucose, insulin resistance and diabetes.
Table 4 Sex-specific, multivariable adjusted association of VAT/SAT ratio with cardiometabolic risk factors. Data are presented as betas (for continuous traits) and odds ratios (for dichotomous traits) with 95% confidence intervals for the condition per 1 SD (more ...)
To assess whether VAT/SAT ratio is associated with cardiometabolic traits independent of BMI, we additionally adjusted our models for BMI. In women, associations of VAT/SAT ratio with cardiometabolic risk factors remained fairly stable after BMI adjustment. In men, BMI-adjustment strengthened several associations, including those with fasting glucose (p=0.0005), impaired fasting glucose (p=0.01) and insulin resistance (p<0.0001). depicts adjusted means of systolic blood pressure, fasting glucose, HDL-cholesterol and log-transformed triacylglycerols, partioned by sex-specific tertiles of VAT/SAT ratio. Additional adjustment for waist circumference did not materially change our findings (data not shown).
Fig. 1 VAT-SAT ratio tertiles and risk factors. Multivariable (including BMI)-adjusted means of fasting glucose (a), systolic blood pressure (SBP) (b), log-transformed triacylglycerols (c) and HDL-cholesterol (HDL) (d), partitioned by sex-specific tertiles of (more ...)
To further explore whether the VAT/SAT ratio is a correlate of cardiometabolic risk independent of VAT, we additionally adjusted our models for VAT (; the effect estimates for VAT in these models are given in electronic supplementary material [ESM] Table 1
). In the VAT-adjusted models, several associations of VAT/SAT ratio in women were attenuated, but those with lower HDL-cholesterol, higher triacylglycerols (both p
<0.0001) and higher prevalence of hypertension (p
=0.02), diabetes (p
=0.01) and the metabolic syndrome (p
=0.005) remained significant. Similarly, in men, additional VAT-adjustment attenuated some associations, but those with higher systolic (p
=0.006) and diastolic blood pressure (p
=0.03), higher fasting glucose (p
=0.0005), lower HDL-cholesterol and higher triacylglycerols (both p<0.0001) and higher prevalence of diabetes (p
=0.006) remained significant.
To assess the potential impact of menopause on our observations in women, we stratified our female study sample by menopausal status in a secondary analysis. We observed similar associations between VAT/SAT ratio and cardiometabolic risk in pre- and postmenopausal women (ESM Table 2