Table 3. Results of multiple regression analysis showing the association of cardio-ankle vascular index and/or each obesity-related index with carotid mean intima-media thickness in all participants (n = 7,750).

a) Analysis using either CAVI or each obesity-related index as an independent variable
Variable	Carotid mean intima-media thickness
	Model 1		Model 2
	Coefficient (β)	P value	Coefficient (β)	P value
CAVI	0.022	0.065	0.013	0.265
Body mass index	0.173	< 0.0001	0.140	< 0.0001
Percent body fat	0.123	< 0.0001	0.075	< 0.0001
Waist circumference	0.159	< 0.0001	0.124	< 0.0001
Visceral fat area	0.128	< 0.0001	0.085	< 0.0001
CAVI, cardio-ankle vascular index. Model 1 was adjusted for age, gender, and smoking status. Model 2 was further adjusted for systolic blood pressure, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglyceride, fasting blood glucose, and creatinine, plus those for Model 1.

b) Analysis using both CAVI and each obesity-related index as independent variables
Variable	Carotid mean intima-media thickness
	Model 1		Model 2
	Coefficient (β)	P value	Coefficient (β)	P value
CAVI	0.074	< 0.0001	0.058	< 0.0001
Body mass index	0.188	< 0.0001	0.155	< 0.0001
CAVI	0.037	< 0.01	0.023	0.056
Percent body fat	0.126	< 0.0001	0.077	< 0.0001
CAVI	0.052	< 0.0001	0.038	< 0.01
Waist circumference	0.166	< 0.0001	0.131	< 0.0001
CAVI	0.030	< 0.05	0.019	0.104
Visceral fat area	0.130	< 0.0001	0.089	< 0.0001
CAVI, cardio-ankle vascular index. Model 1 was adjusted for age, gender, and smoking status. Model 2 was further adjusted for systolic blood pressure, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglyceride, fasting blood glucose, and creatinine, plus those for Model 1.