Table 3.

Multivariable linear regression for brachial artery FMD (%)

	Unstandardized coeffi- cients		Standardized coefficients		P-value
	B	Std. error	Beta	t
CRP
Baseline brachial artery diameter, cm	−19.696	3.256	−0.462	−6.050	0.000
Age, years	−0.077	0.027	−0.215	−2.886	0.005
Systolic blood pressure (BP), mmHg	−0.019	0.014	−0.108	−1.389	0.167
LDL cholesterol, mg/dl	−0.012	0.007	−0.135	−1.792	0.076
Waist to hip ratio	−2.700	4.074	−0.053	−0.663	0.509
Diabetes mellitus	1.194	0.793	0.115	1.506	0.135
Positive smoking history	−0.727	0.437	−0.125	−1.665	0.099
CRP, ng/mla	−0.391	0.291	−0.104	−1.345	0.181
Serum amyloid P (SAP)
Baseline brachial artery diameter, cm	−19.989	3.216	−0.469	−6.215	0.000
Age, years	−0.078	0.026	−0.217	−2.959	0.004
Systolic BP, mmHg	−0.019	0.013	−0.110	−1.448	0.150
LDL cholesterol, mg/dl	−0.010	0.007	−0.118	−1.559	0.122
Waist to hip ratio	−2.445	3.965	−0.048	−0.617	0.539
Diabetes mellitus	1.404	0.787	0.135	1.785	0.077
Positive smoking history	−0.741	0.430	−0.127	−1.722	0.088
SAP, ng/mla	−1.094	0.498	−0.165	−2.198	0.030

LDL, low-density lipoprotein. In the two multivariable linear regression models, the dependent variable was brachial artery FMD (%) and the independent variables that were sequentially added were the baseline brachial artery diameter, age, systolic BP, LDL cholesterol, waist to hip ratio, diabetes mellitus, smoking history, and either CRP or SAP. The standardized beta coefficient is based on a one standard deviation of change in the exposure on the outcome and provides context to each variable’s contribution to brachial artery FMD variance.

^aC-reactive protein (CRP) and SAP are log-transformed for analysis.