Table 3.

Logistic regression models for association of ankle-brachial index (ABI) and femoral-ankle pulse-wave velocity (faPWV) with cardiovascular disease at visit 5.

	Model Coefficients				Overall Model Fit
Continuous	OR	95% CI		P Value	AIC	R2McF	X2 (df)	P Value
Univariable
ABI	0.12	0.07	0.20	<0.001	4407	.015	64.30	<0.001
faPWV	0.91	0.87	0.94	<0.001	4447	.005	23.81a	<0.001
Model 1
CVD risk factors					3971	.117	522.00	<0.001
Model 2
ABI	0.12	0.07	0.21	<0.001	3921	.129	574.00b	<0.001
faPWV	0.92	0.88	0.96	<0.001	3960	.120	534.52a,b	<0.001
Model 3
ABI	0.13	0.07	0.24	<0.001	3917	.130	579.41c	<0.001
faPWV	0.95	0.91	0.99	0.021
Categorical
Univariable
ABI	3.12	2.45	3.98	<0.001	4393	.018	78.20	<0.001
faPWV	1.57	1.34	1.84	0.008	4441	.006	30.02a	<0.001
Model 1
CVD risk factors					3971	.117	522.00	<0.001
Model 2
ABI	2.41	1.85	3.17	<0.001	3932	.126	562.46b	<0.001
faPWV	1.46	1.23	1.74	<0.001	3955	.121	539.89a,b	<0.001
Model 3
ABI	2.25	1.71	2.96	<0.001	3924	.128	572.86c	<0.001
faPWV	1.35	1.12	1.61	0.001

Categorical comparisons: categorical analyses, comparisons represent ≤0.9 vs. >0.9 for ABI, and Q1 (≤9.94 m/s) vs. Q2-Q4 (>9.94 m/s) for faPWV, whereby higher values were set as the reference level following continuous analysis.

Model descriptions: Model 1 - CVD risk factors only including: age, sex, systolic blood pressure, HDL cholesterol, total cholesterol, smoking status and diabetes. Model 2 - CVD risk factors plus ABI or faPWV separately. Model 3 - CVD risk factors plus ABI and faPWV concurrently. Adjustments: All multivariable models were adjusted for antihypertensive medication use, and race/field center.

^aComparisons: vs univariable and multivariable 2 ABI models (non-nested);

^bvs. multivariable 1(P<0.001).

^cvs. multivariable 2~ABI (P<0.001).

Abbreviations: OR, odds ratio; CI, confidence interval; R²McF, McFadden’s Pseudo R²; X², Chi square statistic, df, degrees of freedom.