Table 2.

Results of multivariate analysis

Variable	β	SE (β)	Odds ratio (95% CI)	P value
PCA versus. MCA*
Model based on demographic, vascular risk, clinical, and topographic data*
Hemianopia	1.861	0.212	6.43 (4.25-9.74)	0.000
Lacunar infarction	0.780	0.219	2.18 (1.42-3.35)	0.000
Symptom-free at discharge	0.653	0.230	1.92 (1.22-3.02)	0.005
Limb weakness	-2.322	0.176	0.10 (0.07-0.14)	0.000
Speech disorders (dysarthria)	-1.102	0.175	0.33 (0.24-0.47)	0.000
Cardioembolic infarction	-0.436	0.214	0.65 (0.43-0.98)	0.042
PCA versus ACA
Model based on demographic, vascular risk, clinical, and topographic data†
Sensory symptoms	0.8610	0.363	2.36 (1.16-4.82)	0.018
Limb weakness	-2.258	0.437	0.11 (0.04-0.25)	0.000
Cardioembolic stroke	-0.856	0.352	0.43 (0.21-0.85)	0.015

*β = -0.526; SE (β) = 0.203; goodness-of-fit χ2 = 0.016; df = 8; P = 0.016. Area under the ROC curve = 0.828; sensitivity 71.6%; specificity 80.4%; VPP 38.5%, VPN 94.3%, correct classification 79.1%.

†β = 2,937; SE (β) = 0.429; goodness-of-fit χ2 = 9.751; df = 6; P = 0.136. Area under the ROC curve = 0.803; sensitivity 73.3%; specificity 64.7%; VPP 90.4%, VPN 34.7%, correct classification 71.7%.