Table 2.

Predictors of vertigo in the binary logistic regression models

Characteristic	Univariate regression analysis			Multivariate regression analysis
	OR	95% CI	p	OR	95% CI	p
Age in years	0.99	96–1.03	.72	0.96	0.9–1.02	.17
Women	3.09	0.99–9.68	.053	3.8	0.79–18.47	.1
Active smoking	0.81	0.23–2.87	.74
Hypertension	0.99	0.26–3.67	.98
Diabetes	2.32	0.41–13.03	.34
Atrial fibrillation	1.37	0.42–4.5	.61
Previous stroke	0.41	0.11–1.58	.2
HbA1c (%)	1	0.55–1.8	1
LDL‐cholesterol (mg/dl)	1.01	0.99–1.02	.51
White blood cells*1,000/µl	1.08	0.9–1.29	.41
CRP (mg/dl)	0.71	0.44–1.16	.17
ESR in the first hour (mm)	0.99	0.96–1.02	.62
Total infarction volume of >0.48 cm3	6.75	1.99–22.85	.002*	4.4	1.05–18.58	.043*
Infarction affecting the cerebellum or dorsal brainstem	14.06	3.42–57.88	<.001*	16.97	3.1–92.95	.001*
H‐L testa						0.88

Abbreviations: CRP, C‐reactive protein; ESR, erythrocyte sedimentation rate; HbA1c, hemoglobin A1c; LDL‐cholesterol, low‐density lipoprotein cholesterol.

^aHosmer–Lemeshow “goodness‐of‐fit” test for the multivariate regression analysis showed a nonsignificant p‐value (p = .88) for the difference between our observed results and the expected results. The nonsignificant p‐value for this test means better fit of the model (the higher the value, the better the fit).

^*Statistically significant.