Table 4.

Multivariable logistic regression analysis with backward elimination of factors independently associated with a poor 90-day (3–6)

Independent variable	OR (95% CI)	P-value
Model 1*,†
WMH lesion burden	1.80 (1.18–2.76)	0.007
Non-SVD SSI mechanism	6.51 (1.52–27.94)	0.012
Model 2*,†,‡
WMH lesion burden	2.28 (1.04–4.99)	0.040
Pre-admission mRS score, per point	8.96 (2.65–30.27)	<0.001
Infarct volume, per mL	1.98 (1.14–3.44)	0.016
Age, per year	0.91 (0.82–1.01)	0.063
Non-SVD SSI mechanism	9.20 (1.04–81.39)	0.046

^*Hosmer-Lemeshow for model 1: χ²=7.758, p=0.457. Hosmer-Lemeshow for model 2: χ²=3.059, p=0.931.

^*Entering the infarct volume as mean or total infarct volume did not meaningfully change the results, respectively.

^†The interaction term of white matter hyperintensity (WMH) lesion burden x infarct volume and presence of multiple SSI were not retained. Finally, entering large artery atherosclerosis, cardioaortic embolism, and small arterial occlusion in lieu of the non-small vessel disease (SVD) small subcortical infarct (SSI) mechanism did not meaningfully change the models whereby presence of SVD-related SSI etiology was associated with a favorable outcome (p<0.05 in each model).

^‡The National Institutes of Health Stroke Scale score was not retained in model 2. Further, restricting the analysis to patients with a modified Rankin Scale (mRS) of <6, did not meaningfully change the results.