Table 3.

Univariable and multivariable multinomial logistic regression for the relationship of thrombus composition with stroke etiology

	Univariable model	Multivariable model*
Base outcome: cardioembolic	OR (95% CI)	aOR (95% CI)
Non-cardioembolic
RBC	1.03 (1.01; 1.05)	1.02 (1.00; 1.04)
F/P	0.97 (0.95; 0.99)	0.98 (0.96; 1.00)
Leukocytes	0.91 (0.83; 1.01)	0.90 (0.81; 1.01)
Undetermined
RBC	1.00 (0.99; 1.01)	1.00 (0.98; 1.01)
F/P	1.00 (0.99; 1.01)	1.00 (0.99; 1.02)
Leukocytes	1.00 (0.94; 1.07)	1.00 (0.92; 1.06)

All analyses were done with thrombus components as a continuous variable, expressed as % of the thrombus. Odds ratios for stroke etiology are shown per percentage increase of thrombus components, with 95% confidence intervals. Interpretation: for every 1% increase in RBCs, OR for a non-cardioembolic stroke (compared with cardioembolic stroke) is 1.03 (95% CI 1.01–1.05). Non-cardioembolic indicates TOAST 1 (large artery sclerosis) + TOAST 4 (other determined cause)

RBC red blood cells, F/P fibrin/platelets, and leukocytes white blood cells

^*Adjusted for: age, sex, IV thrombolysis, coumarin/DOAC use, and thrombus location