Table 5.

Results for the univariable and multivariable analysis of statistically significant predictors of experimental virulence for the 84 Escherichia coli ST131 studied isolates.

	Illness severity score (ISS)						“Killer” status
	Univariable		Multivariablea (forced-entry method)		Multivariable (stepwise method)		Univariable		Multivariableb (forced-entry method)		Multivariable(stepwise method)
Predictorc	β coef.	P valued	β coef.	P valued	β coef.	P valued	OR (CI)	P valued	OR (CI)	P valued	OR (CI)	P valued
pap	0.68	0.03	0.16	0.67	0.05	0.68	2.9 (1.0–8.5)	0.049	1.4 (0.3–6.9)	0.65	2.9 (1.0–8.5)	0.049
kpsMII	0.98	0.003	−0.32	0.52	−0.10	0.50	4.3 (1.2–14.7)	0.02	0.5 (0.05–6.1)	0.62	4.3 (1.2–14.7)	0.02
K2	0.82	0.003	0.47	0.15	0.13	0.24	4.7 (1.6–13.5)	0.004	3.7 (0.9–15.5)	0.07	3.4 (1.1–10.0)	0.03
VG score	0.18	0.01	−0.10	0.31	−0.08	0.54	1.2 (1.0–1.5)	0.07	0.7 (0.5–1.1)	0.15	1.2 (1.0–1.5)	0.07
ExPEC	1.33	<0.001	1.72	0.001	1.22	<0.001	8.2 (2.2–31.2)	0.002	23.2 (1.9–278.7)	0.01	5.98 (1.5–23.5)	0.01
H30R1	−0.70	0.02	0.25	0.48	0.08	0.50	0.37 (0.1–1.0)	0.047	1.5 (0.4–6.6)	0.56	0.37 (0.1–1.0)	0.05
Year (2014)e	0.75	0.006	0.47	0.09	0.57	0.02	3.1 (1.2–7.9)	0.02	2.0 (0.6–6.3)	0.24	3.1 (1.2–7.9)	0.02

β coef, Beta coefficient. OR (CI), odds ratio (95% confidence interval). ExPEC: Extraintestinal pathogenic E. coli.

^aMultivariable linear regression: for final model with all variables included, r² = 0.29 (p < 0.001).

^bMultivariable logistic regression: for final model with all variables included, r² = 0.23 (p = 0.002).

^cVariables shown are the significant univariable predictors of one or both virulence outcomes.

^dStatistically significant P values are in bold.

^eYear of testing correlated roughly with country of origin (rho = 0.84, p < 0.001), so can be considered a surrogate for that trait.