Table 4.

Different models of logistic regression to predict sepsis diagnosis on admission * (n = 74 patients)

	AUC (95% CI)	OR (95% CI)	P-value
Model 1 (M1)
ACTH ≤ 233 nM/L	0.875 (0.774 to 0.942)	9.192 (2.466 to 34.265)	0.001
Cortisol baseline ≥ 450 nM/L		4.279 (1.007 to 18.181)	0.049
Sepsis score >7		5.371 (1.567 to 18.404)	0.007
Model 2 (M2)
ACTH ≤ 233 nM/L	0.903 (0.808 to 0.960)	21.16 (4.252 to 105.307)	0.0002
Cortisol baseline ≥ 450 nM/L		8.81 (1.616 to 48.037)	0.012
Procalcitonin (PCT) >2 ng/mL		28.558 (4.393 to 185.636)	0.0004
Model 3 (M3)
ACTH ≤ 233 nM/L	0.805 (0.693 to 0.890)	9.125 (2.667 to 31.218)	0.0004
Cortisol baseline ≥ 450 nM/L		7.457 (1.89 to 29.426)	0.004
Model 4 (M4)
Procalcitonin (PCT) >2 ng/mL	0.726 (0.607 to 0.826)	16.889 (3.523 to 80.959)	0.0004
Model 5 (M5)
Sepsis score >7	0.727 (0.607 to 0.827)	6.481 (2.34 to 17.952)	0.0003

ACTH, adreno corticotropic hormone; AUC, area under the receiver operating characteristic curve; CI, confidence interval; OR: odds ratio; PCT, procalcitonin.