Table 4.

Multivariate analysis by Cox proportional hazards regression model and diagnostic accuracy of 180-day mortality

Markers	180-day mortality
	HRa	95 % CI	p	Sensitivity	95 % CI	Specificity	95 % CI	ROC-AUC	95 % CI	p
PTX3	1.05	0.47–2.35	0.902	0.69	0.53–0.82	0.56	0.50–0.60	0.66	0.56–0.76	0.005
PTX3b	1.34	0.60–3.00	0.480	0.67	0.51–0.80	0.68	0.62–0.72	.	.	.
PCT	0.86	0.38–1.95	0.713	0.66	0.50–0.79	0.55	0.49–0.60	0.65	0.55–0.76	0.007
CRP	0.71	0.34–1.51	0.376	0.34	0.21–0.50	0.41	0.36–0.47	0.32	0.21–0.43	0.001
PTX3 + PCT	.	.	.	.	.	.	.	0.66	0.56–0.76	0.005
PTX3 + CRP	.	.	.	.	.	.	.	0.72	0.61–0.82	<0.0001
PCT + CRP	.	.	.	.	.	.	.	0.71	0.60–0.82	<0.0001
PTX3 + PCT + CRP	.	.	.	.	.	.	.	0.72	0.62–0.83	<0.0001

Hazard ratio, sensitivity and specificity are calculated for high (above median) versus low (below median) baseline levels of the inflammatory biomarkers. Five patients are not included in the analysis due to missing data regarding Simplified Acute Physiology Score II

^aAdjusted for age, sex, Simplified Acute Physiology Score II and chronic disease (yes/no)

^bPTX3 dichotomized by the optimal cutoff (69.8 ng/ml) found by the ROC curve (highest sum of sensitivity and specificity)

CRP C-reactive protein, HR Hazard ratio, PCT Procalcitonin, PTX3 Pentraxin-3, ROC-AUC Receiver operating characteristic-area under the curve, CI Confidence interval

Dots: Values cannot be given