Prothrombin time is predictive of low plasma prothrombin concentration and clinical outcome in patients with trauma hemorrhage: analyses of prospective observational cohort studies

Table 2 Summary of ROC analyses

Database	Biomarker	Low prothrombin		Mortality 24 h		Massive transfusion
Database	Biomarker	AUC (95% CI) p-value		AUC (95% CI)	p-value	AUC (95% CI)	p-value
1 n = 358	PT	0.94 (0.88,0.96)	–	0.90 (0.82,0.97)	–	0.92 (0.89,0.98)	–
	EXTEM CT	0.68 (0.61,0.85)	<0.001	0.66 (0.48, 0.82)	<0.001	0.73 (0.55,0.80)	<0.001
	EXTEM MCF	0.78 (0.71,0.92)	0.003	0.81 (0.66,0.96)	0.04	0.81 (0.67,0.89)	0.04
2 n = 331	PT	Not applicable		0.78 (0.68,0.89)	–	0.81 (0.73,0.89)	–
	EXTEM CT			0.74 (0.62,0.86)	0.44	0.70 (0.61,0.78)	0.008
	EXTEM MCF			0.67 (0.54,0.81)	0.03	0.75 (0.66,0.84)	0.05

The area under the ROC (AUC), was calculated in separate logistic regression models for each of the binary response variables low prothrombin (<60 IU/dL), mortality 24 h and massive transfusion (PRBC 24 h >10 units) for each of the three biomarkers PT, EXTEM CT and EXTEM MCF. The AUC p-value refers to a two-sided test of the null hypothesis of no difference in AUC between PT and each of the two other biomarkers. In database 2 there were too few patients with admission prothrombin concentrations to perform a valid ROC analysis

ISSN: 1757-7241