Urine hepcidin has additive value in ruling out cardiopulmonary bypass-associated acute kidney injury: an observational cohort study

Table 5 Multivariate logistic regression analysis (all patients) of risk factors and biomarkers at six hours after CPB start for the prediction of no AKI^a

	Model 0 (R² = 0.42)		Model 1 (R² = 0.63)		Model 2 (R² = 0.43)		Model 3(R² = 0.50)
Risk factors and biomarkers	Regression coefficient (B) (SE)	P value	Regression coefficient (B) (SE)	P value	Regression coefficient (B) (SE)	P value	Regression coefficient (B) (SE)	P value
Age, years	-0.2 (0.1)	0.015	-0.2 (0.1)	0.06	-0.1 (0.1)	0.06	-0.2 (0.1)	0.036
LVEF <35%	-2.9 (1.1)	0.009	-4.1 (1.6)	0.008	-2.7 (1.1)	0.011	-3.0 (1.2)	0.01
COPD	-2.2 (1.0)	0.036	-2.4 (1.3)	0.06	-2.0 (1.0)	0.044	-2.5 (1.1)	0.026
Log urine hepcidin/urine creatinine, ng/mg			2.9 (1.1)	0.011
Log urine hepcidin, ng/mL					1.2 (0.7)	0.09
Log plasma hepcidin, ng/mL							2.7 (2.5)	0.56

^aAKI, acute kidney injury; CPB, cardiopulmonary bypass; LVEF, left ventricular ejection fraction; COPD, chronic obstructive pulmonary disease; SE, standard error. Variables tested on univariate relation with incidence of acute kidney injury (AKI) included all variables displayed in Table 1, type and duration of surgery and renal biomarkers at six hours after start of CPB. Multivariate logistic regression analysis included relevant variables for (no) AKI with univariate P value <0.1 (age, atrial fibrillation, LVEF <35%, COPD and peripheral vascular disease). Variables were presented in model 0 if their multivariate P value was <0.05 (age, LVEF <35% and COPD). We included each biomarker one after another (models 1 through 3) into multivariate regression analysis to exclude interaction.

ISSN: 1364-8535