Acute kidney injury is an independent risk factor for pediatric intensive care unit mortality, longer length of stay and prolonged mechanical ventilation in critically ill children: a two-center retrospective cohort study

Table 2 Association of any AKI with PICU mortality using two different approaches for determination of bSCr levels^a

	Standard bSCr determination (AKIN_{standard bSCr})^b		Normative values for bSCr in all subjects (AKIN_{all norms bSCr})^b
Characteristics	Model 1^c	Model 2^c	Model 3^c	Model 4^c
AKI	6.5 (4.0 to 10.7)^d	3.7 (2.1 to 6.4)^d	8.4 (5.1 to 14.0)^d	4.5 (2.6 to 7.9)^d
Age under one year	1.6 (1.0 to 2.7)	1.0 (0.6 to 1.7)	1.8 (1.1 to 3.0)^d	1.0 (0.6 to 1.7)
Female	0.6 (0.4 to 1.1)	0.7 (0.4 to 1.5)	0.6 (0.4 to 1.1)	0.7 (0.4 to 1.2)
CHUSJ site	1.5 (0.9 to 2.5)	1.9 (1.1 to 3.3)^d	1.8 (1.1 to 3.0)^d	2.3 (1.3 to 4.0)^d
PRISM score, tertiles
0 to 3 reference group	-	1.0 (reference)	-	1.0 (reference)
4 to 6	-	0.5 (0.2 to 1.5)	-	0.5 (0.2 to 1.5)
> 6	-	2.6 (1.2 to 5.7)^d	-	2.6 (1.2 to 5.7)^d
Mechanically ventilated	-	16.6 (5.9 to 46.7)^d	-	16.6 (5.9 to 46.7)^d
Postoperative status	-	0.2 (0.1 to 0.6)^d	-	0.2 (0.1 to 0.6)^d
bSCr	-	1.5 (0.9 to 2.7)	-	1.7 (1.0 to 2.9)
Documented infection	-	2.3 (1.1 to 4.8)^d	-	2.2 (1.0 to 4.7)^d

^aAKI = acute kidney injury; AKIN = Acute Kidney Injury Network staging system; bSCr = baseline serum creatinine; CHUSJ = Centre Hospitalier Universitaire Ste-Justine; PICU = pediatric intensive care unit; PRISM = Pediatric Risk of Mortality. All results are presented as odds ratios (95% CI).
^bStandard bSCr method on left side of table (or AKIN_{standard bSCr} ) defines bSCr as the lowest level in the three months prior to admission or normative values based on SCr for age and gender when there were no prior SCr data available. The right side of the table lists normative SCr values for age and gender used to estimate bSCr values in all patients (AKIN_{all norms bSCr}) to determine the presence of AKI. ^cTwo different statistical models (or regression analyses) were performed (using stepwise backward multiple logistic regression analysis) to evaluate the independent effect of AKI on PICU mortality. In model 1, we evaluated the effect of AKI on PICU mortality, controlling only for age, gender and study site. In model 2, we evaluated the effect of AKI on PICU mortality, controlling for age, gender, study site, PRISM tertile, ventilation status, postoperative status, presence or absence of bSCr data and presence or absence of documented infection. ^dStatistically significant odds ratios (P < 0.05).

ISSN: 1364-8535