Red cell transfusions as an independent risk for mortality in critically ill children

Table 5 Multiple regression analyses, with PICU LOS as the dependent variable (n = 4832)

Variables	β-coefficient	95 % confidence interval	p value
Dependent variable: ln(PICU LOS)^a
PIM2 score	0.138	0.118–0.159	<0.001
Transfusion	0.833	0.731–0.935	<0.001
Vasoactive agents	0.377	0.302–0.452	<0.001
Day 1 PELOD	0.013	0.008–0.018	<0.001
PRISM III score	−0.011	−0.018 to −0.003	0.017
Reason for PICU admission^b
Cardiovascular	−0.308	−0.447 to −0.170	<0.001
Neurologic	−0.239	−0.322 to −0.156	<0.001
Trauma	−0.441	−0.538 to −0.343	<0.001
Other^c	−0.262	−0.330 to −0.195	<0.001
Gender	−0.027	−0.080 to −0.026	0.324
Age at admission (months)	0.000	0.000 to −0.000	0.716

^aDue to the non-normality of the variable PICU LOS, the variable was transformed by taking its natural logarithm prior to analysis; the β-coefficients are interpreted by calculating the antilog, which for transfusion is e^0.833, meaning that transfused subjects had a 2.3 day longer PICU LOS than non-transfused subjects
^bReason for PICU admission, odds ratio relative to respiratory (reference group)
^cOther includes all other categories (e.g., endocrinologic, gastrointestinal, infectious, oncologic)

ISSN: 2052-0492