Schneider et al used CART analysis to identify a burn patient at risk for late acute kidney injury (AKI). These can help to identify a patient who may benefit from more aggressive management. The authors are from Loyola University in Maywood, Illinois.
Patient selection: severe burn
Outcome: acute kidney injury (AKI)
Method: classification and regression tree (CART) analysis
Parameters:
(1) non-renal organ failure score (NROF)
(2) lowest 24-hour base deficit in mmol/L (within 24 hours of admission)
(3) lowest 24-hour glucose in mg/dL (within 24 hours of admission)
(4) early transfusion (during resuscitative phase)
NROF
|
base deficit
|
glucose
|
early transfusion
|
Percent Late AKI
|
0 or 1
|
NA
|
NA
|
NA
|
12%
|
2 or more
|
<= -11.41
|
NA
|
NA
|
75%
|
2 or more
|
> - 11.41
|
<= 83
|
NA
|
69%
|
2 or more
|
> - 11.41
|
> 83
|
no
|
25%
|
2 or more
|
> - 11.41
|
> 83
|
yes
|
56%
|
Performance:
• The accuracy was 80% for the derivation set and 73% for a validation set.
Risk factors for progressive AKI:
(1) worsening base deficit during the first 24 hours
(2) older with comorbidities
Risk factors for late AKI:
(1) worsening base deficit during the first 24 hours
(2) poor glucose control
(3) intubation on arrival and/or inhalation injury