Nishioka et al reported models for predicting neurological outcome at 90 days after an out-of-hospital cardiac arrest. One model does not use laboratory values while another does. The authors are from multiple institutions in Japan
Patient selection: out-of-hospital cardiac arrest
Outcome: unfavorable (CPC >= 3) at 90 days
Parameters:
(1) age in years
(2) initial rhythm at scene
(3) bystander CPR
(4) shock during cardiac arrest
(5) administration of adrenaline during cardiac arrest
(6) advanced airway management during cardiac arrest
(7) first documented rhythm in hospital
(8) Glasgow coma motor score, from 1 to 6
(9) low flow time in minutes
(10) serum albumin in g/dL
(11) serum potassium in mmol/L
(12) blood glucose in mg/dL
(13) arterial pH
(14) pCO2 in mm Hg
(15) blood lactate in mmol/L
|
Parameter
|
Finding
|
Points
|
|
age in years
|
|
0.0419 * (age)
|
|
initial rhythm
|
VF or pVT
|
0
|
|
|
PEA
|
0.2733
|
|
|
asystole
|
0.3780
|
|
bystander CPR
|
no
|
0
|
|
|
yes
|
-0.1027
|
|
shock during arrest
|
no
|
0
|
|
|
yes
|
-1.1993
|
|
adrenaline given
|
no
|
0
|
|
|
yes
|
1.1486
|
|
airway management
|
no
|
0
|
|
|
yes
|
0.6124
|
|
first documented rhythm
|
VF or pVT
|
0
|
|
|
PEA (pulseless electrical activity)
|
1.0801
|
|
|
asystole
|
1.3109
|
|
|
ROSC
|
0
|
|
GCS motor score
|
1
|
0.8253
|
|
|
>= 2
|
0
|
|
low flow time
|
|
0.4319 * LN(time)
|
|
serum albumin
|
|
-0.832*(albumin)
|
|
serum potassium
|
|
0.0723*(potassium)
|
|
blood glucose
|
|
0.0005*(glucose)
|
|
arterial pH
|
|
-1.8962*(pH)
|
|
pCO2
|
|
0.0059 * (pCO2)
|
|
blood lactate
|
|
0.0241 * (lactate)
|
where:
• The type of log transformation of the low flow time was not given. It was assumed to be natural logarithm.
• VF or pVT are shockable rhythms.
value of X =
= SUM(points for all of the parameters) + 11.0535
probability of poor neurologic outcome =
= 1 / (1+EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.95.
