Xie et al reported a model for predicting 28-day mortality for a patient with sepsis. This can help to identify a patient who may benefit from more aggressive management. The authors are from Jinan University and South University of Science and Technology in Shenzhen, China.
Patient selection: septic adult
Parameters:
(1) serum procalcitonin on day 1, from 0 to 200 ng/mL
(2) serum procalcitonin on day 3, from 0 to 200 ng/mL
(3) serum IL-6 on day 1, from 0 to 4,500 pg/mL
(4) serum IL-6 on day 3, from 0 to 5,000 pg/mL
(5) serum lactate clearance on day 3, from 100 to -300 percent
clearance =
= ((initial value) - (value on day 3)) / (initial value) * 100%
points for procalcitonin on day 1 =
= (-0.1315 * (procalcitonin on day 1)) + 26.3
points for procalcitonin in day 3 =
= (0.063 * (procalcitonin on day 3))
points for IL-6 on day 1 =
= (0.001133 * (IL-6 on day 1))
points for IL-6 on day 3 =
= (0.02 * (IL-6 on day 3))
points for lactate clearance on day 3 =
= (-0.02125 * (lactate clearance)) + 2.125
total score =
= SUM(points for all of the parameters)
Interpretation:
• minimum score: 0
• maximum score: 152.4
value of X =
= (0.335 * (score)) - 11.46
probability of 28-day mortality =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.85 for the derivation and 0.83 for the validation cohorts.
